Abstract
The objective of this study was to determine whether consumption of 100% fruit juice as compared to whole fruit is associated with increased risk of hypertension or diabetes. We analyzed postmenopausal women in the United States enrolled in the Women’s Health Initiative between 1993 and 1998. Whole fruit and 100% fruit juice intake were assessed by baseline food frequency questionnaire. Standardized questionnaires assessed outcomes every 6–12 months during a mean 7.8 years of follow-up. Cox regression estimated hazard ratios (HR) and 95% confidence intervals (CI) for incident hypertension (n= 36,314 incident cases/80,539 total participants) and diabetes (n=11,488 incident cases/114,219 total participants). In multivariable analyses there was no significant association comparing the highest to lowest quintiles of 100% fruit juice consumption (8 ounces/day compared to none) and incident hypertension (HR 1.00, 95% CI 0.97–1.03) or diabetes (HR 0.96, 95% CI 0.90–1.03). There was also no significant association between whole fruit consumption (2.4 servings/day compared to 0.3 servings/day) and incident hypertension (HR 1.02, 95% CI 0.98–1.05) or diabetes (HR 1.03, 95% CI 0.96–1.10). Consuming moderate amounts of 100% fruit juice or whole fruit was not significantly associated with risk of hypertension or diabetes among postmenopausal US women.
Introduction
One hundred percent fruit juice is rich in nutrients, such as polyphenols, but it is also high in naturally occurring sugars and may be associated with adverse cardiometabolic health effects, such as hypertension or diabetes.i,ii Experts disagree on whether 100% fruit juice should be included as a harmful beverage in healthy beverage policies such as taxes on sugary beverages, food warning labels, and traffic light food labeling programs that seek to prevent obesity and diabetes.iii,iv Policy makers need guidance as to where 100% fruit juice falls in the health spectrum of beverage options, and to understand the cardiometabolic health effects of 100% fruit juice versus whole fruit consumption.
One hundred percent fruit juice may be associated with hypertension and diabetes through several possible biological mechanisms: (1) Increased energy intake and weight gain;v (2) 100% fruit juices have moderately high glycemic indices,vi and consumption of large portions of 100% fruit juice may cause a high postprandial insulin response that predisposes to diabetes through a mechanism independent of weight gain;vii (3) Metabolism of the fructose in 100% fruit juice may increase uric acid production, which is associated with both elevated blood pressure and insulin resistance.viii,ix
Despite the theoretical impact of a diet high in 100% fruit juice on cardiometabolic health, meta-analyses have found limited evidence that 100% fruit juice consumption is associated with changes in blood pressure or incident diabetes.x,xi,xii Liu et al. conducted a meta-analysis of randomized controlled trials (RCTs) analyzing the effect of consuming 4–17 oz. per day (/d) of 100% fruit juice, compared to water or non-caloric beverages, on short-term changes in blood pressure.10 They found slightly lower systolic blood pressure (−2.0 mmHg; 95% CI −4.5 mmHg, 0.4 mmHg) and diastolic blood pressure (−2.1 mmHg; 95% CI −3.8 mmHg, −0.4 mmHg) in the group that consumed 100% fruit juice. However, these findings may be systematically biased because the included RCTs had short follow-up durations (5–12 weeks) and participants in some RCTs had baseline chronic diseases that affect blood pressure. Regarding diabetes, two meta-analyses of prospective cohort studies found evidence that consuming 100% fruit juice was associated with a small increase in diabetes risk in adults (relative risk < 1.1), but these findings were unstable in sensitivity analyses.11,12 Studies included in these two meta-analyses did not examine a population of postmenopausal women, who are at higher risk of diabetes compared to premenopausal women.xiii,xiv,xv
The question of how much 100% fruit juice adults should drink is important to consumers and policy makers. We undertook this secondary analysis of the Women’s Health Initiative (WHI) to determine if consumption of fruit juice, compared to whole fruit, is associated with incident hypertension or diabetes in a large and diverse sample of postmenopausal women at higher risk for these outcomes. We selected whole fruit as a comparison group because the question of whether 100% fruit juice and whole fruit have different associations with cardiometabolic disease risk is controversial. We hypothesized that consumption of 100% fruit juice, but not whole fruit, would be associated with a small increase in hypertension and diabetes risk.
Methods
Study Design and Population
The design and methods of the WHI have been described elsewhere.xvi Briefly, the WHI enrolled 161,808 postmenopausal women ages 50–79 years between 1993–1998 into the WHI Observational Study (OS) and 3 RCTs. We analyzed participants in the OS (n = 93,679) and comparison arm of the Dietary Modification Clinical Trial (DM CT, n = 29,294). Participants in the comparison arm of the DM CT did not receive the low-fat, high fruit and vegetable dietary modification intervention.
Our inclusion criterion was baseline food frequency questionnaire (FFQ) completion. Our exclusion criteria were (1) Energy intake outliers on baseline FFQ (defined as ≤ 600 kcal/day or ≥ 5000 kcal/day),xvii (2a) baseline self-reported past or current hypertension for the analyses of incident hypertension, (2b) baseline self-reported past or current diabetes (not counting gestational diabetes) for the analyses of incident diabetes, and (3) missing answers to the two 100% fruit juice questions on the FFQ (Supplementary Tables A and B).
Exposure Assessment
Our exposure of interest was 100% fruit juice consumption, as measured by a semi-quantitative FFQ designed for the WHI.xviii FFQs were administered at baseline to all participants. The FFQ contained 2 separate items asking about (1) 100% orange and grapefruit juices, and (2) all other 100% fruit juice types. Participants were asked to specify their usual serving size as small (3 oz.), medium (6 oz.), or large (12 oz.), and to indicate the frequency of intake. To increase statistical power and reduce measurement error by constraining outliers, the exposure was parametrized as quintiles of 100% fruit juice/d. Twelve percent of participants reported no 100% fruit juice consumption, and therefore we modified the first quintile to only include participants who drank no 100% fruit juice. WHI FFQ estimates are reproducible over time (the mean correlation coefficient for repeated assessment of individual nutrient or food groups is 0.76), and levels of misclassification are modest (means estimated by the FFQ for most nutrient and food groups are within 10% of means estimated by food records).18 We standardized intake of 100% fruit juice intake and whole fruit to 2000 kilocalories (kcal)/d using the residual method.19
Covariate Assessment
Data on demographic characteristics, medical history, and health behaviors were collected at baseline using standardized questionnaires.xix Physical measurements were measured by trained and certified study personnel at baseline using standardized protocols and calibrated equipment.
Outcome Assessment
Our primary outcomes were self-reported incident hypertension or diabetes. Standardized medical history questionnaires asking about new treatment of hypertension and diabetes were completed every 6–12 months until the conclusion of these studies in March 2005. Participants were considered to have incident hypertension or diabetes if they initiated medication to treat hypertension (“pills for high blood pressure”) or diabetes (“pills or insulin shots for diabetes”). Data from a WHI data confirmation study showed that incident diabetes as measured by this question was consistent with medication inventories and fasting plasma glucose levels.23 Nearly 80% of the participants at baseline who self-reported treatment with either insulin or oral medication for diabetes had a diabetes medication in their baseline medication inventory, and nearly 100% of women without self-reported treated diabetes had no diabetes medication in their baseline inventory.xx Self-reported incident hypertension has been used in other WHI analyses.xxi
Statistical Analysis
We used Cox proportional hazards to estimate the univariate and multivariable-adjusted hazard ratio (HR) of incident hypertension and diabetes. Time at risk was calculated from baseline to the date of first diagnosis reported on follow-up questionnaire or among those disease-free, until the participants’ last follow-up visit (April 2004 – March 2005). The proportional-hazards assumption was not rejected based on Schoenfeld residuals. The hazard ratios estimated for quintiles II – V compare the participants in each upper quintile to the participants in the first quintile. Likewise, in exploratory analyses of higher levels of consumption that expressed exposures in terms of servings/time, the higher categories of consumption are each compared to the lowest category of consumption (< 4 servings/week).
A median approach was used to examine the trend across quintiles. The median oz./d of 100% fruit juice in one quintile was assigned to all participants in that quintile. The statistical significance of the trend was tested with Cox regression using the mean intake value as a continuous independent variable. A median approach was also used to examine the trend across the categories of consumption in exploratory analyses.
Baseline variables considered as potential confounders were age (years), education level (4 levels), race/ethnicity (Asian/Pacific, Black, Hispanic/Latino, White; categories mutually exclusive), smoking status (never, past, current), number of alcoholic drinks/week, leisure-time physical activity (metabolic equivalent [MET]-hours/week), body mass index (BMI, as a continuous variable), mean daily sodium intake (mg/d), total energy intake (kcal/d), Healthy Eating Index (HEI) score,xxii WHI study arm (OS or DM CT control-arm), and use of postmenopausal hormone replacement therapy (never, past, current). Selected variables derived from the FFQ that were associated with < 10% change in the regression coefficient comparing the highest to lowest quintiles for 100% fruit juice consumption were dropped from the final regression models for parsimony (alcoholic drinks/week, mean sodium/d, and HEI score). All analyses were conducted with Stata (version 14, StataCorpLP, College Station, Texas). A two-sided p-value < 0.05 was considered statistically significant for all analyses.
We performed several exploratory analyses to determine whether hypertension or diabetes risk varied between subgroups of postmenopausal women. We examined the influence of (1) stratifying by decade of age; (2) stratifying by study arm; (3) stratifying by baseline blood pressure category; (4) stratifying by race/ethnicity; (5) classifying the exposure as tertiles and quartiles; (6) performing change analysis as described Smith et al.,xxiii which measured change in 100% fruit juice consumption and change in incident hypertension and diabetes over the same 3-year time period; (7) using cutpoints defined by Borgi et al.xxiv to categorize the 100% fruit juice exposure (≤ 4 servings/week; 5–6 servings/week; 1 serving/d; 2–3 servings/d; ≥ 4 servings/d); (8) sub-classifying the exposures of 100% fruit juice and whole fruit into categories of citrus and non-citrus; and (9) restricting the analytic sample to the highest quintile of 100% fruit juice consumers.
Ethics
The WHI study protocol was approved by Institutional Review Boards (IRBs) at each participating institution, and all participants provided written informed consent.
Results
The baseline characteristics of the 80,539 participants analyzed for incident hypertension and 114,219 participants analyzed for incident diabetes are reported in Tables 1 & 2. More participants in the highest quintile of 100% fruit juice consumption (median 8 oz./d of 100% fruit juice) were older, had a normal BMI, were African American, had higher educational attainment, and had a higher HEI diet quality score.
Table 1.
Baseline Characteristics According to 100% Fruit Juice Consumption of 80,539 Postmenopausal US Women Analyzed for Hypertension Risk
| Modified Quintiles of 100% Fruit Juice Consumption | |||||
|---|---|---|---|---|---|
| I | II | III | IV | V | |
| Median oz./d of 100% fruit juice | 0 | 1.0 | 2.6 | 4.9 | 8.0 |
| Range oz./d of 100% fruit juice | 0 | 0.06–1.7 | 1.8–3.7 | 3.8–6.5 | 6.6–36.8 |
| N | 10,317 | 17,556 | 17,555 | 17,556 | 17,555 |
| Age, years | |||||
| 50–59 | 37.1% | 41.7% | 36.8% | 33.8% | 35.7% |
| 60–69 | 44.3% | 42.6% | 43.7% | 44.1% | 42.7% |
| 70–79 | 18.5% | 15.8% | 19.5% | 22.1% | 21.6% |
| BMI category, kg/m2 | |||||
| Normal (18.5 – 24.9) | 40.0% | 37.6% | 41.3% | 43.6% | 47.9% |
| Overweight (25.0 – 29.9) | 34.6% | 35.7% | 34.7% | 34.7% | 32.7% |
| Obese (> 30.0) | 23.9% | 25.9% | 22.6% | 20.6% | 16.2% |
| Blood pressure, mmHg | |||||
| Systolic | 122 ± 16 | 122 ± 16 | 123 ± 16 | 123 ± 16 | 123 ± 16 |
| Diastolic | 73 ± 9 | 74 ± 9 | 73 ± 9 | 73 ± 9 | 73 ± 9 |
| Race/Ethnicity | |||||
| Asian/Pacific | 2.6% | 2.5% | 2.7% | 2.7% | 2.4% |
| Black | 3.9% | 4.2% | 6.0% | 6.6% | 7.2% |
| Hispanic/Latino | 3.8% | 4.2% | 4.4% | 3.6% | 3.2% |
| White | 88.1% | 87.6% | 85.4% | 85.8% | 85.9% |
| Education | |||||
| < High school diploma | 5.1% | 4.2% | 4.3% | 3.6% | 3.5% |
| High school diploma | 18.1% | 15.9% | 15.9% | 15.1% | 13.0% |
| School after high school | 37.8% | 36.8% | 36.8% | 36.4% | 34.1% |
| College degree or higher | 39.1% | 43.1% | 43.0% | 44.9% | 49.5% |
| Smoking | |||||
| Never | 50.9% | 50.7% | 51.8% | 51.6% | 51.3% |
| Past | 41.9% | 42.1% | 41.2% | 41.5% | 41.9% |
| Current | 7.3% | 7.1% | 7.1% | 6.9% | 6.8% |
| Use of postmenopausal hormones | |||||
| Never | 43.7% | 43.6% | 44.7% | 44.1% | 44.8% |
| Past | 15.9% | 16.4% | 16.4% | 16.2% | 15.1% |
| Current | 40.4% | 40.0% | 38.9% | 39.7% | 40.1% |
| Recreational physical activity level, MET-hours/week | 12.2 ± 13.5 | 11.6 ± 13.3 | 12.1 ± 13.8 | 11.9 ± 13.5 | 12.1 ± 13.6 |
| Energy intake, Kcal/day | 1492 ± 593 | 1778 ± 680 | 1545 ± 596 | 1621 ± 606 | 1612 ± 576 |
| 2005 HEI diet quality score | 66.7 ± 11.9 | 65.5 ± 11.3 | 67.7 ± 10.4 | 69.1 ± 9.8 | 71.0 ± 9.1 |
All characteristics were assessed at baseline (1993–1998). Values are mean ± SD (continuous variables) or percent (categorical variables). Not all categories sum to 100% due to rounding of data within some categories. Servings/d of 100% fruit juice are adjusted for energy using the residual method, and standardized to 2000 kcal/d. The Women’s Health Initiative enrolled participants between 1993–1998 throughout the United States.
Table 2.
Baseline Characteristics According to 100% Fruit Juice Consumption of 114,219 Postmenopausal US Women Analyzed for Diabetes Risk
| Modified Quintiles of 100% Fruit Juice Consumption | |||||
|---|---|---|---|---|---|
| I | II | III | IV | V | |
| Median oz./d of 100% fruit juice | 0 | 1.0 | 2.7 | 5.1 | 8.0 |
| Range oz./d of 100% fruit juice | 0 | 0.06–1.7 | 1.8–3.8 | 3.9–6.5 | 6.6–36.8 |
| N | 14,008 | 25,053 | 25,053 | 25,053 | 25,052 |
| Age, years | |||||
| 50–59 | 33.7% | 38.2% | 32.9% | 30.1% | 31.6% |
| 60–69 | 45.1% | 44.0% | 44.6% | 45.3% | 43.8% |
| 70–79 | 21.2% | 17.8% | 22.5% | 24.6% | 24.6% |
| BMI category, kg/m2 | |||||
| Normal (18.5 – 24.9) | 35.7% | 32.9% | 37.6% | 38.8% | 42.1% |
| Overweight (25.0 – 29.9) | 34.6% | 35.2% | 35.3% | 34.7% | 33.7% |
| Obese (> 30.0) | 28.4% | 31.0% | 26.6% | 25.5% | 23.3% |
| Blood pressure, mmHg | |||||
| Systolic | 126 ± 18 | 126 ± 17 | 127 ± 18 | 127 ± 18 | 127 ± 18 |
| Diastolic | 75 ± 9 | 75 ± 9 | 75 ± 9 | 75 ± 9 | 75 ± 9 |
| Race/Ethnicity | |||||
| Asian/Pacific | 2.5% | 2.6% | 2.8% | 2.6% | 2.5% |
| Black | 4.6% | 5.6% | 7.7% | 8.8% | 10.2% |
| Hispanic/Latino | 3.4% | 3.9% | 3.9% | 3.3% | 3.1% |
| White | 87.9% | 86.3% | 83.9% | 84.0% | 82.8% |
| Education | |||||
| < High school diploma | 5.3% | 4.6% | 4.8% | 4.2% | 4.1% |
| High school diploma | 19.0% | 16.6% | 16.8% | 16.1% | 13.9% |
| School after high school | 38.5% | 37.6% | 37.7% | 37.1% | 35.3% |
| College degree or higher | 37.2 % | 41.2% | 40.6% | 42.7% | 46.7% |
| Smoking | |||||
| Never | 51.1% | 51.3% | 51.6% | 51.4% | 51.7% |
| Past | 41.9% | 41.7% | 41.3% | 41.4% | 41.4% |
| Current | 7.0% | 7.0% | 7.1% | 7.2% | 6.9% |
| Use of postmenopausal hormones | |||||
| Never | 43.6% | 43.8% | 44.7% | 44.2% | 44.6% |
| Past | 15.9% | 16.4% | 16.1% | 16.4% | 15.6% |
| Current | 40.4% | 39.8% | 39.2% | 39.5% | 39.8% |
| Recreational physical activity level, MET-hours/week | 12.1 ± 13.5 | 11.8 ±13.3 | 12.1 ± 13.7 | 11.9 ± 13.5 | 12.1 ± 13.5 |
| Energy intake, Kcal/day | 1493 ± 597 | 1770 ± 690 | 1556 ± 599 | 1634 ± 614 | 1610 ± 596 |
| 2005 HEI diet quality score | 66.4 ± 11.8 | 65.3 ± 11.4 | 67.6 ± 10.5 | 68.9 ± 9.9 | 70.9 ± 9.2 |
All characteristics were assessed at baseline (1993–1998). Values are mean ± SD (continuous variables) or percent (categorical variables). Not all categories sum to 100% due to rounding of data within some categories. Servings/d of 100% fruit juice are adjusted for energy using the residual method, and standardized to 2000 kcal/d. The Women’s Health Initiative enrolled participants between 1993–1998 throughout the United States.
Higher versus lower 100% fruit juice consumption was not associated with incident hypertension (Table 3) or diabetes (Table 4) over a mean individual follow-up time of 7.8 years. In multivariable-adjusted analyses comparing highest to lowest quintiles, 100% fruit juice intake was not associated with incident hypertension (HR 1.01, 95% CI 0.97–1.04) or diabetes (HR 0.97, 95% CI 0.91–1.03). Relationships were similar in the univariate and the multivariable-adjusted models.
Table 3.
Associations of 100% Fruit Juice and Whole Fruit Consumption with Hypertension Risk in 80,539 Postmenopausal US Women
| Hypertension Risk According to 100% Fruit Juice Consumption | ||||||
|---|---|---|---|---|---|---|
| Modified Quintiles of 100% Fruit Juice Consumption | P for Trend | |||||
| I | II | III | IV | V | ||
| Median oz./d | 0 | 1.0 | 2.6 | 4.9 | 7.8 | |
| Person-years of follow-up | 58,299 | 100,796 | 100,614 | 99,971 | 99,467 | |
| Number of incident cases | 5994 | 10,087 | 9971 | 10,036 | 10,114 | |
| Incidence rate, per 1000 person-years | 103 | 100 | 99.1 | 100 | 102 | |
| Multivariable-adjusted* hazard ratio (95% CI) | 1.0 (ref) | 0.98 (0.94–1.01) | 0.97 (0.94–1.01) | 0.98 (0.94–1.01) | 1.01 (0.97–1.04) | 0.21 |
| Hypertension Risk According to Whole Fruit Consumption | ||||||
|---|---|---|---|---|---|---|
| Quintiles of 100% Whole Fruit Consumption | P for Trend | |||||
| I | II | III | IV | V | ||
| Median servings/d | 0.3 | 0.7 | 0.9 | 1.6 | 2.4 | |
| Person-years of follow-up | 92,241 | 91,812 | 91,647 | 91,953 | 91,494 | |
| Number of incident cases | 9244 | 9239 | 9263 | 9214 | 9242 | |
| Incidence rate, per 1000 person-years | 100 | 101 | 101 | 100 | 101 | |
| Multivariable-adjusted* hazard ratio (95% CI) | 1.0 (ref) | 1.00 (0.97–1.04) | 1.01 (0.98–1.04) | 1.00 (0.97–1.03) | 1.01 (0.98–1.04) | 0.91 |
Abbreviations: CI (confidence interval), /d (per day), P (probability)
Adjusted for age, education level, race/ethnicity, smoking status, physical activity, body mass index, hormone replacement therapy status, study arm, and total energy intake. Univariate and multivariable-adjusted models yielded nearly identical hazard ratios and 95% confidence intervals.
Hazard ratios estimated for quintiles II – V compare the participants in each upper quintile to the participants in the first quintile.
The Women’s Health Initiative enrolled participants between 1993–1998 throughout the United States.
Table 4.
Associations of 100% Fruit Juice and Whole Fruit Consumption and Diabetes Risk in 114,219 Postmenopausal US Women
| Diabetes Risk According to 100% Fruit Juice Consumption | ||||||
|---|---|---|---|---|---|---|
| Modified Quintiles of 100% Fruit Juice Consumption | P for Trend | |||||
| I | II | III | IV | V | ||
| Median oz./d | 0 | 1.0 | 2.7 | 5.1 | 8.0 | |
| Person-years of follow-up | 102,874 | 183,543 | 183,980 | 183,210 | 184,126 | |
| Number of incident cases | 1435 | 2529 | 2522 | 2541 | 2461 | |
| Incidence rate, per 1000 person-years | 13.9 | 13.8 | 13.7 | 13.9 | 13.4 | |
| Multivariable-adjusted* hazard ratio (95% CI) | 1.0 (ref) | 0.98 (0.92, 1.04) | 0.99 (0.93–1.05) | 1.00 (0.93–1.07) | 0.97 (0.91–1.03) | 0.17 |
| Diabetes Risk According to Whole Fruit Consumption | ||||||
|---|---|---|---|---|---|---|
| Quintiles of 100% Whole Fruit Consumption | P for Trend | |||||
| I | II | III | IV | V | ||
| Median servings/d | 0.3 | 0.6 | 0.9 | 1.6 | 2.4 | |
| Person-years of follow-up | 167,457 | 167,857 | 167,421 | 167,277 | 167,721 | |
| Number of incident cases | 2329 | 2406 | 2255 | 2217 | 2281 | |
| Incidence rate, per 1000 person-years | 13.9 | 14.4 | 13.4 | 13.2 | 13.6 | |
| Multivariable-adjusted* hazard ratio (95% CI) | 1.0 (ref) | 1.03 (0.97–1.10) | 0.97 (0.91–1.03) | 0.95 (0.89–1.01) | 1.00 (0.94–1.06) | 0.34 |
Abbreviations: CI (confidence interval), /d (per day), P (probability)
Adjusted for age, education level, race/ethnicity, smoking status, physical activity, body mass index, hormone replacement therapy status, study arm, and total energy intake. Univariate and multivariable-adjusted models yielded nearly identical hazard ratios and 95% confidence intervals.
Hazard ratios estimated for quintiles II – V compare the participants in each upper quintile to the participants in the first quintile.
The Women’s Health Initiative enrolled participants between 1993–1998 throughout the United States.
Higher versus lower whole fruit consumption was also not associated with incident hypertension (Table 3) or diabetes (Table 4) In multivariable-adjusted analyses comparing the highest to lowest quintiles, whole fruit intake was not associated with risk of hypertension (HR 1.01, 95% CI 0.98–1.04) or diabetes (HR 1.00, 95% CI 0.94–1.06), and associations were similar in univariate and multivariable-adjusted models. Exploratory analyses stratified by baseline age, race/ethnicity, study arm, and blood pressure did not change the results or their interpretation. Change analysis, classifying exposures as tertiles or quartiles, and restricting the analytic sample to the highest quintile of consumers also did not change our results. Exploratory analyses categorizing the exposures of 100% fruit juice and whole fruit using cut-points based on serving frequency showed an increase in hypertension risk associated with consuming ≥ 24 oz./d of 100% fruit juice vs. ≤ 4 oz./day (HR 1.29, 95% CI 1.06–1.57), and the trend of increasing serving frequency and greater hypertension risk was significant (P for trend = 0.03; Table 5). Categorizing the exposures using cut-points of serving frequency (< 4 servings/week, 5–6 servings/week, 1 serving/day, etc) otherwise yielded null associations for trend (Table 5). In exploratory analyses comparing citrus and non-citrus fruit and fruit juice, the increase in hypertension risk associated with consuming multiple servings per day of 100% fruit juice was only observed for non-citrus juices (Supplementary Table C).
Table 5.
Exploratory Analysis: Exposures Defined Using Cut-points
| Incidence of Hypertension Between 1993 and 2005 According to 100% Fruit Juice and Whole Fruit Consumption in 80,539 Postmenopausal US Women | ||||||
|---|---|---|---|---|---|---|
| 100% Fruit Juice | Whole Fruit | |||||
| N | IR, per 1000 PY | Multivariable Adj HR (95% CI) | N | IR, per 1000 PY | Multivariable Adj HR (95% CI) | |
| ≤ 4 serv/week | 42,854 | 2.8 | 1.0 (ref) | 19,651 | 2.7 | 1.0 (ref) |
| 5–6 serv/week | 16,530 | 2.7 | 0.99 (0.97–1.01) | 23,933 | 2.8 | 1.02 (1.00–1.04) |
| 1 serv/day | 17,905 | 2.8 | 1.02 (0.99–1.04) | 23,530 | 2.8 | 1.01 (0.98–1.05) |
| 2–3 serv/day | 2647 | 2.9 | 1.04 (0.99–1.10) | 12,397 | 2.8 | 1.02 (0.98–1.05) |
| ≥ 4 serv/day | 159 | 3.4 | 1.29 (1.06–1.56) | 584 | 2.6 | 0.94 (0.83–1.05) |
| P for Trend | 0.03 | 0.76 | ||||
| Incidence of Diabetes Between 1993 and 2005 According to 100% Fruit Juice and Whole Fruit Consumption in 114,219 Postmenopausal US Women | ||||||
|---|---|---|---|---|---|---|
| 100% Fruit Juice | Whole Fruit | |||||
| N | IR, per 1000 PY | Multivariable Adj HR (95% CI) | N | IR, per 1000 PY | Multivariable Adj HR (95% CI) | |
| ≤ 4 serv/week | 59,226 | 13.8 | 1.0 (ref) | 28,633 | 13.8 | 1.0 (ref) |
| 5–6 serv/week | 24,010 | 13.9 | 1.01 (0.97–1.07) | 34,213 | 14.1 | 1.03 (0.97–1.08) |
| 1 serv/day | 26,142 | 13.5 | 0.97 (0.93–1.02) | 32,771 | 13.4 | 1.00 (0.94–1.06) |
| 2–3 serv/day | 3972 | 13.4 | 0.97 (0.87–1.08) | 17,180 | 13.5 | 1.04 (0.96–1.11) |
| ≥ 4 serv/day | 253 | 10.5 | 0.82 (0.53–1.27) | 806 | 13.3 | 0.93 (0.73–1.18) |
| P for Trend | 0.18 | 0.28 | ||||
Abbreviations: adj (adjusted), HR (hazard ratio), IR (incidence rate), PY (person-years), serv (servings)
Adjusted for age, education level, race/ethnicity, smoking status, physical activity, BMI, hormone replacement therapy status, study arm, and total energy intake.
Hazard ratios compare participants in higher categories of consumption (5–6 serv/week, 1 serv/day, etc) to participants in the lowest category of consumption (< 4 servings/week).
One serving of 100% fruit juice was defined as 6 ounces. The Women’s Health Initiative enrolled participants between 1993–1998 throughout the United States.
Discussion
Our objective was to investigate whether 100% fruit juice consumption was associated with incident hypertension and diabetes in postmenopausal US women. Compared to no consumption of 100% fruit juice, we found no evidence that consumption of 8 oz./d of 100% fruit juice was associated with incident hypertension or diabetes. In exploratory analyses, consuming ≥ 24 oz./d of 100% fruit juice, compared to consuming ≤ 4 oz./day, was associated with increased hypertension risk (but not diabetes risk). Non-citrus fruit juices were associated with this increase in hypertension risk, while citrus fruit juices were not. We also analyzed the association of differing levels of whole fruit consumption and hypertension and diabetes risk. Compared to consuming less then 1/3 serving per day of whole fruit, consuming 2.5 servings/d whole fruit was associated not associated with incident hypertension or diabetes.
To our knowledge, this is the first longitudinal analysis of 100% fruit juice intake and hypertension risk. Previous cohort studies have examined the association of whole fruit, but not 100% fruit juice, with risk of hypertension. Borgi and colleagues analyzed 187,453 US adults for a mean follow-up time of 15.7 years.24 They compared participants who consumed ≥ 4 servings/d of whole fruit to participants who consumed ≤ 4 servings/week of whole fruit, and found a multivariable adjusted HR of 0.92 (95% CI 0.87–0.97) for incident hypertension. When we analyzed 100% fruit juice and whole fruit dietary exposures using the same categories as Borgi et al., we found that whole fruit was associated with a similar, but non-significant decreased risk of hypertension (HR 0.94; 95% CI 0.83–1.05), and found that comparing the highest to lowest categories of 100% fruit juice consumption was associated with an increased risk of hypertension (HR 1.29; 95% CI 1.06–1.56), with a significant increasing trend across categories (P for trend = 0.03).
Our findings that neither 100% fruit juice nor whole fruit consumption were associated with increased diabetes risk differ slightly from other cohort studies.11,12,25,27 Muraki and colleagues analyzed 187,382 adults in the same 3 cohort studies as Borgi and colleagues over a mean follow-up time of 18.5 years. They analyzed the associations of changes in consumption over 4-year time periods of (1) 100% fruit juice, and (2) whole fruit, with diabetes risk over the same 4-year time periods.xxv In multivariable analyses, each 3-serving/week increment in 100% fruit juice consumption was associated with an 8% (95% CI 1.05, 1.11) increased risk of incident diabetes. Each 3-serving/week increment in whole fruit consumption was associated with a decreased risk of incident diabetes (HR 0.98, 95% CI 0.96, 0.99). Though statistically significant, the HRs found by Muraki et al. are close to a null HR of 1.0 and do not suggest a large impact.
The WHI is not well suited to study sugar-sweetened beverages (SSBs) due to low consumption of SSBs, but meta-analyses of other prospective cohort studies found that SSBs are consistently associated with increased cardiometabolic disease risk. Consumption of 1 serving/d of SSB is independently associated with gaining 1 lb/4 years among US adults, compared to gaining 0.3 lb/4 years for 100% fruit juice.2 In meta-analyses comparing the highest versus lowest quantiles of SSB consumption, high SSB consumers had a 12% increased risk of hypertension,xxvi 26% increased risk of type 2 diabetes,xxvii and 19% increased risk of cardiovascular disease.xxviii
Our findings are relevant to dietary policy. The 2015–2020 Dietary Guidelines for Americans (DGA)xxix recommend that adults limit 100% fruit juice to 8 oz./d. These guidelines also emphasize that consuming water and whole fruit is preferred to 100% fruit juice, since 100% fruit juice contains less dietary fiber than whole fruit, and when consumed in excess, 100% fruit juice may contribute extra calories to Americans’ diets. Our findings showed that consumption of 8 oz./d of 100% fruit juice was not associated with increased risk of hypertension or diabetes. Given stronger associations of SSB consumption with cardiometabolic diseases, our results do not support treating 100% fruit juice like SSBs in dietary policies.
This analysis has limitations. Exposure misclassification may be present from participants misunderstanding questions on the FFQ about 100% fruit juice versus fruit drinks such as Tang®. Despite adjusting for nine covariates, there may be residual unmeasured confounding of healthy behaviors and 100% fruit juice consumption, which could bias our associations towards the null.16 Fruit juice consumption among WHI participants was relatively low, reducing power to examine whether higher levels of consumption are more strongly associated with risk of diabetes or hypertension. Subgroup analysis of participants with pre-diabetes was not possible. Finally, our outcome measure of incident hypertension was based on self-report, which has been used in other WHI analyses but has not been objectively validated.24
Conclusion
In this secondary analysis of postmenopausal US women in the WHI, consumption of 8 oz./d of 100% fruit juice was not associated with incident hypertension or diabetes. Consumption of high levels of 100% fruit juice (≥ 24 oz./d) was associated with an increased risk of hypertension, but not diabetes. Our findings that consuming moderate amounts of 100% fruit juice is not associated with hypertension and diabetes risk stand in contrast to evidence linking consumption of moderate amounts of sugar-sweetened beverages to these diseases.5, 28–30 Our results suggest that consuming ≤ 8 oz./d of 100% fruit juice, as recommended by the Dietary Guidelines for Americans, does not increase risk of hypertension or diabetes.
Supplementary Material
Highlights.
We examined if drinking 100% fruit juice is linked to hypertension or diabetes risk
Drinking 1 serving/day of 100% fruit juice did not increase risk of either disease
Drinking 4 servings/day of 100% fruit juice increased hypertension risk (but not diabetes)
Acknowledgments
Funding:
We thank the WHI participants, staff, and investigators. The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts HHSN268201600018C, HHSN268201600001C, HHSN268201600002C, HHSN268201600003C, and HHSN268201600004C. Dr. Auerbach is funded by the Ruth L. Kirschstein National Research Service of the National Institutes of Health through the University of Washington (grant T32HP10002). The funding sources had no role in the design, conduct, or analysis of the study, or the decision to submit the manuscript for publication.
Abbreviations
- Adj
adjusted
- BMI
body mass index
- CVD
cardiovascular disease
- /d
per day
- HEI
Healthy Eating Index
- HR
hazard ratio
- HTN
hypertension
- IR
incidence rate
- kcal
kilocalories
- MET
Metabolic Equivalent of Task
- mmHg
millimeters of mercury
- Oz
ounce
- RCT
randomized controlled trial
- Ref
reference
- RR
relative risk
- Serv
servings
- PY
person-years
Footnotes
Contributions:
Dr. Auerbach conceived the study, completed the analysis, and led the writing. Drs. Littman, Tinker, Krieger, and Young helped design the study, contributed to analysis and interpretation of results, and critically review the manuscript for important intellectual content. Mr. Larsen completed the analysis and critically reviewed the manuscript for important intellectual content. Dr. Neuhouser conceived the study, supervised the overall project, and revised the manuscript for important intellectual content.
Disclosures:
The authors have no conflicts of interest to report
Conflicts of Interest: None Declared
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Contributor Information
Brandon J. Auerbach, Department of Epidemiology, University of Washington, Seattle, Washington.
Alyson J. Littman, Department of Epidemiology, University of Washington, Seattle, Washington; VA Puget Sound Health Care System, Seattle, Washington; Northwest VA Health Services Research & Development Center of Excellence, Seattle, Washington.
Lesley Tinker, Women’s Health Initiative Clinical Coordinating Center, Seattle, Washington; Fred Hutchinson Cancer Research Center, Seattle, Washington.
Joseph Larson, Clinical Coordinating Center, Women’s Health Initiative, Seattle, Washington; Fred Hutchinson Cancer Research Center, Seattle, Washington.
James Krieger, Division of General Internal Medicine and Department of Health Services, University of Washington, Seattle; Healthy Food America, Seattle, Washington.
Bessie Young, Division of Nephrology and Department of Epidemiology, University of Washington, Seattle, Washington; VA Puget Sound Health Care System, Seattle, Washington; Northwest VA Health Services Research and Development Center of Excellence, Seattle, Washington.
Marian Neuhouser, Department of Epidemiology, University of Washington, Seattle, Washington; Women’s Health Initiative Clinical Coordinating Center, Seattle, Washington; Fred Hutchinson Cancer Research Center, Seattle, Washington.
References
- i.Wojcicki JM, Heyman MB. Reducing childhood obesity by eliminating 100% fruit juice. Am J Public Health. 2012;102:1630–3. doi: 10.2105/AJPH.2012.300719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- ii.Mozaffarian D, Hao T, Rimm EB, et al. Changes in diet and lifestyle and long-term weight gain in women and men. N Engl J Med. 2011;364:2392–404. doi: 10.1056/NEJMoa1014296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- iii.World Health Organization. Taxes on sugary drinks: Why do it? Geneva: World Health Organization; Oct, 2016. [Google Scholar]
- iv.Gill JMR, Sattar N. Fruit juice: just another sugary drink? Lancet Diabetes Endocrinol. 2014;2(6):444–446. doi: 10.1016/S2213-8587(14)70013-0. [DOI] [PubMed] [Google Scholar]
- v.Malik VS, Hu FB. Fructose and Cardiometabolic Health: What the Evidence From Sugar-Sweetened Beverages Tells Us. J Am Coll Cardiol. 2015;66:1615–24. doi: 10.1016/j.jacc.2015.08.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- vi.Atkinson F, Foster-Powell K, Brand-Miller JC. International Tables of Glycemic Index and Glycemic Load Values. Diabetes Care. 2008;31(12):2281–3. doi: 10.2337/dc08-1239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- vii.Ludwig DS. The glycemic index: physiological mechanisms relating to obesity, diabetes, and cardiovascular disease. JAMA. 2002;287:2414–23. doi: 10.1001/jama.287.18.2414. [DOI] [PubMed] [Google Scholar]
- viii.Stanhope KL, Schwarz JM, Keim NL, et al. Consuming fructose-sweetened, not glucose-sweetened, beverages increases visceral adiposity and lipids and decreases insulin sensitivity in overweight/obese humans. J Clin Invest. 2009;119:1322–34. doi: 10.1172/JCI37385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- ix.Nakagawa T, Tuttle KR, Short RA, Johnson RJ. Hypothesis: fructose-induced hyperuricemia as a causal mechanism for the epidemic of the metabolic syndrome. Nat Clin Pract Nephrol. 2005;1:80–6. doi: 10.1038/ncpneph0019. [DOI] [PubMed] [Google Scholar]
- x.Liu K, Xing A, Chen K, et al. Effect of fruit juice on cholesterol and blood pressure in adults: a meta-analysis of 19 randomized controlled trials. PLoS One. 2013;8:e61420. doi: 10.1371/journal.pone.0061420. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xi.Xi B, Li S, Liu Z, et al. Intake of fruit juice and incidence of type 2 diabetes: a systematic review and meta-analysis. PLoS One. 2014;9:e93471. doi: 10.1371/journal.pone.0093471. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xii.Imamura F, O’Connor L, Ye Z, et al. Consumption of sugar sweetened beverages, artificially sweetened beverages, and fruit juice and incidence of type 2 diabetes: systematic review, meta-analysis, and estimation of population attributable fraction. BMJ. 2015;351:h3576. doi: 10.1136/bmj.h3576. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xiii.Heianza Y, Arase Y, Kodama S, et al. Effect of postmenopausal status and age at menopause on type 2 diabetes and prediabetes in Japanese individuals. Diabetes Care. 2013;36(12):4007–14. doi: 10.2337/dc13-1048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xiv.Lovejoy JC, Champagne CM, de Jonge L, et al. Increased visceral fat and decreased energy expenditure during the menopausal transition. Int J Obes (Lond) 2008;32:949–58. doi: 10.1038/ijo.2008.25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xv.Rurik I, Móczár C, Buono N, et al. Early and Menopausal Weight Gain and its Relationship with the Development of Diabetes and Hypertension. Exp Clin Endocrinol Diabetes. 2016 doi: 10.1055/s-0035-1569365. [DOI] [PubMed] [Google Scholar]
- xvi.The Women’s Health Initiative Study Group. Design of the Women’s Health Initiative clinical trial and observational study. Control Clin Trials. 1998;19:61–109. doi: 10.1016/s0197-2456(97)00078-0. [DOI] [PubMed] [Google Scholar]
- xvii.Prentice RL, Tinker LF, Huang Y, Neuhouser ML. Calibration Of Self-Reported Dietary Measures Using Biomarkers: An Approach To Enhancing Nutritional Epidemiology Reliability. Curr Atheroscler Rep. 2013;15(9) doi: 10.1007/s11883-013-0353-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xviii.Patterson RE, Kristal AR, Tinker LF, et al. Measurement characteristics of the Women’s Health Initiative food frequency questionnaire. Ann Epidemiol. 1999;9:178–87. doi: 10.1016/s1047-2797(98)00055-6. [DOI] [PubMed] [Google Scholar]
- xix.Langer RD, White E, Lewis CE, et al. The Women’s Health Initiative Observational Study: baseline characteristics of participants and reliability of baseline measures. Ann Epidemiol. 2003;13:S107–21. doi: 10.1016/s1047-2797(03)00047-4. [DOI] [PubMed] [Google Scholar]
- xx.Margolis KL, Lihong LS, Brzyski R, et al. Validity of diabetes self-reports in the Women’s Health Initiative: comparison of medication inventories and fasting glucose measurements. Clin Trials. 2008;5:240–7. doi: 10.1177/1740774508091749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xxi.Wang L, Manson JE, Gaziano JM, et al. Plasma adiponectin and the risk of hypertension in white and black postmenopausal women. Clin Chem. 2012;58(10):1438–45. doi: 10.1373/clinchem.2012.191080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xxii.Guenther PM, Reedy J, Krebs-Smith SM, Reeve BB. Evaluation of the Healthy Eating Index-2005. J Am Diet Assoc. 2008;108:1854–64. doi: 10.1016/j.jada.2008.08.011. [DOI] [PubMed] [Google Scholar]
- xxiii.Smith JD, Hou T, Hu FB, et al. A Comparison of Different Methods for Evaluating Diet, Physical Activity, and Long-Term Weight Gain in 3 Prospective Cohort Studies. J Nutr. 2015;145:2527–34. doi: 10.3945/jn.115.214171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xxiv.Borgi L, Muraki I, Satija A. Fruit and Vegetable Consumption and the Incidence of Hypertension in Three Prospective Cohort Studies. Hypertension. 2016;67:288–293. doi: 10.1161/HYPERTENSIONAHA.115.06497. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xxv.Muraki I, Imamura F, Manson JE. Fruit consumption and risk of type 2 diabetes: results from three prospective longitudinal cohort studies. BMJ. 2013;347:f5001. doi: 10.1136/bmj.f5001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xxvi.Jayalath VH, de Souza RJ, Ha V, Mirrahimi A, Blanco-Mejia S, Di Buono M, Jenkins AL, Leiter LA, Wolever TM, Beyene J, Kendall CW, Jenkins DJ, Sievenpiper JL. Sugar-sweetened beverage consumption and incident hypertension: a systematic review and meta-analysis of prospective cohorts. Am J Clin Nutr. 2015;102:914–21. doi: 10.3945/ajcn.115.107243. [DOI] [PubMed] [Google Scholar]
- xxvii.Malik VS, Popkin BM, Bray GA, Després JP, Willett WC, Hu FB. Sugar-sweetened beverages and risk of metabolic syndrome and type 2 diabetes: a meta-analysis. Diabetes Care. 2010;33:2477–83. doi: 10.2337/dc10-1079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- xxviii.Narain A, Kwok CS, Mamas MA. Soft drinks and sweetened beverages and the risk of cardiovascular disease and mortality: a systematic review and meta-analysis. Int J Clin Pract. 2016;70:791–805. doi: 10.1111/ijcp.12841. [DOI] [PubMed] [Google Scholar]
- xxix.US Department of Health and Human Services & US Department of Agriculture. 2015–2020 Dietary Guidelines for Americans. 8. Dec, 2015. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.