European Journal of Preventive Cardiology

Grobbee, Diederick E; Allpress, Rosemary

doi: 10.1177/2047487316642383pmid: 27048543

Recently, a discussion took place among the members of the Editorial Board of the European Journal of Preventive Cardiology about the appropriateness of limiting the number of authors for papers accepted by the Journal. The background for the discussion was the fact that an increasing number of papers submitted to the Journal have an excessive number of authors, and there is mounting concern about whether this trend can be justified. The International Committee of Medical Journal Editors (ICMJE) has clearly spelled out what qualifies for authorship and recommends that authorship be based on the following four criteria: Substantial contributions to the conception or design of the work; or the acquisition, analysis or interpretation of data for the work; AND Drafting the work or revising it critically for important intellectual content; AND Final approval of the version to be published; AND Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.1 While these all-inclusive criteria are clear and specific, the implementation of a policy to assure credible authorship is challenging. When more than, say, 10 authors are listed for a single paper it could be questioned what the real contribution of some of these authors has been, whether they can take responsibility for the full content and, in fact, whether they are fully aware of the contents and implications of their work. Apart from the criteria mentioned above, there are several other reasons why those with final responsibility for the submission of a paper choose to invite or accept authors. Perhaps one of the most common reasons for researchers to offer authorship is the role these putative authors have played in collecting data necessary for the research, typically in an international context. It is easy to understand why this happens as authorship is often the ‘carrot’ held out to entice centres to participate, since it confers credit and has important academic, social, and financial implications. For many clinicians and others involved in data collection in patient care, e.g. investigators providing subsets or biochemical or genetic data or contributing populations of patients to collaborative analyses, there are limited means to acquire authorships other than through their recognition in papers from larger consortia. In a world where scientific reputations are, for better or worse, still largely driven by publication records, it can be understood that these investigators make their contributions conditional on future authorships. Vice-versa, lead investigators will face major problems in getting the data should they not be able to offer these. A paper published in Nature on Drosophila genomics was published with more than 1000 authors.2 Justification was in part found in the handwork done by 900 graduates to edit highly repetitive DNA sequences. It should be acknowledged that, compared to the number of fruit flies needed for the research, the number of authors was likely quite modest! The above mechanisms are common, difficult to detect or prevent but non-compliant with the ICMJE criteria. Yet, even within the bounds of the legitimate criteria for authorship, papers can still sometimes be ‘overcrowded’ with authors. Research has become increasingly multidisciplinary and international, and combinations of expertise and competences tend to drive the quality of research results. The involvement of multiple centres and groups leads to more authors qualifying for inclusion. Not only authorship per se, but also the ranking of authors is an issue. Traditionally, the author that writes the paper heads the list, while the last author tends to be the one who takes overall responsibility for the work. Therefore, in many academic assessments the first and last authorships are the ones that bear most weight. Consequently, we now regularly encounter shared first authorships, shared last authorships or other creative combinations. Indeed, it may be hard to define exactly which of two or more, commonly junior, authors made the largest contribution to the actual writing of a paper. Given these complexities, one may question whether the current system of a simple listing of individual authors is capable of providing appropriate credits to the different roles and responsibilities that researchers may have had in a project leading to scientific publication. Guideline reports are notorious for their long listings of individual authors. They are also articles of major interest to readers and therefore attract a high number of citations. So, the role of authorship here is a delicate issue. Currently, authorship of guidelines is a blend of the actual work carried out, the ‘buy-in’ of several stakeholders (i.e. interested parties whose agreement to support them is fundamental, and who have often been involved in their formulation), and the representativeness needed to maximize visibility (i.e. to assure sufficient exposure to the scientific community). For many of the authors, the typical involvement may be limited to reviewing subsequent versions of the report. A problem is that for many busy opinion leaders in medicine, these papers are fast tracks to scientific visibility and citations and certainly ‘lower-hanging scientific fruit’ than laborious and time-consuming original research. Guidelines are important, but too many guidelines, position papers, consensus reports and similar learned views may not benefit patients and society and confuse rather than confirm. Honouring these scientific products with across-the-board authorships promotes rather than prevents such practices and needs to be carefully reappraised. There are good reasons, in fact, to publish guidelines under the single affiliation of a scientific or professional association or other prestigious and competent institution although, at present, this is uncommon. In our view, journals should collectively decide to have guidelines published on behalf of associations rather than as the work of individual authors. If needed, participants in the preparation of the guidelines and their roles can be listed separately or even in supplementary on-line material. An available solution to long lists of authors heading a paper is that the paper be written by a writing committee (whose individuals are the listed authors) on behalf of a study group, or even simply ‘on behalf of’ a study group (the study group itself being the sole author), listing the multiple individual contributors at the end of the paper. This is an attractive alternative but it also has a limitation. While all authors listed in this manner at the end of the paper will find the paper included in their personal publication overview (e.g. in PubMed), individual citations are not counted in Web of Science or other bibliometric systems. This will reduce the attraction of the option to some. Yet it seems like a fair solution in those circumstances where contributions do not fully comply with the criteria for authorship but where there is, nevertheless, a justified need to give credit to work done. For example, this is the case for guideline papers. Rules on authorship and, more importantly, authorship practices cannot be changed by a single journal in isolation. Should the European Journal of Preventive Cardiology single-mindedly decide to restrict the number of authors, as we have entertained, this may put the Journal at a competitive disadvantage when other journals are prepared to consider papers with an unrestricted number of authors. As this would particularly affect papers that may be well cited, such as guideline reports, international consortia or multi-institution collaborations, and since journals like authors depend on the extent to which their published work is cited, this would harm the position of our Journal. Something, however, needs to be done. First and foremost, the discussion on authorship should be encouraged among the scientific community and a reappraisal is needed for a system that currently is difficult to comprehend, easy to corrupt and has ceased to provide sufficient credibility to the diversity of roles that individuals and groups play in science. In the meantime, we in the European Journal of Preventive Cardiology have decided not to restrict the number of authors on submissions to the Journal, but to perform a more rigorous assessment of the contribution that each author has made. A written assessment has been introduced as part of the submission process and requires that every author explicitly state their role and contributions to the manuscript. Hopefully, in the future, all scientific journals will move to adopt a common policy on authorship, that will clarify for all concerned an issue that at present remains nebulous. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article. Funding The author(s) received no financial support for the research, authorship and/or publication of this article. References 1 International Committee of Medical Journal Editors. Defining the role of authors and contributors, http://www.icmje.org/recommendations/browse/roles-and-responsibilities/defining-the-role-of-authors-and-contributors.html (2016, accessed 2 March 2016) . 2 Woolston C. Fruit-fly paper has 1,000 authors, http://www.nature.com/news/fruit-fly-paper-has-1-000-authors-1.17555 (2015, accessed 2 March 2016) . © The European Society of Cardiology 2016 This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © The European Society of Cardiology 2016

Wang, Yixuan; Peng, Xiaoxia; Nie, Xiaolu; Chen, Li; Weldon, Ryan; Zhang, Weili; Xiao, Dan; Cai, Jun

doi: 10.1177/2047487315617105pmid: 26603746

Abstract Aims To review comprehensively the prevalence, treatment and control of hypertension; and to estimate the burden of hypertension in China, thereby aiding Chinese health policies for better prevention and control of this condition. Methods and results PubMed, EMbase, China National Knowledge Infrastructure, Wanfang and Chongqing VIP databases were searched for population-based studies published in English and Chinese that described prevalence, treatment and control of hypertension in China, as well as deaths and disabilities attributed to hypertension. All research papers were published between January 1999 and May 2014. Data from 178 studies involving over 2,901,464 participants covering 30 provinces were pooled. Overall, rates of prevalence, treatment and control of hypertension were 28.9%, 35.3% and 13.4% in China. A statistically significant association was observed between temperature gradient and the prevalence of hypertension. There were 10,667 (95% confidence interval 8063–13,345) disability-adjusted life years per 100,000 people. In total, 78.3% of disability-adjusted life years were from years lived with disability and 21.7% from years of life lost due to premature mortality. Conclusions Although there has been a slight improvement in rates for the treatment and control of hypertension, these rates were still suboptimal, especially for men and people living in rural areas. Low and middle-income provinces had a comparatively huge burden of hypertension, which is a considerable risk factor for reducing life expectancy. Our analysis may be helpful in generating a current overview of hypertension in China. Hypertension, prevalence, treatment rate, control rate, burden of hypertension Introduction Hypertension is the leading risk factor for global cardiovascular disease burden and mortality. The burden of hypertension worldwide has risen by almost 30% from 1990, and hypertension was responsible for approximately 9.4 million deaths and 7% of all disability-adjusted life years (DALYs) in 2010.1,2 At least 45% of heart disease deaths and 51% of stroke deaths are due to hypertension.3 Similarly, in China, the burden of hypertension has increased rapidly over the past few decades. The age-standardised rate of hypertension was 27,700 per 100,000 in 2000.4 Hypertensive heart disease led to an age-standardised death rate of 12.8 deaths per 100,000 and an age-standardised DALY rate of 194.7 per 100,000 in 2010.5 Large-scale national surveys of hypertension prevalence were conducted in 1958, 1979–1980, 1991 and 2000 in China, with reported hypertension prevalence rates of 5–11%, 7.7%, 13.6% and 24–27%, respectively. However, it should be noted that China used a diagnostic blood pressure threshold of 160/95 mmHg for hypertension and 140/90 mmHg for normal blood pressure in these studies. Therefore, the figures for hypertension prevalence from this era are misleading by current standards. In the 2005 edition of the Chinese guidelines on prevention and control of hypertension, the diagnostic value for hypertension was 140/90 mmHg and that for normal blood pressure was 120/80 mmHg. In addition, since 2005, the Chinese government has initiated a series of measures to manage hypertension, including: the establishment of a national chronic disease control network; development of medium and long-term national plans for chronic disease control; and substantial increases in hypertension detection and treatment.6 Although several population-based epidemiological surveys have been conducted in China recently,7,8 the characteristics of hypertension, including rates of prevalence, treatment and control, have varied among individual studies. For example, a survey that covered 13 provinces including urban and rural areas reported that the prevalence, awareness, treatment and control of hypertension were 29.6%, 42.6%, 34.1% and 9.3% among people aged 18 years or older, respectively.7 Another study that surveyed 46,239 people aged 20 years or older reported that the prevalence, awareness, treatment and control of hypertension were 26.6%, 45.0%, 36.2% and 11.1%, respectively, in 2007–2008.8 Data from the 2002 China National Nutrition and Health Survey covering 31 provinces estimated that the prevalence, awareness, treatment and control of hypertension were 18.8%, 30.2%, 24.7% and 6.1%, respectively.9 The International Collaborative Study of Cardiovascular Disease in ASIA (InterASIA) sampling of 15,540 people aged 35–74 years found that these factors were 27.2%, 44.7%, 28.2% and 8.1%, respectively, in 2000–2001.10 These findings indicate that China and developed countries11 have similar levels of hypertension prevalence, but awareness, treatment and control of hypertension in China is lower than that in developed countries. Therefore, it is imperative to assess comprehensively the burden of disease attributable to hypertension. Although two recent systematic reviews analysed the prevalence of hypertension, the searches were either limited to papers published in Chinese journals only,12 or were unable to provide an overall estimate of hypertension prevalence in China in recent years.13 In addition,our analysis indicates regional disparity in the burden of disease attributable to hypertension, which is a public health issue worth attention and further characterisation. These logistical data have the potential to inform policy makers on how to allocate medical care and primary education efforts efficiently to combat this growing public health challenge. Methods Literature search A standardised review protocol was designed, with detailed descriptions of the search strategy for data sources, eligible criteria, data-extracting methods and assessment of study quality. We searched for all population-based studies published between January 1999 and May 2014 in English and Chinese that described prevalence, awareness, treatment and control of hypertension in China, as well as the burden of disease attributable to hypertension. Hypertension was defined as average systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, and/or self-reported current treatment for hypertension with antihypertensive medication. Treatment of hypertension was defined as the use of prescription medication for management of high blood pressure. Control of hypertension was defined as treatment with an antihypertensive medication associated with average blood pressure <140/90 mmHg. In this systematic review, we performed English literature searches of the MEDLINE (Pubmed) and EMbase databases. A search was also made in the China National Knowledge Infrastructure digital database and Wanfang and Chongqing VIP databases for papers published in Chinese. The search strategy can be found in Supplementary 1. Eligibility criteria for inclusion Studies were included if the following criteria were met: 1. It was a population-based survey using random sampling design; 2. Methods for measurement of blood pressure were described; 3. Hypertension was defined as average systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg and/or self-reported current treatment for hypertension with antihypertensive medication; 4. At least sex-specific and age-specific prevalences of hypertension were reported. Criteria for exclusion Studies were excluded under the following terms: 1. Participants had pregnancy-induced hypertension, chronic kidney disease, or any other type of secondary hypertension; 2. Sample size was <2000, to ensure sufficient sample size after sex and age stratification; 3. Methods for sampling design were not described or were inadequately described; 4. Population-based survey used convenience sampling, such as participants being sampled from the study institution. Study selection Studies were selected based on the inclusion criteria. Screening of literature proceeded at two levels (Supplementary 2 contains details of study selection). When there was more than one record of the same study, we included all potentially eligible records according to the inclusion criteria, but used the most relevant one as the main record. Data extraction The study authors independently collected data using a pre-designed form. Information about the study design, location, study period, authors, publication year, methodology (diagnostic tools and other surveys, blood pressure measurement techniques and diagnostic standards) and population characteristics (sample size, case size, minimum age, maximum age, mean age and sex distribution) were collected in a standardised form. Assessment of risk of bias Two reviewers independently assessed the quality of the selected studies using quality assessment criteria for cross-sectional studies adapted from Agency for Healthcare Research and Quality recommendations.14 There were seven criteria in the checklist (Supplementary 3). Disagreements were resolved by discussion between the two authors and an epidemiological expert (Supplementary Figure 1). Statistical analysis Study-specific data on percentages are presented as forest plots with exact binomial 95% confidence intervals (CIs). These percentages were pooled using the random-effects method (REM) by provinces. Fixed-effects modelling was not applied because it presumes that all studies are functionally equivalent, that is, all factors (such as population and age) that could influence hypertension prevalence are basically identical in all of the studies and, consequently, the difference in prevalence across the studies occurs only due to sampling error. The age standardisation in provinces was conducted according to the Sixth National Population Census of the People’s Republic of China. The heterogeneity between studies was quantified by the I2 statistic. Furthermore, a multivariable meta-regression was performed to determine if hypertension data are significantly affected by altitude gradient, economic level and temperature gradient. Studies were divided into four groups for altitude gradient and six groups for temperature gradient.15 For economic level, studies were divided into five groups according to the Chinese government.16 Statistical analyses were performed with Stata version 12.0. The geographical distribution on prevalence, treatment rate and control rate of hypertension in provinces was performed with ESRI Arc GIS version 10.1. To determine summary measures of population health, DALYs were calculated. DALYs are the sum of years of life lost due to premature mortality (YLLs) and years lived with disability (YLDs).17 YLLs are computed by multiplying the number of deaths in each age group by a reference life expectancy at that age. YLDs are calculated from the prevalence of a sequela multiplied by the disability weight for that sequela. Disability weights are based on surveys of the general population. R version 3.0.3 was used to quantify the burden of disease.18,19 Results Selection of studies In total, 3750 studies were identified: 96 from PubMed, 142 from EMbase, 1310 from the China National Knowledge Infrastructure, 1011 from Chongqing VIP database, and 1191 from Wanfang data. There were 1383 studies after duplicates were removed, and 699 were excluded because they were found to be irrelevant to the topic after screening by title and abstract. We excluded 506 studies, because they met our criteria for exclusion. Finally, 178 studies involving 2,901,464 participants and covering 30 provinces were eligible after undergoing two level screening (Supplementary Figure 2). Their geographical distribution20 is presented in Figure 1, and the main characteristics of the studies are summarised in Supplementary Table 1. Figure 1. Open in new tabDownload slide Geographical distribution of prevalence of hypertension in Chinese provinces. Prevalence of hypertension The overall prevalence of hypertension in China was 28.9% (95% CI 27.5–30.4%) pooled with the REM. The results of the included studies showed that the prevalence of hypertension varied greatly, ranging from 15.0% in Shaanxi to 42.0% in Liaoning. Prevalence was higher among provinces in the north than those in the south (Figure 1). Heterogeneity was statistically significant (I2 = 99.8%). To estimate the reliability of the overall prevalence based on the epidemiology studies from the provinces, six national cross-sectional studies were pooled,7,8 and the pooled prevalence was 26.8% (95% CI 21.1–32.3%), which was similar to the overall prevalence finding.21,22 Overall, the prevalence of hypertension was higher among men (30.5%, 95% CI 29.0–31.9%) compared to women (27.7%, 95% CI 25.9–29.4%), and the results differed by province, as shown in Table 1. Prevalence was higher in rural (30.8%, 95% CI 28.5–33.0%) than in urban (26.9%, 95% CI 23.8–29.9%) areas. In Figure 2, it can be seen that the prevalence of hypertension, stratified by age and gender, increased with age. Based on 2010 population census of China, there are an estimated 206 million adults aged over 20 years, and 194 million adults aged over 30 years who have hypertension. Figure 2. Open in new tabDownload slide The prevalence of hypertension, stratified by age and gender. Table 1. The prevalence of hypertension, by province. Province . Total prevalence (%) . Rank . Male prevalence (%) . Female prevalence (%) . Anhui 32.5 (25.7–39.6) 9 35.29 (30.5–40.1) 29.7 (21.5–38.5) Beijing 27.2 (23.2–32.4) 17 28.2 (23.9–32.7) 26.2 (21.8–30.8) Chongqing 22.6 (16.2–29.8) 23 23.2 (15.8–31.6) 22.1 (16.4–28.3) Fujiana 18.7 24 22.6 15.2 Gansu 30.2 (22.7–38.4) 12 32.8 (24.3–41.8) 28.3 (20.8–36.5) Guangdong 20.4 (17.9–23.0) 27 22.5 (19.3–25.9) 18.3 (15.6–21.2) Guangxi 23.0 (15.9–30.9) 22 25.5 (18.6–33.1) 20.7 (13.2–29.2) Guizhou 23.1 (19.2–27.3) 21 26.6 (19.6–34.3) 20.2 (18.1–22.4) Hebei 36.8 (32.9–40.9) 5 38.3 (33.6–43.1) 35.8 (31.7–40.0) Heilongjiang 28.2 (18.6–39.1) 15 29.6 (20.8–39.3) 26.9 (16.2–39.3) Henan 33.5 (29.2–37.8) 6 32.5 (28.0–37.1) 34.0 (29.3–38.9) Hubei 19.8 (12.5–27.1) 28 21.0 18.6 (8.0–29.2) Hunan 19.3 (11.5–28.7) 29 19.1 (10.5–29.6) 19.6 (12.3–28.2) Jiangsu 28.6 (25.2–32.2) 14 30.7 (26.9–34.8) 27.0 (23.7–30.4) Jiangxi 21.1 (9.5–32.7) 25 21.7 (10.9–32.4) 20.5 (7.9–33.0) Jilina 32.8 8 43.8 28.4 Liaoning 42.0 (36.3–47.8) 1 41.9 (35.8–48.1) 43.0 (36.8–49.2) Inner Mongolia 38.2 (34.1–42.3) 4 45.3 (38.8–51.9) 32.6 (28.9–36.5) Ningxia 27.0 (25.8–28.1) 18 26.5 (22.7–30.4) 27.5 (25.3–29.7) Qinghai 29.7 (12.3–47.1) 13 32.4 (12.5–52.3) 26.2 (13.4–39.0) Shaanxia 15.0 30 16.3 13.7 Shandong 38.2 (30.9–45.8) 3 37.7 (32.5–43.1) 38.4 (29.5–47.9) Shanghai 31.5 (27.0–36.2) 11 32.7 (27.4–38.3) 30.5 (26.5–34.5) Shanxi 24.7 (18.5–31.5) 19 27.3 (21.0–34.1) 22.5 (15.8–30.1) Sichuan 20.9 (16.7–25.3) 26 22.6 (19.4–26.0) 19.8 (14.6–25.5) Tianjina 32.4 10 32.2 32.5 Tibeta 40.7 2 41.3 40.0 Xinjiang 33.4 (31.5–35.2) 7 36.1 (34.6–37.6) 31.3 (27.9–34.8) Yunnan 23.3 (13.3–35.2) 20 26.1 (17.3–36.0) 21.0 (9.8–35.0) Zhejiang 27.9 (23.0–33.1) 16 28.9 (23.5–34.6) 27.1 (22.5–31.9) Province . Total prevalence (%) . Rank . Male prevalence (%) . Female prevalence (%) . Anhui 32.5 (25.7–39.6) 9 35.29 (30.5–40.1) 29.7 (21.5–38.5) Beijing 27.2 (23.2–32.4) 17 28.2 (23.9–32.7) 26.2 (21.8–30.8) Chongqing 22.6 (16.2–29.8) 23 23.2 (15.8–31.6) 22.1 (16.4–28.3) Fujiana 18.7 24 22.6 15.2 Gansu 30.2 (22.7–38.4) 12 32.8 (24.3–41.8) 28.3 (20.8–36.5) Guangdong 20.4 (17.9–23.0) 27 22.5 (19.3–25.9) 18.3 (15.6–21.2) Guangxi 23.0 (15.9–30.9) 22 25.5 (18.6–33.1) 20.7 (13.2–29.2) Guizhou 23.1 (19.2–27.3) 21 26.6 (19.6–34.3) 20.2 (18.1–22.4) Hebei 36.8 (32.9–40.9) 5 38.3 (33.6–43.1) 35.8 (31.7–40.0) Heilongjiang 28.2 (18.6–39.1) 15 29.6 (20.8–39.3) 26.9 (16.2–39.3) Henan 33.5 (29.2–37.8) 6 32.5 (28.0–37.1) 34.0 (29.3–38.9) Hubei 19.8 (12.5–27.1) 28 21.0 18.6 (8.0–29.2) Hunan 19.3 (11.5–28.7) 29 19.1 (10.5–29.6) 19.6 (12.3–28.2) Jiangsu 28.6 (25.2–32.2) 14 30.7 (26.9–34.8) 27.0 (23.7–30.4) Jiangxi 21.1 (9.5–32.7) 25 21.7 (10.9–32.4) 20.5 (7.9–33.0) Jilina 32.8 8 43.8 28.4 Liaoning 42.0 (36.3–47.8) 1 41.9 (35.8–48.1) 43.0 (36.8–49.2) Inner Mongolia 38.2 (34.1–42.3) 4 45.3 (38.8–51.9) 32.6 (28.9–36.5) Ningxia 27.0 (25.8–28.1) 18 26.5 (22.7–30.4) 27.5 (25.3–29.7) Qinghai 29.7 (12.3–47.1) 13 32.4 (12.5–52.3) 26.2 (13.4–39.0) Shaanxia 15.0 30 16.3 13.7 Shandong 38.2 (30.9–45.8) 3 37.7 (32.5–43.1) 38.4 (29.5–47.9) Shanghai 31.5 (27.0–36.2) 11 32.7 (27.4–38.3) 30.5 (26.5–34.5) Shanxi 24.7 (18.5–31.5) 19 27.3 (21.0–34.1) 22.5 (15.8–30.1) Sichuan 20.9 (16.7–25.3) 26 22.6 (19.4–26.0) 19.8 (14.6–25.5) Tianjina 32.4 10 32.2 32.5 Tibeta 40.7 2 41.3 40.0 Xinjiang 33.4 (31.5–35.2) 7 36.1 (34.6–37.6) 31.3 (27.9–34.8) Yunnan 23.3 (13.3–35.2) 20 26.1 (17.3–36.0) 21.0 (9.8–35.0) Zhejiang 27.9 (23.0–33.1) 16 28.9 (23.5–34.6) 27.1 (22.5–31.9) a Only one study was eligible after undergoing two level study screening. Open in new tab Table 1. The prevalence of hypertension, by province. Province . Total prevalence (%) . Rank . Male prevalence (%) . Female prevalence (%) . Anhui 32.5 (25.7–39.6) 9 35.29 (30.5–40.1) 29.7 (21.5–38.5) Beijing 27.2 (23.2–32.4) 17 28.2 (23.9–32.7) 26.2 (21.8–30.8) Chongqing 22.6 (16.2–29.8) 23 23.2 (15.8–31.6) 22.1 (16.4–28.3) Fujiana 18.7 24 22.6 15.2 Gansu 30.2 (22.7–38.4) 12 32.8 (24.3–41.8) 28.3 (20.8–36.5) Guangdong 20.4 (17.9–23.0) 27 22.5 (19.3–25.9) 18.3 (15.6–21.2) Guangxi 23.0 (15.9–30.9) 22 25.5 (18.6–33.1) 20.7 (13.2–29.2) Guizhou 23.1 (19.2–27.3) 21 26.6 (19.6–34.3) 20.2 (18.1–22.4) Hebei 36.8 (32.9–40.9) 5 38.3 (33.6–43.1) 35.8 (31.7–40.0) Heilongjiang 28.2 (18.6–39.1) 15 29.6 (20.8–39.3) 26.9 (16.2–39.3) Henan 33.5 (29.2–37.8) 6 32.5 (28.0–37.1) 34.0 (29.3–38.9) Hubei 19.8 (12.5–27.1) 28 21.0 18.6 (8.0–29.2) Hunan 19.3 (11.5–28.7) 29 19.1 (10.5–29.6) 19.6 (12.3–28.2) Jiangsu 28.6 (25.2–32.2) 14 30.7 (26.9–34.8) 27.0 (23.7–30.4) Jiangxi 21.1 (9.5–32.7) 25 21.7 (10.9–32.4) 20.5 (7.9–33.0) Jilina 32.8 8 43.8 28.4 Liaoning 42.0 (36.3–47.8) 1 41.9 (35.8–48.1) 43.0 (36.8–49.2) Inner Mongolia 38.2 (34.1–42.3) 4 45.3 (38.8–51.9) 32.6 (28.9–36.5) Ningxia 27.0 (25.8–28.1) 18 26.5 (22.7–30.4) 27.5 (25.3–29.7) Qinghai 29.7 (12.3–47.1) 13 32.4 (12.5–52.3) 26.2 (13.4–39.0) Shaanxia 15.0 30 16.3 13.7 Shandong 38.2 (30.9–45.8) 3 37.7 (32.5–43.1) 38.4 (29.5–47.9) Shanghai 31.5 (27.0–36.2) 11 32.7 (27.4–38.3) 30.5 (26.5–34.5) Shanxi 24.7 (18.5–31.5) 19 27.3 (21.0–34.1) 22.5 (15.8–30.1) Sichuan 20.9 (16.7–25.3) 26 22.6 (19.4–26.0) 19.8 (14.6–25.5) Tianjina 32.4 10 32.2 32.5 Tibeta 40.7 2 41.3 40.0 Xinjiang 33.4 (31.5–35.2) 7 36.1 (34.6–37.6) 31.3 (27.9–34.8) Yunnan 23.3 (13.3–35.2) 20 26.1 (17.3–36.0) 21.0 (9.8–35.0) Zhejiang 27.9 (23.0–33.1) 16 28.9 (23.5–34.6) 27.1 (22.5–31.9) Province . Total prevalence (%) . Rank . Male prevalence (%) . Female prevalence (%) . Anhui 32.5 (25.7–39.6) 9 35.29 (30.5–40.1) 29.7 (21.5–38.5) Beijing 27.2 (23.2–32.4) 17 28.2 (23.9–32.7) 26.2 (21.8–30.8) Chongqing 22.6 (16.2–29.8) 23 23.2 (15.8–31.6) 22.1 (16.4–28.3) Fujiana 18.7 24 22.6 15.2 Gansu 30.2 (22.7–38.4) 12 32.8 (24.3–41.8) 28.3 (20.8–36.5) Guangdong 20.4 (17.9–23.0) 27 22.5 (19.3–25.9) 18.3 (15.6–21.2) Guangxi 23.0 (15.9–30.9) 22 25.5 (18.6–33.1) 20.7 (13.2–29.2) Guizhou 23.1 (19.2–27.3) 21 26.6 (19.6–34.3) 20.2 (18.1–22.4) Hebei 36.8 (32.9–40.9) 5 38.3 (33.6–43.1) 35.8 (31.7–40.0) Heilongjiang 28.2 (18.6–39.1) 15 29.6 (20.8–39.3) 26.9 (16.2–39.3) Henan 33.5 (29.2–37.8) 6 32.5 (28.0–37.1) 34.0 (29.3–38.9) Hubei 19.8 (12.5–27.1) 28 21.0 18.6 (8.0–29.2) Hunan 19.3 (11.5–28.7) 29 19.1 (10.5–29.6) 19.6 (12.3–28.2) Jiangsu 28.6 (25.2–32.2) 14 30.7 (26.9–34.8) 27.0 (23.7–30.4) Jiangxi 21.1 (9.5–32.7) 25 21.7 (10.9–32.4) 20.5 (7.9–33.0) Jilina 32.8 8 43.8 28.4 Liaoning 42.0 (36.3–47.8) 1 41.9 (35.8–48.1) 43.0 (36.8–49.2) Inner Mongolia 38.2 (34.1–42.3) 4 45.3 (38.8–51.9) 32.6 (28.9–36.5) Ningxia 27.0 (25.8–28.1) 18 26.5 (22.7–30.4) 27.5 (25.3–29.7) Qinghai 29.7 (12.3–47.1) 13 32.4 (12.5–52.3) 26.2 (13.4–39.0) Shaanxia 15.0 30 16.3 13.7 Shandong 38.2 (30.9–45.8) 3 37.7 (32.5–43.1) 38.4 (29.5–47.9) Shanghai 31.5 (27.0–36.2) 11 32.7 (27.4–38.3) 30.5 (26.5–34.5) Shanxi 24.7 (18.5–31.5) 19 27.3 (21.0–34.1) 22.5 (15.8–30.1) Sichuan 20.9 (16.7–25.3) 26 22.6 (19.4–26.0) 19.8 (14.6–25.5) Tianjina 32.4 10 32.2 32.5 Tibeta 40.7 2 41.3 40.0 Xinjiang 33.4 (31.5–35.2) 7 36.1 (34.6–37.6) 31.3 (27.9–34.8) Yunnan 23.3 (13.3–35.2) 20 26.1 (17.3–36.0) 21.0 (9.8–35.0) Zhejiang 27.9 (23.0–33.1) 16 28.9 (23.5–34.6) 27.1 (22.5–31.9) a Only one study was eligible after undergoing two level study screening. Open in new tab Rate of treatment and control for hypertension Of 176 studies, only 49 studies reported the rate of treatment. The results of these studies revealed the regional disparities of treatment for hypertension among the provinces, ranging from 64.2% in Shanghai to 18.2% in Gansu (Figure 3(a)). The treatment rate was lower in rural (31.5%, 95% CI 22.7–41.1%) than in urban (35.5%, 95% CI 15.4–58.9%) areas. The treatment rate was lower in men (39.3%, 95% CI 31.1–47.5%) compared to women (43.9%, 95% CI 35.7–52.2%). Figure 3. Open in new tabDownload slide (a) Geographial distribution of hypertension treatment rates in Chinese provinces. (b) Geographical distribution of hypertension control rates in Chinese provinces. Among the studies that investigated the control of hypertension, control was achieved in 13.4% of patients (95% CI 11.5–15.2%) (Figure 3(b)), with evidence of a strong urban–rural disparity: 12.8% of men achieved the target measures in urban areas, and 10.6% in rural areas; 16.0% of women had satisfactory response to treatment in urban areas, and 9.4% in rural areas. Trends of prevalence, treatment, and optimal control rates during the study period The trends of prevalence, treatment and control rates during the study period are shown in Figure 4. There was no evidence to suggest that the prevalence of hypertension has been increasing during the past decade. Interestingly, the rate of optimally controlled hypertension did not improve with increased treatment rates. Figure 4. Open in new tabDownload slide Prevalence, treatment and control rate trends during the study period. Multivariable meta-regression analysis Meta-regression was performed to investigate the relationship between the prevalence of hypertension and the altitudinal gradient as well as the economic level. Altitude was divided into four ranks and economic level was divided into five ranks based on the provinces’ annual gross domestic product. There was no evidence to suggest that prevalence had a significant correlation to altitudinal gradient (P = 0.782, 95% CI –5.63–4.28%) or economic level (P = 0.196, 95% CI –3.11–0.69%). A statistically significant effect was observed between hypertension prevalence and temperature gradient, which was divided into six ranks (P = 0.001, 95% CI 4.1–75.9%) (Figure 5). Figure 5. Open in new tabDownload slide Meta-regression of prevalence of hypertension against temperature gradient. Burden of hypertension There was a total of 10.7% (95% CI 8.1–13.3%) of all DALYs or 10,667 (95% CI 8063–13,345) DALYs per 100,000 individuals. Overall, 78.3% of DALYs resulted from YLDs and 21.7% from YLLs. YLDs contributed more to the burden of hypertension in China. The regional disparities in the burden of hypertension are shown in Supplementary Figure 3. The highest number of DALYs occurred in northern China, including Inner Mongolia, Jilin and Liaoning; the southwest, including Tibet; and the northwest, including Xinjiang. The next highest numbers of DALYs were found in the central China region, including Hunan, and the south China region, including Guangdong (Supplementary Figure 3(a), Table 2). Table 2. Years of life lost, years lived with disability and disability-adjusted life years by provinces. . YLDs/DALYs (%) . YLLs/DALYs (%) . DALYs (per 100,000) . YLDs (per 100,000) . YLLs (per 100,000) . Anhui 79 21 12,192 9661 2531 Beijing 75 25 10,907 8205 2702 Chongqing 71 29 9249 6572 2677 Fujian 73 27 9399 6877 2522 Gansu 76 24 10,386 7910 2476 Guangdong 72 28 8956 6404 2552 Guangxi 70 30 8794 6179 2615 Guizhou 75 25 9788 7286 2503 Hebei 83 17 13,452 11,157 2295 Henan 75 25 10,059 7578 2481 Heilongjiang 75 25 10,845 8124 2721 Hubei 70 30 8325 5765 2560 Hunan 69 31 8177 5594 2583 Jilin 81 19 14,772 11,994 2778 Jiangsu 79 21 11,612 9107 2505 Jiangxi 71 29 8169 5751 2418 Liaoning 82 18 14,294 11,695 2599 Inner Mongolia 83 17 15,145 12,522 2623 Ningxia 76 24 9529 7201 2328 Qinghai 78 22 11,415 8887 2529 Shandong 82 18 13,194 10,757 2437 Shanxi 75 25 9587 7214 2373 Shaanxi 65 35 7125 4613 2512 Shanghai 78 22 12,017 9336 2681 Sichuan 69 31 8389 5790 2599 Tianjin 74 26 10,730 7961 2769 Tibet 82 18 13,872 11,338 2534 Xinjiang 77 23 12,339 9480 2859 Yunnan 76 24 9629 7282 2347 Zhejiang 75 25 10,151 7613 2538 . YLDs/DALYs (%) . YLLs/DALYs (%) . DALYs (per 100,000) . YLDs (per 100,000) . YLLs (per 100,000) . Anhui 79 21 12,192 9661 2531 Beijing 75 25 10,907 8205 2702 Chongqing 71 29 9249 6572 2677 Fujian 73 27 9399 6877 2522 Gansu 76 24 10,386 7910 2476 Guangdong 72 28 8956 6404 2552 Guangxi 70 30 8794 6179 2615 Guizhou 75 25 9788 7286 2503 Hebei 83 17 13,452 11,157 2295 Henan 75 25 10,059 7578 2481 Heilongjiang 75 25 10,845 8124 2721 Hubei 70 30 8325 5765 2560 Hunan 69 31 8177 5594 2583 Jilin 81 19 14,772 11,994 2778 Jiangsu 79 21 11,612 9107 2505 Jiangxi 71 29 8169 5751 2418 Liaoning 82 18 14,294 11,695 2599 Inner Mongolia 83 17 15,145 12,522 2623 Ningxia 76 24 9529 7201 2328 Qinghai 78 22 11,415 8887 2529 Shandong 82 18 13,194 10,757 2437 Shanxi 75 25 9587 7214 2373 Shaanxi 65 35 7125 4613 2512 Shanghai 78 22 12,017 9336 2681 Sichuan 69 31 8389 5790 2599 Tianjin 74 26 10,730 7961 2769 Tibet 82 18 13,872 11,338 2534 Xinjiang 77 23 12,339 9480 2859 Yunnan 76 24 9629 7282 2347 Zhejiang 75 25 10,151 7613 2538 DALYs: disability-adjusted life years; YLDs: years lived with disability; YLLs: years of life lost due to premature mortality. Open in new tab Table 2. Years of life lost, years lived with disability and disability-adjusted life years by provinces. . YLDs/DALYs (%) . YLLs/DALYs (%) . DALYs (per 100,000) . YLDs (per 100,000) . YLLs (per 100,000) . Anhui 79 21 12,192 9661 2531 Beijing 75 25 10,907 8205 2702 Chongqing 71 29 9249 6572 2677 Fujian 73 27 9399 6877 2522 Gansu 76 24 10,386 7910 2476 Guangdong 72 28 8956 6404 2552 Guangxi 70 30 8794 6179 2615 Guizhou 75 25 9788 7286 2503 Hebei 83 17 13,452 11,157 2295 Henan 75 25 10,059 7578 2481 Heilongjiang 75 25 10,845 8124 2721 Hubei 70 30 8325 5765 2560 Hunan 69 31 8177 5594 2583 Jilin 81 19 14,772 11,994 2778 Jiangsu 79 21 11,612 9107 2505 Jiangxi 71 29 8169 5751 2418 Liaoning 82 18 14,294 11,695 2599 Inner Mongolia 83 17 15,145 12,522 2623 Ningxia 76 24 9529 7201 2328 Qinghai 78 22 11,415 8887 2529 Shandong 82 18 13,194 10,757 2437 Shanxi 75 25 9587 7214 2373 Shaanxi 65 35 7125 4613 2512 Shanghai 78 22 12,017 9336 2681 Sichuan 69 31 8389 5790 2599 Tianjin 74 26 10,730 7961 2769 Tibet 82 18 13,872 11,338 2534 Xinjiang 77 23 12,339 9480 2859 Yunnan 76 24 9629 7282 2347 Zhejiang 75 25 10,151 7613 2538 . YLDs/DALYs (%) . YLLs/DALYs (%) . DALYs (per 100,000) . YLDs (per 100,000) . YLLs (per 100,000) . Anhui 79 21 12,192 9661 2531 Beijing 75 25 10,907 8205 2702 Chongqing 71 29 9249 6572 2677 Fujian 73 27 9399 6877 2522 Gansu 76 24 10,386 7910 2476 Guangdong 72 28 8956 6404 2552 Guangxi 70 30 8794 6179 2615 Guizhou 75 25 9788 7286 2503 Hebei 83 17 13,452 11,157 2295 Henan 75 25 10,059 7578 2481 Heilongjiang 75 25 10,845 8124 2721 Hubei 70 30 8325 5765 2560 Hunan 69 31 8177 5594 2583 Jilin 81 19 14,772 11,994 2778 Jiangsu 79 21 11,612 9107 2505 Jiangxi 71 29 8169 5751 2418 Liaoning 82 18 14,294 11,695 2599 Inner Mongolia 83 17 15,145 12,522 2623 Ningxia 76 24 9529 7201 2328 Qinghai 78 22 11,415 8887 2529 Shandong 82 18 13,194 10,757 2437 Shanxi 75 25 9587 7214 2373 Shaanxi 65 35 7125 4613 2512 Shanghai 78 22 12,017 9336 2681 Sichuan 69 31 8389 5790 2599 Tianjin 74 26 10,730 7961 2769 Tibet 82 18 13,872 11,338 2534 Xinjiang 77 23 12,339 9480 2859 Yunnan 76 24 9629 7282 2347 Zhejiang 75 25 10,151 7613 2538 DALYs: disability-adjusted life years; YLDs: years lived with disability; YLLs: years of life lost due to premature mortality. Open in new tab In China, the average life expectancy at birth was 75.6 years (73.7 years for men and 77.8 years for women). When expressed in terms of life expectancy (Supplementary Table 2), the ranking of the regions changed (Supplementary Figure 3(b)). Shanghai, Beijing and Tianjin were the top provinces on the life expectancy list, while Tibet, Qinghai and Yunnan had a lower life expectancy at birth. In nearly all regions, YLDs made up a larger share of DALYs. In Table 2, in which the regions are ordered by DALYs, it can be seen that, in general, the share of burden from disability varied according to demographic and epidemiological transition. The fraction of DALYs due to YLDs varied widely, from 82.9% in Hebei and 82.7% in Inner Mongolia to 64.7% in Shaanxi. In the northeast region, the fraction of DALYs due to YLD was noticeably higher than that due to YLL, especially in Liaoning province. Discussion To calculate attributable burden, an accurate assessment of the population distribution of hypertension with which to compare current distributions is needed. When the prevalence of hypertension in the general population was analysed, only studies utilising participants under the age of 40 years at study onset were included in order to provide conservative estimates of hypertension rates. Our study found that the prevalence of hypertension in China was 28.9% according to 146 observational studies and 26.8% according to six national cross-sectional studies.23,24 The prevalence results obtained in this meta-analysis corroborate previous estimates in the literature. Recent cross-sectional studies found that prevalence was 26.6% in 2007–20088 and 29.6% in 2009–2010.7 Overall, the prevalence of hypertension in recent years did not change substantially. Our results showed that treatment of hypertension has nearly doubled compared to 10 years ago. This shows that financial support for the management of hypertension since 2005 has improved at the national level. However, the rate of control of hypertension has not improved significantly. Compared to developed countries,11,25 there is room for improvement in China with respect to hypertension control. Meta-regression was performed to investigate the potential effect of prevalence. Each study was examined for between-study heterogeneity in the altitudinal gradient, economic level and temperature gradient categories. No statistically significant effects were found for altitudinal gradient or economic level, but an increased percentage of patients at lower temperature levels had hypertension. Thus, temperature levels seem to have an inverse correlation with the prevalence of hypertension. Although there was a slight increase in the treatment and control rates of hypertension compared to a report published in 2002,10 these rates were still suboptimal, especially for men and people living in rural areas. Moreover, regional heterogeneity highlights critical differences in the management of hypertension and the importance of understanding the local burden of disease. Life expectancy at birth reflects the overall health level of a population.26–28 There was a large amount of heterogeneity with respect to life expectancy among the provinces, from 83.89 years in Shanghai to 74.23 years in Tibet. This phenomenon may correlate with the local disease spectrum, universal health service coverage, local sanitation and health education. Shanghai and Beijing, as two of the most developed regions in China, are at the top of the life expectancy list. These developed economies appear to be driving the local government to sustain high health investment, which may lead to relatively lower local hypertension burdens. By contrast, low and middle-income provinces, such as Tibet, Qinghai and Inner Mongolia, have a comparatively huge burden of hypertension, which is a considerable factor in reducing life expectancy. According to our results, approximately 143 million DALYs in China were attributable to hypertension. The regional disparity in the burden of disease attributable to hypertension is an issue to monitor. To meet the huge public health challenge of managing hypertension in such a large populace, the government of China, in cooperation with the World Health Organization and other entities, should continue to develop and implement new medium and long-term national plans for the local management of hypertension. A comprehensive overview of the prevalence, treatment and control of hypertension was conducted to estimate the burden of hypertension in China, which will be of benefit to improve Chinese health policies for prevention and control of this condition. There are some limitations in our study. Firstly, the data are based on shorter-term studies. It would have been better to obtain morbidity data from long-term follow-up on populations with hypertension so that DALYs could be calculated more accurately. However, there are few long-term Chinese studies with relevant data. Therefore, DALYs had to be calculated by the disability weight. Secondly, the meta-regression analysis revealed that temperature levels were the only factor correlated with the prevalence of hypertension. This is likely to be due to the fact that few of the original reports detailed other interacting factors such as body weight and sleeping behaviour. Finally, the economic data of the meta-regression was based on the average level in provinces, without urban and rural stratification. Conclusions This meta-analysis included 178 studies published in English and Chinese. It was a systematic review of hypertension prevalence and burden of disease attributable to hypertension in the Chinese adult population over the past few decades. Overall, the prevalence, treatment and control of hypertension in China were 28.9%, 35.3% and 13.4%, respectively. Although there has been a slight increase in rates for the treatment and control of hypertension, these rates were still suboptimal, especially for men and people living in rural areas. Low and middle-income provinces had a comparatively huge burden of hypertension, which is a considerable risk factor for reducing life expectancy. Our data regarding the burden of hypertension and within-region variations in the prevalence and management of hypertension may be helpful in generating an overview of the current situation in China, and may serve as a reliable and informative starting point for an in-depth study of the national burden of hypertension. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by grants from the National Basic Research Program of China (973 Program), no. 2014CB542302, the National Natural Science Foundation of China (nos. 811170244, 81222001, 81470541). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. References 1 Bikbov B , Perico N, Remuzzi G. Mortality landscape in the Global Burden of Diseases, Injuries and Risk Factors Study . Eur J Intern Med 2014 ; 25 : 1 – 5 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Horton R . GBD 2010: understanding disease, injury, and risk . Lancet 2012 ; 380 : 2053 – 2054 . Google Scholar Crossref Search ADS PubMed WorldCat 3 World Health Organization. A global brief on hypertension. Silent killer, global public health crisis. http://www.who.int/cardiovascular_diseases/publications/global_brief_hypertension/en/ (2013, accessed 26 January 2015) . 4 Kearney PM , Whelton M, Reynolds Ket al. . Global burden of hypertension: analysis of worldwide data . Lancet 2005 ; 365 : 217 – 223 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Yang G , Wang Y, Zeng Yet al. . Rapid health transition in China, 1990–2010: findings from the Global Burden of Disease Study 2010 . Lancet 2013 ; 381 : 1987 – 2015 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Wang L , Kong L, Wu Fet al. . Preventing chronic diseases in China . Lancet 2005 ; 366 : 1821 – 1824 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Wang J , Zhang L, Wang Fet al. . China National Survey of Chronic Kidney Disease Working Group . Prevalence, awareness, treatment, and control of hypertension in China: Results from a national survey . Am J Hypertens 2014 ; 27 : 1355 – 1361 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Gao Y , Chen G, Tian Het al. . China National Diabetes and Metabolic Disorders Study Group . Prevalence of hypertension in China: a cross-sectional study . PLoS One 2013 ; 8 : e65938 – e65938 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Li LM , Rao KQ, Kong LZet al. . Technical Working Group of China National Nutrition and Health Survey . A description on the Chinese national nutrition and health survey in 2002 . Zhonghua Liu Xing Bing Xue Za Zhi 2005 ; 26 : 478 – 484 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 10 Gu D , Reynolds K, Wu Xet al. . InterASIA Collaborative Group. The International Collaborative Study of Cardiovascular Disease in ASIA . Prevalence, awareness, treatment, and control of hypertension in China . Hypertension 2002 ; 40 : 920 – 927 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Egan BM , Zhao Y, Axon RN. US trends in prevalence, awareness, treatment, and control of hypertension, 1988–2008 . JAMA 2010 ; 303 : 2043 – 2050 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Bai J , Tang ZL, Li Let al. . A systematic review: prevalence rate of hypertension in China . Shanghai J Prev Med 2012 ; 24 : 604 – 616 . Google Scholar OpenURL Placeholder Text WorldCat 13 Fang L, Song J, Ma Z, et al. Prevalence and characteristics of hypertension in mainland Chinese adults over decades: a systematic review. J Hum Hypertens 2014; 28: 649–656 . 14 Rostom A, Dube C, Cranney A, et al. Evidence Reports/Technology Assessments, No. 104. Appendix D. Quality Assessment Forms. Rockville (MD): Agency for Healthcare Research and Quality (US). http://www.ncbi.nlm.nih.gov/books/NBK35156 (2004, accessed 15 May 2014) . 15 China Meteorological Administration. The China Climate Bulletin in 2013. http://cmdp.ncc-cma.net/Monitoring/cn_moni_china.php?product=cn_moni_china_basic (2013, accessed 15 September 2014) . 16 China Statistics Press. China Statistical Yearbook. Available at: http://www.stats.gov.cn/tjsj/ndsj/2013/indexeh.htm (2013, accessed 15 September 2014) . 17 Mokdad AH , Jaber S, Aziz MIet al. . The state of health in the Arab world, 1990–2010: an analysis of the burden of diseases, injuries, and risk factors . Lancet 2014 ; 383 : 309 – 320 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Lopez AD , Mathers CD, Ezzati Met al. . Global burden of disease and risk factors , New York : Oxford University Press , 2006 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 19 Salomon JA , Vos T, Hogan DRet al. . Common values in assessing health outcomes from disease and injury: disability weights measurement study for the Global Burden of Disease Study 2010 . Lancet 2012 ; 380 : 2129 – 2143 . Google Scholar Crossref Search ADS PubMed WorldCat 20 The Central People’s Government of the People’s Republic of China. Administrative divisions of the People's Republic of China. http://www.gov.cn/test/2005-06/15/content_18253.htm (2014, 19 September 2014) (in Chinese) . 21 Gu DF , Jiang H, Wu XGet al. . Prevalence, awareness, treatment and control of hypertension in Chinese adults . Zhonghua Yu Fang Yi Xue Za Zhi 2003 ; 37 : 84 – 89 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 22 Zhao XL , Chen J, Cui YLet al. . Current status of primary hypertension in China: an epidemiological study of 12 provinces, 1 autonomous regions, and 1 municipality . Zhonghua Yi Xue Za Zhi 2006 ; 86 : 1148 – 1152 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 23 Xi B , Liang Y, Reilly KHet al. . Trends in prevalence, awareness, treatment, and control of hypertension among Chinese adults 1991–2009 . Int J Cardiol 2012 ; 158 : 326 – 329 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Wu Y , Huxley R, Li Let al. . China NNHS Steering Committee; China NNHS Working Group . Prevalence, awareness, treatment, and control of hypertension in China: data from the China National Nutrition and Health Survey 2002 . Circulation 2008 ; 118 : 2679 – 2686 . Google Scholar Crossref Search ADS PubMed WorldCat 25 McAlister FA , Wilkins K, Joffres Met al. . Changes in the rates of awareness, treatment and control of hypertension in Canada over the past two decades . CMAJ 2011 ; 183 : 1007 – 1013 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Ezzati M , Lopez A, Rodgers Aet al. . Comparative quantification of health risks: Global and regional burden of disease attributable to selected major risk factors , Geneva : World Health Organization , 2004 , pp. 2129 – 2140 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 27 World Health Organization . The World Health Report, 2002: Reducing risks, promoting healthy life , Geneva : World Health Organization , 2002 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 28 Sarink D , Nedkoff L, Briffa Tet al. . Projected age- and sex-specific prevalence of cardiovascular diseases in Western Australian adults from 2005–2045 . Eur J Prev Cardiol 2014 ; 20 : 938 – 946 . Google Scholar OpenURL Placeholder Text WorldCat Author notes * These authors contributed equally to this work. They are considered co-first authors. © The European Society of Cardiology 2016 This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © The European Society of Cardiology 2016

Pfaeffli Dale, Leila; Dobson, Rosie; Whittaker, Robyn; Maddison, Ralph

doi: 10.1177/2047487315613462pmid: 26490093

Abstract Background Mobile wireless devices (mHealth) have been used to deliver cardiovascular disease self-management interventions to educate and support patients in making healthy lifestyle changes. This systematic review aimed to determine the effectiveness of mHealth interventions on behavioural lifestyle changes and medication adherence for cardiovascular disease self-management. Methods A comprehensive literature search was conducted from inception through to 3 March 2015 using MEDLINE, PubMed, PsycINFO, EMBASE and The Cochrane Library. Eligible studies used an experimental trial design to determine the effectiveness of an mHealth intervention to change lifestyle behaviours in any cardiovascular disease population. Data extracted included intervention and comparison group characteristics with a specific focus on the use of behaviour change techniques. Results Seven studies met our inclusion criteria and were included in the qualitative synthesis. All interventions were delivered in part by mobile phone text messaging. Three studies were effective at improving adherence to medication and two studies increased physical activity behaviour. No effects were observed on dietary behaviour or smoking cessation, measured in one study each. Simple text messaging interventions appeared to be most effective; however, no clear relationships were found between study findings and intervention dose, duration or behaviour change techniques targeted. Conclusions Our review found mHealth has the potential to change lifestyle behaviour. Results are still limited to a small number of trials, inconsistent outcome measures and ineffective reporting of intervention characteristics. Large scale, longitudinal studies are now warranted to gain a clear understanding of the effects of mHealth on behaviour change in the cardiovascular disease population. Cardiovascular diseases, mHealth, text messaging, cellular phone, lifestyle change, behaviour, medication adherence Introduction Unhealthy lifestyle behaviours and modifiable individual risk factors are responsible for approximately 80% of cardiovascular disease (CVD).1 Self-management or secondary prevention programmes, commonly known as cardiac rehabilitation, educate and support patients to make healthy lifestyle changes to reduce subsequent cardiac events. Lifestyle behaviour changes include starting and maintaining regular physical activity, eating a healthy diet, stopping smoking, reducing harmful alcohol intake and taking medications as per prescribed regimens.2 Through the targeting of multiple behaviour change, secondary prevention programmes have been associated with reduced mortality and repeat cardiac events, and/or improved health related quality of life.3,4 Although the evidence shows that taking part in cardiac rehabilitation programmes benefits patients, attendance rates remain at less than 50% worldwide.5 Typically, cardiac rehabilitation comprises face-to-face group education and/or exercise sessions for outpatients and can be located at a hospital or community setting. Many patients face significant barriers to attending face-to-face centre-based programmes, such as a lack of transportation or embarrassment about participation,6 and calls have been made to offer alternative types of programmes, such as those delivered in the home.7 Home-based cardiac rehabilitation has shown to be as effective as centre-based with similar health care costs.8 One type of home-based secondary prevention programme that has recently gained attention is telehealth. There are several definitions of telehealth but on the whole it includes functions of remote monitoring, health care provider/patient communication, and information/education exchange.9 Previous systematic reviews and meta-analyses have found telehealth interventions were associated with several positive risk factor modifications in patients with coronary heart disease, including significantly lower total cholesterol and blood pressure, and fewer smokers,6 and outcomes did not significantly differ from centre-based cardiac rehabilitation.10 Telehealth paints a promising picture, but on the whole the mechanisms to deliver telehealth were through landline and desktop or bespoke units. The notion of being physically bound to the intervention tool at home may still be a barrier for some patients and is reflective of the existing provider centric care model. The mobility of smart phones and other portable devices offers greater delivery opportunities and possibilities to enhance self-management in people with CVD. Mobile phones are now omnipresent, with worldwide usage rates nearing 100%.11 Mobile wireless devices for health communication, known as mHealth, have been incorporated into various behaviour change and disease management interventions. The most researched form of mHealth has been the use of text or short-message service (SMS), found to facilitate significant positive effects on health outcomes and/or behaviours.12 The majority of mHealth disease management interventions have focused on the diabetes population.13 The uptake of mHealth for CVD self-management has been slow, perhaps due to earlier misconceptions that people with CVD are uncomfortable using technology as they tend to be older, and adults over aged 65 are less likely to own mobile phones than younger age groups.14 Despite lower mobile phone usage among older adults, recent reviews have shown the CVD population are able to use and engage with technology.15,16 Two recent systematic reviews have assessed the effectiveness of digital (web-based, telemedicine, mobile phone and monitoring sensors)15 and mHealth16 interventions on CVD outcomes. Widmer and colleagues' meta-analysis found digital interventions had beneficial effects on CVD outcomes, including CVD events and all-cause mortality (relative risk (RR), 0.60; 95% confidence interval (CI), 0.43 − 0.83; p = .002; I2 = 0%).15 Hamine et al. found mixed evidence with 7/13 (54%) mHealth interventions showing significant improvements on CVD risk factors such as blood pressure, weight and lipid profile.16 While both reviews were of high-quality, neither examined behaviour change. Lifestyle behaviour change is more proximal to the intervention and the antecedent to clinical outcomes,17 and forms the core of secondary prevention programmes. In fact, a need to review behaviour change and motivational techniques was identified as an area of future research by Widmer et al.,15 as such a review may yield meaningful findings on the features of successful interventions with clinical implications. Understanding the components of successful mHealth interventions, including the use of theory and/or behaviour change techniques (BCTs), may help guide future intervention development for the CVD population. Thus, the primary objective of this systematic review was to determine the effectiveness of mHealth interventions on behavioural lifestyle changes and medication adherence for CVD self-management. Methods This systematic review was conducted according to the PRISMA checklist18 (Supplementary file 1 in the Supplementary Material online) and the Cochrane Systematic Review Handbook;19 however, no review protocol has been published. Eligibility criteria Table 1 outlines our review inclusion criteria. Eligible studies were experimental or quasi-experimental trials published in peer-reviewed journals in English. Participants included patients of any age with any type of established CVD (coronary heart disease, acute coronary syndrome, heart failure, stroke, congenital heart disease). Table 1. The systematic review research question as defined by the PICOS approach. Participants . Interventions . Comparator . Outcomes . Study design . Patients of any age diagnosed with any type of CVD Delivered via mHealth Presence of any type of comparison group Individual risk factor behaviour change for CVD self-management (physical activity, diet, smoking, alcohol use, medication adherence) Experimental or quasi-experimental Participants . Interventions . Comparator . Outcomes . Study design . Patients of any age diagnosed with any type of CVD Delivered via mHealth Presence of any type of comparison group Individual risk factor behaviour change for CVD self-management (physical activity, diet, smoking, alcohol use, medication adherence) Experimental or quasi-experimental PICOS: participants/interventions/comparator/outcomes/study design; CVD: cardiovascular disease. Open in new tab Table 1. The systematic review research question as defined by the PICOS approach. Participants . Interventions . Comparator . Outcomes . Study design . Patients of any age diagnosed with any type of CVD Delivered via mHealth Presence of any type of comparison group Individual risk factor behaviour change for CVD self-management (physical activity, diet, smoking, alcohol use, medication adherence) Experimental or quasi-experimental Participants . Interventions . Comparator . Outcomes . Study design . Patients of any age diagnosed with any type of CVD Delivered via mHealth Presence of any type of comparison group Individual risk factor behaviour change for CVD self-management (physical activity, diet, smoking, alcohol use, medication adherence) Experimental or quasi-experimental PICOS: participants/interventions/comparator/outcomes/study design; CVD: cardiovascular disease. Open in new tab Studies were included based on 1) intervention delivery and 2) outcomes. Included interventions were delivered at least in part via mHealth, defined as health communication delivered using a mobile wireless device such as a mobile phone, patient monitoring device, personal digital assistant (PDA) or other wireless device.20 Interventions using solely landline telephone communication were excluded. Web-based interventions were included only if the intervention was specifically delivered on handheld or tablet devices. Studies must have included one of the following lifestyle behaviour outcomes (primary or secondary outcome): physical activity, diet, smoking, alcohol use, or medication adherence. Mobile phone-based telemedicine studies using remote monitoring of symptoms or clinical data were included if behaviour change outcomes were reported. The primary outcome of interest in this review was a change in lifestyle behaviour (see Table 1), as this is the antecedent to clinical outcomes and often the aim of CVD self-management or secondary prevention programmes. Search strategy A comprehensive literature search was conducted from inception through to 3 March 2015 using MEDLINE, PubMed, PsycINFO, EMBASE and The Cochrane Library. Full details of the MEDLINE search strategy can be found in Table 2 (amended for other databases). No language, publication date, or publication status restrictions were imposed on the search strategy. The search strategy followed guidelines from the Cochrane Systematic Review Handbook and was created by the first author (LPD) with guidance from other authors. Reference lists of relevant studies or reviews were also searched manually. If relevant abstracts were found, authors were contacted for full publications where possible. Articles published in languages other than English were excluded during the full text screening. Table 2. Ovid MEDLINE search strategy (inception through to 3 March 2015). Search . Search term . Combination . Result . 1 cellular phone/ 4838 2 (cellular phone* or cell phone* or cellular telephone* or cell telephone*).ab,ti. 2162 3 (mobile phone* or mobile telephone*).ab,ti. 3697 4 (smart phone* or smartphone* or smart-phone* or iphone or i-phone).ab,ti. 1877 5 (text messag* or sms* or short messag*).ab,ti. 5089 6 (texting or texted).ab,ti. 285 7 (multimedia messag* or mms* or multi-media messag*).ab,ti. 11,358 8 application software.ab,ti. 172 9 (mHealth or m-health or mobile-health).ab,ti. 906 10 wireless technolog*/ 1177 11 Bluetooth.ab,ti. 370 12 computers, handheld/ 2397 13 (tablet or ipad or i-pad).ab,ti. 18,013 14 Internet/ 51,169 15 Internet.ab,ti. 30,226 16 (web?based or web-based).ab,ti. 15,699 17 (online or on-line).ab,ti. 63,650 18 (e-health or ehealth or electronic health).ab,ti. 7149 19 or/1–18 175,176 20 Exp cardiovascular diseases/ 1,908,007 21 Cardiac rehabilitation.ab,ti 3913 22 Exp hyperlipidemia/ 56,445 23 Exp cardiology/ 14,504 24 Or/22–25 1,948,637 25 (exercis* or physical activ* or diet or nutrition or tobacco or smoking or adherence).ab,ti. 789,064 26 (interven* or program* or treatment).tw. 3,903,140 27 19 and 24 and 25 and 26 497 Search . Search term . Combination . Result . 1 cellular phone/ 4838 2 (cellular phone* or cell phone* or cellular telephone* or cell telephone*).ab,ti. 2162 3 (mobile phone* or mobile telephone*).ab,ti. 3697 4 (smart phone* or smartphone* or smart-phone* or iphone or i-phone).ab,ti. 1877 5 (text messag* or sms* or short messag*).ab,ti. 5089 6 (texting or texted).ab,ti. 285 7 (multimedia messag* or mms* or multi-media messag*).ab,ti. 11,358 8 application software.ab,ti. 172 9 (mHealth or m-health or mobile-health).ab,ti. 906 10 wireless technolog*/ 1177 11 Bluetooth.ab,ti. 370 12 computers, handheld/ 2397 13 (tablet or ipad or i-pad).ab,ti. 18,013 14 Internet/ 51,169 15 Internet.ab,ti. 30,226 16 (web?based or web-based).ab,ti. 15,699 17 (online or on-line).ab,ti. 63,650 18 (e-health or ehealth or electronic health).ab,ti. 7149 19 or/1–18 175,176 20 Exp cardiovascular diseases/ 1,908,007 21 Cardiac rehabilitation.ab,ti 3913 22 Exp hyperlipidemia/ 56,445 23 Exp cardiology/ 14,504 24 Or/22–25 1,948,637 25 (exercis* or physical activ* or diet or nutrition or tobacco or smoking or adherence).ab,ti. 789,064 26 (interven* or program* or treatment).tw. 3,903,140 27 19 and 24 and 25 and 26 497 Open in new tab Table 2. Ovid MEDLINE search strategy (inception through to 3 March 2015). Search . Search term . Combination . Result . 1 cellular phone/ 4838 2 (cellular phone* or cell phone* or cellular telephone* or cell telephone*).ab,ti. 2162 3 (mobile phone* or mobile telephone*).ab,ti. 3697 4 (smart phone* or smartphone* or smart-phone* or iphone or i-phone).ab,ti. 1877 5 (text messag* or sms* or short messag*).ab,ti. 5089 6 (texting or texted).ab,ti. 285 7 (multimedia messag* or mms* or multi-media messag*).ab,ti. 11,358 8 application software.ab,ti. 172 9 (mHealth or m-health or mobile-health).ab,ti. 906 10 wireless technolog*/ 1177 11 Bluetooth.ab,ti. 370 12 computers, handheld/ 2397 13 (tablet or ipad or i-pad).ab,ti. 18,013 14 Internet/ 51,169 15 Internet.ab,ti. 30,226 16 (web?based or web-based).ab,ti. 15,699 17 (online or on-line).ab,ti. 63,650 18 (e-health or ehealth or electronic health).ab,ti. 7149 19 or/1–18 175,176 20 Exp cardiovascular diseases/ 1,908,007 21 Cardiac rehabilitation.ab,ti 3913 22 Exp hyperlipidemia/ 56,445 23 Exp cardiology/ 14,504 24 Or/22–25 1,948,637 25 (exercis* or physical activ* or diet or nutrition or tobacco or smoking or adherence).ab,ti. 789,064 26 (interven* or program* or treatment).tw. 3,903,140 27 19 and 24 and 25 and 26 497 Search . Search term . Combination . Result . 1 cellular phone/ 4838 2 (cellular phone* or cell phone* or cellular telephone* or cell telephone*).ab,ti. 2162 3 (mobile phone* or mobile telephone*).ab,ti. 3697 4 (smart phone* or smartphone* or smart-phone* or iphone or i-phone).ab,ti. 1877 5 (text messag* or sms* or short messag*).ab,ti. 5089 6 (texting or texted).ab,ti. 285 7 (multimedia messag* or mms* or multi-media messag*).ab,ti. 11,358 8 application software.ab,ti. 172 9 (mHealth or m-health or mobile-health).ab,ti. 906 10 wireless technolog*/ 1177 11 Bluetooth.ab,ti. 370 12 computers, handheld/ 2397 13 (tablet or ipad or i-pad).ab,ti. 18,013 14 Internet/ 51,169 15 Internet.ab,ti. 30,226 16 (web?based or web-based).ab,ti. 15,699 17 (online or on-line).ab,ti. 63,650 18 (e-health or ehealth or electronic health).ab,ti. 7149 19 or/1–18 175,176 20 Exp cardiovascular diseases/ 1,908,007 21 Cardiac rehabilitation.ab,ti 3913 22 Exp hyperlipidemia/ 56,445 23 Exp cardiology/ 14,504 24 Or/22–25 1,948,637 25 (exercis* or physical activ* or diet or nutrition or tobacco or smoking or adherence).ab,ti. 789,064 26 (interven* or program* or treatment).tw. 3,903,140 27 19 and 24 and 25 and 26 497 Open in new tab Selection of studies The search was conducted by the first author (LPD). Search results were merged into EndNote X7 Referencing Software and duplicates were removed. Titles and abstracts were then screened to remove obviously irrelevant articles. Articles identified for full text review were compared with eligibility criteria by two authors (LPD and RD). Disagreements were brought to a third reviewer (RM) for discussion until consensus was reached. Reasons for exclusion of studies were recorded. Data collection process Data were extracted by the first author (LPD) using items informed from the PRISMA checklist and the Cochrane Systematic Review Handbook: 1) population (type of CVD, age, sex, country, setting, sample size, inclusion/exclusion criteria); 2) study design (design, length of follow-up); 3) intervention (technology used, characteristics, dose, theoretical framework and BCTs, duration of intervention and follow-up); 4) comparator (type, description, dose); 5) outcomes (primary and secondary outcome measures and results). Study authors were contacted if additional non-published data were required. Data extraction was verified independently by the second author (RD). Risk of bias assessment Risk of bias was assessed using the Cochrane Collaboration's tool for assessing risk of bias19 and the Jadad score for study quality.21 The data collected included the following: method of randomisation and allocation concealment; blinding of outcome assessors and participants; number of participants randomised, excluded and lost to follow-up; reporting of power calculation; type of analysis (intention to treat or per protocol); reporting of pre-specified outcomes. Results The combined search strategy identified 3456 records, of which eight were identified through manual searching of reference lists and Google Scholar (Figure 1). Once duplicates were removed, 2374 records were screened for eligibility by title and abstract. The full text was obtained for 192 records, of which nine (describing seven trials) met the inclusion criteria. A list of excluded studies, with the primary reason for exclusion, can be found in Supplementary file 2 online. Due to the heterogeneity of study intervention and outcome measures, meta-analysis was not possible,19 therefore we summarised the study characteristics, results, limitations and implications using qualitative synthesis. Figure 1. Open in new tabDownload slide Flow diagram of included studies. CVD: cardiovascular disease. Characteristics of studies The characteristics of included studies can be found in Table 322–30 and will be described according to participants, study design, intervention and comparison groups, risk of bias, and effectiveness. Table 3. Characteristics of included studies. Study . Participants . Study design; device; media . Aim . Intervention . Comparator . Intervention duration; follow-up . Effect on behaviour change? . Antypas 201422 69 adults with CVD Mean age (95% CI) intervention: 60 (56–63) Control: 59 (56–62) Male: 52/67 (78%) Country: Norway Two-group cluster RCT Device: computer + mobile phone Media: website + SMS To assess the effect of a tailored Internet- and mobile phone-based intervention on the maintenance of physical activity levels after a cardiac rehabilitation stay Participants had access to website and received tailored messages via web and SMS. The website contained info about CVD and self-management, and an online discussion forum. Participants could set physical activity goals based on stage of change and received feedback via graph on website. They received SMS reminders about planned physical activity and if activity was completed. Access to non-tailored website and online forum Duration: not stated Follow-up: one and three months Yes (physical activity) Blasco 201223 203 adults with ACS with at least one CVD risk factor Mean age (SD) Intervention: 61 (12) Control: 61 (12) Male: 163/203 (80%) Country: Spain Single blind RCT Device; media: patient: mobile phone; SMS physician: computer; web application To analyse the efficacy of a telemonitoring system for the follow-up of patients with ACS The telemedicine group were provided with clinical measurement devices and sent results via their mobile phone. Physicians accessed the patient data via the web application and sent individualised SMS with recommendations to the patient. Usual care (met with cardiologist 3 × over study period; received verbal and written information about CVD prevention Duration and follow-up: 12 months No (smoking cessation) Khonsari 201524 62 adults with ACS Mean age (SD): 58 (13) Male: 53/62 (86%) Country: Malaysia Open-labelled RCT Device: mobile phone Medium: SMS To investigate the effect of automated SMS-based reminders on post-discharge medication adherence among patients with ACS Participants received automated SMS reminders before every medication intake. Also received SMS reminder for prescription refill after 30 days. Researcher telephoned participants once per fortnight to check in on SMS delivery and whether hospital readmission was needed. Usual care (CR and follow-up appointment with cardiologist) Duration and follow-up: Eight weeks Yes (medication adherence) Maddison 2015;25 201426 171 adults with IHD within previous 3–24 months Mean age (SD): 60 (9.3) Male: 139/171 (81%) Country: New Zealand Parallel Two-arm RCT Device: mobile phone + computer Media: SMS + website To determine the effectiveness and cost-effectiveness of a mHealth delivered exercise CR programme for people with IHD to improve exercise capacity and physical activity levels; to evaluate the mediating effects of self-efficacy on physical activity levels in the mHealth programme The intervention group received exercise prescription and physical activity behaviour change SMS (3–5 per week) + access to a website containing video messages (three per week), CR information, self-monitoring of PA on website Usual care (encouragement to attend CR) Duration and follow-up: 24 weeks Yes (physical activity) Park 2014;27 Park 201528 90 adults with CHD, hospitalised for MI or PCI Mean age (SD): 59 (9.4) Male: 68/90 (76%) Country: USA Three-arm prospective RCT Device: mobile phone Medium: SMS To examine the efficacy of a mHealth intervention to improve adherence to antiplatelet and statin medications among MI and/or PCI patients; to compare medication self-efficacy Arm 1: received two-way SMS medication reminders (requiring response) and one-way health education SMS (74 SMS messages over 30 days). Arm 2: received one-way health education SMS alone (14 messages over 30 days). Arm 3: Control group (no treatment) Duration and follow-up: 30 days Yes (medication adherence) Quilici 201329 521 patients who had coronary stenting for ACS Mean age (SD): 64 (10) Male: 382/499 (77%) Country: not stated but author affiliation is France RCT pilot study Device: mobile phone Medium: SMS To test whether participants receiving daily motivational SMS had better aspirin completion rates compared with standard care Intervention group received one daily personalised and motivational SMS reminder to take aspirin for 30 days Standard care (not described) Duration and follow-up: one month Yes (medication adherence) Varnfield 201430 120 post MI patients referred to CR Mean age (SD): 56 (10) Male: 82/94 (87%) (94/120 completed baseline) Country: Australia Two-group parallel RCT Device: smartphone Media: SMS, multimedia materials (video, audio), application (wellness diary, step counter) To investigate whether CAP-CR is effective in improving CR use in post-MI patients compared with a traditional centre based programme CAP-CR: smartphone for health and exercise monitoring via a wellness diary and step counter applications. Motivational and educational materials were delivered by SMS, audio and visual files. Participants received all core components of CR. Mentors reviewed participants' data via a web platform and provided weekly telephone consultations to feed back on goals. Standard care (traditional centre-based CR) Duration: six weeks; however, the CAP-CR group could use the smartphone and monitoring devices for six months. Follow-up: six weeks and six months. No (diet) Study . Participants . Study design; device; media . Aim . Intervention . Comparator . Intervention duration; follow-up . Effect on behaviour change? . Antypas 201422 69 adults with CVD Mean age (95% CI) intervention: 60 (56–63) Control: 59 (56–62) Male: 52/67 (78%) Country: Norway Two-group cluster RCT Device: computer + mobile phone Media: website + SMS To assess the effect of a tailored Internet- and mobile phone-based intervention on the maintenance of physical activity levels after a cardiac rehabilitation stay Participants had access to website and received tailored messages via web and SMS. The website contained info about CVD and self-management, and an online discussion forum. Participants could set physical activity goals based on stage of change and received feedback via graph on website. They received SMS reminders about planned physical activity and if activity was completed. Access to non-tailored website and online forum Duration: not stated Follow-up: one and three months Yes (physical activity) Blasco 201223 203 adults with ACS with at least one CVD risk factor Mean age (SD) Intervention: 61 (12) Control: 61 (12) Male: 163/203 (80%) Country: Spain Single blind RCT Device; media: patient: mobile phone; SMS physician: computer; web application To analyse the efficacy of a telemonitoring system for the follow-up of patients with ACS The telemedicine group were provided with clinical measurement devices and sent results via their mobile phone. Physicians accessed the patient data via the web application and sent individualised SMS with recommendations to the patient. Usual care (met with cardiologist 3 × over study period; received verbal and written information about CVD prevention Duration and follow-up: 12 months No (smoking cessation) Khonsari 201524 62 adults with ACS Mean age (SD): 58 (13) Male: 53/62 (86%) Country: Malaysia Open-labelled RCT Device: mobile phone Medium: SMS To investigate the effect of automated SMS-based reminders on post-discharge medication adherence among patients with ACS Participants received automated SMS reminders before every medication intake. Also received SMS reminder for prescription refill after 30 days. Researcher telephoned participants once per fortnight to check in on SMS delivery and whether hospital readmission was needed. Usual care (CR and follow-up appointment with cardiologist) Duration and follow-up: Eight weeks Yes (medication adherence) Maddison 2015;25 201426 171 adults with IHD within previous 3–24 months Mean age (SD): 60 (9.3) Male: 139/171 (81%) Country: New Zealand Parallel Two-arm RCT Device: mobile phone + computer Media: SMS + website To determine the effectiveness and cost-effectiveness of a mHealth delivered exercise CR programme for people with IHD to improve exercise capacity and physical activity levels; to evaluate the mediating effects of self-efficacy on physical activity levels in the mHealth programme The intervention group received exercise prescription and physical activity behaviour change SMS (3–5 per week) + access to a website containing video messages (three per week), CR information, self-monitoring of PA on website Usual care (encouragement to attend CR) Duration and follow-up: 24 weeks Yes (physical activity) Park 2014;27 Park 201528 90 adults with CHD, hospitalised for MI or PCI Mean age (SD): 59 (9.4) Male: 68/90 (76%) Country: USA Three-arm prospective RCT Device: mobile phone Medium: SMS To examine the efficacy of a mHealth intervention to improve adherence to antiplatelet and statin medications among MI and/or PCI patients; to compare medication self-efficacy Arm 1: received two-way SMS medication reminders (requiring response) and one-way health education SMS (74 SMS messages over 30 days). Arm 2: received one-way health education SMS alone (14 messages over 30 days). Arm 3: Control group (no treatment) Duration and follow-up: 30 days Yes (medication adherence) Quilici 201329 521 patients who had coronary stenting for ACS Mean age (SD): 64 (10) Male: 382/499 (77%) Country: not stated but author affiliation is France RCT pilot study Device: mobile phone Medium: SMS To test whether participants receiving daily motivational SMS had better aspirin completion rates compared with standard care Intervention group received one daily personalised and motivational SMS reminder to take aspirin for 30 days Standard care (not described) Duration and follow-up: one month Yes (medication adherence) Varnfield 201430 120 post MI patients referred to CR Mean age (SD): 56 (10) Male: 82/94 (87%) (94/120 completed baseline) Country: Australia Two-group parallel RCT Device: smartphone Media: SMS, multimedia materials (video, audio), application (wellness diary, step counter) To investigate whether CAP-CR is effective in improving CR use in post-MI patients compared with a traditional centre based programme CAP-CR: smartphone for health and exercise monitoring via a wellness diary and step counter applications. Motivational and educational materials were delivered by SMS, audio and visual files. Participants received all core components of CR. Mentors reviewed participants' data via a web platform and provided weekly telephone consultations to feed back on goals. Standard care (traditional centre-based CR) Duration: six weeks; however, the CAP-CR group could use the smartphone and monitoring devices for six months. Follow-up: six weeks and six months. No (diet) CVD: cardiovascular disease; CI: confidence interval; RCT: randomised controlled trial; SMS: short-message service; ACS: acute coronary syndrome; CR: cardiac rehabilitation; IHD: ischaemic heart disease; MI: myocardial infarction; PA: physical activity; CHD: coronary heart disease; PCI: percutaneous coronary intervention; CAP: care assessment platform. Open in new tab Table 3. Characteristics of included studies. Study . Participants . Study design; device; media . Aim . Intervention . Comparator . Intervention duration; follow-up . Effect on behaviour change? . Antypas 201422 69 adults with CVD Mean age (95% CI) intervention: 60 (56–63) Control: 59 (56–62) Male: 52/67 (78%) Country: Norway Two-group cluster RCT Device: computer + mobile phone Media: website + SMS To assess the effect of a tailored Internet- and mobile phone-based intervention on the maintenance of physical activity levels after a cardiac rehabilitation stay Participants had access to website and received tailored messages via web and SMS. The website contained info about CVD and self-management, and an online discussion forum. Participants could set physical activity goals based on stage of change and received feedback via graph on website. They received SMS reminders about planned physical activity and if activity was completed. Access to non-tailored website and online forum Duration: not stated Follow-up: one and three months Yes (physical activity) Blasco 201223 203 adults with ACS with at least one CVD risk factor Mean age (SD) Intervention: 61 (12) Control: 61 (12) Male: 163/203 (80%) Country: Spain Single blind RCT Device; media: patient: mobile phone; SMS physician: computer; web application To analyse the efficacy of a telemonitoring system for the follow-up of patients with ACS The telemedicine group were provided with clinical measurement devices and sent results via their mobile phone. Physicians accessed the patient data via the web application and sent individualised SMS with recommendations to the patient. Usual care (met with cardiologist 3 × over study period; received verbal and written information about CVD prevention Duration and follow-up: 12 months No (smoking cessation) Khonsari 201524 62 adults with ACS Mean age (SD): 58 (13) Male: 53/62 (86%) Country: Malaysia Open-labelled RCT Device: mobile phone Medium: SMS To investigate the effect of automated SMS-based reminders on post-discharge medication adherence among patients with ACS Participants received automated SMS reminders before every medication intake. Also received SMS reminder for prescription refill after 30 days. Researcher telephoned participants once per fortnight to check in on SMS delivery and whether hospital readmission was needed. Usual care (CR and follow-up appointment with cardiologist) Duration and follow-up: Eight weeks Yes (medication adherence) Maddison 2015;25 201426 171 adults with IHD within previous 3–24 months Mean age (SD): 60 (9.3) Male: 139/171 (81%) Country: New Zealand Parallel Two-arm RCT Device: mobile phone + computer Media: SMS + website To determine the effectiveness and cost-effectiveness of a mHealth delivered exercise CR programme for people with IHD to improve exercise capacity and physical activity levels; to evaluate the mediating effects of self-efficacy on physical activity levels in the mHealth programme The intervention group received exercise prescription and physical activity behaviour change SMS (3–5 per week) + access to a website containing video messages (three per week), CR information, self-monitoring of PA on website Usual care (encouragement to attend CR) Duration and follow-up: 24 weeks Yes (physical activity) Park 2014;27 Park 201528 90 adults with CHD, hospitalised for MI or PCI Mean age (SD): 59 (9.4) Male: 68/90 (76%) Country: USA Three-arm prospective RCT Device: mobile phone Medium: SMS To examine the efficacy of a mHealth intervention to improve adherence to antiplatelet and statin medications among MI and/or PCI patients; to compare medication self-efficacy Arm 1: received two-way SMS medication reminders (requiring response) and one-way health education SMS (74 SMS messages over 30 days). Arm 2: received one-way health education SMS alone (14 messages over 30 days). Arm 3: Control group (no treatment) Duration and follow-up: 30 days Yes (medication adherence) Quilici 201329 521 patients who had coronary stenting for ACS Mean age (SD): 64 (10) Male: 382/499 (77%) Country: not stated but author affiliation is France RCT pilot study Device: mobile phone Medium: SMS To test whether participants receiving daily motivational SMS had better aspirin completion rates compared with standard care Intervention group received one daily personalised and motivational SMS reminder to take aspirin for 30 days Standard care (not described) Duration and follow-up: one month Yes (medication adherence) Varnfield 201430 120 post MI patients referred to CR Mean age (SD): 56 (10) Male: 82/94 (87%) (94/120 completed baseline) Country: Australia Two-group parallel RCT Device: smartphone Media: SMS, multimedia materials (video, audio), application (wellness diary, step counter) To investigate whether CAP-CR is effective in improving CR use in post-MI patients compared with a traditional centre based programme CAP-CR: smartphone for health and exercise monitoring via a wellness diary and step counter applications. Motivational and educational materials were delivered by SMS, audio and visual files. Participants received all core components of CR. Mentors reviewed participants' data via a web platform and provided weekly telephone consultations to feed back on goals. Standard care (traditional centre-based CR) Duration: six weeks; however, the CAP-CR group could use the smartphone and monitoring devices for six months. Follow-up: six weeks and six months. No (diet) Study . Participants . Study design; device; media . Aim . Intervention . Comparator . Intervention duration; follow-up . Effect on behaviour change? . Antypas 201422 69 adults with CVD Mean age (95% CI) intervention: 60 (56–63) Control: 59 (56–62) Male: 52/67 (78%) Country: Norway Two-group cluster RCT Device: computer + mobile phone Media: website + SMS To assess the effect of a tailored Internet- and mobile phone-based intervention on the maintenance of physical activity levels after a cardiac rehabilitation stay Participants had access to website and received tailored messages via web and SMS. The website contained info about CVD and self-management, and an online discussion forum. Participants could set physical activity goals based on stage of change and received feedback via graph on website. They received SMS reminders about planned physical activity and if activity was completed. Access to non-tailored website and online forum Duration: not stated Follow-up: one and three months Yes (physical activity) Blasco 201223 203 adults with ACS with at least one CVD risk factor Mean age (SD) Intervention: 61 (12) Control: 61 (12) Male: 163/203 (80%) Country: Spain Single blind RCT Device; media: patient: mobile phone; SMS physician: computer; web application To analyse the efficacy of a telemonitoring system for the follow-up of patients with ACS The telemedicine group were provided with clinical measurement devices and sent results via their mobile phone. Physicians accessed the patient data via the web application and sent individualised SMS with recommendations to the patient. Usual care (met with cardiologist 3 × over study period; received verbal and written information about CVD prevention Duration and follow-up: 12 months No (smoking cessation) Khonsari 201524 62 adults with ACS Mean age (SD): 58 (13) Male: 53/62 (86%) Country: Malaysia Open-labelled RCT Device: mobile phone Medium: SMS To investigate the effect of automated SMS-based reminders on post-discharge medication adherence among patients with ACS Participants received automated SMS reminders before every medication intake. Also received SMS reminder for prescription refill after 30 days. Researcher telephoned participants once per fortnight to check in on SMS delivery and whether hospital readmission was needed. Usual care (CR and follow-up appointment with cardiologist) Duration and follow-up: Eight weeks Yes (medication adherence) Maddison 2015;25 201426 171 adults with IHD within previous 3–24 months Mean age (SD): 60 (9.3) Male: 139/171 (81%) Country: New Zealand Parallel Two-arm RCT Device: mobile phone + computer Media: SMS + website To determine the effectiveness and cost-effectiveness of a mHealth delivered exercise CR programme for people with IHD to improve exercise capacity and physical activity levels; to evaluate the mediating effects of self-efficacy on physical activity levels in the mHealth programme The intervention group received exercise prescription and physical activity behaviour change SMS (3–5 per week) + access to a website containing video messages (three per week), CR information, self-monitoring of PA on website Usual care (encouragement to attend CR) Duration and follow-up: 24 weeks Yes (physical activity) Park 2014;27 Park 201528 90 adults with CHD, hospitalised for MI or PCI Mean age (SD): 59 (9.4) Male: 68/90 (76%) Country: USA Three-arm prospective RCT Device: mobile phone Medium: SMS To examine the efficacy of a mHealth intervention to improve adherence to antiplatelet and statin medications among MI and/or PCI patients; to compare medication self-efficacy Arm 1: received two-way SMS medication reminders (requiring response) and one-way health education SMS (74 SMS messages over 30 days). Arm 2: received one-way health education SMS alone (14 messages over 30 days). Arm 3: Control group (no treatment) Duration and follow-up: 30 days Yes (medication adherence) Quilici 201329 521 patients who had coronary stenting for ACS Mean age (SD): 64 (10) Male: 382/499 (77%) Country: not stated but author affiliation is France RCT pilot study Device: mobile phone Medium: SMS To test whether participants receiving daily motivational SMS had better aspirin completion rates compared with standard care Intervention group received one daily personalised and motivational SMS reminder to take aspirin for 30 days Standard care (not described) Duration and follow-up: one month Yes (medication adherence) Varnfield 201430 120 post MI patients referred to CR Mean age (SD): 56 (10) Male: 82/94 (87%) (94/120 completed baseline) Country: Australia Two-group parallel RCT Device: smartphone Media: SMS, multimedia materials (video, audio), application (wellness diary, step counter) To investigate whether CAP-CR is effective in improving CR use in post-MI patients compared with a traditional centre based programme CAP-CR: smartphone for health and exercise monitoring via a wellness diary and step counter applications. Motivational and educational materials were delivered by SMS, audio and visual files. Participants received all core components of CR. Mentors reviewed participants' data via a web platform and provided weekly telephone consultations to feed back on goals. Standard care (traditional centre-based CR) Duration: six weeks; however, the CAP-CR group could use the smartphone and monitoring devices for six months. Follow-up: six weeks and six months. No (diet) CVD: cardiovascular disease; CI: confidence interval; RCT: randomised controlled trial; SMS: short-message service; ACS: acute coronary syndrome; CR: cardiac rehabilitation; IHD: ischaemic heart disease; MI: myocardial infarction; PA: physical activity; CHD: coronary heart disease; PCI: percutaneous coronary intervention; CAP: care assessment platform. Open in new tab Participants and design Participants in each included study were relatively homogenous in terms of their disease (coronary heart disease), mean age (late 50s to early 60s), and gender (>75% male). All participants were required to be somewhat familiar with the technology used to deliver the intervention, either by owning a mobile phone22–24,27,29,30 and/or having access to the Internet.22,25 Mobile phones were provided to participants if needed in one study.25 Sample sizes ranged from 69 to 521 and included a total of 1236 participants. Studies were conducted in Europe,22,23,29 North America,27 Asia24 and Australasia.25,30 All seven studies used a two-arm randomised controlled trial (RCT) design, with the exception of Park et al.,27 who conducted a three arm RCT. Four studies had behaviour change as primary outcomes: one intervened on and measured physical activity22 and three targeted adherence to medication.24,27,29 Maddison and colleagues'25 intervention aimed to increase exercise capacity with physical activity as a secondary outcome. Varnfield and colleagues'30 intervention targeted all core components of cardiac rehabilitation and measured dietary habits (including alcohol) as a secondary outcome. Blasco and colleagues'23 intervention monitored cardiac risk factors, which included smoking cessation as a secondary outcome. More detail on study outcomes can be found below. Intervention Six interventions were delivered by any type of mobile phone capable of receiving SMS, and one intervention was delivered specifically by smartphone.30 Every study used SMS to deliver health education, recommendations or motivational messages. Three studies used SMS alone to encourage medication adherence.24,27,29 Two studies used SMS in combination with a website to increase physical activity participation: Antypas and Wangberg's22 intervention was predominantly delivered via the web with SMS reminder to engage in planned physical activities; Maddison et al.25 delivered exercise prescription and behaviour change support via SMS with additional information and support provided on a website. The remaining two studies used remote monitoring of participants' clinical data entered on mobile phones, where physicians or mentors viewed participant data via a web application and provided feedback to the participant by SMS.23,30 One of these (Varnfield et al.30) also included a smartphone application and audio/visual files to deliver educational and motivational materials. Intervention dose and duration varied considerably and was not always reported. Intervention duration ranged from 30 days27,29 to 12 months.23 Only four studies specified the number of SMS messages sent: 3–5 per week for 24 weeks,25 a total of 74 or 14 over 30 days,27 daily for 30 days,29 and one message before every medication intake over eight weeks.24 Two interventions used automated two-way SMS (i.e. interactive communication between participant and provider) to monitor behaviour22,27 while the remainder used one-way SMS (messages pushed to participant without a response expected). Each intervention was described as ‘personalised’; however, not all described how this was done. Personalisation ranged from including the participant's name in the SMS,24 tailoring to the time of day participants’ wished to receive messages27 to offering personalised feedback based on achieving behavioural goals25,30 or clinical outcomes.23 Three studies used a theory to frame their intervention: the transtheoretical model in one22 and self-efficacy theory in the other two.25,27 While not always mentioned, specific techniques, or BCTs, were also used to support behaviour change.31 As part of our review, the first author (LPD) coded the interventions according to BCTs as described in each article. The findings are provided in Table 4. The most commonly used BCTs included goal setting, self-monitoring of behaviour, information about health consequences, and prompts/cues. Three studies used seven or more BCTS,22,25,30 while two studies appeared to use only one BTC (prompts to take medications).24,29 Table 4. Behaviour change techniques used in interventions included in the review.31 Behaviour change technique . Antypas 201422 . Blasco 201223 . Khonsari 201524 . Maddison 2015;25 201426 . Park 2014;27 201528 . Quilici 201329 . Varnfield 201430 . 1.1 Goal setting (behaviour) ✓ ✓ ✓ 1.4 Action planning ✓ 1.5 Review behavioural goal(s) ✓ 2.1 Monitoring of behaviour by others without feedback ✓ ✓ 2.2 Feedback on behaviour ✓ ✓ 2.3 Self-monitoring of behaviour ✓ ✓ ✓ ✓ 2.4 Self-monitoring of outcome(s) of behaviour ✓ ✓ 2.5 Monitoring outcome(s) of behaviour by others without feedback 2.6 Biofeedback ✓ ✓ 2.7 Feedback on outcomes of behaviour ✓ ✓ 3.1 Social support (unspecified) ✓ ✓ 4.1 Instruction on how to perform a behaviour ✓ 5.1 Information about health consequences ✓ ✓ ✓ ✓ ✓ 5.6 Information about emotional consequences ✓ 6.1 Demonstration of the behaviour ✓ 7.1 Prompts/cues ✓ ✓ ✓ ✓ 8.1 Behavioural practice/rehearsal ✓ 11.2 Reduce negative emotions ✓ Behaviour change technique . Antypas 201422 . Blasco 201223 . Khonsari 201524 . Maddison 2015;25 201426 . Park 2014;27 201528 . Quilici 201329 . Varnfield 201430 . 1.1 Goal setting (behaviour) ✓ ✓ ✓ 1.4 Action planning ✓ 1.5 Review behavioural goal(s) ✓ 2.1 Monitoring of behaviour by others without feedback ✓ ✓ 2.2 Feedback on behaviour ✓ ✓ 2.3 Self-monitoring of behaviour ✓ ✓ ✓ ✓ 2.4 Self-monitoring of outcome(s) of behaviour ✓ ✓ 2.5 Monitoring outcome(s) of behaviour by others without feedback 2.6 Biofeedback ✓ ✓ 2.7 Feedback on outcomes of behaviour ✓ ✓ 3.1 Social support (unspecified) ✓ ✓ 4.1 Instruction on how to perform a behaviour ✓ 5.1 Information about health consequences ✓ ✓ ✓ ✓ ✓ 5.6 Information about emotional consequences ✓ 6.1 Demonstration of the behaviour ✓ 7.1 Prompts/cues ✓ ✓ ✓ ✓ 8.1 Behavioural practice/rehearsal ✓ 11.2 Reduce negative emotions ✓ Open in new tab Table 4. Behaviour change techniques used in interventions included in the review.31 Behaviour change technique . Antypas 201422 . Blasco 201223 . Khonsari 201524 . Maddison 2015;25 201426 . Park 2014;27 201528 . Quilici 201329 . Varnfield 201430 . 1.1 Goal setting (behaviour) ✓ ✓ ✓ 1.4 Action planning ✓ 1.5 Review behavioural goal(s) ✓ 2.1 Monitoring of behaviour by others without feedback ✓ ✓ 2.2 Feedback on behaviour ✓ ✓ 2.3 Self-monitoring of behaviour ✓ ✓ ✓ ✓ 2.4 Self-monitoring of outcome(s) of behaviour ✓ ✓ 2.5 Monitoring outcome(s) of behaviour by others without feedback 2.6 Biofeedback ✓ ✓ 2.7 Feedback on outcomes of behaviour ✓ ✓ 3.1 Social support (unspecified) ✓ ✓ 4.1 Instruction on how to perform a behaviour ✓ 5.1 Information about health consequences ✓ ✓ ✓ ✓ ✓ 5.6 Information about emotional consequences ✓ 6.1 Demonstration of the behaviour ✓ 7.1 Prompts/cues ✓ ✓ ✓ ✓ 8.1 Behavioural practice/rehearsal ✓ 11.2 Reduce negative emotions ✓ Behaviour change technique . Antypas 201422 . Blasco 201223 . Khonsari 201524 . Maddison 2015;25 201426 . Park 2014;27 201528 . Quilici 201329 . Varnfield 201430 . 1.1 Goal setting (behaviour) ✓ ✓ ✓ 1.4 Action planning ✓ 1.5 Review behavioural goal(s) ✓ 2.1 Monitoring of behaviour by others without feedback ✓ ✓ 2.2 Feedback on behaviour ✓ ✓ 2.3 Self-monitoring of behaviour ✓ ✓ ✓ ✓ 2.4 Self-monitoring of outcome(s) of behaviour ✓ ✓ 2.5 Monitoring outcome(s) of behaviour by others without feedback 2.6 Biofeedback ✓ ✓ 2.7 Feedback on outcomes of behaviour ✓ ✓ 3.1 Social support (unspecified) ✓ ✓ 4.1 Instruction on how to perform a behaviour ✓ 5.1 Information about health consequences ✓ ✓ ✓ ✓ ✓ 5.6 Information about emotional consequences ✓ 6.1 Demonstration of the behaviour ✓ 7.1 Prompts/cues ✓ ✓ ✓ ✓ 8.1 Behavioural practice/rehearsal ✓ 11.2 Reduce negative emotions ✓ Open in new tab Comparison Five studies compared their interventions with standard/usual care. Usual care included encouragement to attend or attending traditional centre-based cardiac rehabilitation,24,25,30 visits to cardiologists to discuss CVD secondary prevention,23 or was not described.29 One three-arm study compared their health education and medication reminder SMS with health education SMS alone and to a control (no SMS).27 The remaining study compared a tailored website with SMS with a non-tailored website.22 Quality and risk of bias Table 5 describes risk of bias and study quality. Four of the seven included studies had a low risk of selection bias. The remaining three studies23,24,29 did not describe their random allocation methods. Detection bias was assessed based on single blinding of outcome assessors, as the nature of mHealth interventions renders it difficult for participants to be blinded. Three studies had a low risk of detection bias, including one study22 which also blinded participants to treatment condition. Three studies were high risk as the studies were unblinded24,25,27 and one was unclear as blinding was not described.29 Table 5. Risk of bias. . . Selection bias . Detection bias . Attrition bias . Reporting bias . Study . Jadad score (/5) . Random sequence generation . Allocation concealment . Blinding of participant, personnel, assessors . Attrition . Selective outcome reporting . . . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Antypas 201422 5 Low risk True random number online service Low risk Concealed until participant entered code online Low risk Participants, investigators and outcome assessors blinded High risk High rates of attrition. 50/69 (72%) lost to follow-up; uneven among groups. High risk Some outcomes reported in protocol are missing from results Blasco 201223 4 Not clear Method not described Not clear Method not described Low risk Outcome assessors blinded Low risk Low attrition reported. 26/203 (13%) withdrew or lost to follow-up, similar across groups. High risk Some outcomes reported in results but not pre-specified in methods (no protocol or trial registration) Khonsari 201524 2 Not clear Method not described Not clear Method not described High risk Primary outcome assessors not blinded Low risk Low attrition reported. 2/62 (3%) lost to follow-up (death). Low risk Nothing to suggest selective reporting (but no protocol or trial registration) Maddison 2015;25 201426 5 Low risk Computer randomisation Low risk Minimisation sequence Low risk Outcome assessors blinded Low risk Low attrition reported. 18/171 (11%) lost to follow-up, similar across groups. High risk One secondary outcome in protocol not reported in results Park 2014;27 201528 3 Low risk Random sequence in blocks of six Low risk Consecutive opaque sealed envelopes High risk Outcome assessors not blinded Low risk Low attrition reported. 6/90 (7%) lost to follow-up, similar across groups. Low risk Nothing to suggest selective reporting, but outcomes reported over two papers (no protocol or trial registration) Quilici 201329 2 Not clear Method not described Not clear Method not described Not clear Method not described Low risk Low attrition reported. 22/521 (4%) withdrew, similar across groups. Low risk Nothing to suggest selective reporting (but no protocol or trial registration) Varnfield 201430 2 Low risk Computer generated permuted-block randomisation (block sizes of four, six, eight) Low risk Sequentially numbered opaque sealed envelopes High risk Outcome assessors not blinded High risk High rates of attrition (only 76/120; 63% assessed at six weeks; not split evenly over groups. Even greater dropout for secondary outcomes. High risk Primary outcome changed from protocol to results paper; some secondary outcomes not reported in results . . Selection bias . Detection bias . Attrition bias . Reporting bias . Study . Jadad score (/5) . Random sequence generation . Allocation concealment . Blinding of participant, personnel, assessors . Attrition . Selective outcome reporting . . . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Antypas 201422 5 Low risk True random number online service Low risk Concealed until participant entered code online Low risk Participants, investigators and outcome assessors blinded High risk High rates of attrition. 50/69 (72%) lost to follow-up; uneven among groups. High risk Some outcomes reported in protocol are missing from results Blasco 201223 4 Not clear Method not described Not clear Method not described Low risk Outcome assessors blinded Low risk Low attrition reported. 26/203 (13%) withdrew or lost to follow-up, similar across groups. High risk Some outcomes reported in results but not pre-specified in methods (no protocol or trial registration) Khonsari 201524 2 Not clear Method not described Not clear Method not described High risk Primary outcome assessors not blinded Low risk Low attrition reported. 2/62 (3%) lost to follow-up (death). Low risk Nothing to suggest selective reporting (but no protocol or trial registration) Maddison 2015;25 201426 5 Low risk Computer randomisation Low risk Minimisation sequence Low risk Outcome assessors blinded Low risk Low attrition reported. 18/171 (11%) lost to follow-up, similar across groups. High risk One secondary outcome in protocol not reported in results Park 2014;27 201528 3 Low risk Random sequence in blocks of six Low risk Consecutive opaque sealed envelopes High risk Outcome assessors not blinded Low risk Low attrition reported. 6/90 (7%) lost to follow-up, similar across groups. Low risk Nothing to suggest selective reporting, but outcomes reported over two papers (no protocol or trial registration) Quilici 201329 2 Not clear Method not described Not clear Method not described Not clear Method not described Low risk Low attrition reported. 22/521 (4%) withdrew, similar across groups. Low risk Nothing to suggest selective reporting (but no protocol or trial registration) Varnfield 201430 2 Low risk Computer generated permuted-block randomisation (block sizes of four, six, eight) Low risk Sequentially numbered opaque sealed envelopes High risk Outcome assessors not blinded High risk High rates of attrition (only 76/120; 63% assessed at six weeks; not split evenly over groups. Even greater dropout for secondary outcomes. High risk Primary outcome changed from protocol to results paper; some secondary outcomes not reported in results Open in new tab Table 5. Risk of bias. . . Selection bias . Detection bias . Attrition bias . Reporting bias . Study . Jadad score (/5) . Random sequence generation . Allocation concealment . Blinding of participant, personnel, assessors . Attrition . Selective outcome reporting . . . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Antypas 201422 5 Low risk True random number online service Low risk Concealed until participant entered code online Low risk Participants, investigators and outcome assessors blinded High risk High rates of attrition. 50/69 (72%) lost to follow-up; uneven among groups. High risk Some outcomes reported in protocol are missing from results Blasco 201223 4 Not clear Method not described Not clear Method not described Low risk Outcome assessors blinded Low risk Low attrition reported. 26/203 (13%) withdrew or lost to follow-up, similar across groups. High risk Some outcomes reported in results but not pre-specified in methods (no protocol or trial registration) Khonsari 201524 2 Not clear Method not described Not clear Method not described High risk Primary outcome assessors not blinded Low risk Low attrition reported. 2/62 (3%) lost to follow-up (death). Low risk Nothing to suggest selective reporting (but no protocol or trial registration) Maddison 2015;25 201426 5 Low risk Computer randomisation Low risk Minimisation sequence Low risk Outcome assessors blinded Low risk Low attrition reported. 18/171 (11%) lost to follow-up, similar across groups. High risk One secondary outcome in protocol not reported in results Park 2014;27 201528 3 Low risk Random sequence in blocks of six Low risk Consecutive opaque sealed envelopes High risk Outcome assessors not blinded Low risk Low attrition reported. 6/90 (7%) lost to follow-up, similar across groups. Low risk Nothing to suggest selective reporting, but outcomes reported over two papers (no protocol or trial registration) Quilici 201329 2 Not clear Method not described Not clear Method not described Not clear Method not described Low risk Low attrition reported. 22/521 (4%) withdrew, similar across groups. Low risk Nothing to suggest selective reporting (but no protocol or trial registration) Varnfield 201430 2 Low risk Computer generated permuted-block randomisation (block sizes of four, six, eight) Low risk Sequentially numbered opaque sealed envelopes High risk Outcome assessors not blinded High risk High rates of attrition (only 76/120; 63% assessed at six weeks; not split evenly over groups. Even greater dropout for secondary outcomes. High risk Primary outcome changed from protocol to results paper; some secondary outcomes not reported in results . . Selection bias . Detection bias . Attrition bias . Reporting bias . Study . Jadad score (/5) . Random sequence generation . Allocation concealment . Blinding of participant, personnel, assessors . Attrition . Selective outcome reporting . . . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Judgement . Evidence . Antypas 201422 5 Low risk True random number online service Low risk Concealed until participant entered code online Low risk Participants, investigators and outcome assessors blinded High risk High rates of attrition. 50/69 (72%) lost to follow-up; uneven among groups. High risk Some outcomes reported in protocol are missing from results Blasco 201223 4 Not clear Method not described Not clear Method not described Low risk Outcome assessors blinded Low risk Low attrition reported. 26/203 (13%) withdrew or lost to follow-up, similar across groups. High risk Some outcomes reported in results but not pre-specified in methods (no protocol or trial registration) Khonsari 201524 2 Not clear Method not described Not clear Method not described High risk Primary outcome assessors not blinded Low risk Low attrition reported. 2/62 (3%) lost to follow-up (death). Low risk Nothing to suggest selective reporting (but no protocol or trial registration) Maddison 2015;25 201426 5 Low risk Computer randomisation Low risk Minimisation sequence Low risk Outcome assessors blinded Low risk Low attrition reported. 18/171 (11%) lost to follow-up, similar across groups. High risk One secondary outcome in protocol not reported in results Park 2014;27 201528 3 Low risk Random sequence in blocks of six Low risk Consecutive opaque sealed envelopes High risk Outcome assessors not blinded Low risk Low attrition reported. 6/90 (7%) lost to follow-up, similar across groups. Low risk Nothing to suggest selective reporting, but outcomes reported over two papers (no protocol or trial registration) Quilici 201329 2 Not clear Method not described Not clear Method not described Not clear Method not described Low risk Low attrition reported. 22/521 (4%) withdrew, similar across groups. Low risk Nothing to suggest selective reporting (but no protocol or trial registration) Varnfield 201430 2 Low risk Computer generated permuted-block randomisation (block sizes of four, six, eight) Low risk Sequentially numbered opaque sealed envelopes High risk Outcome assessors not blinded High risk High rates of attrition (only 76/120; 63% assessed at six weeks; not split evenly over groups. Even greater dropout for secondary outcomes. High risk Primary outcome changed from protocol to results paper; some secondary outcomes not reported in results Open in new tab Power calculations were reported in five studies.22–25,30 All seven studies reported the number of participants randomised and analysed; five had low risk of attrition bias as few participants withdrew or were lost to follow-up. Two studies22,30 had high rates of attrition split unevenly across groups at follow-up and failed to recruit enough participants as specified in protocol power calculations. Intention to treat analysis was used in five studies.23–25,27,30 Reporting bias occurred in four studies, as some outcomes reported in the protocol or methods section were not included in the results,22,25 or reported findings were not pre-specified in the measures section.23 One study changed the primary outcome due to low recruitment rates.30 Overall no study was completely free of bias, but three studies were classified as high quality (≥4 out of 5) according to the Jadad score23,25,27 (see Supplementary file 3 for details). Effectiveness of interventions on behaviour change Table 6 describes study outcomes. Five of seven studies found a positive treatment effect on behaviour change. Both physical activity studies saw an increase in self-reported physical activity.22,25 All three medication adherence studies found a treatment effect;24,27,29 however, Park et al.27 found an improvement in adherence to antiplatelet medication but no effect was seen on statin medication. No treatment effects were observed on dietary habits and alcohol30 or smoking cessation.23 Table 6. Outcomes of included studies. Study . Arms . Behaviour change outcome . Additional outcomes . Process outcomes/adherence to intervention . Antypas 201422 Arm 1: tailored website and SMS PA reminders Arm 2: non-tailored website, no SMS Three month: tailored group: significantly higher median PA score than control (5613 vs. 1356, p = .02)a No differences in moderate-to-vigorous PA, vigorous PA, or sitting No differences between groups for self-efficacy, social support, anxiety, depression at one month or three months No differences in number of website visits between groups Email and SMS reminders and messages were least useful elements of the intervention; activity calendar was the most useful Blasco 201223 Arm 1: telemonitoring group, SMS recommendations from physician to patient based on clinical data Arm 2: usual care No differences in smoking cessation TMG more likely to improve cardiac risk factor profilea (RR = 1.4; 95% CI = 1.1-1.7, p = .01); achieve BP goal (RR = 1.4; 95% CI = 1.1–1.9, p = .012); HbA1c goal (RR = 1.6; 95% CI = 1.11–2.4, p = .018) No differences in LDL, HRQoL or anxiety 89% (SD 16) of remote monitoring sessions were completed 0.5 messages per patient were missed due to phone turned off Khonsari 201524 Arm 1: SMS medication reminder Arm 2: usual care Sig higher medication adherence in intervention group (x2(2) = 18.614,p < 0.001); RR of being low adherent in control group = 4.09, 95% CI 1.82-9.18* % with no symptoms (Heart functional status): 84% intervention vs. 32% control (x2(1) = 16.957,p<0.001). No sig differences in ACS readmissions or deaths 93.5% found intervention useful and 64.5% felt it helped them take their medications; 80.6% wanted SMS reminder continued and 83.9% would recommend programme to other patients Maddison 2015;25 201426 Arm 1: SMS PA prescription and behaviour strategies + website Arm 2: usual care Intervention group significantly increased leisure time PA (110 min/week; 40%) and walking (151 min/week; 42%), but not moderate-to-vigorous PA, vigorous PA No significant differences in exercise capacitya. Intervention group significantly increased self-efficacy (6%) and general health domain of SF36. Change in task self-efficacy mediated treatment effect on leisure time PA but not on walking. Likely cost-effective. No differences in weight, waist–hip ratio, BP 82% of int participants read some or all of their SMS; 57% viewed some or all of the video messages on the website. Participants logged on an average of once per fortnight. Park 2014;27 201528 Arm 1: TM Reminder (TM R) + TM Education (TM E) received two-way SMS medication reminders and one-way health education SMS Arm 2: TM Education group received one-way health education SMS alone Arm 3: control Objective medication adherence (electronic pill bottle)a: TM R + TM E and TM E alone both had significantly better % of prescribed doses taken (F(2,42) = 3.84, p = .03); correct doses taken (F(2,41) = 3.29, p = .047); prescribed doses taken on schedule (F(2,41) = 3.53, p = .04); than control using for antiplatelets (no differences between TM groups). No differences for statins. Self-reported medication adherence: no differences between all three groups. No sig differences in medication self-efficacy. Lower levels of depression and higher social support positively predicted medication adherence Response rate to two-way SMS for TM R + TM E was higher for antiplatelets than statins Both TM groups reported high satisfaction with receiving TM Both groups felt with TM: ‘felt someone cared by receiving TM’ Few technical issues Quilici 201329 Arm 1: daily SMS reminding aspirin intake Arm 2: standard care SMS group had significantly better adherence than control: platelet testing – non-adherence in SMS 5.2% vs. 11.2% in control, OR (95% CI): 0.43 (0.22–0.86); p = .01)a SMS greater self-reported adherence: OR (95% CI): 0.37 (0.15–0.90); p = .02) N/A 92% of intervention group thought SMS support was valuable Varnfield 201430 Arm 1: CAP-CR used smartphone app to monitor wellness and exercise. Received educational/motivational SMS, video and audio messages Arm 2: traditional CR (TCR) At six weeks: no differences in nutrition (fat, fibre, sodium, alcohol) At six weeks: significant differences between groups for diastolic BP, triglycerides, EQ5D-Index. No differences in 6MWT, mental health, systolic BP, heart rate, weight, body mass index, waist circumference, lipids At six months: no differences between groups for 6MWT, LDL, HDL, HRQoL, or psychological distress At six weeks: uptake (RR 1.3), adherence (RR 1.4) and completion (RR 1.7) was higher in CAP-CR than TCRa 85% of participants found step counter to be motivating Daily exercise entries recorded by 89% of CAP-CR adherers (n = 45) N = 3 dropout from CAP-CR due to difficulty with technology Study . Arms . Behaviour change outcome . Additional outcomes . Process outcomes/adherence to intervention . Antypas 201422 Arm 1: tailored website and SMS PA reminders Arm 2: non-tailored website, no SMS Three month: tailored group: significantly higher median PA score than control (5613 vs. 1356, p = .02)a No differences in moderate-to-vigorous PA, vigorous PA, or sitting No differences between groups for self-efficacy, social support, anxiety, depression at one month or three months No differences in number of website visits between groups Email and SMS reminders and messages were least useful elements of the intervention; activity calendar was the most useful Blasco 201223 Arm 1: telemonitoring group, SMS recommendations from physician to patient based on clinical data Arm 2: usual care No differences in smoking cessation TMG more likely to improve cardiac risk factor profilea (RR = 1.4; 95% CI = 1.1-1.7, p = .01); achieve BP goal (RR = 1.4; 95% CI = 1.1–1.9, p = .012); HbA1c goal (RR = 1.6; 95% CI = 1.11–2.4, p = .018) No differences in LDL, HRQoL or anxiety 89% (SD 16) of remote monitoring sessions were completed 0.5 messages per patient were missed due to phone turned off Khonsari 201524 Arm 1: SMS medication reminder Arm 2: usual care Sig higher medication adherence in intervention group (x2(2) = 18.614,p < 0.001); RR of being low adherent in control group = 4.09, 95% CI 1.82-9.18* % with no symptoms (Heart functional status): 84% intervention vs. 32% control (x2(1) = 16.957,p<0.001). No sig differences in ACS readmissions or deaths 93.5% found intervention useful and 64.5% felt it helped them take their medications; 80.6% wanted SMS reminder continued and 83.9% would recommend programme to other patients Maddison 2015;25 201426 Arm 1: SMS PA prescription and behaviour strategies + website Arm 2: usual care Intervention group significantly increased leisure time PA (110 min/week; 40%) and walking (151 min/week; 42%), but not moderate-to-vigorous PA, vigorous PA No significant differences in exercise capacitya. Intervention group significantly increased self-efficacy (6%) and general health domain of SF36. Change in task self-efficacy mediated treatment effect on leisure time PA but not on walking. Likely cost-effective. No differences in weight, waist–hip ratio, BP 82% of int participants read some or all of their SMS; 57% viewed some or all of the video messages on the website. Participants logged on an average of once per fortnight. Park 2014;27 201528 Arm 1: TM Reminder (TM R) + TM Education (TM E) received two-way SMS medication reminders and one-way health education SMS Arm 2: TM Education group received one-way health education SMS alone Arm 3: control Objective medication adherence (electronic pill bottle)a: TM R + TM E and TM E alone both had significantly better % of prescribed doses taken (F(2,42) = 3.84, p = .03); correct doses taken (F(2,41) = 3.29, p = .047); prescribed doses taken on schedule (F(2,41) = 3.53, p = .04); than control using for antiplatelets (no differences between TM groups). No differences for statins. Self-reported medication adherence: no differences between all three groups. No sig differences in medication self-efficacy. Lower levels of depression and higher social support positively predicted medication adherence Response rate to two-way SMS for TM R + TM E was higher for antiplatelets than statins Both TM groups reported high satisfaction with receiving TM Both groups felt with TM: ‘felt someone cared by receiving TM’ Few technical issues Quilici 201329 Arm 1: daily SMS reminding aspirin intake Arm 2: standard care SMS group had significantly better adherence than control: platelet testing – non-adherence in SMS 5.2% vs. 11.2% in control, OR (95% CI): 0.43 (0.22–0.86); p = .01)a SMS greater self-reported adherence: OR (95% CI): 0.37 (0.15–0.90); p = .02) N/A 92% of intervention group thought SMS support was valuable Varnfield 201430 Arm 1: CAP-CR used smartphone app to monitor wellness and exercise. Received educational/motivational SMS, video and audio messages Arm 2: traditional CR (TCR) At six weeks: no differences in nutrition (fat, fibre, sodium, alcohol) At six weeks: significant differences between groups for diastolic BP, triglycerides, EQ5D-Index. No differences in 6MWT, mental health, systolic BP, heart rate, weight, body mass index, waist circumference, lipids At six months: no differences between groups for 6MWT, LDL, HDL, HRQoL, or psychological distress At six weeks: uptake (RR 1.3), adherence (RR 1.4) and completion (RR 1.7) was higher in CAP-CR than TCRa 85% of participants found step counter to be motivating Daily exercise entries recorded by 89% of CAP-CR adherers (n = 45) N = 3 dropout from CAP-CR due to difficulty with technology SMS: short-message service; PA: physical activity; TMG: telemonitoring group; RR: relative risk; CI: confidence interval; BP: blood pressure; HbA1C: glycosylated haemoglobin level; LDL: low-density lipoprotein; HRQoL: health related quality of life; ACS: acute coronary syndrome; SF36: short form 36 health survey; CAP: care assessment platform; CR: cardiac rehabilitation; EQ5D: health status questionnaire; 6MWT: six minute walk test; HDL: high-density lipoprotein a Primary outcome. Open in new tab Table 6. Outcomes of included studies. Study . Arms . Behaviour change outcome . Additional outcomes . Process outcomes/adherence to intervention . Antypas 201422 Arm 1: tailored website and SMS PA reminders Arm 2: non-tailored website, no SMS Three month: tailored group: significantly higher median PA score than control (5613 vs. 1356, p = .02)a No differences in moderate-to-vigorous PA, vigorous PA, or sitting No differences between groups for self-efficacy, social support, anxiety, depression at one month or three months No differences in number of website visits between groups Email and SMS reminders and messages were least useful elements of the intervention; activity calendar was the most useful Blasco 201223 Arm 1: telemonitoring group, SMS recommendations from physician to patient based on clinical data Arm 2: usual care No differences in smoking cessation TMG more likely to improve cardiac risk factor profilea (RR = 1.4; 95% CI = 1.1-1.7, p = .01); achieve BP goal (RR = 1.4; 95% CI = 1.1–1.9, p = .012); HbA1c goal (RR = 1.6; 95% CI = 1.11–2.4, p = .018) No differences in LDL, HRQoL or anxiety 89% (SD 16) of remote monitoring sessions were completed 0.5 messages per patient were missed due to phone turned off Khonsari 201524 Arm 1: SMS medication reminder Arm 2: usual care Sig higher medication adherence in intervention group (x2(2) = 18.614,p < 0.001); RR of being low adherent in control group = 4.09, 95% CI 1.82-9.18* % with no symptoms (Heart functional status): 84% intervention vs. 32% control (x2(1) = 16.957,p<0.001). No sig differences in ACS readmissions or deaths 93.5% found intervention useful and 64.5% felt it helped them take their medications; 80.6% wanted SMS reminder continued and 83.9% would recommend programme to other patients Maddison 2015;25 201426 Arm 1: SMS PA prescription and behaviour strategies + website Arm 2: usual care Intervention group significantly increased leisure time PA (110 min/week; 40%) and walking (151 min/week; 42%), but not moderate-to-vigorous PA, vigorous PA No significant differences in exercise capacitya. Intervention group significantly increased self-efficacy (6%) and general health domain of SF36. Change in task self-efficacy mediated treatment effect on leisure time PA but not on walking. Likely cost-effective. No differences in weight, waist–hip ratio, BP 82% of int participants read some or all of their SMS; 57% viewed some or all of the video messages on the website. Participants logged on an average of once per fortnight. Park 2014;27 201528 Arm 1: TM Reminder (TM R) + TM Education (TM E) received two-way SMS medication reminders and one-way health education SMS Arm 2: TM Education group received one-way health education SMS alone Arm 3: control Objective medication adherence (electronic pill bottle)a: TM R + TM E and TM E alone both had significantly better % of prescribed doses taken (F(2,42) = 3.84, p = .03); correct doses taken (F(2,41) = 3.29, p = .047); prescribed doses taken on schedule (F(2,41) = 3.53, p = .04); than control using for antiplatelets (no differences between TM groups). No differences for statins. Self-reported medication adherence: no differences between all three groups. No sig differences in medication self-efficacy. Lower levels of depression and higher social support positively predicted medication adherence Response rate to two-way SMS for TM R + TM E was higher for antiplatelets than statins Both TM groups reported high satisfaction with receiving TM Both groups felt with TM: ‘felt someone cared by receiving TM’ Few technical issues Quilici 201329 Arm 1: daily SMS reminding aspirin intake Arm 2: standard care SMS group had significantly better adherence than control: platelet testing – non-adherence in SMS 5.2% vs. 11.2% in control, OR (95% CI): 0.43 (0.22–0.86); p = .01)a SMS greater self-reported adherence: OR (95% CI): 0.37 (0.15–0.90); p = .02) N/A 92% of intervention group thought SMS support was valuable Varnfield 201430 Arm 1: CAP-CR used smartphone app to monitor wellness and exercise. Received educational/motivational SMS, video and audio messages Arm 2: traditional CR (TCR) At six weeks: no differences in nutrition (fat, fibre, sodium, alcohol) At six weeks: significant differences between groups for diastolic BP, triglycerides, EQ5D-Index. No differences in 6MWT, mental health, systolic BP, heart rate, weight, body mass index, waist circumference, lipids At six months: no differences between groups for 6MWT, LDL, HDL, HRQoL, or psychological distress At six weeks: uptake (RR 1.3), adherence (RR 1.4) and completion (RR 1.7) was higher in CAP-CR than TCRa 85% of participants found step counter to be motivating Daily exercise entries recorded by 89% of CAP-CR adherers (n = 45) N = 3 dropout from CAP-CR due to difficulty with technology Study . Arms . Behaviour change outcome . Additional outcomes . Process outcomes/adherence to intervention . Antypas 201422 Arm 1: tailored website and SMS PA reminders Arm 2: non-tailored website, no SMS Three month: tailored group: significantly higher median PA score than control (5613 vs. 1356, p = .02)a No differences in moderate-to-vigorous PA, vigorous PA, or sitting No differences between groups for self-efficacy, social support, anxiety, depression at one month or three months No differences in number of website visits between groups Email and SMS reminders and messages were least useful elements of the intervention; activity calendar was the most useful Blasco 201223 Arm 1: telemonitoring group, SMS recommendations from physician to patient based on clinical data Arm 2: usual care No differences in smoking cessation TMG more likely to improve cardiac risk factor profilea (RR = 1.4; 95% CI = 1.1-1.7, p = .01); achieve BP goal (RR = 1.4; 95% CI = 1.1–1.9, p = .012); HbA1c goal (RR = 1.6; 95% CI = 1.11–2.4, p = .018) No differences in LDL, HRQoL or anxiety 89% (SD 16) of remote monitoring sessions were completed 0.5 messages per patient were missed due to phone turned off Khonsari 201524 Arm 1: SMS medication reminder Arm 2: usual care Sig higher medication adherence in intervention group (x2(2) = 18.614,p < 0.001); RR of being low adherent in control group = 4.09, 95% CI 1.82-9.18* % with no symptoms (Heart functional status): 84% intervention vs. 32% control (x2(1) = 16.957,p<0.001). No sig differences in ACS readmissions or deaths 93.5% found intervention useful and 64.5% felt it helped them take their medications; 80.6% wanted SMS reminder continued and 83.9% would recommend programme to other patients Maddison 2015;25 201426 Arm 1: SMS PA prescription and behaviour strategies + website Arm 2: usual care Intervention group significantly increased leisure time PA (110 min/week; 40%) and walking (151 min/week; 42%), but not moderate-to-vigorous PA, vigorous PA No significant differences in exercise capacitya. Intervention group significantly increased self-efficacy (6%) and general health domain of SF36. Change in task self-efficacy mediated treatment effect on leisure time PA but not on walking. Likely cost-effective. No differences in weight, waist–hip ratio, BP 82% of int participants read some or all of their SMS; 57% viewed some or all of the video messages on the website. Participants logged on an average of once per fortnight. Park 2014;27 201528 Arm 1: TM Reminder (TM R) + TM Education (TM E) received two-way SMS medication reminders and one-way health education SMS Arm 2: TM Education group received one-way health education SMS alone Arm 3: control Objective medication adherence (electronic pill bottle)a: TM R + TM E and TM E alone both had significantly better % of prescribed doses taken (F(2,42) = 3.84, p = .03); correct doses taken (F(2,41) = 3.29, p = .047); prescribed doses taken on schedule (F(2,41) = 3.53, p = .04); than control using for antiplatelets (no differences between TM groups). No differences for statins. Self-reported medication adherence: no differences between all three groups. No sig differences in medication self-efficacy. Lower levels of depression and higher social support positively predicted medication adherence Response rate to two-way SMS for TM R + TM E was higher for antiplatelets than statins Both TM groups reported high satisfaction with receiving TM Both groups felt with TM: ‘felt someone cared by receiving TM’ Few technical issues Quilici 201329 Arm 1: daily SMS reminding aspirin intake Arm 2: standard care SMS group had significantly better adherence than control: platelet testing – non-adherence in SMS 5.2% vs. 11.2% in control, OR (95% CI): 0.43 (0.22–0.86); p = .01)a SMS greater self-reported adherence: OR (95% CI): 0.37 (0.15–0.90); p = .02) N/A 92% of intervention group thought SMS support was valuable Varnfield 201430 Arm 1: CAP-CR used smartphone app to monitor wellness and exercise. Received educational/motivational SMS, video and audio messages Arm 2: traditional CR (TCR) At six weeks: no differences in nutrition (fat, fibre, sodium, alcohol) At six weeks: significant differences between groups for diastolic BP, triglycerides, EQ5D-Index. No differences in 6MWT, mental health, systolic BP, heart rate, weight, body mass index, waist circumference, lipids At six months: no differences between groups for 6MWT, LDL, HDL, HRQoL, or psychological distress At six weeks: uptake (RR 1.3), adherence (RR 1.4) and completion (RR 1.7) was higher in CAP-CR than TCRa 85% of participants found step counter to be motivating Daily exercise entries recorded by 89% of CAP-CR adherers (n = 45) N = 3 dropout from CAP-CR due to difficulty with technology SMS: short-message service; PA: physical activity; TMG: telemonitoring group; RR: relative risk; CI: confidence interval; BP: blood pressure; HbA1C: glycosylated haemoglobin level; LDL: low-density lipoprotein; HRQoL: health related quality of life; ACS: acute coronary syndrome; SF36: short form 36 health survey; CAP: care assessment platform; CR: cardiac rehabilitation; EQ5D: health status questionnaire; 6MWT: six minute walk test; HDL: high-density lipoprotein a Primary outcome. Open in new tab The primary outcome of interest for this review was behaviour change; however, a variety of clinical outcomes (four studies) and psychological outcomes (five studies) were reported. For clinical outcomes, a positive effect was found on blood pressure in two of three studies,23,30 but no differences were seen on lipids,23,30 body mass index, and exercise or functional capacity.25,30 No study assessing anxiety22,23 or depression22 found any effects, but quality of life measures showed positive effects in two25,30 of three studies.23 The self-efficacy theoretical construct improved in one25 of three studies28,30 in which it was measured and was found to partially mediate physical activity behaviour. Only one study examined cost and found its intervention to be cost-effective for increasing leisure time physical activity.25 One serious adverse event was found to be related to study treatment (cycling accident) in the one study in which adverse events were reported.25 Process outcomes Each study included some type of process outcome. Three studies reported high fidelity, or the degree of participants' adherence, to the SMS component of their interventions,23–25 and one to the smartphone application (wellness diary) among adherers.30 The number of SMS messages received or read was not reported in four studies. In studies with Internet components, Antypas and Wangberg22 found no differences in the number of website visits between the tailored and control group, while Maddison et al.25 reported lower use of the web compared with the SMS component of the intervention. All three medication adherence studies reported high satisfaction with the SMS support;24,27,29 overall satisfaction was not reported in the other studies. Few technical difficulties were reported in any of the seven trials. Discussion To our knowledge this was the first review examining mHealth interventions to change behaviour specifically in the CVD population. We identified seven recent studies (published in 2012–2015) that met our criteria. The objective of five trials was specifically to change behaviours, of which four had this as their primary outcome, and all appeared to have a positive effect on either medication adherence or physical activity. No effects were observed on dietary behaviours or smoking cessation; however, these outcomes were secondary and were measured in only one study each. It is possible that changing dietary and/or smoking behaviours may be more difficult via mHealth; however, there is strong evidence for the use of SMS-based interventions to stop smoking in non-clinical populations.32 As dietary change and smoking cessation were measured in only one study each in the current review, more research is needed before conclusions can be drawn on the effectiveness of mHealth interventions to change these behaviours. Due to the low number of studies and the variable quality of included studies, it remains difficult to draw conclusions on the effectiveness of these interventions. Data were too heterogeneous to conduct meta-analyses; however, using the narrative synthesis methodology, findings show potential for the use of mHealth to change behaviour as part of CVD self-management. Previous reviews have also found mHealth or SMS interventions to have positive effects on behaviour change or disease self-management in other populations.12,16,33 A review of reviews found the majority of SMS interventions delivered to healthy and clinical populations were effective when addressing physical activity, smoking cessation or medication adherence,12 which is consistent with our conclusions with the exceptions of smoking cessation, which remains under-studied in the CVD population. All studies included in our review reported either high fidelity or high satisfaction with mHealth intervention and few technical issues, indicating that the CVD population examined in this review is comfortable using technology in their rehabilitation. It is important to note that participants were likely to be mobile phone or Internet literate, as owning a mobile phone was part of the eligibility criteria in six studies. More research is needed to determine whether these effects translate to other CVD patients; there is potential that mHealth could inadvertently cause health disparities if some have limited access or ability to use technologies.16 Despite these potential shortcomings, a recent report found 77% of adults aged 65 + owned a mobile phone and over half (59%) used the Internet.14 Characteristics of effective interventions All interventions were delivered in part via SMS on mobile phones. The five effective interventions were relatively simple to deliver and operate. The two studies with null behaviour findings involved more complex remote monitoring and feedback from clinicians23 or mentors;30 however, the primary outcome for these studies was not the behavioural outcome of interest for this review. In the two studies which used a website in addition to SMS,22,25 the web-based components appeared to be underutilised by participants. From our findings, it appears that simple, easy to use SMS interventions may help patients change and/or adhere to recommended lifestyle behaviours. It is unknown whether more complex interventions involving remote monitoring by mobile phone are effective at changing behaviour as more studies are needed. Cost-effectiveness is also an important consideration as complex mHealth interventions may have greater set-up costs than simple SMS interventions, particularly if patients need to be provided with at-home monitoring equipment33 or if interventions require patients to use mobile-broadband data plans, of which prices vary considerably worldwide.11 It is frequently argued that a benefit of mHealth may be its cost-effectiveness; the only study that evaluated costs did support this claim.25 It was difficult to determine the intervention dose–response relationship, as three of seven studies did not report the frequency of messaging. For the studies that did report dose, receiving at least one message every three days appeared to have an effect, and all four studies that reported dose found a positive effect. Only one study compared two different frequencies and types of messages, and found no differences between the high- and low-dose groups.27 Intervention duration also varied considerably. We found evidence for short term effects (one to two months) on medication adherence in three studies24,27,29 and evidence for maintenance of physical activity at three22 and six months.25 But only one study had a long follow-up (12 months) and so sustainability of effect is unknown. While short term effects were observed in studies with shorter follow-up (one to six months) conclusions on sustained effect beyond this time period cannot be made. One study with a longer-term follow-up (12 months) showed no effect on behaviour change (smoking cessation).23 Studies with longer follow up are needed to determine the sustained effect on behaviour change. Previous reviews have reported that the effective mHealth behaviour change interventions were those that were theory-based34 and offered personalised, tailored and bi-directional messaging.12 Our findings were inconsistent and although each intervention was personalised, the degree of personalisation or tailoring did not appear to influence outcomes. Three studies were underpinned by a theoretical framework with measured constructs; however, only one study saw a significant difference in the theoretical outcome measured (self-efficacy) and found it to partially mediate physical activity behaviour.26 It is difficult to draw conclusions on theoretical outcomes as each study used a different self-efficacy measure and measurement tool. As all studies had a behavioural outcome, an examination of theoretical constructs and BCTs was warranted. Each intervention incorporated some type of BCT, even if it was not specifically mentioned. A greater number of BCTs did not correspond to greater behaviour change in our review; however, it is important to note that we included different behavioural outcomes in our review. It is easier to adhere to some behaviours than others, for instance a large multi-centre trial found that the percentage of acute coronary syndrome patients adhering to medication regimen was higher than adherence to smoking, diet or exercise behaviour.35 A simple one-way SMS reminder or prompt might be effective to see changes in medication adherence but may not be effective in changing more complex behaviour like eating a healthy diet. Authors of a large systematic review examining the effectiveness of mHealth behaviour change and disease management interventions argued the need for coding intervention content according to BCTs to investigate which are most effective at changing behaviour.13 We coded intervention content according to the BCT taxonomy31 to attempt meaningful interpretation of results but found no consistent pattern; however, the coding of BCTs was limited to information available in each publication and to our interpretations. Access to SMS or other intervention content sent to participants would be needed to increase coding accuracy. As with other reviews, we were unable to draw strong conclusions on features of effective interventions due to the small number of studies (with a combined sample size of just over 1200 participants) and the need for better reporting of intervention characteristics.12,33 More agreement from researchers is needed on the use of the BCT taxonomy or similar to guide intervention development and measurement31 and the use of appropriate theoretical measures and measurement tools. Only then can we tease out which mHealth content or features resonate best with people, determine the mechanism of action and pool theoretical data for meta-analysis. Strengths and limitations of studies included in the review Previous mHealth reviews have called for the need for more robust study design with adequate power before claims of effectiveness can truly be made.12,13 Strengths of the current review were that each study used an RCT design and five of seven studies included power calculations for their primary outcome; however only two studies were powered for behavioural outcomes.22,24 All but one study accounted for all randomised participants30 and all but two studies described using intention to treat analyses.22,29 A limitation was that two of the five studies with pre-specified power calculations failed to achieve the target sample size as stated in protocol papers due to recruitment issues.22,30 The same two studies also had high levels of attrition; therefore recruitment and retention of participants were limitations in their study designs. Detection bias was another limitation as three studies were unblinded and one did not describe their blinding methodology.29 Only three studies had registered their trials; the same three also published protocols and development papers.22,25,30 It is disappointing that new studies emerging after the release of the CONSORT statement for reporting parallel group RCTs in 201036 are being published even though key methodology and intervention details appear to be missing. The studies in this review that had unclear risk of bias were not registered or did not publish protocols. It is well known that studies with unclear reporting of methods tend to over-estimate treatment effects.37 A final limitation of the included studies was the lack of diversity across almost all demographic characteristics, that is, gender, age, ethnicity. For example, the majority of participants were men around age 60. While studies were undertaken across many countries, they mostly represented higher income countries. There is a need to conduct similar studies in low or middle income countries where CVD has high rates with few secondary prevention resources.38 Women, elderly and/or patients of low socio-economic status have not been adequately represented and future studies should determine how these groups respond to mHealth programmes. Limitations of this review Our review may suffer publication and language bias, as only studies published in English were included. We did not limit our search by publication type or language, but conference excluded abstracts and four studies not published in English during the full text screening. One study published in Korean appeared to meet our criteria and had a positive effect on smoking cessation;39 however, the inclusion of one additional study would not have greatly altered our findings. We also may have missed some literature during our search; however, we aimed to minimise this through pre-testing the search criteria to find known studies and manually searching relevant references lists. As mentioned, we were unable to conduct a meta-analysis due to the heterogeneity of the outcome data. With time, more mHealth studies will be published meeting our review criteria as seven published protocols for registered RCTs were found during our search.40–46 A review update with meta-analysis will be warranted in a few years. Implications for future research Our review revealed that there is the potential for mHealth to increase medication adherence and physical activity participation among CVD patients at least in the short term; however, more research is needed before it could be recommended to include technology into existing cardiac rehabilitation or secondary prevention programmes. As argued, there is a need for better reporting of intervention characteristics and a consensus on the measures used to detect changes in behaviour and potential moderators/mediators of behaviour. It is still unclear what type of mHealth interventions are most effective, although simple SMS has shown the most promise in the CVD population to date. More empirical research enabling meta-analysis is required before we can say mHealth interventions can change key behaviours in the CVD population. Definitive research using high quality, robust design with long term follow-up (such as 2–5 years) is needed to examine sustainability of behaviour change and size of the effect on clinical outcomes. Conclusions This review examined seven trials using mHealth to deliver behaviour change interventions to patients with CVD and found small but positive effects on medication adherence and physical activity. While study design and quality has improved since earlier reviews of mHealth, the review is still limited by the small number of trials, inconsistent outcome measures and ineffective reporting of intervention characteristics. Large scale, longitudinal studies are warranted to gain a clear understanding of the effects of mHealth on behaviour change (and by proxy the clinical impact) in the CVD population. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) received no financial support for the research, authorship, and/or publication of this article. References 1 World Health Organisation . Cardiovascular diseases (CVDs) , Geneva : World Health Organisation , 2013 . Fact sheet N 317 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 2 Piepoli MF , Corra U, Adamopoulos Set al. . Secondary prevention in the clinical management of patients with cardiovascular diseases. Core components, standards and outcome measures for referral and delivery: A Policy Statement from the Cardiac Rehabilitation Section of the European Association for Cardiovascular Prevention & Rehabilitation. Endorsed by the Committee for Practice Guidelines of the European Society of Cardiology . Eur J Prev Cardiol 2014 ; 21 : 664 – 681 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Müller-Riemenschneider F , Meinhard C, Damm Ket al. . Effectiveness of nonpharmacological secondary prevention of coronary heart disease . Eur J Cardiovasc Prev Rehabil 2010 ; 17 : 688 – 700 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Whalley B , Thompson D, Taylor R. Psychological interventions for coronary heart disease: Cochrane systematic review and meta-analysis . Int J Behav Med 2014 ; 21 : 109 – 121 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Turk-Adawi KI , Grace SL. Narrative review comparing the benefits of and participation in cardiac rehabilitation in high-, middle- and low-income countries . Heart Lung Circ 2015 ; 24 : 510 – 520 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Neubeck L , Freedman SB, Clark AMet al. . Participating in cardiac rehabilitation: A systematic review and meta-synthesis of qualitative data . Eur J Prev Cardiol 2012 ; 19 : 494 – 503 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Dalal HM , Zawada A, Jolly Ket al. . Home based versus centre based cardiac rehabilitation: Cochrane review and meta-analysis . BMJ 2010 ; 340 : b5631 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Taylor RS , Dalal H, Jolly Ket al. . Home-based versus centre-based cardiac rehabilitation . Cochrane Database Syst Rev 2010 . CD007130-CD . Google Scholar OpenURL Placeholder Text WorldCat 9 Neubeck L , Redfern J, Fernandez Ret al. . Telehealth interventions for the secondary prevention of coronary heart disease: A systematic review . Eur J Cardiovasc Prev Rehabil 2009 ; 16 : 281 – 289 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Huang K , Liu W, He Det al. . Telehealth interventions versus center-based cardiac rehabilitation of coronary artery disease: A systematic review and meta-analysis . Eur J Prev Cardiol 2015 ; 22 : 959 – 971 . Google Scholar Crossref Search ADS PubMed WorldCat 11 International Telecommunication Union . ICT facts and figures , Geneva, Switzerland : International Telecommunication Union , 2013 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 12 Hall AK , Cole-Lewis H, Bernhardt JM. Mobile text messaging for health: A systematic review of reviews . Ann Rev Public Health 2015 ; 36 : 393 – 415 . Google Scholar Crossref Search ADS WorldCat 13 Free C , Phillips G, Galli Let al. . The effectiveness of mobile-health technology-based health behaviour change or disease management interventions for health care consumers: A systematic review . PLoS Med 2013 ; 10 : e1001362 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Smith A . Older adults and technology use , Washington DC : Pew Research Center , 2014 . . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 15 Widmer RJ , Collins NM, Collins CSet al. . Digital health interventions for the prevention of cardiovascular disease: A systematic review and meta-analysis . Mayo Clinic Proc , 90 : 469 – 480 . Crossref Search ADS WorldCat 16 Hamine S , Gerth-Guyette E, Faulx Det al. . Impact of mHealth chronic disease management on treatment adherence and patient outcomes: A systematic review . J Med Internet Res 2015 ; 17 : e52 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Fjeldsoe BS , Marshall AL, Miller YD. Behavior change interventions delivered by mobile phone telephone short-message service . Am J Prev Med 2009 ; 36 : 165 – 173 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Liberati A , Altman DG, Tetzlaff Jet al. . The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration . PLoS Medicine 2009 ; 6 : e1000100 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Higgins J , Green S. Cochrane handbook for systematic reviews of interventions , Chichester, UK : John Wiley & Sons Ltd , 2008 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 20 World Health Organisation . mHealth: New horizons for health through mobile technologies , Geneva : World Health Organisation , 2011 . Global Observatory for eHealth series, vol. 3 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 21 Jadad AR , Moore RA, Carroll Det al. . Assessing the quality of reports of randomized clinical trials: Is blinding necessary? . Control Clin Trials 1996 ; 17 : 1 – 12 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Antypas K , Wangberg SC. An Internet- and mobile-based tailored intervention to enhance maintenance of physical activity after cardiac rehabilitation: Short-term results of a randomized controlled trial . J Med Internet Res 2014 ; 16 : e77 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Blasco A , Carmona M, Fernández-Lozano Iet al. . Evaluation of a telemedicine service for the secondary prevention of coronary artery disease . J Cardiopulm Rehabil Prev 2012 ; 32 : 25 – 31 . . Google Scholar Crossref Search ADS PubMed WorldCat 24 Khonsari S , Subramanian P, Chinna Ket al. . Effect of a reminder system using an automated short message service on medication adherence following acute coronary syndrome . Eur J Cardiovasc Nurs 2015 ; 14 : 170 – 179 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Maddison R , Pfaeffli L, Whittaker Ret al. . A mobile phone intervention increases physical activity in people with cardiovascular disease: Results from the HEART randomized controlled trial . Eur J Prev Cardiol 2015 ; 22 : 701 – 709 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Maddison R , Pfaeffli L, Stewart Ret al. . The HEART mobile phone trial: The partial mediating effects of self-efficacy on physical activity among cardiac patients . Front Public Health 2014 ; 2 : 56 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Park LG , Howie-Esquivel J, Chung MLet al. . A text messaging intervention to promote medication adherence for patients with coronary heart disease: A randomized controlled trial . Patient Educ Couns 2014 ; 94 : 261 – 268 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Park LG , Howie-Esquivel J, Whooley MAet al. . Psychosocial factors and medication adherence among patients with coronary heart disease: A text messaging intervention . Eur J Cardiovasc Nurs 2014 ; 14 : 264 – 273 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Quilici J , Fugon L, Beguin Set al. . Effect of motivational mobile phone short message service on aspirin adherence after coronary stenting for acute coronary syndrome . Int J Cardiol 2013 ; 168 : 568 – 569 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Varnfield M , Karunanithi M, Lee CKet al. . Smartphone-based home care model improved use of cardiac rehabilitation in postmyocardial infarction patients: Results from a randomised controlled trial . Heart 2014 ; 100 : 1770 – 1779 . . Google Scholar Crossref Search ADS PubMed WorldCat 31 Michie S , Richardson M, Johnston Met al. . The behavior change technique taxonomy (v1) of 93 hierarchically clustered techniques: Building an international consensus for the reporting of behavior change interventions . Ann Behav Med 2013 ; 46 : 81 – 95 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Whittaker R , McRobbie H, Bullen Cet al. . Mobile phone-based interventions for smoking cessation . Cochrane Database Syst Rev 2012 ; 11 : Cd006611 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 33 Anglada-Martinez H , Riu-Viladoms G, Martin-Conde Met al. . Does mHealth increase adherence to medication? Results of a systematic review . Int J Clin Pract 2015 ; 69 : 9 – 32 . Google Scholar Crossref Search ADS PubMed WorldCat 34 Cole-Lewis H , Kershaw T. Text messaging as a tool for behavior change in disease prevention and management . Epidemiol Rev 2010 ; 32 : 56 – 69 . Google Scholar Crossref Search ADS PubMed WorldCat 35 Chow C , Jolly S, Rao-Melacini Pet al. . Association of diet and exercise, and smoking modification with risk of early cardiovascular events after acute coronary syndromes . Circulation 2010 ; 121 : 750 – 758 . Google Scholar Crossref Search ADS PubMed WorldCat 36 Schulz KF , Altman DG, Moher D. CONSORT 2010 Statement: Updated guidelines for reporting parallel group randomized trials . Ann Intern Med 2010 ; 152 : 726 – 732 . Google Scholar Crossref Search ADS PubMed WorldCat 37 Moher D , Hopewell S, Schulz KFet al. . CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials . J Clin Epidemiol 2010 ; 63 : e1 – e37 . Google Scholar Crossref Search ADS PubMed WorldCat 38 Gaziano TA , Bitton A, Anand Set al. . Growing epidemic of coronary heart disease in low- and middle-income countries . Curr Probl Cardiol 2010 ; 35 : 72 – 115 . Google Scholar Crossref Search ADS PubMed WorldCat 39 Kong JH , Ha Y. [Effects of a smoking cessation program including telephone counseling and text messaging using stages of change for outpatients after a myocardial infarction] . J Korean Acad Nurs 2013 ; 43 : 557 – 567 . Google Scholar Crossref Search ADS PubMed WorldCat 40 Dale LP , Whittaker R, Jiang Yet al. . Improving coronary heart disease self-management using mobile technologies (Text4Heart): A randomised controlled trial protocol . Trials 2014 ; 15 : 71 . Google Scholar Crossref Search ADS PubMed WorldCat 41 Maddison R , Rawstorn JC, Rolleston Aet al. . The remote exercise monitoring trial for exercise-based cardiac rehabilitation (REMOTE-CR): A randomised controlled trial protocol . BMC Public Health 2014 ; 14 : 1236 . Google Scholar Crossref Search ADS PubMed WorldCat 42 Eyles H , McLean R, Neal Bet al. . Using mobile technology to support lower-salt food choices for people with cardiovascular disease: Protocol for the SaltSwitch randomized controlled trial . BMC Public Health 2014 ; 14 : 950 . Google Scholar Crossref Search ADS PubMed WorldCat 43 Chow C , Redfern J, Thiagalingam Aet al. . Design and rationale of the tobacco, exercise and diet messages (TEXT ME) trial of a text-message-based intervention for ongoing prevention of cardiovascular disease in people with coronary disease: A randomised controlled trial . BMJ Open 2012 ; 2 : e000606 . Google Scholar Crossref Search ADS PubMed WorldCat 44 Redfern J , Usherwood T, Harris MFet al. . A randomised controlled trial of a consumer-focused e-health strategy for cardiovascular risk management in primary care: the Consumer Navigation of Electronic Cardiovascular Tools (CONNECT) study protocol . BMJ Open 2014 ; 4 : e004523 . Google Scholar Crossref Search ADS PubMed WorldCat 45 Alsaleh E , Blake H, Windle R. Behavioural intervention to increase physical activity among patients with coronary heart disease: Protocol for a randomised controlled trial . Int J Nurs Stud 2012 ; 49 : 1489 – 1493 . Google Scholar Crossref Search ADS PubMed WorldCat 46 Saywell N , Vandal AC, Brown Pet al. . Telerehabilitation to improve outcomes for people with stroke: Study protocol for a randomised controlled trial . Trials 2012 ; 13 : 233 . Google Scholar Crossref Search ADS PubMed WorldCat © The European Society of Cardiology 2016 This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © The European Society of Cardiology 2016

Prince, Stephanie A; Blanchard, Christopher M; Grace, Sherry L; Reid, Robert D

doi: 10.1177/2047487315617101pmid: 26607698

Abstract Background Sedentary time is an independent risk factor for cardiometabolic disease and mortality. It is unknown how much time individuals with coronary artery disease spend being sedentary or how their sedentary time relates to markers of health. The objectives of this study were to: (a) quantify sedentary time in a post-cardiac rehabilitation (CR) population, and (b) assess association with cardiometabolic risk, independent of moderate-to-vigorous physical activity. Design Cross-sectional. Methods As part of a larger trial, 263 recent CR graduates (∼10 days post-CR, mean age 63.6 ± 9.3 years, 75% male) wore an ActiGraph GT3X accelerometer during waking hours (≥4 days, ≥10 hours/day) to quantify sedentary time (≤150 counts per minute). Spearman correlations were computed to assess relationships between sedentary time (adjusted for wear time) with markers of cardiometabolic health and fitness. Significant markers were examined using multiple linear regressions. Results Participants spent an average of 8 hours/day sedentary (∼14 bouts/day). Sedentary time was negatively correlated with high-density lipoprotein and V·O2peak and positively correlated with triglycerides, body mass index and waist circumference. After adjusting for age, sex, medications and moderate-to-vigorous physical activity, hours/day of sedentary time remained significantly associated with log V·O2peak (β = −0.02, p = 0.001) and body mass index (β = 0.49, p = 0.02). Conclusions Findings suggest that even among a group of post-CR individuals who are already probably more active than patients who have not undergone CR, sedentary time remains high and is associated with poorer cardiorespiratory fitness, suggesting a possible new area of focus among CR programs. Sedentary lifestyle, motor activity, heart diseases, rehabilitation Introduction The majority of Canadian adults’ waking time is spent being sedentary.1 Sedentary behavior refers to an energy expenditure ≤1.5 metabolic equivalents while in a sitting or reclining posture during waking hours and not simply the absence of physical activity.2 Research consistently identifies sedentary behavior as an independent risk factor for chronic diseases including cardiovascular disease, diabetes and cancer, as well as premature mortality.3,4 Cardiovascular diseases, including coronary artery disease (CAD), are the leading cause of mortality globally.5 Importantly, CAD outcomes are largely dependent on gains in cardiometabolic fitness achieved via exercise interventions (e.g. cardiac rehabilitation [CR]).6,7 CR is an evidence-based standard of care for those who have experienced an acute coronary syndrome, coronary revascularization and other cardiac conditions. It is recommended in CR that patients achieve 30–60 min of moderate-to-vigorous physical activity (MVPA) on most, or preferably all, days of the week.6 While the benefits of regular MVPA have been established, there has been little research examining to what extent sedentary time has an independent effect on cardiometabolic health and fitness in patients with CAD. Further, CR guidelines do not include any recommendations for sedentary time in patients with CAD, although they do recognize sedentary lifestyle (including sedentary time and physical inactivity) as a risk factor,8 and research has shown that physical activity-focused programs/interventions are not likely to yield meaningful reductions in sedentary time.9 While there is limited data to suggest that individuals with cardiovascular disease have greater sedentary behavior than those without,10 there is currently no published data to quantify sedentary time in patients with CAD or how sedentary time relates to health outcomes in this population. Sedentary time is hypothesized to be greater in this population because it is both a risk factor for CAD and because it can lead to metabolic dysfunction and vascular damage.11 In addition to the lack of descriptive data on sedentary time in this population, there is very limited research describing the relationship between sedentary time and cardiorespiratory fitness.12 Although CR guidelines promote that patients achieve 30–60 min of MVPA per day,6 they may also need to consider evidence around sedentary time from general populations, as the lessons learned are likely to have a greater impact in a CAD population. Therefore, the objectives of this cross-sectional study were to: (a) quantify sedentary behaviors in a post-CR CAD population, and (b) assess associations between sedentary behaviors and measures of cardiometabolic health and fitness, independent of MVPA. The hypothesis of the study was that sedentary time would be related to markers of cardiometabolic health and fitness, with greater sedentary time being associated with poorer measures of health. Methods This cross-sectional study was conducted at two sites: the University of Ottawa Heart Institute (UOHI) and University Health Network (UHN) in Ontario, Canada. Secondary analysis of data from cardiac patients enrolled in an ongoing randomized controlled trial (RCT) named ECologically Optimizing exercise maintenance in men and women Post-CR (ECO-PCR) was undertaken. The trial evaluates the efficacy of an exercise facilitator-led intervention on long-term exercise maintenance among patients who have completed CR in comparison to usual care. This research received ethical approval from the UOHI Human Research Ethics Board (UOHI #201200579-01) and the UHN Research Ethics Board (UHN 12-5018-AE9l). Participants CR graduates from supervised programs were approached to participate. ECO-PCR inclusion and exclusion criteria have been described elsewhere (http://clinicaltrials.gov/ct2/show/NCT01658683). All participants had established CAD, with the majority referred to CR following a percutaneous coronary intervention, coronary artery bypass graft surgery or myocardial infarction. A total of 1736 patients were screened for eligibility, 541 were deemed eligible and 300 were randomized to the RCT. Adults with complete baseline accelerometer data at the time of this sub-study were selected (N = 278). All participants provided written informed consent prior to participating. Procedures CR participants were approached at one of their last classes to solicit participation. Consenting participants were provided a self-report survey assessing sociodemographic characteristics. Clinical characteristics were extracted from charts, including indicators of cardiometabolic health and fitness from their CR discharge assessment. Participants were asked to wear an ActiGraph GT3X accelerometer (ActiGraph, Pensacola, FL) on their right hip during waking hours for 9 days. Participants were instructed to remove the monitor while sleeping or during water-related activities (e.g. swimming or bathing). The GT3X accelerometer is a lightweight and compact triaxial accelerometer that captures movement about the vertical (y-axis), horizontal (x-axis) and perpendicular (z-axis) axes. It also provides output for the vector magnitude; a composite measure of all three axes (vector magnitude = √(x2  +  y2 + z2)). A 15-s sampling epoch was used and converted into counts per minute (cpm). Counts are a reflection of the frequency and intensity of the raw acceleration values and are summed based on epochs (i.e. ‘chunks’ of time). Accelerometer data reduction A valid day was defined as ≥10 hours of wear time, and participants were required to have a minimum of 4 valid days to be retained in the analyses.1 For participants with >7 valid days, the first day was removed (to minimize reactivity), and the subsequent 7 days used for the average. Wear time was calculated by subtracting non-wear time from 24 hours. Non-wear time was defined as at least 60 min of consecutive zeros for counts, with an allowance of up to 2 min of counts between zero and 150. Sedentary time was defined using a previously validated cut-point of ≤150 cpm for use with vector magnitude.13 Sedentary bouts are defined as the number of bouts (≥10 consecutive minutes) of sedentary time detected per day. Each interruption in sedentary time (>150 cpm) was considered a break. The data analysis used the ‘ignore first sedentary break of each day’ to remove sedentary time accrued while the device was theoretically removed at night. If this was not removed the total time in sedentary breaks would be greater than the total wear time. MVPA was defined using a cut-point of ≥2690 cpm for use with vector magnitude.14 Minutes spent in bouts (≥10 consecutive minutes) were used to quantify MVPA.15 Weekly averages were calculated by multiplying the daily average (min/day) by seven. Steps per day were also captured by the accelerometers. Markers of cardiometabolic health Baseline data were extracted from CR charts for several cardiometabolic markers of health, including: measured height and weight to determine body mass index (BMI) in kg/m2, measured waist circumference, waist-to-height ratio, systolic and diastolic blood pressure (BP), HbA1c percentage, triglycerides, total cholesterol, high-density lipoprotein (HDL) and low-density lipoprotein (LDL). Markers of cardiorespiratory fitness After initial randomization to treatment group in ECO-PCR, participants were stratified by treatment group and randomly assigned to one of two exercise testing conditions: graded exercise tests required (50% of sample); or graded exercise tests not required (50% of sample). A symptom-limited graded exercise test with electrocardiographic monitoring using a ramp protocol on a treadmill to obtain V·O2peak (ml/kg/min) as an indicator of cardiorespiratory fitness. Exercise tolerance was also assessed among all participants using the Duke Activity Status Index (DASI), a self-administered questionnaire that measures functional capacity.16 Statistical analysis All analyses were conducted using SAS version 9.3 (SAS Institute Inc., Cary, NC, USA). Summary data were tested for normality using plots and Shapiro–Wilk’s test statistics. Descriptive data were reported using means ± standard deviation or frequencies and percentages. Mean values between males and females were compared using the Wilcoxon–Mann–Whitney test. Spearman correlations were computed to assess relationships between wear time-adjusted sedentary time, bouts of sedentary time and breaks in sedentary time with markers of cardiometabolic health and fitness assessed at CR exit. Cardiometabolic markers that were significantly correlated with sedentary time were then examined using linear regression, controlling for age, sex, medications, wear time and MVPA. All normal variables were log-transformed using base-10 logs prior to entry into the linear regression models. Adjustments were made for age and sex as cardiorespiratory fitness levels are known to decline with age and are known to be more strongly linked with risk factors and cardiovascular disease among males than females.17,18 Sedentary time was adjusted for wear time by using the residuals produced by regressing sedentary time on total wear time.19 Results Of the 278 trial participants, 263 (95%) with valid accelerometer results were used for the analyses. The majority (81%) had a minimum of 7 valid days of data. Results were similar regardless of whether missing days were imputed based on daily averages or when only using those with 7 days of valid data. Table 1 provides participant characteristics. On average, participants were considered to be well educated, married, non-smokers, overweight, normotensive, appeared to have well-managed lipids, and had a moderate-to-high functional capacity.20,21 Women were older, had greater total cholesterol, HDL and LDL and were more likely to be retired. Men had greater BMI, waist circumference, diastolic BP, DASI scores, and V·O2peak values. Table 1. Participant characteristics. Characteristic . Total (N = 263) . Males (n = 198) . Females (n = 65) . Males vs females . Sociodemographics  Age (years) 63.6 ± 9.3 62.9 ± 9.3 65.7 ± 9.0 0.03  College or university, n (%) 183 (74%) 134 (72%) 49 (80%) 0.35  Married or equivalent, n (%) 183 (74%) 149 (80%) 34 (56%) 0.005  Working status 0.03  Employed full time, n (%) 82 (31%) 66 (33%) 16 (25%)  Employed part time, n (%) 11 (4%) 7 (4%) 4 (6%)  Self-employed 24 (9%) 23 (12%) 1 (1.5%)  Unemployed 10 (4%) 9 (4%) 1 (1.5%)  Retired 122 (47%) 83 (42%) 39 (60%)  Missing 14 (5%) 10 (5%) 4 (6%) Clinical  Smoker, n (%) 4 (2%) 3 (2%) 1 (2%) 0.51  Type 2 diabetes, n (%) 45 (18%) 39 (20%) 6 (9%) 0.05 Medications  ACE-inhibitors, n (%) 153 (58%) 117 (59%) 36 (55%) 0.60  Beta-blockers, n (%) 207 (79%) 157 (79%) 50 (77%) 0.69  Statins, n (%) 242 (92%) 186 (94%) 56 (86%) 0.05 Markers of cardiometabolic health and fitness  Body mass index (kg/m2) [n = 262] 28.2 ± 4.6 28.5 ± 4.4 27.1 ± 5.0 0.008  Waist circumference (cm) [n = 262] 99.5 ± 12.6 101.4 ± 12.2 93.6 ± 12.2 <0.0001  Waist-to-height ratio [n = 262] 0.58 ± 0.08 0.58 ± 0.08 0.58 ± 0.08 0.44  Systolic BP (mm Hg) [n = 261] 121 ± 18 121 ± 17 120 ± 19 0.85  Diastolic BP (mm Hg) [n = 262] 72 ± 10 73 ± 9 70 ± 10 0.03  HbA1c (%) [n = 245] 5.9 ± 0.7 5.9 ± 0.8 5.8 ± 0.6 0.45  Total cholesterol (mmol/l) [n = 259] 3.40 ± 0.82 3.28 ± 0.69 3.77 ± 0.57 0.0002  HDL (mmol/l) [n = 259] 1.18 ± 0.32 1.09 ± 0.28 1.45 ± 0.29 <0.0001  LDL (mmol/l) [n = 258] 1.68 ± 0.65 1.62 ± 0.56 1.87 ± 0.86 0.05  Triglycerides (mmol/l) [n = 259] 1.23 ± 0.65 1.28 ± 0.72 1.11 ± 0.39 0.10  DASI score [n = 216] 47.0 ± 11.5 48.5 ± 10.5 41.4 ± 13.5 0.001  Peak V·O2 before CR (ml/kg/min) [n = 110] 22.1 ± 5.8 23.0 ± 5.6 18.8 ± 5.6 0.003  Peak V·O2 after CR (ml/kg/min) [n = 168] 25.2 ± 6.6 26.2 ± 6.7 21.1 ± 4.3 <0.0001  Level of functional capacity [n = 168]20,21 0.0005  Low (<16 ml/kg/min), n (%) 12 (7%) 7 (5%) 5 (14%)  Moderate (16–21 ml/kg/min), n (%) 35 (21%) 21 (16%) 14 (40%)  High (≥22 ml/kg/min), n (%) 121 (72%) 105 (79%) 16 (46%) Characteristic . Total (N = 263) . Males (n = 198) . Females (n = 65) . Males vs females . Sociodemographics  Age (years) 63.6 ± 9.3 62.9 ± 9.3 65.7 ± 9.0 0.03  College or university, n (%) 183 (74%) 134 (72%) 49 (80%) 0.35  Married or equivalent, n (%) 183 (74%) 149 (80%) 34 (56%) 0.005  Working status 0.03  Employed full time, n (%) 82 (31%) 66 (33%) 16 (25%)  Employed part time, n (%) 11 (4%) 7 (4%) 4 (6%)  Self-employed 24 (9%) 23 (12%) 1 (1.5%)  Unemployed 10 (4%) 9 (4%) 1 (1.5%)  Retired 122 (47%) 83 (42%) 39 (60%)  Missing 14 (5%) 10 (5%) 4 (6%) Clinical  Smoker, n (%) 4 (2%) 3 (2%) 1 (2%) 0.51  Type 2 diabetes, n (%) 45 (18%) 39 (20%) 6 (9%) 0.05 Medications  ACE-inhibitors, n (%) 153 (58%) 117 (59%) 36 (55%) 0.60  Beta-blockers, n (%) 207 (79%) 157 (79%) 50 (77%) 0.69  Statins, n (%) 242 (92%) 186 (94%) 56 (86%) 0.05 Markers of cardiometabolic health and fitness  Body mass index (kg/m2) [n = 262] 28.2 ± 4.6 28.5 ± 4.4 27.1 ± 5.0 0.008  Waist circumference (cm) [n = 262] 99.5 ± 12.6 101.4 ± 12.2 93.6 ± 12.2 <0.0001  Waist-to-height ratio [n = 262] 0.58 ± 0.08 0.58 ± 0.08 0.58 ± 0.08 0.44  Systolic BP (mm Hg) [n = 261] 121 ± 18 121 ± 17 120 ± 19 0.85  Diastolic BP (mm Hg) [n = 262] 72 ± 10 73 ± 9 70 ± 10 0.03  HbA1c (%) [n = 245] 5.9 ± 0.7 5.9 ± 0.8 5.8 ± 0.6 0.45  Total cholesterol (mmol/l) [n = 259] 3.40 ± 0.82 3.28 ± 0.69 3.77 ± 0.57 0.0002  HDL (mmol/l) [n = 259] 1.18 ± 0.32 1.09 ± 0.28 1.45 ± 0.29 <0.0001  LDL (mmol/l) [n = 258] 1.68 ± 0.65 1.62 ± 0.56 1.87 ± 0.86 0.05  Triglycerides (mmol/l) [n = 259] 1.23 ± 0.65 1.28 ± 0.72 1.11 ± 0.39 0.10  DASI score [n = 216] 47.0 ± 11.5 48.5 ± 10.5 41.4 ± 13.5 0.001  Peak V·O2 before CR (ml/kg/min) [n = 110] 22.1 ± 5.8 23.0 ± 5.6 18.8 ± 5.6 0.003  Peak V·O2 after CR (ml/kg/min) [n = 168] 25.2 ± 6.6 26.2 ± 6.7 21.1 ± 4.3 <0.0001  Level of functional capacity [n = 168]20,21 0.0005  Low (<16 ml/kg/min), n (%) 12 (7%) 7 (5%) 5 (14%)  Moderate (16–21 ml/kg/min), n (%) 35 (21%) 21 (16%) 14 (40%)  High (≥22 ml/kg/min), n (%) 121 (72%) 105 (79%) 16 (46%) Data presented as mean ± standard deviation unless otherwise specified. CR: cardiac rehabilitation; DASI: Duke Activity Status Index; HDL: high-density lipoprotein; LDL: low-density lipoprotein. Open in new tab Table 1. Participant characteristics. Characteristic . Total (N = 263) . Males (n = 198) . Females (n = 65) . Males vs females . Sociodemographics  Age (years) 63.6 ± 9.3 62.9 ± 9.3 65.7 ± 9.0 0.03  College or university, n (%) 183 (74%) 134 (72%) 49 (80%) 0.35  Married or equivalent, n (%) 183 (74%) 149 (80%) 34 (56%) 0.005  Working status 0.03  Employed full time, n (%) 82 (31%) 66 (33%) 16 (25%)  Employed part time, n (%) 11 (4%) 7 (4%) 4 (6%)  Self-employed 24 (9%) 23 (12%) 1 (1.5%)  Unemployed 10 (4%) 9 (4%) 1 (1.5%)  Retired 122 (47%) 83 (42%) 39 (60%)  Missing 14 (5%) 10 (5%) 4 (6%) Clinical  Smoker, n (%) 4 (2%) 3 (2%) 1 (2%) 0.51  Type 2 diabetes, n (%) 45 (18%) 39 (20%) 6 (9%) 0.05 Medications  ACE-inhibitors, n (%) 153 (58%) 117 (59%) 36 (55%) 0.60  Beta-blockers, n (%) 207 (79%) 157 (79%) 50 (77%) 0.69  Statins, n (%) 242 (92%) 186 (94%) 56 (86%) 0.05 Markers of cardiometabolic health and fitness  Body mass index (kg/m2) [n = 262] 28.2 ± 4.6 28.5 ± 4.4 27.1 ± 5.0 0.008  Waist circumference (cm) [n = 262] 99.5 ± 12.6 101.4 ± 12.2 93.6 ± 12.2 <0.0001  Waist-to-height ratio [n = 262] 0.58 ± 0.08 0.58 ± 0.08 0.58 ± 0.08 0.44  Systolic BP (mm Hg) [n = 261] 121 ± 18 121 ± 17 120 ± 19 0.85  Diastolic BP (mm Hg) [n = 262] 72 ± 10 73 ± 9 70 ± 10 0.03  HbA1c (%) [n = 245] 5.9 ± 0.7 5.9 ± 0.8 5.8 ± 0.6 0.45  Total cholesterol (mmol/l) [n = 259] 3.40 ± 0.82 3.28 ± 0.69 3.77 ± 0.57 0.0002  HDL (mmol/l) [n = 259] 1.18 ± 0.32 1.09 ± 0.28 1.45 ± 0.29 <0.0001  LDL (mmol/l) [n = 258] 1.68 ± 0.65 1.62 ± 0.56 1.87 ± 0.86 0.05  Triglycerides (mmol/l) [n = 259] 1.23 ± 0.65 1.28 ± 0.72 1.11 ± 0.39 0.10  DASI score [n = 216] 47.0 ± 11.5 48.5 ± 10.5 41.4 ± 13.5 0.001  Peak V·O2 before CR (ml/kg/min) [n = 110] 22.1 ± 5.8 23.0 ± 5.6 18.8 ± 5.6 0.003  Peak V·O2 after CR (ml/kg/min) [n = 168] 25.2 ± 6.6 26.2 ± 6.7 21.1 ± 4.3 <0.0001  Level of functional capacity [n = 168]20,21 0.0005  Low (<16 ml/kg/min), n (%) 12 (7%) 7 (5%) 5 (14%)  Moderate (16–21 ml/kg/min), n (%) 35 (21%) 21 (16%) 14 (40%)  High (≥22 ml/kg/min), n (%) 121 (72%) 105 (79%) 16 (46%) Characteristic . Total (N = 263) . Males (n = 198) . Females (n = 65) . Males vs females . Sociodemographics  Age (years) 63.6 ± 9.3 62.9 ± 9.3 65.7 ± 9.0 0.03  College or university, n (%) 183 (74%) 134 (72%) 49 (80%) 0.35  Married or equivalent, n (%) 183 (74%) 149 (80%) 34 (56%) 0.005  Working status 0.03  Employed full time, n (%) 82 (31%) 66 (33%) 16 (25%)  Employed part time, n (%) 11 (4%) 7 (4%) 4 (6%)  Self-employed 24 (9%) 23 (12%) 1 (1.5%)  Unemployed 10 (4%) 9 (4%) 1 (1.5%)  Retired 122 (47%) 83 (42%) 39 (60%)  Missing 14 (5%) 10 (5%) 4 (6%) Clinical  Smoker, n (%) 4 (2%) 3 (2%) 1 (2%) 0.51  Type 2 diabetes, n (%) 45 (18%) 39 (20%) 6 (9%) 0.05 Medications  ACE-inhibitors, n (%) 153 (58%) 117 (59%) 36 (55%) 0.60  Beta-blockers, n (%) 207 (79%) 157 (79%) 50 (77%) 0.69  Statins, n (%) 242 (92%) 186 (94%) 56 (86%) 0.05 Markers of cardiometabolic health and fitness  Body mass index (kg/m2) [n = 262] 28.2 ± 4.6 28.5 ± 4.4 27.1 ± 5.0 0.008  Waist circumference (cm) [n = 262] 99.5 ± 12.6 101.4 ± 12.2 93.6 ± 12.2 <0.0001  Waist-to-height ratio [n = 262] 0.58 ± 0.08 0.58 ± 0.08 0.58 ± 0.08 0.44  Systolic BP (mm Hg) [n = 261] 121 ± 18 121 ± 17 120 ± 19 0.85  Diastolic BP (mm Hg) [n = 262] 72 ± 10 73 ± 9 70 ± 10 0.03  HbA1c (%) [n = 245] 5.9 ± 0.7 5.9 ± 0.8 5.8 ± 0.6 0.45  Total cholesterol (mmol/l) [n = 259] 3.40 ± 0.82 3.28 ± 0.69 3.77 ± 0.57 0.0002  HDL (mmol/l) [n = 259] 1.18 ± 0.32 1.09 ± 0.28 1.45 ± 0.29 <0.0001  LDL (mmol/l) [n = 258] 1.68 ± 0.65 1.62 ± 0.56 1.87 ± 0.86 0.05  Triglycerides (mmol/l) [n = 259] 1.23 ± 0.65 1.28 ± 0.72 1.11 ± 0.39 0.10  DASI score [n = 216] 47.0 ± 11.5 48.5 ± 10.5 41.4 ± 13.5 0.001  Peak V·O2 before CR (ml/kg/min) [n = 110] 22.1 ± 5.8 23.0 ± 5.6 18.8 ± 5.6 0.003  Peak V·O2 after CR (ml/kg/min) [n = 168] 25.2 ± 6.6 26.2 ± 6.7 21.1 ± 4.3 <0.0001  Level of functional capacity [n = 168]20,21 0.0005  Low (<16 ml/kg/min), n (%) 12 (7%) 7 (5%) 5 (14%)  Moderate (16–21 ml/kg/min), n (%) 35 (21%) 21 (16%) 14 (40%)  High (≥22 ml/kg/min), n (%) 121 (72%) 105 (79%) 16 (46%) Data presented as mean ± standard deviation unless otherwise specified. CR: cardiac rehabilitation; DASI: Duke Activity Status Index; HDL: high-density lipoprotein; LDL: low-density lipoprotein. Open in new tab Movement patterns are described in Table 2. On average, participants spent 8 hours/day or 56% of their waking day being sedentary and only 3% of their day in bouts of MVPA. Sedentary time was accumulated in 14 bouts per day with an average of 13 breaks per day. Males accumulated significantly more sedentary time than females (mean difference = 58.9 min/day). Table 2. Wear time and movement patterns by sex. . Mean ± standard deviation . . . Total (N = 263) . Males (n = 198) . Females (n = 65) . Males vs females . Total wear time, min/day 848.6 ± 77.7 852.8 ± 76.9 835.6 ± 79.1 p = 0.15 Sedentary behavior  Sedentary time, min/day 478.8 ± 94.0 493.4 ± 90.6 434.5 ± 90.6 p < 0.0001  Sedentary time, hours/day 8.0 ± 1.6 8.2 ± 1.5 7.2 ± 1.5 p < 0.0001  Sedentary bouts, number/day 14.1 ± 3.8 14.7 ± 3.6 12.3 ± 3.7 p < 0.0001  Sedentary breaks, number/day 13.1 ± 3.7 13.7 ± 3.6 11.2 ± 3.7 p < 0.0001 Activity  Light activity, min/day 324.9 ± 81.3 312.5 ± 78.2 362.5 ± 79.7 p < 0.0001  Total MVPA, min/day 44.9 ± 26.7 46.9 ± 27.4 38.7 ± 23.7 p = 0.04  Time in MVPA bouts, min/day 24.9 ± 19.9 26.3 ± 20.2 20.9 ± 18.5 p = 0.05  Steps per day 7296 ± 2732 7290 ± 2709 7313 ± 2822 p = 0.76 . Mean ± standard deviation . . . Total (N = 263) . Males (n = 198) . Females (n = 65) . Males vs females . Total wear time, min/day 848.6 ± 77.7 852.8 ± 76.9 835.6 ± 79.1 p = 0.15 Sedentary behavior  Sedentary time, min/day 478.8 ± 94.0 493.4 ± 90.6 434.5 ± 90.6 p < 0.0001  Sedentary time, hours/day 8.0 ± 1.6 8.2 ± 1.5 7.2 ± 1.5 p < 0.0001  Sedentary bouts, number/day 14.1 ± 3.8 14.7 ± 3.6 12.3 ± 3.7 p < 0.0001  Sedentary breaks, number/day 13.1 ± 3.7 13.7 ± 3.6 11.2 ± 3.7 p < 0.0001 Activity  Light activity, min/day 324.9 ± 81.3 312.5 ± 78.2 362.5 ± 79.7 p < 0.0001  Total MVPA, min/day 44.9 ± 26.7 46.9 ± 27.4 38.7 ± 23.7 p = 0.04  Time in MVPA bouts, min/day 24.9 ± 19.9 26.3 ± 20.2 20.9 ± 18.5 p = 0.05  Steps per day 7296 ± 2732 7290 ± 2709 7313 ± 2822 p = 0.76 MVPA: moderate-to-vigorous intensity physical activity. Open in new tab Table 2. Wear time and movement patterns by sex. . Mean ± standard deviation . . . Total (N = 263) . Males (n = 198) . Females (n = 65) . Males vs females . Total wear time, min/day 848.6 ± 77.7 852.8 ± 76.9 835.6 ± 79.1 p = 0.15 Sedentary behavior  Sedentary time, min/day 478.8 ± 94.0 493.4 ± 90.6 434.5 ± 90.6 p < 0.0001  Sedentary time, hours/day 8.0 ± 1.6 8.2 ± 1.5 7.2 ± 1.5 p < 0.0001  Sedentary bouts, number/day 14.1 ± 3.8 14.7 ± 3.6 12.3 ± 3.7 p < 0.0001  Sedentary breaks, number/day 13.1 ± 3.7 13.7 ± 3.6 11.2 ± 3.7 p < 0.0001 Activity  Light activity, min/day 324.9 ± 81.3 312.5 ± 78.2 362.5 ± 79.7 p < 0.0001  Total MVPA, min/day 44.9 ± 26.7 46.9 ± 27.4 38.7 ± 23.7 p = 0.04  Time in MVPA bouts, min/day 24.9 ± 19.9 26.3 ± 20.2 20.9 ± 18.5 p = 0.05  Steps per day 7296 ± 2732 7290 ± 2709 7313 ± 2822 p = 0.76 . Mean ± standard deviation . . . Total (N = 263) . Males (n = 198) . Females (n = 65) . Males vs females . Total wear time, min/day 848.6 ± 77.7 852.8 ± 76.9 835.6 ± 79.1 p = 0.15 Sedentary behavior  Sedentary time, min/day 478.8 ± 94.0 493.4 ± 90.6 434.5 ± 90.6 p < 0.0001  Sedentary time, hours/day 8.0 ± 1.6 8.2 ± 1.5 7.2 ± 1.5 p < 0.0001  Sedentary bouts, number/day 14.1 ± 3.8 14.7 ± 3.6 12.3 ± 3.7 p < 0.0001  Sedentary breaks, number/day 13.1 ± 3.7 13.7 ± 3.6 11.2 ± 3.7 p < 0.0001 Activity  Light activity, min/day 324.9 ± 81.3 312.5 ± 78.2 362.5 ± 79.7 p < 0.0001  Total MVPA, min/day 44.9 ± 26.7 46.9 ± 27.4 38.7 ± 23.7 p = 0.04  Time in MVPA bouts, min/day 24.9 ± 19.9 26.3 ± 20.2 20.9 ± 18.5 p = 0.05  Steps per day 7296 ± 2732 7290 ± 2709 7313 ± 2822 p = 0.76 MVPA: moderate-to-vigorous intensity physical activity. Open in new tab Spearman correlations between sedentary time and the markers of cardiometabolic health and fitness identified significant positive correlations with BMI (r = 0.17, p = 0.006), waist circumference (r = 0.17, p = 0.005) and triglycerides (r = 0.14, p = 0.03) and significant negative correlations with HDL (r = −0.17, p = 0.005) and V·O2peak (r = −0.19, p = 0.02). Correlations with sedentary bouts and sedentary breaks were similar for BMI (bouts: r = 0.20, p = 0.002; breaks: r = 0.19, p = 0.002) and waist circumference (r = 0.17, p = 0.005 for both), were lower for HDL (−0.14, p = 0.02 for both) and not significant for triglycerides or V·O2peak. Sedentary time was highly correlated with breaks (r = 0.88) and bouts (r = 0.89). Average time in bouts, HDL, triglycerides and V·O2peak were not normally distributed and log-transformed prior to entry into the regression models. Results of the linear regression analyses (Table 3) identified that wear time-adjusted sedentary time was significantly associated with BMI, waist circumference, logHDL and log V·O2peak, but not logTriglycerides. For example, for every 1 hour increase in sedentary time per day, waist circumference increased by ∼2 cm and V·O2peak decreased by 1.04 ml/kg/min. Adjusting for MVPA did not change the strength of the associations. Adding age and sex to the models rendered waist circumference and logHDL no longer significant, but strengthened the association with log V·O2peak; further adjusting for medication did not change the associations. Based on results from the correlation analysis, as well as the fully adjusted models, compared to the other outcomes assessed, sedentary time showed the strongest association with log V·O2peak. Table 3. Standardized regression coefficients of total sedentary time with selected markers of cardiometabolic health and fitness. . ST adjusted for wear time . ST adjusted for wear time and MVPA . ST adjusted for age, sex, MVPA . ST adjusted for age, sex, medications, MVPA . Metabolic variables . β, p-value . β, p-value . β, p-value . β, p-value . Adjusted R2 . Body mass index 0.63, p = 0.002 0.59, p = 0.004 0.51, p = 0.01 0.49, p = 0.02 0.09 Waist circumference 1.86, p = 0.001 1.77, p = 0.002 1.16, p = 0.04 1.13, p = 0.05 0.10 logTriglyceridesa 1.03, p = 0.08 1.03, p = 0.12 1.02, p = 0.29 1.02, p = 0.27 0.02 logHDLa −1.03, p = 0.008 −1.03, p = 0.006 −1.01, p = 0.44 −1.01, p = 0.38 0.28 log |${\dot {\rm V}}$| O2peaka (n = 168) −1.04, p = 0.006 −1.03, p = 0.02 −1.04, p = 0.0009 −1.04, p = 0.001 0.41 . ST adjusted for wear time . ST adjusted for wear time and MVPA . ST adjusted for age, sex, MVPA . ST adjusted for age, sex, medications, MVPA . Metabolic variables . β, p-value . β, p-value . β, p-value . β, p-value . Adjusted R2 . Body mass index 0.63, p = 0.002 0.59, p = 0.004 0.51, p = 0.01 0.49, p = 0.02 0.09 Waist circumference 1.86, p = 0.001 1.77, p = 0.002 1.16, p = 0.04 1.13, p = 0.05 0.10 logTriglyceridesa 1.03, p = 0.08 1.03, p = 0.12 1.02, p = 0.29 1.02, p = 0.27 0.02 logHDLa −1.03, p = 0.008 −1.03, p = 0.006 −1.01, p = 0.44 −1.01, p = 0.38 0.28 log |${\dot {\rm V}}$| O2peaka (n = 168) −1.04, p = 0.006 −1.03, p = 0.02 −1.04, p = 0.0009 −1.04, p = 0.001 0.41 a Unstandardized regression coefficients for triglycerides, HDL and |${\dot {\rm V}}$| O2peak presented as 10eb to enable the interpretation of these outcomes in their original units. HDL: high-density lipoprotein; MVPA: moderate-to-vigorous intensity physical activity; ST: sedentary time. Open in new tab Table 3. Standardized regression coefficients of total sedentary time with selected markers of cardiometabolic health and fitness. . ST adjusted for wear time . ST adjusted for wear time and MVPA . ST adjusted for age, sex, MVPA . ST adjusted for age, sex, medications, MVPA . Metabolic variables . β, p-value . β, p-value . β, p-value . β, p-value . Adjusted R2 . Body mass index 0.63, p = 0.002 0.59, p = 0.004 0.51, p = 0.01 0.49, p = 0.02 0.09 Waist circumference 1.86, p = 0.001 1.77, p = 0.002 1.16, p = 0.04 1.13, p = 0.05 0.10 logTriglyceridesa 1.03, p = 0.08 1.03, p = 0.12 1.02, p = 0.29 1.02, p = 0.27 0.02 logHDLa −1.03, p = 0.008 −1.03, p = 0.006 −1.01, p = 0.44 −1.01, p = 0.38 0.28 log |${\dot {\rm V}}$| O2peaka (n = 168) −1.04, p = 0.006 −1.03, p = 0.02 −1.04, p = 0.0009 −1.04, p = 0.001 0.41 . ST adjusted for wear time . ST adjusted for wear time and MVPA . ST adjusted for age, sex, MVPA . ST adjusted for age, sex, medications, MVPA . Metabolic variables . β, p-value . β, p-value . β, p-value . β, p-value . Adjusted R2 . Body mass index 0.63, p = 0.002 0.59, p = 0.004 0.51, p = 0.01 0.49, p = 0.02 0.09 Waist circumference 1.86, p = 0.001 1.77, p = 0.002 1.16, p = 0.04 1.13, p = 0.05 0.10 logTriglyceridesa 1.03, p = 0.08 1.03, p = 0.12 1.02, p = 0.29 1.02, p = 0.27 0.02 logHDLa −1.03, p = 0.008 −1.03, p = 0.006 −1.01, p = 0.44 −1.01, p = 0.38 0.28 log |${\dot {\rm V}}$| O2peaka (n = 168) −1.04, p = 0.006 −1.03, p = 0.02 −1.04, p = 0.0009 −1.04, p = 0.001 0.41 a Unstandardized regression coefficients for triglycerides, HDL and |${\dot {\rm V}}$| O2peak presented as 10eb to enable the interpretation of these outcomes in their original units. HDL: high-density lipoprotein; MVPA: moderate-to-vigorous intensity physical activity; ST: sedentary time. Open in new tab Discussion The results of the study indicate that even among a group of post-CR CAD patients, to whom active living has been promoted, the majority of waking time was spent engaged in sedentary behaviors. Further, the majority of these patients were not meeting MVPA guidelines.6 Males were more sedentary than females. Sedentary time was negatively associated with V·O2peak and positively associated with BMI independent of age, sex, wear time, medications and MVPA. Total sedentary time was more consistently associated with markers of cardiometabolic health and fitness than breaks in sedentary time and bouts of sedentary time. The CR graduates in this study spent 56% of their waking day being sedentary, with much of this time spent in prolonged bouts. Evidence suggests that those participating in physical activity-focused interventions are not likely to reduce their sedentary time by a meaningful amount.9 It is postulated that those who exercise may actually compensate for this good behavior by feeling less guilt for sitting once the exercise is complete. As this study is cross-sectional it is not possible to gauge whether sedentary time was reduced as a result of CR. Future research is needed to compare sedentary behavior in patients pre-CR and among those who have not participated in CR to ascertain whether CR participation is associated with lower sedentary behavior. This may inform sedentary behavior recommendations in CR, if evolving evidence continues to demonstrate that sedentary behavior is independently associated with poorer health outcomes in cardiac and non-cardiac patients. Regardless, given the current state of the evidence around sedentary behavior, replacing sedentary time with light or greater intensity movements has beneficial effects on health risk.22 Sex differences in activity patterns were observed. Male CAD patients spent a significantly greater proportion of their day engaged in sedentary behavior (58% vs 52%); while females spent a greater amount of time in light physical activity or incidental movement. This finding is consistent with population-level data showing that after the age of 60 the trend reverses, with males being more sedentary.23 Similar to other studies, a gradient in association was observed between greater sedentary time and greater BMI and waist circumference and lower HDL and V·O2peak.12,24–26 Interestingly, waist-to-height ratio, which has been shown to be a better predictor of cardiovascular disease risk,27 was not significantly associated with sedentary time. The ratio appears to remove issues associated with age and sex for measuring risk27 and helps to explain how waist circumference is no longer significant after adjustment for age and sex. The lack of association between blood glucose control and sedentary time was consistent with previous studies.19,24 This lack of association is probably explained by the fact that associations are usually observed for insulin rather than glucose, which mechanistically has been shown to be more acutely affected by reduced muscle contractility observed during bouts of sedentary time.11 Previous studies in apparently healthy populations have shown mixed findings, with some showing significant associations19,28 and others showing no association19,24,26 between sedentary time and total cholesterol, LDL or BP. Sedentary time is likely related to these factors through similar pathways as exercise. Literature has consistently documented that exercise is related to HDL and triglyceride values, but not LDL.29,30 Further, the findings are also likely explained by the large number of participants taking statins, ACE-inhibitors and beta-blockers for lowering cholesterol (specifically LDL) and BP.31,32 The majority had well-controlled lipids and BP. The strongest relationship was observed between sedentary time and V·O2peak, independent of MVPA. This finding is of particular importance given the prognostic value of cardiorespiratory fitness on cardiovascular health and mortality,33 especially among CR patients.34 The findings herein are also consistent with previous studies showing negative associations between sedentary time and cardiorespiratory fitness in general population samples.12,25 Given that gains in cardiorespiratory fitness is one of the key indicators of CR success,35 and that the present study identified a significant association between sedentary time and V·O2peak in a post-CR CAD population, further longitudinal or interventional study investigations are warranted to establish the directional effects of sedentary time on cardiorespiratory fitness. Further, while MVPA and V·O2peak were more strongly correlated (r = 0.35), reducing sedentary time probably provides additional benefits and may be an easier target for patients. This study is not without limitations. The greatest limitation is that this study was cross-sectional in design, and therefore no causal conclusions can be drawn. As such, the identification of directionality between sedentary time and the markers of cardiometabolic health and fitness, or whether sedentary time has a compounding effect over time on these outcomes, requires future research. This is an important area for future research. Although this study is cross-sectional, it is the first known study to have quantified sedentary time and its relationship with health indicators in a CAD population, and does suggest a new and important area of study for this population. Further, measures of sedentary time were collected post-CR, with an average of 10 days post-graduation. This does introduce an issue with the temporality of the data collection. While this is a limitation, movement patterns are not likely to have changed dramatically within this time frame. Secondly, this study relied on movement cut-points that were developed in healthy adult populations. Moreover, they were developed in controlled laboratory settings, whereas participants in the current study wore the accelerometers in free-living situations. It is thus possible that these cut-points may lead to some misclassification of sedentary time and MVPA. However, the use of accelerometers provides objective measurement of sedentary time and activity in a continuous pattern and is one of the greater strengths of the study. It has been shown that self-reported sedentary time may underestimate the effects of sedentary time on cardiometabolic health.28 The sample size, although large for a clinical study, is limited in its power to detect associations between some of the variables (power 74–96% for correlations). Finally, the participants were classified as having controlled lipids and BP, having a high functional capacity, were highly educated and were comprised of recent CR graduates; therefore, the findings herein are not necessarily generalizable to CAD patients who have not undergone such physical activity intervention or whose risk factors are not well controlled. Conclusions CR graduates spent almost 8 waking hours/day being sedentary, accrued in approximately 14 bouts/day (∼30 min at a time). Similar to observations from healthy populations, greater sedentary time was associated with lower functional capacity and greater BMI, independent of MVPA. Findings suggest that even among a group of post-CR individuals who are already likely more active than patients who have not undergone CR, sedentary time remains high and is associated with poorer functional capacity, suggesting a possible new area of focus for CR programs. Sedentary time may offer a unique opportunity to improve outcomes, specifically gains in cardiorespiratory fitness, one of the key indicators of CR program success. Further investigation is needed using a longitudinal or interventional study design. Acknowledgements We would like to thank the participants of the ECO-PCR trial for their participation. The authors would like to thank Evyanne Wooding, Ashley Armstrong, Golnoush Taherzadeh and Kujaany Kana for their assistance with participant recruitment and data collection. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: SAP is funded by a Strategic Endowed Research Fellowship from the UOHI Foundation and a Canadian Institutes of Health Research Fellowship. The ECO-PCR study is funded by the Heart and Stroke Foundation of Ontario (Grant #000109). References 1 Colley RC , Garriguet D, Janssen Iet al. . Physical activity of Canadian adults: accelerometer results from the 2007 to 2009 Canadian Health Measures Survey . Health Rep 2011 ; 22 : 7 – 14 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 2 Sedentary Behaviour Research Network . Letter to the editor: standardized use of the terms “sedentary” and “sedentary behaviours” . Appl Physiol Nutr Metab 2012 ; 37 : 540 – 542 . Crossref Search ADS PubMed WorldCat 3 Biswas A , Oh PI, Faulkner GEet al. . Sedentary time and its association with risk for disease incidence, mortality, and hospitalization in adults: a systematic review and meta-analysis . Ann Intern Med 2015 ; 162 : 123 – 132 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Wilmot EG , Edwardson CL, Achana FAet al. . Sedentary time in adults and the association with diabetes, cardiovascular disease and death: systematic review and meta-analysis . Diabetologia 2012 ; 55 : 2895 – 2905 . Google Scholar Crossref Search ADS PubMed WorldCat 5 World Health Organization . Global status report on noncommunicable diseases 2010 , Geneva : WHO , 2011 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 6 Stone JA , Campbell NR, Genest Jet al. Health behaviour interventions and cardiovascular disease risk factor modifications . In: Stone JA, Arthur HM, Suskin NG (eds). Canadian Guidelines for Cardiac Rehabilitation and Cardiovascular Disease Prevention , 3rd ed. Winnipeg, MB : Canadian Association of Cardiac Rehabilitation , 2009 , pp. 251 – 340 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 7 Lavie CJ , Thomas RJ, Squires RWet al. . Exercise training and cardiac rehabilitation in primary and secondary prevention of coronary heart disease . Mayo Clin Proc 2009 ; 84 : 373 – 383 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Piepoli MF , Corra U, Adamopoulos Set al. . Secondary prevention in the clinical management of patients with cardiovascular diseases. Core components, standards and outcome measures for referral and delivery: a policy statement from the cardiac rehabilitation section of the European Association for Cardiovascular Prevention & Rehabilitation . Eur J Prev Cardiol 2014 ; 21 : 664 – 681 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Prince SA , Saunders TJ, Gresty Ket al. . A comparison of the effectiveness of physical activity and sedentary behaviour interventions in reducing sedentary time in adults: a systematic review and meta-analysis of controlled trials . Obes Rev 2014 ; 15 : 905 – 919 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Evenson KR , Butler EN, Rosamond WD. Prevalence of physical activity and sedentary behavior among adults with cardiovascular disease in the United States . J Cardiopulm Rehabil Prev 2014 ; 34 : 406 – 419 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Tremblay MS , Colley RC, Saunders TJet al. . Physiological and health implications of a sedentary lifestyle . Appl Physiol Nutr Metab 2010 ; 35 : 725 – 740 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Kulinski JP , Khera A, Ayers CRet al. . Association between cardiorespiratory fitness and accelerometer-derived physical activity and sedentary time in the general population . Mayo Clin Proc 2014 ; 89 : 1063 – 1071 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Kozey-Keadle S , Libertine A, Lyden Ket al. . Validation of wearable monitors for assessing sedentary behavior . Med Sci Sports Exerc 2011 ; 43 : 1561 – 1567 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Sasaki JE , John D, Freedson PS. Validation and comparison of ActiGraph activity monitors . J Sci Med Sport 2011 ; 14 : 411 – 416 . Google Scholar Crossref Search ADS PubMed WorldCat 15 World Health Organization . Global recommendations on physical activity for health , Geneva : WHO , 2010 . Google Scholar PubMed OpenURL Placeholder Text Google Preview WorldCat COPAC 16 Hlatky MA , Boineau RE, Higginbotham MBet al. . A brief self-administered questionnaire to determine functional capacity (the Duke Activity Status Index) . Am J Cardiol 1989 ; 64 : 651 – 654 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Buskirk ER , Hodgson JL. Age and aerobic power: the rate of change in men and women . Fed Proc 1987 ; 46 : 1824 – 1829 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 18 Parker BA , Kalasky MJ, Proctor DN. Evidence for sex differences in cardiovascular aging and adaptive responses to physical activity . Eur J Appl Physiol 2010 ; 110 : 235 – 246 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Carson V , Wong SL, Winkler Eet al. . Patterns of sedentary time and cardiometabolic risk among Canadian adults . Prev Med 2014 ; 65 : 23 – 27 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Weber KT , Janicki JS, McElroy PA. Determination of aerobic capacity and the severity of chronic cardiac and circulatory failure . Circulation 1987 ; 76 : VI40 – 45 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 21 Patterson JA , Naughton J, Pietras RJet al. . Treadmill exercise in assessment of the functional capacity of patients with cardiac disease . Am J Cardiol 1972 ; 30 : 757 – 762 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Buman MP , Winkler EA, Kurka JMet al. . Reallocating time to sleep, sedentary behaviors, or active behaviors: associations with cardiovascular disease risk biomarkers, NHANES 2005–2006 . Am J Epidemiol 2014 ; 179 : 323 – 334 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Matthews CE , Chen KY, Freedson PSet al. . Amount of time spent in sedentary behaviors in the United States, 2003–2004 . Am J Epidemiol 2008 ; 167 : 875 – 881 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Healy GN , Matthews CE, Dunstan DWet al. . Sedentary time and cardio-metabolic biomarkers in US adults: NHANES 2003–06 . Eur Heart J 2011 ; 32 : 590 – 597 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Wientzek A , Tormo Diaz MJ, Castano JMet al. . Cross-sectional associations of objectively measured physical activity, cardiorespiratory fitness and anthropometry in European adults . Obesity 2014 ; 22 : E127 – 134 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Shuval K , Finley CE, Barlow CEet al. . Sedentary behavior, cardiorespiratory fitness, physical activity, and cardiometabolic risk in men: the Cooper Center Longitudinal Study . Mayo Clin Proc 2014 ; 89 : 1052 – 1062 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Browning LM , Hsieh SD, Ashwell M. A systematic review of waist-to-height ratio as a screening tool for the prediction of cardiovascular disease and diabetes: 0· 5 could be a suitable global boundary value . Nutr Res Rev 2010 ; 23 : 247 – 269 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Celis-Morales CA , Perez-Bravo F, Ibanez Let al. . Objective vs. self-reported physical activity and sedentary time: effects of measurement method on relationships with risk biomarkers . PLoS One 2012 ; 7 : e36345 – e36345 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Hartung GH , Squires WG, Gotto AM Jr. Effect of exercise training on plasma high-density lipoprotein cholesterol in coronary disease patients . Am Heart J 1981 ; 101 : 181 – 184 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Durstine JL , Grandjean PW, Davis PGet al. . Blood lipid and lipoprotein adaptations to exercise . Sports Med 2001 ; 31 : 1033 – 1062 . Google Scholar Crossref Search ADS PubMed WorldCat 31 Davidson MH , Toth PP. Comparative effects of lipid-lowering therapies . Prog Cardiovasc Dis 2004 ; 47 : 73 – 104 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Daskalopoulou SS , Rabi DM, Zarnke KBet al. . The 2015 Canadian Hypertension Education Program recommendations for blood pressure measurement, diagnosis, assessment of risk, prevention, and treatment of hypertension . Can J Cardiol 2015 ; 31 : 549 – 568 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Berry JD , Willis B, Gupta Set al. . Lifetime risks for cardiovascular disease mortality by cardiorespiratory fitness levels measured at ages 45, 55, and 65 years in men. The Cooper Center Longitudinal Study . J Am Coll Cardiol 2011 ; 57 : 1604 – 1610 . Google Scholar Crossref Search ADS PubMed WorldCat 34 Kavanagh T , Mertens DJ, Hamm LFet al. . Prediction of long-term prognosis in 12 169 men referred for cardiac rehabilitation . Circulation 2002 ; 106 : 666 – 671 . Google Scholar Crossref Search ADS PubMed WorldCat 35 Grace SL , Poirier P, Norris CMet al. . Pan-Canadian development of cardiac rehabilitation and secondary prevention quality indicators . Can J Cardiol 2014 ; 30 : 945 – 948 . Google Scholar Crossref Search ADS PubMed WorldCat © The European Society of Cardiology 2016 This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © The European Society of Cardiology 2016

Christensen, Anne V; Koch, Mette B; Davidsen, Michael; Jensen, Gorm B; Andersen, Lisbeth V; Juel, Knud

doi: 10.1177/2047487315613665pmid: 26538614

Abstract Background Social inequality is present in the morbidity as well as the mortality of cardiovascular diseases. This paper aims to quantify and compare the level of educational inequality across different cardiovascular diagnoses. Design Register based study. Methods Comparison of the extent of inequality across different cardiovascular diagnoses requires a measure of inequality which is comparable across subgroups with different educational distributions. The slope index of inequality and the relative index of inequality were applied for measuring inequalities in incidence of six cardiovascular diagnoses: ischaemic heart disease, acute myocardial infarction, valvular heart disease, congestive heart failure, atrial fibrillation and stroke in the period 2005–2009. All individuals in the general Danish population aged 35–84 years were followed in national registers regarding hospitalisation, death and education from 1985 to 2009 (annual average of 2.9 million people) to define incident cases. Results Marked educational inequality was found in the incidence of ischaemic heart disease, acute myocardial infarction, heart failure and stroke (relative index of inequality: 0.37 (95% confidence interval 0.34; 0.40) to 0.60 (0.57; 0.63), absolute index of inequality: −241 (−254.4; −227.4) to −37 (−42.7; −31.1)) while inequality in atrial fibrillation and, in particular, in valvular heart disease was small and insignificant (relative index of inequality: 0.57 (0.49; 0.65) to 0.97 (0.88; 1.08), absolute index of inequality: −29 (−35.1; −21.9) to −1 (−4.8; −3.8)). Conclusion The degree of educational inequality in cardiovascular diseases depends on the diagnosis, with the highest inequality in ischaemic heart disease, acute myocardial infarction, heart failure and stroke. Small differences were found between men and women. Cardiovascular diseases, inequalities, registers, Denmark Introduction There is a universal, strong and consistent relationship between socioeconomic position, as assessed, for example, by educational level, and the incidence and mortality of cardiovascular disease (CVD).1–5 This relationship is especially apparent in ischaemic heart disease (IHD), including acute myocardial infarction (AMI),6–8 stroke9 and heart failure.10 Previous studies have shown socioeconomic inequalities in major cardiovascular risk factors such as diet, smoking, obesity and hypertension.11,12 Utilisation of the health care system, for example, access to invasive cardiac procedures, also seems to differ markedly among different educational groups.13,14 Furthermore, psychosocial factors, such as lack of personal control over events and stress, may also be risk factors for CVD.5,14 As these risk factors are socially patterned they may contribute to socioeconomic inequalities. The aim of the current study is to quantify and compare levels of educational inequality within six CVD diagnoses (IHD, AMI, valvular heart disease (VHD), heart failure, atrial fibrillation and stroke) by utilising national public registers. The diagnoses in question have different risk factors and, therefore, levels of educational inequality may vary. As educational inequality in incidence of CVD indicates differences in levels of risk factors, the potential for prevention may differ between diagnostic groups. Methods Registers The analyses in this study are based on data from nationwide Danish registers. Public registers such as the National Patient Register15 (NPR), the National Register of Causes of Death16 (CDR), the National Population Register17 and the Population Education Register18 are maintained and validated in a unified manner. By using the unique Danish identification number17 all individuals in the Danish population aged 35–84 years can be tracked in these registers and information regarding education, hospitalisation and death during the period 1985–2009 can be found. In order to identify incident cases of each of the six diagnoses, each individual was followed in the NPR15 and the CDR.16 The NPR was established in 1977 and contains information about all hospital admissions (1977 to the present) and outpatient contacts (1995 to the present) in Danish hospitals. Each admission is characterised by a primary diagnosis classified according to the International Classification of Diseases (ICD) version 8 (1977–1993) and version 10 (1994 to the present). The CDR contains information on all deaths in Denmark since 1970. Each death is characterised by an underlying cause of death classified, as above, according to the ICD. Diagnoses By utilising the registers, six CVD diagnoses were defined in a standardised way: IHD, AMI, VHD, heart failure, atrial fibrillation and stroke. Each cardiovascular diagnosis was defined using the following ICD codes: all CVDs: I00–I99 (ICD 10), 390–458 (ICD 8); IHD: I20–I25 (ICD 10), 410–414 (ICD 8); AMI: I21–I22 (ICD 10), 410 (ICD 8); VHD: I05–I08, I34–I37 (ICD 10), 394–397, 424 (ICD 8); heart failure: I11.0, I13.0, I13.2, I50 (ICD 10), 427.0 (ICD 8); atrial fibrillation: I48 (ICD 10), 427.4 (ICD 8); stroke: I60–I69 (ICD 10), 430–438 (ICD 8). Definition of incidence A case was selected if a relevant diagnosis was registered in either register. For atrial fibrillation and heart failure, admissions as well as outpatient contacts in the NPR were used, while for the other diagnoses only hospital admissions were considered. Treatment of and examination for atrial fibrillation and heart failure is to a large degree carried out in outpatient clinics while this is not the case for the remaining CVDs. A case was considered to be an incident if it occurred in the period 2005–2009 and no previous registrations of the diagnosis in question were found in the preceding 20 years dating back to 1985. For atrial fibrillation and heart failure a case was considered as incident if no previous registrations were found the last 10 years as registration of outpatient contacts was not initiated until 1995. Definition of education The study comprises all Danish residents aged 35–84 where there is information about educational group during the period 2005–2009. This is an annual average of 2.9 million people and comprises approximately 56% of the total population. Information on educational group as a measure of social position was obtained from the Population Education Register.18 The length of time of the highest completed education was categorised into four groups, labelled: lowest (primary school) (24.3%), low (vocational education or high school) (43.4%), high (secondary education) (22.4%), and highest (tertiary or research education) (7.8%). More women are in the lowest, low and high educational group whereas more men are in the highest educational group.19 Statistical analysis All analyses were carried out separately for men and women. Incidence rates were calculated per 100,000 person-years at risk. The population at risk consists of the whole population with the prevalent cases deducted. Comparison of educational inequality between different groups requires a summary measure. For this purpose the Relative Index of Inequality (RII) and the Slope Index of Inequality (SII) were applied. These measures meet the requirements set by the guidelines on measuring social inequality in health and making comparisons across populations. The RII is a relative measure of inequality that can be interpreted as the ratio of the morbidity and mortality rates of those at the bottom of the educational hierarchy compared with those at the top of the hierarchy. The SII is an absolute measure that expresses the health inequality between the top and bottom of the educational hierarchy in terms of rate differences instead of rate ratios. The RII and SII are regression-based measures that take into account all educational groups, not only the highest and the lowest. Furthermore, these measures also take into account the size of the educational groups. More specifically, each educational group was transformed into an educational score scaled from 0 (shortest time in education) to 1 (longest time in education).20–22 This educational score was weighted to reflect the share of the sample in each educational group. Both the RII and the SII, and their 95% confidence intervals (CIs), were computed by use of a generalised linear model (GLM) using the proc genmod procedure in SAS 9.3 with the binomial distribution. RII is the coefficient to the educational score using the log as link-function, while the SII is the same coefficient using the identity as link-function. All analyses were adjusted for age. The values of RII are between 0 and 1 if incidence rates decrease with increasing education, and greater than 1 if incidence rates increase with increasing education. If RII is 1 there is no relative inequality. The SII is negative if incidence rates decrease with increasing level of education and, conversely, it is positive if the incidence rates increase with increasing education. If SII is 0 there is no absolute inequality. Results A yearly average of 38,000 new cases of CVD was registered in the population during the period 2005–2009. The most numerous CVD is IHD including AMI, where an average of 14,000 new cases was registered in the period 2005–2009 (Table 1). Table 1. Number of incident cases in the period 2005–2009 among the Danish population aged 35–84 years with information about educational group. Yearly average. . CVD . IHD . AMI . VHDa . HF . AF . Stroke . Men 21,198 8830 5029 1045 3251 7239 5457 Women 16,856 5163 2575 715 1887 5306 4550 Total 38,053 13,993 7605 1760 5137 12,545 10,007 . CVD . IHD . AMI . VHDa . HF . AF . Stroke . Men 21,198 8830 5029 1045 3251 7239 5457 Women 16,856 5163 2575 715 1887 5306 4550 Total 38,053 13,993 7605 1760 5137 12,545 10,007 CVD: cardiovascular disease; IHD: ischaemic heart disease; AMI: acute myocardial infarction; VHD: valvular heart disease; HF: heart failure; AF: atrial fibrillation a The VHD group consists of aortic valve disorders (80%), mitral valve disorders (19%) and tricuspid valve disorders (1%). Open in new tab Table 1. Number of incident cases in the period 2005–2009 among the Danish population aged 35–84 years with information about educational group. Yearly average. . CVD . IHD . AMI . VHDa . HF . AF . Stroke . Men 21,198 8830 5029 1045 3251 7239 5457 Women 16,856 5163 2575 715 1887 5306 4550 Total 38,053 13,993 7605 1760 5137 12,545 10,007 . CVD . IHD . AMI . VHDa . HF . AF . Stroke . Men 21,198 8830 5029 1045 3251 7239 5457 Women 16,856 5163 2575 715 1887 5306 4550 Total 38,053 13,993 7605 1760 5137 12,545 10,007 CVD: cardiovascular disease; IHD: ischaemic heart disease; AMI: acute myocardial infarction; VHD: valvular heart disease; HF: heart failure; AF: atrial fibrillation a The VHD group consists of aortic valve disorders (80%), mitral valve disorders (19%) and tricuspid valve disorders (1%). Open in new tab There is a clear educational gradient in CVD among both men and women. Among men, the age-standardised incidence in the group with the lowest education is approximately 2000 per 100,000 person-years. In the group with the highest education, the age-standardised incidence is 1400 per 100,000 person-years. Among women, the age-standardised incidence is approximately 1400 per 100,000 person-years in the lowest educated group and 850 per 100,000 person-years in the highest educated group (data not shown). There are differences in the level, and in the educational gradient, dependent on the specific diagnosis in CVD (Figure 1). Figure 1. Open in new tabDownload slide Incidence of different cardiovascular diseases by educational groups and sex, 2005–2009. Age-standardised rate per 100,000 person-years. IHD: ischaemic heart disease; AMI: acute myocardial infarction; VHD: valvular heart disease; HF: heart failure; AF: atrial fibrillation; Str: stroke Relative inequality With the exception of VHD among men, there was significant relative inequality in the different CVDs as measured by RII. The RII was higher among women than among men in all CVDs, mostly reflecting the lower incidence level among women (Figure 2). In all diagnoses a RII lower than 1 was observed, reflecting decreasing incidence with increasing educational level. Among men, a large relative inequality (RII: 0.5 to 0.6) was seen for IHD, AMI, heart failure and stroke. Atrial fibrillation showed a small relative inequality among men (RII = 0.88, 95% CI: 0.84; 0.91). For VHD no significant inequality was observed. Among women, the same pattern was seen for all diagnoses except VHD, where a rather large relative inequality was observed (RII = 0.57, 95% CI: 0.49; 0.65). Among both men and women the patterns of inequality in the different CVDs were present in all age groups (data not shown). Figure 2. Open in new tabDownload slide Inequality in incidence of different cardiovascular diseases for men and women 2005–2009. Relative index of inequality (1: no inequality, decreasing index: increasing inequality). The relative index of inequality may be interpreted as the rate ratio at the top and the bottom of the educational hierarchy. IHD: ischaemic heart disease; AMI: acute myocardial infarction; VHD: valvular heart disease; HF: heart failure; AF: atrial fibrillation; Str: stroke Absolute inequality Absolute inequality as measured by SII was also found across all CVDs. All SIIs were negative, reflecting decreasing incidence with increasing educational level (Figure 3). The largest absolute inequality was found in IHD for both men and women (SII = −241, 95% CI: −254; −227 and SSI = −142, 95% CI: −150; −133). Among both sexes large inequalities were seen in AMI and stroke (SII: −75 to −140). Heart failure and atrial fibrillation showed a smaller absolute inequality (SII: −26 to −55), reflecting both a flatter gradient and a low incidence of heart failure and atrial fibrillation. There was no significant absolute inequality in VHD among men and a very small absolute inequality among women. These patterns in SII were present in all age groups for both men and women (data not shown). Figure 3. Open in new tabDownload slide Inequality in incidence of different cardiovascular diseases for men and women 2005–2009. Slope index of inequality (0: no inequality, increasing (numerical) index: increasing inequality). The slope index of inequality may be interpreted as the rate difference between the top and the bottom of the educational hierarchy. IHD: ischaemic heart disease; AMI: acute myocardial infarction; VHD: valvular heart disease; HF: heart failure; AF: atrial fibrillation; Str: stroke Discussion Educational inequality was found in CVD and to a varying degree across six cardiovascular diagnoses (IHD, AMI, VHD, heart failure, atrial fibrillation and stroke). In both sexes and in all age groups, considerable educational inequalities were demonstrated in AMI, IHD, heart failure and stroke. However, the inequalities in atrial fibrillation and, in particular, in VHD were small and insignificant. Small differences were found between men and women. It is a well-established fact that social inequality exists in both the morbidity and the mortality of IHD, AMI, heart failure and stroke1–3,5–7,9,10 and this is confirmed in the present study. The incidence of these diagnoses is closely related to lifestyle factors that occur unequally across different social strata.2,10,23,24 Risk factors for IHD and AMI include lifestyle factors, in particular, smoking, obesity, physical inactivity, high alcohol consumption, low intake of fruit and vegetables, and the secondary effects of hypertension, high cholesterol and diabetes. Similarly, the occurrence of stroke and heart failure is related to the same lifestyle factors as well as to the secondary effects of heart rhythm disorders such as atrial fibrillation.12 The fact that lifestyle factors play a significant role in the occurrence of CVDs has been shown in previous studies. In an international study of 52 countries it has been suggested that around 90% of IHD cases could be averted by the normalisation of risk factors.25 In a Norwegian study, adjustment for risk factors (smoking, physical activity, marital status, body mass index (BMI), blood pressure and cholesterol) reduced excess mortality in IHD by 91% for men and 67% for women with short education (primary or no schooling).26 Similarly, other European studies have found that lifestyle factors (diet, smoking and alcohol) explained 70% of the educational differences in coronary heart disease and stroke24 and that psychosocial stress, unhealthy lifestyle, and occupational factors mediated the excess risk of coronary heart disease found in women with short education, and men in low social classes.27,28 Very little educational inequality was found in atrial fibrillation and VHD in this study, indicating that lifestyle factors play only a minor role in the development of these diseases. Risk factors for atrial fibrillation include, in particular, age, hypertension, height, diabetes, obesity, alcohol over-consumption and the presence of another CVD.12,29–31 These risk factors are connected to different social factors: height and alcohol use is more prevalent among the affluent and well-educated,32,33 while obesity and hypertension are more prevalent in lower social strata.34 The overall effect of these risk factors on the occurrence of atrial fibrillation may, therefore, be neutral. The primary cause of VHD is age-associated calcific valve changes and congenital conditions.35 Social inequality in atrial fibrillation and VHD has not been studied to the same extent as other CVDs and results are conflicting. In a Swedish study, socioeconomic factors were not independent risk factors for atrial fibrillation,36 while an American study found the risk of atrial fibrillation to be associated with low family income. However, education was associated with risk of atrial fibrillation only among women.37 Social inequality in VHD has only been sparsely studied. A British study found that a disadvantaged social background had a negative influence on long-term survival after heart valve surgery, especially among women.38 We found small differences in educational inequality in CVD between men and women. The relative educational inequality was generally higher and the absolute educational inequality was lower for women than for men. The largest sex-difference was seen in VHD. These results were affected by the lower prevalence of CVDs among women. An American study has found stronger associations between low income, short time in education and the risk of coronary heart disease among women than among men.7 Women with a short education had greater social and psychological risks than men with a similarly short education in age-adjusted models. The association between short education and increased coronary heart disease risk in women was largely explained by BMI, which was the single most significant factor explaining the association.7 This was also found in a Swiss study which concluded that social inequalities in cardiovascular risk factors were greater among women than among men despite a higher prevalence of risk factors among men.39 The overall incidence of IHD, in general, and AMI, in particular, has declined markedly in Denmark and in many other European countries.40 This development has been brought about by structural and individual efforts to reduce the prevalence of risk factors and by the continuing development of new treatments for CVD. With regard to the incidence of atrial fibrillation, in contrast to IHD and AMI, there seems to be a general increase in occurrence41 which is to a high degree, but not exclusively, caused by the aging of the Danish population. Increasing body size (height and BMI) in the population may also play a role.29,30,40 Little is known about the time frame in the incidence of VHD because register information may be insufficient. Strengths and limitations of this study The analyses for this study are based on national register data covering a homogenous population, which allows for a uniform definition of CVD diagnoses. Cardiovascular diagnoses in the registers have been extensively validated, except for the diagnosis of VHD.42–44 However, there is no reason to suspect lower register quality for VHD as compared with the other diagnoses. This study focuses on incidence defined as the first ever event of CVD. We consider this as a more relevant measure of disease activity than the standard measurement of mortality because most patients survive the first attack. The incidence estimates depend on the length of the latency period – the shorter the latency period, the higher the incidence rate. This is because as the latency period shortens, more subjects with prior disease will be considered as incident cases. It is possible that different latency periods for each diagnosis might have increased the validity of the results.40 The latency period chosen was based on available register information. Information on education is missing for 2%–4% of the population but, as missing data is only a problem for such a small proportion of the total population, we do not consider that it has affected our results. No information concerning risk factors is registered. If risk factor levels had been available, a better understanding of the differential causes of inequality might have been possible. Implications for future research In the face of declining mortality and incidence rates of CVDs, it is notable that inequalities in CVD are still present and growing.1,3,4 So far, prevention of CVDs has largely focused on reducing known risk factors. Even though smoking prevalence has been declining, the decline is less pronounced among low socioeconomic groups in Danish society45 and large inequalities still exist in obesity and overweight,34 indicating that existing prevention and policy efforts may not be effective in reducing inequalities in health. It is of particular interest that decline in smoking prevalence is not associated with declining incidence of atrial fibrillation and VHD. The results of this study indicate that educational inequality exists across different CVD diagnoses, but that the level of inequality depends on the specific CVD diagnosis. It is, therefore, important that the different CVDs are considered separately when developing prevention strategies. By addressing lifestyle changes in population prevention strategies, and, thereby, aiming to reduce the incidence of CVDs, we may also reduce the effect of educational and social inequality. However, we know that, traditionally, the effect of health prevention campaigns is not equally distributed across educational groups.46 Health professionals should bear this in mind when planning prevention strategies or public health policy in relation to CVD. Addressing educational inequality in CVD would, therefore, require socially differentiated efforts targeted towards people with short education. In the present analysis we have not considered the effect of secular trends in educational attainment in the population. The proportion of the population receiving secondary and higher education has increased markedly over the last three decades. The effect of this trend on cardiovascular health warrants further study. This study examined educational inequality across six cardiovascular diagnoses in Danish residents aged 35–84 years. The degree of inequality depends on the specific diagnosis. The effect of educational inequality is likely to be explained by differences in levels of risk factors between educational groups. Future prevention of CVD should differentiate between the CVD diagnoses and provide targeted strategies to people with short education. Acknowledgements KJ generated the core idea for the study. MD and MBK conducted the statistical analyses. AVC and MBK wrote the first draft of the paper. AVC, MBK, MD, GBJ, LVA and KJ contributed to the analyses of the results, and to the writing and final approval of the paper. KJ is the guarantor of this work. The Danish national registers used in this study are approved by the National Data Protection Agency. Hence, no specific ethical approval is required. Data used in this study cannot be shared, as it contains personal identification numbers (encrypted). Technical details and aggregated intermediate datasets are available from the corresponding author. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) received no financial support for the research, authorship, and/or publication of this article. References 1 Dalstra JA , Kunst AE, Borrell Cet al. . Socioeconomic differences in the prevalence of common chronic diseases: An overview of eight European countries . Int J Epidemiol 2005 ; 34 : 316 – 326 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Kaplan GA , Keil JE. Socioeconomic factors and cardiovascular disease: A review of the literature . Circulation 1993 ; 88 : 1973 – 1998 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Mackenbach JP , Cavelaars AE, Kunst AEet al. . Socioeconomic inequalities in cardiovascular disease mortality; an international study . Eur Heart J 2000 ; 21 : 1141 – 1151 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Mackenbach JP , Bos V, Andersen Oet al. . Widening socioeconomic inequalities in mortality in six Western European countries . Int J Epidemiol 2003 ; 32 : 830 – 837 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Marmot MG , Shipley MJ, Rose G. Inequalities in death-specific explanations of a general pattern? Lancet 1984 ; 1 : 1003 – 1006 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Andersen I , Osler M, Petersen Let al. . Income and risk of ischaemic heart disease in men and women in a Nordic welfare country . Int J Epidemiol 2003 ; 32 : 367 – 374 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Thurston RC , Kubzansky LD, Kawachi Iet al. . Is the association between socioeconomic position and coronary heart disease stronger in women than in men? Am J Epidemiol 2005 ; 162 : 57 – 65 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Igland J , Vollset SE, Nygard OKet al. . Educational inequalities in acute myocardial infarction incidence in Norway: A nationwide cohort study . PloS One 2014 ; 9 : e106898 – e106898 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Bennett S . Socioeconomic inequalities in coronary heart disease and stroke mortality among Australian men, 1979–1993 . Int J Epidemiol 1996 ; 25 : 266 – 275 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Blair AS , Lloyd-Williams F, Mair FS. What do we know about socioeconomic status and congestive heart failure? A review of the literature . J Fam Pract 2002 ; 51 : 169 – 169 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 11 Osler M , Gerdes LU, Davidsen Met al. . Socioeconomic status and trends in risk factors for cardiovascular diseases in the Danish MONICA population, 1982–1992 . J Epidemiol Community Health 2000 ; 54 : 108 – 113 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Roger VL , Go AS, Lloyd-Jones DMet al. . Heart disease and stroke statistics – 2012 update: A report from the American Heart Association . Circulation 2012 ; 125 : e2 – e220 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Alter DA , Naylor CD, Austin Pet al. . Effects of socioeconomic status on access to invasive cardiac procedures and on mortality after acute myocardial infarction . N Engl J Med 1999 ; 341 : 1359 – 1367 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Marmot MG , Kogevinas M, Elston MA. Social/economic status and disease . Annu Rev Public Health 1987 ; 8 : 111 – 135 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Lynge E , Sandegaard JL, Rebolj M. The Danish National Patient Register . Scand J Public Health 2011 ; 39 : 30 – 33 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Helweg-Larsen K . The Danish Register of Causes of Death . Scand J Public Health 2011 ; 39 : 26 – 29 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Pedersen CB . The Danish Civil Registration System . Scand J Public Health 2011 ; 39 : 22 – 25 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Jensen VM , Rasmussen AW. Danish education registers . Scand J Public Health 2011 ; 39 : 91 – 94 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Statistics Denmark. Statbank Denmark, www.statistikbanken.dk/statbank5a/default.asp (2015, accessed 23 September 2015) . 20 Ernstsen L , Strand BH, Nilsen SMet al. . Trends in absolute and relative educational inequalities in four modifiable ischaemic heart disease risk factors: Repeated cross-sectional surveys from the Nord-Trondelag Health Study (HUNT) 1984–2008 . BMC Public Health 2012 ; 12 : 266 – 266 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Mackenbach JP , Kunst AE. Measuring the magnitude of socio-economic inequalities in health: An overview of available measures illustrated with two examples from Europe . Soc Sci Med 1997 ; 44 : 757 – 771 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Wagstaff A , Paci P, van Doorslaer E. On the measurement of inequalities in health . Soc Sci Med 1991 ; 33 : 545 – 557 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Koch CG , Li L, Kaplan GAet al. . Socioeconomic position, not race, is linked to death after cardiac surgery . Circ Cardiovasc Qual Outcomes 2010 ; 3 : 267 – 276 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Mejean C , Droomers M, van der Schouw YTet al. . The contribution of diet and lifestyle to socioeconomic inequalities in cardiovascular morbidity and mortality . Int J Cardiol 2013 ; 168 : 5190 – 5195 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Yusuf S , Hawken S, Ounpuu Set al. . Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): Case-control study . Lancet 2004 ; 364 : 937 – 952 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Strand BH , Tverdal A. Can cardiovascular risk factors and lifestyle explain the educational inequalities in mortality from ischaemic heart disease and from other heart diseases? 26 year follow up of 50,000 Norwegian men and women . J Epidemiol Community Health 2004 ; 58 : 705 – 709 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Wamala SP , Mittleman MA, Schenck-Gustafsson Ket al. . Potential explanations for the educational gradient in coronary heart disease: A population-based case-control study of Swedish women . Am J Public Health 1999 ; 89 : 315 – 321 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Suadicani P , Hein HO, Gyntelberg F. Strong mediators of social inequalities in risk of ischaemic heart disease: A six-year follow-up in the Copenhagen Male Study . Int J Epidemiol 1997 ; 26 : 516 – 522 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Rosengren A , Hauptman PJ, Lappas Get al. . Big men and atrial fibrillation: Effects of body size and weight gain on risk of atrial fibrillation in men . Eur Heart J 2009 ; 30 : 1113 – 1120 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Schmidt M , Botker HE, Pedersen Let al. . Adult height and risk of ischemic heart disease, atrial fibrillation, stroke, venous thromboembolism, and premature death: A population based 36-year follow-up study . Eur J Epidemiol 2014 ; 29 : 111 – 118 . Google Scholar Crossref Search ADS PubMed WorldCat 31 Mukamal KJ , Tolstrup JS, Friberg Jet al. . Alcohol consumption and risk of atrial fibrillation in men and women: The Copenhagen City Heart Study . Circulation 2005 ; 112 : 1736 – 1742 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Cavelaars AE , Kunst AE, Geurts JJet al. . Persistent variations in average height between countries and between socio-economic groups: An overview of 10 European countries . Ann Hum Biol 2000 ; 27 : 407 – 421 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Tolstrup JS , Jensen MK, Tjonneland Aet al. . Drinking pattern and mortality in middle-aged men and women . Addiction 2004 ; 99 : 323 – 330 . Google Scholar Crossref Search ADS PubMed WorldCat 34 Devaux M , Sassi F. Social inequalities in obesity and overweight in 11 OECD countries . Eur J Public Health 2013 ; 23 : 464 – 469 . Google Scholar Crossref Search ADS PubMed WorldCat 35 Otto CM , Bonow RO Valvular heart disease . In: Zipes DP, Libby P, Bonow ROet al. (eds). Braunwald’s heart disease: A textbook of cardiovascular medicine , Philadelphia : Elsevier , 2012 , pp. 1468 – 1539 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 36 Zoller B , Li X, Sundquist Jet al. . Neighbourhood deprivation and hospitalization for atrial fibrillation in Sweden . Europace 2013 ; 15 : 1119 – 1127 . Google Scholar Crossref Search ADS PubMed WorldCat 37 Misialek JR , Rose KM, Everson-Rose SAet al. . Socioeconomic status and the incidence of atrial fibrillation in whites and blacks: The Atherosclerosis Risk In Communities (ARIC) study . J Am Heart Assoc 2014 ; 3 : e001159 – e001159 . Google Scholar Crossref Search ADS PubMed WorldCat 38 Bagger JP , Edwards MB, Taylor KM. Influence of socioeconomic status on survival after primary aortic or mitral valve replacement . Heart 2008 ; 94 : 182 – 185 . Google Scholar Crossref Search ADS PubMed WorldCat 39 Stringhini S , Spencer B, Marques-Vidal Pet al. . Age and gender differences in the social patterning of cardiovascular risk factors in Switzerland: The CoLaus study . PloS One 2012 ; 7 : e49443 – e49443 . Google Scholar Crossref Search ADS PubMed WorldCat 40 Koch MB , Davidsen M, Andersen LVet al. . Increasing prevalence despite decreasing incidence of ischaemic heart disease and myocardial infarction. A national register based perspective in Denmark, 1980–2009 . Eur J Prev Cardiol 2015 ; 22 : 189 – 195 . Google Scholar Crossref Search ADS PubMed WorldCat 41 Friberg J , Buch P, Scharling Het al. . Rising rates of hospital admissions for atrial fibrillation . Epidemiology 2003 ; 14 : 666 – 672 . Google Scholar Crossref Search ADS PubMed WorldCat 42 Madsen M , Davidsen M, Rasmussen Set al. . The validity of the diagnosis of acute myocardial infarction in routine statistics: A comparison of mortality and hospital discharge data with the Danish MONICA registry . J Clin Epidemiol 2003 ; 56 : 124 – 130 . Google Scholar Crossref Search ADS PubMed WorldCat 43 Joensen AM , Jensen MK, Overvad Ket al. . Predictive values of acute coronary syndrome discharge diagnoses differed in the Danish National Patient Registry . J Clin Epidemiol 2009 ; 62 : 188 – 194 . Google Scholar Crossref Search ADS PubMed WorldCat 44 Kumler T , Gislason GH, Kirk Vet al. . Accuracy of a heart failure diagnosis in administrative registers . Eur J Heart Fail 2008 ; 10 : 658 – 660 . Google Scholar Crossref Search ADS PubMed WorldCat 45 Jorgensen T , Capewell S, Prescott Eet al. . Population-level changes to promote cardiovascular health . Eur J Prev Cardiol 2013 ; 20 : 409 – 421 . Google Scholar Crossref Search ADS PubMed WorldCat 46 Bender AM , Jorgensen T, Helbech Bet al. . Socioeconomic position and participation in baseline and follow-up visits: The Inter99 study . Eur J Prev Cardiol 2012 ; 21 : 899 – 905 . Google Scholar Crossref Search ADS PubMed WorldCat © The European Society of Cardiology 2016 This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © The European Society of Cardiology 2016

Pussinen, Pirkko J; Könönen, Eija

doi: 10.1177/2047487316636506pmid: 26915578

During the past decades, there has been a growing interest directed to potential associations between oral infections and general health. In this context, cardiovascular diseases have been extensively studied since 1989.1,2 Tooth loss has been related to chronic, non-communicable systemic diseases. It is a common event in middle-aged and elderly populations with compromised dental health, and missing teeth can be seen as a surrogate measure of past and/or present oral disease(s). In the current issue of the European Journal of Preventive Cardiology, there is an interesting article by Vedin and co-workers,3 providing large multi-national data from the STABILITY trial of cardiovascular outcomes in over 15,000 patients with stable coronary heart disease (CHD). Besides a physical examination and blood sampling, the study participants completed a lifestyle questionnaire, including information on dental health. The authors used self-reported tooth loss, categorized in five tooth level groups (no teeth, and 1–14, 15–19, 20–25, and >25 teeth present), and determined the association with major adverse cardiovascular events (MACEs) during a median follow-up time of 3.7 years. MACE was a collective term for outcomes including cardiovascular death, non-fatal myocardial infarction or non-fatal stroke. A 6% increased risk of MACE was found with each tooth loss level so that the highest level, that is, no teeth, had a 27% increased risk in comparison with those having more than 25 teeth at baseline. The authors concluded that tooth loss, independent of the confounding factors, can predict adverse cardiovascular events in stable CHD. The principal reasons leading to missing teeth are: 1) apical periodontitis; profound caries and its sequelae, particularly root canal infection dispersed extraradicularly, and 2) marginal periodontitis; periodontal disease causing destruction in soft and hard tooth-supporting tissues. In adults, periodontal disease is considered the leading cause of tooth loss (http://www.nidcr.nih.gov/DataStatistics/FindDataByTopic/GumDisease/). In a study using full-mouth radiographic measurements,4 periapical bone lesions (apical periodontitis) and considerably reduced marginal bone level (marginal periodontitis) were the major risk factors for tooth loss. It is noteworthy that most of the literature exploring the association between cardiovascular diseases and ‘periodontitis’ denote marginal periodontitis. Despite the different pathogenesis and tissues affected, these two biofilm-associated infections share certain similarities. Marginal periodontitis, a chronic infection-induced inflammatory process, characteristically results in the activation of osteoclastogenesis and, consequently, irreversible destruction of alveolar bone, while apical periodontitis is an inflammatory reaction around the apex of a root in order to detain pulp infection within the root canal system. However, microbial findings as well as findings in infection-induced host response with increased levels of proinflammatory cytokines and neutrophil-derived proteolytic enzymes in tissues resemble each other in these two diseases.5,6 An independent association between marginal periodontitis and increased risk for cardiovascular diseases has been established, although a causal relationship is not supported by the data.2 Periodontal intervention has been shown to decrease systemic inflammation and improve endothelial function but whether it prevents cardiovascular outcomes has not been shown so far.2 Marginal periodontitis is typically a ‘silent disease’, even in advanced cases, and it may cause irreversible harm long before its end point, loss of teeth. The hypothetical mechanisms behind the association include common genetic susceptibility, direct infection of the endothelium by oral bacteria, cross-reactivity between the bacteria- and host-derived antigens called molecular mimicry, and systemic inflammation induced by locally produced cytokines and inflammatory mediators. Infectious agents and their biologically active components and toxins as well as the inflammatory mediators have a substantial access via inflamed and bleeding gingival tissues to the circulation. Repeated bacteraemia and low-grade systemic inflammation develops, the latter being considered the most plausible indirect mechanism linking this chronic infection to cardiovascular diseases.7,8 As regards the relationship between cardiovascular events and apical periodontitis, the literature is scarce.9 Similar mechanisms as suggested above for marginal periodontitis, where a local inflammatory process can give rise to an inflammatory response at distant sites, may link apical periodontitis to cardiovascular events. Indeed, a systematic review by Gomes and co-workers shows that apical periodontitis is associated with elevated levels of systemic inflammatory markers.10 In another recent study of young adults with periapical bone lesions but without marginal periodontitis, cardiovascular events and cardiovascular risk factors, the selected inflammatory and oxidative stress markers correlated inversely with endothelial flow reserve, suggesting that apical periodontitis associates with endothelial dysfunction.11 Whether there is a causal relationship between oral infections and cardiovascular events or whether there are just confounding factors that connect these diseases is still under debate.2,3,10,12 Potentially the strongest confounding factors for the association between marginal periodontitis and cardiovascular diseases include age, smoking and low socioeconomic status, but diabetes, obesity, diet, metabolic syndrome, gender, microbiota and unfavourable lipid profile may also play important roles. In a recent article on the capability of the number of missing teeth in predicting incident cardiovascular events, diabetes and mortality, several of the factors listed above were associated with missing-teeth categories: a direct association was found with age, body mass index, male gender, prevalent diabetes, family history of cardiovascular disease and diabetes, and serum triglyceride and C-reactive protein (CRP) concentrations, and an indirect association with socioeconomic status and high-density lipoprotein cholesterol concentration.13 At the same time, the number of teeth did not associate with self-reported macronutrient intake, including energy, carbohydrate, fat or protein, although unhealthy diet could be behind the associations of missing teeth and cardiovascular diseases. Especially edentulous subjects appear to consume foods which are rich in unhealthy fats and poor in fibres and vitamins.14,15 Moreover, subjects who have lost more teeth are more likely to change their diet to a proatherogenic direction.14,15 Smoking, in a dose-dependent manner, has a major impact on deteriorated periodontal health, seen as marginal alveolar bone loss as well as a reduced number of teeth.12,16 Among the STABILITY population, almost 70% were current or former smokers and over 80% were males.3 In a large German population-based prospective cohort study, a consistently stronger association between smoking and tooth loss was observed in men than women and in younger (<50 years) men than older men and women.16 The strong relation between smoking and tooth loss is seen already in early adulthood.17 Instead, conflicting evidence exists on the impact of smoking on the pathogenesis of apical periodontitis.18 Maybe the link between oral infections and systemic diseases should be further studied among healthy never-smokers solely. In large studies, subanalysis of the outcomes among never-smokers would be possible but is rarely done.19 Recently, adiposity and subclinical inflammation was reported to expose to incident tooth loss, and CRP was suggested to act as a mediator.20 Furthermore, the waist–hip ratio in women but body mass index in men was connected to tooth loss, and these relationships remained significant after adjustments for smoking, socioeconomic factors, dental treatment, caries and periodontal measures. Indeed, periodontitis has been reported to be independently associated with increased odds for overweight (2.56, 95% confidence interval (CI) 1.21–5.40) and obesity (3.11, 95% CI 1.05–6.48) compared with normal weight individuals.21 Also the impact of the individual microbiota and a direct infection has been raised, since oral bacteria may translocate into circulation and induce systemic inflammatory and immunological responses.7 Multiple oral species have been identified in atherosclerotic plaque and coronary thrombi.2,22,23 Patients with symptomatic atherosclerotic vascular disease have increased oral mucosa Anaeroglobus24 as well as salivary and subgingival Aggregatibacter actinomycetemcomitans and systemic immune reaction against it.25,26 Further, the common carotid artery intima-medial thickness progression has been reported to attenuate with improvement in clinical and microbiological periodontal status.27 In addition to transient but repeated bacteraemia, the oral microbiota is also a potential source of endotoxaemia,28 which is associated with an increased risk of cardiometabolic disorders.29 However, further longitudinal and larger studies employing new technologies as well as clinical interventions are needed to establish the role of oral microbiota in atherogenesis. Although self-reported measures have certain shortcomings, they offer a feasible approach to gather health data from large-scale surveys. They are also widely used in dental health comparisons between different countries. Self-reported dental measures, using relatively simple questions, are considered useful.30 In general, there is a good correlation between questionnaire findings and clinical conditions that can be easily recognized. The number of remaining teeth is demonstrated to be reliably calculated and reported by individuals themselves.30 The STABILITY study questionnaire, gathering data from 39 countries, included two questions on participants’ dental health: ‘How many teeth do you have in your mouth?’ and ‘Do your gums bleed when brushing your teeth or at other times?’ It was shown that gum bleeding and tooth loss rates were highest in Eastern Europe.31 This is well in line with a cross-sectional survey of 31 European countries where three self-reported oral health outcomes, edentulousness, no functional dentition and oral impacts on daily living, were evaluated.32 Oral health status of adults varied considerably between European countries clustered by welfare state regimes; having fewer than 20 natural teeth (lack of functional dentition) and edentulousness (no teeth) are most prevalent in Eastern and least prevalent in Northern European countries, especially in Sweden.32 However, this difference may partly be due to different approaches in treatment decisions. The article by Vedin et al. is among the few prospective studies analysing the association of missing teeth with the risk of cardiovascular diseases.3 Interestingly, they investigated this association in subjects with existing CHD when the earlier studies have been performed with general populations19,33,34 or subjects who were free from cardiovascular diseases at the baseline.13,35 In these studies the risk estimates have been on a similar level as in the Vedin et al. study: in a meta-analysis, the summary estimate among subjects with 0–10 remaining teeth at baseline compared with those with only 0–8 missing teeth indicated a 34% increased risk for all CVD events.36 The number of missing teeth could be added into the current general cardiovascular risk factors profiled in the Framingham Heart Study.13 Most of the prospective investigations are based on cardiovascular disease-related mortality,33,37 whereas only a few studies include also morbidity data as outcomes.3,13,34 On the other hand, plenty of cross-sectional studies have reported corresponding results and recognized missing teeth among cardiovascular disease risk factors.38,39 Although several articles report the association of ‘tooth loss’ with cardiovascular outcome, there is only limited information on the actual loss of teeth,35 merely ‘missing teeth’, which have been counted only at the baseline. Furthermore, whether the use of dentures or replacing teeth will affect cardiovascular mortality is still an open question.40 It is noteworthy that the oral infections considered in the present article are highly prevalent in adult-aged populations. However, they are mainly preventable and also treatable with successful outcomes, especially when a proper diagnosis has been made before advanced tissue destruction has occurred.41,42 Basic diagnostics relies on clinical and radiological examinations, while treatment protocols include careful instrumentation with or without local antimicrobial agents, and in certain cases antibiotics or surgery. In marginal periodontitis patients who smoke, both non-surgical and surgical treatment outcome is compromised and, therefore, smoking cessation is currently considered an essential part of their periodontal therapy.43 Furthermore, all periodontitis patients need regular supportive periodontal therapy to maintain the treatment outcome. In compliant patients, a well-maintained periodontal condition can be achieved, whereas in erratic compliers, the disease can continue or relapse exposing patients to the risk of tooth loss.41,44 In cases of asymptomatic apical periodontitis, there is a high frequency of persistence. According to a recent Swedish study,42 more than one-half of persistent lesions in root filled teeth were not retreated and they can remain as such for decades. Unlike smoking cessation, a complete elimination of dental infections by extracting all teeth does not decrease the CHD risk.12 This may be due to late diagnosing of the infections, which have already caused damage to general health, or the outcome of replacing natural teeth with dentures or implants, which may cause problems to general health corresponding with the original infections. In conclusion, the current evidence indicates that the number of missing teeth is linked to an increased risk for cardiovascular events and all-cause mortality. Clearly, in large-scale studies the number of teeth reflects past or even present oral health status and can be used as an easily accessible general marker, when information acquired by expensive and laborious clinical or radiological dental measures are missing. The degree of confounding by common risk factors in oral and cardiovascular diseases may also be decreased when larger epidemiological studies can be exploited. The question about the causality, however, cannot be solved before carefully designed periodontal interventions with sufficient statistical power are available. On an individual level, tooth loss may be useful as an additional indicator when estimating the cardiovascular risk of middle-aged and elderly individuals either with or without prevalent cardiovascular disease. The medical practitioners should be aware of the increased cardiovascular disease risk associated with poor oral health and refer their patients to a dentist, since the oral infections are treatable. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) received no financial support for the research, authorship, and/or publication of this article. References 1 Mattila K , Nieminen MS, Valtonen VVet al. . Association between dental health and acute myocardial infarction . Brit Med J 1989 ; 298 : 779 – 781 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Lockhart PB , Bolger AF, Papapanou PNet al. . Periodontal disease and atherosclerotic vascular disease: Does the evidence support an independent association?: A scientific statement from the American Heart Association . Circulation 2012 ; 125 : 2520 – 2544 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Vedin O, Hagström E, Budaj A, et al. Tooth loss is independently associated with poor outcomes in stable coronary heart disease. Eur J Prev Cardiol. Epub ahead of print 16 December 2015. pii: 2047487315621978 . 4 Bahrami G , Vaeth M, Kirkevang LLet al. . Risk factors for tooth loss in an adult population: A radiographic study . J Clin Periodontol 2008 ; 35 : 1059 – 1065 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Hernandez M , Dutzan N, Garcia-Sesnich Jet al. . Host–pathogen interactions in progressive chronic periodontitis . J Dent Res 2011 ; 90 : 1164 – 1170 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Martinho FC , Leite FR, Nascimento GGet al. . Clinical investigation of bacterial species and endotoxin in endodontic infection and evaluation of root canal content activity against macrophages by cytokine production . Clin Oral Investig 2014 ; 18 : 2095 – 2102 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Pussinen PJ , Tuomisto K, Jousilahti Pet al. . Endotoxemia, immune response to periodontal pathogens, and systemic inflammation associate with incident cardiovascular disease events . Arterioscler Thromb Vasc Biol 2007 ; 27 : 1433 – 1439 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Bokhari SA , Khan AA, Butt AKet al. . Periodontitis in coronary heart disease patients: Strong association between bleeding on probing and systemic biomarkers . J Clin Periodontol 2014 ; 41 : 1048 – 1054 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Cotti E , Dessì C, Piras Aet al. . Can a chronic dental infection be considered a cause of cardiovascular disease? A review of the literature . Int J Cardiol 2011 ; 148 : 4 – 10 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Gomes MS , Blattner TC, Sant’Ana Filho Met al. . Can apical periodontitis modify systemic levels of inflammatory markers? A systematic review and meta-analysis . J Endod 2013 ; 39 : 1205 – 1217 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Cotti E , Zedda A, Deidda Met al. . Endodontic infection and endothelial dysfunction are associated with different mechanisms in men and women . J Endod 2015 ; 41 : 594 – 600 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Hujoel PP , Drangsholt M, Spiekerman Cet al. . Periodontitis–systemic disease associations in the presence of smoking – causal or coincidental? Periodontology 2000 2002 ; 30 : 51 – 60 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Liljestrand JM , Havulinna AS, Paju Set al. . Missing teeth predict incident cardiovascular events, diabetes, and death . J Dent Res 2015 ; 94 : 1055 – 1062 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Hung HC , Willett W, Ascherio Aet al. . Tooth loss and dietary intake . J Am Dent Assoc 2003 ; 134 : 1185 – 1192 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Hung HC , Colditz G, Joshipura KJ. The association between tooth loss and the self-reported intake of selected CVD-related nutrients and foods among US women . Community Dent Oral Epidemiol 2005 ; 33 : 167 – 173 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Dietrich T , Walter C, Oluwagbemigun Ket al. . Smoking, smoking cessation, and risk of tooth loss: The EPIC-Potsdam Study . J Dent Res 2015 ; 94 : 1369 – 1375 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Ylöstalo P , Sakki T, Laitinen Jet al. . The relation of tobacco smoking to tooth loss among young adults . Eur J Oral Sci 2004 ; 112 : 121 – 126 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Walter C , Rodriguez FR, Taner Bet al. . Association of tobacco use and periapical pathosis – a systematic review . Int Endod J 2012 ; 45 : 1065 – 1073 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Watt RG , Tsakos G, de Oliveira Cet al. . Tooth loss and cardiovascular disease mortality risk – results from the Scottish Health Survey . PLoS One 2012 ; 7 : e30797 – e30797 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Meisel P , Holtfräter B, Völzke Het al. . Sex differences of tooth loss and obesity on systemic markers of inflammation . J Dent Res 2014 ; 93 : 774 – 779 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Suvan JE, Petrie A, Nibali L, et al. Association between overweight/obesity and increased risk of periodontitis. J Clin Periodontol. Epub ahead of print 9 June 2015. DOI: 10.1111/jcpe.12421 . 22 Haraszthy VI , Zambon JJ, Trevisan Met al. . Identification of periodontal pathogens in atheromatous plaques . J Periodontol 2000 ; 71 : 1554 – 1560 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Ohki T , Itabashi Y, Kohno Tet al. . Detection of periodontal bacteria in thrombi of patients with acute myocardial infarction by polymerase chain reaction . Am Heart J 2012 ; 163 : 164 – 167 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Fåk F , Tremaroli V, Bergström Get al. . Oral microbiota in patients with atherosclerosis . Atherosclerosis 2015 ; 243 : 573 – 578 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Hyvärinen K , Mäntylä P, Buhlin Ket al. . A common periodontal pathogen has an adverse association with both acute and stable coronary artery disease . Atherosclerosis 2012 ; 223 : 478 – 484 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Mäntylä P , Buhlin K, Paju Set al. . Subgingival Aggregatibacter actinomycetemcomitans associates with the risk of coronary artery disease . J Clin Periodontol 2013 ; 40 : 583 – 590 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Desvarieux M , Demmer RT, Jacobs DRet al. . Changes in clinical and microbiological periodontal profiles relate to progression of carotid intima-media thickness: The Oral Infections and Vascular Disease Epidemiology study . J Am Heart Assoc 2013 ; 2 : e000254 – e000254 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Pussinen PJ , Vilkuna-Rautiainen T, Alfthan Get al. . Severe periodontitis enhances macrophage activation via increased serum lipopolysaccharide . Arterioscler Thromb Vasc Biol 2004 ; 24 : 2174 – 2180 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Kallio KA , Hätönen KA, Lehto Met al. . Endotoxemia, nutrition, and cardiometabolic disorders . Acta Diabetol 2015 ; 52 : 395 – 404 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Buhlin K , Gustafsson A, Andersson Ket al. . Validity and limitations of self-reported periodontal health . Community Dent Oral Epidemiol 2002 ; 30 : 431 – 437 . Google Scholar Crossref Search ADS PubMed WorldCat 31 Vedin O , Hagström E, Gallup Det al. . Periodontal disease in patients with chronic coronary heart disease: Prevalence and association with cardiovascular risk factors . Eur J Prev Cardiol 2015 ; 22 : 771 – 778 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Guarnizo-Herreño CC , Tsakos G, Sheiham Aet al. . Oral health and welfare state regimes: A cross-national analysis of European countries . Eur J Oral Sci 2013 ; 121 : 169 – 175 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Schwahn C , Polzer I, Haring Ret al. . Missing, unreplaced teeth and risk of all-cause and cardiovascular mortality . Int J Cardiol 2013 ; 167 : 1430 – 1437 . Google Scholar Crossref Search ADS PubMed WorldCat 34 Oluwagbemigun K , Dietrich T, Pischon Net al. . Association between number of teeth and chronic systemic diseases: A cohort study followed for 13 years . PLoS One 2015 ; 10 : e0123879 – e0123879 . Google Scholar Crossref Search ADS PubMed WorldCat 35 Hung HC , Joshipura KJ, Colditz Get al. . The association between tooth loss and coronary heart disease in men and women . J Public Health Dent 2004 ; 64 : 209 – 215 . Google Scholar Crossref Search ADS PubMed WorldCat 36 Humphrey LL , Fu R, Buckley DIet al. . Periodontal disease and coronary heart disease incidence: A systematic review and meta-analysis . J Gen Intern Med 2008 ; 23 : 2079 – 2086 . Google Scholar Crossref Search ADS PubMed WorldCat 37 Holmlund A , Holm G, Lind L. Number of teeth as a predictor of cardiovascular mortality in a cohort of 7,674 subjects followed for 12 years . J Periodontol 2010 ; 81 : 870 – 876 . Google Scholar Crossref Search ADS PubMed WorldCat 38 Medina-Solis CE , Pontigo-Loyola AP, Perez-Campos Eet al. . Association between edentulism and angina pectoris in Mexican adults 35 years of age and older: A multivariate analysis of a population-based survey . J Periodontol 2014 ; 85 : 406 – 416 . Google Scholar Crossref Search ADS PubMed WorldCat 39 Wiener RC , Sambamoorthi U. Cross-sectional association between the number of missing teeth and cardiovascular disease among adults aged 50 or older: BRFSS 2010 . Int J Vasc Med 2014 ; 2014 : 421567 – 421567 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 40 Polzer I , Schwahn C, Völzke Het al. . The association of tooth loss with all-cause and circulatory mortality. Is there a benefit of replaced teeth? A systematic review and meta-analysis . Clin Oral Investig 2012 ; 16 : 333 – 351 . Google Scholar Crossref Search ADS PubMed WorldCat 41 Eickholz P , Kaltschmitt J, Berbig Jet al. . Tooth loss after active periodontal therapy. 1: Patient-related factors for risk, prognosis, and quality of outcome . J Clin Periodontol 2008 ; 35 : 165 – 174 . Google Scholar Crossref Search ADS PubMed WorldCat 42 Petersson K, Fransson H, Wolf E, et al. Twenty-year follow-up of root filled teeth in a Swedish population receiving high-cost dental care. Int Endod J. Epub ahead of print 3 July 2015. DOI: 10.1111/iej.12495 . 43 Bunaes DF , Lie SA, Enersen Met al. . Site-specific treatment outcome in smokers following non-surgical and surgical periodontal therapy . J Clin Periodontol 2015 ; 42 : 933 – 942 . Google Scholar Crossref Search ADS PubMed WorldCat 44 Lee CT , Huang HY, Sun TCet al. . Impact of patient compliance on tooth loss during supportive periodontal therapy. A systematic review and meta-analysis . J Dent Res 2015 ; 94 : 777 – 786 . Google Scholar Crossref Search ADS PubMed WorldCat © The European Society of Cardiology 2016 This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © The European Society of Cardiology 2016

Vedin, Ola; Hagström, Emil; Budaj, Andrzej; Denchev, Stephan; Harrington, Robert A; Koenig, Wolfgang; Soffer, Joseph; Sritara, Piyamitr; Stebbins, Amanda; Stewart, Ralph HA; Swart, Henk P; Viigimaa, Margus; Vinereanu, Dragos; Wallentin, Lars; White, Harvey D;

Özyilmaz, Akin; Boersma, Cornelis; Visser, Sipke T; Postma, Maarten J; de Jong-van den Berg, Lolkje TW; Lambers-Heerspink, Hiddo J; de Jong, Paul E; Gansevoort, Ron T

doi: 10.1177/2047487315604310pmid: 26358992

Linhart, Christine; Tivollier, Jean-Michel; Taylor, Richard; Barguil, Yann; Magliano, Dianna J; Bourguignon, Chloé; Zimmet, Paul

doi: 10.1177/2047487315604833pmid: 26346757

Abstract Background Wallis Island is part of a French Territory in the South Pacific. In 1980 the prevalence of hypertension and type 2 diabetes mellitus (T2DM) was low, consistent with a subsistence economy. Considerable social and economic changes have occurred over the last 30 years. Methods Survey data from 1980 and 2009 were analysed by sex in 10-year age groups, and 25–64 years age-standardised to the 2008 Census. Means and prevalences were calculated for blood pressure, fasting plasma glucose, body mass index (BMI), blood cholesterol and triglycerides as risk factors contributing to cardiovascular disease. Results During 1980–2009 there were significant increases (p < 0.05) in age-standardised means and prevalences of blood pressure and hypertension, fasting plasma glucose and T2DM, BMI and obesity, blood cholesterol (men) and triglycerides; and non-significant increases in mean diastolic blood pressure and fasting plasma glucose in women. Mean cholesterol and the prevalence of elevated cholesterol declined in women. Hypertension prevalence increased from 12% to 43% in men and from 15% to 30% in women, with 42% of the increase in men and 33% of the increase in women statistically explained by increases in BMI. T2DM increased from 2.3% to 12.2% in men and from 4.0% to 15.8% in women, with 35% of the increase in men and 26% of the increase in women statistically explained by increases in BMI. Conclusions Risk factors for cardiovascular disease have increased considerably in Wallis Island over the past 30 years, consistent with modernisation in way of life. Cardiovascular diseases, diabetes mellitus, obesity, Polynesia, hypertension, risk factors, cholesterol, triglycerides Introduction Cardiovascular disease (CVD), type 2 diabetes mellitus (T2DM) and their risk factors, including hypertension and obesity, have reached considerable proportions among adults in many developing nations, particularly in some Pacific Island populations.1–3 Wallis and Futuna is a Polynesian French Territory in the South Pacific between Fiji and Samoa. It is relatively isolated with limited natural resources and a population of 13,445 (2008 Census).4 Population growth has been moderated by considerable out—migration to New Caledonia (also a French Territory), with 21,300 people designating Wallisian or Futunan ethnicity in the 2009 New Caledonia Census.5 In view of changes in availability of goods and services, and food consumption habits,6,7 it is anticipated that CVD and its risk factors will have increased in recent decades. In 1980 a survey of T2DM, hypertension and other risk factors for CVD was conducted on samples of Polynesian adults on Wallis Island,8 and in 2009 a survey with similar methodology was undertaken involving samples of Polynesian adults from the islands of both Wallis and Futuna.9 Due to greater isolation and differences in culture, Futunans from the 2009 survey were not included in the present analyses to ensure comparability between the two surveys. Using these population-based cross-sectional surveys of adults aged 25–64 years the present study examines over 30 years on Wallis Island: (1) changes in CVD risk factors in the population, and the trajectory and magnitude of any change; (2) variation in changes by age and sex; and (3) the extent to which changes in body mass index (BMI) statistically explain changes in the prevalence of hypertension and T2DM. There are no other studies of changes in CVD risk factors in a Pacific Island population over such a long period, and from a baseline of such relatively low hypertension and T2DM prevalence. A similar study in the Indian Ocean island of Mauritius found secular increases from 1987 to 2009 in T2DM and fasting plasma glucose (FPG).10 The ethnic composition of the surveyed Mauritian population was predominantly of South Asian and African (Creole) descent,10,11 and the prevalence of T2DM was 12.8% in 1987, consistent with an already modernised population. Methods Study populations In the 1980 survey a village was randomly selected from each of the three main districts of Wallis Island, and the entire population aged ≥20 years (on the 1979 electoral roll) was invited to participate. The response rate was 97% of those present in the villages at the time of the survey. Due to a larger than expected number of people temporarily absent from their village, the sample was augmented by randomly selecting a further 76 participants from two additional villages (response rate 100%). The age structure of the survey sample, the 1976 Wallis Census, and the electoral roll of each village was generally similar. Analysis in the present study is based on 213 men and 228 women aged 25–64 years;8 a sampling fraction of 25% (based on the 1976 Census). The 2009 survey was conducted on the islands of Wallis and Futuna, however, for the purposes of the present study only participants from Wallis Island were included in analysis to facilitate comparisons with the 1980 survey. In 2009 a population list by village and by household was available, and a random selection of households by district was followed by random selection of one person from each household aged ≥18 years and >2 years’ residence. The quota method was used to fill required strata by sex, district and age group to ensure representativeness with the 2008 Census, with a response rate of 86%. Analysis in the present study is based on 116 men and 154 women aged 25–64 years from the Wallis Island sample;9 a sampling fraction of 7% (based on the 2008 Census). Both surveys aimed to include only Wallis Polynesians. The population of Wallis Island is over 97% Polynesian. Inclusion of non-Polynesians is considered to be minimal. Data collection In both surveys questionnaires were administered through interview by the survey team. In 1980 blood pressure (BP) was measured with random zero muddler mercury sphygmomanometers (which reset baselines between successive readings), and in 2009 OMRON M3 electronic BP monitors were employed. The mean of the first and second BP measurements (unaffected by observer bias) were used for analysis in both surveys. Measurements of height (by measuring stick with a movable perpendicular head baton) and weight (by scale) were taken once in both surveys, and used to determine BMI: weight (kg)/height (m)2. In 1980 T2DM hyperglycaemia was based on venous plasma glucose 2 hours after a 75 grams oral glucose load. However, a fasting blood specimen was also taken and the plasma frozen and analysed a week later in Melbourne (Australia); the latter data are used in the present study to ensure comparability. In 2009, a fasting venous blood specimen was taken and the frozen plasma analysed in Noumea (New Caledonia).8,9 Definitions of morbidity Hypertension is defined as systolic blood pressure (SBP) ≥140 mmHg and/or diastolic blood pressure (DBP) ≥90 mmHg and/or currently taking medication for hypertension.12 T2DM is defined as FPG ≥7.0 mmol/l and/or currently taking medication for T2DM.12 Obesity is defined as; BMI ≥30 according to the standard definition by the World Health Organisation (WHO);12 and BMI >32 as suggested by Swinburn et al.13 since Polynesians have a significantly higher ratio of lean mass:fat mass compared with Europeans. Statistical analysis Data were analysed by sex by 10-year age groups; and as one age group (25–64 years) directly age-standardised to the 2008 Census population of Wallis Island (a locally relevant standard) because of alteration in the age structure of the Wallis population between the 1976 and 2008 Censuses from changes in fertility, mortality and out-migration. The effect of BMI on increases in hypertension and T2DM is assessed using logistic regression by comparing the odds ratio (OR) for period after adjusting for age and BMI, compared with adjusting for age alone. Statistical significance was assessed at p < 0.05. Changes in the frequency distributions of BMI between 1980 and 2009 were assessed using histograms, mean and standard deviation (SD), and median and inter-quartile range (IQR). Data were analysed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Biological analysis Biological parameters in 1980 were measured by a SMAC Technicon 20 autoanalyser (Technicon Instrument Corp., Tarrytown, NY 10591, USA) in Melbourne, and in 2009 by an Architect autoanalyser (Abbott Laboratories, Abbott Park, IL, USA) in Noumea. In 1980 FPG was measured by glucose oxidase/peroxidase whereas in 2009 the hexokinase/glucose-6-phosphate dehydrogenase reaction was employed. Cholesterol measurement was with the same enzymatic method (esterase/oxidase/peroxidase with phenolic chromophore) in 198014 and 2009. For plasma triglycerides, in 1980 a glycerol-3-phosphate dehydrogenase reaction on glycerol released after the action of the lipase was used, whereas in 2009 the reaction was based on glycerol phosphate oxidase on total glycerol. Internal and external standards for chemical pathologies laboratories have been in use in Australia from the 1960s, and were in use in Noumea during biochemical analysis of the 2009 survey measurements. Results Between 1980 and 2009 in Wallisian men and women aged 25–64 years there was a statistically significant increase in the means and prevalences (age-standardised) of BP and hypertension, FPG and T2DM, BMI and obesity, and triglycerides; except for non-significant increases in mean DBP and FPG in women (Tables 1 to 4). Table 1. Mean systolic/diastolic blood pressure and hypertension prevalence, Wallis Polynesians, 1980 and 2009. Age . Sample . Mean SBP . Mean DBP . % Hypertension . group . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 118.8 (115.9–121.7) 131.0*** (126.6–135.3) 71.6 (69.0–74.2) 74.5 (71.1–77.8) 10.0 (4.4–18.8) 17.4 (5.0–38.8) 35–44 56 23 115.6 (111.9–119.3) 130.0*** (124.5–135.6) 72.4 (70.0–74.7) 78.3** (74.4–82.2) 7.1 (2.0–17.3) 30.4* (13.2–52.9) 45–54 41 32 117.3 (112.7–121.8) 141.0*** (134.5–147.5) 75.6 (71.9–79.4) 83.1** (78.9–87.2) 12.2 (4.1–26.2) 57.6*** (39.2–74.5) 55–64 36 37 119.9 (113.9–125.9) 141.5*** (133.8–149.2) 75.1 (70.4–79.9) 82.1* (77.7–86.6) 19.4 (8.2–36.0) 64.9*** (47.5–79.8) Linear age trend p-value ns <0.05 <0.05 <0.05 ns <0.05 25–64a 213 115 117.8 (115.8–119.8) 136.2*** (133.0–139.5) 73.5 (71.9–75.1) 79.6*** (77.5–81.6) 11.7 (7.4–16.1) 43.0*** (34.0–52.3) Female 25–34 81 34 110.2 (107.2–113.2) 121.1*** (117.5–124.8) 71.6 (69.3-73.8) 72.3 (68.6-76.0) 7.4 (2.7–15.4) 15.2 (5.1–31.9) 35–44 61 51 114.2 (109.7–118.7) 123.2** (118.9–127.4) 74.7 (71.4–78.0) 75.0 (72.2–77.9) 19.7 (10.6–31.8) 9.8 (3.3–21.4) 45–54 52 40 120.7 (116.1–125.4) 141.3*** (135.5–147.1) 76.3 (73.4–79.2) 81.6* (78.3–84.8) 11.5 (4.4–23.4) 58.5*** (42.1–73.7) 55–64 34 28 127.7 (120.7–134.7) 137.7 (129.3–146.1) 75.6 (71.4–79.7) 75.7 (70.9–80.5) 23.5 (10.7–41.1) 57.1** (37.2–75.5) Linear age trend p-value <0.05 <0.05 <0.05 <0.05 ns <0.05 25–64a 228 153 117.4 (115.1–119.8) 129.0*** (126.1–131.9) 74.5 (73.0–76.0) 75.9 (74.1–77.6) 15.1 (10.4–19.7) 29.6*** (22.4–36.9) Age . Sample . Mean SBP . Mean DBP . % Hypertension . group . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 118.8 (115.9–121.7) 131.0*** (126.6–135.3) 71.6 (69.0–74.2) 74.5 (71.1–77.8) 10.0 (4.4–18.8) 17.4 (5.0–38.8) 35–44 56 23 115.6 (111.9–119.3) 130.0*** (124.5–135.6) 72.4 (70.0–74.7) 78.3** (74.4–82.2) 7.1 (2.0–17.3) 30.4* (13.2–52.9) 45–54 41 32 117.3 (112.7–121.8) 141.0*** (134.5–147.5) 75.6 (71.9–79.4) 83.1** (78.9–87.2) 12.2 (4.1–26.2) 57.6*** (39.2–74.5) 55–64 36 37 119.9 (113.9–125.9) 141.5*** (133.8–149.2) 75.1 (70.4–79.9) 82.1* (77.7–86.6) 19.4 (8.2–36.0) 64.9*** (47.5–79.8) Linear age trend p-value ns <0.05 <0.05 <0.05 ns <0.05 25–64a 213 115 117.8 (115.8–119.8) 136.2*** (133.0–139.5) 73.5 (71.9–75.1) 79.6*** (77.5–81.6) 11.7 (7.4–16.1) 43.0*** (34.0–52.3) Female 25–34 81 34 110.2 (107.2–113.2) 121.1*** (117.5–124.8) 71.6 (69.3-73.8) 72.3 (68.6-76.0) 7.4 (2.7–15.4) 15.2 (5.1–31.9) 35–44 61 51 114.2 (109.7–118.7) 123.2** (118.9–127.4) 74.7 (71.4–78.0) 75.0 (72.2–77.9) 19.7 (10.6–31.8) 9.8 (3.3–21.4) 45–54 52 40 120.7 (116.1–125.4) 141.3*** (135.5–147.1) 76.3 (73.4–79.2) 81.6* (78.3–84.8) 11.5 (4.4–23.4) 58.5*** (42.1–73.7) 55–64 34 28 127.7 (120.7–134.7) 137.7 (129.3–146.1) 75.6 (71.4–79.7) 75.7 (70.9–80.5) 23.5 (10.7–41.1) 57.1** (37.2–75.5) Linear age trend p-value <0.05 <0.05 <0.05 <0.05 ns <0.05 25–64a 228 153 117.4 (115.1–119.8) 129.0*** (126.1–131.9) 74.5 (73.0–76.0) 75.9 (74.1–77.6) 15.1 (10.4–19.7) 29.6*** (22.4–36.9) % Hypertension: prevalence of hypertension SBP ≥140 and/or DBP ≥90 mmHg and/or self-report taking medication for hypertension; 95% confidence intervals in parentheses. a Age-standardised to the 2008 Census of Wallis Island. * p < 0.05. ** p < 0.01. *** p < 0.001. SBP: systolic blood pressure; DBP: diastolic blood pressure; ns: not significant (p ≥ 0.05). Open in new tab Table 1. Mean systolic/diastolic blood pressure and hypertension prevalence, Wallis Polynesians, 1980 and 2009. Age . Sample . Mean SBP . Mean DBP . % Hypertension . group . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 118.8 (115.9–121.7) 131.0*** (126.6–135.3) 71.6 (69.0–74.2) 74.5 (71.1–77.8) 10.0 (4.4–18.8) 17.4 (5.0–38.8) 35–44 56 23 115.6 (111.9–119.3) 130.0*** (124.5–135.6) 72.4 (70.0–74.7) 78.3** (74.4–82.2) 7.1 (2.0–17.3) 30.4* (13.2–52.9) 45–54 41 32 117.3 (112.7–121.8) 141.0*** (134.5–147.5) 75.6 (71.9–79.4) 83.1** (78.9–87.2) 12.2 (4.1–26.2) 57.6*** (39.2–74.5) 55–64 36 37 119.9 (113.9–125.9) 141.5*** (133.8–149.2) 75.1 (70.4–79.9) 82.1* (77.7–86.6) 19.4 (8.2–36.0) 64.9*** (47.5–79.8) Linear age trend p-value ns <0.05 <0.05 <0.05 ns <0.05 25–64a 213 115 117.8 (115.8–119.8) 136.2*** (133.0–139.5) 73.5 (71.9–75.1) 79.6*** (77.5–81.6) 11.7 (7.4–16.1) 43.0*** (34.0–52.3) Female 25–34 81 34 110.2 (107.2–113.2) 121.1*** (117.5–124.8) 71.6 (69.3-73.8) 72.3 (68.6-76.0) 7.4 (2.7–15.4) 15.2 (5.1–31.9) 35–44 61 51 114.2 (109.7–118.7) 123.2** (118.9–127.4) 74.7 (71.4–78.0) 75.0 (72.2–77.9) 19.7 (10.6–31.8) 9.8 (3.3–21.4) 45–54 52 40 120.7 (116.1–125.4) 141.3*** (135.5–147.1) 76.3 (73.4–79.2) 81.6* (78.3–84.8) 11.5 (4.4–23.4) 58.5*** (42.1–73.7) 55–64 34 28 127.7 (120.7–134.7) 137.7 (129.3–146.1) 75.6 (71.4–79.7) 75.7 (70.9–80.5) 23.5 (10.7–41.1) 57.1** (37.2–75.5) Linear age trend p-value <0.05 <0.05 <0.05 <0.05 ns <0.05 25–64a 228 153 117.4 (115.1–119.8) 129.0*** (126.1–131.9) 74.5 (73.0–76.0) 75.9 (74.1–77.6) 15.1 (10.4–19.7) 29.6*** (22.4–36.9) Age . Sample . Mean SBP . Mean DBP . % Hypertension . group . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 118.8 (115.9–121.7) 131.0*** (126.6–135.3) 71.6 (69.0–74.2) 74.5 (71.1–77.8) 10.0 (4.4–18.8) 17.4 (5.0–38.8) 35–44 56 23 115.6 (111.9–119.3) 130.0*** (124.5–135.6) 72.4 (70.0–74.7) 78.3** (74.4–82.2) 7.1 (2.0–17.3) 30.4* (13.2–52.9) 45–54 41 32 117.3 (112.7–121.8) 141.0*** (134.5–147.5) 75.6 (71.9–79.4) 83.1** (78.9–87.2) 12.2 (4.1–26.2) 57.6*** (39.2–74.5) 55–64 36 37 119.9 (113.9–125.9) 141.5*** (133.8–149.2) 75.1 (70.4–79.9) 82.1* (77.7–86.6) 19.4 (8.2–36.0) 64.9*** (47.5–79.8) Linear age trend p-value ns <0.05 <0.05 <0.05 ns <0.05 25–64a 213 115 117.8 (115.8–119.8) 136.2*** (133.0–139.5) 73.5 (71.9–75.1) 79.6*** (77.5–81.6) 11.7 (7.4–16.1) 43.0*** (34.0–52.3) Female 25–34 81 34 110.2 (107.2–113.2) 121.1*** (117.5–124.8) 71.6 (69.3-73.8) 72.3 (68.6-76.0) 7.4 (2.7–15.4) 15.2 (5.1–31.9) 35–44 61 51 114.2 (109.7–118.7) 123.2** (118.9–127.4) 74.7 (71.4–78.0) 75.0 (72.2–77.9) 19.7 (10.6–31.8) 9.8 (3.3–21.4) 45–54 52 40 120.7 (116.1–125.4) 141.3*** (135.5–147.1) 76.3 (73.4–79.2) 81.6* (78.3–84.8) 11.5 (4.4–23.4) 58.5*** (42.1–73.7) 55–64 34 28 127.7 (120.7–134.7) 137.7 (129.3–146.1) 75.6 (71.4–79.7) 75.7 (70.9–80.5) 23.5 (10.7–41.1) 57.1** (37.2–75.5) Linear age trend p-value <0.05 <0.05 <0.05 <0.05 ns <0.05 25–64a 228 153 117.4 (115.1–119.8) 129.0*** (126.1–131.9) 74.5 (73.0–76.0) 75.9 (74.1–77.6) 15.1 (10.4–19.7) 29.6*** (22.4–36.9) % Hypertension: prevalence of hypertension SBP ≥140 and/or DBP ≥90 mmHg and/or self-report taking medication for hypertension; 95% confidence intervals in parentheses. a Age-standardised to the 2008 Census of Wallis Island. * p < 0.05. ** p < 0.01. *** p < 0.001. SBP: systolic blood pressure; DBP: diastolic blood pressure; ns: not significant (p ≥ 0.05). Open in new tab Table 2. Mean fasting plasma glucose and the prevalence of diabetes, Wallis Polynesians, 1980 and 2009. Age . Sample . Mean FPG . % Impaired FPG . % Diabetes mellitus . group . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 4.7 (4.5–4.8) 5.8 (4.6–7.1) 1.3 (0.0–6.7) 13.0* (2.8–33.6) 0.0 (0.0–5.0) 8.7* (1.1–28.0) 35–44 56 23 5.0 (4.5–5.4) 5.0 (4.7–5.3) 3.6 (0.0–12.3) 17.4 (5.0–38.8) 1.8 (0.0–9.6) 0.0 (0.0–16.0) 45–54 40 33 5.1 (4.9–5.3) 5.8* (5.1–6.5) 5.0 (0.0–16.9) 9.1 (1.9–24.3) 0.0 (0.0–9.0) 18.2** (7.0–35.5) 55–64 36 37 5.4 (4.6–6.2) 5.3 (5.0–5.6) 5.6 (0.0–18.7) 5.4 (0.6–18.2) 8.3 (1.8–22.5) 18.9 (8.0–35.1) Linear age trend p-value <0.05 ns ns ns <0.05 <0.05 25–64a 212 116 5.0 (4.8–5.2) 5.5** (5.2–5.8) 3.7 (1.1.–6.2) 10.8** (5.1–16.5) 2.3 (0.3–4.4) 12.2*** (6.2–18.2) Female 25–34 81 33 4.4 (4.2–4.6) 4.8 (4.4–5.1) 0.0 (0.0–5.0) 6.1 (0.7–20.2) 0.0 (0.0–0.5) 3.0 (0.1–15.8) 35–44 61 51 5.0 (4.7–5.4) 4.9 (4.7–5.2) 4.9 (1.0–13.7) 2.0 (0.1–10.5) 4.9 (1.0–13.7) 7.8 (2.2–18.9) 45–54 52 41 5.2 (4.8–5.6) 6.1 (5.3–6.8) 7.7 (2.1–18.5) 12.2 (4.1–26.2) 5.8 (1.2–16.0) 26.8** (14.2–42.9) 55–64 34 28 5.7 (4.6–6.8) 6.2 (5.5–6.9) 5.9 (0.7–19.7) 7.1 (0.9–23.5) 5.9 (0.7–19.7) 39.3** (21.5–59.4) Linear age trend p-value <0.05 <0.05 <0.05 ns ns <0.05 25–64a 228 153 5.0 (4.8–5.3) 5.3 (5.1–5.6) 4.5 (1.9–7.2) 6.3 (2.5–10.1) 4.0(1.5–6.6) 15.8*** (10.1–21.6) Age . Sample . Mean FPG . % Impaired FPG . % Diabetes mellitus . group . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 4.7 (4.5–4.8) 5.8 (4.6–7.1) 1.3 (0.0–6.7) 13.0* (2.8–33.6) 0.0 (0.0–5.0) 8.7* (1.1–28.0) 35–44 56 23 5.0 (4.5–5.4) 5.0 (4.7–5.3) 3.6 (0.0–12.3) 17.4 (5.0–38.8) 1.8 (0.0–9.6) 0.0 (0.0–16.0) 45–54 40 33 5.1 (4.9–5.3) 5.8* (5.1–6.5) 5.0 (0.0–16.9) 9.1 (1.9–24.3) 0.0 (0.0–9.0) 18.2** (7.0–35.5) 55–64 36 37 5.4 (4.6–6.2) 5.3 (5.0–5.6) 5.6 (0.0–18.7) 5.4 (0.6–18.2) 8.3 (1.8–22.5) 18.9 (8.0–35.1) Linear age trend p-value <0.05 ns ns ns <0.05 <0.05 25–64a 212 116 5.0 (4.8–5.2) 5.5** (5.2–5.8) 3.7 (1.1.–6.2) 10.8** (5.1–16.5) 2.3 (0.3–4.4) 12.2*** (6.2–18.2) Female 25–34 81 33 4.4 (4.2–4.6) 4.8 (4.4–5.1) 0.0 (0.0–5.0) 6.1 (0.7–20.2) 0.0 (0.0–0.5) 3.0 (0.1–15.8) 35–44 61 51 5.0 (4.7–5.4) 4.9 (4.7–5.2) 4.9 (1.0–13.7) 2.0 (0.1–10.5) 4.9 (1.0–13.7) 7.8 (2.2–18.9) 45–54 52 41 5.2 (4.8–5.6) 6.1 (5.3–6.8) 7.7 (2.1–18.5) 12.2 (4.1–26.2) 5.8 (1.2–16.0) 26.8** (14.2–42.9) 55–64 34 28 5.7 (4.6–6.8) 6.2 (5.5–6.9) 5.9 (0.7–19.7) 7.1 (0.9–23.5) 5.9 (0.7–19.7) 39.3** (21.5–59.4) Linear age trend p-value <0.05 <0.05 <0.05 ns ns <0.05 25–64a 228 153 5.0 (4.8–5.3) 5.3 (5.1–5.6) 4.5 (1.9–7.2) 6.3 (2.5–10.1) 4.0(1.5–6.6) 15.8*** (10.1–21.6) % Impaired FPG: prevalence of FPG ≥6.1 and <7.0 mmol/l and not taking medication for type 2 diabetes mellitus (T2DM); % diabetes mellitus: prevalence of FPG ≥7.0 mmol/l and/or taking medication for T2DM; 95% confidence intervals in parentheses. a Age-standardised to the 2008 Census of Wallis Island. * p < 0.05. ** p < 0.01. *** p < 0.001. FPG: fasting plasma glucose; ns: not significant (p ≥ 0.05). Open in new tab Table 2. Mean fasting plasma glucose and the prevalence of diabetes, Wallis Polynesians, 1980 and 2009. Age . Sample . Mean FPG . % Impaired FPG . % Diabetes mellitus . group . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 4.7 (4.5–4.8) 5.8 (4.6–7.1) 1.3 (0.0–6.7) 13.0* (2.8–33.6) 0.0 (0.0–5.0) 8.7* (1.1–28.0) 35–44 56 23 5.0 (4.5–5.4) 5.0 (4.7–5.3) 3.6 (0.0–12.3) 17.4 (5.0–38.8) 1.8 (0.0–9.6) 0.0 (0.0–16.0) 45–54 40 33 5.1 (4.9–5.3) 5.8* (5.1–6.5) 5.0 (0.0–16.9) 9.1 (1.9–24.3) 0.0 (0.0–9.0) 18.2** (7.0–35.5) 55–64 36 37 5.4 (4.6–6.2) 5.3 (5.0–5.6) 5.6 (0.0–18.7) 5.4 (0.6–18.2) 8.3 (1.8–22.5) 18.9 (8.0–35.1) Linear age trend p-value <0.05 ns ns ns <0.05 <0.05 25–64a 212 116 5.0 (4.8–5.2) 5.5** (5.2–5.8) 3.7 (1.1.–6.2) 10.8** (5.1–16.5) 2.3 (0.3–4.4) 12.2*** (6.2–18.2) Female 25–34 81 33 4.4 (4.2–4.6) 4.8 (4.4–5.1) 0.0 (0.0–5.0) 6.1 (0.7–20.2) 0.0 (0.0–0.5) 3.0 (0.1–15.8) 35–44 61 51 5.0 (4.7–5.4) 4.9 (4.7–5.2) 4.9 (1.0–13.7) 2.0 (0.1–10.5) 4.9 (1.0–13.7) 7.8 (2.2–18.9) 45–54 52 41 5.2 (4.8–5.6) 6.1 (5.3–6.8) 7.7 (2.1–18.5) 12.2 (4.1–26.2) 5.8 (1.2–16.0) 26.8** (14.2–42.9) 55–64 34 28 5.7 (4.6–6.8) 6.2 (5.5–6.9) 5.9 (0.7–19.7) 7.1 (0.9–23.5) 5.9 (0.7–19.7) 39.3** (21.5–59.4) Linear age trend p-value <0.05 <0.05 <0.05 ns ns <0.05 25–64a 228 153 5.0 (4.8–5.3) 5.3 (5.1–5.6) 4.5 (1.9–7.2) 6.3 (2.5–10.1) 4.0(1.5–6.6) 15.8*** (10.1–21.6) Age . Sample . Mean FPG . % Impaired FPG . % Diabetes mellitus . group . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 4.7 (4.5–4.8) 5.8 (4.6–7.1) 1.3 (0.0–6.7) 13.0* (2.8–33.6) 0.0 (0.0–5.0) 8.7* (1.1–28.0) 35–44 56 23 5.0 (4.5–5.4) 5.0 (4.7–5.3) 3.6 (0.0–12.3) 17.4 (5.0–38.8) 1.8 (0.0–9.6) 0.0 (0.0–16.0) 45–54 40 33 5.1 (4.9–5.3) 5.8* (5.1–6.5) 5.0 (0.0–16.9) 9.1 (1.9–24.3) 0.0 (0.0–9.0) 18.2** (7.0–35.5) 55–64 36 37 5.4 (4.6–6.2) 5.3 (5.0–5.6) 5.6 (0.0–18.7) 5.4 (0.6–18.2) 8.3 (1.8–22.5) 18.9 (8.0–35.1) Linear age trend p-value <0.05 ns ns ns <0.05 <0.05 25–64a 212 116 5.0 (4.8–5.2) 5.5** (5.2–5.8) 3.7 (1.1.–6.2) 10.8** (5.1–16.5) 2.3 (0.3–4.4) 12.2*** (6.2–18.2) Female 25–34 81 33 4.4 (4.2–4.6) 4.8 (4.4–5.1) 0.0 (0.0–5.0) 6.1 (0.7–20.2) 0.0 (0.0–0.5) 3.0 (0.1–15.8) 35–44 61 51 5.0 (4.7–5.4) 4.9 (4.7–5.2) 4.9 (1.0–13.7) 2.0 (0.1–10.5) 4.9 (1.0–13.7) 7.8 (2.2–18.9) 45–54 52 41 5.2 (4.8–5.6) 6.1 (5.3–6.8) 7.7 (2.1–18.5) 12.2 (4.1–26.2) 5.8 (1.2–16.0) 26.8** (14.2–42.9) 55–64 34 28 5.7 (4.6–6.8) 6.2 (5.5–6.9) 5.9 (0.7–19.7) 7.1 (0.9–23.5) 5.9 (0.7–19.7) 39.3** (21.5–59.4) Linear age trend p-value <0.05 <0.05 <0.05 ns ns <0.05 25–64a 228 153 5.0 (4.8–5.3) 5.3 (5.1–5.6) 4.5 (1.9–7.2) 6.3 (2.5–10.1) 4.0(1.5–6.6) 15.8*** (10.1–21.6) % Impaired FPG: prevalence of FPG ≥6.1 and <7.0 mmol/l and not taking medication for type 2 diabetes mellitus (T2DM); % diabetes mellitus: prevalence of FPG ≥7.0 mmol/l and/or taking medication for T2DM; 95% confidence intervals in parentheses. a Age-standardised to the 2008 Census of Wallis Island. * p < 0.05. ** p < 0.01. *** p < 0.001. FPG: fasting plasma glucose; ns: not significant (p ≥ 0.05). Open in new tab Table 3. Mean body mass index and obesity prevalence, Wallis Polynesians, 1980 and 2009. Age group . Sample . Mean BMI . % Obese BMI≥30 . %Obese BMI>32 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 26.6 (25.9–27.3) 32.3*** (29.7–35.0) 11.3 (5.3–20.3) 52.2*** (30.6–73.2) 6.3 (2.1–14.0) 47.8*** (26.8–69.4) 35–44 56 23 27.8 (26.5–29.2) 33.5*** (30.7–36.3) 26.8 (15.8–40.3) 69.6*** (47.1–86.8) 16.1 (7.6–28.3) 52.2*** (30.6–73.2) 45–54 41 33 28.1 (26.3–30.0) 32.7** (30.3–35.0) 31.7 (18.1–48.1) 72.7*** (54.5–86.7) 24.4 (12.4–40.0) 45.5 28.1–63.7) 55–64 36 37 27.3 (25.5–29.1) 31.7*** (30.0–33.4) 33.3 (18.6–51.0) 54.1 (36.9–70.5) 11.1 (3.1–26.1) 37.8* (22.5–55.2) Linear age trend p-value ns ns <0.05 ns ns ns 25–64a 213 116 27.4 (26.8–28.1) 32.5*** (31.4–33.6) 25.0 (19.2–30.8) 61.7*** (52.8–70.6) 14.1 (9.4–18.8) 45.4*** (36.2–54.5) Female 25–34 74 34 29.0 (27.9–30.0) 36.5*** (33.5–39.5) 37.8 (26.8–49.9) 73.5*** (55.6–87.1) 29.7 (19.6–41.5) 73.5*** (55.6–87.1) 35–44 52 51 30.9 (29.4–32.4) 33.9** (32.3–35.5) 53.8 (39.5–67.7) 80.4** (66.9–90.2) 40.4 (27.0–54.9) 70.6** (56.2–82.5) 45–54 45 41 29.5 (27.8–31.3) 35.3*** (33.4–37.2) 53.3 (37.9–68.3) 85.4*** (70.8–94.4) 37.8 (23.8–53.5) 73.2*** (57.1–85.8) 55–64 30 28 30.3 (28.3–32.3) 32.5 (30.1–34.8) 50.0 (31.3–68.7) 64.3 (44.1–81.4) 36.7 (19.9–56.1) 39.3 (21.5–59.4) Linear age trend p-value ns ns ns ns ns <0.05 25–64a 201 154 29.9 (29.2–30.6) 34.7*** (33.6–35.8) 48.6 (41.7–55.5) 77.2*** (70.7–83.8) 36.1 (29.5–42.7) 67.3*** (59.9–74.7) Age group . Sample . Mean BMI . % Obese BMI≥30 . %Obese BMI>32 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 26.6 (25.9–27.3) 32.3*** (29.7–35.0) 11.3 (5.3–20.3) 52.2*** (30.6–73.2) 6.3 (2.1–14.0) 47.8*** (26.8–69.4) 35–44 56 23 27.8 (26.5–29.2) 33.5*** (30.7–36.3) 26.8 (15.8–40.3) 69.6*** (47.1–86.8) 16.1 (7.6–28.3) 52.2*** (30.6–73.2) 45–54 41 33 28.1 (26.3–30.0) 32.7** (30.3–35.0) 31.7 (18.1–48.1) 72.7*** (54.5–86.7) 24.4 (12.4–40.0) 45.5 28.1–63.7) 55–64 36 37 27.3 (25.5–29.1) 31.7*** (30.0–33.4) 33.3 (18.6–51.0) 54.1 (36.9–70.5) 11.1 (3.1–26.1) 37.8* (22.5–55.2) Linear age trend p-value ns ns <0.05 ns ns ns 25–64a 213 116 27.4 (26.8–28.1) 32.5*** (31.4–33.6) 25.0 (19.2–30.8) 61.7*** (52.8–70.6) 14.1 (9.4–18.8) 45.4*** (36.2–54.5) Female 25–34 74 34 29.0 (27.9–30.0) 36.5*** (33.5–39.5) 37.8 (26.8–49.9) 73.5*** (55.6–87.1) 29.7 (19.6–41.5) 73.5*** (55.6–87.1) 35–44 52 51 30.9 (29.4–32.4) 33.9** (32.3–35.5) 53.8 (39.5–67.7) 80.4** (66.9–90.2) 40.4 (27.0–54.9) 70.6** (56.2–82.5) 45–54 45 41 29.5 (27.8–31.3) 35.3*** (33.4–37.2) 53.3 (37.9–68.3) 85.4*** (70.8–94.4) 37.8 (23.8–53.5) 73.2*** (57.1–85.8) 55–64 30 28 30.3 (28.3–32.3) 32.5 (30.1–34.8) 50.0 (31.3–68.7) 64.3 (44.1–81.4) 36.7 (19.9–56.1) 39.3 (21.5–59.4) Linear age trend p-value ns ns ns ns ns <0.05 25–64a 201 154 29.9 (29.2–30.6) 34.7*** (33.6–35.8) 48.6 (41.7–55.5) 77.2*** (70.7–83.8) 36.1 (29.5–42.7) 67.3*** (59.9–74.7) % Obese BMI ≥30: prevalence obesity according to the standard definition of the World Health Organisation;12 % Obese BMI>32: prevalence of obesity as suggested by Swinburn et al.13 for Polynesian populations; 95% confidence intervals in parentheses; pregnant women excluded. a Age-standardised to the 2008 Census of Wallis Island. * p < 0.05. ** p < 0.01. *** p < 0.001. BMI: body mass index; ns: not significant (p ≥ 0.05). Open in new tab Table 3. Mean body mass index and obesity prevalence, Wallis Polynesians, 1980 and 2009. Age group . Sample . Mean BMI . % Obese BMI≥30 . %Obese BMI>32 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 26.6 (25.9–27.3) 32.3*** (29.7–35.0) 11.3 (5.3–20.3) 52.2*** (30.6–73.2) 6.3 (2.1–14.0) 47.8*** (26.8–69.4) 35–44 56 23 27.8 (26.5–29.2) 33.5*** (30.7–36.3) 26.8 (15.8–40.3) 69.6*** (47.1–86.8) 16.1 (7.6–28.3) 52.2*** (30.6–73.2) 45–54 41 33 28.1 (26.3–30.0) 32.7** (30.3–35.0) 31.7 (18.1–48.1) 72.7*** (54.5–86.7) 24.4 (12.4–40.0) 45.5 28.1–63.7) 55–64 36 37 27.3 (25.5–29.1) 31.7*** (30.0–33.4) 33.3 (18.6–51.0) 54.1 (36.9–70.5) 11.1 (3.1–26.1) 37.8* (22.5–55.2) Linear age trend p-value ns ns <0.05 ns ns ns 25–64a 213 116 27.4 (26.8–28.1) 32.5*** (31.4–33.6) 25.0 (19.2–30.8) 61.7*** (52.8–70.6) 14.1 (9.4–18.8) 45.4*** (36.2–54.5) Female 25–34 74 34 29.0 (27.9–30.0) 36.5*** (33.5–39.5) 37.8 (26.8–49.9) 73.5*** (55.6–87.1) 29.7 (19.6–41.5) 73.5*** (55.6–87.1) 35–44 52 51 30.9 (29.4–32.4) 33.9** (32.3–35.5) 53.8 (39.5–67.7) 80.4** (66.9–90.2) 40.4 (27.0–54.9) 70.6** (56.2–82.5) 45–54 45 41 29.5 (27.8–31.3) 35.3*** (33.4–37.2) 53.3 (37.9–68.3) 85.4*** (70.8–94.4) 37.8 (23.8–53.5) 73.2*** (57.1–85.8) 55–64 30 28 30.3 (28.3–32.3) 32.5 (30.1–34.8) 50.0 (31.3–68.7) 64.3 (44.1–81.4) 36.7 (19.9–56.1) 39.3 (21.5–59.4) Linear age trend p-value ns ns ns ns ns <0.05 25–64a 201 154 29.9 (29.2–30.6) 34.7*** (33.6–35.8) 48.6 (41.7–55.5) 77.2*** (70.7–83.8) 36.1 (29.5–42.7) 67.3*** (59.9–74.7) Age group . Sample . Mean BMI . % Obese BMI≥30 . %Obese BMI>32 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 80 23 26.6 (25.9–27.3) 32.3*** (29.7–35.0) 11.3 (5.3–20.3) 52.2*** (30.6–73.2) 6.3 (2.1–14.0) 47.8*** (26.8–69.4) 35–44 56 23 27.8 (26.5–29.2) 33.5*** (30.7–36.3) 26.8 (15.8–40.3) 69.6*** (47.1–86.8) 16.1 (7.6–28.3) 52.2*** (30.6–73.2) 45–54 41 33 28.1 (26.3–30.0) 32.7** (30.3–35.0) 31.7 (18.1–48.1) 72.7*** (54.5–86.7) 24.4 (12.4–40.0) 45.5 28.1–63.7) 55–64 36 37 27.3 (25.5–29.1) 31.7*** (30.0–33.4) 33.3 (18.6–51.0) 54.1 (36.9–70.5) 11.1 (3.1–26.1) 37.8* (22.5–55.2) Linear age trend p-value ns ns <0.05 ns ns ns 25–64a 213 116 27.4 (26.8–28.1) 32.5*** (31.4–33.6) 25.0 (19.2–30.8) 61.7*** (52.8–70.6) 14.1 (9.4–18.8) 45.4*** (36.2–54.5) Female 25–34 74 34 29.0 (27.9–30.0) 36.5*** (33.5–39.5) 37.8 (26.8–49.9) 73.5*** (55.6–87.1) 29.7 (19.6–41.5) 73.5*** (55.6–87.1) 35–44 52 51 30.9 (29.4–32.4) 33.9** (32.3–35.5) 53.8 (39.5–67.7) 80.4** (66.9–90.2) 40.4 (27.0–54.9) 70.6** (56.2–82.5) 45–54 45 41 29.5 (27.8–31.3) 35.3*** (33.4–37.2) 53.3 (37.9–68.3) 85.4*** (70.8–94.4) 37.8 (23.8–53.5) 73.2*** (57.1–85.8) 55–64 30 28 30.3 (28.3–32.3) 32.5 (30.1–34.8) 50.0 (31.3–68.7) 64.3 (44.1–81.4) 36.7 (19.9–56.1) 39.3 (21.5–59.4) Linear age trend p-value ns ns ns ns ns <0.05 25–64a 201 154 29.9 (29.2–30.6) 34.7*** (33.6–35.8) 48.6 (41.7–55.5) 77.2*** (70.7–83.8) 36.1 (29.5–42.7) 67.3*** (59.9–74.7) % Obese BMI ≥30: prevalence obesity according to the standard definition of the World Health Organisation;12 % Obese BMI>32: prevalence of obesity as suggested by Swinburn et al.13 for Polynesian populations; 95% confidence intervals in parentheses; pregnant women excluded. a Age-standardised to the 2008 Census of Wallis Island. * p < 0.05. ** p < 0.01. *** p < 0.001. BMI: body mass index; ns: not significant (p ≥ 0.05). Open in new tab Table 4. Mean cholesterol and elevated cholesterol and triglyceride prevalence, Wallis Polynesians, 1980 and 2009. Age group . Sample . Mean cholesterol . % Cholesterol ≥5.2 . % Triglycerides ≥1.7 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 78 23 3.7 (3.5–3.9) 4.5** (4.0–5.0) 6.4 (2.1–14.3) 30.4** (13.2–52.9) 1.3 (0.0–6.9) 26.1*** (10.2–48.4) 35–44 56 23 3.8 (3.6–4.0) 4.3** (3.9–4.7) 3.6 (0.4–12.3) 13.0 (2.8–33.6) 1.8 (0.0–9.5) 43.5*** (23.2–65.5) 45–54 39 32 4.0 (3.7–4.3) 4.5* (4.2–4.9) 12.8 (4.3–27.4) 21.9 (9.3–40.0) 2.5 (0.0–13.2) 34.4*** (18.6–53.2) 55–64 35 37 4.0 (3.7–4.3) 4.3 (4.1–4.6) 5.7 (0.7–19.2) 16.2 (6.2–32.0) 2.9 (0.0–14.9) 27.0** (13.8–44.1) Linear age trend p-value <0.05 ns ns ns ns ns 25–64a 208 115 3.9 (3.8–4.0) 4.4*** (4.2–4.6) 6.8 (3.4–10.3) 20.7*** (13.2–28.2) 2.0 (0.1–4.0) 32.1*** (23.5–40.7) Female 25–34 71 34 4.2 (3.9–4.5) 3.8 (3.5–4.1) 15.5 (8.0–26.0) 2.9 (0.0–15.3) 9.7 (4.0–19.0) 20.6 (8.7–37.9) 35–44 53 51 4.2 (3.9–4.4) 3.8* (3.6–4.0) 17.0 (8.0–29.8) 9.8 (3.3–21.4) 1.9 (0.0–10.0) 15.7* (7.0–28.6) 45–54 50 41 4.3 (4.0–4.6) 4.0 (3.8–4.3) 20.0 (10.0–33.7) 7.3 (1.5–19.9) 6.0 (1.3–16.6) 26.8** (14.2–42.9) 55–64 32 28 4.6 (4.2–4.9) 4.0* (3.7–4.4) 15.6 (5..3–32.8) 10.7 (2.3–28.2) 3.1 (0.0–16.2) 10.7 (2.3–28.2) Linear age trend p-value ns ns ns ns ns ns 25–64a 206 154 4.3 (4.2–4.4) 3.9** (3.8–4.0) 17.2 (12.1–22.3) 7.5** (3.3–11.6) 5.3 (2.2–8.3) 18.9*** (12.8–25.1) Age group . Sample . Mean cholesterol . % Cholesterol ≥5.2 . % Triglycerides ≥1.7 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 78 23 3.7 (3.5–3.9) 4.5** (4.0–5.0) 6.4 (2.1–14.3) 30.4** (13.2–52.9) 1.3 (0.0–6.9) 26.1*** (10.2–48.4) 35–44 56 23 3.8 (3.6–4.0) 4.3** (3.9–4.7) 3.6 (0.4–12.3) 13.0 (2.8–33.6) 1.8 (0.0–9.5) 43.5*** (23.2–65.5) 45–54 39 32 4.0 (3.7–4.3) 4.5* (4.2–4.9) 12.8 (4.3–27.4) 21.9 (9.3–40.0) 2.5 (0.0–13.2) 34.4*** (18.6–53.2) 55–64 35 37 4.0 (3.7–4.3) 4.3 (4.1–4.6) 5.7 (0.7–19.2) 16.2 (6.2–32.0) 2.9 (0.0–14.9) 27.0** (13.8–44.1) Linear age trend p-value <0.05 ns ns ns ns ns 25–64a 208 115 3.9 (3.8–4.0) 4.4*** (4.2–4.6) 6.8 (3.4–10.3) 20.7*** (13.2–28.2) 2.0 (0.1–4.0) 32.1*** (23.5–40.7) Female 25–34 71 34 4.2 (3.9–4.5) 3.8 (3.5–4.1) 15.5 (8.0–26.0) 2.9 (0.0–15.3) 9.7 (4.0–19.0) 20.6 (8.7–37.9) 35–44 53 51 4.2 (3.9–4.4) 3.8* (3.6–4.0) 17.0 (8.0–29.8) 9.8 (3.3–21.4) 1.9 (0.0–10.0) 15.7* (7.0–28.6) 45–54 50 41 4.3 (4.0–4.6) 4.0 (3.8–4.3) 20.0 (10.0–33.7) 7.3 (1.5–19.9) 6.0 (1.3–16.6) 26.8** (14.2–42.9) 55–64 32 28 4.6 (4.2–4.9) 4.0* (3.7–4.4) 15.6 (5..3–32.8) 10.7 (2.3–28.2) 3.1 (0.0–16.2) 10.7 (2.3–28.2) Linear age trend p-value ns ns ns ns ns ns 25–64a 206 154 4.3 (4.2–4.4) 3.9** (3.8–4.0) 17.2 (12.1–22.3) 7.5** (3.3–11.6) 5.3 (2.2–8.3) 18.9*** (12.8–25.1) % Cholesterol ≥5.2: prevalence of serum cholesterol ≥5.2 mmol/l; % Triglycerides ≥1.7 prevalence of triglycerides ≥1.7 mmol/l; 95% confidence intervals in parentheses. a Age-standardised to the 2008 Census of Wallis Island. * p < 0.05. ** p < 0.01. *** p < 0.001. ns: not significant (p ≥ 0.05). Open in new tab Table 4. Mean cholesterol and elevated cholesterol and triglyceride prevalence, Wallis Polynesians, 1980 and 2009. Age group . Sample . Mean cholesterol . % Cholesterol ≥5.2 . % Triglycerides ≥1.7 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 78 23 3.7 (3.5–3.9) 4.5** (4.0–5.0) 6.4 (2.1–14.3) 30.4** (13.2–52.9) 1.3 (0.0–6.9) 26.1*** (10.2–48.4) 35–44 56 23 3.8 (3.6–4.0) 4.3** (3.9–4.7) 3.6 (0.4–12.3) 13.0 (2.8–33.6) 1.8 (0.0–9.5) 43.5*** (23.2–65.5) 45–54 39 32 4.0 (3.7–4.3) 4.5* (4.2–4.9) 12.8 (4.3–27.4) 21.9 (9.3–40.0) 2.5 (0.0–13.2) 34.4*** (18.6–53.2) 55–64 35 37 4.0 (3.7–4.3) 4.3 (4.1–4.6) 5.7 (0.7–19.2) 16.2 (6.2–32.0) 2.9 (0.0–14.9) 27.0** (13.8–44.1) Linear age trend p-value <0.05 ns ns ns ns ns 25–64a 208 115 3.9 (3.8–4.0) 4.4*** (4.2–4.6) 6.8 (3.4–10.3) 20.7*** (13.2–28.2) 2.0 (0.1–4.0) 32.1*** (23.5–40.7) Female 25–34 71 34 4.2 (3.9–4.5) 3.8 (3.5–4.1) 15.5 (8.0–26.0) 2.9 (0.0–15.3) 9.7 (4.0–19.0) 20.6 (8.7–37.9) 35–44 53 51 4.2 (3.9–4.4) 3.8* (3.6–4.0) 17.0 (8.0–29.8) 9.8 (3.3–21.4) 1.9 (0.0–10.0) 15.7* (7.0–28.6) 45–54 50 41 4.3 (4.0–4.6) 4.0 (3.8–4.3) 20.0 (10.0–33.7) 7.3 (1.5–19.9) 6.0 (1.3–16.6) 26.8** (14.2–42.9) 55–64 32 28 4.6 (4.2–4.9) 4.0* (3.7–4.4) 15.6 (5..3–32.8) 10.7 (2.3–28.2) 3.1 (0.0–16.2) 10.7 (2.3–28.2) Linear age trend p-value ns ns ns ns ns ns 25–64a 206 154 4.3 (4.2–4.4) 3.9** (3.8–4.0) 17.2 (12.1–22.3) 7.5** (3.3–11.6) 5.3 (2.2–8.3) 18.9*** (12.8–25.1) Age group . Sample . Mean cholesterol . % Cholesterol ≥5.2 . % Triglycerides ≥1.7 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . 1980 . 2009 . Male 25–34 78 23 3.7 (3.5–3.9) 4.5** (4.0–5.0) 6.4 (2.1–14.3) 30.4** (13.2–52.9) 1.3 (0.0–6.9) 26.1*** (10.2–48.4) 35–44 56 23 3.8 (3.6–4.0) 4.3** (3.9–4.7) 3.6 (0.4–12.3) 13.0 (2.8–33.6) 1.8 (0.0–9.5) 43.5*** (23.2–65.5) 45–54 39 32 4.0 (3.7–4.3) 4.5* (4.2–4.9) 12.8 (4.3–27.4) 21.9 (9.3–40.0) 2.5 (0.0–13.2) 34.4*** (18.6–53.2) 55–64 35 37 4.0 (3.7–4.3) 4.3 (4.1–4.6) 5.7 (0.7–19.2) 16.2 (6.2–32.0) 2.9 (0.0–14.9) 27.0** (13.8–44.1) Linear age trend p-value <0.05 ns ns ns ns ns 25–64a 208 115 3.9 (3.8–4.0) 4.4*** (4.2–4.6) 6.8 (3.4–10.3) 20.7*** (13.2–28.2) 2.0 (0.1–4.0) 32.1*** (23.5–40.7) Female 25–34 71 34 4.2 (3.9–4.5) 3.8 (3.5–4.1) 15.5 (8.0–26.0) 2.9 (0.0–15.3) 9.7 (4.0–19.0) 20.6 (8.7–37.9) 35–44 53 51 4.2 (3.9–4.4) 3.8* (3.6–4.0) 17.0 (8.0–29.8) 9.8 (3.3–21.4) 1.9 (0.0–10.0) 15.7* (7.0–28.6) 45–54 50 41 4.3 (4.0–4.6) 4.0 (3.8–4.3) 20.0 (10.0–33.7) 7.3 (1.5–19.9) 6.0 (1.3–16.6) 26.8** (14.2–42.9) 55–64 32 28 4.6 (4.2–4.9) 4.0* (3.7–4.4) 15.6 (5..3–32.8) 10.7 (2.3–28.2) 3.1 (0.0–16.2) 10.7 (2.3–28.2) Linear age trend p-value ns ns ns ns ns ns 25–64a 206 154 4.3 (4.2–4.4) 3.9** (3.8–4.0) 17.2 (12.1–22.3) 7.5** (3.3–11.6) 5.3 (2.2–8.3) 18.9*** (12.8–25.1) % Cholesterol ≥5.2: prevalence of serum cholesterol ≥5.2 mmol/l; % Triglycerides ≥1.7 prevalence of triglycerides ≥1.7 mmol/l; 95% confidence intervals in parentheses. a Age-standardised to the 2008 Census of Wallis Island. * p < 0.05. ** p < 0.01. *** p < 0.001. ns: not significant (p ≥ 0.05). Open in new tab BMI increased in men from 1980 (mean 27.4, SD 4.8; median 26.5, IQR 24.0–30.0) to 2009 (mean 32.5, SD 6.0; median 31.3, IQR 28.7–35.4); and in women from 1980 (mean 29.9, SD 5.3; median 29.8, IQR 25.9–33.6) to 2009 (mean 34.7, SD 6.8; median 34.6, IQR 30.7–39.0). Frequency distributions broadened and shifted to the right (Figure 1). Figure 1. Open in new tabDownload slide Body mass index Wallis Polynesians 1980 and 2009. Normal (Gaussian) distributions derived from frequency distributions of body mass index expressed as proportions of participants by sex and survey. Participants: 1980 men n = 213, women n = 201; 2009 men n = 116, women n = 154. Means and prevalences of cholesterol (age-standardised) significantly increased in men and decreased in women during 1980–2009, and by 2009 cholesterol had become significantly higher in men (Table 4). Sex-specific differences in BP emerged during 1980–2009 with significantly higher means and prevalences in men in 2009 (Table 1). Mean BMI and obesity prevalence, using BMI ≥3012 or BMI >32,13 remained significantly higher in women than men in 1980 and 2009 (Table 3). In 1980 there was no significant increase in mean SBP with age in men, and no significant increase in hypertension prevalence with age for either sex; whilst in 2009 both sexes demonstrated significant increases in mean SBP and DBP, and hypertension prevalence, with increasing age (Table 1). For hypertension, the OR for the period effect in men in 2009, compared with 1980 (referent), after adjusting for age, was 5.9 (95% confidence interval (CI) 3.3–10.5), and after adjusting for age and BMI was 3.4 (1.8–6.4): a 42% change. For hypertension in women, the OR for period, after adjusting for age, was 3.3 (1.9–5.8), and after adjusting for age and BMI was 2.2 (1.2–4.1): a 33% change. For T2DM the OR for the period effect in men for 2009, compared with 1980 (referent), after adjusting for age, was 6.5 (2.1–20.3), and after adjusting for age and BMI was 4.3 (1.3–14.0): a 35% change. For T2DM in women the OR after adjusting for age was 7.4 (3.0–18.7), and after adjusting for age and BMI was 5.5 (2.1–14.3): a 26% change. Discussion Between 1980 and 2009 the prevalence of hypertension increased more than three-fold in men and two-fold in women, with 42% of the increase in men and 33% of the increase in women statistically explained by increases in BMI. In 1980 the prevalence of hypertension in both sexes, and mean SPB in males, did not rise with age, whilst in 2009 there was a significant rise with age in BP means and prevalences in both sexes. Previous studies have found a greater rise in BP with age in urbanised populations compared with rural populations with a mainly subsistence economy.15,16 Sodium intake is a known population risk factor for hypertension.17 Although not an accurate measure of sodium intake and excretion, urinary sodium concentrations in the 1980 survey suggests that sodium consumption in Wallis was low (men: 1.8; women: 2.1 g/l) in comparison with Wallis Polynesians living in the more modernised Noumea (men: 3.8; women: 2.9 g/l).8 This is consistent with higher BP in populations in modernised societies, which has occurred in Wallis over the past 30 years. The 2009 survey did not collect data on sodium consumption. In 2009, alcohol consumption on Wallis Island, also a known risk factor for hypertension,18,19 was 22% for drinking alcohol at least once a week, and 14% were classified as heavy drinkers – at least three drinks at a time at least once a week.9 The 1980 survey did not collect data on alcohol consumption. In 1980 the prevalence of T2DM in Wallis was relatively low, and comparable to other rural and traditional Polynesian populations at that time, including rural Western Samoa1 and Tuvalu.20 It was suggested that in light of the prevalence of obesity in Wallis, this relatively low prevalence of T2DM may indicate that physical activity and/or the traditional diet were protective in these circumstances,21 although the duration of obesity may yet have been insufficient to produce a corresponding increase in T2DM. Between 1980 and 2009 the prevalence of T2DM increased more than five-fold in men and almost four-fold in women, with 35% of the increase in men and 26% of the increase in women statistically explained by increases in BMI. This is similar to previous comparable studies using repeated cross-sectional samples of individuals in Mauritius,10 the United States of America22 and Europe,23 which found that increases in FPG and T2DM were partly or largely explicable by increases in BMI. Using the WHO standard definition for obesity (BMI ≥30) or the cut-point suggested by Swinburn et al.13 for Polynesian populations (BMI >32), obesity was already an issue in 1980 in Wallisian women of all age groups and in middle aged men. This indicates that modernisation of way of life was occurring in Wallis in the early 1980s, and as observed in other Polynesian populations, such as Tuvalu,20 women were affected by obesity earlier than men. BMI frequency distributions indicate that the population distribution of BMI in both sexes between 1980 and 2009 has shifted to the right (higher values) with broadening of the shape of the distribution, but no change in the difference between the mean and median, suggesting a total population phenomenon, rather than development of obesity in a minority. Although no dietary survey data are available for Wallis to examine changes in consumption of saturated fats during 1980–2009, a possible explanation for the decline in blood cholesterol in women during this period could be a substantial decline in the consumption of coconut products, an important component of calories (energy intake) in some Pacific populations. Dietary studies of other Pacific Island populations around the 1980s, including Tokelau, Pukapuka and Kiribati, identified women as having higher dietary intake of saturated fats24 and higher serum cholesterol levels24–26 in comparison with men, which was considered to be related to greater consumption of coconut products.25,26 In Pacific Island populations where coconut was traditionally a staple dietary component, consumption decreased when more convenient cooked foods (especially rice and flour) become readily available,27,28 and a similar trend may have occurred in Wallis. Further investigation is needed to identify sex-specific explanations of the observed changes in serum cholesterol. Tobacco consumption is a known risk factor for CVD. In the 2009 survey nearly 50% of participants reported smoking tobacco over the previous 12 months, and 43% reported daily smoking. The 1980 survey did not collect data on smoking habits. In 2009, 25% of men and 47% of women reported low levels of physical activity (less than 600 MET-minutes of physical activity per week), also a risk factor for CVD. These data were not collected in 1980, although with less subsistence activity it is likely that physical activity has decreased over the past 30 years. The 1980 and 2009 surveys of CVD risk factors in Wallis are both based on randomly selected samples from a homogenous population on one island, used similar methods for data collection, and are analysed according to the same definitions for hypertension, T2DM and obesity. Potential limitations of the present study are that random sample selection was employed at the village level in 1980 and at the household level in 2009. However, both surveys were representative of the population at the nearest previous Census. The 2009 sample in the present study is smaller than the 1980 sample because of exclusion of Futunans from the 2009 survey to ensure comparability of participants from Wallis Island only. However, statistical testing is used to assess differences between samples. Modifications over time have occurred in biological analysis of FPG and plasma cholesterol and triglycerides, mainly resulting in the reduction of interferences.29,30 There are larger coefficients of variation for some 1980 measures compared with 2009, with the 2009 readings likely to be more precise, especially for cholesterol and triglycerides. Comparison of the 1980 and 2009 surveys show that in the formerly traditional island society of Wallis, considerable increases in risk factors for CVD have occurred over the past 30 years consistent with changes in way of life. The decline in mean cholesterol in women may be due to lower coconut intake; however, this requires further investigation. Intensive preventive interventions are indicated to lower risk factors for CVD. These increases have implications for future premature morbidity and mortality, and costs to the health care system. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was funded through the School of Public Health and Community Medicine, University of New South Wales. References 1 Taylor R , Zimmet PZ. Obesity and diabetes in Western Samoa . Int J Obes 1981 ; 5 : 367 – 376 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 2 Ram P , Banuve S, Zimmet Pet al. . Hypertension and its correlates in Fiji: The results of the 1980 national survey . Fiji Med J 1982 ; 10 : 99 – 105 . Google Scholar OpenURL Placeholder Text WorldCat 3 Lin S, Tukana I, Linhart C, et al. Diabetes and obesity trends in Fiji over 30 years. J Diabetes 2015: in press. 2015 . 4 Service Territorial de la Statistique et des Etudes Economiques. Recensement de la population 2008, http://www.spc.int/prism/Country/WF/stats/Statistics/Census/census-index.html (2012, accessed 10 October 2014) . 5 Institut de la Statistique et des Etudes Economiques (ISEE) Nouvelle-Caledonie. Synthese 1 Recensement de la Population 2009. Report, ISEE, Noumea, 2011 . 6 Doumenge AJP, Villenave D and Chapuis O. Agriculture, food and nutrition in four South Pacific archipelagos. New Caledonia, Vanuatu, French Polynesia, Wallis and Futuna, http://www.uncapsa.org/Publication/cg14.pdf (1995, accessed 10 November 2014) . 7 Service Territorial de la Statistique et des Etudes Economiques (STSEE). Enquete budget des familles Wallis et Futuna 2005-2006 Synthese des principaux resultats. Report, STSEE, Mata-Utu, November 2006 . 8 Taylor R, Bennett PH and Zimmet P. Epidemiological studies of diabetes and cardiovascular disease in Wallis Polynesians – a comparison of residents of Wallis Island and first generation migrants to Nouméa, New Caledonia. Report, South Pacific Commission Technical Paper, No. 181. South Pacific Commission, Noumea, 1984 . 9 Secretariat of the Pacific Community (SPC). Study of risk factors for chronic non-communicable diseases in Wallis and Futuna. Report, SPC, Noumea, 2010 . 10 Magliano DJ , Söderberg S, Zimmet PZet al. . Explaining the increase of diabetes prevalence and plasma glucose in Mauritius . Diabetes Care 2012 ; 35 : 87 – 91 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Dowse G , Gareeboo H, Zimmet Pet al. . High prevalence of NIDDM and impaired glucose tolerance in Indian, Creole, and Chinese Mauritians . Diabetes 1990 ; 39 : 390 – 396 . Google Scholar Crossref Search ADS PubMed WorldCat 12 World Health Organisation. WHO STEPS chronic disease risk factor surveillance data book. http://www.who.int/chp/steps/Part6_Section3D.pdf (2014, accessed 10 October 2014) . 13 Swinburn BA , Ley SJ, Carmichael HEet al. . Body size and composition in Polynesians . Int J Obesity 1999 ; 23 : 1178 – 1183 . Google Scholar Crossref Search ADS WorldCat 14 O’Leary N , Duggan PF. Some improvements in operation of the Technicon SMAC system . Clin Chem 1980 ; 26 : 793 – 793 . Google Scholar Crossref Search ADS PubMed WorldCat 15 McGarvey S , Baker P. The effects of modernisation and migration on Samoan blood pressures . Hum Biol 1979 ; 51 : 451 – 479 . Google Scholar OpenURL Placeholder Text WorldCat 16 Zimmet P , Jackson L, Whitehouse S. Blood pressure studies in two populations with varying degrees of modernisation . NZ Med J 1980 ; 91 : 249 – 252 . Google Scholar OpenURL Placeholder Text WorldCat 17 Thuesen B, Toft U, Buhelt LP, et al. Estimated daily salt intake in relation to blood pressure and blood lipids: The role of obesity. Eur J Prev Cardiol. Epub ahead of print 3 October 2014. pii: 2047487314553201 . 18 Zatu MC, Van Rooyen JM, Kruger A, et al. Alcohol intake, hypertension development and mortality in black South Africans. Eur J Prev Cardiol. Epub ahead of print 10 December 2014. pii: 2047487314563447 . 19 Shankarishan P , Borah PK, Mohapatra PKet al. . Population attributable risk estimates for risk factors associated with hypertension in an Indian population . Eur J Prev Cardiol 2013 ; 20 : 963 – 971 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Taylor R , Zimmet P. The influence of variation in obesity on the sex difference in the prevalence of abnormal glucose tolerance in Tuvalu . NZ Med J 1981 ; 94 : 176 – 178 . Google Scholar OpenURL Placeholder Text WorldCat 21 Taylor R , Bennett P, LeGonidec Get al. . The prevalence of diabetes mellitus in a traditional-living Polynesian population: The Wallis Island survey . Diabetes Care 1983 ; 6 : 334 – 340 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Cowie CC , Rust KF, Ford ESet al. . Full accounting of diabetes and pre-diabetes in the U.S . population in 1988–1994 and 2005–2006. Diabetes Care 2009 ; 32 : 287 – 294 . Google Scholar OpenURL Placeholder Text WorldCat 23 Drivsholm T , Ibsen H, Schroll Met al. . Increasing prevalence of diabetes mellitus and impaired glucose tolerance among 60-year-old Danes . Diabet Med 2001 ; 18 : 126 – 113 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Prior IA , Davidson F, Salmond CEet al. . Cholesterol, coconuts, and diet on Polynesian atolls: A natural experiment: The Pukapuka and Tokelau Island studies . Am J Clin Nutr 1981 ; 34 : 1552 – 1561 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Prior I . Cardiovascular epidemiology in Polynesians in the Pacific . Singapore Med J 1973 ; 14 : 223 – 227 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 26 Taylor R , Badcock J, Pargeter Ket al. . Dietary intake, exercise, obesity and non-communicable disease in rural and urban populations of three Pacific Island countries . Am J Clin Nutr 1992 ; 11 : 283 – 293 . Google Scholar OpenURL Placeholder Text WorldCat 27 Lipoeto NI , Agus Z, Oenzil Fet al. . Contemporary Minangkabua food culture in West Sumatra, Indonesia . Asia Pacific J Clin Nutr 2001 ; 10 : 10 – 16 . Google Scholar Crossref Search ADS WorldCat 28 World Health Organisation Western Pacific Region. Diet, food supply and obesity in the Pacific. Report, World Health Organisation, Geneva, 2003 . 29 Kaufmann-Raab I , Jonen HG, Jähnchen Eet al. . Interference by acetaminophen in the glucose oxidase-peroxidase method for blood glucose determination . Clin Chem 1976 ; 22 : 1729 – 1731 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Klotzsch SG , McNamara JR. Triglyceride measurements: A review of methods and interferences . Clin Chem 1990 ; 36 : 1605 – 1613 . Google Scholar Crossref Search ADS PubMed WorldCat © The European Society of Cardiology 2016 This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © The European Society of Cardiology 2016

Bongers, Bart C; Hulzebos, Erik HJ; Helbing, Willem A; ten Harkel, Arend DJ; van Brussel, Marco; Takken, Tim

doi: 10.1177/2047487315611769pmid: 26464293

Abstract Background Oxygen uptake efficiency (OUE), the relation between oxygen uptake (VO2) and minute ventilation (VE), differs between healthy children and children with heart disease. This study aimed to investigate the normal response profiles of OUE during a progressive cardiopulmonary exercise test. Design: Cross-sectional. Methods Healthy children between eight and 19 years of age (114 boys and 100 girls, mean ± SD age 12.7 ± 2.8 years) performed a maximal cardiopulmonary exercise test. Peak VO2 (VO2peak), ventilatory threshold and peak VE were determined. OUE was determined by the OUE plateau (OUEP), OUE at the ventilatory threshold (OUE@VT) and OUE slope (OUES). Results OUEP (42.4 ± 4.6) and OUE@VT (41.9 ± 4.7) were similar and less variable than OUES (2138 ± 703). OUEP correlated strongly with OUE@VT (r = 0.974); however, OUEP was weak-to-moderately correlated with VO2peak (r = 0.646), the ventilatory threshold (r = 0.548) and OUES (r = 0.589). OUES correlated strongly with VO2peak (r = 0.948) and the ventilatory threshold (r = 0.856). Reference centiles for OUEP show an almost linear increase from about 37 in eight-year olds to about 47 in 18-year olds, with no sex-difference. OUES increased from about 1400 in eight-year-old boys to approximately 3500 in 18-year-old boys. OUES increased from roughly 1250 in eight-year-old girls to about 2650 in 18-year-old girls. Conclusions This study provides sex- and age-related normative values for both OUEP and OUES, which facilitates the interpretation of OUE in children. OUEP and OUES are objective and non-invasive cardiopulmonary exercise test parameters which do not require a maximal effort and might be indicative of cardiorespiratory function during exercise. Exercise testing, child, oxygen uptake efficiency plateau, oxygen uptake efficiency slope, cardiovascular health, reference values Introduction Measuring oxygen uptake (VO2), carbon dioxide output (VCO2), minute ventilation (VE) and heart rate (HR) during an incremental cardiopulmonary exercise test (CPET) up to maximal exertion provides a non-invasive assessment of the integrated physiological response of the pulmonary, cardiovascular and metabolic system to progressive exercise. The non-invasive and dynamic nature of the performed measurements during the progressive load on the cardiorespiratory system provides clinicians with important information that can be used for diagnostic, prognostic and evaluative purposes. Evaluating physiological parameters obtained at specific time points during the CPET (e.g. at peak exercise or at the ventilatory threshold) remains the most commonly used clinical test in paediatric exercise medicine.1 Assessing only a single value for the parameters of interest leads to substantial loss of physiological data as provided by evaluating cardiopulmonary response profiles throughout the test.1–4 Moreover, considering the trending of certain CPET parameters optimizes the use of the enormous amount of data generated during a routine CPET. Evaluating the response profile of certain CPET parameters has been considered as a crucial component of the interpretative strategy.5–7 The oxygen uptake efficiency (OUE) measurement, which is the reciprocal of the ventilatory equivalent for oxygen (VE/VO2), provides an estimation of the efficiency of the VE with respect to VO2. The OUE slope (OUES) concept is based on the curvilinear relationship between the VE and VO2 throughout a progressive CPET. Baba et al.8 introduced a logarithmic transformation of VE over the entire exercise period, resulting in a linear relationship between the VE and VO2 during the last part of the CPET. The regression coefficient of the regression line describing this linear relationship represents the OUES. More recently, Sun et al.9,10 proposed plotting the OUE ratio against time during the CPET. The 90-s average of the highest consecutive OUE values is termed OUE plateau (OUEP), which typically occurs around the ventilatory threshold. OUES and OUEP have been found indicative of functional impairment and prognosis in adult patients with heart failure.10–15 Little is known concerning the normalcy of the trending of OUE measurements during progressive exercise in healthy children. The OUEP has never been investigated in children, and although the OUES is known to differ between healthy children and children with heart disease,8,16,17 its prognostic value has never been investigated in children with a chronic disease. Before the prognostic value of OUE measurements can be investigated adequately in paediatric patient populations, knowledge concerning its characteristics in healthy children is essential. Therefore, the current study aims to evaluate the OUE response patterns in a large sample of healthy children. Methods Participants Two hundred and fourteen healthy Dutch children, ranging in age from eight to 19 years, performed a maximal effort during the CPET. Children and adolescents were recruited from primary and secondary schools, or were family members of hospital staff. All children were free from cardiovascular, pulmonary, neurological, or musculoskeletal disease. Informed consent was obtained from the parents, as well as from children ≥12 years of age. The study protocol was approved by the Medical Ethics Committees of the Erasmus Medical Centre Rotterdam and the University Medical Centre Utrecht, the Netherlands. CPET All participants performed a progressive CPET up to maximal exertion in upright position on an electronically braked cycle ergometer (Jaeger ER900 (Viasys Healthcare GmbH, Höchberg, Germany) at Erasmus Medical Centre Rotterdam and Lode Corival (Lode BV, Groningen, the Netherlands) at the University Medical Centre Utrecht). Participants performed a three-minute warm-up phase (unloaded cycling), where after the work rate was increased by constant increments of 10, 15 or 20 W/min, depending on the participant’s body height (<125 cm, between 125 and 150 cm, and >150 cm respectively) according to the Godfrey protocol.18 Throughout the CPET, participants had to maintain a pedalling frequency between 60 and 80 revolutions/min. The protocol continued until the participant’s pedalling frequency fell definitely below 60 revolutions/min, despite strong verbal encouragement. During the CPET, participants breathed through a facemask (Hans Rudolph, Kansas City, MO, USA) connected to a calibrated respiratory gas analysis system (Jaeger Oxycon Champion, Viasys Healthcare GmbH, Höchberg, Germany). Expired gas was passed through a flow meter (Triple V volume transducer), an oxygen analyser and a carbon dioxide analyser. The flow meter and gas analysers were connected to a computer, which calculated breath-by-breath VE, VO2, VCO2 and the respiratory exchange ratio (RER) averaged at 10-s intervals. HR was measured by continuous twelve-lead electrocardiography. A test was considered to be at or near the maximal level when participants showed clinical signs of intense effort (e.g. unsteady biking, sweating, and clear unwillingness to continue exercising despite strong encouragement), were unable to maintain the required pedalling speed, and when at least one of the following criteria was met: an HR at peak exercise (HRpeak) of >180 beats min−1 or an RER at peak exercise (RERpeak) of >1.0.19 Data analysis Absolute values at peak exercise were calculated as the average value over the last 30 s prior to termination of the test. HRpeak was defined as the highest HR achieved during the CPET. The ventilatory threshold was defined as the point at which the ventilatory equivalent for oxygen and the partial end-tidal oxygen tension reached a minimum and thereafter began to rise in a consistent manner, coinciding with an unchanged ventilatory equivalent for carbon dioxide and partial end-tidal carbon dioxide tension.5 When this ventilatory equivalents method appeared to provide uncertain results for a participant’s ventilatory threshold, the point at which the linear slope of the relation between the VCO2 and VO2 changed was taken as the ventilatory threshold, according to the V-slope method.20 The ventilatory threshold was expressed as an absolute value and as a percentage of VO2peak. The graphical presentation of VE as a function of VCO2 during the progressive CPET was used to determine the point at which VE increased out of proportion to VCO2, the respiratory compensation point. The slope of the relation between VE and VCO2 (VE/VCO2-slope) was calculated by linear least squares regression of the relation between VE and VCO2 up to the respiratory compensation point. OUE throughout the CPET was graphically presented by dividing each VO2 value (ml/min) by the corresponding VE value (l/min). The OUEP was then determined as the 90-s average of the highest consecutive values for OUE (VO2/VE, see Figure 1).9,10 OUE at the ventilatory threshold (OUE@VT) was the 60-s average of consecutive values at and immediately before the ventilatory threshold.9 Finally, the OUES was calculated using all exercise data by a linear least squares regression of the VO2 (ml/min) on the common logarithm of the VE (l/min).8 Figure 1. Open in new tabDownload slide OUE measurements throughout a CPET in a 13-year-old healthy girl: the OUEP (graph a) occurred around the ventilatory threshold and was 42.3 (horizontal black bar), whereas the value of the OUES (graph b) was 2191.9. CPET: cardiopulmonary exercise test; Log VE: common logarithm of the minute ventilation (l/min); OUE: oxygen uptake efficiency; OUEP: oxygen uptake efficiency plateau; OUES: oxygen uptake efficiency slope; VE: minute ventilation (l/min); VO2: oxygen uptake (ml/min) Statistical analysis All data were analysed using the Statistical Package for the Social Sciences for Windows (version 20.0; IBM, SPSS Inc., Chicago, IL, USA). Data are presented as mean value ± standard deviation (SD) and corresponding ranges. Variability of OUE measurements was evaluated by calculating the coefficient of variation, defined as 100 × (SD/mean), for the OUES, OUEP and OUE@VT. Independent samples t-tests were used to examine differences between boys and girls. Pearson correlation coefficients were calculated to examine associations between OUE measurements and different anthropometric and exercise parameters. Age- and sex-related reference centiles for the OUES and OUEP were derived using the lambda, mu, sigma (LMS) method as introduced by Cole,21,22 using Growth Analyser 3.5 (Growth Analyser BV, Rotterdam, the Netherlands). Based on these reference centiles, prediction equations were established for the OUES and OUEP with sex and age as independent predictors. Significance was set a priori at the 0.05 level. Results Two hundred and fourteen healthy, non-athletic Dutch children, 114 boys and 100 girls, performed a maximal effort during the CPET. Participant characteristics and CPET results are depicted in Table 1.23,24 All tests were completed without adverse effects, such as dizziness, fainting or vomiting. Table 1. Participant characteristics and CPET results. . Boys (n = 114) . Girls (n = 100) . p-value . . Anthropometric parameters:  Age, years 12.7 ± 2.9 (8.0–18.5) 12.6 ± 2.8 (8.0–18.2) 0.841  Body mass, kg 47.1 ± 14.8 (24.1–84.0) 47.6 ± 13.4 (25.0–81.7) 0.816  Body mass for age, SDSa 0.03 ± 0.83 (–2.13–2.18) 0.19 ± 0.92 c (–1.96–2.74) N/A  Body height, cm 158 ± 17 (124–197) 157 ± 13 (128–179) 0.550  Body height for age, SDSa 0.00 ± 0.86 (–1.91–2.15) 0.30 ± 0.89 d (–1.43–2.93) N/A  BMI, kg/m2 18.3 ± 2.5 (13.7–27.2) 18.9 ± 3.2 (13.7–30.1) 0.113  BMI for age, SDSa 0.04 ± 0.89 (–1.81–2.51) 0.06 ± 1.01 (–2.51–2.70) N/A  BSAb, m2 1.43 ± 0.30 (0.92–2.07) 1.43 ± 0.26 (0.95–2.02) 0.927 CPET parameters:  WRpeak, W 183 ± 68 (75–400) 161 ± 57 (70–400) 0.011 *  WRpeak, W/kg 3.8 ± 0.7 (2.6–5.6) 3.4 ± 0.6 (2.2–5.3) <0.001 ***  HRpeak, beats/min 189 ± 9 (171–211) 190 ± 9 g (171–212) 0.536  RERpeak 1.15 ± 0.07 e (0.98–1.34) 1.17 ± 0.08 h (1.01–1.37) 0.052  VO2peak, l/min 2.25 ± 0.74 e (1.13–4.10) 1.95 ± 0.61 h (0.97–4.15) 0.001 **  VO2peak, ml/kg per min 48.3 ± 6.2 e (34.2–62.3) 41.2 ± 5.7 h (28.4–55.6) <0.001 ***  Ventilatory threshold, l/min 1.33 ± 0.46f (0.53–2.77) 1.16 ± 0.37 h (0.56–2.91) 0.003 **  Ventilatory threshold, %VO2peak 59 ± 8f (40–86) 60 ± 9 h (40–81) 0.411  VEpeak, l/min 80 ± 25 e (42–157) 71 ± 21 h (34–152) 0.007 **  VEpeak, l/kg per min 1.7 ± 0.3 e (0.9–2.5) 1.5 ± 0.3 h (0.8–2.1) <0.001 ***  VE/VCO2-slope 26.6 ± 3.4 e (14.9–35.1) 27.0 ± 3.4 h (17.3–36.0) 0.370  OUES 2284 ± 764 e (1158–4256) 1970 ± 585 h (949–3816) 0.001 **  OUEP 42.6 ± 4.7 e (32.6–60.9) 42.3 ± 4.6 h (31.3–58.4) 0.686  OUE@VT 42.0 ± 4.6 e (32.5–58.9) 41.9 ± 4.7 h (30.9–57.8) 0.845 . Boys (n = 114) . Girls (n = 100) . p-value . . Anthropometric parameters:  Age, years 12.7 ± 2.9 (8.0–18.5) 12.6 ± 2.8 (8.0–18.2) 0.841  Body mass, kg 47.1 ± 14.8 (24.1–84.0) 47.6 ± 13.4 (25.0–81.7) 0.816  Body mass for age, SDSa 0.03 ± 0.83 (–2.13–2.18) 0.19 ± 0.92 c (–1.96–2.74) N/A  Body height, cm 158 ± 17 (124–197) 157 ± 13 (128–179) 0.550  Body height for age, SDSa 0.00 ± 0.86 (–1.91–2.15) 0.30 ± 0.89 d (–1.43–2.93) N/A  BMI, kg/m2 18.3 ± 2.5 (13.7–27.2) 18.9 ± 3.2 (13.7–30.1) 0.113  BMI for age, SDSa 0.04 ± 0.89 (–1.81–2.51) 0.06 ± 1.01 (–2.51–2.70) N/A  BSAb, m2 1.43 ± 0.30 (0.92–2.07) 1.43 ± 0.26 (0.95–2.02) 0.927 CPET parameters:  WRpeak, W 183 ± 68 (75–400) 161 ± 57 (70–400) 0.011 *  WRpeak, W/kg 3.8 ± 0.7 (2.6–5.6) 3.4 ± 0.6 (2.2–5.3) <0.001 ***  HRpeak, beats/min 189 ± 9 (171–211) 190 ± 9 g (171–212) 0.536  RERpeak 1.15 ± 0.07 e (0.98–1.34) 1.17 ± 0.08 h (1.01–1.37) 0.052  VO2peak, l/min 2.25 ± 0.74 e (1.13–4.10) 1.95 ± 0.61 h (0.97–4.15) 0.001 **  VO2peak, ml/kg per min 48.3 ± 6.2 e (34.2–62.3) 41.2 ± 5.7 h (28.4–55.6) <0.001 ***  Ventilatory threshold, l/min 1.33 ± 0.46f (0.53–2.77) 1.16 ± 0.37 h (0.56–2.91) 0.003 **  Ventilatory threshold, %VO2peak 59 ± 8f (40–86) 60 ± 9 h (40–81) 0.411  VEpeak, l/min 80 ± 25 e (42–157) 71 ± 21 h (34–152) 0.007 **  VEpeak, l/kg per min 1.7 ± 0.3 e (0.9–2.5) 1.5 ± 0.3 h (0.8–2.1) <0.001 ***  VE/VCO2-slope 26.6 ± 3.4 e (14.9–35.1) 27.0 ± 3.4 h (17.3–36.0) 0.370  OUES 2284 ± 764 e (1158–4256) 1970 ± 585 h (949–3816) 0.001 **  OUEP 42.6 ± 4.7 e (32.6–60.9) 42.3 ± 4.6 h (31.3–58.4) 0.686  OUE@VT 42.0 ± 4.6 e (32.5–58.9) 41.9 ± 4.7 h (30.9–57.8) 0.845 Values are presented as mean ± SD (range). a Calculated using Dutch normative values.23 b Calculated using the equation of Haycock et al.24 c Significantly different from 0 (p = 0.045). d Significantly different from 0 (p = 0.001). e Respiratory gas analysis measurements were invalid in two boys, so in this case n = 112. f Respiratory gas analysis measurements were invalid in two boys and the ventilatory threshold was not determinable in one boy, so in this case n = 111. g Heart rate could not be measured in one girl, so in this case n = 99. h Respiratory gas analysis measurements were invalid in two girls, so in this case n = 98. CPET: cardiopulmonary exercise test; SDS: standard deviation score; N/A: not applicable; BMI: body mass index; BSA: body surface area; WRpeak: peak work rate; HRpeak: heart rate at peak exercise; RERpeak: respiratory exchange ratio at peak exercise; VO2peak: oxygen uptake at peak exercise; VEpeak: minute ventilation at peak exercise; VE/VCO2-slope: slope of the relation between minute ventilation and carbon dioxide production up to the respiratory compensation point; OUES: oxygen uptake efficiency slope; OUEP: oxygen uptake efficiency plateau; OUE@VT: oxygen uptake efficiency at the ventilatory threshold * p < 0.05 ** p < 0.01 *** p < 0.001 Open in new tab Table 1. Participant characteristics and CPET results. . Boys (n = 114) . Girls (n = 100) . p-value . . Anthropometric parameters:  Age, years 12.7 ± 2.9 (8.0–18.5) 12.6 ± 2.8 (8.0–18.2) 0.841  Body mass, kg 47.1 ± 14.8 (24.1–84.0) 47.6 ± 13.4 (25.0–81.7) 0.816  Body mass for age, SDSa 0.03 ± 0.83 (–2.13–2.18) 0.19 ± 0.92 c (–1.96–2.74) N/A  Body height, cm 158 ± 17 (124–197) 157 ± 13 (128–179) 0.550  Body height for age, SDSa 0.00 ± 0.86 (–1.91–2.15) 0.30 ± 0.89 d (–1.43–2.93) N/A  BMI, kg/m2 18.3 ± 2.5 (13.7–27.2) 18.9 ± 3.2 (13.7–30.1) 0.113  BMI for age, SDSa 0.04 ± 0.89 (–1.81–2.51) 0.06 ± 1.01 (–2.51–2.70) N/A  BSAb, m2 1.43 ± 0.30 (0.92–2.07) 1.43 ± 0.26 (0.95–2.02) 0.927 CPET parameters:  WRpeak, W 183 ± 68 (75–400) 161 ± 57 (70–400) 0.011 *  WRpeak, W/kg 3.8 ± 0.7 (2.6–5.6) 3.4 ± 0.6 (2.2–5.3) <0.001 ***  HRpeak, beats/min 189 ± 9 (171–211) 190 ± 9 g (171–212) 0.536  RERpeak 1.15 ± 0.07 e (0.98–1.34) 1.17 ± 0.08 h (1.01–1.37) 0.052  VO2peak, l/min 2.25 ± 0.74 e (1.13–4.10) 1.95 ± 0.61 h (0.97–4.15) 0.001 **  VO2peak, ml/kg per min 48.3 ± 6.2 e (34.2–62.3) 41.2 ± 5.7 h (28.4–55.6) <0.001 ***  Ventilatory threshold, l/min 1.33 ± 0.46f (0.53–2.77) 1.16 ± 0.37 h (0.56–2.91) 0.003 **  Ventilatory threshold, %VO2peak 59 ± 8f (40–86) 60 ± 9 h (40–81) 0.411  VEpeak, l/min 80 ± 25 e (42–157) 71 ± 21 h (34–152) 0.007 **  VEpeak, l/kg per min 1.7 ± 0.3 e (0.9–2.5) 1.5 ± 0.3 h (0.8–2.1) <0.001 ***  VE/VCO2-slope 26.6 ± 3.4 e (14.9–35.1) 27.0 ± 3.4 h (17.3–36.0) 0.370  OUES 2284 ± 764 e (1158–4256) 1970 ± 585 h (949–3816) 0.001 **  OUEP 42.6 ± 4.7 e (32.6–60.9) 42.3 ± 4.6 h (31.3–58.4) 0.686  OUE@VT 42.0 ± 4.6 e (32.5–58.9) 41.9 ± 4.7 h (30.9–57.8) 0.845 . Boys (n = 114) . Girls (n = 100) . p-value . . Anthropometric parameters:  Age, years 12.7 ± 2.9 (8.0–18.5) 12.6 ± 2.8 (8.0–18.2) 0.841  Body mass, kg 47.1 ± 14.8 (24.1–84.0) 47.6 ± 13.4 (25.0–81.7) 0.816  Body mass for age, SDSa 0.03 ± 0.83 (–2.13–2.18) 0.19 ± 0.92 c (–1.96–2.74) N/A  Body height, cm 158 ± 17 (124–197) 157 ± 13 (128–179) 0.550  Body height for age, SDSa 0.00 ± 0.86 (–1.91–2.15) 0.30 ± 0.89 d (–1.43–2.93) N/A  BMI, kg/m2 18.3 ± 2.5 (13.7–27.2) 18.9 ± 3.2 (13.7–30.1) 0.113  BMI for age, SDSa 0.04 ± 0.89 (–1.81–2.51) 0.06 ± 1.01 (–2.51–2.70) N/A  BSAb, m2 1.43 ± 0.30 (0.92–2.07) 1.43 ± 0.26 (0.95–2.02) 0.927 CPET parameters:  WRpeak, W 183 ± 68 (75–400) 161 ± 57 (70–400) 0.011 *  WRpeak, W/kg 3.8 ± 0.7 (2.6–5.6) 3.4 ± 0.6 (2.2–5.3) <0.001 ***  HRpeak, beats/min 189 ± 9 (171–211) 190 ± 9 g (171–212) 0.536  RERpeak 1.15 ± 0.07 e (0.98–1.34) 1.17 ± 0.08 h (1.01–1.37) 0.052  VO2peak, l/min 2.25 ± 0.74 e (1.13–4.10) 1.95 ± 0.61 h (0.97–4.15) 0.001 **  VO2peak, ml/kg per min 48.3 ± 6.2 e (34.2–62.3) 41.2 ± 5.7 h (28.4–55.6) <0.001 ***  Ventilatory threshold, l/min 1.33 ± 0.46f (0.53–2.77) 1.16 ± 0.37 h (0.56–2.91) 0.003 **  Ventilatory threshold, %VO2peak 59 ± 8f (40–86) 60 ± 9 h (40–81) 0.411  VEpeak, l/min 80 ± 25 e (42–157) 71 ± 21 h (34–152) 0.007 **  VEpeak, l/kg per min 1.7 ± 0.3 e (0.9–2.5) 1.5 ± 0.3 h (0.8–2.1) <0.001 ***  VE/VCO2-slope 26.6 ± 3.4 e (14.9–35.1) 27.0 ± 3.4 h (17.3–36.0) 0.370  OUES 2284 ± 764 e (1158–4256) 1970 ± 585 h (949–3816) 0.001 **  OUEP 42.6 ± 4.7 e (32.6–60.9) 42.3 ± 4.6 h (31.3–58.4) 0.686  OUE@VT 42.0 ± 4.6 e (32.5–58.9) 41.9 ± 4.7 h (30.9–57.8) 0.845 Values are presented as mean ± SD (range). a Calculated using Dutch normative values.23 b Calculated using the equation of Haycock et al.24 c Significantly different from 0 (p = 0.045). d Significantly different from 0 (p = 0.001). e Respiratory gas analysis measurements were invalid in two boys, so in this case n = 112. f Respiratory gas analysis measurements were invalid in two boys and the ventilatory threshold was not determinable in one boy, so in this case n = 111. g Heart rate could not be measured in one girl, so in this case n = 99. h Respiratory gas analysis measurements were invalid in two girls, so in this case n = 98. CPET: cardiopulmonary exercise test; SDS: standard deviation score; N/A: not applicable; BMI: body mass index; BSA: body surface area; WRpeak: peak work rate; HRpeak: heart rate at peak exercise; RERpeak: respiratory exchange ratio at peak exercise; VO2peak: oxygen uptake at peak exercise; VEpeak: minute ventilation at peak exercise; VE/VCO2-slope: slope of the relation between minute ventilation and carbon dioxide production up to the respiratory compensation point; OUES: oxygen uptake efficiency slope; OUEP: oxygen uptake efficiency plateau; OUE@VT: oxygen uptake efficiency at the ventilatory threshold * p < 0.05 ** p < 0.01 *** p < 0.001 Open in new tab OUES, OUEP and OUE@VT values were obtained in all participants. The average OUES, OUEP and OUE@VT values were 2138 ± 703 (range: 949 to 4256), 42.4 ± 4.6 (range: 31.3 to 60.9) and 41.9 ± 4.7 (range: 30.9 to 58.9), respectively. Coefficients of variation equalled 32.9%, 10.9% and 11.1% for the OUES, OUEP and OUE@VT, respectively. Even when OUES values were normalized for body mass (45.9 ± 7.9) or BSA (1483 ± 272), the variability of the OUES was higher (17.2% and 18.3%, respectively) than for the OUEP and the OUE@VT. Figure 2 presents reference centiles for absolute OUES values in boys (left upper graph) and girls (left lower graph) from eight to 18 years of age. Mean absolute OUES values increase from approximately 1400 at eight years of age to 3500 at 18 years of age in boys (+150%, Figure 2; Supplementary Table 1 in Supplementary Material online), whereas values increase from about 1250 at eight years of age to 2650 at 18 years of age in girls (+112%, Figure 2; Supplementary Table 1 in Supplementary Material online). Reference centiles for absolute OUEP values in boys (right upper graph) and girls (right lower graph) from eight to 18 years of age are also depicted in Figure 2. Mean OUEP values in both boys and girls increase from roughly 37 at eight years of age up to approximately 47 at 18 years of age (+27%, Figure 2; Supplementary Table 1 in Supplementary Material online), with no significant sex-differences. Table 2 depicts prediction equations for the OUES and the OUEP for healthy children and adolescents. Figure 2. Open in new tabDownload slide Age- and sex-related reference centiles for OUE: the OUES against age in boys (left upper graph) and girls (left lower graph) and the OUEP against age in boys (right upper graph) and girls (right lower graph). OUE: oxygen uptake efficiency; OUEP: oxygen uptake efficiency plateau; OUES: oxygen uptake efficiency slope Table 2. OUES and OUEP prediction equations for healthy children. Dependent . . Constant and independent predictors . Statistical analysis . . Sex . Constant . Age (years) . R2 . OUESa Boys 577.208 6.172 × age2 52.069 × age 0.9997 Girls −342.403 −2.589 × age2 214.606 × age 0.9993 OUES/kgb Boys 21.757 −0.0011 × age4 0.0562 × age3 −1.0675 × age2 8.8991 × age 0.9063 Girls 41.276 −0.0006 × age4 0.0247 × age3 −0.3252 × age2 1.4446 × age 0.9910 OUEPa Boys 26.340 −0.029 × age2 1.641 × age 0.9998 Girls 28.437 −0.0036 × age2 1.1409 × age 0.9999 Dependent . . Constant and independent predictors . Statistical analysis . . Sex . Constant . Age (years) . R2 . OUESa Boys 577.208 6.172 × age2 52.069 × age 0.9997 Girls −342.403 −2.589 × age2 214.606 × age 0.9993 OUES/kgb Boys 21.757 −0.0011 × age4 0.0562 × age3 −1.0675 × age2 8.8991 × age 0.9063 Girls 41.276 −0.0006 × age4 0.0247 × age3 −0.3252 × age2 1.4446 × age 0.9910 OUEPa Boys 26.340 −0.029 × age2 1.641 × age 0.9998 Girls 28.437 −0.0036 × age2 1.1409 × age 0.9999 a For the absolute OUES and OUEP, the presented constants and independent predictors are based on polynomial regression (order 2) of the P50 values depicted in Figure 2. b For the OUES/kg, the presented constants and independent predictors are based on polynomial regression (order 4). OUE: oxygen uptake efficiency; OUES: oxygen uptake efficiency slope; OUEP: oxygen uptake efficiency plateau. Open in new tab Table 2. OUES and OUEP prediction equations for healthy children. Dependent . . Constant and independent predictors . Statistical analysis . . Sex . Constant . Age (years) . R2 . OUESa Boys 577.208 6.172 × age2 52.069 × age 0.9997 Girls −342.403 −2.589 × age2 214.606 × age 0.9993 OUES/kgb Boys 21.757 −0.0011 × age4 0.0562 × age3 −1.0675 × age2 8.8991 × age 0.9063 Girls 41.276 −0.0006 × age4 0.0247 × age3 −0.3252 × age2 1.4446 × age 0.9910 OUEPa Boys 26.340 −0.029 × age2 1.641 × age 0.9998 Girls 28.437 −0.0036 × age2 1.1409 × age 0.9999 Dependent . . Constant and independent predictors . Statistical analysis . . Sex . Constant . Age (years) . R2 . OUESa Boys 577.208 6.172 × age2 52.069 × age 0.9997 Girls −342.403 −2.589 × age2 214.606 × age 0.9993 OUES/kgb Boys 21.757 −0.0011 × age4 0.0562 × age3 −1.0675 × age2 8.8991 × age 0.9063 Girls 41.276 −0.0006 × age4 0.0247 × age3 −0.3252 × age2 1.4446 × age 0.9910 OUEPa Boys 26.340 −0.029 × age2 1.641 × age 0.9998 Girls 28.437 −0.0036 × age2 1.1409 × age 0.9999 a For the absolute OUES and OUEP, the presented constants and independent predictors are based on polynomial regression (order 2) of the P50 values depicted in Figure 2. b For the OUES/kg, the presented constants and independent predictors are based on polynomial regression (order 4). OUE: oxygen uptake efficiency; OUES: oxygen uptake efficiency slope; OUEP: oxygen uptake efficiency plateau. Open in new tab Correlations (Supplementary Table 2 in Supplementary Material online) between the OUEP and absolute VO2peak and between the OUEP and the ventilatory threshold were much lower than those between the OUES and absolute VO2peak (r = 0.646 vs. r = 0.948; with p < 0.001 for both coefficients) and between the OUES and the ventilatory threshold (r = 0.548 vs. r = 0.856; with p < 0.001 for both coefficients). The VE/VCO2-slope was moderately to strongly correlated with both the OUEP (r = –0.719; p < 0.001) and the OUES (r = –0.641; p < 0.001). The OUEP correlated highly with OUE@VT (r = 0.974; p < 0.001), whereas the associations between the OUEP and the OUES and between the OUE@VT and the OUES were lower (r = 0.589 and r = 0.578, respectively; p < 0.001 for both coefficients). Discussion This study describes the characteristics of the OUE in a large healthy paediatric population, aged 8–19 years. The OUES and OUEP are easy, non-invasive and objectively determinable from CPET data and provide information about the function of the cardiorespiratory system during progressive exercise. Calculation of the OUES does not require a maximal effort of the child, as the relationship between the common logarithm of the VE and the VO2 is linear during the last part of the CPET. The OUEP also does not require a maximal effort of the child, as it occurs around the ventilatory threshold at moderate exercise intensity. Similar values were found for the OUEP and OUE@VT, which were less variable between participants than the OUES. These coefficients of variation were slightly lower than those described for healthy adults.9 Absolute OUES values increased with age in boys and girls, with higher values attained by boys. This finding is in agreement with previous studies in children.25–27 OUEP values also increased with age in boys and girls, with no significant sex-difference. The OUES correlated strongly with VO2peak and the ventilatory threshold in this study, which is in line with previous research in children.8,16,25–29 The OUEP was weak-to-moderately correlated with VO2peak, the ventilatory threshold and the OUES, which was also reported in healthy adults by Sun et al.9 Sun et al.9 reported higher test–retest reproducibility for the OUEP compared with the OUES. The authors reported an average variability between paired values of the OUEP and OUES in 24 healthy adults of 3.9 ± 2.5% and 11.3 ± 8.6%, respectively (p < 0.001). Despite this higher variability within participants, the OUES has previously been reported to show good reproducibility in healthy adults.30,31 Baba et al.30 found a coefficient of repeatability (COR) for the OUES, VO2peak and ventilatory threshold of 20%, 16% and 31%, respectively. Similarly, van Laethem et al.31 reported an intraclass correlation coefficient (ICC) for the OUES of 0.93 and a COR of 18.7%, which was comparable to VO2peak (ICC of 0.95 and COR of 17.3%) and superior to the ventilatory threshold (ICC of 0.86, COR not reported). De Groot et al.32 found an ICC of 0.80 and a coefficient of variation within participants of 24.3% for the OUES in 23 children with spina bifida. Data in children addressing test–retest reproducibility for the OUEP are currently lacking. The OUES and OUEP are indicative of cardiorespiratory function during exercise. Without significant lung disease, both the OUES and OUEP are indicative of cardiovascular function throughout exercise. The curvilinear response of VE during progressive exercise, caused by the progressively increasing contribution of the anaerobic glycolysis to energy metabolism, provides the conceptual basis for the OUES. Baba et al.8 observed that the logarithmic transformation of the VE makes the relation between VE and VO2 during progressive exercise linear, in which the slope of the regression line describing this linear relation represents the OUES. In essence, the OUES provides an estimation of the efficiency of the VE with respect to the VO2. Higher OUES values indicate a more efficient VO2. OUEP represents the maximal (most efficient) OUE, which occurs during submaximal exercise intensities around the ventilatory threshold, and not at rest or at maximal exercise. At rest, the high mixed venous oxygen content requires less oxygen extraction from alveolar gas. Moreover, ventilation is less efficient at rest for VO2, as low VE values result in a high physiologic dead space resulting in suboptimal ventilation perfusion matching. During maximal exercise, mixed venous oxygen content is lowest. Moreover, high VE values result in an optimal ratio of physiologic dead space to tidal volume and therefore in an optimal ventilation–perfusion matching. However, anaerobic glycolysis, with lactic acid and carbon dioxide as its by-products, results in an excessive ventilatory response to metabolic acidosis. The latter will significantly reduce the ratio of VO2 to VE (OUE). Therefore, the OUEP typically occurs around the ventilatory threshold, before OUE values decline caused by the increased VE as a result of anaerobic glycolysis. Higher OUEP values indicate a more efficient VO2, whereas lower values represent a lower VO2 for any given VE. Hence, both the OUES and OUEP are dependent on the ventilatory threshold and the ratio of physiologic dead space to tidal volume.8,9 By definition, a maximal CPET is highly dependent on the motivation of the child to continue exercising against high intensities, when dyspnoea, muscle fatigue and other stress sensations are commonly experienced. When the delivered effort of the child during a CPET cannot be classified as maximal, the complete test may be deemed a failure despite the wealth of data successfully collected.1 The latter can provide important information concerning the normalcy of the child’s pulmonary, cardiovascular and metabolic response to progressive exercise, even in the absence of a ‘true’ VO2peak. A maximal effort of the child is not required to determine the OUES and the OUEP. This makes OUE measurements an appealing alternative in patients with heart failure or cardiac rhythm abnormalities in which maximal exercise testing is sometimes contraindicated. When the OUES is calculated, the VE is logarithmically transformed to produce a linear slope when plotted against VO2. This makes the OUES theoretically an exercise intensity independent measure which is resistant to disruption by early termination of the CPET. Several studies in children confirmed the linearity of the OUES;16,25–27 however, other studies found OUES values calculated using exercise data up to submaximal exercise intensities to be slightly, but significantly, lower than OUES values calculated using exercise data up to peak exercise.8,28,33 The OUEP occurs around the ventilatory threshold, just before OUE values start to decline due to the increased VE as a result of the bicarbonate buffering of the proton from lactate resulting in the non-metabolic production of carbon dioxide. Hence, OUE measurements can be calculated from submaximal exercise testing as long as children are able to reach their ventilatory threshold. If the child does not reach his or her ventilatory threshold, OUEP values should be interpreted with caution. OUES values should be interpreted with caution when the ventilatory threshold is not reached or when a plateau in VO2 is observed despite an increase in work rate (VO2max).34 Despite the fact that the ventilatory threshold involves only a single value of a CPET, it has been found to be strongly correlated with VO2peak.35 This makes the ventilatory threshold a useful indicator for aerobic exercise capacity in children unable or unwilling to perform a maximal effort. An important shortcoming of the ventilatory threshold is the fact that it is sometimes not clearly identifiable, as well as the continuing debate concerning its reproducibility.36 The ventilatory threshold depends on the mode of exercise testing, the utilized exercise protocol and the method of detection, and is a subjective measurement and is thus subject to substantial intra- and inter-observer variability.37–40 Both the OUES and OUEP are easy determinable in each CPET and free from interobserver and intraobserver variability, since they are mathematically determined by a set of CPET data. Study limitations Reference centiles were generated with respect to age and sex, without taking other relevant factors (e.g. race, maturation and anthropometrics) into account. Moreover, we did not measure physical activity levels of the included participants. Finally, a longitudinal study design would have provided a more secure analysis of the development of OUE throughout the paediatric age range. Future research It is currently unknown whether the OUEP is able to differentiate between healthy children and children with cardiovascular or severe respiratory disease (forced expiratory volume in 1 s <70% of predicted). Previous findings suggest that the OUEP has discriminative and prognostic value in adults with heart disease.10 Further research is required to assess its discriminative and prognostic properties in different paediatric patient populations. Test–retest reproducibility, as well as the evaluative properties of the OUES and OUEP, in paediatric populations remains also the subject of further research. Conclusion This study provides sex- and age-related normative values for both OUEP and OUES, which facilitates the interpretation of OUE in children. The OUEP and OUES increase with age in boys and girls, in which boys attain higher OUES values than girls. Both the OUEP and OUES are objective and non-invasive CPET parameters which do not require a maximal effort and might be indicative of cardiorespiratory function during exercise. From OUE measurements, the OUES seems to be a valid submaximal alternative for VO2peak. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) received no financial support for the research, authorship, and/or publication of this article. References 1 Cooper DM , Leu SY, Galassetti Pet al. . Dynamic interactions of gas exchange, body mass, and progressive exercise in children . Med Sci Sports Exerc 2014 ; 46 : 877 – 886 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Neder JA , Nery LE, Peres Cet al. . Reference values for dynamic responses to incremental cycle ergometry in males and females aged 20 to 80 . Am J Respir Crit Care Med 2001 ; 164 : 1481 – 1486 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Sun XG , Hansen JE, Garatachea Net al. . Ventilatory efficiency during exercise in healthy subjects . Am J Respir Crit Care Med 2002 ; 166 : 1443 – 1448 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Ramos RP , Alencar MC, Treptow Eet al. . Clinical usefulness of response profiles to rapidly incremental cardiopulmonary exercise testing . Pulm Med 2013 , pp. 359021 – 359021 . Google Scholar OpenURL Placeholder Text WorldCat 5 American Thoracic Society; American College of Chest Physicians . ATS/ACCP statement on cardiopulmonary exercise testing . Am J Respir Crit Care Med 2003 ; 167 : 211 – 277 . Crossref Search ADS PubMed WorldCat 6 Palange P , Ward SA, Carlsen KHet al. . Recommendations on the use of exercise testing in clinical practice . Eur Respir J 2007 ; 29 : 185 – 209 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Wasserman K , Hansen JE, Sue DYet al. . Principles of exercise testing and interpretation: Including pathophysiology and clinical applications , 5th ed. Philadelphia : Lippincott Williams & Wilkins , 2012 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 8 Baba R , Nagashima M, Goto Met al. . Oxygen uptake efficiency slope: A new index of cardiorespiratory functional reserve derived from the relation between oxygen uptake and minute ventilation during incremental exercise . J Am Coll Cardiol 1996 ; 28 : 1567 – 1572 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Sun XG , Hansen JE, Stringer WW. Oxygen uptake efficiency plateau: Physiology and reference values . Eur J Appl Physiol 2012 ; 112 : 919 – 928 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Sun XG , Hansen JE, Stringer WW. Oxygen uptake efficiency plateau best predicts early death in heart failure . Chest 2012 ; 141 : 1284 – 1294 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Davies LC , Wensel R, Georgiadou Pet al. . Enhanced prognostic value from cardiopulmonary exercise testing in chronic heart failure by non–linear analysis: Oxygen uptake efficiency slope . Eur Heart J 2006 ; 27 : 684 – 690 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Arena R , Brubaker P, Moore Bet al. . The oxygen uptake efficiency slope is reduced in older patients with heart failure and a normal ejection fraction . Int J Cardiol 2010 ; 144 : 101 – 102 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Arena R , Myers J, Abella Jet al. . Prognostic significance of the oxygen uptake efficiency slope: Percent-predicted versus actual value . Am J Cardiol 2010 ; 105 : 757 – 758 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Antoine-Jonville S , Pichon A, Vazir Aet al. . Oxygen uptake efficiency slope, aerobic fitness, and V(E)-VCO2 slope in heart failure . Med Sci Sports Exerc 2012 ; 44 : 428 – 434 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Arena R , Guazzi M, Myers Jet al. . The relationship between minute ventilation and oxygen consumption in heart failure: Comparing peak VE/VO2 and the oxygen uptake efficiency slope . Int J Cardiol 2012 ; 154 : 384 – 385 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Bongers BC , Hulzebos HJ, Blank ACet al. . The oxygen uptake efficiency slope in children with congenital heart disease: Construct and group validity . Eur J Cardiovasc Prev Rehabil 2011 ; 18 : 384 – 392 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Motonaga KS , Punn R, Axelrod DMet al. . Diminished exercise capacity and chronotropic incompetence in pediatric patients with congenital complete heart block and chronic right ventricular pacing . Heart Rhythm 2015 ; 12 : 560 – 565 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Godfrey S . Exercise testing in children: Applications in health and disease , London : WB Saunders Company Ltd , 1974 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 19 Armstrong N , Welsman JR Aerobic fitness . In: Armstrong N, van Mechelen W (eds). Paediatric exercise science and medicine , Oxford : Oxford University Press , 2008 , pp. 97 – 108 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 20 Beaver WL , Wasserman K, Whipp BJ. A new method for detecting anaerobic threshold by gas exchange . J Appl Physiol 1986 ; 60 : 2020 – 2027 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Cole TJ . Fitting smoothed centile curves to reference data . J R Stat Soc 1988 ; 151 : 385 – 418 . Google Scholar Crossref Search ADS WorldCat 22 Cole TJ . The LMS method for constructing normalized growth standards . Eur J Clin Nutr 1990 ; 44 : 45 – 60 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 23 Fredriks AM , van Buuren S, Wit JMet al. . Body index measurements in 1996-7 compared with 1980 . Arch Dis Child 2000 ; 82 : 107 – 112 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Haycock GB , Schwartz GJ, Wisotsky DH. Geometric method for measuring body surface area: A height-weight formula validated in infants, children, and adults . J Pediatr 1978 ; 93 : 62 – 66 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Marinov B , Kostianev S. Exercise performance and oxygen uptake efficiency slope in obese children performing standardized exercise . Acta Physiol Pharmacol Bulg 2003 ; 27 : 59 – 64 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 26 Marinov B , Mandadzhieva S, Kostianev S. Oxygen-uptake efficiency slope in healthy 7- to 18-year-old children . Pediatr Exerc Sci 2007 ; 19 : 159 – 170 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Akkerman M , van Brussel M, Bongers BCet al. . Oxygen uptake efficiency slope in healthy children . Pediatr Exerc Sci 2010 ; 22 : 431 – 441 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Bongers BC , Hulzebos EH, Arets BGet al. . Validity of the oxygen uptake efficiency slope in children with cystic fibrosis and mild-to-moderate airflow obstruction . Pediatr Exerc Sci 2012 ; 24 : 129 – 141 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Breithaupt PG , Colley RC, Adamo KB. Using the oxygen uptake efficiency slope as an indicator of cardiorespiratory fitness in the obese pediatric population . Pediatr Exerc Sci 2012 ; 24 : 357 – 368 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Baba R , Kubo N, Morotome Yet al. . Reproducibility of the oxygen uptake efficiency slope in normal healthy subjects . J Sports Med Phys Fitness 1999 ; 39 : 202 – 206 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 31 van Laethem C , de Sutter J, Peersman Wet al. . Intratest reliability and test-retest reproducibility of the oxygen uptake efficiency slope in healthy participants . Eur J Cardiovasc Prev Rehabil 2009 ; 16 : 493 – 498 . Google Scholar Crossref Search ADS PubMed WorldCat 32 de Groot JF , Takken T, Gooskens RHet al. . Reproducibility of maximal and submaximal exercise testing in ‘normal ambulatory’ and ‘community ambulatory’ children and adolescents with spina bifida: Which is best for the evaluation and application of exercise training? Phys Ther 2011 ; 91 : 267 – 276 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Drinkard B , Roberts MD, Ranzenhofer LMet al. . Oxygen-uptake efficiency slope as a determinant of fitness in overweight adolescents . Med Sci Sports Exerc 2007 ; 39 : 1811 – 1816 . Google Scholar Crossref Search ADS PubMed WorldCat 34 Niemeijer VM , van’t Veer M, Schep Get al. . Causes of nonlinearity of the oxygen uptake efficiency slope: A prospective study in patients with chronic heart failure . Eur J Prev Cardiol 2014 ; 21 : 347 – 353 . Google Scholar Crossref Search ADS PubMed WorldCat 35 Hebestreit H , Staschen B, Hebestreit A. Ventilatory threshold: A useful method to determine aerobic fitness in children? Med Sci Sports Exerc 2000 ; 32 : 1964 – 1969 . Google Scholar Crossref Search ADS PubMed WorldCat 36 Bongers BC , van Brussel M, Hulzebos HJet al. . Pediatric norms for cardiopulmonary exercise testing. In relation to sex and age , 2nd ed. ’s Hertogenbosch : Uitgeverij BOXpress , 2014 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 37 Yeh MP , Gardner RM, Adams TDet al. . ‘Anaerobic threshold’: Problems of determination and validation . J Appl Physiol 1983 ; 55 : 1178 – 1186 . Google Scholar Crossref Search ADS PubMed WorldCat 38 Gladden LB , Yates JW, Stremel RWet al. . Gas exchange and lactate anaerobic thresholds: Inter- and intraevaluator agreement . J Appl Physiol 1985 ; 58 : 2082 – 2089 . Google Scholar Crossref Search ADS PubMed WorldCat 39 Shimizu M , Myers J, Buchanan Net al. . The ventilatory threshold: method, protocol, and evaluator agreement . Am Heart J 1991 ; 122 : 509 – 516 . Google Scholar Crossref Search ADS PubMed WorldCat 40 Vanhees L , Lefevre J, Philippaerts Ret al. . How to assess physical activity? How to assess physical fitness? Eur J Cardiovasc Prev Rehabil 2005 ; 12 : 102 – 114 . Google Scholar Crossref Search ADS PubMed WorldCat © The European Society of Cardiology 2016 This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © The European Society of Cardiology 2016