|Home | About | Journals | Submit | Contact Us | Français|
Dietary intake of various fats may have different effects on blood pressure. We conducted a prospective cohort study to examine the association between intake of subtype and individual fatty acids (FAs) and the risk of developing hypertension among 28,100 US women aged ≥39 years and free of cardiovascular disease and cancer. Baseline intake of FAs was assessed using semiquantitative food frequency questionnaires. Incident hypertension was identified from annual follow-up questionnaires based on self-reported physician diagnosis, medication use, and blood pressure levels. A total of 13,633 women developed incident hypertension during 12.9 years of follow-up. After adjusting for demographic, lifestyle, and other dietary factors, intake of saturated FAs (SFAs), monounsaturated FAs (MUFAs), trans-unsaturated FAs (trans FAs) was positively associated with the risk of hypertension. The multivariable relative risks (RRs) and 95% confidence intervals (CIs) of hypertension in the highest compared to the lowest quintile of intake were 1.12 (1.05–1.20) for SFAs, 1.11 (1.04–1.18) for MUFAs, and 1.15 (1.08–1.22) for trans FAs. After additional adjustment for body mass index and history of diabetes and hypercholesterolemia, these associations were attenuated and remained statistically significant only for trans FAs (RR in the highest quintile: 1.08, 95% CI: 1.01–1.15). Intake of polyunsaturated FAs (PUFAs), including ω3 and ω6 PUFAs, was not significantly associated with the risk of hypertension. In conclusion, higher intake of SFAs, MUFAs, and trans FAs was each associated with increased risk of hypertension among middle-aged and older women, whereas only association for trans FAs remained statistically significant after adjustment for obesity-related factors.
Dietary fat is an important modifiable risk factor for hypertension.1 Interventions that reduce total fat intake can effectively lower systolic and diastolic blood pressure (BP).2–4 Recent research further suggest that subtypes of fat with different molecular structure may have different effects on BP.5, 6 Investigations using animal models have shown that diets high in saturated fats increase BP7, 8 while diets enriched with ω3 polyunsaturated fats protect against induced BP elevations.9–11
In previous epidemiologic studies, BP has been shown positively correlated with saturated fats intake12, 13 and inversely correlated with monounsaturated13 and polyunsaturated fats intake.14, 15 Only a few studies have examined the prospective association between dietary fat intake16, 17 or its biomarkers in blood18, 19 and the risk of developing hypertension, with inconsistent results reported. In randomized trials, reduction of saturated fats intake alone did not seem to affect BP;20 increase of polyunsaturated (particularly ω3) fats intake can lower BP, whereas the effect was primarily found among hypertensive but not normotensive individuals.21–23
To better understand the different roles of various dietary fats in the development of hypertension, we conducted a prospective analysis in a large cohort of middle-aged and older US women to investigate the association between baseline intake of saturated fatty acids (SFAs), monounsaturated FAs (MUFAs), polyunsaturated FAs (PUFAs), and trans-unsaturated FAs (trans FAs), and the subsequent risk of hypertension during long-term follow-up.
The Women’s Health Study (WHS) is a randomized, double-blind, placebo-controlled, 2×2 factorial trial evaluating the risks and benefits of low-dose aspirin and vitamin E in the primary prevention of cardiovascular disease and cancer.24, 25 A third component, β-carotene, was initially included in the trial but terminated after a median treatment of 2.1 years.26 Written informed consent was obtained from all participants. The trial and ongoing cohort follow-up was approved by the institutional review board of Brigham and Women’s Hospital, Boston, MA (please see http://hyper.ahajournals.org). From September 1992 to May 1995, 39,876 female US health professionals, aged ≥39 years and free from cardiovascular disease and cancer (except non-melanoma skin cancer), were randomized into the WHS. Of the 39,876 women randomized, 39,310 completed a 131-item validated semiquantitative food frequency questionnaire (FFQ). For this study, we excluded 10,751 women who had hypertension at baseline, defined as having a self-reported physician diagnosis of hypertension, self-reported current systolic BP ≥140 mmHg or diastolic BP ≥90 mmHg, or use of antihypertensive treatment. We also excluded 829 women who reported implausible total daily energy intake, 21 women who provided incomplete information on the FFQ, and 41 women who had pre-randomization cardiovascular disease or cancer. After these exclusions, a baseline population of 28,100 women remained for analysis.
On the baseline FFQ, a commonly used unit or portion size was specified for each food item. Participants were asked how often they had consumed that amount, on average, during the previous year. Nine possible responses ranging from “never or less than once per month” to “6+ per day” were recorded. Nutrient intake was computed by multiplying the intake frequency of each unit of food by the nutrient content of the specified portion size according to food composition tables from the US Department of Agriculture,27 supplemented with information obtained from the manufacturers and published reports. The estimation of fat intake also took into account the types of fat or oil used during food preparation. The trans isomers contents of unsaturated fats were estimated based on the method proposed by Sacks and Willett.28 Intake of SFAs, MUFAs, PUFAs, and trans FAs was each calculated as sum of the respective individual FAs. All FAs intake presented in the current study have adjusted for total energy intake using the residual method.29 In similar cohorts of health professionals, Pearson’s correlation coefficients comparing responses from the FFQ with those from four 1-week dietary records spaced over a year were 0.70 for SFAs, 0.69 for MUFAs, and 0.64 for PUFAs.30, 31 FAs intake estimated from the FFQ also correlated well with FA composition of the adipose tissue (Spearman’s r=0.51 for trans FAs and 0.48 for ω3 PUFAs).32
Incident hypertension was ascertained from annual follow-up questionnaires by meeting at least 1 of 4 criteria: a new physician diagnosis of hypertension; newly initiated antihypertensive treatment; self-reported current systolic BP ≥140 mmHg; or self-reported current diastolic BP ≥90 mmHg. Women reported the month and year of hypertension diagnosis. For incident cases missing dates of physician diagnosis or defined by other criteria, time of event was assigned by randomly selecting a date between the questionnaires with and without hypertension. Individuals who developed cardiovascular disease during follow-up, the management of which may affect BP, were censored on the date of cardiovascular disease diagnosis. In health professionals, self-reported BP correlates well with measured systolic BP (r=0.72) and diastolic BP (r=0.60),33 and the validity of self-reported hypertension is high.34 In a random sample of WHS participants, self-reported incident hypertension was confirmed in 48 of 50 (96%) women and absence of hypertension was confirmed in 45 of 50 (90%) women through telephone interviews.
Statistical analyses were performed using SAS software (SAS Institute, Cary, NC, USA) version 9.1. Intake of FAs was divided into quintiles. Distribution of hypertension risk factors was compared across quintiles of FAs intake. Person-years of follow-up were calculated for each participant from randomization to the date of incident hypertension, the last day in the study, or 29 February 2007, whichever came first. After verifying the assumption of constant proportional hazard, we used Cox regression model to estimate the hazard ratio (presented as relative risk, RR) and 95% confidence interval (CI) of hypertension across quintiles of FAs intake, with the lowest quintile as the reference. Models first adjusted for age, race, total energy intake, and randomized treatment assignment; then additionally adjusted for lifestyle factors including smoking, total alcohol intake, physical activity, postmenopausal status, postmenopausal hormone use, and several hypertension-related nutritional factors including dietary intake of calcium, potassium, sodium, and fiber (multivariable model 1); and finally adjusted for obesity-related metabolic factors that may also serve as intermediate factors linking dietary FAs to hypertension development, including body mass index (BMI), history of diabetes, and history of hypercholesterolemia (multivariable model 2). Analyses were further stratified by known hypertension risk factors including age (<55, ≥55 years), BMI (<25, ≥25 kg/m2), smoking status (current, non-current), alcohol intake (none, any), and physical activity (<600, ≥600 kcal/week). We also performed several sensitivity analyses. First, we repeated all analyses with dietary FAs represented as the percentage of total energy intake; second, we considered incident hypertension using alternative definitions, such as self-reported elevated BP only, physician diagnosis or antihypertensive treatment only, or multiple indications. The results of sensitivity analyses were similar to main analyses (data not shown).
Among 28,100 women free of hypertension at baseline, energy-adjusted FA intake ranged from 2.55 to 51.4 g/d for SFAs, 3.39 to 52.0 g/d for MUFAs, 2.10 to 40.5 g/d for PUFAs, and 0.01 to 12.4 g/d for trans FAs. SFAs and MUFAs intake was highly correlated (Pearson r=0.78). PUFAs Intake was moderately correlated with MUFAs intake (r=0.55) and weakly but significantly correlated with SFAs intake (r=0.26). Trans FAs intake was moderately correlated with SFAs (r=0.52) and MUFAs (r=0.63) intake and weakly correlated with PUFAs intake (r=0.31).
Table 1 shows the baseline characteristics of participants by quintiles of SFAs, MUFAs, and PUFAs intake. For all three FA subtypes, women consuming greater amounts were heavier, more likely to be current smokers, less physically active, and had lower alcohol intake. Women consuming more SFAs and MUFAs were younger, less likely to be postmenopausal and use postmenopausal hormones, less likely to have hypercholesterolemia, while more likely to be diabetic. Intake of all three FA subtypes also was positively associated with intake of sodium, and inversely associated with intake of calcium, potassium, and fiber. Baseline systolic and diastolic BP increased with increasing intake of all three FA subtypes. The associations of hypertension risk factors with trans FAs intake were similar to the associations with SFAs and MUFAs intake (data not shown).
A total of 13,633 cases of incident hypertension were identified during an average of 12.9 years follow-up, with 2,427 cases identified by elevated systolic or diastolic BP only, 61 cases identified by physician diagnosis only, 1,540 cases identified by newly initiated antihypertensive medications only, and the remaining cases identified by multiple indications. After adjusting for age, race, total energy intake, and randomized treatment, the risk of hypertension significantly increased across increasing intake of all FA subtypes except ω3 PUFAs. Additional djustment for lifestyle factors and nutritional factors (multivariable model 1) attenuated these associations. Intake of SFAs, MUFAs, and trans FAs remained significantly and positively associated with risk of hypertension, while intake of PUFAs was only marginally significantly associated with risk of hypertension (Table 2). After further adjustment for BMI, diabetes, and hypercholesterolemia (multivariable model 2), the associations for SFAs, MUFAs, and PUFAs intake were all attenuated to null, only the positive association for trans FAs intake remained significant, with the fully adjusted RRs of hypertension across increasing quintiles of 1.00, 1.04, 1.07, 1.06, and 1.08 (p, trend: 0.04). There was an inverse association of polyunsaturated-to-saturated FAs ratio (P to S ratio) and a positive association of ω6 to ω3 PUFAs ratio with risk of hypertension in the reduced model (data not shown), but these associations were no longer significant in the multivariable models (Table 2).
We further stratified the analyses by baseline age (Figure 1). The positive associations of SFAs, MUFAs, and trans FAs intake with risk of hypertension all appeared to be stronger for women aged <55 years than for older women, albeit the test for age-related interactions reached statistical significance only for trans FAs. The multivariable model 2 RRs of hypertension across increasing quintiles of trans FAs intake among women aged 39-<55 years were 1.00, 1.11, 1.12, 1.14, and 1.21 (p, trend: <0.0001). The corresponding RRs among women aged 55–89 years were 1.00, 0.95, 1.01, 0.96, and 0.91 (p, trend: 0.09). The associations between dietary FAs and hypertension risk only slightly varied by baseline BMI, smoking status, alcohol consumption, and physical activity (all p, interaction > 0.05).
The associations of major individual FA intake with risk of hypertension largely followed the associations for their respective FA subtypes (Table S1, please see http://hyper.ahajournals.org). Among women younger than 55 years, intake of individual SFA (including 16:0 and 18:0), MUFA (including 16:1 and 18:1), PUFA (including 20:4ω6 and 22:6ω3), and trans FA (including trans 16:1, trans 18:1, and trans 18:2) was each positively associated with risk of hypertension in multivariable model 1. These associations were attenuated in multivariable model 2, with the associations for 18:1, 20:4ω6, and trans 16:1 no longer significant. Among women aged ≥55 years, intake of individual FA was generally not associated with the risk of hypertension.
In this large-scale prospective cohort study of middle-aged and older women, intake of SFAs, MUFAs, and trans FAs was each positively associated with risk of hypertension. The associations for SFAs and MUFAs were largely attenuated by adjustment for potential intermediate factors including BMI, diabetes, and hypercholesterolemia, while the associations for trans FAs remained significant even after these adjustments. Intake of PUFAs was generally not associated with risk of hypertension.
Various fats potentially have different effects on BP. Experimental studies found that feeding rats with SFAs resulted in impaired endothelial function7 and enhanced sympathetic nervous system activities,8 which will increase BP. In contrast, consumption of long-chain ω3 PUFAs modulated plasma phospholipid composition and cell membrane fluidity, increased the production of vasodilators, and reduced cardiac adrenergic activity,5, 35 which will lower BP. The incorporation of ω6 PUFAs into cell membrane changed the balance between vasoconstrictors and vasodilators,5 the subsequent net effects on BP have varied in different animal models.36, 37 Similarly, MUFAs also modify membrane phospholipid composition38 and vascular reactivity,39 the net effects may either raise or lower BP. Direct effects of trans FAs on BP remain largely unclear. Due to the lack of flexible structure of their parent unsaturated FAs, trans FAs display biological features more similar to SFAs.40 Furthermore, since trans FAs compete with other unsaturated FAs for enzymatic desaturation, the presence of trans FAs may increase the demand for essential PUFAs.41
Previous prospective studies linking FAs intake with incident hypertension have yielded inconsistent results. In the Nurses’ Health Study,16 a cohort of 121,700 US women aged 34–59 years, and the Health Professionals Follow-up Study,17 a cohort of 51,529 US men aged 40–75 years, no association was found between baseline intake of SFAs, MUFAs, PUFAs, or trans FAs assessed from FFQ and incident hypertension during a short follow-up of 4 years. Two studies investigated the association between plasma FAs, as biomarker for dietary FAs, and risk of hypertension.18, 19 The Atherosclerosis Risk in Communities study found that higher levels of SFAs and MUFAs and lower levels of PUFAs and P to S ratio in baseline plasma phospholipids and cholesterol ester were significantly associated with an increased 6-year incidence of hypertension.18 The Uppsala Longitudinal Study of Adult Men showed that the baseline plasma levels of SFA 16:0 and 18:0 were significantly higher and the plasma level of PUFA 18:2ω6 was significantly lower in men who developed sustained hypertension compared with men who remained normotensive or had white-coat hypertension.19 Both studies did not examine trans FAs in plasma. In randomized trials, dietary interventions that lowered saturated fat intake alone did not significantly affect BP;20 diets rich in MUFAs reduced BP levels42 or antihypertensive medication required for hypertensive patients in some,43 but not all20 studies; ω3 PUFAs supplementation lowered BP in hypertensive individuals but not in normotensives, and the substantial BP reductions usually occurred at relatively high doses (≥3 g/d);21–23 and ω6 PUFAs (mainly 18:2ω6) supplements have not demonstrated apparent effects on BP.20 We are not aware of any clinical trial that tested the effect of reducing dietary trans fats on BP or risk of developing hypertension.
In our study, the positive associations of SFAs, MUFAs, and trans FAs intake with risk of hypertension were substantially attenuated after adjustment for BMI. This finding underscored a potential confounding effect of obesity in the dietary fat and hypertension relation. Alternatively, since dietary fat is a major source of energy that may contribute to the development of obesity, a strong impact of BMI in the FA-hypertension relation also possibly reflects that obesity and its related pathophysiological processes are important intermediate steps linking fat intake to BP change. Another intriguing finding in our study is that the associations between subtype and individual FAs intake with risk of hypertension have appeared stronger among younger women versus older women. However, the test for age-related interactions only reached statistical significance for trans FAs. The lack of a clear association between intake of ω3 PUFA, ω6 PUFA, and ω6 to ω3 ratio with risk of hypertension, though consistent with findings from previous studies that measured plasma FAs as biomarker for dietary FAs,18, 19 was unexpected. Although in vivo and in vitro experimental studies demonstrated potential effects of PUFAs on BP control, the amount of PUFAs intake in our study population is probably insufficient to strongly affect the risk of developing hypertension among initially normotensive individuals. Another possible explanation for the lack of association for PUFAs is residual confounding by unknown factors that correlate with high PUFAs intake or supplement use.
Strengths of the current study include the prospective study design, large sample size, and minimal loss to follow up. However, several limitations of this study also deserve comment. First, the assessment of dietary FAs intake and ascertainment of incident hypertension are based on self-reported information in our study.
Nevertheless, the validity and reproducibility of FFQ as a measure of long-term dietary intake30–32 and the accuracy of self-reported hypertension in health professionals33, 34 have been demonstrated in previous validation studies. For lifestyle and clinical covariates, including BMI, a single baseline assessment is subject to misclassification considering possible change during follow-up, while random misclassification will typically lead to an underestimation of true association. Second, since correlations between intake of SFAs, MUFAs, and trans FAs were moderate to high in our study, we can not completely separate their effects from each other. Third, despite comprehensive adjustment for multiple lifestyle and clinical factors, residual confounding by unmeasured or imprecisely measured hypertension risk factors may persist. Fourth, although our large sample size allowed us to assess the association between dietary FAs and risk of hypertension in various subgroups, the potential for false positive findings imply that we should interpret any significant findings in subgroup analysis with caution. Finally, WHS participants are predominantly white female health professionals, which minimized potential confounding by race/ethnicity and socioeconomic factors but also may limit study generalizability. Yet similar associations noted in prior community-based studies18, 19 suggest that our results may indeed be generalizable to other populations.
Current dietary guidelines for US adults recommend total fat intake between 20% to 35% of calories, saturated FAs intake less than 10% of calories, and trans FAs consumption as low as possible.44 Epidemiologic evidence on the relevance of such dietary recommendations to the prevention of hypertension is surprisingly limited. With total fat intake as a percentage of calories falling over recent decades,45 more specific recommendations on the optimal amount and type of FAs intake for hypertension and cardiovascular disease prevention are needed. Our study comprehensively examined the association between intake of subtype and individual FAs and risk of hypertension. Our findings support the importance of current dietary fat recommendations for middle-aged and older women. Our findings also suggest that an adverse diet profile, along with other unhealthy lifestyle, may increase the risk of hypertension through promoting obesity. More studies are warranted to further elucidate the interrelation between dietary fat, the development of obesity, and the pathogenesis of hypertension.
We are indebted to the 39,876 participants in the Women's Health Study for their dedicated and conscientious collaboration, and to the entire staff of the Women's Health Study for their assistance in designing and conducting the trial.
Sources of Funding:
This study was supported by a scientist development grant funded by American Heart Association (0735390N) and research grants CA047988, HL043851, and HL080467 from the National Institutes of Health, Bethesda, MD. These grants provided funding for study conduct, data collection, and data analyses.
Dr. Wang received a Scientist Development Grant from American Heart Association.