Sex-specific associations between adolescent categories of BMI with cardiovascular and non-cardiovascular mortality in midlife

Sex-specific associations between adolescent categories of BMI with cardiovascular and... Context: Most studies linking long‑ term consequences of adolescent underweight and obesity are limited to men. Objective: To assess the sex‑ specific association of adolescent BMI with cardiovascular ‑ and non‑ cardiovascular‑ related mortality in young adulthood and midlife. Setting: A nationwide cohort. Participants: 927,868 women, 1,366,271 men. Interventions: Medical examination data at age 17, including BMI, were linked to the national death registry. Main outcomes: Death attributed to cardiovascular (CVD) and non‑ CVD causes. Results: During 17,346,230 women‑ years and 28,367,431 men‑ years of follow‑ up, there were 451 and 3208 CVD deaths, respectively, and 6235 and 22,223 non‑ CVD deaths, respectively. Compared to low‑ normal BMI (18.5–22.0 kg/ m ), underweight women had a lower adjusted risk for CVD mortality (Cox hazard ratio (HR) = 0.68; 95% CI 0.46–0.98) in contrast to underweight men (HR = 0.99; 0.88–1.13). The latter were at higher risk for non‑ CVD mortality (HR = 1.04; 1.00–1.09), unlike underweight women (HR = 1.01; 0.93–1.10). Findings, which persisted when the study sample was limited to those with unimpaired health, were accentuated for the obese with ≥ 30 years follow‑ up. Both sexes exhibited similarly higher risk estimates already in the high‑ normal BMI range (22.0 ≤ BMI < 25.0 kg/m ) with overall no interaction between sex and BMI (p = 0.62). Adjusted spline models suggested lower BMI values for minimal mortality 2 2 risk among women (16.8 and 18.2 kg/m ) than men (18.8 and 20.0 kg/m ), for CVD and non‑ CVD death, respectively. Conclusions: Underweight adolescent females have favorable cardiovascular outcomes in adulthood. Otherwise the risk patterns were similar between the sexes. The optimal BMI value for women and men with respect to future CVD outcomes is within or below the currently accepted low‑ normal BMI range. Keywords: Adolescence, Body mass index, Cohort study, Obesity, Cardiovascular death, Sex, Women, Underweight *Correspondence: Gilad.Twig@gmail.com The Israel Defense Forces Medical Corps, Tel Hashomer, Ramat Gan, Israel Full list of author information is available at the end of the article © The Author(s) 2018. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creat iveco mmons .org/ publi cdoma in/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 2 of 10 Primary outcomes and documentation of the cause Introduction of death Cardiovascular disease remains the leading cause of The primary outcomes of the study were either death death in the Western world [1, 2]. Over the last few dec- attributed to cardiovascular causes according to the ades, the prevalence of obesity has risen worldwide to as International Classification of Diseases (ICD-9: 390– high as 20% of the adolescent population in the US [3]. 459; ICD-10: I00–I99) or deaths attributed to non-car- There is increasing evidence for the link between adoles - diovascular etiologies. Given that deaths attributed to cent overweight and obesity and cardiovascular and all- diabetes (ICD-9: 250; ICD-10: E08–E13) are closely asso- cause mortality [4–6]. ciated with cardiovascular deaths [15], we grouped these Many of the studies included only men [6–8] or together with cardiovascular deaths as was performed in reported on a small number of deaths among women previous studies [16–18]. The underlying cause of death, [9]. Furthermore, the majority of the studies used World as officially coded from death notifications by the Israel Health Organization (WHO) or the Centers for Disease Central Bureau of Statistics, was linked to the database Control (CDC) classifications which include a broad nor - using the participants’ national ID number. The cause of mal BMI range as a reference category and consequently death was available to us only from 1981 onwards. Deaths may attenuate the risk attributed to obesity and overes- among recruits are documented in IDF computer records timate the threshold BMI associated with increased risk since 1967 with an indicator whether service-related (no for death. Notably, the risk among underweight women other causes are provided). Follow-up terminated on June and men at adolescence is controversial and it is unclear 30th 2011, or at the date of death, whichever came first. if the association is confounded by coexisting chronic ill- ness [10, 11], or biased by recalled (vs. measured) BMI data [12]. We recently studied the association between Data collection and study variables BMI in late adolescence and future risk for death attrib- Health examinations and review of the medical history uted to cardiovascular disease [13] or diabetes [14] in a were performed by trained military physicians. Stand- cohort of 2.3  million adolescents. Here, we compared ardized measurements of weight and height were under- the sex-specific relationships between adolescent BMI taken by trained personnel with each participant barefoot and cardiovascular and non-cardiovascular mortality in and in underwear; BMI was calculated. Data regarding midlife. Furthermore, we aimed at identifying and com- education, residential socioeconomic status, country of paring the sex-specific BMI threshold values associated birth and country of origin were recorded. Education was with increased risk for future mortality. grouped into ≤ 9, 10, 11 or 12 years of formal schooling. Socioeconomic status (SES), based on locality of resi- dence at the time of study enrolment [19], was grouped Materials and methods into low, medium and high. Country of origin (classified Study population by father’s or grandfather’s country of birth if the father All citizens obligated for military service in Israel are was born abroad) and country of birth were grouped as required to undergo a compulsory medical assessment reported previously [13]. at age of 17  years. Figure  1 displays the examination process and study design. Between January 1 1967 and Statistical analysis December 31 2010 2,454,693 adolescents were examined BMI was treated as a continuous variable and was also at ages 16–19  years. Participants with missing BMI data grouped according to a modification of the World (n = 64,186) and non-Jewish minorities who were unrep- Health Organization classification by splitting of the nor - resentative of their source population (n = 92,377) were mal range, with the following subgroups: BMI < 18.50 excluded from the analyses. Included in the final sam - (underweight), 18.50 ≤ BMI < 22.0 (low-normal), ple were 927,868 women and 1,366,271 men for a total 22.0 ≤ BMI < 25.0 (high-normal), 25.0 ≤ BMI < 30.0 (over- of 2,294,139 examinees. As noted in previous studies of weight) and BMI ≥ 30  kg/m (obese), as at age 17  years this cohort [13, 14], orthodox and ultra-orthodox Jew- adolescents have completed > 98% of their growth [20]. ish women are not legally obligated for military service Person-year mortality rates were calculated with follow- and therefore, may not be examined, and therefore may up commencing from 1981. Cox proportional hazard be under-represented here. The Jewish adolescent men models stratified by sex were used to estimate the haz - in this study can be considered a nationally representa- ard ratios (HRs) and 95% confidence intervals (CI) for tive sample [13]. The Israeli Medical Corps Institutional CVD outcomes and non-CVD related outcomes with Review Board provided ethical approval for the study and the low-normal BMI group as the reference category. We waived the need of informed consent given strict mainte- considered as potential confounders in the multivariable nance of participants’ anonymity (Fig. 1). Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 3 of 10 Pre-recruitment evaluation between 1967-2010 (n=2,454,693) Medical assessment • Review of health summary by participant’s family physician • Detailed medical interview and physical examination by a physician • Anthropometric measurement Sociodemographic assessment • Education • Country of origin • Socio-economic status Excluded • Missing BMI data (n=64,186) • Deaths between 1967 and 1981 (n=3,991) • Non-Jewish minorities that were unrepresentative of their source population (n=92,377) 1,366,271 men and 927,868 women studied 32,127 deaths (1981-2011) Men Women Cardiovascular death (n=3,208) Cardiovascular deaths (n=451) Non-cardiovascular deaths (n=22,233) Non-cardiovascular deaths (n=6,235) Fig. 1 Flow diagram of study design and outcomes. Cardiovascular deaths were considered as those attributed to an underlying cardiovascular cause (ICD‑9: 390–459; ICD ‑10: I00–I99) or diabetes (ICD ‑9: 250; ICD ‑10: E08–E13), whereas the remaining deaths were classified as non‑ cardiovascular. There were 1662 deaths (4.9% of the total N of deaths) for whom the cause of death was not available model all available variables that were significantly asso - health status at baseline (i.e. no indication of any medi- ciated with BMI and cardiovascular outcomes (p < 0.05; cal diagnosis in the medical review that would require age, birth year, sex, residential SES, education, country of chronic medical treatment or would limit ability to origin and height). Adjusted Cox regression spline mod- serve in a combat unit) [13, 21] to avert the possibility els (SAS, version 9.4) were fit to estimate the BMI value of reverse causality. In a separate analysis, we limited associated with minimum mortality risk for each of the the analysis to participants with a follow-up of at least study outcomes. Cubic splines with three equally spaced 3 decades (enrolled between 1967 and 1981) to allow a knots positioned between the minimum and maximum meaningful and equal period of follow-up between the values of the variable were presented. Spline models for sexes. We also analyzed the association between ado- Cox proportional hazard were performed with SAS/ lescent BMI and mortality with the reference group STAT and SAS/GRAPH software version 9.4 SAS insti- defined as the standard normal range (18.5–25.0  kg/ tute Inc., Cary, NC, USA. Sex interaction was computed m ). Multiple imputation was applied to those with with BMI as a continuous variable both in unadjusted missing data (1.4% of examinees) as reported previ- and multivariable-adjusted models. ously [13]. Analyses were performed using SPSS (ver- Several sensitivity analyses were conducted. We sion 23.0), unless mentioned otherwise. restricted the Cox analysis to those with unimpaired Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 4 of 10 Women Cardiovascular mortality Non-cardiovascular mortality HR=1.5 HR=3.0 HR=2.0 HR=1.2 BMI (kg/m ) BMI (kg/m ) Men Cardiovascular mortality Non-cardiovascular mortality HR=1.5 HR=3.0 HR=1.2 HR=2.0 BMI (kg/m ) BMI (kg/m ) Fig. 2 The relationship between adolescent BMI and mortality. Spline analysis demonstrating the non‑linear relationship between BMI at adolescence and adulthood mortality among women (upper panels) and men (lower panels) comparing cardiovascular disease‑related mortality (left) and non‑ cardiovascular disease mortality (right). The Cox models were adjusted for age, birth year, sex, residential socioeconomic status, education, country of origin and height. Dashed lines show Hazard ratios of 1.2, 1.5, 2.0 and 3.0 and their matching BMI levels (dashed arrows). The BMI level where minimal risk exists is marked by a vertical line and a full arrow During 17,346,230 and 28,367,431 person-years of Results follow up among women and men, respectively, there Baseline characteristics are shown in Table 1. The mean were 451 and 3208 CVD deaths (mean ages at death age at enrolment was 17.3 ± 0.4  years for women and 41.9 ± 10.9  years and 47.0 ± 9.1  years), respectively, and 17.4 ± 0.4  years for men, with over 85% of inductees 6235 and 22,233 non-CVD deaths (mean ages at death enrolled at age 17. The sample was heterogeneous as to the country of origin. The mean BMI values at base 39.3 ± 11.8  years and 37.5 ± 12.5  years), respectively. Notably, there were sex differences in the proportions line were 21.7 and 21.6  kg/m for women and men, of cardiovascular mortality; coronary heart disease, respectively. Ln of Hazard Ratio Ln of Hazard Ratio Ln of Hazard Ratio Ln of Hazard Ratio Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 5 of 10 Table 1 Baseline characteristics of the study cohort BMI (kg/m ) Total < 18.5 18.50–21.99 22.0–24.99 25.0–29.99 ≥ 30.0 Women Number of participants 128,876 448,712 224,937 99,867 25,476 927,868 Age (years) 17.3 17.3 17.3 17.3 17.3 17.3 Height (cm) 162.8 162.1 161.8 161.8 162.3 162.1 Mean BMI (kg/m ) 17.5 20.3 23.3 26.8 32.9 21.7 12 years schooling (%) 91 91 90 89 90 91 Low residential SES (%) 20 20 21 22 23 21 Born in Israel (%) 87 87 86 87 87 87 Country of origin (%) Israel 13.7 48.9 23.9 10.7 2.8 USSR 13.7 47.3 24.5 11.4 3.1 Asia 16.4 49.0 22.6 9.7 2.4 Africa 12.5 47.1 25.2 12.0 3.3 Europe 12.7 49.2 25.0 10.6 2.5 Ethiopia 29.3 45.4 16.5 7.1 1.7 Men Number of participants 186,190 678,005 324,665 139,556 37,855 1,366,271 Age (years) 17.3 17.3 17.4 17.4 17.4 17.4 Height (cm) 173.2 173.4 173.7 174.0 174.0 173.5 Mean BMI (kg/m ) 17.5 20.3 23.3 26.8 32.9 21.6 12 years schooling (%) 71 72 75 76 76 73 Low residential SES (%) 27 27 26 27 30 27 Born in Israel (%) 84 83 82 83 84 83 Country of origin (%) Israel 13.2 48.6 24.1 10.9 3.2 USSR 11.2 46.0 26.4 12.7 3.7 Asia 16.9 50.7 21.4 8.7 2.3 Africa 12.7 51.6 23.5 9.5 2.7 Europe 11.9 48.6 25.5 11.2 2.8 Ethiopia 31.2 52.6 11.2 3.9 1.1 21.3% in women vs. 43.7% in men; and stroke, 21.3% vs. to low-normal BMI in women of 0.68 (95% CI 0.46–0.98, 13.1%, respectively (Additional file  1: Table S1). The over - p = 0.04). This finding persisted when the study sam - all median follow-up of women and men was 17.4  years ple was limited to women with unimpaired health (347 (intra-quartile range [IQR], 9.3–27.0) and 19.4  years deaths; HR = 0.63, 95% CI 0.40–1.00, p = 0.049; Addi- (intra-quartile range 10.4–30.8), respectively. Obese ado- tional file  1: Table  S2), and was accentuated when the lescents had a shorter follow-up in both women (11.9, standard normal range (18.5–24.9  kg/m ) was set as IQR 6.0–19.5) and men (11.5, IQR 5.9–20.1), reflecting the reference category (HR = 0.59, 95% CI 0.41–0.86, the rise in prevalence of overweight and obesity in more p = 0.006). However, the strength of the association recent years [22]. Table  2 shows person-year incidence among obese women was reduced. Limiting the study rates and hazard ratios (HRs) for death among women sample to participants with at least 3 decades of follow- and men, the latter both unadjusted and multivariable- up was associated with an equal duration of follow- adjusted. CVD mortality in men showed a fivefold excess up between the sexes (mean for men, 36.9 ± 5.2  years; across the BMI categories and total mortality a twofold women 36.8 ± 4.5  years), reduced the sex differences excess compared with women. The underweight group in the age of death and accentuated the findings (Addi - was associated with the lowest rates of cardiovascular tional file  1: Table  S3). Both women and men exhibited death among women and men, and was associated with increased risk for cardiovascular mortality among the an unadjusted (cardiovascular protective) HR compared high-normal BMI group [HR of 1.42 (95% CI 1.14–1.77) Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 6 of 10 Table 2 Hazard ratios for CVD and non-CVD mortality in women and men stratified by BMI categories BMI (kg/m ) Total or BMI as a continuous < 18.5 18.50–21.99 22.0–24.99 25.0–29.99 ≥ 30.0 variable Women N of participants 128,876 448,712 224,937 99,867 25,476 927,868 Median follow‑up (25th; 75th) 16.4 (10.0, 25.6) 18.0 (9.7, 27.9) 17.9 (9.6, 27.9) 15.9 (8.3, 24.8) 11.9 (6.0, 19.5) 17.4 (9.3, 27.0) Cumulative follow‑up (person‑ years) 2,301,449 8,641,690 4,305,855 1,745,031 352,203 17,346,230 Cardiovascular mortality (451 deaths) N of deaths 31 189 134 76 21 451 Crude incidence (per 10 person years) 1.34 2.18 3.11 4.35 5.96 2.59 HR 0.68 1 (ref ) 1.42 2.13 3.90 1.12 3 95% CI 0.46–0.98 1.14–1.77 1.63–2.78 2.47–6.14 1.096–1.151 −8 −9 −20 p 0.043 0.002 2.9*10 4.4*10 3.3*10 Non-cardiovascular mortality (6235 deaths) N of deaths 741 3008 1621 707 158 6235 Crude incidence 32.19 34.80 37.64 40.51 44.86 35.94 HR 1.01 1 (ref ) 1.07 1.23 1.66 1.031 95% CI 0.93–1.10 1.01–1.14 1.13–1.33 1.42–1.95 1.02–1.04 −6 −10 −14 P 0.80 0.021 10 5.3*10 4.8*10 Men N of participants 186,190 678,005 324,665 139,556 37,855 1,366,271 Median follow‑up (25th; 75th) 19.4 (10.9, 29.8) 20.4 (11.3, 31.9) 19.3 (10.1, 31.1) 16.1 (8.1, 26.9) 11.5 (5.9, 20.1) 19.4 (10.4, 30.8) Cumulative follow‑up (person‑ years) 3,834,407 14,685,322 6,744,963 2,555,251 547,485 28,367,431 Cardiovascular mortality (3208 deaths) N of deaths 306 1284 865 578 175 3208 Crude incidence (per 10 person years) 7.98 8.74 12.82 22.62 31.96 11.30 HR 0.99 1 (ref ) 1.53 2.99 5.40 1.14 95% CI 0.88–1.13 1.40–1.67 2.71–3.31 4.60–6.33 1.13–1.15 −22 −103 −95 −169 p 0.93 8.7*10 10 1.8*10 3.7*10 Non-cardiovascular mortality (22,233 deaths) N of deaths 2970 11,184 5358 2158 563 22,233 Crude incidence (per 10 person years) 77.45 76.15 79.43 84.45 102.83 78.37 HR 1.04 1 (ref ) 1.06 1.16 1.48 1.019 95% CI 1.00–1.09 1.03–1.10 1.11–1.21 1.36–1.61 1.01–1.02 −10 −19 −17 p 0.040 0.0003 6.1*10 1.8*10 2.6*10 The association was assessed with Cox models adjusted for age, sex, birth year, residential socio-economic status, education, country of origin, and height vs. 1.53 (95% CI 1.40–1.67), respectively], overweight obese groups in both sexes with an interaction between [2.13 (95% CI 1.63–2.78) vs. 2.99 (95% CI 2.71–3.31), sex and BMI in an unadjusted model (p for interac- respectively] and obese adolescents [3.90 (95% CI tion = 0.012) reflecting the stronger associations among 2.47–6.14) vs. 5.40 (95% CI 4.60–6.33), respectively] women, that became less significant in the multivari - compared with low-normal BMI (18.50–21.99  kg/m ) able model (p for interaction = 0.069). Multivariable group. Nevertheless, there was no statistically significant spline models indicated a minimum risk for cardiovascu- interaction between sex and BMI for CVD death (p for lar death at adolescent BMI values of 16.8 and 18.8  kg/ interaction = 0.62). m for women and men, respectively (corresponding to Underweight women did not exhibit excess risk for the 3rd and 13th BMI US-CDC percentiles for men and non-CVD death, whereas underweight men had a small, women aged 17.4 years, respectively; Fig. 2). The thresh - but significant, increased adjusted risk (HR = 1.04; 95% old adolescent BMI values for a significantly increased CI = 1.00–1.09, p = 0.04). For this outcome, an increased cardiovascular mortality risk were 21.0 and 21.4  kg/m risk was observed for the high-normal, overweight and for women and men, respectively. For non-CVD death, Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 7 of 10 the minimum risk BMI values computed were 18. 2 and whereas in other studies the entire underweight group 20.0  kg/m for women and men, respectively (corre- was excluded from analysis [9, 39]. Several methodologi- sponding to the 12th and 29th BMI US-CDC percentiles cal inconsistencies may contribute to the controversy. A for men and women aged 17.4  years, respectively), with wide age range at enrollment that can span over four [30, significantly increased non-CVD mortality risk at BMI 37] or even seven decades [35] may blunt a strong asso- values of 20.0 and 22.0 kg/m , respectively. ciation reported at a young age [4, 40]. Lack of system- atic evaluation of health status at baseline [31, 34, 36] and Discussion use of recalled vs. measured BMI data [35, 38] are addi- Our study presents sex-specific associations of adoles - tional potential problems that were shown to particularly cent BMI with CVD and non-CVD mortality by mid- influence the underweight group [12]. In this regard, adulthood in a national cohort of young Israeli women our study sample was homogenous in age of enrollment, and men. The main findings of the study are that under - included systematic evaluation of health status at base- weight was protective for CVD mortality among ado- line and was based exclusively on measured weight and lescent women, but was associated with excess all-cause height data. Additionally, a sensitivity analysis that was mortality among men. The optimal BMI was lower in limited to participants with unimpaired health accentu- 2 2 women than in men; 16.8 kg/m vs. 18.8 kg/m for CVD ated the lower risk for both CVD and non-CVD death 2 2 mortality, and 18.2  kg/m vs. 20.0  kg/m for non-CVD among underweight women, but not among men. mortality. There was a significant interaction between sex The definition of abnormal BMI in childhood has gar - and BMI for non-CVD death. The association between nered considerable attention due to the rise of obesity BMI and non-CVD death was stronger for women than epidemic [41]. Yet, only a few studies have revisited the men in all groups except for underweight. entire so-called normal BMI range and suggested sex- CVD mortality has declined substantially over the past specific optimal BMI values in childhood with respect several decades [23, 24], but there is evidence indicating to fatal outcomes in adulthood. Here, the optimal that this positive trend among women is lagging behind BMI values for what appear to be well below the upper that of men [25], especially at young age, and even some underweight cutoff for women and within the low nor - evidence for increased CVD mortality in young women mal range for men (note that the 5th BMI CDC percen- [26, 27]. Our finding of an overall similar association tile at age 17.4  years is 17.3 and 17.9  kg/m for women between adolescent BMI and CVD mortality in both and men, respectively). Data from large meta-analyses sexes is in agreement with previous reports [5, 28]. We that were limited to adults (with age at enrolment that report that overweight and obesity at adolescence were spanned at least 4 decades) pointed to an optimal range associated with an increased risk for cardiovascular between 23.0 and 25.0 kg/m for all-cause mortality, and death with lower risk estimates for cardiovascular mor- 22.5  kg/m for cardiovascular mortality [28, 35, 37]. In tality among women than men. Most previous studies an Austrian cohort that investigated the association of showed a non-linear, ‘U’ or ‘J’-curve shaped, relationship age specific BMI and all-cause mortality a clear trend of between BMI and mortality [29–31]. Baker et  al. found increasing optimal  BMI was witnessed but children and linear associations between childhood BMI and coronary adolescence were not included [42]. A pooled analy- heart disease with a mildly stronger association among sis of 239 studies indicated an optimal BMI value lower boys [5]. In a study that included 238 overweight or obese than 22.0 kg/m for all-cause mortality when the sample adolescents that were followed for 60  years, Must et  al. was limited to young adults (age 35–49  years) with sig- reported a HR for coronary mortality of 2.3 (1.4–4.1) and nificantly higher risk estimates for underweight men than 13.2 (1.6–108.0) for overweight and obese boys, but no underweight women [40]. risk for overweight and obese girls [9]. While this find - Several lines of evidence support associations between ing was reported elsewhere [32],a large Norwegian study underweight and metabolic fitness or longevity. Caloric that included 230,000 adolescents that were followed for restriction, defined as a reduced intake of calories not approximately 3 decades, showed a similar HR for car- causing malnutrition, has been demonstrated as an diovascular death in both sexes [33]. These controversies intervention that can prolong life and health span [43] may be attributed to differences in follow-up, definition by activating cellular pathways such as autophagy [44, of the reference group, exclusion of participants with 45]. Epidemiological studies on lifespan of Okinawans in BMI in the low normal range [9], or the definition used Japan reported low caloric intake, lower mean BMI, lower for cardiovascular mortality. There are conflicting results cardiovascular- and age-related morbidity and prolonged on the association between underweight and mortal- longevity with an accentuated association between lower ity risk [12, 31, 34, 35], with reports of excess mortality BMI and cardiovascular morbidity among women than [30, 36–38] or reduced risk (only among women) [11], men in this population [46]. We have shown recently in Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 8 of 10 this cohort that healthy underweight women, but not between adolescent BMI and cardiovascular-specific underweight men of any other BMI group, exceed by up mortality. Finally, conclusions from this cohort are to 4.1  cm their expected height at age 17  years, a well- based on midlife mortality in which cardiovascular established risk marker for cardiovascular health [47]. death constituted 11.2%, similarly to a fraction of 15% The mechanisms underlying these sex-specific differ - that was reported in a nationwide Swedish study whose ences are complex and may include sex hormones, gut design was similar to ours [57], and should not be microbiome-related mechanisms, nutrition and stress extrapolated to older ages. [48]. Additional link between lower BMI and long-term The strengths of the study include its large sample size, outcomes, most notably cardiovascular, was demon- systematic collection of measured anthropometric data, strated following bariatric surgery, showing that both long follow up and large person-years database. This diabetic and non-diabetic patients enjoy a dramatic allowed the analysis to focus specifically on women, with reduction in the risk for cardiovascular mortality decades adequate power despite a relatively low rate of events in following surgery [49]. women in midlife. Several limitations of the current study should be addressed. First, BMI data were collected at a single time point with no available adulthood BMI measure- Conclusions ments, preventing us from determining whether the To conclude, underweight at adolescence appears to association presented here is independent of adulthood engender a protective effect among women but not men BMI. This is a limitation as there is evidence to support with respect to adulthood CVD-related mortality. The that trajectory of BMI from childhood to adulthood has lowest risk for CVD related mortality is seen with lower affects outcomes [50, 51]. In a sub-sample of this cohort than normal BMI values in women and in the low-nor- the correlation of adolescent BMI at age 17  years and mal BMI range in men. adult BMI at age 50  years was 0.53 (i.e. an R of about Additional file 25%) [52]. This would be expected to dilute the asso - ciation of adolescent BMI with mortality if all the asso- Additional file 1. Additional tables. ciation was mediated through adult BMI. Nevertheless, we find a substantive point estimate of the hazard ratio with lower 95% confidence bounds of 1.6 and 2.5 for Authors’ contributions overweight and obese women, respectively, suggest- AF and GT conceived and designed the study, analyze and interpreted the data and drafted the manuscript. ED performed the statistical analysis. All ing BMI at adolescence to be a remarkably strong risk authors interpreted the results and critically review the manuscript. GT had full marker for fatal outcomes in midlife. Second, we were access for the data and take full responsibility for the integrity of the data. All unable to adjust the reported results for established authors read and approved the final manuscript. risk factors such as smoking and exercise level, and Author details could also not adjust for the effect of adverse events or 1 The Israel Defense Forces Medical Corps, Tel Hashomer, Ramat Gan, Israel. 2 3 comorbidities which evolved from adolescence to early The Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel. Pediat‑ ric Endocrine and Diabetes Unit, Edmond and Lily Safra Children’s Hospital, adulthood and may carry increased risk of adverse out- Sheba Medical Center, Ramat‑Gan, Israel. The Dr. Pinchas Bornstein Talpiot comes [53] as these were not collected in our cohort Medical Leadership Program, Sheba Medical Center, Ramat‑Gan, Israel. at age 17  years. Yet, the association remained nearly Institute of Endocrinology, Sheba Medical Center, Tel Hashomer, Ramat‑Gan, Israel. The Division of Endocrinology, Diabetes and Hypertension, Brigham unchanged following adjustment for socio-demo- and Women’s Hospital, Harvard Medical School, Boston, MA, USA. Hebrew graphic variables, which were previously shown to be University‑Hadassah School of Public Health and Community Medicine, Ein strongly related to smoking in a subgroup of this cohort Kerem, Jerusalem, Israel. Department of Medicine, Sheba Medical Center, Tel Hashomer, Ramat‑Gan, Israel. [54]. Third, other than BMI and its components, this cohort did not include other anthropometric meas- Acknowledgements urements that were found to be strong predictors of The author wish to thank Mrs. Dorit Tzur for technical assistance during the study. CVD mortality, independent of BMI [55]. Nevertheless, BMI is a well-studied measure that is currently recom- Competing interests mended by the US Preventive Services Task Force as The authors declare that they have no competing interests. the screening measure of choice for childhood and ado- Availability of data and materials lescent obesity [56]. In addition, as mentioned, we were The current database is not publically available due to individual privacy of the unable to allocate the specific cause of death for indi - participants. However, it may be available from the corresponding author on reasonable request. viduals who died during the years 1967–1981. Never- theless, we previously showed by simulations that this Consent for publication gap of knowledge is unlikely to change the association Not applicable. Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 9 of 10 Ethics approval and consent to participate 17. Afshin A, Micha R, Khatibzadeh S, Fahimi S, Shi P, Powles J, Singh G, The Israeli Medical Corps Institutional Review Board provided ethical approval Yakoob MY, Abdollahi M, Al‑Hooti S, et al. The impact of dietary habits for the study and waived the need of informed consent given strict mainte‑ and metabolic risk factors on cardiovascular and diabetes mortality in nance of participants’ anonymity. countries of the Middle East and North Africa in 2010: a comparative risk assessment analysis. BMJ Open. 2015;5(5):e006385. Funding 18. Russell‑ Jones D, Khan R. Insulin‑associated weight gain in dia‑ None. betes–causes, effects and coping strategies. Diab Obes Metab. 2007;9(6):799–812. 19. Tzibel N. Characterization and classification of local authorities by the Publisher’s Note socio‑ economic level of the population 2006. Hebrew: Central Bureau of Springer Nature remains neutral with regard to jurisdictional claims in pub‑ Statistics; 2009. lished maps and institutional affiliations. 20. Greulich WW, Pyle SI. Radiographic atlas of skeletal development of the hand and wrist. Stanford: Stanford University Press; 1959. Received: 6 March 2018 Accepted: 28 May 2018 21. Twig G, Ben‑Ami Shor D, Furer A, Levine H, Derazne E, Goldberger N, Haklai Z, Levy M, Afek A, Leiba A, et al. Adolescent body mass index and cardiovascular disease–specific mortality by midlife. J Clin Endocrinol Metab. 2017;102(8):3011–20. 22. Twig G, Afek A, Shamiss A, Derazne E, Landau Rabbi M, Tzur D, Gordon B, References Tirosh A. Adolescence BMI and trends in adulthood mortality: a study of 1. Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, Abraham 2.16 million adolescents. J Clin Endocrinol Metab. 2014;99(6):2095–103. J, Adair T, Aggarwal R, Ahn SY, et al. Global and regional mortality from 23. Sidney S, Quesenberry CP Jr, Jaffe MG, et al. Recent trends in cardiovas‑ 235 causes of death for 20 age groups in 1990 and 2010: a system‑ cular mortality in the united states and public health goals. JAMA Cardiol. atic analysis for the Global Burden of Disease Study 2010. Lancet. 2016;1:594–9. 2012;380(9859):2095–128. 24. Mensah GA, Wei GS, Sorlie PD, Fine LJ, Rosenberg Y, Kaufmann PG, Mus‑ 2. Mathers CD, Loncar D. Projections of global mortality and burden of solino ME, Hsu LL, Addou E, Engelgau MM, et al. Decline in cardiovascular disease from 2002 to 2030. PLoS Med. 2006;3(11):e442. mortality. Possible Causes Implic. 2017;120(2):366–80. 3. Ogden CL, Carroll MD, Lawman HG, Fryar CD, Kruszon‑Moran D, Kit BK, 25. Gupta A, Wang Y, Spertus JA, Geda M, Lorenze N, Nkonde‑Price C, Flegal KM. Trends in obesity prevalence among children and adolescents D’Onofrio G, Lichtman JH, Krumholz HM. Trends in acute myocardial in the United States, 1988‑1994 through 2013‑2014. J Am Med Assoc. infarction in young patients and differences by sex and race, 2001 to 2016;315(21):2292–9. 2010. J Am Coll Cardiol. 2014;64(4):337–45. 4. Bender R, Jockel KH, Trautner C, Spraul M, Berger M. Eec ff t of age on 26. Ford ES, Capewell S. Coronary heart disease mortality among young excess mortality in obesity. J Am Med Assoc. 1999;281(16):1498–504. adults in the U.S. from 1980 through 2002: concealed leveling of mortal‑ 5. Baker JL, Olsen LW, Sorensen TI. Childhood body‑mass index and ity rates. J Am Coll Cardiol. 2007;2007(50):2128–32. the risk of coronary heart disease in adulthood. N Engl J Med. 27. Bairey Merz CN, Andersen H, Sprague E, Burns A, Keida M, Walsh MN, 2007;357(23):2329–37. Greenberger P, Campbell S, Pollin I, McCullough C, et al. Knowledge, 6. Tirosh A, Shai I, Afek A, Dubnov‑Raz G, Ayalon N, Gordon B, Derazne E, attitudes, and beliefs regarding cardiovascular disease in women: the Tzur D, Shamis A, Vinker S, et al. Adolescent BMI trajectory and risk of women’s heart alliance. J Am Coll Cardiol. 2017;70(2):123–32. diabetes versus coronary disease. N Engl J Med. 2011;364(14):1315–25. 28. Mongraw‑ Chaffin ML, Peters SAE, Huxley RR, Woodward M. The sex ‑ 7. Falkstedt D, Hemmingsson T, Rasmussen F, Lundberg I. Body mass index specific association between BMI and coronary heart disease: a system‑ in late adolescence and its association with coronary heart disease atic review and meta‑analysis of 95 cohorts with 1.2 million participants. and stroke in middle age among Swedish men. Int J Obes (Lond). Lancet Diab Endocrinol. 2015;3(6):437–49. 2007;31(5):777–83. 29. Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW. Body‑mass index 8. Ohlsson C, Bygdell M, Sonden A, Rosengren A, Kindblom JM. Association and mortality in a prospective cohort of US adults. New Engl J Med. between excessive BMI increase during puberty and risk of cardiovascu‑ 1999;341:1097–105. lar mortality in adult men: a population‑based cohort study. Lancet Diab 30. Katzmarzyk PT, Craig CL, Bouchard C. Original article underweight, over‑ Endocrinol. 2016;4(12):1017–24. weight and obesity: relationships with mortality in the 13‑ year follow‑up 9. Must A, Jacques P, Dallal G, Bajema C, Dietz W. Long‑term morbidity and of the Canada Fitness Survey. J Clin Epidemiol. 2001;54(9):916–20. mortality of overweight adolescents. A follow‑up of the Harvard growth 31. Seidell JC, Verschuren WM, van Leer EM, Kromhout D. Overweight, under‑ study of 1922 to 1935. N Engl J Med. 1992;327(19):1350–5. weight, and mortality: a prospective study of 48,287 men and women. 10. Strand BH, Kuh D, Shah I, Guralnik J, Hardy R. Childhood, adolescent Arch Intern Med. 1996;156:958–63. and early adult body mass index in relation to adult mortality: results 32. Gunnell DJ, Frankel SJ, Nanchahal K, Peters TJ, Davey Smith G. Childhood from the British 1946 birth cohort. J Epidemiol Community Health. obesity and adult cardiovascular mortality: a 57‑ y follow‑up study based 2012;66(3):225–32. on the Boyd Orr cohort. Am J Clin Nutr. 1998;67:1111–8. 11. Wandell PE, Carlsson AC, Theobald H. The association between BMI value 33. Bjorge T, Engeland A, Tverdal A, Smith GD. Body mass index in adoles‑ and long‑term mortality. Int J Obes (Lond). 2009;33(5):577–82. cence in relation to cause‑specific mortality: a follow‑up of 230,000 12. Roh L, Braun J, Chiolero A, Bopp M, Rohrmann S, Faeh D. Mortality risk Norwegian adolescents. Am J Epidemiol. 2008;168:30–7. associated with underweight: a census‑linked cohort of 31,578 individu‑ 34. Manson JE, Stampfer MJ, Hennekens CH, Willett WC. Body weight and als with up to 32 years of follow‑up. BMC Public Health. 2014;14(1):371. longevity. A reassessment. JAMA, J Am Med Assoc. 1987;257(3):353–8. 13. Twig G, Yaniv G, Levine H, Leiba A, Goldberger N, Derazne E, Ben‑Ami 35. Berrington de Gonzalez A, Hartge P, Cerhan JR, Flint AJ, Hannan L, MacIn‑ Shor D, Tzur D, Afek A, Shamiss A, et al. Body‑mass index in 2.3 million nis RJ, Moore SC, Tobias GS, Anton‑ Culver H, Freeman LB, et al. Body‑mass adolescents and cardiovascular death in adulthood. N Engl J Med. index and mortality among 1.46 million white adults. N Engl J Med. 2016;374(25):2430–40. 2010;363(23):2211–19. 14. Twig G, Tirosh A, Leiba A, Levine H, Ben‑Ami Shor D, Derazne E, Haklai Z, 36. Aune D, Sen A, Prasad M, Norat T, Janszky I, Tonstad S, Romundstad P, Goldberger N, Kasher‑Meron M, Yifrach D, et al. BMI at age 17 years and Vatten LJ. BMI and all cause mortality: systematic review and non‑linear diabetes mortality in midlife: a nationwide cohort of 2.3 million adoles‑ dose‑response meta‑analysis of 230 cohort studies with 3.74 million cents. Diab Care. 2016;39(11):1996–2003. deaths among 30.3 million participants. BMJ. 2016;353:i2156. 15. Collaboration ERF. Diabetes mellitus, fasting glucose, and risk of cause‑ 37. Collaboration PS. Body‑mass index and cause ‑specific mortality in 900 specific death. N Engl J Med. 2011;2011(364):829–41. 000 adults: collaborative analyses of 57 prospective studies. Lancet. 16. Kaati G, Bygren LO, Edvinsson S. Cardiovascular and diabetes mortality 2009;373:1083–96. determined by nutrition during parents’ and grandparents’ slow growth period. Eur J Hum Genet. 2002;10:682. Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 10 of 10 38. Calle E, Thun M, Petrelli J, Rodriguez C, Heath C Jr. Body‑mass index 49. Pontiroli AE, Zakaria AS, Mantegazza E, et al. Long‑term mortality and and mortality in a prospective cohort of US adults. N Engl J Med. incidence of cardiovascular diseases and type 2 diabetes in diabetic and 1999;341:1097–105. nondiabetic obese patients undergoing gastric banding: a controlled 39. Song X, Tabák AG, Zethelius B, Yudkin JS, Söderberg S, Laatikainen study. Cardiovasc Diabetol. 2016;15:39. T, Stehouwer CD, Dankner R, Jousilahti P, Onat A. Obesity attenuates 50. Bjerregaard LG, Jensen BW, Ängquist L, Osler M, Sørensen TIA, Baker JL. gender differences in cardiovascular mortality. Cardiovasc Diabetol. Change in overweight from childhood to early adulthood and risk of 2014;13(1):144. type 2 diabetes. N Engl J Med. 2018;378(14):1302–a12. 40. Collaboration GBM. Body‑mass index and all‑ cause mortality: individual‑ 51. Peter RS, Keller F, Klenk J, Concin H, Nagel G. Body mass trajectories, participant‑ data meta‑analysis of 239 prospective studies in four conti‑ diabetes mellitus, and mortality in a large cohort of Austrian adults. nents. Lancet. 2016;388:776–86. Medicine. 2016;95:e5608. 41. Krebs NF, Himes JH, Jacobson D, Nicklas TA, Guilday P, Styne D. Assess‑ 52. Cohen‑Manheim I, Doniger GM, Sinnreich R, Simon ES, Murad H, Pinchas‑ ment of child and adolescent overweight and obesity. Pediatrics. Mizrachi R, Kark JD. Body mass index, height and socioeconomic position 2007;120(Suppl 4):S193–228. in adolescence, their trajectories into adulthood, and cognitive function 42. Peter RS, Mayer B, Concin H, Nagel G. The effect of age on the shape of in midlife. Alzheimer’s Dis. 2017;55(3):1207–21. the BMI‑mortality relation and BMI associated with minimum all‑ cause 53. Zucker I, Shohat T, Dankner R, Chodick G. New onset diabetes in adult‑ mortality in a large Austrian cohort. Int J Obes (Lond). 2015;39:530–4. hood is associated with a substantial risk for mortality at all ages: a 43. Lopez‑ Otin C, Galluzzi L, Freije JMP, Madeo F, Kroemer G. Metabolic con‑ population based historical cohort study with a decade‑long follow‑up. trol of longevity. Cell. 2016;166:802–21. Cardiovasc Diabetol. 2017;16:105. 44. Madeo F, Zimmermann A, Maiuri MC, Kroemer G. Essential role for 54. Kark JD, Laor A. Cigarette smoking and educational level among young autophagy in life span extension. J Clin Investig. 2015;125(1):85–93. Israelis upon release from military service in 1988—a public health chal‑ 45. Kark JD, Goldberger N, Kimura M, Sinnreich R, Aviv A. Energy intake lenge. Isr J Med Sci. 1992;28(1):33–7. and leukocyte telomere length in young adults. Am J Clin Nutr. 55. Pischon T, Boeing H, Hoffmann K, Bergmann M, Schulze MB, Overvad K, 2012;95:479–87. van der Schouw Y T, Spencer E, Moons KG, Tjonneland A, et al. General 46. Willcox BJ, Willcox DC, Suzuki M. Demographic, phenotypic, and genetic and abdominal adiposity and risk of death in Europe. N Engl J Med. characteristics of centenarians in Okinawa and Japan: part 1‑ centenarians 2008;359(20):2105–20. in Okinawa. Mech Ageing Dev. 2017;165(Pt B):75–9. 56. Force USPST, Grossman DC, Bibbins‑Domingo K, Curry SJ, Barry MJ, David‑ 47. Pinhas‑Hamiel O, Reichman B, Shina A, Derazne E, Tzur D, Yifrach D, Wiser son KW, Doubeni CA, Epling JW, Jr., Kemper AR, Krist AH, et al. Screening I, Afek A, Shamis A, Tirosh A, et al. Sex differences in the impact of thin‑ for obesity in children and adolescents: US preventive services task force ness, overweight, obesity, and parental height on adolescent height. J recommendation statement. J Am Med Assoc. 2017;317(23):2417–26. Adolesc Health. 2017;61(2):233–9. 57. Hogstrom G, Nordstrom A, Nordstrom P. Aerobic fitness in late ado ‑ 48. Ostan R, Monti D, Gueresi P, Bussolotto M, Franceschi C, Baggio G. Gender, lescence and the risk of early death: a prospective cohort study of 1.3 aging and longevity in humans: an update of an intriguing/neglected million Swedish men. Int J Epidemiol. 2015;2015;45:1159–68. scenario paving the way to a gender‑specific medicine. Clin Sci (London, England: 1979). 2016;130(19):1711–25. Ready to submit your research ? Choose BMC and benefit from: fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Cardiovascular Diabetology Springer Journals

Sex-specific associations between adolescent categories of BMI with cardiovascular and non-cardiovascular mortality in midlife

Free
10 pages

Loading next page...
 
/lp/springer_journal/sex-specific-associations-between-adolescent-categories-of-bmi-with-ejf00TSING
Publisher
BioMed Central
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Diabetes; Angiology; Cardiology
eISSN
1475-2840
D.O.I.
10.1186/s12933-018-0727-7
Publisher site
See Article on Publisher Site

Abstract

Context: Most studies linking long‑ term consequences of adolescent underweight and obesity are limited to men. Objective: To assess the sex‑ specific association of adolescent BMI with cardiovascular ‑ and non‑ cardiovascular‑ related mortality in young adulthood and midlife. Setting: A nationwide cohort. Participants: 927,868 women, 1,366,271 men. Interventions: Medical examination data at age 17, including BMI, were linked to the national death registry. Main outcomes: Death attributed to cardiovascular (CVD) and non‑ CVD causes. Results: During 17,346,230 women‑ years and 28,367,431 men‑ years of follow‑ up, there were 451 and 3208 CVD deaths, respectively, and 6235 and 22,223 non‑ CVD deaths, respectively. Compared to low‑ normal BMI (18.5–22.0 kg/ m ), underweight women had a lower adjusted risk for CVD mortality (Cox hazard ratio (HR) = 0.68; 95% CI 0.46–0.98) in contrast to underweight men (HR = 0.99; 0.88–1.13). The latter were at higher risk for non‑ CVD mortality (HR = 1.04; 1.00–1.09), unlike underweight women (HR = 1.01; 0.93–1.10). Findings, which persisted when the study sample was limited to those with unimpaired health, were accentuated for the obese with ≥ 30 years follow‑ up. Both sexes exhibited similarly higher risk estimates already in the high‑ normal BMI range (22.0 ≤ BMI < 25.0 kg/m ) with overall no interaction between sex and BMI (p = 0.62). Adjusted spline models suggested lower BMI values for minimal mortality 2 2 risk among women (16.8 and 18.2 kg/m ) than men (18.8 and 20.0 kg/m ), for CVD and non‑ CVD death, respectively. Conclusions: Underweight adolescent females have favorable cardiovascular outcomes in adulthood. Otherwise the risk patterns were similar between the sexes. The optimal BMI value for women and men with respect to future CVD outcomes is within or below the currently accepted low‑ normal BMI range. Keywords: Adolescence, Body mass index, Cohort study, Obesity, Cardiovascular death, Sex, Women, Underweight *Correspondence: Gilad.Twig@gmail.com The Israel Defense Forces Medical Corps, Tel Hashomer, Ramat Gan, Israel Full list of author information is available at the end of the article © The Author(s) 2018. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creat iveco mmons .org/ publi cdoma in/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 2 of 10 Primary outcomes and documentation of the cause Introduction of death Cardiovascular disease remains the leading cause of The primary outcomes of the study were either death death in the Western world [1, 2]. Over the last few dec- attributed to cardiovascular causes according to the ades, the prevalence of obesity has risen worldwide to as International Classification of Diseases (ICD-9: 390– high as 20% of the adolescent population in the US [3]. 459; ICD-10: I00–I99) or deaths attributed to non-car- There is increasing evidence for the link between adoles - diovascular etiologies. Given that deaths attributed to cent overweight and obesity and cardiovascular and all- diabetes (ICD-9: 250; ICD-10: E08–E13) are closely asso- cause mortality [4–6]. ciated with cardiovascular deaths [15], we grouped these Many of the studies included only men [6–8] or together with cardiovascular deaths as was performed in reported on a small number of deaths among women previous studies [16–18]. The underlying cause of death, [9]. Furthermore, the majority of the studies used World as officially coded from death notifications by the Israel Health Organization (WHO) or the Centers for Disease Central Bureau of Statistics, was linked to the database Control (CDC) classifications which include a broad nor - using the participants’ national ID number. The cause of mal BMI range as a reference category and consequently death was available to us only from 1981 onwards. Deaths may attenuate the risk attributed to obesity and overes- among recruits are documented in IDF computer records timate the threshold BMI associated with increased risk since 1967 with an indicator whether service-related (no for death. Notably, the risk among underweight women other causes are provided). Follow-up terminated on June and men at adolescence is controversial and it is unclear 30th 2011, or at the date of death, whichever came first. if the association is confounded by coexisting chronic ill- ness [10, 11], or biased by recalled (vs. measured) BMI data [12]. We recently studied the association between Data collection and study variables BMI in late adolescence and future risk for death attrib- Health examinations and review of the medical history uted to cardiovascular disease [13] or diabetes [14] in a were performed by trained military physicians. Stand- cohort of 2.3  million adolescents. Here, we compared ardized measurements of weight and height were under- the sex-specific relationships between adolescent BMI taken by trained personnel with each participant barefoot and cardiovascular and non-cardiovascular mortality in and in underwear; BMI was calculated. Data regarding midlife. Furthermore, we aimed at identifying and com- education, residential socioeconomic status, country of paring the sex-specific BMI threshold values associated birth and country of origin were recorded. Education was with increased risk for future mortality. grouped into ≤ 9, 10, 11 or 12 years of formal schooling. Socioeconomic status (SES), based on locality of resi- dence at the time of study enrolment [19], was grouped Materials and methods into low, medium and high. Country of origin (classified Study population by father’s or grandfather’s country of birth if the father All citizens obligated for military service in Israel are was born abroad) and country of birth were grouped as required to undergo a compulsory medical assessment reported previously [13]. at age of 17  years. Figure  1 displays the examination process and study design. Between January 1 1967 and Statistical analysis December 31 2010 2,454,693 adolescents were examined BMI was treated as a continuous variable and was also at ages 16–19  years. Participants with missing BMI data grouped according to a modification of the World (n = 64,186) and non-Jewish minorities who were unrep- Health Organization classification by splitting of the nor - resentative of their source population (n = 92,377) were mal range, with the following subgroups: BMI < 18.50 excluded from the analyses. Included in the final sam - (underweight), 18.50 ≤ BMI < 22.0 (low-normal), ple were 927,868 women and 1,366,271 men for a total 22.0 ≤ BMI < 25.0 (high-normal), 25.0 ≤ BMI < 30.0 (over- of 2,294,139 examinees. As noted in previous studies of weight) and BMI ≥ 30  kg/m (obese), as at age 17  years this cohort [13, 14], orthodox and ultra-orthodox Jew- adolescents have completed > 98% of their growth [20]. ish women are not legally obligated for military service Person-year mortality rates were calculated with follow- and therefore, may not be examined, and therefore may up commencing from 1981. Cox proportional hazard be under-represented here. The Jewish adolescent men models stratified by sex were used to estimate the haz - in this study can be considered a nationally representa- ard ratios (HRs) and 95% confidence intervals (CI) for tive sample [13]. The Israeli Medical Corps Institutional CVD outcomes and non-CVD related outcomes with Review Board provided ethical approval for the study and the low-normal BMI group as the reference category. We waived the need of informed consent given strict mainte- considered as potential confounders in the multivariable nance of participants’ anonymity (Fig. 1). Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 3 of 10 Pre-recruitment evaluation between 1967-2010 (n=2,454,693) Medical assessment • Review of health summary by participant’s family physician • Detailed medical interview and physical examination by a physician • Anthropometric measurement Sociodemographic assessment • Education • Country of origin • Socio-economic status Excluded • Missing BMI data (n=64,186) • Deaths between 1967 and 1981 (n=3,991) • Non-Jewish minorities that were unrepresentative of their source population (n=92,377) 1,366,271 men and 927,868 women studied 32,127 deaths (1981-2011) Men Women Cardiovascular death (n=3,208) Cardiovascular deaths (n=451) Non-cardiovascular deaths (n=22,233) Non-cardiovascular deaths (n=6,235) Fig. 1 Flow diagram of study design and outcomes. Cardiovascular deaths were considered as those attributed to an underlying cardiovascular cause (ICD‑9: 390–459; ICD ‑10: I00–I99) or diabetes (ICD ‑9: 250; ICD ‑10: E08–E13), whereas the remaining deaths were classified as non‑ cardiovascular. There were 1662 deaths (4.9% of the total N of deaths) for whom the cause of death was not available model all available variables that were significantly asso - health status at baseline (i.e. no indication of any medi- ciated with BMI and cardiovascular outcomes (p < 0.05; cal diagnosis in the medical review that would require age, birth year, sex, residential SES, education, country of chronic medical treatment or would limit ability to origin and height). Adjusted Cox regression spline mod- serve in a combat unit) [13, 21] to avert the possibility els (SAS, version 9.4) were fit to estimate the BMI value of reverse causality. In a separate analysis, we limited associated with minimum mortality risk for each of the the analysis to participants with a follow-up of at least study outcomes. Cubic splines with three equally spaced 3 decades (enrolled between 1967 and 1981) to allow a knots positioned between the minimum and maximum meaningful and equal period of follow-up between the values of the variable were presented. Spline models for sexes. We also analyzed the association between ado- Cox proportional hazard were performed with SAS/ lescent BMI and mortality with the reference group STAT and SAS/GRAPH software version 9.4 SAS insti- defined as the standard normal range (18.5–25.0  kg/ tute Inc., Cary, NC, USA. Sex interaction was computed m ). Multiple imputation was applied to those with with BMI as a continuous variable both in unadjusted missing data (1.4% of examinees) as reported previ- and multivariable-adjusted models. ously [13]. Analyses were performed using SPSS (ver- Several sensitivity analyses were conducted. We sion 23.0), unless mentioned otherwise. restricted the Cox analysis to those with unimpaired Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 4 of 10 Women Cardiovascular mortality Non-cardiovascular mortality HR=1.5 HR=3.0 HR=2.0 HR=1.2 BMI (kg/m ) BMI (kg/m ) Men Cardiovascular mortality Non-cardiovascular mortality HR=1.5 HR=3.0 HR=1.2 HR=2.0 BMI (kg/m ) BMI (kg/m ) Fig. 2 The relationship between adolescent BMI and mortality. Spline analysis demonstrating the non‑linear relationship between BMI at adolescence and adulthood mortality among women (upper panels) and men (lower panels) comparing cardiovascular disease‑related mortality (left) and non‑ cardiovascular disease mortality (right). The Cox models were adjusted for age, birth year, sex, residential socioeconomic status, education, country of origin and height. Dashed lines show Hazard ratios of 1.2, 1.5, 2.0 and 3.0 and their matching BMI levels (dashed arrows). The BMI level where minimal risk exists is marked by a vertical line and a full arrow During 17,346,230 and 28,367,431 person-years of Results follow up among women and men, respectively, there Baseline characteristics are shown in Table 1. The mean were 451 and 3208 CVD deaths (mean ages at death age at enrolment was 17.3 ± 0.4  years for women and 41.9 ± 10.9  years and 47.0 ± 9.1  years), respectively, and 17.4 ± 0.4  years for men, with over 85% of inductees 6235 and 22,233 non-CVD deaths (mean ages at death enrolled at age 17. The sample was heterogeneous as to the country of origin. The mean BMI values at base 39.3 ± 11.8  years and 37.5 ± 12.5  years), respectively. Notably, there were sex differences in the proportions line were 21.7 and 21.6  kg/m for women and men, of cardiovascular mortality; coronary heart disease, respectively. Ln of Hazard Ratio Ln of Hazard Ratio Ln of Hazard Ratio Ln of Hazard Ratio Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 5 of 10 Table 1 Baseline characteristics of the study cohort BMI (kg/m ) Total < 18.5 18.50–21.99 22.0–24.99 25.0–29.99 ≥ 30.0 Women Number of participants 128,876 448,712 224,937 99,867 25,476 927,868 Age (years) 17.3 17.3 17.3 17.3 17.3 17.3 Height (cm) 162.8 162.1 161.8 161.8 162.3 162.1 Mean BMI (kg/m ) 17.5 20.3 23.3 26.8 32.9 21.7 12 years schooling (%) 91 91 90 89 90 91 Low residential SES (%) 20 20 21 22 23 21 Born in Israel (%) 87 87 86 87 87 87 Country of origin (%) Israel 13.7 48.9 23.9 10.7 2.8 USSR 13.7 47.3 24.5 11.4 3.1 Asia 16.4 49.0 22.6 9.7 2.4 Africa 12.5 47.1 25.2 12.0 3.3 Europe 12.7 49.2 25.0 10.6 2.5 Ethiopia 29.3 45.4 16.5 7.1 1.7 Men Number of participants 186,190 678,005 324,665 139,556 37,855 1,366,271 Age (years) 17.3 17.3 17.4 17.4 17.4 17.4 Height (cm) 173.2 173.4 173.7 174.0 174.0 173.5 Mean BMI (kg/m ) 17.5 20.3 23.3 26.8 32.9 21.6 12 years schooling (%) 71 72 75 76 76 73 Low residential SES (%) 27 27 26 27 30 27 Born in Israel (%) 84 83 82 83 84 83 Country of origin (%) Israel 13.2 48.6 24.1 10.9 3.2 USSR 11.2 46.0 26.4 12.7 3.7 Asia 16.9 50.7 21.4 8.7 2.3 Africa 12.7 51.6 23.5 9.5 2.7 Europe 11.9 48.6 25.5 11.2 2.8 Ethiopia 31.2 52.6 11.2 3.9 1.1 21.3% in women vs. 43.7% in men; and stroke, 21.3% vs. to low-normal BMI in women of 0.68 (95% CI 0.46–0.98, 13.1%, respectively (Additional file  1: Table S1). The over - p = 0.04). This finding persisted when the study sam - all median follow-up of women and men was 17.4  years ple was limited to women with unimpaired health (347 (intra-quartile range [IQR], 9.3–27.0) and 19.4  years deaths; HR = 0.63, 95% CI 0.40–1.00, p = 0.049; Addi- (intra-quartile range 10.4–30.8), respectively. Obese ado- tional file  1: Table  S2), and was accentuated when the lescents had a shorter follow-up in both women (11.9, standard normal range (18.5–24.9  kg/m ) was set as IQR 6.0–19.5) and men (11.5, IQR 5.9–20.1), reflecting the reference category (HR = 0.59, 95% CI 0.41–0.86, the rise in prevalence of overweight and obesity in more p = 0.006). However, the strength of the association recent years [22]. Table  2 shows person-year incidence among obese women was reduced. Limiting the study rates and hazard ratios (HRs) for death among women sample to participants with at least 3 decades of follow- and men, the latter both unadjusted and multivariable- up was associated with an equal duration of follow- adjusted. CVD mortality in men showed a fivefold excess up between the sexes (mean for men, 36.9 ± 5.2  years; across the BMI categories and total mortality a twofold women 36.8 ± 4.5  years), reduced the sex differences excess compared with women. The underweight group in the age of death and accentuated the findings (Addi - was associated with the lowest rates of cardiovascular tional file  1: Table  S3). Both women and men exhibited death among women and men, and was associated with increased risk for cardiovascular mortality among the an unadjusted (cardiovascular protective) HR compared high-normal BMI group [HR of 1.42 (95% CI 1.14–1.77) Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 6 of 10 Table 2 Hazard ratios for CVD and non-CVD mortality in women and men stratified by BMI categories BMI (kg/m ) Total or BMI as a continuous < 18.5 18.50–21.99 22.0–24.99 25.0–29.99 ≥ 30.0 variable Women N of participants 128,876 448,712 224,937 99,867 25,476 927,868 Median follow‑up (25th; 75th) 16.4 (10.0, 25.6) 18.0 (9.7, 27.9) 17.9 (9.6, 27.9) 15.9 (8.3, 24.8) 11.9 (6.0, 19.5) 17.4 (9.3, 27.0) Cumulative follow‑up (person‑ years) 2,301,449 8,641,690 4,305,855 1,745,031 352,203 17,346,230 Cardiovascular mortality (451 deaths) N of deaths 31 189 134 76 21 451 Crude incidence (per 10 person years) 1.34 2.18 3.11 4.35 5.96 2.59 HR 0.68 1 (ref ) 1.42 2.13 3.90 1.12 3 95% CI 0.46–0.98 1.14–1.77 1.63–2.78 2.47–6.14 1.096–1.151 −8 −9 −20 p 0.043 0.002 2.9*10 4.4*10 3.3*10 Non-cardiovascular mortality (6235 deaths) N of deaths 741 3008 1621 707 158 6235 Crude incidence 32.19 34.80 37.64 40.51 44.86 35.94 HR 1.01 1 (ref ) 1.07 1.23 1.66 1.031 95% CI 0.93–1.10 1.01–1.14 1.13–1.33 1.42–1.95 1.02–1.04 −6 −10 −14 P 0.80 0.021 10 5.3*10 4.8*10 Men N of participants 186,190 678,005 324,665 139,556 37,855 1,366,271 Median follow‑up (25th; 75th) 19.4 (10.9, 29.8) 20.4 (11.3, 31.9) 19.3 (10.1, 31.1) 16.1 (8.1, 26.9) 11.5 (5.9, 20.1) 19.4 (10.4, 30.8) Cumulative follow‑up (person‑ years) 3,834,407 14,685,322 6,744,963 2,555,251 547,485 28,367,431 Cardiovascular mortality (3208 deaths) N of deaths 306 1284 865 578 175 3208 Crude incidence (per 10 person years) 7.98 8.74 12.82 22.62 31.96 11.30 HR 0.99 1 (ref ) 1.53 2.99 5.40 1.14 95% CI 0.88–1.13 1.40–1.67 2.71–3.31 4.60–6.33 1.13–1.15 −22 −103 −95 −169 p 0.93 8.7*10 10 1.8*10 3.7*10 Non-cardiovascular mortality (22,233 deaths) N of deaths 2970 11,184 5358 2158 563 22,233 Crude incidence (per 10 person years) 77.45 76.15 79.43 84.45 102.83 78.37 HR 1.04 1 (ref ) 1.06 1.16 1.48 1.019 95% CI 1.00–1.09 1.03–1.10 1.11–1.21 1.36–1.61 1.01–1.02 −10 −19 −17 p 0.040 0.0003 6.1*10 1.8*10 2.6*10 The association was assessed with Cox models adjusted for age, sex, birth year, residential socio-economic status, education, country of origin, and height vs. 1.53 (95% CI 1.40–1.67), respectively], overweight obese groups in both sexes with an interaction between [2.13 (95% CI 1.63–2.78) vs. 2.99 (95% CI 2.71–3.31), sex and BMI in an unadjusted model (p for interac- respectively] and obese adolescents [3.90 (95% CI tion = 0.012) reflecting the stronger associations among 2.47–6.14) vs. 5.40 (95% CI 4.60–6.33), respectively] women, that became less significant in the multivari - compared with low-normal BMI (18.50–21.99  kg/m ) able model (p for interaction = 0.069). Multivariable group. Nevertheless, there was no statistically significant spline models indicated a minimum risk for cardiovascu- interaction between sex and BMI for CVD death (p for lar death at adolescent BMI values of 16.8 and 18.8  kg/ interaction = 0.62). m for women and men, respectively (corresponding to Underweight women did not exhibit excess risk for the 3rd and 13th BMI US-CDC percentiles for men and non-CVD death, whereas underweight men had a small, women aged 17.4 years, respectively; Fig. 2). The thresh - but significant, increased adjusted risk (HR = 1.04; 95% old adolescent BMI values for a significantly increased CI = 1.00–1.09, p = 0.04). For this outcome, an increased cardiovascular mortality risk were 21.0 and 21.4  kg/m risk was observed for the high-normal, overweight and for women and men, respectively. For non-CVD death, Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 7 of 10 the minimum risk BMI values computed were 18. 2 and whereas in other studies the entire underweight group 20.0  kg/m for women and men, respectively (corre- was excluded from analysis [9, 39]. Several methodologi- sponding to the 12th and 29th BMI US-CDC percentiles cal inconsistencies may contribute to the controversy. A for men and women aged 17.4  years, respectively), with wide age range at enrollment that can span over four [30, significantly increased non-CVD mortality risk at BMI 37] or even seven decades [35] may blunt a strong asso- values of 20.0 and 22.0 kg/m , respectively. ciation reported at a young age [4, 40]. Lack of system- atic evaluation of health status at baseline [31, 34, 36] and Discussion use of recalled vs. measured BMI data [35, 38] are addi- Our study presents sex-specific associations of adoles - tional potential problems that were shown to particularly cent BMI with CVD and non-CVD mortality by mid- influence the underweight group [12]. In this regard, adulthood in a national cohort of young Israeli women our study sample was homogenous in age of enrollment, and men. The main findings of the study are that under - included systematic evaluation of health status at base- weight was protective for CVD mortality among ado- line and was based exclusively on measured weight and lescent women, but was associated with excess all-cause height data. Additionally, a sensitivity analysis that was mortality among men. The optimal BMI was lower in limited to participants with unimpaired health accentu- 2 2 women than in men; 16.8 kg/m vs. 18.8 kg/m for CVD ated the lower risk for both CVD and non-CVD death 2 2 mortality, and 18.2  kg/m vs. 20.0  kg/m for non-CVD among underweight women, but not among men. mortality. There was a significant interaction between sex The definition of abnormal BMI in childhood has gar - and BMI for non-CVD death. The association between nered considerable attention due to the rise of obesity BMI and non-CVD death was stronger for women than epidemic [41]. Yet, only a few studies have revisited the men in all groups except for underweight. entire so-called normal BMI range and suggested sex- CVD mortality has declined substantially over the past specific optimal BMI values in childhood with respect several decades [23, 24], but there is evidence indicating to fatal outcomes in adulthood. Here, the optimal that this positive trend among women is lagging behind BMI values for what appear to be well below the upper that of men [25], especially at young age, and even some underweight cutoff for women and within the low nor - evidence for increased CVD mortality in young women mal range for men (note that the 5th BMI CDC percen- [26, 27]. Our finding of an overall similar association tile at age 17.4  years is 17.3 and 17.9  kg/m for women between adolescent BMI and CVD mortality in both and men, respectively). Data from large meta-analyses sexes is in agreement with previous reports [5, 28]. We that were limited to adults (with age at enrolment that report that overweight and obesity at adolescence were spanned at least 4 decades) pointed to an optimal range associated with an increased risk for cardiovascular between 23.0 and 25.0 kg/m for all-cause mortality, and death with lower risk estimates for cardiovascular mor- 22.5  kg/m for cardiovascular mortality [28, 35, 37]. In tality among women than men. Most previous studies an Austrian cohort that investigated the association of showed a non-linear, ‘U’ or ‘J’-curve shaped, relationship age specific BMI and all-cause mortality a clear trend of between BMI and mortality [29–31]. Baker et  al. found increasing optimal  BMI was witnessed but children and linear associations between childhood BMI and coronary adolescence were not included [42]. A pooled analy- heart disease with a mildly stronger association among sis of 239 studies indicated an optimal BMI value lower boys [5]. In a study that included 238 overweight or obese than 22.0 kg/m for all-cause mortality when the sample adolescents that were followed for 60  years, Must et  al. was limited to young adults (age 35–49  years) with sig- reported a HR for coronary mortality of 2.3 (1.4–4.1) and nificantly higher risk estimates for underweight men than 13.2 (1.6–108.0) for overweight and obese boys, but no underweight women [40]. risk for overweight and obese girls [9]. While this find - Several lines of evidence support associations between ing was reported elsewhere [32],a large Norwegian study underweight and metabolic fitness or longevity. Caloric that included 230,000 adolescents that were followed for restriction, defined as a reduced intake of calories not approximately 3 decades, showed a similar HR for car- causing malnutrition, has been demonstrated as an diovascular death in both sexes [33]. These controversies intervention that can prolong life and health span [43] may be attributed to differences in follow-up, definition by activating cellular pathways such as autophagy [44, of the reference group, exclusion of participants with 45]. Epidemiological studies on lifespan of Okinawans in BMI in the low normal range [9], or the definition used Japan reported low caloric intake, lower mean BMI, lower for cardiovascular mortality. There are conflicting results cardiovascular- and age-related morbidity and prolonged on the association between underweight and mortal- longevity with an accentuated association between lower ity risk [12, 31, 34, 35], with reports of excess mortality BMI and cardiovascular morbidity among women than [30, 36–38] or reduced risk (only among women) [11], men in this population [46]. We have shown recently in Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 8 of 10 this cohort that healthy underweight women, but not between adolescent BMI and cardiovascular-specific underweight men of any other BMI group, exceed by up mortality. Finally, conclusions from this cohort are to 4.1  cm their expected height at age 17  years, a well- based on midlife mortality in which cardiovascular established risk marker for cardiovascular health [47]. death constituted 11.2%, similarly to a fraction of 15% The mechanisms underlying these sex-specific differ - that was reported in a nationwide Swedish study whose ences are complex and may include sex hormones, gut design was similar to ours [57], and should not be microbiome-related mechanisms, nutrition and stress extrapolated to older ages. [48]. Additional link between lower BMI and long-term The strengths of the study include its large sample size, outcomes, most notably cardiovascular, was demon- systematic collection of measured anthropometric data, strated following bariatric surgery, showing that both long follow up and large person-years database. This diabetic and non-diabetic patients enjoy a dramatic allowed the analysis to focus specifically on women, with reduction in the risk for cardiovascular mortality decades adequate power despite a relatively low rate of events in following surgery [49]. women in midlife. Several limitations of the current study should be addressed. First, BMI data were collected at a single time point with no available adulthood BMI measure- Conclusions ments, preventing us from determining whether the To conclude, underweight at adolescence appears to association presented here is independent of adulthood engender a protective effect among women but not men BMI. This is a limitation as there is evidence to support with respect to adulthood CVD-related mortality. The that trajectory of BMI from childhood to adulthood has lowest risk for CVD related mortality is seen with lower affects outcomes [50, 51]. In a sub-sample of this cohort than normal BMI values in women and in the low-nor- the correlation of adolescent BMI at age 17  years and mal BMI range in men. adult BMI at age 50  years was 0.53 (i.e. an R of about Additional file 25%) [52]. This would be expected to dilute the asso - ciation of adolescent BMI with mortality if all the asso- Additional file 1. Additional tables. ciation was mediated through adult BMI. Nevertheless, we find a substantive point estimate of the hazard ratio with lower 95% confidence bounds of 1.6 and 2.5 for Authors’ contributions overweight and obese women, respectively, suggest- AF and GT conceived and designed the study, analyze and interpreted the data and drafted the manuscript. ED performed the statistical analysis. All ing BMI at adolescence to be a remarkably strong risk authors interpreted the results and critically review the manuscript. GT had full marker for fatal outcomes in midlife. Second, we were access for the data and take full responsibility for the integrity of the data. All unable to adjust the reported results for established authors read and approved the final manuscript. risk factors such as smoking and exercise level, and Author details could also not adjust for the effect of adverse events or 1 The Israel Defense Forces Medical Corps, Tel Hashomer, Ramat Gan, Israel. 2 3 comorbidities which evolved from adolescence to early The Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel. Pediat‑ ric Endocrine and Diabetes Unit, Edmond and Lily Safra Children’s Hospital, adulthood and may carry increased risk of adverse out- Sheba Medical Center, Ramat‑Gan, Israel. The Dr. Pinchas Bornstein Talpiot comes [53] as these were not collected in our cohort Medical Leadership Program, Sheba Medical Center, Ramat‑Gan, Israel. at age 17  years. Yet, the association remained nearly Institute of Endocrinology, Sheba Medical Center, Tel Hashomer, Ramat‑Gan, Israel. The Division of Endocrinology, Diabetes and Hypertension, Brigham unchanged following adjustment for socio-demo- and Women’s Hospital, Harvard Medical School, Boston, MA, USA. Hebrew graphic variables, which were previously shown to be University‑Hadassah School of Public Health and Community Medicine, Ein strongly related to smoking in a subgroup of this cohort Kerem, Jerusalem, Israel. Department of Medicine, Sheba Medical Center, Tel Hashomer, Ramat‑Gan, Israel. [54]. Third, other than BMI and its components, this cohort did not include other anthropometric meas- Acknowledgements urements that were found to be strong predictors of The author wish to thank Mrs. Dorit Tzur for technical assistance during the study. CVD mortality, independent of BMI [55]. Nevertheless, BMI is a well-studied measure that is currently recom- Competing interests mended by the US Preventive Services Task Force as The authors declare that they have no competing interests. the screening measure of choice for childhood and ado- Availability of data and materials lescent obesity [56]. In addition, as mentioned, we were The current database is not publically available due to individual privacy of the unable to allocate the specific cause of death for indi - participants. However, it may be available from the corresponding author on reasonable request. viduals who died during the years 1967–1981. Never- theless, we previously showed by simulations that this Consent for publication gap of knowledge is unlikely to change the association Not applicable. Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 9 of 10 Ethics approval and consent to participate 17. Afshin A, Micha R, Khatibzadeh S, Fahimi S, Shi P, Powles J, Singh G, The Israeli Medical Corps Institutional Review Board provided ethical approval Yakoob MY, Abdollahi M, Al‑Hooti S, et al. The impact of dietary habits for the study and waived the need of informed consent given strict mainte‑ and metabolic risk factors on cardiovascular and diabetes mortality in nance of participants’ anonymity. countries of the Middle East and North Africa in 2010: a comparative risk assessment analysis. BMJ Open. 2015;5(5):e006385. Funding 18. Russell‑ Jones D, Khan R. Insulin‑associated weight gain in dia‑ None. betes–causes, effects and coping strategies. Diab Obes Metab. 2007;9(6):799–812. 19. Tzibel N. Characterization and classification of local authorities by the Publisher’s Note socio‑ economic level of the population 2006. Hebrew: Central Bureau of Springer Nature remains neutral with regard to jurisdictional claims in pub‑ Statistics; 2009. lished maps and institutional affiliations. 20. Greulich WW, Pyle SI. Radiographic atlas of skeletal development of the hand and wrist. Stanford: Stanford University Press; 1959. Received: 6 March 2018 Accepted: 28 May 2018 21. Twig G, Ben‑Ami Shor D, Furer A, Levine H, Derazne E, Goldberger N, Haklai Z, Levy M, Afek A, Leiba A, et al. Adolescent body mass index and cardiovascular disease–specific mortality by midlife. J Clin Endocrinol Metab. 2017;102(8):3011–20. 22. Twig G, Afek A, Shamiss A, Derazne E, Landau Rabbi M, Tzur D, Gordon B, References Tirosh A. Adolescence BMI and trends in adulthood mortality: a study of 1. Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, Abraham 2.16 million adolescents. J Clin Endocrinol Metab. 2014;99(6):2095–103. J, Adair T, Aggarwal R, Ahn SY, et al. Global and regional mortality from 23. Sidney S, Quesenberry CP Jr, Jaffe MG, et al. Recent trends in cardiovas‑ 235 causes of death for 20 age groups in 1990 and 2010: a system‑ cular mortality in the united states and public health goals. JAMA Cardiol. atic analysis for the Global Burden of Disease Study 2010. Lancet. 2016;1:594–9. 2012;380(9859):2095–128. 24. Mensah GA, Wei GS, Sorlie PD, Fine LJ, Rosenberg Y, Kaufmann PG, Mus‑ 2. Mathers CD, Loncar D. Projections of global mortality and burden of solino ME, Hsu LL, Addou E, Engelgau MM, et al. Decline in cardiovascular disease from 2002 to 2030. PLoS Med. 2006;3(11):e442. mortality. Possible Causes Implic. 2017;120(2):366–80. 3. Ogden CL, Carroll MD, Lawman HG, Fryar CD, Kruszon‑Moran D, Kit BK, 25. Gupta A, Wang Y, Spertus JA, Geda M, Lorenze N, Nkonde‑Price C, Flegal KM. Trends in obesity prevalence among children and adolescents D’Onofrio G, Lichtman JH, Krumholz HM. Trends in acute myocardial in the United States, 1988‑1994 through 2013‑2014. J Am Med Assoc. infarction in young patients and differences by sex and race, 2001 to 2016;315(21):2292–9. 2010. J Am Coll Cardiol. 2014;64(4):337–45. 4. Bender R, Jockel KH, Trautner C, Spraul M, Berger M. Eec ff t of age on 26. Ford ES, Capewell S. Coronary heart disease mortality among young excess mortality in obesity. J Am Med Assoc. 1999;281(16):1498–504. adults in the U.S. from 1980 through 2002: concealed leveling of mortal‑ 5. Baker JL, Olsen LW, Sorensen TI. Childhood body‑mass index and ity rates. J Am Coll Cardiol. 2007;2007(50):2128–32. the risk of coronary heart disease in adulthood. N Engl J Med. 27. Bairey Merz CN, Andersen H, Sprague E, Burns A, Keida M, Walsh MN, 2007;357(23):2329–37. Greenberger P, Campbell S, Pollin I, McCullough C, et al. Knowledge, 6. Tirosh A, Shai I, Afek A, Dubnov‑Raz G, Ayalon N, Gordon B, Derazne E, attitudes, and beliefs regarding cardiovascular disease in women: the Tzur D, Shamis A, Vinker S, et al. Adolescent BMI trajectory and risk of women’s heart alliance. J Am Coll Cardiol. 2017;70(2):123–32. diabetes versus coronary disease. N Engl J Med. 2011;364(14):1315–25. 28. Mongraw‑ Chaffin ML, Peters SAE, Huxley RR, Woodward M. The sex ‑ 7. Falkstedt D, Hemmingsson T, Rasmussen F, Lundberg I. Body mass index specific association between BMI and coronary heart disease: a system‑ in late adolescence and its association with coronary heart disease atic review and meta‑analysis of 95 cohorts with 1.2 million participants. and stroke in middle age among Swedish men. Int J Obes (Lond). Lancet Diab Endocrinol. 2015;3(6):437–49. 2007;31(5):777–83. 29. Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW. Body‑mass index 8. Ohlsson C, Bygdell M, Sonden A, Rosengren A, Kindblom JM. Association and mortality in a prospective cohort of US adults. New Engl J Med. between excessive BMI increase during puberty and risk of cardiovascu‑ 1999;341:1097–105. lar mortality in adult men: a population‑based cohort study. Lancet Diab 30. Katzmarzyk PT, Craig CL, Bouchard C. Original article underweight, over‑ Endocrinol. 2016;4(12):1017–24. weight and obesity: relationships with mortality in the 13‑ year follow‑up 9. Must A, Jacques P, Dallal G, Bajema C, Dietz W. Long‑term morbidity and of the Canada Fitness Survey. J Clin Epidemiol. 2001;54(9):916–20. mortality of overweight adolescents. A follow‑up of the Harvard growth 31. Seidell JC, Verschuren WM, van Leer EM, Kromhout D. Overweight, under‑ study of 1922 to 1935. N Engl J Med. 1992;327(19):1350–5. weight, and mortality: a prospective study of 48,287 men and women. 10. Strand BH, Kuh D, Shah I, Guralnik J, Hardy R. Childhood, adolescent Arch Intern Med. 1996;156:958–63. and early adult body mass index in relation to adult mortality: results 32. Gunnell DJ, Frankel SJ, Nanchahal K, Peters TJ, Davey Smith G. Childhood from the British 1946 birth cohort. J Epidemiol Community Health. obesity and adult cardiovascular mortality: a 57‑ y follow‑up study based 2012;66(3):225–32. on the Boyd Orr cohort. Am J Clin Nutr. 1998;67:1111–8. 11. Wandell PE, Carlsson AC, Theobald H. The association between BMI value 33. Bjorge T, Engeland A, Tverdal A, Smith GD. Body mass index in adoles‑ and long‑term mortality. Int J Obes (Lond). 2009;33(5):577–82. cence in relation to cause‑specific mortality: a follow‑up of 230,000 12. Roh L, Braun J, Chiolero A, Bopp M, Rohrmann S, Faeh D. Mortality risk Norwegian adolescents. Am J Epidemiol. 2008;168:30–7. associated with underweight: a census‑linked cohort of 31,578 individu‑ 34. Manson JE, Stampfer MJ, Hennekens CH, Willett WC. Body weight and als with up to 32 years of follow‑up. BMC Public Health. 2014;14(1):371. longevity. A reassessment. JAMA, J Am Med Assoc. 1987;257(3):353–8. 13. Twig G, Yaniv G, Levine H, Leiba A, Goldberger N, Derazne E, Ben‑Ami 35. Berrington de Gonzalez A, Hartge P, Cerhan JR, Flint AJ, Hannan L, MacIn‑ Shor D, Tzur D, Afek A, Shamiss A, et al. Body‑mass index in 2.3 million nis RJ, Moore SC, Tobias GS, Anton‑ Culver H, Freeman LB, et al. Body‑mass adolescents and cardiovascular death in adulthood. N Engl J Med. index and mortality among 1.46 million white adults. N Engl J Med. 2016;374(25):2430–40. 2010;363(23):2211–19. 14. Twig G, Tirosh A, Leiba A, Levine H, Ben‑Ami Shor D, Derazne E, Haklai Z, 36. Aune D, Sen A, Prasad M, Norat T, Janszky I, Tonstad S, Romundstad P, Goldberger N, Kasher‑Meron M, Yifrach D, et al. BMI at age 17 years and Vatten LJ. BMI and all cause mortality: systematic review and non‑linear diabetes mortality in midlife: a nationwide cohort of 2.3 million adoles‑ dose‑response meta‑analysis of 230 cohort studies with 3.74 million cents. Diab Care. 2016;39(11):1996–2003. deaths among 30.3 million participants. BMJ. 2016;353:i2156. 15. Collaboration ERF. Diabetes mellitus, fasting glucose, and risk of cause‑ 37. Collaboration PS. Body‑mass index and cause ‑specific mortality in 900 specific death. N Engl J Med. 2011;2011(364):829–41. 000 adults: collaborative analyses of 57 prospective studies. Lancet. 16. Kaati G, Bygren LO, Edvinsson S. Cardiovascular and diabetes mortality 2009;373:1083–96. determined by nutrition during parents’ and grandparents’ slow growth period. Eur J Hum Genet. 2002;10:682. Furer et al. Cardiovasc Diabetol (2018) 17:80 Page 10 of 10 38. Calle E, Thun M, Petrelli J, Rodriguez C, Heath C Jr. Body‑mass index 49. Pontiroli AE, Zakaria AS, Mantegazza E, et al. Long‑term mortality and and mortality in a prospective cohort of US adults. N Engl J Med. incidence of cardiovascular diseases and type 2 diabetes in diabetic and 1999;341:1097–105. nondiabetic obese patients undergoing gastric banding: a controlled 39. Song X, Tabák AG, Zethelius B, Yudkin JS, Söderberg S, Laatikainen study. Cardiovasc Diabetol. 2016;15:39. T, Stehouwer CD, Dankner R, Jousilahti P, Onat A. Obesity attenuates 50. Bjerregaard LG, Jensen BW, Ängquist L, Osler M, Sørensen TIA, Baker JL. gender differences in cardiovascular mortality. Cardiovasc Diabetol. Change in overweight from childhood to early adulthood and risk of 2014;13(1):144. type 2 diabetes. N Engl J Med. 2018;378(14):1302–a12. 40. Collaboration GBM. Body‑mass index and all‑ cause mortality: individual‑ 51. Peter RS, Keller F, Klenk J, Concin H, Nagel G. Body mass trajectories, participant‑ data meta‑analysis of 239 prospective studies in four conti‑ diabetes mellitus, and mortality in a large cohort of Austrian adults. nents. Lancet. 2016;388:776–86. Medicine. 2016;95:e5608. 41. Krebs NF, Himes JH, Jacobson D, Nicklas TA, Guilday P, Styne D. Assess‑ 52. Cohen‑Manheim I, Doniger GM, Sinnreich R, Simon ES, Murad H, Pinchas‑ ment of child and adolescent overweight and obesity. Pediatrics. Mizrachi R, Kark JD. Body mass index, height and socioeconomic position 2007;120(Suppl 4):S193–228. in adolescence, their trajectories into adulthood, and cognitive function 42. Peter RS, Mayer B, Concin H, Nagel G. The effect of age on the shape of in midlife. Alzheimer’s Dis. 2017;55(3):1207–21. the BMI‑mortality relation and BMI associated with minimum all‑ cause 53. Zucker I, Shohat T, Dankner R, Chodick G. New onset diabetes in adult‑ mortality in a large Austrian cohort. Int J Obes (Lond). 2015;39:530–4. hood is associated with a substantial risk for mortality at all ages: a 43. Lopez‑ Otin C, Galluzzi L, Freije JMP, Madeo F, Kroemer G. Metabolic con‑ population based historical cohort study with a decade‑long follow‑up. trol of longevity. Cell. 2016;166:802–21. Cardiovasc Diabetol. 2017;16:105. 44. Madeo F, Zimmermann A, Maiuri MC, Kroemer G. Essential role for 54. Kark JD, Laor A. Cigarette smoking and educational level among young autophagy in life span extension. J Clin Investig. 2015;125(1):85–93. Israelis upon release from military service in 1988—a public health chal‑ 45. Kark JD, Goldberger N, Kimura M, Sinnreich R, Aviv A. Energy intake lenge. Isr J Med Sci. 1992;28(1):33–7. and leukocyte telomere length in young adults. Am J Clin Nutr. 55. Pischon T, Boeing H, Hoffmann K, Bergmann M, Schulze MB, Overvad K, 2012;95:479–87. van der Schouw Y T, Spencer E, Moons KG, Tjonneland A, et al. General 46. Willcox BJ, Willcox DC, Suzuki M. Demographic, phenotypic, and genetic and abdominal adiposity and risk of death in Europe. N Engl J Med. characteristics of centenarians in Okinawa and Japan: part 1‑ centenarians 2008;359(20):2105–20. in Okinawa. Mech Ageing Dev. 2017;165(Pt B):75–9. 56. Force USPST, Grossman DC, Bibbins‑Domingo K, Curry SJ, Barry MJ, David‑ 47. Pinhas‑Hamiel O, Reichman B, Shina A, Derazne E, Tzur D, Yifrach D, Wiser son KW, Doubeni CA, Epling JW, Jr., Kemper AR, Krist AH, et al. Screening I, Afek A, Shamis A, Tirosh A, et al. Sex differences in the impact of thin‑ for obesity in children and adolescents: US preventive services task force ness, overweight, obesity, and parental height on adolescent height. J recommendation statement. J Am Med Assoc. 2017;317(23):2417–26. Adolesc Health. 2017;61(2):233–9. 57. Hogstrom G, Nordstrom A, Nordstrom P. Aerobic fitness in late ado ‑ 48. Ostan R, Monti D, Gueresi P, Bussolotto M, Franceschi C, Baggio G. Gender, lescence and the risk of early death: a prospective cohort study of 1.3 aging and longevity in humans: an update of an intriguing/neglected million Swedish men. Int J Epidemiol. 2015;2015;45:1159–68. scenario paving the way to a gender‑specific medicine. Clin Sci (London, England: 1979). 2016;130(19):1711–25. Ready to submit your research ? Choose BMC and benefit from: fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions

Journal

Cardiovascular DiabetologySpringer Journals

Published: Jun 5, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off