Consumption of fructose-containing food and beverage sources in childhood through to adulthood and risk of hypertension: a prospective cohort study

Abstract

Background:

High intakes of fructose-containing sugars among children and adolescents is implicated in obesity and related comorbidities, including hypertension. However, sugar-sweetened beverages (SSBs), fruit juices, and whole fruit, have different nutritional profiles and matrices, which may confer different effects on blood pressure.

Methods:

The Growing Up Today Study (GUTS) is a longitudinal cohort of 25,749 individuals (55% female) drawn from two enrollment waves, GUTS 1 (n=16,875; baseline 1996) and GUTS II (n=10,918; baseline 2004) followed prospectively through 2021 (age at enrollment mean=12 years, age at end of follow-up mean=36 years). Participants provided updated information on lifestyle, health status, and habitual diet via validated food frequency questionnaires every 1–4 years. We conducted multivariable adjusted Cox proportional hazards regression models to estimate the associations of total fructose and SSB, fruit juice, and whole fruit intake (cumulative averages) with incident hypertension (hazard ratios [HR] and 95% confidence intervals [CI]), adjusting for major diet and lifestyle factors. We also modeled substitutions of SSBs or fruit juice with whole fruit, milk, and water.

Results:

During up to 25 years of follow-up, 1,625 (6.3%) participants reported a hypertension diagnosis. Total fructose intake was not associated with incident hypertension (highest vs. lowest quintile HR, 1.07; 95% CI, 0.92, 1.25; P-trend<0.001). However, participants with the highest intake of SSBs (≥2 servings/day vs. <3 servings/week) and fruit juice (≥1.5 servings/day vs. <1 serving/week) had a higher risk of hypertension (HR, 1.52; 95% CI, 1.27, 1.83; P-trend<0.001 and HR, 1.35; 95% CI, 1.06, 1.71; P-trend=0.018, respectively). In contrast, whole fruit was not associated with hypertension (highest vs. lowest category HR, 0.79; 95% CI, 0.59, 1.05; P-trend=0.08). Replacing one serving/day of SSB with milk, water, or whole fruit was associated with a 13% (95% CI, 5%, 20%), 9% (95% CI, 3%, 15%), and 22% (95% CI, 11%, 31%) lower risk of hypertension, respectively. Additionally, replacing fruit juice with whole fruit was associated with a 19% (95% CI, 3%, 32%) lower risk of hypertension.

Conclusions:

SSBs and fruit juice intakes were positively associated with a higher risk of hypertension, independent of overall diet quality, physical activity, and other factors. Our findings support public health guidelines to limit the overconsumption of SSBs and fruit juice starting in childhood to protect against the development of hypertension.

INTRODUCTION

Elevated blood pressure remains a leading risk factor for the global burden of disease, contributing significantly to morbidity and mortality.^{1, 2} Hypertension rates have risen dramatically, affecting over one billion people worldwide, contributing to an estimated 9.4 million deaths annually.^{2, 3} Importantly, high blood pressure is also emerging earlier in life, with growing rates being seen in younger adults, and children and adolescents also being affected, thus highlighting the importance of early detection and prevention.^4–6 Hypertension often coexists with obesity, type 2 diabetes (T2D), and dyslipidemia, which in tandem, significantly elevates the risk of cardiovascular disease (CVD) and related conditions.^{3, 7, 8} Despite its widespread prevalence, hypertension is largely modifiable through lifestyle factors, particularly through improving diet quality, including reductions in sugar intake.⁷ Early-life dietary patterns play a critical role as childhood and adolescence represent key periods for shaping long-term cardiometabolic health. Identifying modifiable dietary risk factors in early life may support long-term reductions in the risk of hypertension and downstream cardiometabolic disease risk later in life.⁸

The overconsumption of sugars has been implicated as a driver in cardiometabolic disease, with a special focus on fructose due to its unique metabolic response in the liver.^{9, 10} However, considerations should be made for the whole food source rather than the nutrient in isolation, as different food matrices may influence absorption and bioavailability differently. This consideration is in line with the shifts away from nutrient-based recommendations towards dietary pattern and food-based recommendations. Sugar-sweetened beverages (SSBs) are a major source of fructose in the diet and have been consistently linked to adverse health outcomes.^11–13 In contrast, whole fruit, another important source of fructose, also provides dietary fibre and polyphenols, which may independently protect against the development of CVD.¹⁴ Fruit juices, which contain high amounts of free sugars and fructose, akin to those of SSBs, also contain many of the vitamins and minerals found in their whole fruit form; however, they lack dietary fiber. The evidence on fruit juice and hypertension has yielded mixed findings from epidemiological and dietary intervention trials.^15–17 International dietary guidelines on fruit juice consumption also remain inconsistent.^18–21

The existing literature on fructose-containing foods and beverages and hypertension risk has largely focused on short-term trials or assessments in adult populations. Few studies have assessed long-term consumption patterns across critical life stages and how they may influence disease risk later in life. Thus, this study aims to examine the associations of daily total fructose intake, and major fructose food and beverage sources, including SSBs, fruit juice, and whole fruit consumption, and risk of hypertension in a large prospective cohort following participants from childhood through to adulthood.

METHODS

Study Population

Our analysis was conducted in the Growing Up Today Study (GUTS), an ongoing prospective cohort that recruited the children of participants in the Nurses’ Health Study II (NHSII) cohort. The first phase of enrollment, GUTS I, was initiated in 1996 and included 16,875 children from across the U.S. who were aged 9–14 years at baseline. The second phase of enrollment, GUTS II, was initiated in 2004 and included 10, 918 children aged 9–16 years at baseline. In 2013, GUTS I and GUTS II were merged as the GUTS cohort for continued follow-up. All participants were invited to complete questionnaires annually from 1996–2001, then every 2 to 4 years thereafter, to report usual diet, anthropometrics, lifestyle factors, and health status. For our analyses, we excluded individuals reporting hypertension at baseline, those with missing information on dietary intake data at baseline, or with implausible estimates for total energy (<500 or >5000 kcal).²² After exclusions, our final analysis included 25,749 participants. This study was approved by the Institutional Review Board of the Brigham and Women’s Hospital and the Harvard T.H. Chan School of Public Health. Informed consent was implied by the return of the cohort questionnaires.

Assessment of Diet

A validated 132-item semiquantitative food frequency questionnaire (FFQ) was administered annually from 1996 to 1998, then in 2001, 2004, 2006, 2008, 2011, and 2015. On each FFQ, participants were asked how often, on average, they consumed a standard serving of a food or beverage, ranging from “never or less than once per month” to “6 or more per day”. A serving was specified as a can or glass for SSBs and a glass for fruit juices. We defined SSBs to include sodas, fruit drinks (fruit punches, lemonades, iced teas, or other non-carbonated fruit drinks), and sport drinks. Fruit juices were defined as orange juice or apple juice and “other” 100% fruit juices. Whole fruits were defined as apples, oranges, bananas, mangos, grapes, pears, melons, strawberries, and peaches. Nutrient and energy intakes were calculated by multiplying the frequency of consumption of each unit of food and beverage by the estimates of nutrient and energy contents derived from external composition databases and summing across all items. Total fructose equivalents (TFE) were calculated as the sum of the fructose monosaccharaide and the fructose from sucrose across all fructose-containing FFQ items. The FFQ has been validated against three 24-hour recalls with good validity (average Pearson correlation coefficient = 0.54).²³ The correlation coefficients using a similar FFQ were r=0.84 for colas, r=0.75–0.90 for fruit juices, and r=0.38–0.95 for individual whole fruits.²⁴

Assessment of Hypertension

Hypertension was ascertained via self-report in the 2010 to 2021 questionnaires. Participants were asked if they were ever diagnosed by a health care provider with hypertension and to indicate the year of diagnosis. In the 2010 questionnaire, the earliest response option for the year of diagnosis was “before 1996” and ranged until “2010+”. A previous study showed high validity of self-reported hypertension compared to medical records in the GUTS.²⁵

Assessment of Covariates

Information on anthropometrics, diet, lifestyle factors, and dietary supplementation use were collected at baseline and updated through follow-up questionnaires. Body mass index (BMI) was calculated using self-reported height and weight from each questionnaire, which has been shown to have good validity.²⁶ Race/ethnicity was self-reported at baseline. Information on dietary covariates were obtained using updated FFQs, and we adjusted for other dietary factors as potential confounders, including red and processed meats, fruits and vegetables, and whole grains. Physical activity was converted to metabolic equivalent (MET) hours per week by multiplying the time spent on each activity (e.g. various sports and aerobic exercises) by its corresponding MET score. Screen time was measured as the average number of hours spent per week using electronic devices (e.g. watching television, using a computer, or playing video game). Sleep duration was measured as the average number of hours spent sleeping per day.

Statistical Analysis

Baseline characteristics were summarized across categories of exposure. Continuous variables were assessed visually for their distribution and presented as means and standard deviations (SDs) for consistency and comparability with prior epidemiologic studies. Follow-up duration and person-years were calculated from the return of the baseline questionnaire to the end of follow-up (June 2022) or until the date of hypertension diagnosis, whichever came first. We used Cox proportional hazards regression models to estimate the hazard ratios (HRs) and 95% confidence intervals (CIs) of the associations between TFE (grams/day), SSBs, fruit juice, and whole fruit (servings/day) and risk of hypertension. We evaluated the proportional hazards assumption by testing interactions between each exposure category and follow-up time in fully adjusted Cox models. There was no evidence of non-proportional hazards for any exposure (SSBs: p = 0.58; fruit juice: p = 0.90; whole fruit: p = 0.71; TFE: p = 0.98), indicating that the proportional hazards assumption was satisfied in all models. To represent long-term average diet, we used the cumulative averages of dietary data from baseline through each follow-up cycle, updated over time. For TFE, we compared each quintile of intake to the lowest quintile. Similarly, for SSBs, fruit juice, and whole fruit, we categorized the frequency of intake, with the lowest consumption category serving as the reference group. We conducted tests for linear trends of the diet-hypertension associations using the Wald test by assigning the median value to each quintile or category of consumption and modeling the variable as a continuous variable. For TFE, we also assessed hypertension risk per 5% increment in energy from total fructose. Missing dietary or covariate data were carried forward from data reported in the preceding questionnaire cycle. All analyses were stratified by age (years) and follow-up intervals. We adjusted the multivariable models for race (white or non-white), physical activity (METs/week, in tertiles), screen time (hrs/week, in tertiles), sleep (hrs/day, in tertiles), multivitamin use (yes or no), smoking (yes or no), total energy (quintiles), and intakes of fruits, vegetables, whole grains, and red and processed meat (quintiles). We subsequently added BMI (quintiles) to the model (multivariable model 2). SSBs, fruit juice, and whole fruit were mutually adjusted for. For analyses of TFE and whole fruit and risk of hypertension, adjustment for whole fruit intake was removed from the model. Analyses were conducted separately for GUTS I and GUTS II and pooled HRs were obtained from pooling the cohort data.

Dose-response analyses using the fully adjusted model were conducted using restricted cubic spline regression with 3 knots to assess possible non-linear relationships between TFE (g/day), SSBs, fruit juice, and whole fruit (servings/day) and risk of hypertension. Tests for non-linearity were determined from likelihood ratio tests.²⁷

Substitution analyses were conducted by contrasting 1 serving/day of SSB with a serving of fruit juice, total milk, water, and whole fruit, and substituting fruit juice with an equivalent amount of total milk, water, and whole fruit. The items were added to multivariable adjusted model 2 as continuous variables and the difference in β coefficients and variance, and the covariance were used to estimate the HRs and 95% CIs for a 1:1 substitution with risk of developing hypertension.²⁸

We conducted sensitivity analyses by assessing baseline intake levels and risk of hypertension to represent intakes in childhood/adolescence and potential impact later in life. Additionally, we assessed the association using the most recent measure of dietary intake with hypertension risk (simple update method). We also assessed associations by including baseline BMI in multivariable model 2, which accounts for early-life BMI as a potential influence on hypertension risk. Stratified analyses and potential effect modification by age (< 25 years vs. ≥ 25 years), sex (female, male), racial or ethnic groups (non-Hispanic white vs. not non-Hispanic white), BMI (< median vs. ≥ median), physical activity (< median vs. ≥ median), and multivitamin use (no vs. yes) were evaluated, and all models were multivariable adjusted. We also examined associations by type of SSB (sodas, fruit punches, and sport drinks), fruit juice (orange juice, and apple and “other” juices), and whole fruits, categorized as temperate fruits (apples, pears, grapes, berries, apricots) and tropical fruits (oranges, banana, melons). Analyses were performed using SAS version 9.4 (SAS Institute) and all statistical tests were 2-sided and P values < 0.05 were considered statistically significant. This study was reported in accordance with the STROBE guidelines.

RESULTS

During up to 25 years of follow-up, 1,625 (6.3%) participants reported a diagnosis of hypertension. At baseline, participants had a mean age of 12 years (SD = 1.86) and a mean BMI of 19.6 kg/m² (SD = 3.73). Participants ranged in age from 7 to 40 years over the study period, and by end of follow-up the mean age was 36 years (SD = 1.66). The cohort was 55% female and 96% non-Hispanic white. The median age at hypertension onset was 36 years (IQR range, 32 – 38 years). Age-adjusted baseline characteristics according to quintiles of TFE, and by categories of SSBs, fruit juice, and whole fruit in Table 1. Demographics across quintiles of TFE and across categories of SSBs, fruit juice, and whole fruit were largely consistent. However, those with the highest intakes of TFE, tended to have higher intakes of SSBs, fruit juice, and whole fruit. Those with the highest intakes of SSBs, fruit juice, and whole fruit, tended to have higher total energy intake, physical activity, and intake of TFE. Higher BMI was observed among individuals with the highest intake levels of SSBs and TFE and the lowest intake levels of fruit juice and whole fruit. Those who consumed more fruit juices and whole fruit were also more likely to take a multivitamin. Time trends in average consumption levels of TFE, SSBs, fruit juice, and whole fruit are presented in Figures S1 and S2.

Table 1.

Age-adjusted baseline characteristics of children in the GUTS across quintiles of TFE and categories of SSB, fruit juice, and whole fruit consumption

	Category 1	Category 2	Category 3	Category 4	Category 5
TFE	Q1	Q2	Q3	Q4	Q5
Median TFE, g/day	44.1	52.8	59.5	67.0	80.7
No of participants	5417	4179	4230	4783	7140
Age, years*	11.8 (1.9)	11.9 (1.8)	12.0 (1.9)	12.2 (1.9)	12.4 (1.8)
BMI, kg/m2	19.8 (4.0)	19.5 (3.8)	19.5 (3.6)	19.5 (3.7)	19.7 (3.7)
Non-Hispanic white, %	96	96	96	96	95
Boys, %	46	47	46	44	43
Physical Activity, METs/wk	99.9 (68.7)	105.0 (70.1)	106.4 (70.1)	105.5 (70.9)	107.6 (74.6)
Current smoker, %	9	10	9	10	12
Multivitamin use, %	39	42	43	44	43
Screen time, hrs/wk	12.2 (16.1)	12.7 (15.8)	13.4 (16.1)	13.5 (16.2)	15.1 (18.1)
Total energy, kcal/day	2088 (681)	2199 (696)	2209 (697)	2191 (725)	2087 (709)
SSB, servings/day	0.4 (0.4)	0.7 (0.6)	0.9 (0.7)	1.2 (0.9)	1.9 (1.3)
Diet soda, servings/day	0.2 (0.5)	0.2 (0.4)	0.2 (0.4)	0.2 (0.4)	0.2 (0.4)
Milk, servings/day	2.0 (1.0)	1.8 (1.0)	1.6 (1.0)	1.5 (1.0)	1.2 (1.0)
Fruit juice, servings/day	0.4 (0.4)	0.6 (0.5)	0.7 (0.6)	0.8 (0.6)	0.9 (0.7)
Whole fruit, servings/day	0.7 (0.5)	0.8 (0.6)	0.9 (0.6)	1.0 (0.7)	1.0 (0.7)
SSB	<3/wk	3/wk - <1/d	1 - <1.5/d	1.5/d - <2/d	2+/d
No of participants	7275	6755	4661	2722	4336
Age, years*	11.9 (1.9)	12.1 (1.9)	12.0 (1.8)	12.3 (1.8)	12.2 (1.8)
BMI, kg/m2	19.7 (3.9)	19.5 (3.7)	19.4 (3.6)	19.4 (3.5)	19.8 (3.9)
Non-Hispanic white, %	96	96	97	96	95
Boys, %	36	43	48	53	52
Physical Activity, METs/wk	92.7 (66.1)	100.5 (67.0)	109.0 (72.9)	117.8 (73.6)	119.2 (77.9)
Current smoker, %	8	9	11	12	13
Multivitamin use, %	43	43	42	40	41
Screen time, hrs/wk	12.8 (15.4)	13.9 (16.3)	13.3 (16.3)	14.3 (17.6)	13.8 (18.6)
Total energy, kcal/day	1805 (596)	2053 (621)	2199 (627)	2417 (690)	2622 (731)
TFE, g/day	52.8 (15.5)	60.1 (14.3)	65.1 (14.3)	69.5 (14.2)	86.0 (20.2)
Diet soda, servings/day	0.2 (0.5)	0.2 (0.4)	0.2 (0.4)	0.1 (0.4)	0.2 (0.4)
Milk, servings/day	1.6 (1.0)	1.6 (1.0)	1.6 (1.0)	1.5 (1.0)	1.4 (1.0)
Fruit juice, servings/day	0.6 (0.6)	0.6 (0.6)	0.7 (0.6)	0.8 (0.6)	0.8 (0.7)
Whole fruit, servings/day	0.9 (0.7)	0.9 (0.6)	0.9 (0.6)	0.9 (0.7)	0.9 (0.7)
Fruit Juice	<1/wk	1 - <4/wk	4/wk - <1/d	1 - <1.5/d	1.5+/d
No of participants	6953	4415	5533	5800	3048
Age, years*	12.1 (1.9)	12.2 (1.8)	12.1 (1.9)	12.0 (1.9)	12.0 (1.8)
BMI, kg/m2	19.9 (4.1)	19.7 (3.7)	19.5 (3.6)	19.4 (3.6)	19.3 (3.3)
Non-Hispanic white, %	96	96	95	96	96
Boys, %	43	43	46	46	47
Physical Activity, METs/wk	90.7 (64.6)	101.6 (68.5)	105.4 (68.9)	111.38 (73.0)	130.2 (81.9)
Current smoker, %	10	11	10	10	11
Multivitamin use, %	38	39	43	46	48
Screen time, hrs/wk	15.7 (19.2)	13.5 (16.5)	13.5 (16.0)	12.2 (15.0)	10.9 (14.5)
Total energy, kcal/day	1830 (632)	2008 (628)	2184 (652)	2345 (670)	2620 (724)
TFE, g/day	59.0 (20.6)	59.5 (17.9)	63.8 (16.7)	68.8 (18.0)	76.1 (18.6)
SSB, servings/day	1.0 (1.0)	1.1 (1.0)	1.1 (1.0)	1.2 (1.1)	1.4 (1.2)
Diet soda, servings/day	0.2 (0.5)	0.2 (0.4)	0.2 (0.4)	0.2 (0.4)	0.1 (0.4)
Milk, servings/day	1.4 (1.0)	1.5 (1.0)	1.6 (1.0)	1.7 (1.0)	1.7 (1.0)
Whole fruit, servings/day	0.6 (0.5)	0.8 (0.5)	0.9 (0.6)	1.0 (0.7)	1.3 (0.8)
Whole Fruit	<1/wk	1 - <4/wk	4/wk - <1/d	1 - <1.5/d	1.5+/d
No of participants	1723	8850	6357	4850	3969
Age, years*	12.2 (1.9)	12.2 (1.8)	12.1 (1.9)	12.0 (1.9)	12.1 (1.9)
BMI, kg/m2	19.9 (4.0)	19.8 (3.9)	19.6 (3.6)	19.4 (3.5)	19.4 (3.5)
Non-Hispanic white, %	95	96	96	96	96
Boys, %	56	46	45	43	41
Physical Activity, METs/wk	87.8	95.7	102.1	112.3	129.5
Current smoker, %	10	11	10	10	9
Multivitamin use, %	39	39	42	44	48
Screen time, hrs/wk	16.2 (21.6)	14.0 (17.4)	14.1 (16.8)	12.2 (14.7)	11.8 (14.1)
Total energy, kcal/day	1763 (664)	1908 (626)	2152 (631)	2339 (673)	2600 (723)
TFE, g/day	63.2 (25.1)	62.6 (20.5)	63.6 (18.4)	65.2 (17.4)	68.7 (16.8)
SSB, servings/day	1.3 (1.1)	1.1 (1.1)	1.1 (1.0)	1.1 (1.0)	1.1 (1.0)
Diet soda, servings/day	0.2 (0.5)	0.2 (0.5)	0.2 (0.4)	0.2 (0.4)	0.2 (0.4)
Milk, servings/day	1.3 (1.0)	1.5 (1.0)	1.6 (1.0)	1.7 (1.0)	1.7 (1.0)
Fruit juice, servings/day	0.3 (0.5)	0.5 (0.5)	0.7 (0.6)	0.8 (0.6)	1.0 (0.7)

Values are means (SD) or percentages, unless otherwise indicated. Table 1 presents age-standardized baseline characteristics standardized to the age distribution of the study population. TFE are presented across quintiles. METs/wk, metabolic equivalent hours per week; TFE, total fructose equivalents.

^*Value is not age adjusted

In multivariable models adjusting for lifestyle and other dietary factors, TFE was not associated with hypertension risk. Comparing the highest quintile of intake (> 81g/day) to the reference, the HR was 1.04 (95% CI, 0.89, 1.22; P trend < 0.001) (Table 2). After further adjusting for BMI (multivariable model 2), the association remained null (HR, 1.07; 95% CI, 0.92, 1.25; P trend < 0.001). Similarly, each 5% increment in energy from TFE showed no significant association with hypertension risk (HR, 1.01; 95% CI, 0.99, 1.04). Dose-response analysis using restricted cubic splines indicates no significant departure from linearity (P = 0.22) (Figure 1).

Table 2.

Hypertension risk according to intake of total fructose equivalents in the GUTS

TFE	Q1	Q2	Q3	Q4	Q5	p Trend	HR (95% CI) per 5% increment in energy from TFE
Median TFE (g/day)	44.1	52.8	59.5	67.0	80.7
Cases/ person-years	305 / 76,870	296 / 78,215	328 / 78,561	334 / 78,801	362 / 78,537
age-adjusted model	1	0.89 (0.76, 1.04)	0.94 (0.80, 1.09)	0.92 (0.79, 1.08)	1.03 (0.88, 1.20)	<0.001	1.01 (1.00, 1.03)
multivariable model 1	1	0.90 (0.76, 1.05)	0.95 (0.81, 1.11)	0.94 (0.80, 1.10)	1.04 (0.89, 1.22)	<0.001	1.02 (0.99, 1.04)
multivariable model 2	1	0.90 (0.77, 1.06)	0.96 (0.82, 1.13)	0.97 (0.83, 1.13)	1.07 (0.92, 1.25)	<0.001	1.01 (0.99, 1.04)

Data from 25,749 participants in the GUTS. Multivariable model 1 was adjusted for age (years), race (white, non-white), physical activity (METs/wk, tertiles), screen time (hrs/wk, tertiles), sleep (hrs/d, tertiles), multivitamin use (yes or no), smoking (yes or no), total energy (quintiles), and intake of vegetables, whole grains, red and processed meat (quintiles), and alcohol intake (continuous). Multivariable model 2 was further adjusted for body mass index (quintiles). HR, hazard ratio; CI, confidence interval.

Figure 1. Dose-response analysis for TFE, SSB, fruit juice, and whole fruit and risk of hypertension in the GUTS.

Multivariable adjusted restricted cubic spline analysis of the association between TFE, SSBs, fruit juice, and whole fruit on risk of hypertension. P values were linear p < 0.001 and non-linear p = 0.22 for TFE; linear p < 0.001 and non-linear p = 0.23 for SSBs; linear p < 0.005 and non-linear p = 0.09 for fruit juice; linear p = 0.046 and non-linear p = 0.74 for whole fruit. Hazard ratios were adjusted for age, race, physical activity, screen time, sleep, multivitamin use, smoking, total energy, intake of fruits, vegetables, whole grains, red and processed meat, alcohol intake and body mass index. SSBs, fruit juice, and whole fruit were mutually adjusted for. The y-axis is displayed on a logarithmic scale.

In multivariable adjusted models, higher intake of SSBs was significantly associated with a higher risk of hypertension (Table 3). Compared to the reference group (< 3/week), individuals in the highest category of SSB intake (≥ 2/day) had a higher risk of developing hypertension (HR, 1.61; 95% CI, 1.34, 1.93; P trend < 0.001). After further adjusting for BMI (multivariable model 2), the association was slightly attenuated but remained significant (HR, 1.52; 95% CI, 1.27, 1.83; P trend < 0.001). Each serving per day increment of SSB was associated with a 14% higher risk of hypertension (HR, 1.14; 95% CI, 1.07, 1.22). Dose-response analysis from spline regression models showed no significant evidence of departure from linearity (P = 0.23) (Figure 1). An analysis of SSB subtypes showed that each daily serving of soda and sports drinks was significantly associated with a 23% and 36% higher risk of hypertension, respectively (HR, 1.23; 95% CI, 1.11, 1.37 and HR, 1.36; 95% CI, 1.13, 1.63) (Table S1). In contrast, fruit punch consumption was not associated with hypertension risk.

Table 3.

Hypertension risk according to intake of SSBs, fruit juice, and whole fruit in the GUTS

	Category 1	Category 2	Category 3	Category 4	Category 5	p Trend	HR (95% CI) per 1 serving per day increment
SSB	<3/wk	3/wk - <1/day	1 - <1.5/day	1.5 - <2/day	2+/day
Cases/ person-years	406 / 116,074	549 / 132,072	298 / 67,234	151 / 36,068	221 / 39,536
age-adjusted model	1	1.13 (1.00, 1.29)	1.21 (1.04, 1.41)	1.18 (0.98, 1.42)	1.71 (1.45, 2.02)	<0.001	1.19 (1.12, 1.26)
multivariable model 1	1	1.09 (0.96, 1.25)	1.15 (0.98, 1.34)	1.12 (0.92, 1.36)	1.61 (1.34, 1.93)	<0.001	1.17 (1.09, 1.25)
multivariable model 2	1	1.09 (0.96, 1.25)	1.14 (0.97, 1.33)	1.12 (0.92, 1.37)	1.52 (1.27, 1.83)	<0.001	1.14 (1.07, 1.22)
Fruit Juice	<1/wk	1/wk - <4/wk	4/wk - <1/day	1/day - <1.5/day	1.5+/day
Cases/ person-years	288 / 75,380	580 / 129,184	385 / 98,684	258 / 62,242	114 / 25,493
age-adjusted model	1	1.03 (0.89, 1.18)	0.92 (0.79, 1.07)	1.03 (0.87, 1.22)	1.14 (0.92, 1.42)	0.51	1.02 (0.92, 1.12)
multivariable model 1	1	1.05 (0.91, 1.22)	0.98 (0.83, 1.15)	1.12 (0.93, 1.35)	1.29 (1.01, 1.63)	0.07	1.09 (0.97, 1.22)
multivariable model 2	1	1.08 (0.93, 1.25)	1.00 (0.85, 1.18)	1.20 (1.00, 1.44)	1.35 (1.06, 1.71)	0.018	1.12 (1.00, 1.25)
Whole Fruit	<1/wk	1/wk - <4/wk	4/wk - <1/day	1/day - <1.5/day	1.5+/day
Cases/ person-years	83 / 16,986	537 / 120,798	522 / 116,373	301 / 81,769	182 / 55,057
age-adjusted model	1.00	0.89 (0.71, 1.13)	0.89 (0.71, 1.13)	0.76 (0.59, 0.97)	0.71 (0.55, 0.93)	<0.001	0.85 (0.78, 0.93)
multivariable model 1	1.00	0.89 (0.70, 1.12)	0.90 (0.71, 1.15)	0.80 (0.61, 1.04)	0.79 (0.59, 1.06)	0.07	0.91 (0.82, 1.02)
multivariable model 2	1.00	0.88 (0.70, 1.11)	0.90 (0.71, 1.15)	0.80 (0.62, 1.04)	0.79 (0.59, 1.05)	0.08	0.91 (0.82, 1.02)

Data from 25,749 participants in the GUTS. Multivariable model 1 was adjusted for age (years), race (white, non-white), physical activity (METs/wk, tertiles), screen time (hrs/wk, tertiles), sleep (hrs/d, tertiles), multivitamin use (yes or no), smoking (yes or no), total energy (quintiles), and intake of fruits, vegetables, whole grains, red and processed meat (quintiles), and alcohol intake (continuous). Multivariable model 2 was further adjusted for body mass index (quintiles). SSBs, fruit juice, and whole fruit were mutually adjusted for. HR, hazard ratio; CI, confidence interval.

In multivariable adjusted models, higher intake of fruit juice was significantly associated with a higher risk of hypertension (Table 3). Compared to the reference group (< 1/week), individuals in the highest category of fruit juice intake (≥ 1.5/day) had a higher risk of developing hypertension (HR, 1.29; 95% CI, 1.01, 1.63; P trend = 0.06). The significant association remained after further adjusting for BMI (HR, 1.35; 95% CI, 1.06, 1.71; P trend = 0.02) (multivariable model 2). Each serving per day increment of fruit juice was marginally associated with hypertension risk (HR, 1.12; 95% CI, 1.00, 1.25). Dose-response analysis showed no significant evidence of departure from linearity (P = 0.09); however, the curve resembled a J-shaped like pattern, with risk estimates trending towards a lower risk at low intakes of fruit juice (Figure 1). As for subtypes of fruit juice, each daily serving of orange juice was significantly associated with a 20% higher risk (HR, 1.20; 95% CI, 1.02, 1.41), while apple and other juices was not associated with hypertension (Table S2).

In multivariable adjusted models, higher intake of whole fruit (≥ 1.5/day) compared to reference (< 1/week) was not associated with hypertension risk (HR, 0.79; 95% CI, 0.59, 1.06; P trend = 0.07) (Table 3). When further adjusting for BMI (multivariable model 2), the association remained the same (HR, 0.79; 95% CI, 0.59, 1.06; P trend = 0.08). Each daily serving of whole fruit was not associated with hypertension risk (HR, 0.91; 95% CI, 0.82, 1.02). Dose-response analyses showed no significant evidence of departure from linearity (P = 0.74) (Figure 1). Analysis by whole fruit subtype showed no significant association of temperate or tropical fruit and risk of hypertension (Table S3).

Stratified analyses for TFE, SSBs, fruit juice, and whole fruit by age, sex, racial or ethnic groups, BMI, physical activity, and multivitamin use are presented in Table S4-S7. Stratification by physical activity for TFE and risk of hypertension showed that individuals below the median level of physical activity had a higher risk of hypertension compared to those above the median level (HR, 1.07; 95% CI, 1.00, 1.14 and HR, 0.95; 95% CI, 0.85, 1.05, respectively, p-interaction=0.01). Sub-group analyses for SSBs, fruit juice, and whole fruit were largely consistent across strata.

Our substitution analysis for SSBs and fruit juice is presented in Figure 2. Replacing one serving per day of SSB with one serving per day of whole fruit, milk, or water was associated with lower risk of hypertension by 22%, 13%, and 9%, respectively (HR, 0.78; 95% CI, 0.69, 0.89, HR, 0.87; 95% CI, 0.80, 0.95, and HR, 0.91; 95% CI, 0.85, 0.97, respectively). Replacing one serving per day of fruit juice with one serving per day of whole fruit was associated with 19% lower risk of hypertension (HR, 0.81; 95% CI, 0.68, 0.97). No association was found for replacing SSB with fruit juice or replacing fruit juice with milk or water.

Figure 2. Substitution analysis for serving-to-serving substitution of SSBs and fruit juice for other foods and beverages on hypertension risk in the GUTS.

Hazard ratios were adjusted for age, race, physical activity, screen time, sleep, multivitamin use, smoking, total energy, intake of fruits, vegetables, whole grains, red and processed meat, alcohol intake and body mass index. The x-axis is displayed on a logarithmic scale.

In our sensitivity analyses of assessing baseline intake, to represent intakes in childhood and adolescence, and the most recent measure of intake (simple updates), the results remained largely consistent with our primary analyses (Table S8-S9). Results were also largely consistent when baseline BMI was added to the multivariable adjusted models.

DISCUSSION

In this large prospective cohort study following participants from childhood through to adulthood, we found higher intakes of TFE from all sources were not associated with hypertension risk. However, when examining major sources of fructose, we found higher intakes of SSBs and fruit juice were associated with a higher risk of hypertension, while higher intakes of whole fruit were not related to hypertension. Substituting one serving per day of SSBs for whole fruit, milk, and water, and substituting fruit juice for whole fruit, were associated with a lower risk of hypertension.

Our largely null findings for total fructose consumption and hypertension risk are likely due to the differing nutrient matrices of fructose containing food and beverage sources. In this cohort, SSBs, fruit juice, and whole fruit were the major contributors to total fructose intake, yet these foods differ substantially in their nutrient compositions and metabolic effects. Our findings are in line with results from the NHS and the Health Professionals Follow-up Study (HPFS) which found that fructose from added sugar and juice later in midlife were associated with a higher risk of coronary heart disease, while fructose from whole fruits was not.²⁹ Similarly, another study using participants from the UK Biobank found that fructose from beverages was positively associated with hypertension risk, while fructose from solid food sources was inversely associated.³⁰ These findings underscore the need to consider the whole food source rather than the nutrient alone, as different nutrient matrixes may lead to distinct effects on health outcomes.

SSBs are the primary source of total free sugars in the diet, typically composed of high-fructose corn syrup or sucrose. A large body of evidence has consistently linked SSB consumption to adverse cardiometabolic health outcomes, including hypertension.^{12, 13, 31–33} Our findings are consistent with results from the NHS, NHSII, and HPFS cohorts which indicate that 1 serving/day of SSB intake is associated with a 13% higher risk of hypertension.³³ A previous analysis using the GUTS I cohort alone was largely consistent with ours.³⁴ The present analysis builds on this work by increasing sample size and follow-up time, as well as examining associations across categories of intake. We further assessed dose-response and substitution analyses to provide a more comprehensive understanding of beverage intake and hypertension risk. Additionally, systematic reviews and meta-analyses in children and adults further support these findings.^{35, 36} A systematic review and meta-analysis of observational studies in children found that high SSB intake was associated with a 1.67 mmHg higher SBP and a 36% higher risk of hypertension, with no significant difference in DBP.³⁶ However, the majority of studies were cross sectional or over short follow-up periods, and therefore did not evaluate the long-term risk of hypertension into adulthood. A meta-analysis of cohort studies in adults found a significant linear dose-response association between SSB intake and hypertension risk.³⁵

Our study, which spans a critical life stage from childhood through to adulthood, provides evidence that SSB consumption throughout early life contributes to a higher risk of hypertension in later years. We found a significant linear association, indicating that each serving per day of SSB was associated with a higher risk of hypertension. In line with a recent study assessing SSBs and CVD, we found no significant difference in the association between SSB consumption and hypertension risk across low and high physical activity groups, suggesting that even among physically active individuals, SSB intake remains a risk factor.³⁷ Our study also found that replacing SSBs with milk, water, or whole fruit was associated with lower hypertension risk. The observed association with milk may be attributed in part to the calcium, potassium, and bioactive peptides in milk, which have been shown to support vascular function and lower blood pressure.³⁸ Replacing SSBs with water, the recommended beverage of choice, and whole fruit, a nutrient-dense food, were also suitable alternatives. Interestingly, our analysis of SSB subtypes revealed that both soda and sports drinks were positively associated with hypertension risk. While the association with soda is well established, the link between sport drinks and hypertension is particularly noteworthy, as these beverages are often marketed as health-promoting. To our knowledge, this study represents one of the largest prospective evaluations of sports drink consumption and hypertension risk to date. These beverages are often consumed in high quantities and contain added sugars delivering fructose and sodium, which may contribute to hypertension through excess energy, increased sodium retention, and altered fluid balance.³⁹ These findings highlight the importance of considering all sources of SSBs when evaluating dietary risk factors for hypertension.

The association between fruit juice consumption and cardiometabolic risk, including hypertension, remains an area of ongoing debate, with mixed findings in the literature. Some studies have reported positive associations with several cardiometabolic disease risk factors,^{16, 40–42} while others have found no association,^{15, 32, 43} and some have reported inverse associations.^{14, 17} A recent study found that children introduced to fruit juice at an earlier age had higher blood pressure, heart rate, mean arterial pressure, and greater odds of developing hypertension later in childhood compared to those introduced to juice at a later age.¹⁶ A systematic review and meta-analysis of prospective cohort studies in adults found a dose-dependent association between fruit juice intake and hypertension, showing protective associations at moderate intake levels and harmful associations at high intake levels.³⁵ In our study, we found a significant association between high fruit juice consumption (≥ 1.5 servings/ day) and hypertension risk. Our dose-response analysis suggests a possible J-shaped pattern, indicating a trend where low fruit juice consumption was not associated with hypertension risk and may even confer potential benefits, while higher intake levels may be linked to harm. Of note, while this was not a significant non-linear association, the trend indicates the importance of dose when considering fruit juice intake. At low to moderate doses of fruit juice, the potential benefit from vitamins, minerals, and polyphenols may counteract the adverse metabolic effects of excess fructose and energy at higher doses. When we assessed fruit juice subtypes, we found that a one serving/day increase of orange juice was associated with a higher risk of hypertension, whereas apple and other juice types were not. This finding may be due to potential misclassification of orange juice intake during childhood. Orange-flavoured fruit drinks with added sugars are commonly consumed by younger children and may have been reported as orange juice intake, thus biasing the association towards one more reflective of SSBs. Further research exploring specific fruit juices and hypertension risk is warranted. Importantly, our study found that replacing fruit juice with whole fruit was associated with a lower risk of hypertension, further reinforcing the benefits of consuming whole fruit over the fruit juice form.

Whole fruit consumption has consistently been associated with a reduced risk of hypertension and cardiometabolic disease, supporting its role in a healthful diet.^{14, 35, 41} Numerous guidelines including those from the Dietary Guidelines for Americans, Canadian Dietary Guidelines, and the Mediterranean and DASH diets, emphasize whole fruit as a key pillar of a healthy dietary pattern.^{18, 20, 44, 45} In our study, we observed a trend towards a lower risk of hypertension with whole fruit intake, although the association did not reach statistical significance. When we assessed whole fruit subtypes, we found no differences in hypertension risk between temperate and tropical fruits. A recent study suggested differences in whole fruit subtypes and risk of T2D, possibly due to variations in glycemic index and phytochemical contents.⁴⁶ Our findings reinforce dietary recommendations that prioritize whole fruit consumption over fruit juice or SSBs.

Several biological mechanisms may explain the differential associations observed for SSBs, fruit juice, and whole fruit and hypertension risk. Excessive fructose intake, particularly from liquid sources, is metabolized in the liver, where it can lead to increased production of uric acid. Elevated uric acid has been shown to result in endothelial dysfunction, contributing to increased blood pressure.⁴⁷ Additionally, excess fructose intake promotes hepatic do novo lipogenesis, increasing the production of triglycerides and very-low-density lipoproteins (VLDLs).^{48, 49} This process can contribute to dyslipidemia and the accumulation of ectopic fat, which further exacerbates cardiometabolic dysfunction.^{48, 49} As liquid sources of fructose, both SSBs and fruit juice share similar metabolic effects. One key mechanism is their potential to promote excess weight gain, a well-established risk factor for hypertension.⁷ Unlike solid foods, liquid calories are less satiating and can lead to an incomplete compensation of energy at subsequent meals.⁵⁰ For example, a typical 8 oz glass of orange juice contains the equivalent of about 3 whole oranges, which is more than typically consumed in a single sitting. Additionally, studies have shown that the consumption of liquid calories result in greater weight gain than the consumption of solid calories.⁵¹ While fruit juice retains some of the benefits from whole fruit, it contains lower amounts of dietary fibre, polyphenols, and other nutrients that are lost in processing.

The strengths of our study include the long study duration, large sample size, and repeated measures of diet and lifestyle, which enhance the reliability of our findings. We also conducted several sensitivity and subgroup analyses, as well as a substitution analysis to further explore the robustness of our results. Our study also includes several limitations that should be acknowledged. Due to the observational nature of our study, residual confounding cannot be ruled out despite adjusting for numerous potential confounders in our analyses. Dietary intake was assessed through FFQs, which are subject to measurement error and recall bias. To account for this, we utilized the cumulative average intakes across multiple questionnaire cycles. Additionally, the use of last observation carried forward for missing data and cumulative averaging may introduce some bias if missingness is informative or if dietary intake changes over time. Furthermore, the FFQs have been validated against dietary recalls with moderate validity.²³ Additionally, hypertension and BMI were self-reported in this cohort, however, prior validation studies have shown self-reported hypertension and self-reported height and weight to have good validity.^{25, 26} Lastly, our study population was predominantly non- Hispanic white which limits the generalizability of the findings. Future research in more racially and ethnically diverse cohorts is warranted.

In conclusion, our study provides evidence that high consumption of SSBs and fruit juice is associated with a higher risk of hypertension. Importantly, replacing SSBs with whole fruit, water, or milk, and replacing fruit juice with whole fruit, may help lower hypertension risk. By assessing diet across early life stages, our study provides evidence linking life-long dietary patterns and lasting health consequences. Given the rising prevalence of hypertension, our findings support dietary guidelines that emphasize whole fruit consumption while limiting the intake of free sugars from SSBs and fruit juice from an early age.

CLINICAL PERSPECTIVE.

What Is New?

• In a large prospective cohort, higher intake of sugar-sweetened beverages and fruit juice was associated with a higher risk of hypertension, whereas total fructose intake and whole fruit consumption were not.

• Substituting sugar-sweetened beverages with milk, water, or whole fruit, and substituting fruit juice with whole fruit, was associated with a lower risk of hypertension.

• By following participants from childhood into adulthood, this study provides evidence that early dietary patterns are associated with long-term hypertension risk.

What Are the Clinical Implications?

• Our findings support clinical guidelines and public health policies that limit the overconsumption of sugar-sweetened beverages and fruit juice to reduce long-term hypertension risk.

• Whole fruit, water, and milk represent suitable alternatives to sugar-sweetened beverages and fruit juice and may be encouraged as part of hypertension prevention strategies beginning early in life.

• Dietary recommendations should emphasize reducing liquid sources of free sugars while promoting whole fruit consumption rather than the fruit juice form.

Non-standard Abbreviations and Acronyms

BMI

body mass index

CI

confidence intervals

CVD

cardiovascular disease

FFQ

food frequency questionnaire

GUTS

Growing Up Today Study

HPFS

Health Professionals Follow-up Study

HR

hazard ratios

MET

metabolic equivalent

NHS

Nurses’ Health Study

SD

standard deviations

SSB

sugar-sweetened beverages

T2D

type 2 diabetes

TFE

total fructose equivalents

Data Sharing:

Data in this manuscript will be made available upon request.