Eating patterns of Australian adults: associations with blood pressure and hypertension prevalence

Eating patterns of Australian adults: associations with blood pressure and hypertension prevalence Purpose Eating patterns have been linked to obesity, an established risk factor for hypertension; however, their contribution to hypertension is poorly understood. This study aimed to examine associations of frequency of meals, snacks and all eating occasions (EO), and temporal eating patterns, with blood pressure (BP) and hypertension. Methods Dietary data collected via two 24-h recalls during the 2011–2012 Australian National Nutrition and Physical Activ- ity Survey (n = 4482 adults, ≥ 19 years) were analysed. Frequencies of EO, meals, and snacks were calculated. Temporal eating patterns were determined using latent class analysis. Multivariate regression models assessed associations of eating patterns with systolic BP (SBP), diastolic BP (DBP), and hypertension prevalence. Results Among men, a higher snack frequency was inversely associated with DBP [β = − 0.59, 95% confidence interval (CI) (− 1.12, − 0.07)] and hypertension [odds ratio (OR) 0.86, 95% CI (0.75, 0.98)] after adjustment for covariates and BMI. However, these associations disappeared after additional adjustment for total energy intake and overall diet quality. Among women, a temporal eating pattern characterized by a later “lunch” meal was associated with SBP [β = 2.45, 95% CI (0.05, 4.84)], DBP [β = 1.69, 95% CI (0.25, 3.13)], and hypertension [OR = 1.49, 95% CI (1.00, 2.22)], when compared to a “conventional” eating pattern. Conclusions In this study, an inverse association found between snack frequency and BP among men disappeared after adjustment for dietary factors and a “later lunch” pattern was associated with higher BP in women. Future research is needed to understand the relationship and potential mechanistic pathways between eating patterns and BP. Keywords Blood pressure · Circadian rhythms · Eating frequency · Eating patterns · Meals · Meal timing · Snacks Introduction foods has been identified as an important strategy for the prevention of cardiovascular disease and hypertension [2, Poor dietary habits, obesity, and high systolic blood pres- 3]. However, the contribution of eating patterns to hyperten- sure (SBP) are three major risk factors contributing to sion is poorly understood [4]. Eating patterns, including the the global burden of disease [1]. As these risk factors are frequency and temporality of eating occasions [(EO), e.g., strongly interrelated, it has been postulated that diet is a meals and snacks], may represent a unique cardiometabolic key driver of global increases in BMI and subsequent SBP risk factor [5]. This is because the metabolic and physiologi- [1]. The adoption of a dietary pattern rich in plant-based cal functions in responses to food are also, in part, regulated foods but low in red meat and discretionary (e.g., non-core) by circadian clock systems. Thus, variations in the temporal distribution of food consumption may have important health implications [6]. Electronic supplementary material The online version of this Epidemiologic evidence [7–11] suggests that skipping article (https ://doi.org/10.1007/s0039 4-018-1741-y) contains supplementary material, which is available to authorized users. breakfast and irregular meal habits are associated with poorer cardiometabolic health in adults, including incident * Rebecca M. Leech hypertension [7]. However, evidence for associations with rebecca.leech@deakin.edu.au EO frequency is less consistent [4]. In three studies, a higher Institute for Physical Activity and Nutrition (IPAN), School EO frequency (e.g., ≥ 5 EO per day) was associated with of Exercise and Nutrition Sciences, Deakin University, 75 lower SBP [12, 13] and diastolic blood pressure (DBP) [13, Pigdons Rd, Waurn Ponds, Geelong 3216, Australia Vol.:(0123456789) 1 3 European Journal of Nutrition 14]. In support of these findings, a review of short-term approval to conduct the survey components of the NNPAS experimental studies found that six or more meals per day, [25]. compared to one meal, resulted in favourable decreases in cardiometabolic risk markers [15]. However, most of the Blood pressure reviewed studies did not control for differences in dietary intake and other population-based studies have found no All BP measurements were voluntary, excluded pregnant evidence for a relation of EO frequency with blood pressure women, and were taken by trained ABS staff during the (BP) [16, 17]. Furthermore, research suggests that meals and home visit. Participants were asked to sit comfortably and snacks are differentially related to dietary intakes [18, 19] relax their left arm. Two SBP and DBP measurements were and obesity [19, 20], yet studies examining meals and snacks taken on the left arm, with the palm facing upwards, using an as separate EO in relation to hypertension are rare [12]. automated BP monitor. The second reading was used, except The temporal distribution of EO in relation to BP has where there was greater than 10 mmHg difference between rarely been examined [21, 22]. In the UK Cohort Study, the first and second readings, in which case, a third reading Almoosawi et al. [21] found a positive association for higher was taken and the second and third readings averaged. If total energy intake in the evening with 10-year increases in all three readings differed from each other by > 20 mmHg, SBP and DBP. In a study of Spanish adult volunteers, con- then these readings were considered invalid. Participants sumption of an afternoon meal was associated with lower were classified as hypertensive (≥ 140/90 mmHg) or non- SBP and DBP [22]. However, these studies focussed on tim- hypertensive (< 140/90 mmHg) [26]. No data about use of ing of EO/energy intake at isolated periods of the day and BP medications were collected; however, participants’ self- did not simultaneously capture EO at other times of the day reported whether they had current or previous hypertensive [4]. Data-driven statistical techniques (e.g., cluster and latent disease. class analysis) have recently been identified as useful tools for capturing temporal eating patterns across the day, but Dietary assessment have only been used to examine relations with diet quality and obesity [23, 24]. Therefore, the aim of this study was Dietary data were collected over two 24-h recalls, conducted to examine the associations of frequency of meals, snacks approximately 9 days apart, on a different day of the week. and all EO, and temporal eating patterns, determined using During each dietary recall, based on the validated USDA a latent class analysis approach, with BP and hypertension, automated multiple 5-pass method, participants were asked in Australian men and women. Based on the findings of our to identify the type of EO (e.g., breakfast, lunch, dinner, or earlier studies that found a positive relation for snacking snack) and the time when each EO commenced. The Aus- frequency and a “grazing” temporal eating pattern with tralian Supplement and Nutrient Database 2011–2013 was measures of adiposity, we hypothesised that that these same used to determine energy and nutrient intakes from all foods eating patterns would be related to (obesity-associated) BP and beverages and dietary information was averaged across and hypertension [20, 23]. the two recall days to obtain mean estimates of eating pat- terns, total energy intakes, and food intakes. Subjects and methods Frequencies of all EO, meals, and snacks Sample and study design The methods used to calculate mean total frequencies of all EO, meals, and snacks have been described previously This study is a secondary analysis of data drawn from the [18, 27]. Briefly, an EO constituted any eating event that nationally representative, cross-sectional 2011–2012 Aus- provided a minimum energy content of 210 kJ (50 kcal) and tralian National Nutrition and Physical Activity Survey was separated in time from the surrounding EO by 15 min. (NNPAS 2011–12) and was registered at anzctr.org.au as This approach was informed by the previous research and ACTRN12617001029381. The NNPAS, conducted between current recommendations for defining EO in eating patterns May 2011 and June 2012, was administered by the Aus- research [4, 27]. EO were then classie fi d as meals and snacks tralian Bureau of Statistics (ABS); full details of the study according to participants’ self-report of EO. Meals included design and methods have been described elsewhere [25]. EO reported as breakfast, brunch, lunch, dinner, and supper, Briefly, 12,153 participants aged 2 or more years (includ- whereas snacks included EO reported as morning/afternoon ing 9338 adults, aged ≥ 19 years; 77% response rate) were tea and beverage break were classified as snack EO. Based selected using a multistage, probability sampling design of on the sample distribution, frequencies of all EO, meals and private dwellings (Fig. 1). The Australian Government Cen- snacks were divided into categories of 1–3, 4–5, or ≥ 6 EO, sus and Statistics Act 1905 provided the ABS with ethics 1–2, 3, or > 3 meals, and 0–1, 2–3, or > 3 snacks. 1 3 European Journal of Nutrition Fig. 1 Flowchart of included NNPAS 2011-12 participants from the 2011– Total survey participants 2012 Australian National Nutrition and Physical Activity n = 12153 Survey (NNPAS 2011–12) Adults aged ≥19 years n = 9341 Completed two 24 hour dietary recalls n = 6053 (65%) Not pregnant, breastfeeding or undertaking shift work in past 4 weeks n = 5366 Missing data on eating occasion type or time of eating occasion consumption n = 116 excluded Reported no energy intake during a 24 hour recall n = 8 excluded Missing data for blood pressure n = 578 excluded Missing data for BMI or covariates n = 182 excluded Final analytic sample n = 4482 characterized by frequent but less distinct meal occasions Temporal eating patterns and a higher probability of having an EO after 8 p.m. These temporal eating patterns have been associated with socio- Temporal eating patterns were determined using latent class demographic and eating pattern characteristics [28], diet analysis, as reported in detail previously [23, 28]. Briefly, quality, and, among women, adiposity outcomes [23]. three latent classes of temporal eating patterns were identi- fied for men and women based on frequency and timing of Covariates EO across the day and labelled according to their defining characteristics. The optimal number of classes was selected The following socio-demographic, health behaviours, and based on model fit indices, likelihood ratio tests comparing −1 anthropometric variables, collected during the household k with k class models, and pattern interpretability [29]. survey, were considered as potential covariates due to their The first pattern was labelled as “conventional” due to the previous reported relation with hypertension risk [1, 26]. probability of participants having three EO at “conven- tional” mealtimes in Australia (e.g., between 7 and 8 a.m., Socio‑demographics 12–1 p.m., and 6–7 p.m.). The second pattern was distin- guished by a > 0.9 probability of a “lunch” meal approxi- Education level was categorised as: low (completed some mately 1 h later than the “conventional pattern” and labelled high-school or less), medium (completed high-school or the “later lunch” pattern. The probability of a “dinner” meal completed some high-school and/or certificate/diploma) 1 h later than the conventional pattern was also higher (e.g., or high (having a tertiary qualification). Country of birth > 0.6) in participants with a “later lunch” pattern. The third was categorised by the ABS as: Australia, predominantly pattern was labelled the “Grazing” pattern, because it was 1 3 European Journal of Nutrition English-speaking countries (other than Australia) and all Statistics other countries. All statistical analyses were stratified by sex and used Stata statistical software, Version 14.2 (Stata Inc., College Sta- Health behaviours tion, TX, USA). Point estimates and standard errors were determined by applying person and replicate weights that Smoking status was self-reported and categorised as cur- accounted for the probability of participant selection and rent smoker, ex-smoker, and never smoked. Participants the clustered survey design, respectively. Descriptive sta- were categorised as meeting or not meeting current Aus- tistics for sample characteristics are presented as weighted tralian physical activity guidelines (150 min and 5 sessions), means (95% CI) or weighted percentages. After examining based on self-reported frequency and duration of walking the distribution of the data, the following variables were for recreation or transport and moderate or vigorous leisure- log-transformed to improve normality: BMI, daily total sed- time physical activity [30, 31]. Self-reported information entary time, and total energy intake. Weighted geometric on sleep duration the night before the survey and how much means (95% CI) were used for all log-transformed variables. time participants spent sitting or lying down at work, during The F test (for continuous data) and adjusted Pearson χ transport and leisure activities in the past week, were used to test (for categorical data) were used to determine sex-spe- calculate (per day) total minutes spent sleeping per day and cific differences in sample characteristics by hypertension in sedentary behaviour, respectively. Participants reported status. Multiple linear regression (for continuous outcomes) whether they were currently on a weight-loss diet for health and logistic regression (for binary outcomes) were used to reasons (yes/no). Average daily total energy intake and diet test for associations of frequencies of all EO, meal and quality scores were calculated from the 2 days of recall. The snacks (continuous), and temporal eating patterns, with SBP established food-based Dietary Guidelines Index (DGI) was and DBP (continuous) and hypertension prevalence (binary). used as a measure of overall diet quality [32–34]. The DGI Four models were tested: model 1 was an unadjusted model; assesses compliance with recommendations outlined in the model 2 adjusted for age (years, continuous), education level Australian Dietary Guidelines, and is the sum of 13 com- (low, medium or high), country of birth (Australia, other pre- ponents (score range of 0–130), each corresponding to an dominantly English-speaking countries, all other countries), Australian Dietary guideline and scored proportionally out smoking status (never, former or current), daily, meeting of 10. The components include meeting recommendations physical activity guidelines (yes/no), daily sedentary time for food variety, intakes of fruits, vegetables (including leg- (min; continuous), sleep duration (h, continuous), dieting for umes), grain foods, dairy and alternatives, meat and alterna- health reasons (yes/no); model 3 further adjusted for BMI, tives, unsaturated fat, fluids, discretionary foods, saturated and model 4 further adjusted for total energy intake and DGI fat, salt, added sugar, and alcohol. Higher scores indicate a scores (both continuous). In light of the previous research better diet quality. Measurement of height (cm) and weight that reported a positive association among participants with (kg) were taken to one decimal point by trained ABS staff the highest EO frequencies (i.e., > 5 EO) [13], in the pre- using a portable stadiometer and digital scales. BMI (weight sent study, any observed statistically significant (P < 0.05) [kg]/height [m] ) was calculated. adjusted association between the continuous measures of frequencies of EO, meals, or snacks, and the outcome vari- ables were further explored by examining associations for Analytic sample eating pattern frequency categories (e.g., 1–3 [reference], 4–5 or ≥ 6 EO; 1–2 [reference], 3 or > 3 meals and 0–1 [ref- The analytic sample included the 65% of adult participants erence], and 2–3 or > 3 snacks). Finally, the effect of energy who completed both dietary recalls (n = 6053; Fig. 1). Par- misreporting, defined as the ratio of total energy intake to ticipants were eligible for this analysis if they were not preg- total energy expenditure was considered [35]; however, its nant, breastfeeding, or undertaking shift-work in the past inclusion did not improve its predictive power when BMI 4 weeks (n = 5366) and were excluded if they reported no was already in the model. A previous study has also shown energy intake during either dietary recall (n = 8 excluded) that energy misreporting bias can be statistically corrected or did not report the time at which an EO commenced or using predictors of energy misreporting (i.e., dieting behav- the type of EO (n = 116 excluded). Of the remaining 5242 iours and BMI) [36]. participants, 578 (11%) had missing data for BP and a fur- ther 182 (3.9%) were missing data for BMI and covariates: Sensitivity analysis BMI (n = 149), physical activity (n = 28), and sedentary time (n = 5). The final analytic sample was 2099 men and 2383 As data on BP medications use were not collected in the women. NNPAS [25], a sensitivity analysis was conducted that 1 3 European Journal of Nutrition included only participants who self-reported no current or men, a significant inverse association was found for fre- previous hypertensive disease (n = 1612 men and n = 1853 quency of all EO and snacks but not meals. women). After further analysis that examined these associations by categories of snack and EO frequency, the inverse adjusted associations (model 3) were only observed among men who reported > 3 snacks [OR 0.52, 95% CI (0.30, 0.81); Results P = 0.006] and ≥ 6 EO [OR 0.54, 95% CI (0.34, 0.84); P = 0.008], compared to those who reporting < 2 snacks and Table 1 presents the characteristics of men and women par- ≤ 3 EO. However, again, these associations attenuated after ticipants in the NNPAS 2011–12 by hypertension status. further adjustment for total energy intake and DGI scores, Of the participants, 24% of men and 20% of women were and no significant associations were found among women. classified as having hypertension. Among both sexes, there Compared to a “conventional” temporal eating pattern, a were significant differences by hypertension status for age, “later lunch temporal” eating pattern was also associated education level, meeting physical activity guidelines, BMI, with hypertension prevalence among women, but only after and self-report status of previous or current hypertensive adjustment for covariates and BMI scores. disease (P < 0.05). Differences were also found for total Results of the sensitivity analyses which included only daily energy intake, DGI scores, and frequency of all EO men and women with no self-reported current or previ- and snacks between non-hypertensive and hypertensive men ous hypertensive disease, showed similar null associations (P < 0.05). For the snack frequency categories, a higher and of frequency of EO, snacks, and meals with SBP, DBP or lower proportion of hypertensive men reported having fewer hypertension prevalence, after adjustment for covariates, than two snacks and more than three snacks per day, respec- BMI, and dietary intakes (Supplementary Table 1). How- tively, compared to non-hypertensives (P < 0.05). No signifi- ever, among women, the finding of a positive association cant differences were found among women for the dietary between a “later lunch” pattern and DBP (but not SBP or or eating pattern variables according to hypertension status. hypertension prevalence) persisted after exclusion of those The sex-specific associations of frequency of all EO, with no self-reported current/previous hypertensive disease. meals, and snacks (continuous) with SBP and DBP are pre- sented in Table 2. In the basic model, a statistically signifi- cant positive association was found between meal frequency Discussion and SBP among women which disappeared after adjustment for the covariates in model 2. Among men, inverse associa- To our knowledge, this is one of the first studies among tions were found between frequency of all EO and snacks adults to examine associations of meal and snack frequency and DBP, and after adjustment for covariates (model 2) and and temporal eating patterns, based on the timing and fre- BMI (model 3). quency of EO across the day, with BP and hypertension After further analysis that examined these associations by prevalence [12]. Among men, frequency of all EO and categories of snack and EO frequency, the inverse associa- snacks was inversely associated with DBP and lower odds tions were observed among men who reported > 3 snacks of hypertension prevalence, but these associations disap- [DBP: β = − 2.23, 95% CI (− 4.14, − 0.32); P = 0.023] peared after adjustment for overall diet quality scores and and ≥ 6 EO [DBP: β = − 2.32, 95% CI (− 4.45, − 0.20); total energy intakes. Among women, a “later lunch” pattern, P = 0.032], compared to those who reporting < 2 snacks and identified using latent class analysis in our earlier study [28], ≤ 3 EO. However, these associations with DBP for men were and characterized by a later lunch EO (e.g., between 1 and attenuated after further adjustment for the total energy intake 2 p.m.) was associated with higher SBP, DBP, and hyper- and DGI scores: β = − 1.56, 95% CI (− 3.59, 0.47); P = 0.13 tension prevalence. However, only associations with DBP and β = − 1.37, 95% CI (− 3.62, 0.89); P = 0.23, respectively. persisted after exclusion of persons with the self-reported Results of the regression analyses showed no associa- previous/current hypertension. tions between latent classes of temporal eating patterns and Only a few studies have examined the relationship SBP or DBP among men (Table 2). Among women, a “later between EO frequency and BP among adults [12, 13, 16, lunch” temporal eating pattern was positively associated 17], with conflicting findings. However, it is difficult to com- with SBP and DBP, when compared to a “conventional” pare the results of these studies, because they define EO pattern, after adjustment for covariates and BMI, and after using different approaches. For example, EO have mostly further adjustment for dietary intakes. been self-reported by participants in response to a single Associations of frequency of all EO, meals, and snacks survey question where an EO is not further defined [12, (continuous) and latent classes of temporal eating patterns 17]. Whereas in another study, in a small sample of healthy with hypertension prevalence are shown in Table 3. Among volunteers (n = 115), EO was defined as any eating event 1 3 European Journal of Nutrition Table 1 Characteristics of men and women in the NNPAS by hypertension status Men (n = 2099) Women (n = 2383) Non-hypertensive Hypertensive (n = 561) P value Non-hypertensive Hypertensive (n = 500) P value (n = 1538) (n = 1883) Socio-demographics  Age (years) 43.1 (42.4, 44.8) 56.2 (54.2, 58.2) < 0.0001 44.7 (44.0, 45.5) 58.4 (56.8, 60.0) < 0.0001  Education level (%) < 0.05 < 0.0001   Low 18 25 25 43   Medium 53 54 44 33   High 29 21 32 24  Country of birth (%) 0.95 0.12   Australia 69 69 70 61   Predominantly 13 13 11 15 English-speaking countries   All other countries 18 18 19 24 Health behaviours or characteristics  Smoking status 0.14 0.56   Never 47 39 59 55   Former 35 40 27 31   Current 18 20 14 14  Meets physical activ- 48 41 0.04 45 35 < 0.01 ity guidelines (%)  Daily sedentary time 306.7 (289.8, 324.6) 297.2 (273.0, 323.4) 0.54 260.3 (247.9, 273.2) 263.1 (237.0, 292.1) 0.86 (min)  Currently on a diet 12 8 0.06 17 15 0.43 for health reasons (%)  Sleep duration (h) 7.9 (7.8, 8.0) 7.9 (7.7, 8.0) 0.81 8.0 (7.9, 8.1) 8.0 (7.7, 8.2) 0.52  Total energy intake 9498 (9304, 9696) 8463 (8170, 8766) < 0.0001 7087 (6928, 7250) 6890 (6587, 7206) 0.31 (kJ)  Dietary Guidelines 80.1 (79.0, 81.2) 77.6 (75.7, 79.6) < 0.05 80.7 (79.5, 82.0) 81.8 (80.1, 83.5) 0.31 Index (score)  BMI (score) 26.8 (26.5, 27.2) 28.9 (28.2, 29.5) < 0.0001 25.9 (25.5, 26.3) 28.7 (27.9, 29.6) < 0.0001  Systolic blood pres- 118.7 (117.9, 119.5) 148.3 (145.6, 151.0) < 0.0001 112.5 (111.7, 113.1) 148.8 (146.9, 150.8) < 0.0001 sure (mmHg)  Diastolic blood pres- 73.4 (72.8, 74.0) 88.7 (87.5, 89.8) < 0.0001 72.9 (72.2, 73.6) 89.0 (87.9, 90.2) < 0.0001 sure (mmHg)  No current or previ- 85 63 < 0.0001 86 56 < 0.0001 ous hypertensive disease (%) Eating patterns  Eating occasion 4.9 (4.8, 5.0) 4.7 (4.5, 4.8) < 0.01 4.8 (4.7, 4.9) 4.7 (4.6, 4.9) 0.37 frequency  Meal frequency 2.9 (2.8, 2.9) 2.9 (2.8, 2.9) 0.79 2.9 (2.9, 3.0) 3.0 (2.9, 3.0) 0.55  Snack frequency 2.1 (1.98, 2.2) 1.8 (1.7, 2.0) < 0.01 1.9 (1.9, 2.0) 1.8 (1.7, 2.0) 0.19  Categories of eating 0.07 0.54 occasion frequency (%)   1–3 19 24 18 20   4–5 57 60 61 62   ≥ 6 23 17 21 18  Categories of meal 0.27 0.89 frequency (%) 1 3 European Journal of Nutrition Table 1 (continued) Men (n = 2099) Women (n = 2383) Non-hypertensive Hypertensive (n = 561) P value Non-hypertensive Hypertensive (n = 500) P value (n = 1538) (n = 1883)   < 3 31 32 26 25   3 53 48 54 54   > 3 15 20 20 21  Categories of snack < 0.05 0.59 frequency (%)   < 2 45 51 48 51   2–3 37 38 39 38   > 3 18 11 13 11  Latent classes of 0.09 0.17 temporal eating pat- terns (%)   Conventional 40 47 41 37   Later lunch 36 30 32 39   Grazing 24 23 27 24 Values are weighted means (95% confidence intervals) or weighted percentages. Significant sex-specific differences by hypertension status assessed using an F test for continuous variables or design-adjusted Pearson χ test Whether met physical activity guidelines of 150 min and 5 sessions/week Values are geometric means (95% CI) DGI represents a total diet quality score (score range 0–130) with higher scores indicating better overall diet quality (including kilojoule-free events) separated in time by 15 min associated with better diet quality scores for intakes of fruits [13]. A higher EO frequency has been associated with lower and dairy products, two food groups recommended as part of hypertension prevalence [12] and incidence [13], whereas the Dietary Approaches to Stop Hypertension (DASH) diet other studies have found no associations with BP [16, 17]. [3]. However, in the same study, snack frequency was also In the present study, EO frequency was inversely associ- associated with poorer scores for intakes of discretionary ated with DBP and hypertension prevalence among men, foods and added sugars among men [18]. Future research but these associations attenuated after further adjustment that examines the role of diet quality on the relation between for total energy intakes and overall diet quality. EO frequency and hypertension is warranted. Studies examining the separate effects of meal and snack Epidemiological evidence suggests a positive association frequency on BP outcomes are rare [12]. Compared to par- between evening energy intakes or the later timing of an EO ticipants who reported no snacks, Kim et al. [12] found that and obesity [38], but studies examining temporal patterns of a snack frequency of three per day was associated with lower eating in relation to BP are rare. In one study, higher energy odds of hypertension, but associations attenuated (e.g., 95% intake at breakfast was associated with lower hypertension CI included one) after adjustment for adiposity measures. In prevalence but not 10-year incidence, and higher energy the present study, snack frequency (specifically > 3 snacks) intake in the evening was associated with higher hyper- was inversely associated with BP outcomes among men, but tension incidence, which remained borderline significant again associations attenuated after adjustment for overall after adjustment for baseline BMI [21]. Another study, in diet quality and total energy intakes. Notably, with respect to the same cohort, found no association between time of day adjustment of dietary factors, the previous studies on eating macronutrient intakes and BP [39]. Keller et al. found that patterns and BP have only adjusted for either total energy the consumption of an afternoon meal, but not other conven- intakes [13] or energy and nutrient intakes [12, 16, 17], and tional Spanish meals, was modestly associated with lower not a measure of overall diet quality based on food intakes. SBP and DBP, even after adjustment for dietary intake and The findings from the present study suggest that diet qual- waist circumference [22]. In the present study, a temporal ity may be an important factor in the relation between EO eating pattern characterized by having a later “lunch” meal frequency and BP, and is supported by the previous studies was associated with SBP, DBP, and hypertension prevalence that have shown a beneficial effect of healthful dietary pat- among women, after adjustment for potential covariates, terns for the prevention of hypertension [37]. In a previous BMI, and diet quality. However, rather than examining the study of NNPAS participants. [18], snack frequency was timing of a single meal or energy intake across stratified 1 3 European Journal of Nutrition Table 2 Associations of eating patterns with systolic and diastolic blood pressure in Australian men and women Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) a b c d a b c d Model 1Model 2Model 3Model 4Model 1Model 2Model 3Model 4 Men (n = 2099)  Eating fre- quency   Eating − 0.49 − 0.65 − 0.47 − 0.18 − 0.80 − 0.88 − 0.63 − 0.39 (− 0.98, occasion (− 1.22, (− 1.41, (− 1.24, (− 1.11, (− 1.36, (− 1.45, (− 1.17, 0.20) fre- 0.25) 0.11) 0.30) 0.74) − 0.23)** − 0.31)** − 0.09)* quency   Meal fre- 0.47 (− 1.96, − 1.29 − 0.86 0.13 (− 2.35, − 1.04 − 1.46 − 0.88 − 0.46 (− 2.18, quency 2.91) (− 3.41, (− 3.13, 2.62) (− 2.43, (− 2.99, (− 2.48, 1.26) 0.84) 1.40) 0.35) 0.08) 0.71) − 0.77 − 0.59 − 0.41 (− 0.99,   Snack fre- − 0.69 − 0.59 − 0.46 − 0.32 − 0.73 (− 1.35, (− 1.12, 0.18) quency (− 1.46, (− 1.37, (− 1.22, (− 1.24, (− 1.33, − 0.13)* − 0.20)** − 0.07)* 0.09) 0.19) 0.30) 0.59)  Temporal eating pat- terns   Conven- – – – – – – – – tional (refer- ence, n = 941)   Later − 2.44 − 1.13 − 0.80 − 0.68 − 0.97 − 0.63 − 0.19 − 0.10 (− 1.65, lunch (− 4.93, (− 3.42, (− 2.94, (− 2.82, (− 2.66, (− 2.37, (− 1.74, 1.45) (n = 702) 0.05) 1.16) 1.33) 1.46) 0.71) 1.11) 1.36)   Grazing − 2.72 0.63 (− 2.19, 0.77 (− 1.89, 0.75 (− 1.86, − 1.61 − 0.99 − 0.81 − 0.58 (− 2.49, (n = 456) (− 5.59, 3.46) 3.42) 3.36) (− 3.86, (− 3.19, (− 2.68, 1.34) 0.15) 0.64) 1.21) 1.05) Women (n = 2383)  Eating fre- quency   Eating − 0.07 − 0.07 − 0.06 − 0.51 − 0.02 − 0.03 − 0.10 − 0.27 (− 0.88, occasion (− 1.06, (− 0.88, (− 0.75, (− 1.46, (− 0.58, (− 0.58, (− 0.47, 0.35) fre- 0.91) 0.75) 0.86) 0.43) 0.54) 0.52) 0.67) quency   Meal fre- 2.33 (0.06, − 0.13 0.14 (− 1.82, − 0.76 − 0.07 − 0.28 − 0.01 − 0.65 (− 2.15, quency 4.59)* (− 2.03, 2.09) (− 2.86, (− 1.72, (− 1.83, (− 1.55, 0.85) 1.78) 1.33) 1.58) 1.27) 1.53) − 0.08 0.00 (− 0.92, − 0.52 0.03 (− 0.54, 0.07 (− 0.51, 0.15 (− 0.43, − 0.16 (− 0.81,   Snack fre- − 0.62 quency (− 1.73, (− 1.01, 0.92) (− 1.56, 0.61) 0.64) 0.74) 0.49) 0.50) 0.84) 0.52)  Temporal eating pat- terns   Conven- – – – – – – – – tional (refer- ence, n = 1001)   Later 2.13 (− 0.72, 2.38 (− 0.01, 2.55 (0.12, 2.45 (0.05, 1.69 (0.16, 1.56 (0.11, 1.73 (0.27, 1.69 (0.25, lunch 4.97) 4.77) 4.97)* 4.84)* 3.21)* 3.01)* 3.19)* 3.13)* (n = 807)   Grazing 0.08 (− 2.76, 2.41 (− 0.27, 2.20 (− 0.52, 1.93 (− 0.87, 0.98 (− 0.69, 0.85 (− 0.83, 0.64 (− 1.08, 0.50 (− 1.22, (n = 575) 2.93) 5.10) 4.92) 4.72) 2.66) 2.54) 2.35) 2.22) 1 3 European Journal of Nutrition Table 2 (continued) Values are presented as β coefficients (95% confidence intervals). Associations were examined using the Wald tests of associations for linear regression; *P < 0.05, **P < 0.01 Cr ude analysis Adjusted for age (years, continuous), sedentary time (min/day, continuous), education level (low/medium/high), country of birth (Australia/ other mainly English-speaking countries/all other countries), meets PA guidelines (yes/no), smoking status (never smoked/past smoker/current smoker), and dieting (yes/no) Model 2 and additionally adjusted for BMI scores Model 3 and additionally adjusted for Dietary Guideline Index scores and total energy intake Table 3 Associations of eating patterns with hypertension prevalence in Australian men and women a b c d Model 1Model 2Model 3Model 4 Men (n = 2099)  Eating frequency   Eating occasion frequency 0.87 (0.80, 0.97)* 0.85 (0.75, 0.96)** 0.87 (0.76, 0.98)* 0.94 (0.81, 1.08)   Meal frequency 1.03 (0.76, 1.40) 0.88 (0.64, 1.21) 0.92 (0.66, 1.30) 1.12 (0.77, 1.61)   Snack frequency 0.86 (0.76, 0.96)* 0.85 (0.75, 0.96)* 0.86 (0.75, 0.98)* 0.91 (0.79, 1.05)  Temporal eating patterns   Conventional (reference, n = 941) – – – –   Later lunch (n = 702) 0.70 (0.52, 0.93)* 0.78 (0.56, 1.08) 0.81 (0.59, 1.12) 0.82 (0.59, 1.13)   Grazing (n = 456) 0.81 (0.55, 1.20) 1.14 (0.74, 1.74) 1.15 (0.76, 1.75) 1.22 (0.81, 1.82) Women (n = 2383)  Eating frequency   Eating occasion frequency 0.94 (0.83, 1.08) 0.95 (0.83, 1.08) 0.97 (0.84, 1.10) 0.93 (0.79, 1.09)   Meal frequency 1.10 (0.81, 1.50) 0.94 (0.70, 1.27) 0.96 (0.70, 1.31) 0.93 (0.79, 1.09)   Snack frequency 0.91 (0.79, 1.05) 0.95 (0.83, 1.09) 0.95 (0.82, 1.10) 0.93 (0.79, 1.09)  Temporal eating patterns   Conventional (reference, n = 1001) – – – –   Later lunch (n = 807) 1.34 (0.94, 1.90) 1.47 (1.00, 2.16) 1.51 (1.01, 2.25)* 1.49 (1.00, 2.22)*   Grazing (n = 575) 0.97 (0.64, 1.45) 1.11 (0.71, 1.73) 1.06 (0.67, 1.68) 1.04 (0.64, 1.67) Values are presented as odds ratios (95% confidence intervals). Associations were examined using the Wald tests of associations for logistic regression; *P < 0.05, **P < 0.01 Cr ude analysis Adjusted for age (years, continuous), sedentary time (min/day, continuous), education level (low/medium/high), country of birth (Australia/ other mainly English-speaking countries/all other countries), meets PA guidelines (yes/no), smoking status (never smoked/past smoker/current smoker), and dieting (yes/no) Model 2 and additionally adjusted for BMI scores Model 3 and additionally adjusted for Dietary Guideline Index scores and total energy intake time-periods, the present study examined temporal eating higher blood glucose levels, when compared to participants patterns based on a novel latent class analysis approach, who had the same meal in the morning [41]. Insulin metabo- which captures the timing multiple EO across the day and lism may contribute to the association between temporal eat- the likely correlations of energy intakes between EO [4]. ing patterns and BP due to its role in modulating vasodilator The possible mechanisms by which the later timing of a effects on the endothelium via nitric oxide bioavailability “lunch” meal might increase SBP and DBP among women [40]. Measures of insulin sensitivity and blood glucose lev- are unclear. The timing of the “dinner” (evening) meal also els in future epidemiological research examining temporal tended to be later in women with this pattern and research eating patterns and BP are needed. has shown that insulin sensitivity gradually lowers across Strengths of this study include the examination of asso- the day, into the evening [40]. In addition, in an experimen- ciations with BP in a large nationally representative sam- tal trial that controlled for dietary intakes, the timing of an ple, adjusted for BMI, and multiple important confounders, evening meal high in energy and carbohydrates but low in including a measure of overall diet quality. Eating patterns fibre was associated with reduced insulin sensitivity and were determined from 2 days of dietary recall, and an EO was 1 3 European Journal of Nutrition defined using an evidence-based approach [ 27]. While the References novel methodology used to determine temporal eating patterns 1. Collaborators GBDRF. (2017) Global, regional, and national com- is also considered a study strength, it should be noted that parative risk assessment of 84 behavioural, environmental and nd fi ings from data-driven, exploratory methods may not be occupational, and metabolic risks or clusters of risks, 1990–2016: generalizable to populations from other countries. For exam- a systematic analysis for the Global Burden of Disease Study ple, the temporal eating patterns found in the present study 2016. Lancet 390(10100):1345–1422. h t t p s : / / d o i . o rg / 1 0 . 1 0 1 6 / S0140 -6736(17)32366 -8 may reflect the sociocultural and lifestyle characteristics of 2. Estruch R, Ros E, Martinez-Gonzalez MA (2013) Mediterranean the Australian adult population; further research is needed to diet for primary prevention of cardiovascular disease. N Engl J better understand differences in temporal eating patterns in Med 369(7):676–677. https ://doi.org/10.1056/NEJMc 13066 59 other populations. Limitations of the present study include the 3. Appel LJ, Brands MW, Daniels SR, Karanja N, Elmer PJ, Sacks FM, American Heart A (2006) Dietary approaches to prevent cross-sectional study design which does not permit the assess- and treat hypertension: a scientific statement from the Ameri- ment of temporal relationships and the lack of data on anti- can Heart Association. Hypertension 47(2):296–308. https ://doi. hypertensive medications which may have led to an underes- org/10.1161/01.HYP.00002 02568 .01167 .B6 timation of associations with BP. However, in the analysis that 4. St-Onge MP, Ard J, Baskin ML, Chiuve SE, Johnson HM, Kris- Etherton P, Varady K (2017) Meal timing and frequency: implica- excluded participants with no current/previous self-reported tions for cardiovascular disease prevention: a scientific statement hypertension, the association between a “later lunch pattern” from the American Heart Association. Circulation 135(9):e96– with DBP persisted. e121. https ://doi.org/10.1161/CIR.00000 00000 00047 6 In conclusion, a temporal eating pattern distinguished by a 5. Leech RM, Worsley A, Timperio A, McNaughton SA (2015) Understanding meal patterns: definitions, methodology and later “lunch” meal when compared to a “conventional” Aus- impact on nutrient intake and diet quality. Nutr Res Rev 28(1):1– tralian mealtime pattern was associated with higher BP in 21. https ://doi.org/10.1017/S0954 42241 40002 62 women. Among men, a higher snack frequency was inversely 6. Asher G, Sassone-Corsi P (2015) Time for food: the intimate associated with BP, but these associations disappeared after interplay between nutrition, metabolism, and the circadian clock. Cell 161(1):84–92. https ://doi.org/10.1016/j.cell.2015.03.015 adjustment for total energy intake and overall diet quality. 7. Odegaard AO, Jacobs DR Jr, Steffen LM, Van Horn L, Ludwig Future research that considers the role of overall diet quality DS, Pereira MA (2013) Breakfast frequency and development and cardiometabolic indicators (e.g. insulin and glucose levels) of metabolic risk. Diabetes Care 36(10):3100–3106. https ://doi. on the relation between eating patterns and BP is warranted. org/10.2337/dc13-0316 8. Mekary RA, Giovannucci E, Cahill L, Willett WC, van Dam RM, Hu FB (2013) Eating patterns and type 2 diabetes risk in older Acknowledgements A.T. was supported by a National Heart Founda- women: breakfast consumption and eating frequency. Am J Clin tion of Australia Future Leader Fellowship (Award 100046). S.A.M. is Nutr 98(2):436–443. https ://doi.org/10.3945/ajcn.112.05752 1 supported by a National Health and Medical Research Council Career 9. Pot GK, Almoosawi S, Stephen AM (2016) Meal irregularity and Development Fellowship (ID1104636). R.M.L. is supported by an cardiometabolic consequences: results from observational and Alfred Deakin Postdoctoral Research Fellowship. The authors would intervention studies. Proc Nutr Soc 75(4):475–486. https ://doi. like to acknowledge Dr. Katherine Livingstone for the data preparation org/10.1017/S0029 66511 60002 39 of the diet quality score variable used in this study. 10. Deshmukh-Taskar P, Nicklas TA, Radcliffe JD, O’Neil CE, Liu Y (2013) The relationship of breakfast skipping and type of break- Compliance with ethical standards fast consumed with overweight/obesity, abdominal obesity, other cardiometabolic risk factors and the metabolic syndrome in young Research involving human participants and/or animals This research adults. The National Health and Nutrition Examination Survey does not include patient data or data from a clinical trial. This research (NHANES): 1999–2006. Public Health Nutr 16(11):2073–2082. is secondary analysis of Australian (national) Health Survey (AHS) https ://doi.org/10.1017/S1368 98001 20042 96 data commissioned by the Commonwealth of Australia and conducted 11. Sierra-Johnson J, Unden AL, Linestrand M, Rosell M, Sjogren P, by the Australian Bureau of Statistics (ABS). This secondary analysis Kolak M, De Faire U, Fisher RM, Hellenius ML (2008) Eating was registered at anzctr.org.au as ACTRN12617001029381. The web meals irregularly: a novel environmental risk factor for the meta- address of this trial is: http://www .ANZCT R.org.au/ACTRN 12617 bolic syndrome. Obesity (Silver Spring) 16(6):1302–1307. https 00102 9381.aspx. Ethical approval for the ABS to conduct the house- ://doi.org/10.1038/oby.2008.203 hold components of the AHS is provided by the Australian Federal 12. Kim S, Park GH, Yang JH, Chun SH, Yoon HJ, Park MS (2014) Government Census and Statistics Act, 1905 (https ://www.legis latio Eating frequency is inversely associated with blood pressure n.gov.au/Detai ls/C2006 C0017 8). and hypertension in Korean adults: analysis of the Third Korean National Health and Nutrition Examination Survey. Eur J Clin Nutr. https ://doi.org/10.1038/ejcn.2014.9 Conflict of interest On behalf of all authors, the corresponding author 13. Karatzi K, Georgiopoulos G, Yannakoulia M, Efthimiou E, Voi- states that there is no conflict of interest. donikola P, Mitrakou A, Manios E, Alevizaki M, Papamichael C, Stamatelopoulos K (2016) Eating frequency predicts new Open Access This article is distributed under the terms of the Crea- onset hypertension and the rate of progression of blood pressure, tive Commons Attribution 4.0 International License (http://creat iveco arterial stiffness, and wave reflections. J Hypertens. https ://doi. mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- org/10.1097/hjh.00000 00000 00082 2 tion, and reproduction in any medium, provided you give appropriate 14. Karatzi K, Yannakoulia M, Psaltopoulou T, Voidonikola P, Kol- credit to the original author(s) and the source, provide a link to the lias G, Sergentanis TN, Retsas T, Alevizaki M, Papamichael C, Creative Commons license, and indicate if changes were made. 1 3 European Journal of Nutrition Stamatelopoulos K (2014) Meal patterns in healthy adults: inverse report. JAMA 289 (19):2560–2572. https ://doi.or g/10.1001/ association of eating frequency with subclinical atherosclerosis jama.289.19.2560 indexes. Clin Nutr. https ://doi.org/10.1016/j.clnu.2014.04.022 27. Leech RM, Worsley A, Timperio A, McNaughton SA (2015) 15. Bhutani S, Varady KA (2009) Nibbling versus feasting: which Characterizing eating patterns: a comparison of eating occa- meal pattern is better for heart disease prevention? Nutr Rev sion definitions. Am J Clin Nutr 102(5):1229–1237. https ://doi. 67(10):591–598. https://doi.or g/10.1111/j.1753-4887.2009.00231 org/10.3945/ajcn.115.11466 0 .x 28. Leech RM, Worsley A, Timperio A, McNaughton SA (2017) 16. Titan SM, Bingham S, Welch A, Luben R, Oakes S, Day N, Khaw Temporal eating patterns: a latent class analysis approach. Int KT (2001) Frequency of eating and concentrations of serum cho- J Behav Nutr Phys Act 14(1):3. h tt p s : / / do i . o rg/ 1 0 . 1 1 86 / s 1 2 9 6 lesterol in the Norfolk population of the European prospective 6-016-0459-6 investigation into cancer (EPIC-Norfolk): cross sectional study. 29. Lo Y, Mendell NR, Rubin DB (2001) Testing the number of com- BMJ 323(7324):1286–1288 ponents in a normal mixture. Biometrika 88(3):767–778. https :// 17. Edelstein SL, Barrett-Connor EL, Wingard DL, Cohn BA (1992) doi.org/10.1093/biome t/88.3.767 Increased meal frequency associated with decreased cholesterol 30. Australian Institute of Health and Welfare (2003) The Active Aus- concentrations; Rancho Bernardo, CA, 1984–1987. Am J Clin tralia Survey: a guide and manual for implementation, analysis Nutr 55(3):664–669 and reporting. Cat. no. CVD 22. AIHW, Canberra 18. Leech RM, Livingstone KM, Worsley A, Timperio A, 31. The Department of Health (2014) Australia’s physical activity and McNaughton SA (2016) Meal frequency but not snack frequency sedentary behaviour guidelines [Webpage on the Internet]. DoH, is associated with micronutrient intakes and overall diet quality Canberra (cited 26 Nov 2017) in Australian men and women. J Nutr 146(10):2027–2034. https 32. National Health and Medical Research Council (2013) Australian ://doi.org/10.3945/jn.116.23407 0 dietary guidelines. NHMRC, Canberra 19. Murakami K, Livingstone MB (2015) Associations between meal 33. Thorpe MG, Milte CM, Crawford D, McNaughton SA (2016) A and snack frequency and diet quality and adiposity measures in revised Australian dietary guideline index and its association with British adults: findings from the National Diet and Nutrition Sur - key sociodemographic factors, health behaviors and body mass vey. Public Health Nutr 19(9):1624–1634. https: //doi.org/10.1017/ index in peri-retirement aged adults. Nutrients 8(3):160. https :// S1368 98001 50029 79doi.org/10.3390/nu803 0160 20. Leech RM, Worsley A, Timperio A, McNaughton SA (2017) The 34. Livingstone KM, McNaughton SA (2016) Diet quality is associ- role of energy intake and energy misreporting in the associations ated with obesity and hypertension in Australian adults: a cross between eating patterns and adiposity. Eur J Clin Nutr. https: //doi. sectional study. BMC Public Health 16(1):1037. https ://doi. org/10.1038/ejcn.2017.90org/10.1186/s1288 9-016-3714-5 21. Almoosawi S, Prynne CJ, Hardy R, Stephen AM (2013) Time- 35. Banna JC, McCrory MA, Fialkowski MK, Boushey C (2017) of-day of energy intake: association with hypertension and blood Examining plausibility of self-reported energy intake data: con- pressure 10 years later in the 1946 British Birth Cohort. J Hyper- siderations for method selection. Front Nutr 4:45. https ://doi. tens 31(5):882–892 org/10.3389/fnut.2017.00045 22. Keller K, Rodriguez Lopez S, Carmenate Moreno M (2017) Asso- 36. Vainik U, Konstabel K, Latt E, Maestu J, Purge P, Jurimae J ciation between meal intake behavior and blood pressure in Span- (2016) Diet misreporting can be corrected: confirmation of the ish adults. Nutr Hosp 34(3):654–660. https ://doi.org/10.20960 / association between energy intake and fat-free mass in adoles- nh.487 cents. Br J Nutr 116(8):1425–1436. https://doi.or g/10.1017/s0007 23. Leech RM, Timperio A, Livingstone KM, Worsley A, 11451 60033 17 McNaughton SA (2017) Temporal eating patterns: associations 37. Folsom AR, Parker ED, Harnack LJ (2007) Degree of concord- with nutrient intakes, diet quality, and measures of adiposity. ance with DASH diet guidelines and incidence of hypertension Am J Clin Nutr 106(4):1121–1130. h t t p s : / / d o i . o r g / 1 0 . 3 9 4 5 / and fatal cardiovascular disease. Am J Hypertens 20(3):225–232. ajcn.117.15658 8https ://doi.org/10.1016/j.amjhy per.2006.09.003 24. Eicher-Miller HA, Khanna N, Boushey CJ, Gelfand SB, Delp EJ 38. Almoosawi S, Vingeliene S, Karagounis LG, Pot GK (2016) (2016) Temporal dietary patterns derived among the adult par- Chrono-nutrition: a review of current evidence from observational ticipants of the National Health and Nutrition Examination Sur- studies on global trends in time-of-day of energy intake and its vey 1999–2004 are associated with diet quality. J Acad Nutr Diet association with obesity. Proc Nutr Soc 75(4):487–500. https :// 116(2):283–291. https ://doi.org/10.1016/j.jand.2015.05.014doi.org/10.1017/S0029 66511 60003 06 25. Australian Bureau of Statistics (2013) Australian health survey: 39. Almoosawi S, Prynne CJ, Hardy R, Stephen AM (2013) Time- users’ guide, 2011–2013. Cat. no. 4363.0.55.001. ABS, Canberra of-day and nutrient composition of eating occasions: prospective 26. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green association with the metabolic syndrome in the 1946 British birth LA, Izzo JL Jr, Jones DW, Materson BJ, Oparil S, Wright JT cohort. Int J Obes 37(5):725–731 Jr, Roccella EJ, National Heart, Lung, and Blood Institute Joint 40. Steinberg HO, Baron AD (2002) Vascular function, insulin resist- National Committee on Prevention, Detection, Evaluation, and ance and fatty acids. Diabetologia 45(5):623–634. h t t p s : / / d o i . Treatment of High Blood Pressure, National High Blood Pressure org/10.1007/s0012 5-002-0800-2 Education Program Coordinating Committee (2003) The seventh 41. Morgan LM, Shi J-W, Hampton SM, Frost G (2012) Effect of report of the Joint National Committee on Prevention, Detection, meal timing and glycaemic index on glucose control and insulin Evaluation, and Treatment of High Blood Pressure: the JNC 7 secretion in healthy volunteers. Br J Nutr 108(7):1286–1291 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png European Journal of Nutrition Springer Journals

Eating patterns of Australian adults: associations with blood pressure and hypertension prevalence

Free
11 pages
Loading next page...
 
/lp/springer_journal/eating-patterns-of-australian-adults-associations-with-blood-pressure-yHKcC0zENZ
Publisher
Springer Berlin Heidelberg
Copyright
Copyright © 2018 by The Author(s)
Subject
Chemistry; Nutrition
ISSN
1436-6207
eISSN
1436-6215
D.O.I.
10.1007/s00394-018-1741-y
Publisher site
See Article on Publisher Site

Abstract

Purpose Eating patterns have been linked to obesity, an established risk factor for hypertension; however, their contribution to hypertension is poorly understood. This study aimed to examine associations of frequency of meals, snacks and all eating occasions (EO), and temporal eating patterns, with blood pressure (BP) and hypertension. Methods Dietary data collected via two 24-h recalls during the 2011–2012 Australian National Nutrition and Physical Activ- ity Survey (n = 4482 adults, ≥ 19 years) were analysed. Frequencies of EO, meals, and snacks were calculated. Temporal eating patterns were determined using latent class analysis. Multivariate regression models assessed associations of eating patterns with systolic BP (SBP), diastolic BP (DBP), and hypertension prevalence. Results Among men, a higher snack frequency was inversely associated with DBP [β = − 0.59, 95% confidence interval (CI) (− 1.12, − 0.07)] and hypertension [odds ratio (OR) 0.86, 95% CI (0.75, 0.98)] after adjustment for covariates and BMI. However, these associations disappeared after additional adjustment for total energy intake and overall diet quality. Among women, a temporal eating pattern characterized by a later “lunch” meal was associated with SBP [β = 2.45, 95% CI (0.05, 4.84)], DBP [β = 1.69, 95% CI (0.25, 3.13)], and hypertension [OR = 1.49, 95% CI (1.00, 2.22)], when compared to a “conventional” eating pattern. Conclusions In this study, an inverse association found between snack frequency and BP among men disappeared after adjustment for dietary factors and a “later lunch” pattern was associated with higher BP in women. Future research is needed to understand the relationship and potential mechanistic pathways between eating patterns and BP. Keywords Blood pressure · Circadian rhythms · Eating frequency · Eating patterns · Meals · Meal timing · Snacks Introduction foods has been identified as an important strategy for the prevention of cardiovascular disease and hypertension [2, Poor dietary habits, obesity, and high systolic blood pres- 3]. However, the contribution of eating patterns to hyperten- sure (SBP) are three major risk factors contributing to sion is poorly understood [4]. Eating patterns, including the the global burden of disease [1]. As these risk factors are frequency and temporality of eating occasions [(EO), e.g., strongly interrelated, it has been postulated that diet is a meals and snacks], may represent a unique cardiometabolic key driver of global increases in BMI and subsequent SBP risk factor [5]. This is because the metabolic and physiologi- [1]. The adoption of a dietary pattern rich in plant-based cal functions in responses to food are also, in part, regulated foods but low in red meat and discretionary (e.g., non-core) by circadian clock systems. Thus, variations in the temporal distribution of food consumption may have important health implications [6]. Electronic supplementary material The online version of this Epidemiologic evidence [7–11] suggests that skipping article (https ://doi.org/10.1007/s0039 4-018-1741-y) contains supplementary material, which is available to authorized users. breakfast and irregular meal habits are associated with poorer cardiometabolic health in adults, including incident * Rebecca M. Leech hypertension [7]. However, evidence for associations with rebecca.leech@deakin.edu.au EO frequency is less consistent [4]. In three studies, a higher Institute for Physical Activity and Nutrition (IPAN), School EO frequency (e.g., ≥ 5 EO per day) was associated with of Exercise and Nutrition Sciences, Deakin University, 75 lower SBP [12, 13] and diastolic blood pressure (DBP) [13, Pigdons Rd, Waurn Ponds, Geelong 3216, Australia Vol.:(0123456789) 1 3 European Journal of Nutrition 14]. In support of these findings, a review of short-term approval to conduct the survey components of the NNPAS experimental studies found that six or more meals per day, [25]. compared to one meal, resulted in favourable decreases in cardiometabolic risk markers [15]. However, most of the Blood pressure reviewed studies did not control for differences in dietary intake and other population-based studies have found no All BP measurements were voluntary, excluded pregnant evidence for a relation of EO frequency with blood pressure women, and were taken by trained ABS staff during the (BP) [16, 17]. Furthermore, research suggests that meals and home visit. Participants were asked to sit comfortably and snacks are differentially related to dietary intakes [18, 19] relax their left arm. Two SBP and DBP measurements were and obesity [19, 20], yet studies examining meals and snacks taken on the left arm, with the palm facing upwards, using an as separate EO in relation to hypertension are rare [12]. automated BP monitor. The second reading was used, except The temporal distribution of EO in relation to BP has where there was greater than 10 mmHg difference between rarely been examined [21, 22]. In the UK Cohort Study, the first and second readings, in which case, a third reading Almoosawi et al. [21] found a positive association for higher was taken and the second and third readings averaged. If total energy intake in the evening with 10-year increases in all three readings differed from each other by > 20 mmHg, SBP and DBP. In a study of Spanish adult volunteers, con- then these readings were considered invalid. Participants sumption of an afternoon meal was associated with lower were classified as hypertensive (≥ 140/90 mmHg) or non- SBP and DBP [22]. However, these studies focussed on tim- hypertensive (< 140/90 mmHg) [26]. No data about use of ing of EO/energy intake at isolated periods of the day and BP medications were collected; however, participants’ self- did not simultaneously capture EO at other times of the day reported whether they had current or previous hypertensive [4]. Data-driven statistical techniques (e.g., cluster and latent disease. class analysis) have recently been identified as useful tools for capturing temporal eating patterns across the day, but Dietary assessment have only been used to examine relations with diet quality and obesity [23, 24]. Therefore, the aim of this study was Dietary data were collected over two 24-h recalls, conducted to examine the associations of frequency of meals, snacks approximately 9 days apart, on a different day of the week. and all EO, and temporal eating patterns, determined using During each dietary recall, based on the validated USDA a latent class analysis approach, with BP and hypertension, automated multiple 5-pass method, participants were asked in Australian men and women. Based on the findings of our to identify the type of EO (e.g., breakfast, lunch, dinner, or earlier studies that found a positive relation for snacking snack) and the time when each EO commenced. The Aus- frequency and a “grazing” temporal eating pattern with tralian Supplement and Nutrient Database 2011–2013 was measures of adiposity, we hypothesised that that these same used to determine energy and nutrient intakes from all foods eating patterns would be related to (obesity-associated) BP and beverages and dietary information was averaged across and hypertension [20, 23]. the two recall days to obtain mean estimates of eating pat- terns, total energy intakes, and food intakes. Subjects and methods Frequencies of all EO, meals, and snacks Sample and study design The methods used to calculate mean total frequencies of all EO, meals, and snacks have been described previously This study is a secondary analysis of data drawn from the [18, 27]. Briefly, an EO constituted any eating event that nationally representative, cross-sectional 2011–2012 Aus- provided a minimum energy content of 210 kJ (50 kcal) and tralian National Nutrition and Physical Activity Survey was separated in time from the surrounding EO by 15 min. (NNPAS 2011–12) and was registered at anzctr.org.au as This approach was informed by the previous research and ACTRN12617001029381. The NNPAS, conducted between current recommendations for defining EO in eating patterns May 2011 and June 2012, was administered by the Aus- research [4, 27]. EO were then classie fi d as meals and snacks tralian Bureau of Statistics (ABS); full details of the study according to participants’ self-report of EO. Meals included design and methods have been described elsewhere [25]. EO reported as breakfast, brunch, lunch, dinner, and supper, Briefly, 12,153 participants aged 2 or more years (includ- whereas snacks included EO reported as morning/afternoon ing 9338 adults, aged ≥ 19 years; 77% response rate) were tea and beverage break were classified as snack EO. Based selected using a multistage, probability sampling design of on the sample distribution, frequencies of all EO, meals and private dwellings (Fig. 1). The Australian Government Cen- snacks were divided into categories of 1–3, 4–5, or ≥ 6 EO, sus and Statistics Act 1905 provided the ABS with ethics 1–2, 3, or > 3 meals, and 0–1, 2–3, or > 3 snacks. 1 3 European Journal of Nutrition Fig. 1 Flowchart of included NNPAS 2011-12 participants from the 2011– Total survey participants 2012 Australian National Nutrition and Physical Activity n = 12153 Survey (NNPAS 2011–12) Adults aged ≥19 years n = 9341 Completed two 24 hour dietary recalls n = 6053 (65%) Not pregnant, breastfeeding or undertaking shift work in past 4 weeks n = 5366 Missing data on eating occasion type or time of eating occasion consumption n = 116 excluded Reported no energy intake during a 24 hour recall n = 8 excluded Missing data for blood pressure n = 578 excluded Missing data for BMI or covariates n = 182 excluded Final analytic sample n = 4482 characterized by frequent but less distinct meal occasions Temporal eating patterns and a higher probability of having an EO after 8 p.m. These temporal eating patterns have been associated with socio- Temporal eating patterns were determined using latent class demographic and eating pattern characteristics [28], diet analysis, as reported in detail previously [23, 28]. Briefly, quality, and, among women, adiposity outcomes [23]. three latent classes of temporal eating patterns were identi- fied for men and women based on frequency and timing of Covariates EO across the day and labelled according to their defining characteristics. The optimal number of classes was selected The following socio-demographic, health behaviours, and based on model fit indices, likelihood ratio tests comparing −1 anthropometric variables, collected during the household k with k class models, and pattern interpretability [29]. survey, were considered as potential covariates due to their The first pattern was labelled as “conventional” due to the previous reported relation with hypertension risk [1, 26]. probability of participants having three EO at “conven- tional” mealtimes in Australia (e.g., between 7 and 8 a.m., Socio‑demographics 12–1 p.m., and 6–7 p.m.). The second pattern was distin- guished by a > 0.9 probability of a “lunch” meal approxi- Education level was categorised as: low (completed some mately 1 h later than the “conventional pattern” and labelled high-school or less), medium (completed high-school or the “later lunch” pattern. The probability of a “dinner” meal completed some high-school and/or certificate/diploma) 1 h later than the conventional pattern was also higher (e.g., or high (having a tertiary qualification). Country of birth > 0.6) in participants with a “later lunch” pattern. The third was categorised by the ABS as: Australia, predominantly pattern was labelled the “Grazing” pattern, because it was 1 3 European Journal of Nutrition English-speaking countries (other than Australia) and all Statistics other countries. All statistical analyses were stratified by sex and used Stata statistical software, Version 14.2 (Stata Inc., College Sta- Health behaviours tion, TX, USA). Point estimates and standard errors were determined by applying person and replicate weights that Smoking status was self-reported and categorised as cur- accounted for the probability of participant selection and rent smoker, ex-smoker, and never smoked. Participants the clustered survey design, respectively. Descriptive sta- were categorised as meeting or not meeting current Aus- tistics for sample characteristics are presented as weighted tralian physical activity guidelines (150 min and 5 sessions), means (95% CI) or weighted percentages. After examining based on self-reported frequency and duration of walking the distribution of the data, the following variables were for recreation or transport and moderate or vigorous leisure- log-transformed to improve normality: BMI, daily total sed- time physical activity [30, 31]. Self-reported information entary time, and total energy intake. Weighted geometric on sleep duration the night before the survey and how much means (95% CI) were used for all log-transformed variables. time participants spent sitting or lying down at work, during The F test (for continuous data) and adjusted Pearson χ transport and leisure activities in the past week, were used to test (for categorical data) were used to determine sex-spe- calculate (per day) total minutes spent sleeping per day and cific differences in sample characteristics by hypertension in sedentary behaviour, respectively. Participants reported status. Multiple linear regression (for continuous outcomes) whether they were currently on a weight-loss diet for health and logistic regression (for binary outcomes) were used to reasons (yes/no). Average daily total energy intake and diet test for associations of frequencies of all EO, meal and quality scores were calculated from the 2 days of recall. The snacks (continuous), and temporal eating patterns, with SBP established food-based Dietary Guidelines Index (DGI) was and DBP (continuous) and hypertension prevalence (binary). used as a measure of overall diet quality [32–34]. The DGI Four models were tested: model 1 was an unadjusted model; assesses compliance with recommendations outlined in the model 2 adjusted for age (years, continuous), education level Australian Dietary Guidelines, and is the sum of 13 com- (low, medium or high), country of birth (Australia, other pre- ponents (score range of 0–130), each corresponding to an dominantly English-speaking countries, all other countries), Australian Dietary guideline and scored proportionally out smoking status (never, former or current), daily, meeting of 10. The components include meeting recommendations physical activity guidelines (yes/no), daily sedentary time for food variety, intakes of fruits, vegetables (including leg- (min; continuous), sleep duration (h, continuous), dieting for umes), grain foods, dairy and alternatives, meat and alterna- health reasons (yes/no); model 3 further adjusted for BMI, tives, unsaturated fat, fluids, discretionary foods, saturated and model 4 further adjusted for total energy intake and DGI fat, salt, added sugar, and alcohol. Higher scores indicate a scores (both continuous). In light of the previous research better diet quality. Measurement of height (cm) and weight that reported a positive association among participants with (kg) were taken to one decimal point by trained ABS staff the highest EO frequencies (i.e., > 5 EO) [13], in the pre- using a portable stadiometer and digital scales. BMI (weight sent study, any observed statistically significant (P < 0.05) [kg]/height [m] ) was calculated. adjusted association between the continuous measures of frequencies of EO, meals, or snacks, and the outcome vari- ables were further explored by examining associations for Analytic sample eating pattern frequency categories (e.g., 1–3 [reference], 4–5 or ≥ 6 EO; 1–2 [reference], 3 or > 3 meals and 0–1 [ref- The analytic sample included the 65% of adult participants erence], and 2–3 or > 3 snacks). Finally, the effect of energy who completed both dietary recalls (n = 6053; Fig. 1). Par- misreporting, defined as the ratio of total energy intake to ticipants were eligible for this analysis if they were not preg- total energy expenditure was considered [35]; however, its nant, breastfeeding, or undertaking shift-work in the past inclusion did not improve its predictive power when BMI 4 weeks (n = 5366) and were excluded if they reported no was already in the model. A previous study has also shown energy intake during either dietary recall (n = 8 excluded) that energy misreporting bias can be statistically corrected or did not report the time at which an EO commenced or using predictors of energy misreporting (i.e., dieting behav- the type of EO (n = 116 excluded). Of the remaining 5242 iours and BMI) [36]. participants, 578 (11%) had missing data for BP and a fur- ther 182 (3.9%) were missing data for BMI and covariates: Sensitivity analysis BMI (n = 149), physical activity (n = 28), and sedentary time (n = 5). The final analytic sample was 2099 men and 2383 As data on BP medications use were not collected in the women. NNPAS [25], a sensitivity analysis was conducted that 1 3 European Journal of Nutrition included only participants who self-reported no current or men, a significant inverse association was found for fre- previous hypertensive disease (n = 1612 men and n = 1853 quency of all EO and snacks but not meals. women). After further analysis that examined these associations by categories of snack and EO frequency, the inverse adjusted associations (model 3) were only observed among men who reported > 3 snacks [OR 0.52, 95% CI (0.30, 0.81); Results P = 0.006] and ≥ 6 EO [OR 0.54, 95% CI (0.34, 0.84); P = 0.008], compared to those who reporting < 2 snacks and Table 1 presents the characteristics of men and women par- ≤ 3 EO. However, again, these associations attenuated after ticipants in the NNPAS 2011–12 by hypertension status. further adjustment for total energy intake and DGI scores, Of the participants, 24% of men and 20% of women were and no significant associations were found among women. classified as having hypertension. Among both sexes, there Compared to a “conventional” temporal eating pattern, a were significant differences by hypertension status for age, “later lunch temporal” eating pattern was also associated education level, meeting physical activity guidelines, BMI, with hypertension prevalence among women, but only after and self-report status of previous or current hypertensive adjustment for covariates and BMI scores. disease (P < 0.05). Differences were also found for total Results of the sensitivity analyses which included only daily energy intake, DGI scores, and frequency of all EO men and women with no self-reported current or previ- and snacks between non-hypertensive and hypertensive men ous hypertensive disease, showed similar null associations (P < 0.05). For the snack frequency categories, a higher and of frequency of EO, snacks, and meals with SBP, DBP or lower proportion of hypertensive men reported having fewer hypertension prevalence, after adjustment for covariates, than two snacks and more than three snacks per day, respec- BMI, and dietary intakes (Supplementary Table 1). How- tively, compared to non-hypertensives (P < 0.05). No signifi- ever, among women, the finding of a positive association cant differences were found among women for the dietary between a “later lunch” pattern and DBP (but not SBP or or eating pattern variables according to hypertension status. hypertension prevalence) persisted after exclusion of those The sex-specific associations of frequency of all EO, with no self-reported current/previous hypertensive disease. meals, and snacks (continuous) with SBP and DBP are pre- sented in Table 2. In the basic model, a statistically signifi- cant positive association was found between meal frequency Discussion and SBP among women which disappeared after adjustment for the covariates in model 2. Among men, inverse associa- To our knowledge, this is one of the first studies among tions were found between frequency of all EO and snacks adults to examine associations of meal and snack frequency and DBP, and after adjustment for covariates (model 2) and and temporal eating patterns, based on the timing and fre- BMI (model 3). quency of EO across the day, with BP and hypertension After further analysis that examined these associations by prevalence [12]. Among men, frequency of all EO and categories of snack and EO frequency, the inverse associa- snacks was inversely associated with DBP and lower odds tions were observed among men who reported > 3 snacks of hypertension prevalence, but these associations disap- [DBP: β = − 2.23, 95% CI (− 4.14, − 0.32); P = 0.023] peared after adjustment for overall diet quality scores and and ≥ 6 EO [DBP: β = − 2.32, 95% CI (− 4.45, − 0.20); total energy intakes. Among women, a “later lunch” pattern, P = 0.032], compared to those who reporting < 2 snacks and identified using latent class analysis in our earlier study [28], ≤ 3 EO. However, these associations with DBP for men were and characterized by a later lunch EO (e.g., between 1 and attenuated after further adjustment for the total energy intake 2 p.m.) was associated with higher SBP, DBP, and hyper- and DGI scores: β = − 1.56, 95% CI (− 3.59, 0.47); P = 0.13 tension prevalence. However, only associations with DBP and β = − 1.37, 95% CI (− 3.62, 0.89); P = 0.23, respectively. persisted after exclusion of persons with the self-reported Results of the regression analyses showed no associa- previous/current hypertension. tions between latent classes of temporal eating patterns and Only a few studies have examined the relationship SBP or DBP among men (Table 2). Among women, a “later between EO frequency and BP among adults [12, 13, 16, lunch” temporal eating pattern was positively associated 17], with conflicting findings. However, it is difficult to com- with SBP and DBP, when compared to a “conventional” pare the results of these studies, because they define EO pattern, after adjustment for covariates and BMI, and after using different approaches. For example, EO have mostly further adjustment for dietary intakes. been self-reported by participants in response to a single Associations of frequency of all EO, meals, and snacks survey question where an EO is not further defined [12, (continuous) and latent classes of temporal eating patterns 17]. Whereas in another study, in a small sample of healthy with hypertension prevalence are shown in Table 3. Among volunteers (n = 115), EO was defined as any eating event 1 3 European Journal of Nutrition Table 1 Characteristics of men and women in the NNPAS by hypertension status Men (n = 2099) Women (n = 2383) Non-hypertensive Hypertensive (n = 561) P value Non-hypertensive Hypertensive (n = 500) P value (n = 1538) (n = 1883) Socio-demographics  Age (years) 43.1 (42.4, 44.8) 56.2 (54.2, 58.2) < 0.0001 44.7 (44.0, 45.5) 58.4 (56.8, 60.0) < 0.0001  Education level (%) < 0.05 < 0.0001   Low 18 25 25 43   Medium 53 54 44 33   High 29 21 32 24  Country of birth (%) 0.95 0.12   Australia 69 69 70 61   Predominantly 13 13 11 15 English-speaking countries   All other countries 18 18 19 24 Health behaviours or characteristics  Smoking status 0.14 0.56   Never 47 39 59 55   Former 35 40 27 31   Current 18 20 14 14  Meets physical activ- 48 41 0.04 45 35 < 0.01 ity guidelines (%)  Daily sedentary time 306.7 (289.8, 324.6) 297.2 (273.0, 323.4) 0.54 260.3 (247.9, 273.2) 263.1 (237.0, 292.1) 0.86 (min)  Currently on a diet 12 8 0.06 17 15 0.43 for health reasons (%)  Sleep duration (h) 7.9 (7.8, 8.0) 7.9 (7.7, 8.0) 0.81 8.0 (7.9, 8.1) 8.0 (7.7, 8.2) 0.52  Total energy intake 9498 (9304, 9696) 8463 (8170, 8766) < 0.0001 7087 (6928, 7250) 6890 (6587, 7206) 0.31 (kJ)  Dietary Guidelines 80.1 (79.0, 81.2) 77.6 (75.7, 79.6) < 0.05 80.7 (79.5, 82.0) 81.8 (80.1, 83.5) 0.31 Index (score)  BMI (score) 26.8 (26.5, 27.2) 28.9 (28.2, 29.5) < 0.0001 25.9 (25.5, 26.3) 28.7 (27.9, 29.6) < 0.0001  Systolic blood pres- 118.7 (117.9, 119.5) 148.3 (145.6, 151.0) < 0.0001 112.5 (111.7, 113.1) 148.8 (146.9, 150.8) < 0.0001 sure (mmHg)  Diastolic blood pres- 73.4 (72.8, 74.0) 88.7 (87.5, 89.8) < 0.0001 72.9 (72.2, 73.6) 89.0 (87.9, 90.2) < 0.0001 sure (mmHg)  No current or previ- 85 63 < 0.0001 86 56 < 0.0001 ous hypertensive disease (%) Eating patterns  Eating occasion 4.9 (4.8, 5.0) 4.7 (4.5, 4.8) < 0.01 4.8 (4.7, 4.9) 4.7 (4.6, 4.9) 0.37 frequency  Meal frequency 2.9 (2.8, 2.9) 2.9 (2.8, 2.9) 0.79 2.9 (2.9, 3.0) 3.0 (2.9, 3.0) 0.55  Snack frequency 2.1 (1.98, 2.2) 1.8 (1.7, 2.0) < 0.01 1.9 (1.9, 2.0) 1.8 (1.7, 2.0) 0.19  Categories of eating 0.07 0.54 occasion frequency (%)   1–3 19 24 18 20   4–5 57 60 61 62   ≥ 6 23 17 21 18  Categories of meal 0.27 0.89 frequency (%) 1 3 European Journal of Nutrition Table 1 (continued) Men (n = 2099) Women (n = 2383) Non-hypertensive Hypertensive (n = 561) P value Non-hypertensive Hypertensive (n = 500) P value (n = 1538) (n = 1883)   < 3 31 32 26 25   3 53 48 54 54   > 3 15 20 20 21  Categories of snack < 0.05 0.59 frequency (%)   < 2 45 51 48 51   2–3 37 38 39 38   > 3 18 11 13 11  Latent classes of 0.09 0.17 temporal eating pat- terns (%)   Conventional 40 47 41 37   Later lunch 36 30 32 39   Grazing 24 23 27 24 Values are weighted means (95% confidence intervals) or weighted percentages. Significant sex-specific differences by hypertension status assessed using an F test for continuous variables or design-adjusted Pearson χ test Whether met physical activity guidelines of 150 min and 5 sessions/week Values are geometric means (95% CI) DGI represents a total diet quality score (score range 0–130) with higher scores indicating better overall diet quality (including kilojoule-free events) separated in time by 15 min associated with better diet quality scores for intakes of fruits [13]. A higher EO frequency has been associated with lower and dairy products, two food groups recommended as part of hypertension prevalence [12] and incidence [13], whereas the Dietary Approaches to Stop Hypertension (DASH) diet other studies have found no associations with BP [16, 17]. [3]. However, in the same study, snack frequency was also In the present study, EO frequency was inversely associ- associated with poorer scores for intakes of discretionary ated with DBP and hypertension prevalence among men, foods and added sugars among men [18]. Future research but these associations attenuated after further adjustment that examines the role of diet quality on the relation between for total energy intakes and overall diet quality. EO frequency and hypertension is warranted. Studies examining the separate effects of meal and snack Epidemiological evidence suggests a positive association frequency on BP outcomes are rare [12]. Compared to par- between evening energy intakes or the later timing of an EO ticipants who reported no snacks, Kim et al. [12] found that and obesity [38], but studies examining temporal patterns of a snack frequency of three per day was associated with lower eating in relation to BP are rare. In one study, higher energy odds of hypertension, but associations attenuated (e.g., 95% intake at breakfast was associated with lower hypertension CI included one) after adjustment for adiposity measures. In prevalence but not 10-year incidence, and higher energy the present study, snack frequency (specifically > 3 snacks) intake in the evening was associated with higher hyper- was inversely associated with BP outcomes among men, but tension incidence, which remained borderline significant again associations attenuated after adjustment for overall after adjustment for baseline BMI [21]. Another study, in diet quality and total energy intakes. Notably, with respect to the same cohort, found no association between time of day adjustment of dietary factors, the previous studies on eating macronutrient intakes and BP [39]. Keller et al. found that patterns and BP have only adjusted for either total energy the consumption of an afternoon meal, but not other conven- intakes [13] or energy and nutrient intakes [12, 16, 17], and tional Spanish meals, was modestly associated with lower not a measure of overall diet quality based on food intakes. SBP and DBP, even after adjustment for dietary intake and The findings from the present study suggest that diet qual- waist circumference [22]. In the present study, a temporal ity may be an important factor in the relation between EO eating pattern characterized by having a later “lunch” meal frequency and BP, and is supported by the previous studies was associated with SBP, DBP, and hypertension prevalence that have shown a beneficial effect of healthful dietary pat- among women, after adjustment for potential covariates, terns for the prevention of hypertension [37]. In a previous BMI, and diet quality. However, rather than examining the study of NNPAS participants. [18], snack frequency was timing of a single meal or energy intake across stratified 1 3 European Journal of Nutrition Table 2 Associations of eating patterns with systolic and diastolic blood pressure in Australian men and women Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) a b c d a b c d Model 1Model 2Model 3Model 4Model 1Model 2Model 3Model 4 Men (n = 2099)  Eating fre- quency   Eating − 0.49 − 0.65 − 0.47 − 0.18 − 0.80 − 0.88 − 0.63 − 0.39 (− 0.98, occasion (− 1.22, (− 1.41, (− 1.24, (− 1.11, (− 1.36, (− 1.45, (− 1.17, 0.20) fre- 0.25) 0.11) 0.30) 0.74) − 0.23)** − 0.31)** − 0.09)* quency   Meal fre- 0.47 (− 1.96, − 1.29 − 0.86 0.13 (− 2.35, − 1.04 − 1.46 − 0.88 − 0.46 (− 2.18, quency 2.91) (− 3.41, (− 3.13, 2.62) (− 2.43, (− 2.99, (− 2.48, 1.26) 0.84) 1.40) 0.35) 0.08) 0.71) − 0.77 − 0.59 − 0.41 (− 0.99,   Snack fre- − 0.69 − 0.59 − 0.46 − 0.32 − 0.73 (− 1.35, (− 1.12, 0.18) quency (− 1.46, (− 1.37, (− 1.22, (− 1.24, (− 1.33, − 0.13)* − 0.20)** − 0.07)* 0.09) 0.19) 0.30) 0.59)  Temporal eating pat- terns   Conven- – – – – – – – – tional (refer- ence, n = 941)   Later − 2.44 − 1.13 − 0.80 − 0.68 − 0.97 − 0.63 − 0.19 − 0.10 (− 1.65, lunch (− 4.93, (− 3.42, (− 2.94, (− 2.82, (− 2.66, (− 2.37, (− 1.74, 1.45) (n = 702) 0.05) 1.16) 1.33) 1.46) 0.71) 1.11) 1.36)   Grazing − 2.72 0.63 (− 2.19, 0.77 (− 1.89, 0.75 (− 1.86, − 1.61 − 0.99 − 0.81 − 0.58 (− 2.49, (n = 456) (− 5.59, 3.46) 3.42) 3.36) (− 3.86, (− 3.19, (− 2.68, 1.34) 0.15) 0.64) 1.21) 1.05) Women (n = 2383)  Eating fre- quency   Eating − 0.07 − 0.07 − 0.06 − 0.51 − 0.02 − 0.03 − 0.10 − 0.27 (− 0.88, occasion (− 1.06, (− 0.88, (− 0.75, (− 1.46, (− 0.58, (− 0.58, (− 0.47, 0.35) fre- 0.91) 0.75) 0.86) 0.43) 0.54) 0.52) 0.67) quency   Meal fre- 2.33 (0.06, − 0.13 0.14 (− 1.82, − 0.76 − 0.07 − 0.28 − 0.01 − 0.65 (− 2.15, quency 4.59)* (− 2.03, 2.09) (− 2.86, (− 1.72, (− 1.83, (− 1.55, 0.85) 1.78) 1.33) 1.58) 1.27) 1.53) − 0.08 0.00 (− 0.92, − 0.52 0.03 (− 0.54, 0.07 (− 0.51, 0.15 (− 0.43, − 0.16 (− 0.81,   Snack fre- − 0.62 quency (− 1.73, (− 1.01, 0.92) (− 1.56, 0.61) 0.64) 0.74) 0.49) 0.50) 0.84) 0.52)  Temporal eating pat- terns   Conven- – – – – – – – – tional (refer- ence, n = 1001)   Later 2.13 (− 0.72, 2.38 (− 0.01, 2.55 (0.12, 2.45 (0.05, 1.69 (0.16, 1.56 (0.11, 1.73 (0.27, 1.69 (0.25, lunch 4.97) 4.77) 4.97)* 4.84)* 3.21)* 3.01)* 3.19)* 3.13)* (n = 807)   Grazing 0.08 (− 2.76, 2.41 (− 0.27, 2.20 (− 0.52, 1.93 (− 0.87, 0.98 (− 0.69, 0.85 (− 0.83, 0.64 (− 1.08, 0.50 (− 1.22, (n = 575) 2.93) 5.10) 4.92) 4.72) 2.66) 2.54) 2.35) 2.22) 1 3 European Journal of Nutrition Table 2 (continued) Values are presented as β coefficients (95% confidence intervals). Associations were examined using the Wald tests of associations for linear regression; *P < 0.05, **P < 0.01 Cr ude analysis Adjusted for age (years, continuous), sedentary time (min/day, continuous), education level (low/medium/high), country of birth (Australia/ other mainly English-speaking countries/all other countries), meets PA guidelines (yes/no), smoking status (never smoked/past smoker/current smoker), and dieting (yes/no) Model 2 and additionally adjusted for BMI scores Model 3 and additionally adjusted for Dietary Guideline Index scores and total energy intake Table 3 Associations of eating patterns with hypertension prevalence in Australian men and women a b c d Model 1Model 2Model 3Model 4 Men (n = 2099)  Eating frequency   Eating occasion frequency 0.87 (0.80, 0.97)* 0.85 (0.75, 0.96)** 0.87 (0.76, 0.98)* 0.94 (0.81, 1.08)   Meal frequency 1.03 (0.76, 1.40) 0.88 (0.64, 1.21) 0.92 (0.66, 1.30) 1.12 (0.77, 1.61)   Snack frequency 0.86 (0.76, 0.96)* 0.85 (0.75, 0.96)* 0.86 (0.75, 0.98)* 0.91 (0.79, 1.05)  Temporal eating patterns   Conventional (reference, n = 941) – – – –   Later lunch (n = 702) 0.70 (0.52, 0.93)* 0.78 (0.56, 1.08) 0.81 (0.59, 1.12) 0.82 (0.59, 1.13)   Grazing (n = 456) 0.81 (0.55, 1.20) 1.14 (0.74, 1.74) 1.15 (0.76, 1.75) 1.22 (0.81, 1.82) Women (n = 2383)  Eating frequency   Eating occasion frequency 0.94 (0.83, 1.08) 0.95 (0.83, 1.08) 0.97 (0.84, 1.10) 0.93 (0.79, 1.09)   Meal frequency 1.10 (0.81, 1.50) 0.94 (0.70, 1.27) 0.96 (0.70, 1.31) 0.93 (0.79, 1.09)   Snack frequency 0.91 (0.79, 1.05) 0.95 (0.83, 1.09) 0.95 (0.82, 1.10) 0.93 (0.79, 1.09)  Temporal eating patterns   Conventional (reference, n = 1001) – – – –   Later lunch (n = 807) 1.34 (0.94, 1.90) 1.47 (1.00, 2.16) 1.51 (1.01, 2.25)* 1.49 (1.00, 2.22)*   Grazing (n = 575) 0.97 (0.64, 1.45) 1.11 (0.71, 1.73) 1.06 (0.67, 1.68) 1.04 (0.64, 1.67) Values are presented as odds ratios (95% confidence intervals). Associations were examined using the Wald tests of associations for logistic regression; *P < 0.05, **P < 0.01 Cr ude analysis Adjusted for age (years, continuous), sedentary time (min/day, continuous), education level (low/medium/high), country of birth (Australia/ other mainly English-speaking countries/all other countries), meets PA guidelines (yes/no), smoking status (never smoked/past smoker/current smoker), and dieting (yes/no) Model 2 and additionally adjusted for BMI scores Model 3 and additionally adjusted for Dietary Guideline Index scores and total energy intake time-periods, the present study examined temporal eating higher blood glucose levels, when compared to participants patterns based on a novel latent class analysis approach, who had the same meal in the morning [41]. Insulin metabo- which captures the timing multiple EO across the day and lism may contribute to the association between temporal eat- the likely correlations of energy intakes between EO [4]. ing patterns and BP due to its role in modulating vasodilator The possible mechanisms by which the later timing of a effects on the endothelium via nitric oxide bioavailability “lunch” meal might increase SBP and DBP among women [40]. Measures of insulin sensitivity and blood glucose lev- are unclear. The timing of the “dinner” (evening) meal also els in future epidemiological research examining temporal tended to be later in women with this pattern and research eating patterns and BP are needed. has shown that insulin sensitivity gradually lowers across Strengths of this study include the examination of asso- the day, into the evening [40]. In addition, in an experimen- ciations with BP in a large nationally representative sam- tal trial that controlled for dietary intakes, the timing of an ple, adjusted for BMI, and multiple important confounders, evening meal high in energy and carbohydrates but low in including a measure of overall diet quality. Eating patterns fibre was associated with reduced insulin sensitivity and were determined from 2 days of dietary recall, and an EO was 1 3 European Journal of Nutrition defined using an evidence-based approach [ 27]. While the References novel methodology used to determine temporal eating patterns 1. Collaborators GBDRF. (2017) Global, regional, and national com- is also considered a study strength, it should be noted that parative risk assessment of 84 behavioural, environmental and nd fi ings from data-driven, exploratory methods may not be occupational, and metabolic risks or clusters of risks, 1990–2016: generalizable to populations from other countries. For exam- a systematic analysis for the Global Burden of Disease Study ple, the temporal eating patterns found in the present study 2016. Lancet 390(10100):1345–1422. h t t p s : / / d o i . o rg / 1 0 . 1 0 1 6 / S0140 -6736(17)32366 -8 may reflect the sociocultural and lifestyle characteristics of 2. Estruch R, Ros E, Martinez-Gonzalez MA (2013) Mediterranean the Australian adult population; further research is needed to diet for primary prevention of cardiovascular disease. N Engl J better understand differences in temporal eating patterns in Med 369(7):676–677. https ://doi.org/10.1056/NEJMc 13066 59 other populations. Limitations of the present study include the 3. Appel LJ, Brands MW, Daniels SR, Karanja N, Elmer PJ, Sacks FM, American Heart A (2006) Dietary approaches to prevent cross-sectional study design which does not permit the assess- and treat hypertension: a scientific statement from the Ameri- ment of temporal relationships and the lack of data on anti- can Heart Association. Hypertension 47(2):296–308. https ://doi. hypertensive medications which may have led to an underes- org/10.1161/01.HYP.00002 02568 .01167 .B6 timation of associations with BP. However, in the analysis that 4. St-Onge MP, Ard J, Baskin ML, Chiuve SE, Johnson HM, Kris- Etherton P, Varady K (2017) Meal timing and frequency: implica- excluded participants with no current/previous self-reported tions for cardiovascular disease prevention: a scientific statement hypertension, the association between a “later lunch pattern” from the American Heart Association. Circulation 135(9):e96– with DBP persisted. e121. https ://doi.org/10.1161/CIR.00000 00000 00047 6 In conclusion, a temporal eating pattern distinguished by a 5. Leech RM, Worsley A, Timperio A, McNaughton SA (2015) Understanding meal patterns: definitions, methodology and later “lunch” meal when compared to a “conventional” Aus- impact on nutrient intake and diet quality. Nutr Res Rev 28(1):1– tralian mealtime pattern was associated with higher BP in 21. https ://doi.org/10.1017/S0954 42241 40002 62 women. Among men, a higher snack frequency was inversely 6. Asher G, Sassone-Corsi P (2015) Time for food: the intimate associated with BP, but these associations disappeared after interplay between nutrition, metabolism, and the circadian clock. Cell 161(1):84–92. https ://doi.org/10.1016/j.cell.2015.03.015 adjustment for total energy intake and overall diet quality. 7. Odegaard AO, Jacobs DR Jr, Steffen LM, Van Horn L, Ludwig Future research that considers the role of overall diet quality DS, Pereira MA (2013) Breakfast frequency and development and cardiometabolic indicators (e.g. insulin and glucose levels) of metabolic risk. Diabetes Care 36(10):3100–3106. https ://doi. on the relation between eating patterns and BP is warranted. org/10.2337/dc13-0316 8. Mekary RA, Giovannucci E, Cahill L, Willett WC, van Dam RM, Hu FB (2013) Eating patterns and type 2 diabetes risk in older Acknowledgements A.T. was supported by a National Heart Founda- women: breakfast consumption and eating frequency. Am J Clin tion of Australia Future Leader Fellowship (Award 100046). S.A.M. is Nutr 98(2):436–443. https ://doi.org/10.3945/ajcn.112.05752 1 supported by a National Health and Medical Research Council Career 9. Pot GK, Almoosawi S, Stephen AM (2016) Meal irregularity and Development Fellowship (ID1104636). R.M.L. is supported by an cardiometabolic consequences: results from observational and Alfred Deakin Postdoctoral Research Fellowship. The authors would intervention studies. Proc Nutr Soc 75(4):475–486. https ://doi. like to acknowledge Dr. Katherine Livingstone for the data preparation org/10.1017/S0029 66511 60002 39 of the diet quality score variable used in this study. 10. Deshmukh-Taskar P, Nicklas TA, Radcliffe JD, O’Neil CE, Liu Y (2013) The relationship of breakfast skipping and type of break- Compliance with ethical standards fast consumed with overweight/obesity, abdominal obesity, other cardiometabolic risk factors and the metabolic syndrome in young Research involving human participants and/or animals This research adults. The National Health and Nutrition Examination Survey does not include patient data or data from a clinical trial. This research (NHANES): 1999–2006. Public Health Nutr 16(11):2073–2082. is secondary analysis of Australian (national) Health Survey (AHS) https ://doi.org/10.1017/S1368 98001 20042 96 data commissioned by the Commonwealth of Australia and conducted 11. Sierra-Johnson J, Unden AL, Linestrand M, Rosell M, Sjogren P, by the Australian Bureau of Statistics (ABS). This secondary analysis Kolak M, De Faire U, Fisher RM, Hellenius ML (2008) Eating was registered at anzctr.org.au as ACTRN12617001029381. The web meals irregularly: a novel environmental risk factor for the meta- address of this trial is: http://www .ANZCT R.org.au/ACTRN 12617 bolic syndrome. Obesity (Silver Spring) 16(6):1302–1307. https 00102 9381.aspx. Ethical approval for the ABS to conduct the house- ://doi.org/10.1038/oby.2008.203 hold components of the AHS is provided by the Australian Federal 12. Kim S, Park GH, Yang JH, Chun SH, Yoon HJ, Park MS (2014) Government Census and Statistics Act, 1905 (https ://www.legis latio Eating frequency is inversely associated with blood pressure n.gov.au/Detai ls/C2006 C0017 8). and hypertension in Korean adults: analysis of the Third Korean National Health and Nutrition Examination Survey. Eur J Clin Nutr. https ://doi.org/10.1038/ejcn.2014.9 Conflict of interest On behalf of all authors, the corresponding author 13. Karatzi K, Georgiopoulos G, Yannakoulia M, Efthimiou E, Voi- states that there is no conflict of interest. donikola P, Mitrakou A, Manios E, Alevizaki M, Papamichael C, Stamatelopoulos K (2016) Eating frequency predicts new Open Access This article is distributed under the terms of the Crea- onset hypertension and the rate of progression of blood pressure, tive Commons Attribution 4.0 International License (http://creat iveco arterial stiffness, and wave reflections. J Hypertens. https ://doi. mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- org/10.1097/hjh.00000 00000 00082 2 tion, and reproduction in any medium, provided you give appropriate 14. Karatzi K, Yannakoulia M, Psaltopoulou T, Voidonikola P, Kol- credit to the original author(s) and the source, provide a link to the lias G, Sergentanis TN, Retsas T, Alevizaki M, Papamichael C, Creative Commons license, and indicate if changes were made. 1 3 European Journal of Nutrition Stamatelopoulos K (2014) Meal patterns in healthy adults: inverse report. JAMA 289 (19):2560–2572. https ://doi.or g/10.1001/ association of eating frequency with subclinical atherosclerosis jama.289.19.2560 indexes. Clin Nutr. https ://doi.org/10.1016/j.clnu.2014.04.022 27. Leech RM, Worsley A, Timperio A, McNaughton SA (2015) 15. Bhutani S, Varady KA (2009) Nibbling versus feasting: which Characterizing eating patterns: a comparison of eating occa- meal pattern is better for heart disease prevention? Nutr Rev sion definitions. Am J Clin Nutr 102(5):1229–1237. https ://doi. 67(10):591–598. https://doi.or g/10.1111/j.1753-4887.2009.00231 org/10.3945/ajcn.115.11466 0 .x 28. Leech RM, Worsley A, Timperio A, McNaughton SA (2017) 16. Titan SM, Bingham S, Welch A, Luben R, Oakes S, Day N, Khaw Temporal eating patterns: a latent class analysis approach. Int KT (2001) Frequency of eating and concentrations of serum cho- J Behav Nutr Phys Act 14(1):3. h tt p s : / / do i . o rg/ 1 0 . 1 1 86 / s 1 2 9 6 lesterol in the Norfolk population of the European prospective 6-016-0459-6 investigation into cancer (EPIC-Norfolk): cross sectional study. 29. Lo Y, Mendell NR, Rubin DB (2001) Testing the number of com- BMJ 323(7324):1286–1288 ponents in a normal mixture. Biometrika 88(3):767–778. https :// 17. Edelstein SL, Barrett-Connor EL, Wingard DL, Cohn BA (1992) doi.org/10.1093/biome t/88.3.767 Increased meal frequency associated with decreased cholesterol 30. Australian Institute of Health and Welfare (2003) The Active Aus- concentrations; Rancho Bernardo, CA, 1984–1987. Am J Clin tralia Survey: a guide and manual for implementation, analysis Nutr 55(3):664–669 and reporting. Cat. no. CVD 22. AIHW, Canberra 18. Leech RM, Livingstone KM, Worsley A, Timperio A, 31. The Department of Health (2014) Australia’s physical activity and McNaughton SA (2016) Meal frequency but not snack frequency sedentary behaviour guidelines [Webpage on the Internet]. DoH, is associated with micronutrient intakes and overall diet quality Canberra (cited 26 Nov 2017) in Australian men and women. J Nutr 146(10):2027–2034. https 32. National Health and Medical Research Council (2013) Australian ://doi.org/10.3945/jn.116.23407 0 dietary guidelines. NHMRC, Canberra 19. Murakami K, Livingstone MB (2015) Associations between meal 33. Thorpe MG, Milte CM, Crawford D, McNaughton SA (2016) A and snack frequency and diet quality and adiposity measures in revised Australian dietary guideline index and its association with British adults: findings from the National Diet and Nutrition Sur - key sociodemographic factors, health behaviors and body mass vey. Public Health Nutr 19(9):1624–1634. https: //doi.org/10.1017/ index in peri-retirement aged adults. Nutrients 8(3):160. https :// S1368 98001 50029 79doi.org/10.3390/nu803 0160 20. Leech RM, Worsley A, Timperio A, McNaughton SA (2017) The 34. Livingstone KM, McNaughton SA (2016) Diet quality is associ- role of energy intake and energy misreporting in the associations ated with obesity and hypertension in Australian adults: a cross between eating patterns and adiposity. Eur J Clin Nutr. https: //doi. sectional study. BMC Public Health 16(1):1037. https ://doi. org/10.1038/ejcn.2017.90org/10.1186/s1288 9-016-3714-5 21. Almoosawi S, Prynne CJ, Hardy R, Stephen AM (2013) Time- 35. Banna JC, McCrory MA, Fialkowski MK, Boushey C (2017) of-day of energy intake: association with hypertension and blood Examining plausibility of self-reported energy intake data: con- pressure 10 years later in the 1946 British Birth Cohort. J Hyper- siderations for method selection. Front Nutr 4:45. https ://doi. tens 31(5):882–892 org/10.3389/fnut.2017.00045 22. Keller K, Rodriguez Lopez S, Carmenate Moreno M (2017) Asso- 36. Vainik U, Konstabel K, Latt E, Maestu J, Purge P, Jurimae J ciation between meal intake behavior and blood pressure in Span- (2016) Diet misreporting can be corrected: confirmation of the ish adults. Nutr Hosp 34(3):654–660. https ://doi.org/10.20960 / association between energy intake and fat-free mass in adoles- nh.487 cents. Br J Nutr 116(8):1425–1436. https://doi.or g/10.1017/s0007 23. Leech RM, Timperio A, Livingstone KM, Worsley A, 11451 60033 17 McNaughton SA (2017) Temporal eating patterns: associations 37. Folsom AR, Parker ED, Harnack LJ (2007) Degree of concord- with nutrient intakes, diet quality, and measures of adiposity. ance with DASH diet guidelines and incidence of hypertension Am J Clin Nutr 106(4):1121–1130. h t t p s : / / d o i . o r g / 1 0 . 3 9 4 5 / and fatal cardiovascular disease. Am J Hypertens 20(3):225–232. ajcn.117.15658 8https ://doi.org/10.1016/j.amjhy per.2006.09.003 24. Eicher-Miller HA, Khanna N, Boushey CJ, Gelfand SB, Delp EJ 38. Almoosawi S, Vingeliene S, Karagounis LG, Pot GK (2016) (2016) Temporal dietary patterns derived among the adult par- Chrono-nutrition: a review of current evidence from observational ticipants of the National Health and Nutrition Examination Sur- studies on global trends in time-of-day of energy intake and its vey 1999–2004 are associated with diet quality. J Acad Nutr Diet association with obesity. Proc Nutr Soc 75(4):487–500. https :// 116(2):283–291. https ://doi.org/10.1016/j.jand.2015.05.014doi.org/10.1017/S0029 66511 60003 06 25. Australian Bureau of Statistics (2013) Australian health survey: 39. Almoosawi S, Prynne CJ, Hardy R, Stephen AM (2013) Time- users’ guide, 2011–2013. Cat. no. 4363.0.55.001. ABS, Canberra of-day and nutrient composition of eating occasions: prospective 26. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green association with the metabolic syndrome in the 1946 British birth LA, Izzo JL Jr, Jones DW, Materson BJ, Oparil S, Wright JT cohort. Int J Obes 37(5):725–731 Jr, Roccella EJ, National Heart, Lung, and Blood Institute Joint 40. Steinberg HO, Baron AD (2002) Vascular function, insulin resist- National Committee on Prevention, Detection, Evaluation, and ance and fatty acids. Diabetologia 45(5):623–634. h t t p s : / / d o i . Treatment of High Blood Pressure, National High Blood Pressure org/10.1007/s0012 5-002-0800-2 Education Program Coordinating Committee (2003) The seventh 41. Morgan LM, Shi J-W, Hampton SM, Frost G (2012) Effect of report of the Joint National Committee on Prevention, Detection, meal timing and glycaemic index on glucose control and insulin Evaluation, and Treatment of High Blood Pressure: the JNC 7 secretion in healthy volunteers. Br J Nutr 108(7):1286–1291 1 3

Journal

European Journal of NutritionSpringer Journals

Published: Jun 6, 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