Sugar-Sweetened Beverage, Sugar Intake of Individuals, and Their Blood Pressure
The obesity epidemic has focused attention on relationships of sugars and sugar-sweetened beverages (SSBs) to cardiovascular risk factors. Here we report cross-sectional associations of SSBs, diet beverages, and sugars with blood pressure (BP) for United Kingdom and US participants of the International Study of Macro/Micronutrients and Blood Pressure. Data collected include four 24-hour dietary recalls, two 24-hour urine collections, 8 BP readings, and questionnaire data for 2696 people ages 40 to 59 years of age from 10 US/United Kingdom population samples. Associations of SSBs, diet beverages, and sugars (fructose, glucose, and sucrose) with BP were assessed by multiple linear regression. SSB intake related directly to BP, with P values of 0.005 to <0.001 (systolic BP) and 0.14 to <0.001 (diastolic BP). SSB intake higher by 1 serving per day (355 mL/24 hours) was associated with systolic/diastolic BP differences of +1.6/+0.8 mm Hg (both P<0.001) and +1.1/+0.4 mm Hg (P<0.001/<0.05) with adjustment for weight and height. Diet beverage intake was inversely associated with BP (P 0.41 to 0.003). Fructose- and glucose-BP associations were direct, with significant sugar-sodium interactions: for individuals with above-median 24-hour urinary sodium excretion, fructose intake higher by 2 SD (5.6% kcal) was associated with systolic/diastolic BP differences of +3.4/+2.2 mm Hg (both P<0.001) and +2.5/+1.7 mm Hg (both P=0.002) with adjustment for weight and height. Observed independent, direct associations of SSB intake and BP are consistent with recent trial data. These findings, plus adverse nutrient intakes among SSB consumers, and greater sugar-BP differences for persons with higher sodium excretion lend support to recommendations that intake of SSBs, sugars, and salt be substantially reduced.
See Editorial Commentary, pp 676–678
Adverse blood pressure (BP), prevalent worldwide, is an independent major risk factor for cardiovascular diseases.1 Public health measures are needed to address this problem, with an emphasis on primary and primordial prevention.2 Established modifiable risk factors for elevated BP are high sodium intake, inadequate potassium intake, high body mass index (BMI), and excessive alcohol intake.3,4 Other dietary factors possibly related to adverse BP levels include lower intakes of calcium, magnesium, phosphorus,5 iron,6 vegetable protein,7 glutamic acid,8 polyunsaturated fatty acids,9,10 and starch11 and higher intakes of cholesterol,12 animal protein, and red meat.6,7
The Western obesity epidemic has focused attention on the relationships to cardiovascular disease risk factors of diets rich in added sugars, particularly glucose, sucrose, and fructose, for example, as high-fructose corn syrup, abundant in sugar-sweetened beverages (SSBs).13–15 Animal data indicate direct pressor effects of glucose, fructose, and sucrose on BP.16–20 Limited short-term human trial data are mostly compatible with animal findings21–23; observational and long-term trial data are inconsistent.24–29 The most compelling evidence to date comes from the PREMIER Study, a behavioral intervention trial of 810 prehypertensive and hypertensive individuals, where reduced intake of SSBs or sugar over 18 months was associated with reduced BP.30
Here we report cross-sectional associations with BP of SSBs, diet (noncaloric-sweetened) beverages, and sugars (fructose, glucose, and sucrose) for 2696 participants of the International Study of Macro/Micronutrients and Blood Pressure (INTERMAP) from 10 population samples in the United States and the United Kingdom.
Population Samples, Field Methods (1996–1999)
INTERMAP surveyed 4680 men and women ages 40 to 59 years from Japan (4 samples), People's Republic of China (3 samples), United Kingdom (2 samples), and the United States (8 samples). We focus here on the 2696 US and United Kingdom participants, because SSB and diet beverage intake was negligible in the Japanese and Chinese samples. Participants were randomly recruited from general and occupational populations.31 Each participant attended 4 times, the first 2 visits on consecutive days, the second 2 visits on consecutive days on average 3 weeks later. For BP measurement, each participant, having emptied his/her bladder, was seated for 5 minutes, feet flat on the floor, in a quiet room, with no physical activity and no eating, drinking, or smoking in the preceding 30 minutes. BP was measured twice at each visit with a random-zero sphygmomanometer; Korotkoff sounds I and V were criteria for systolic BP and diastolic BP. Measurements of height and weight and questionnaire data on daily alcohol consumption over the previous 7 days were obtained at 2 visits. Dietary data were collected at each visit by a trained interviewer with use of the in-depth multi-pass 24-hour recall method.32 Questionnaire data were obtained on possible confounders. Each participant provided two 24-hour urine collections, start and end timed at the research center; measurements included urinary volume and sodium, potassium, calcium, magnesium, urea, and creatinine levels.31 Urinary sodium, potassium, and urea excretion were used to validate dietary intake of sodium, potassium, and protein; correlations ranged from r=0.42 to r=0.55.32 The study received institutional ethics committee approval for each site; all of the participants gave written consent; study procedures were in accordance with institutional guidelines.
Sugar-sweetened and diet beverage intakes were estimated from food records. SSBs included uncarbonated and carbonated soft drinks (eg, soda), fruit drinks (excluding 100% fruit juices), and lemonade but excluded diet beverages. Diet beverages included uncarbonated and carbonated drinks sweetened with artificial (noncaloric) sweeteners. Dietary data were converted to nutrient intakes (83 nutrients) with use of enhanced country-specific food tables, standardized across countries by the Nutrition Coordinating Center, University of Minnesota.32,33 Measurements per person were averaged across the 4 visits for beverage, nutrient, and BP variables and across the 2 collections for 24-hour urinary variables.
Reliability as a measure of possible regression dilution bias34 for beverage, nutrient, and BP variables (expressed as the observed univariate regression coefficient as a percentage of the theoretical “true” coefficient) was estimated by the following formula: 1/[1+(ratio/2)]×100. The ratio is intraindividual variance divided by interindividual variance, calculated from mean intakes/BP levels of the first and second 2 visits to account for higher correlation between intakes/BP levels on consecutive days.35
Associations among dietary variables were explored by partial Pearson correlation, adjusted for age, sex, and sample, pooled by country. Multiple regression analyses assessed relations to systolic and diastolic BPs of each person's intake of SSBs and diet beverages (milliliters per 24 hours, models adjusted for energy intake), fructose, glucose, and sucrose (percentage of kilocalories). Four models were used, and each controlled successively for a larger number of possible nondietary and dietary confounders, with and without adjustment for weight and height, followed by an additional series of sensitivity analyses that included censored normal regression to adjust for potential antihypertensive treatment bias.36 US and United Kingdom regression coefficients were pooled (weighted by inverse of their variance). A test for heterogeneity was done to examine differences between US and United Kingdom regression coefficients. Age, sex, BMI, and sodium interactions were assessed by interaction terms in regression models. Departure from linearity was tested with squared terms.
Analyses were done with SAS 9.1 (SAS Institute; by I.J.B.). Statistical tests were 2 sided. Main findings are presented as BP differences associated with beverage intake higher by 1 serving (355 mL/24 hours) or with sugar intake higher by 2 SD; statistical significance is expressed as z scores (regression coefficient/SE; z≥1.96, P≤0.05; z≥2.58, P<0.01; z≥3.29, P<0.001), uncorrected for regression dilution bias or multiple testing.
Mean systolic/diastolic BP was 118.6/73.4 mm Hg in the United States and 120.4/77.3 mm Hg in the United Kingdom (please see Table S1 in the online Data Supplement at http://hyper.ahajournals.org). Mean SSB and diet beverage intakes were higher in the United States than the United Kingdom, with a mean SSB intake of 0.9 servings per day (306 mL/24 hours) in the United States and 0.2 servings per day (66 mL/24 hours) in the United Kingdom. Expressed as milliliters per 1000 kcal, SSB intake was higher in men and diet beverage intake higher in women. Fructose, glucose, and sucrose intakes (percentage of kilocalories) were higher in the United States than the United Kingdom, and similar in men and women.
Nutrient Intakes and Other Variables by Category of Beverage Intake
Compared with participants who consumed no SSBs, adjusted mean energy intake was higher by 120 kcal/24 hours for those who consumed ≤1 serving per day (≤355 mL/24 hours) and was higher by 397 kcal/24 hours for those who consumed >1 serving per day (Table 1). Mean intakes of starch, fiber, protein (animal and vegetable), polyunsaturated and monounsaturated fatty acids, alcohol, minerals, caffeine (variables expressed as percentage of kilocalories or amount per 1000 kcal), and urinary potassium excretion were lowest in those consuming >1 serving per day. Fructose, glucose, sucrose intake, and urinary sodium/potassium ratio were highest in those consuming >1 serving per day. Mean BMI was lowest for nonconsumers (28.4 kg/m2) and highest for those consuming >1 serving per day (30.2 kg/m2). Mean participant age and years of education completed were lowest in the highest category of SSB consumption; physical activity, BMI, and BP (systolic and diastolic) were highest in the same category. Findings were consistent for men and women analyzed separately (data not tabulated).
|Variable||SSB Intake*||F Score||P|
|0 Servings per Day (N=808)||≤1 Serving per Day (N=1153)†||>1 Serving per Day (N=735)†|
|Mean||95% CI||Mean||95% CI||Mean||95% CI|
|SSBs, mL/1000 kcal||0.0||… to …||69.1||62.2 to 76.0||288.2||279.7 to 296.6||…||…|
|Diet beverages, mL/1000 kcal||151.8||136.9 to 166.7||69.1||54.3 to 84.0||2.0||0.0 to 20.2||129.87||<0.001|
|Energy, kcal/24 h||2043||1994 to 2092||2163||2114 to 2212||2440||2380 to 2499||86.59||<0.001|
|Fructose, % kcal||3.1||2.9 to 3.3||4.1||3.9 to 4.3||7.4||7.2 to 7.7||635.65||<0.001|
|Glucose, % kcal||3.4||3.2 to 3.6||4.3||4.1 to 4.5||7.3||7.1 to 7.5||643.62||<0.001|
|Sucrose, % kcal||9.2||8.7 to 9.6||10.1||9.6 to 10.5||11.5||11.0 to 12.1||40.74||<0.001|
|Starch, % kcal||25.5||25.1 to 26.0||24.3||23.8 to 24.8||20.9||20.4 to 21.5||132.91||<0.001|
|Fiber, g/1000 kcal||11.8||11.6 to 12.1||10.8||10.5 to 11.1||9.1||8.7 to 9.4||124.25||<0.001|
|Animal protein, % kcal||10.5||10.2 to 10.8||10.2||9.9 to 10.4||9.6||9.3 to 10.0||12.76||<0.001|
|Vegetable protein, % kcal||6.2||6.0 to 6.3||5.7||5.6 to 5.9||4.9||4.7 to 5.0||145.91||<0.001|
|Total SFA, % kcal||11.2||10.9 to 11.5||11.2||11.0 to 11.5||10.9||10.6 to 11.3||2.33||0.10|
|Total MFA, % kcal||11.5||11.3 to 11.8||11.6||11.4 to 11.9||11.3||11.0 to 11.6||3.02||0.049|
|Total PFA, % kcal||6.6||6.4 to 6.8||6.7||6.5 to 6.9||6.4||6.2 to 6.6||4.75||0.009|
|Omega-3 PFA, % kcal||0.75||0.72 to 0.77||0.75||0.73 to 0.78||0.69||0.66 to 0.72||10.49||<0.001|
|Omega-6 PFA, % kcal||5.9||5.8 to 6.1||6.0||5.9 to 6.2||5.7||5.5 to 6.0||4.65||0.01|
|Trans-fatty acids, % kcal||1.6||1.5 to 1.7||1.6||1.5 to 1.7||1.6||1.5 to 1.6||1.32||0.27|
|Dietary cholesterol, mg/1000 kcal||126.9||121.9 to 131.9||127.4||122.4 to 132.4||131.2||125.1 to 137.3||1.16||0.31|
|Keys dietary lipid score‡||37.7||36.9 to 38.6||37.8||36.9 to 38.6||37.7||36.6 to 38.7||0.02||0.98|
|Phosphorus, mg/1000 kcal||679.5||669.4 to 689.6||635.9||625.9 to 646.0||567.2||554.9 to 579.5||158.19||<0.001|
|Magnesium, mg/1000 kcal||166.8||163.8 to 169.8||153.5||150.5 to 156.5||126.5||122.8 to 130.2||232.14||<0.001|
|Calcium, mg/1000 kcal||444.1||432.6 to 455.6||412.3||400.8 to 423.8||350.2||336.2 to 364.3||86.22||<0.001|
|Iron, mg/1000 kcal||7.8||7.6 to 8.0||7.2||7.0 to 7.4||5.8||5.6 to 6.1||118.01||<0.001|
|Vitamin C, mg/1000 kcal||48.8||45.8 to 51.7||50.5||47.6 to 53.5||47.7||44.1 to 51.3||1.59||0.20|
|Caffeine, mg/1000 kcal||151.1||140.4 to 161.9||120.4||109.7 to 131.1||91.4||78.2 to 104.5||39.15||<0.001|
|14-d alcohol, g/24 h||10.8||9.5 to 12.1||10.1||8.9 to 11.4||8.6||7.1 to 10.2||3.93||0.02|
|Urinary sodium, mmol/24 h||156.0||151.3 to 160.8||154.7||149.9 to 159.4||153.1||147.3 to 158.9||0.47||0.63|
|Urinary potassium, mmol/24 h||66.2||64.5 to 67.8||62.5||60.9 to 64.2||56.0||53.9 to 58.0||47.96||<0.001|
|Urinary sodium/potassium ratio||2.5||2.4 to 2.6||2.7||2.6 to 2.8||3.1||3.0 to 3.2||40.47||<0.001|
|Age, y||50.4||49.9 to 50.8||49.9||49.5 to 50.4||48.6||48.0 to 49.2||19.09||<0.001|
|Education, years completed||13.8||13.6 to 14.1||13.6||13.3 to 13.8||12.9||12.6 to 13.2||19.27||<0.001|
|Moderate and heavy physical activity, h/24 h||2.3||2.1 to 2.6||2.5||2.2 to 2.7||3.4||3.0 to 3.7||23.99||<0.001|
|Body mass index, kg/m2||28.4||27.9 to 28.9||28.6||28.1 to 29.1||30.2||29.6 to 30.8||22.35||<0.001|
|Systolic BP, mm Hg||118.8||117.6 to 120.0||119.5||118.3 to 120.6||122.5||121.1 to 123.9||14.54||<0.001|
|Diastolic BP, mm Hg||73.5||72.7 to 74.3||74.0||73.2 to 74.9||75.5||74.5 to 76.5||8.13||<0.001|
|Variable||%||95% CI||%||95% CI||%||95% CI||F Score||P|
|Obese§||31.4||27.4 to 35.5||32.5||28.5 to 36.5||44.8||39.9 to 49.8||17.87||<0.001|
|Hypertensive∥||28.7||25.0 to 32.4||29.5||25.8 to 33.2||34.5||30.0 to 39.0||3.82||0.02|
Nutrient intakes among diet beverage consumers were mostly higher than nonconsumers (please see Table S2); exceptions were sugars and vitamin C (lowest in those consuming >1 serving per day), as well as energy, fiber, omega-3 polyunsaturated fatty acids, cholesterol, alcohol, urinary sodium/potassium ratio (no difference). Diet beverage consumers had higher mean BMI than nonconsumers and lower physical activity. No differences in participant age, education, or BP were observed.
Reliability estimates for SSBs were 80% (United States) and 58% (United Kingdom); for diet beverages the estimates were 91% (United States) and 85% (United Kingdom; please see Table S3). Reliability estimates for fructose, glucose, and sucrose ranged from 68% (sucrose, United States) to 81% (sucrose, United Kingdom). BP reliability estimates were uniformly high (>90%).
Intakes of SSBs, fructose, and glucose (amount per 24 hours, adjusted for sample, age, and sex) were positively correlated, SSBs with fructose or glucose were r=0.72; fructose and glucose were r=0.94 (Table S4); correlations with sucrose were positive, smaller than the foregoing. Expressed as a proportion of energy intake, SSB intake was similarly correlated with sugars and inversely correlated with starch, fiber, vegetable protein, minerals, and urinary potassium (r=−0.25 to −0.37; Table S5). No correlations r>0.25 were observed for diet beverage intake (Table S5). Expressed as percentage of kilocalories, fructose and glucose intake were positively correlated with vitamin C (r=0.40 and 0.43, respectively; Table S5). Fructose, glucose, and sucrose were inversely correlated with starch, animal protein, fatty acids, alcohol, minerals, and urinary electrolytes (r=−0.02 to −0.37).
Associations with systolic BP were consistently direct (z=2.82 to 4.98; P=0.005 to <0.001), in models adjusted separately for potential confounders including vegetable protein, minerals, and caffeine (Table 2). In model 3, adjusted for energy, urinary sodium, potassium, dietary alcohol, cholesterol, and polyunsaturated and saturated fatty acids, SSB intake higher by 1 serving per day (355 mL/24 hours) was associated with a systolic BP difference of +1.6 mm Hg (z=4.98; P<0.001) and +1.1 mm Hg (z=3.40; P<0.001) with control for weight and height. Associations with diastolic BP were direct (z=1.47 to 3.42; P=0.14 to <0.001). BP differences/z scores were larger in censored normal regressions and subgroup analyses excluding individuals with high day-to-day variability in nutrient intakes or BP, smaller in subgroup analyses of nonhypertensive participants (please see Table S6), and similar in models adjusted for fructose, glucose, or sucrose intake (data not tabulated). SSB-BMI interactions with P<0.05 were observed for 7 of 8 systolic BP models; in stratified analyses, direct SSB-BP associations were stronger for individuals with lower BMI (Table S7). SSB-sodium interactions were nonsignificant; in stratified analyses, direct SSB-BP associations were stronger for individuals with higher 24-hour urinary sodium excretion (Table S8).
|Beverage Model||Systolic BP||Diastolic BP|
|Not Adjusted for Weight and Height||Adjusted for Weight and Height||Not Adjusted for Weight and Height||Adjusted for Weight and Height|
|Difference, mm Hg||95% CI||z||Difference, mm Hg||95% CI||z||Difference, mm Hg||95% CI||z||Difference, mm Hg||95% CI||z|
|1||1.26||0.68 to 1.84||4.28||1.05||0.50 to 1.60||3.74||0.61||0.21 to 1.01||3.02||0.49||0.11 to 0.87||2.51|
|2||1.35||0.75 to 1.95||4.38||0.94||0.36 to 1.52||3.20||0.61||0.20 to 1.02||2.90||0.37||−0.03 to 0.77||1.81|
|3||1.60||0.97 to 2.23||4.98||1.05||0.44 to 1.66||3.40||0.76||0.32 to 1.20||3.42||0.43||0.004 to 0.86||1.98|
|4a (vegetable protein)||1.36||0.67 to 2.05||3.87||0.99||0.33 to 1.65||2.95||0.58||0.11 to 1.05||2.42||0.36||−0.10 to 0.82||1.54|
|4b (calcium)||1.42||0.78 to 2.06||4.36||0.91||0.29 to 1.53||2.90||0.66||0.22 to 1.10||2.95||0.35||−0.08 to 0.78||1.61|
|4c (magnesium)||1.25||0.57 to 1.93||3.59||0.96||0.30 to 1.62||2.86||0.60||0.13 to 1.07||2.52||0.43||−0.03 to 0.89||1.83|
|4d (phosphorus)||1.41||0.76 to 2.06||4.25||0.90||0.27 to 1.53||2.82||0.63||0.18 to 1.08||2.77||0.33||−0.11 to 0.77||1.47|
|4e (caffeine)||1.61||0.98 to 2.24||4.98||1.08||0.47 to 1.69||3.49||0.74||0.31 to 1.17||3.36||0.43||0.01 to 0.85||2.00|
|1||−0.27||−0.78 to 0.24||−1.03||−0.68||−1.16 to −0.20||−2.75||−0.26||−0.60 to 0.08||−1.49||−0.50||−0.83 to −0.17||−2.94|
|2||−0.32||−0.82 to 0.18||−1.25||−0.58||−1.07 to −0.09||−2.34||−0.28||−0.63 to 0.07||−1.59||−0.43||−0.76 to −0.10||−2.52|
|3||−0.35||−0.86 to 0.16||−1.34||−0.58||−1.07 to −0.09||−2.32||−0.30||−0.65 to 0.05||−1.70||−0.44||−0.78 to −0.10||−2.52|
|4a (vegetable protein)||−0.28||−0.80 to 0.24||−1.06||−0.54||−1.03 to −0.05||−2.17||−0.26||−0.61 to 0.09||−1.46||−0.42||−0.76 to −0.08||−2.40|
|4b (calcium)||−0.31||−0.83 to 0.21||−1.17||−0.54||−1.03 to −0.05||−2.15||−0.28||−0.63 to 0.07||−1.57||−0.42||−0.76 to −0.08||−2.40|
|4c (magnesium)||−0.28||−0.79 to 0.23||−1.08||−0.55||−1.04 to −0.06||−2.18||−0.27||−0.62 to 0.08||−1.53||−0.43||−0.77 to −0.09||−2.48|
|4d (phosphorus)||−0.22||−0.74 to 0.30||−0.83||−0.48||−0.97 to 0.01||−1.91||−0.22||−0.57 to 0.13||−1.23||−0.38||−0.72 to −0.04||−2.16|
|4e (caffeine)||−0.36||−0.87 to 0.15||−1.38||−0.60||−1.09 to −0.11||−2.41||−0.30||−0.65 to 0.05||−1.68||−0.44||−0.78 to −0.10||−2.55|
Associations with systolic and diastolic BPs were consistently inverse (z=−0.83 to −2.94; P=0.41 to 0.003). In model 3, diet beverage intake higher by 1 serving per day was associated with a systolic BP difference of −0.35 mm Hg (z=−1.34; P=0.18) and −0.58 mm Hg (z=−2.32; P=0.02) with control for weight and height (Table 2). BP differences/z scores were larger when diet beverage intake was expressed as a proportion of energy intake and for censored normal regressions, and was smaller in subgroup analyses of nonhypertensive participants (Table S9). Diet beverage-BMI interactions (P<0.05) were detected for 3 of 8 diastolic BP models. In stratified analyses, inverse diet beverage-BP associations were stronger for individuals with higher BMI (Table S7).
Associations of fructose and glucose with BP were direct, and BP differences and z scores were smaller than those observed for SSB (z=0.23 to 3.14; P=0.82 to 0.002; Table S10). Sucrose-BP associations were bidirectional and BP differences and z scores small (Table S10). Fructose- and glucose-sodium interactions (P<0.05) were observed for all of the models. In stratified analyses, fructose- and glucose-related BP differences were observed only for individuals with higher urinary sodium excretion. BP differences and z scores were large: in model 3, fructose intake higher by 2 SD (5.6% kcal) was associated with a systolic BP difference of +3.4 mm Hg (z=4.01; P<0.001) and +2.5 mm Hg (z=3.10; P=0.002) with control for weight and height (Table 3). Glucose-BP associations were of a similar magnitude (Table 3).
|Sugar (2 SD) Subcohort, Model||Systolic BP*||Diastolic BP*|
|Not Adjusted for Weight or Height||Adjusted for Weight and Height||Not Adjusted for Weight or Height||Adjusted for Weight and Height|
|Difference, mm Hg||95% CI||z||Difference, mm Hg||95% CI||z||Difference, mm Hg||95% CI||z||Difference, mm Hg||95% CI||z|
|Fructose (5.6% kcal)|
|Lower sodium (N=1349)|
|3||0.36||−1.17 to 1.89||0.46||−0.09||−1.56 to 1.38||−0.12||0.00||… to …||0.00||−0.32||−1.33 to 0.69||−0.62|
|4c (magnesium)||−0.13||−1.72 to 1.46||−0.16||−0.33||−1.87 to 1.21||−0.42||−0.26||−1.37 to 0.85||−0.46||−0.42||−1.49 to 0.65||−0.77|
|Higher sodium (N=1347)|
|3||3.40||1.74 to 5.06||4.01||2.50||0.92 to 4.08||3.10||2.20||1.07 to 3.33||3.81||1.71||0.61 to 2.81||3.05|
|4c (magnesium)||2.70||0.97 to 4.43||3.06||2.27||0.63 to 3.91||2.71||1.97||0.79 to 3.15||3.28||1.74||0.60 to 2.88||2.98|
|Glucose (5.1% kcal)|
|Lower sodium (N=1349)|
|3||0.42||−1.10 to 1.94||0.54||−0.15||−1.70 to 1.40||−0.19||0.13||1.15 to −0.89||−0.25||−0.27||−1.29 to 0.75||−0.52|
|4c (magnesium)||−0.15||−1.78 to 1.48||−0.18||−0.45||−2.03 to 1.13||−0.56||−0.16||−1.28 to 0.96||−0.28||−0.38||−1.44 to 0.68||−0.70|
|Higher sodium (N=1347)|
|3||3.68||2.03 to 5.33||4.36||2.69||1.11 to 4.27||3.34||2.14||1.02 to 3.26||3.73||1.61||0.51 to 2.71||2.88|
|4c (magnesium)||2.94||1.20 to 4.68||3.31||2.47||0.81 to 4.13||2.92||1.91||0.72 to 3.10||3.15||1.66||0.51 to 2.81||2.82|
The main findings here are a direct association of SSB consumption with BP, and direct associations of fructose and glucose intake with BP that were stronger among individuals with higher urinary sodium excretion.
Observed direct associations of SSB with BP are compatible with the findings of the PREMIER intervention trial, where reduction in SSB consumption by 355 mL/d was associated with systolic/diastolic BP lower by 1.8/1.1 mm Hg and 0.7/0.4 mm Hg, with adjustment for change in body weight.30 For diet beverages, findings were similar to INTERMAP, that is, inverse and nonsignificant in multivariate models; and caffeine had no association. In INTERMAP, SSB-BP associations were independent of caffeine, and caffeine intake was inversely associated with SSB consumption. Although some SSBs, for example, cola, are important sources of caffeine, it is likely that SSB consumption displaced coffee and tea consumption (main dietary sources of caffeine) for many individuals. In analyses of women from the Nurses Health Studies I and II, sugared and diet cola consumptions, but not caffeine consumption, were associated with risk of incident hypertension.37 Among adolescents of the National Health and Nutrition Examination Survey 1999–2004, SSB intake was associated with systolic BP and serum uric acid concentration (see below).28
To our knowledge, no observational studies have reported associations of glucose intake with BP. Forman et al27 found no link between fructose intake (assessed by food frequency questionnaire) and incident hypertension among >200 000 women and men of the Nurses' Health Study I and II, and the Health Professionals Follow-up Study; Jalal et al29 reported a direct association between fructose intake and odds of elevated BP in cross-sectional analysis of 4528 adults from the National Health and Nutrition Examination Survey 2003–2006.
The direct associations reported here for SSB/fructose intake and BP are consistent with the hypothesized effect on the uric acid pathway. Fructose consumption may lead to increased serum uric acid via phosphorylation of fructose by hepatocytes and generation of adenosine diphosphate, which is metabolized to uric acid38; raised serum uric acid may influence BP by reducing levels of NO, a potent vasodilator.39 Sugar consumption has also been linked to enhanced sympathetic nervous system activity and sodium retention.21,40 Detection of significant interaction with sodium excretion, that is, direct fructose- and glucose-BP associations stronger for individuals with higher urinary sodium excretion, is compatible with the findings of several animal studies.41–44 He et al45 reported that SSB consumption was directly associated with salt intake (assessed by 7-day dietary record) in United Kingdom children and adolescents. Here, sodium excretion was not associated with SSB consumption in United States and United Kingdom adults; however, urinary sodium/potassium ratio was directly associated with SSB. Significant interaction with BMI, that is, direct SSB-BP associations weaker for individuals with higher BMI, could be attributed to greater misclassification of SSB intake in this subgroup because of differential underreporting of SSB intake.46
Limitations of the INTERMAP findings include their cross-sectional nature; underestimation of effect size, attributable to limited reliability in the measurement of nutrients (ie, regression dilution bias, despite repeated measures, although observed BP differences were of similar magnitude to the PREMIER intervention trial); possible systematic bias (likely minimized by observer training, standardization, multipass methods, open nonleading questioning, and extensive ongoing quality control); and possible residual confounding. There was little evidence from multiple sensitivity analyses to indicate substantial bias. SSB-, glucose-, and fructose-BP associations were reduced with control for weight and height. Interpretation of this finding is problematic: if intakes of SSBs/sugars act on BP through positive energy balance and increased body mass, then body mass is in the causal pathway, and statistical control for weight (standardized for height) is overadjustment.34 Findings adjusted for BMI (not presented here) were quantitatively similar to those adjusted for weight and height. We are presently unable to quantify high-fructose corn syrup; however, SSB intake may be a good proxy, because high-fructose corn syrup is the most common caloric sweetener used by the US beverage industry.14 Fructose intake was higher and urinary potassium and fiber intake lower for participants consuming >1 serving per day of SSBs compared with those consuming ≤1 serving per day, indicating that higher fructose intake in SSB consumers likely reflects high-fructose corn syrup consumption rather than fruit intake. Because INTERMAP was designed primarily as a study of individual-level diet-BP associations, the samples were not intended to be nationally representative, but, given the heterogeneity of the 8 US samples in particular, and the similarity of US and United Kingdom SSB-BP associations, it is reasonable to infer that findings may be applicable to middle-aged US and United Kingdom men and women.
Higher intake of SSBs was associated with more adverse overall nutritional quality, and there were independent direct associations of SSBs, fructose, and glucose with BP; sugar-BP associations were stronger among higher-sodium consumers. These findings are consistent with recent trial data30 and lend support to recommendations for reducing intake of SSBs, added sugars, and salt for the improvement cardiovascular health.
The INTERMAP Study was accomplished through the fine work of staff at local, national, and international centers; a partial listing of colleagues is published in J Hum Hypertens (2003;17:603–606).
Sources of Funding
The PhD work of I.J.B. was supported by a
Lawes CMM, Vander Hoorn S, Law MR, Elliott P, MacMahon S, Rodgers A. Blood pressure and the global burden of disease 2000: part I–estimates of blood pressure levels; part II–estimates of attributable burden. J Hypertens. 2006; 24:413–430.CrossrefMedlineGoogle Scholar
Lenfant C. Task force on research in epidemiology and prevention of cardiovascular diseases. Circulation. 1994; 90:2609–2617.LinkGoogle Scholar
Stamler J, Neaton JD, Garside DB, Daviglus ML. Current status: six established major risk factors–and low risk. In: , Marmot M, Elliott P, eds. Coronary Heart Disease Epidemiology. Oxford, United Kingdom: Oxford University Press; 2005:32–70.Google Scholar
Elliott P, Stamler J. Primary prevention of high blood pressure. In: , Marmot M, Elliott P, eds. Coronary Heart Disease Epidemiology. Oxford, United Kingdom: Oxford University Press; 2005:751–768.Google Scholar
Elliott P, Kesteloot H, Appel LJ, Dyer AR, Ueshima H, Chan Q, Brown IJ, Zhao L, Stamler J. Dietary phosphorus and blood pressure: INTERMAP Study. Hypertension. 2008; 51:669–675.LinkGoogle Scholar
Tzoulaki I, Brown IJ, Chan Q, Van Horn L, Ueshima H, Zhao L, Stamler J, Elliott P. Relation of iron and red meat intake to blood pressure: cross sectional epidemiological study. Brit Med J. 2008; 337:a258.CrossrefMedlineGoogle Scholar
Elliott P, Stamler J, Dyer AR, Appel L, Dennis B, Kesteloot H, Ueshima H, Okayama A, Chan Q, Garside DB, Zhou B. Association between protein intake and blood pressure: the INTERMAP Study. Arch Intern Med. 2006; 166:79–87.CrossrefMedlineGoogle Scholar
Stamler J, Brown IJ, Daviglus ML, Chan Q, Kesteloot H, Ueshima H, Zhao L, Elliott P. Glutamic acid, the main dietary amino acid, and blood pressure: the INTERMAP Study (International Collaborative Study of Macronutrients, Micronutrients and Blood Pressure). Circulation. 2009; 120:221–228.LinkGoogle Scholar
Ueshima H, Stamler J, Elliott P, Chan Q, Brown IJ, Carnethon M, Daviglus ML, He K, Moag-Stahlberg A, Rodriguez BL, Steffen LM, Van Horn L, Yarnell J, Zhou B. Food omega-3 fatty acid intake of individuals (total, linolenic acid, long-chain) and their blood pressure: INTERMAP Study. Hypertension. 2007; 50:313–319.LinkGoogle Scholar
Miura K, Stamler J, Nakagawa H, Elliott P, Ueshima H, Chan Q, Brown IJ, Tzoulaki I, Saitoh S, Dyer AR, Daviglus ML, Kesteloot H, Okayama A, Curb JD, Rodriguez BL, Elmer PJ, Steffen LM, Robertson C, Zhao L. Relationship of dietary linoleic acid to blood pressure: the International Study of Macro-Micronutrients and Blood Pressure. Hypertension. 2008; 52:408–414.LinkGoogle Scholar
Brown IJ, Elliott P, Robertson CE, Chan Q, Daviglus ML, Dyer AR, Huang C-C, Rodriguez BL, Sakata K, Ueshima H, Van Horn L, Zhao L, Stamler J. Dietary starch intake of individuals and their blood pressure: the international study of macronutrients and micronutrients and blood pressure. J Hypertens. 2009; 27:231–236.CrossrefMedlineGoogle Scholar
Sakurai M, Stamler J, Miura K, Brown IJ, Nakagawa H, Elliott P, Ueshima H, Chan Q, Tzoulaki I, Dyer AR, Okayama A, Zhao L. Relationship of dietary cholesterol to blood pressure: the INTERMAP study. J Hypertens. 2011; 29:222–228.CrossrefMedlineGoogle Scholar
Johnson RJ, Segal MS, Sautin Y, Nakagawa T, Feig DI, Kang D-H, Gersch MS, Benner S, Sanchez-Lozada LG. Potential role of sugar (fructose) in the epidemic of hypertension, obesity and the metabolic syndrome, diabetes, kidney disease, and cardiovascular disease. Am J Clin Nutr. 2007; 86:899–906.MedlineGoogle Scholar
Bray GA. Fructose: should we worry?Int J Obes. 2008; 32:S127–S131.CrossrefGoogle Scholar
Johnson RK, Appel LJ, Brands M, Howard BV, Lefevre M, Lustig RH, Sacks F, Steffen LM, Wylie-Rosett J. Dietary sugars intake and cardiovascular health: a scientific statement from the American Heart Association. Circulation. 2009; 120:1011–1020.LinkGoogle Scholar
Fournier RD, Chiueh CC, Kopin IJ, Knapka JJ, DiPette D, Preuss HG. Refined carbohydrate increases blood pressure and catecholamine excretion in SHR and WKY. Am J Physiol. 1986; 250:E381–E385.MedlineGoogle Scholar
el Zein M, Areas JL, Knapka J, MacCarthy P, Yousufi AK, DiPette D, Holland B, Goel R, Preuss HG. Excess sucrose and glucose ingestion acutely elevate blood pressure in spontaneously hypertensive rats. Am J Hypertens. 1990; 3:380–386.CrossrefMedlineGoogle Scholar
el Zein M, Areas JL, Preuss HG. Long-term effects of excess sucrose ingestion on three strains of rats. Am J Hypertens. 1990; 3:560–562.CrossrefMedlineGoogle Scholar
Preuss HG, Knapka JJ, MacCarthy P, Yousufi AK, Sabnis SG, Antonovych TT. High sucrose diets increase blood pressure of both salt-sensitive and salt-resistant rats. Am J Hypertens. 1992; 5:585–591.CrossrefMedlineGoogle Scholar
Song D, Hutchings S, Pang CCY. Chronic N-acetylcysteine prevents fructose-induced insulin resistance and hypertension in rats. Eur J Pharmacol. 2005; 508:205–210.CrossrefMedlineGoogle Scholar
Rebello T, Hodges RE, Smith JL. Short-term effects of various sugars on antinatriuresis and blood pressure changes in normotensive young men. Am J Clin Nutr. 1983; 38:84–94.CrossrefMedlineGoogle Scholar
Black RNA, Spence M, McMahon RO, Cuskelly GJ, Ennis CN, McCance DR, Young IS, Bell PM, Hunter SJ. Effect of eucaloric high- and low-sucrose diets with identical macronutrient profile on insulin resistance and vascular risk: a randomized controlled trial. Diabetes. 2006; 55:3566–3572.CrossrefMedlineGoogle Scholar
Brown CM, Dulloo AG, Yepuri G, Montani J-P. Fructose ingestion acutely elevates blood pressure in healthy young humans. Am J Physiol Regul Integr Comp Physiol. 2008; 294:730–737.CrossrefMedlineGoogle Scholar
Israel KD, Michaelis OE, Reiser S, Keeney M. Serum uric acid, inorganic phosphorus, and glutamic-oxalacetic transaminase and blood pressure in carbohydrate-sensitive adults consuming three different levels of sucrose. Ann Nutr Metab. 1983; 27:425–435.CrossrefMedlineGoogle Scholar
Joffres MR, Reed DM, Yano K. Relationship of magnesium intake and other dietary factors to blood pressure: the Honolulu Heart Study. Am J Clin Nutr. 1987; 45:469–475.CrossrefMedlineGoogle Scholar
Stamler J, Caggiula AW, Grandits GA. Relation of body mass and alcohol, nutrient, fiber, and caffeine intakes to blood pressure in the special intervention and usual care groups in the Multiple Risk Factor Intervention Trial. Am J Clin Nutr. 1997; 65:338S–365S.CrossrefMedlineGoogle Scholar
Forman JP, Choi H, Curhan GC. Fructose and vitamin C intake do not influence risk for developing hypertension. J Am Soc Nephrol. 2009; 20:863–871.CrossrefMedlineGoogle Scholar
Nguyen S, Choi HK, Lustig RH, Hsu C-Y. Sugar-sweetened beverages, serum uric acid, and blood pressure in adolescents. J Pediatr. 2009; 154:807–813.CrossrefMedlineGoogle Scholar
Jalal DI, Smits G, Johnson RJ, Chonchol M. Increased fructose associates with elevated blood pressure. J Am Soc Nephrol. 2010; 21:1543–1549.CrossrefMedlineGoogle Scholar
Chen L, Caballero B, Mitchell DC, Loria C, Lin P-H, Champagne CM, Elmer PJ, Ard JD, Batch BC, Anderson CAM, Appel LJ. Reducing consumption of sugar-sweetened beverages is associated with reduced blood pressure: a prospective study among United States adults. Circulation. 2010; 121:2398–2406.LinkGoogle Scholar
Stamler J, Elliott P, Dennis B, Dyer AR, Kesteloot H, Liu K, Ueshima H, Zhou B. INTERMAP: background, aims, design, methods, and descriptive statistics (non-dietary). J Hum Hypertens. 2003; 17:591–608.CrossrefMedlineGoogle Scholar
Dennis B, Stamler J, Buzzard M, Conway R, Elliott P, Moag-Stahlberg A, Okayama A, Okuda N, Robertson C, Robinson F, Schakel S, Stevens M, Van Heel N, Zhao LC, Zhou BF. INTERMAP: the dietary data–process and quality control. J Hum Hypertens. 2003; 17:609–622.CrossrefMedlineGoogle Scholar
Schakel SF, Dennis BH, Wold AC, Conway R, Zhao L, Okuda N, Okayama A, Moag-Stahlberg A, Robertson C, Van Heel N, Buzzard IM, Stamler J. Enhancing data on nutrient composition of foods eaten by the participants in the INTERMAP Study in China, Japan, the United Kingdom, and the United States. J Food Comp Anal. 2003; 16:395–408.CrossrefGoogle Scholar
Liu K. Measurement error and its impact on partial correlation and multiple linear regression analyses. Am J Epidemiol. 1988; 127:864–874.CrossrefMedlineGoogle Scholar
Grandits GA, Bartsch GE, Stamler J. Chapter 4: method issues in dietary data analysed in the Multiple Risk Factor Intervention Trial. Am J Clin Nutr. 1997; 65(suppl):211S–227S.CrossrefMedlineGoogle Scholar
Tobin MD, Sheehan NA, Scurrah KJ, Burton PR. Adjusting for treatment effects in studies of quantitative traits: antihypertensive therapy and systolic blood pressure. Stat Med. 2005; 24:2911–2935.CrossrefMedlineGoogle Scholar
Winkelmayer WC, Stampfer MJ, Willett WC, Curhan GC. Habitual caffeine intake and the risk of hypertension in women. JAMA. 2005; 294:2330–2335.CrossrefMedlineGoogle Scholar
Hallfrisch J. Metabolic effects of dietary fructose. FASEB J. 1990; 4:2652–2660.CrossrefMedlineGoogle Scholar
Nakagawa T, Hu H, Zharikov S, Tuttle K, Short RA, Glushakova O, Ouyang X, Feig DI, Block ER, Herrera-Acosta J, Patel JM, Johnson RJ. A causal role for uric acid in fructose-induced metabolic syndrome. Am J Physiol Renal. 2006; 290:F625–F631.CrossrefMedlineGoogle Scholar
Rowe JW, Young JB, Minaker KL, Stevens AL, Pallotta J, Landsberg L. Effect of insulin and glucose infusions on sympathetic nervous system activity in normal man. Diabetes. 1981; 30:219–225.CrossrefMedlineGoogle Scholar
Ahrens RA, Demuth P, Lee MK, Majkowski JW. Moderate sucrose ingestion and blood pressure in the rat. J Nutr. 1980; 110:725–731.CrossrefMedlineGoogle Scholar
Martinez FJ, Rizza RA, Romero JC. High-fructose feeding elicits insulin resistance, hyperinsulinism, and hypertension in normal mongrel dogs. Hypertension. 1994; 23:456–463.LinkGoogle Scholar
Preuss HG. Interplay between sugar and salt on blood pressure in spontaneously hypertensive rats. Nephron. 1994; 68:385–387.CrossrefMedlineGoogle Scholar
Kotchen TA, Morley Kotchen J. Dietary sodium and blood pressure: interactions with other nutrients. Am J Clin Nutr. 1997; 65(suppl):708S–711S.CrossrefMedlineGoogle Scholar
He FJ, Marrero NM, MacGregor GA. Salt intake is related to soft drink consumption in children and adolescents: a link to obesity?Hypertension. 2008; 51:629–634.LinkGoogle Scholar
Pryer JA, Vrijheid M, Nichols R, Kiggins M, Elliott P. Who are the ‘low energy reporters’ in the dietary and nutritional survey of British adults?Int J Epidemiol. 1997; 26:146–154.CrossrefMedlineGoogle Scholar