Sugars and risk of mortality in the NIH-AARP Diet and Health Study^,

Abstract

Background: Although previous studies have linked intake of sugars with incidence of cancer and other chronic diseases, its association with mortality remains unknown.

Objective: We investigated the association of total sugars, added sugars, total fructose, added fructose, sucrose, and added sucrose with the risk of all-cause, cardiovascular disease, cancer, and other-cause mortality in the NIH-AARP Diet and Health Study.

Design: The participants (n = 353,751), aged 50–71 y, were followed for up to 13 y. Intake of individual sugars over the previous 12 mo was assessed at baseline by using a 124-item NIH Diet History Questionnaire.

Results: In fully adjusted models (fifth quartile compared with first quartile), all-cause mortality was positively associated with the intake of total sugars [HR (95% CI): 1.13 (1.06, 1.20); P-trend < 0.0001], total fructose [1.10 (1.04, 1.17); P-trend < 0.0001], and added fructose [1.07 (1.01, 1.13); P-trend = 0.005) in women and total fructose [1.06 (1.01, 1.10); P-trend = 0.002] in men. In men, a weak inverse association was found between other-cause mortality and dietary added sugars (P-trend = 0.04), sucrose (P-trend = 0.03), and added sucrose (P-trend = 0.006). Investigation of consumption of sugars by source showed that the positive association with mortality risk was confined only to sugars from beverages, whereas the inverse association was confined to sugars from solid foods.

Conclusions: In this large prospective study, total fructose intake was weakly positively associated with all-cause mortality in both women and men, whereas added sugar, sucrose, and added sucrose intakes were inversely associated with other-cause mortality in men. In our analyses, intake of added sugars was not associated with an increased risk of mortality. The NIH-AARP Diet and Health Study was registered at clinicaltrials.gov as NCT00340015.

INTRODUCTION

Consumption of sugars has been increasing worldwide and parallels the global rising prevalence of chronic diseases, such as cancer, type 2 diabetes, and cardiovascular disease (CVD)⁵. There has been growing epidemiologic and experimental evidence to suggest that diets high in sugars increase triglyceride concentrations, stimulate fatty acid synthesis, decrease HDL cholesterol, raise uric acid concentrations, and therefore increase CVD risk (1–4). Long-term exposure of the pancreatic β cells to high concentrations of fatty acids may impair β-cell function and lead to hyperinsulinemia and insulin resistance (5, 6). Postprandial hyperglycemia induced by intake of readily available sugars can also lead to hyperinsulinemia (7). Diets high in sugars may promote carcinogenesis by inducing the synthesis of insulin and insulin-like growth factor I (8), producing advanced glycation end products (AGEs) (9), increasing oxidative stress (10, 11), or promoting weight gain (12, 13). Most of the adverse metabolic effects of sugars have been attributed to fructose, and multiple plausible mechanisms have been suggested (14). The widespread presence of sucrose (50% fructose) and high-fructose corn syrup (42–55% fructose) as sweeteners in the food supply, especially in the United States (15), has stimulated more interest in the possible health effects of added sugars and fructose on obesity (16, 17), the metabolic syndrome (18, 19), hyperuricemia (20, 21), hypertension (22, 23), type 2 diabetes (24, 25), nonalcoholic fatty liver disease (26), and cancer (27).

Although all sugars are chemically indistinguishable, added sugars (ie, sugars added at the table or used as ingredients in processed or prepared foods and drinks) have been more often associated with adverse health effects (16, 19, 28, 29), given that they are sourced from foods that are energy-dense and nutrient- and phytochemical-poor. Moreover, the physiologic effect of sugars per se may also depend on the physical properties of the food itself (30). For example, sugars that are inherently built into a food's cellular structure and accompanied with micronutrients and bioactive compounds, such as naturally occurring sugars in fruit and vegetables, may have metabolic consequences different from those of readily available sugars occurring in high concentration or free in solution in highly processed, fiber-depleted, rapidly digestible foods (31, 32). No study thus far has examined the association of total and added sugars or of individual sugars with mortality. Therefore, we investigated the effect of different types of sugars, including total sugars, sucrose, and fructose, and the effect of added sugars on all-cause and cause-specific mortality with the use of data from a large US prospective study. We hypothesize that intake of sugars, particularly fructose and added sugars, is associated with an increased risk of mortality.

SUBJECTS AND METHODS

Study population

The NIH-AARP Diet and Health Study is a prospective cohort of approximately half a million men and women aged 50–71 y from 6 states in the United States (California, Florida, Louisiana, New Jersey, North Carolina, and Pennsylvania) and 2 metropolitan areas (Atlanta, GA, and Detroit, MI). During 1995–1996, a self-administered baseline questionnaire with questions on demographic characteristics, personal and family medical history, diet, and other lifestyle factors, was mailed to 3.5 million members of the AARP, formerly known as the American Association for Retired Persons. Of the 617,119 members who returned the questionnaire, 567,169 completed it satisfactorily. More detail on the study design has already been reported (33). The NIH-AARP Diet and Health Study was approved by the Special Studies Institutional Review Board of the US National Cancer Institute.

From 567,169 participants, we excluded subjects with duplicate questionnaires (n = 179), those who moved out of the study areas or died before entry (n = 582), withdrew from the study (n = 6), were proxy responders (n = 15,760), who reported having poor health (n = 8365), or were prevalent cases of any cancer except non-melanoma skin cancer (n = 51,154), end-stage renal disease (n = 769), heart disease or stroke (n = 70,533), diabetes (n = 29,965), or gallbladder disease (n = 32,913) at baseline. We further excluded subjects with extreme energy intake (ie, beyond twice the IQR above the 75th or below the 25th percentile of sex-specific Box-Cox transformed energy intake) (n = 3188). The final baseline cohort included 353,751 participants (206,371 men and 147,380 women).

Cohort follow-up and case ascertainment

Follow-up was calculated from baseline (1995–1996) until censoring on 31 December 2008 or death, whichever came first. Cohort members were followed periodically for change of address, via the US Postal Service or by using the information sent from participants themselves. Deaths were ascertained by annual linkage to the US Social Security Administration Death Master File. Confirmation of the vital status and information on underlying causes of death were then obtained through follow-up searches of the National Death Index.

Cancer mortality was compiled from deaths as a result of any cancer [International Classification of Diseases, 9th revision (ICD9): 140–203.8; International Classification of Diseases, 10th revision (ICD10): C00–C44, C45.0, C45.1, C45.7, C45.9, C48–C97). CVD mortality included deaths from diseases of the heart, hypertension (without heart disease), cerebrovascular diseases, atherosclerosis, aortic aneurysm, and dissection and other diseases of the arteries, arterioles, and capillaries (ie, ICD9: 390–398, 401–404, 410–438, 440–448; ICD10: I00–I09, I10–I13, I20–I51, I60–I78). Other-cause mortality included death from any known causes, except for cancer and CVD. All-cause mortality is a combination of all of the above-mentioned causes of deaths, including deaths with an unknown cause.

Dietary assessment

Participants’ usual diet over the previous 12 mo was assessed at baseline by a self-administered semiquantitative food-frequency questionnaire with 124 food items developed at the National Cancer Institute—the Diet History Questionnaire (DHQ) (34). A question in the DHQ inquired about whether participants usually drank sugar-free or regular-calorie type beverages and what kind of sweetener they usually added to coffee and tea (sugar or honey, Equal or aspartame, saccharin or Sweet-n-Low, or other sweetener). DHQ data were processed by using the USDA 1994–1996 Continuing Survey of Food Intakes by Individuals (35), whereas individual monosaccharides and disaccharides were estimated based on the University of Minnesota's Nutrition Data System for Research database version 5.0_35 (2004). Total sugars represent the sum of monosaccharides (ie, glucose, fructose, and galactose) and disaccharides (ie, sucrose, lactose, and maltose; see Supplemental figure 1 under “Supplemental data” in the online issue). Added sugars are the sum of any monosaccharides and disaccharides used as ingredients in processed and prepared foods, soft and alcoholic drinks, jams and jellies, candies, and ice cream as well as sugars eaten separately or added to foods at the table and were estimated by using the USDA MyPyramid Equivalents Database, version 1 (36). The database contains added sugars values for 2038 foods across 23 food groups and disaggregates individual foods to their basic ingredients and assigns them into 32 MyPyramid major groups and subgroups. The added sugars MyPyramid subgroup, expressed in terms of teaspoon equivalents, was converted into grams (1 tsp = 4.2 g added sugars) (36). Fructose represents total fructose and included free fructose and fructose from sucrose. We also estimated added sucrose and added fructose, which were generated by summing sucrose and fructose from added sugars, plus sucrose and fructose from fruit juice, plus half of the sucrose and half of the fructose from apple sauce and dried fruit, respectively. Naturally occurring sucrose and fructose in fruit juice was considered as added, because during processing, fruits’ cellular structure is being disrupted and thereby makes sugars freely available in solution. Half of total fructose and sucrose in applesauce and dried fruit were also considered as added, given that those foods are semiprocessed and retain some physical characteristics of the original fruit (37). In an earlier DHQ validation study, the correlations between added sugars intake from the DHQ and four 24-h dietary recalls, adjusted for total energy intake, were 0.68 in men and 0.79 in women (38).

Statistical analysis

HRs and 95% CIs were estimated by sex-specific Cox proportional hazards regression models, with age as the underlying time metric. All variables of sugars were categorized into quintiles based on sex-specific cutoffs. To test linear trends, participants were assigned a score by using the median value for their quintiles. The scores were then entered as a continuous variable in a regression model. All multivariable models were adjusted for age, race, education, marital status, BMI, smoking (summary variable combining smoking status, smoking intensity, and time since quitting), physical activity, and dietary intakes of total energy, vegetables, and alcohol. In addition, multivariable models for cancer mortality included total fat and red meat intakes and first-degree family history of cancer, and models for CVD mortality included intake of saturated and polyunsaturated fats, history of hypertension, hypercholesterolemia, and use of aspirin. Models for all-cause mortality included all covariates present in any of the models. All dietary variables were categorized into quintiles, except for total energy intake, the distribution of which was normalized by using Box-Cox transformation, and entered in the models as a continuous variable. All dietary variables, except for alcohol, were energy adjusted by using the nutrient-density method. Our models investigated the substitution effect in which total energy intake was kept constant; thus, the reported findings represent the effect of substituting sugars for other energy-contributing nutrients. We created separate models for total sugars, added sugars, total fructose, and sucrose, whereas models with added fructose were adjusted for naturally occurring fructose (which equaled total fructose minus added fructose). Similarly, models with added sucrose were adjusted for naturally occurring sucrose (which equaled sucrose minus added sucrose). In these models, we observed the effect of fructose or sucrose from added sugar sources only, while keeping the intake of naturally occurring fructose or sucrose constant. Models with added sugars were adjusted for fruit as a major source of naturally occurring sugars. To investigate the effect of sugars by source, main models were repeated with individual sugars from solid foods versus sugars from beverages. Overall, we found no significant interactions between our exposures of interest and smoking or alcohol consumption. However, to evaluate the residual confounding effect from these 2 lifestyle factors, known to be strongly associated with mortality, we conducted stratified analyses by smoking status (never and ever smokers) and by alcohol consumption [low (0 to <5 g alcohol/d), moderate (5 to <15 g alcohol/d), or high (≥15 g alcohol/d) in women; low (0 to <5 g alcohol/d), moderate (5 to <30 g alcohol/d), or high (≥30 g alcohol/d) in men]. In further sensitivity analyses, we investigated whether BMI or total energy intake mediated the effect of sugars. The effects of hormone-replacement therapy use in females and total fiber intake were also investigated as potential confounders. To investigate the effect of energy underreporting on the observed associations, all analyses were repeated after energy underreporters were excluded by using the Goldberg cutoffs (39, 40). The participants’ basal metabolic rate was calculated by the Mufflin equation (41), whereas their physical activity level was assumed to be 1.55 (39, 42). All analyses were conducted by using SAS version 9.1 (SAS Institute). The P values for the statistical tests were 2 tailed and were considered statistically significant at a level of <0.05.

RESULTS

During up to 13 y of follow-up, 15,062 deaths in women and 28,617 deaths in men were identified. Most of the deaths were a result of cancer (42% in women and 39.9% in men), whereas 22.6% of women and 26.2% of men died of CVD, 29.6% and 28.5% from other causes, and 5.8% and 5.4% from unknown causes, respectively. Median intakes of added and total sugars were 20.6 g/1000 kcal (IQR: 17.2) and 61.3 g/1000 kcal (IQR: 26.9) in women and 21.0 g/1000 kcal (IQR: 18.9) and 55.2 g/1000 kcal (IQR: 26.2) in men, respectively. Major food sources of added sugars in this population were soft drinks (including sodas, sport drinks, and fruit drinks) (30.5%), followed by sugars, preserves, and candies (20.4%) and cookies, biscuits, cakes, and pies (15.4%). Fruit juices (16.8%) and fruit (16.1%) were leading contributors of total sugars, followed by soft drinks (13.8%) and dairy products (excluding frozen yogurts and ice cream) (13.7%), whereas the leading contributors to fructose were fruit (21.3%), fruit juices (20.9%), soft drinks (16.9%), and sugars, preserves, and candies (10.4%).

In Table 1, we report demographic, behavioral, and health characteristics by sex and quintiles of intake of added and total sugars. Those in the highest compared with the lowest quintile of intake of added and total sugars were less likely to be white, be college educated, have a history of hypertension, and be hormone replacement therapy users (among women). They also consumed less alcohol, vegetables, and red meat and had higher total energy intakes (among women). For some baseline characteristics, participants had opposite trends across quintiles of added compared with total sugars intake. For example, in contrast with the trend for participants with high compared with low total sugars intake, those in the highest compared with the lowest intake of added sugars tended to be younger, to be married (among men), to have a higher BMI (among women), to have ever smoked, to be less physically active, to have a higher total energy intake (among men), and to consume less fruit and dietary fiber but more saturated fat.

TABLE 1.

Baseline characteristics of women (n = 147,380) and men (n = 206,371) in the NIH-AARP Diet and Health study by quintile of intake of added and total sugars¹

Characteristics	Added sugars (g/1000 kcal)				Total sugars (g/1000 kcal)
	Women		Men		Women		Men
	Q1	Q5	Q1	Q5	Q1	Q5	Q1	Q5
≤12.6	≥34.4	≤12.2	≥35.7	≤45.9	≥79.7	≤40.3	≥73.2
Added sugars (g/1000 kcal)	9.7 ± 0.012	51.7 ± 0.11	8.8 ± 0.01	52.9 ± 0.09	14.7 ± 0.04	39.5 ± 0.16	13.0 ± 0.03	42.4 ± 0.13
Total sugars (g/1000 kcal)	51.7 ± 0.12	81.9 ± 0.13	44.2 ± 0.09	78.1 ± 0.10	37.1 ± 0.04	96.1 ± 0.10	32.1 ± 0.03	89.5 ± 0.08
Age (y)	61.8 ± 0.03	60.9 ± 0.03	61.8 ± 0.03	60.9 ± 0.03	61.1 ± 0.03	61.6 ± 0.03	61.2 ± 0.03	61.5 ± 0.03
BMI (kg/m2)	25.8 ± 0.03	26.3 ± 0.03	27.2 ± 0.02	26.8 ± 0.02	26.5 ± 0.03	25.8 ± 0.03	27.6 ± 0.02	26.7 ± 0.02
First-degree relative with cancer (%)	49.6	50.4	46.5	46.6	51.3	49.2	47.6	45.8
Married (%)	46.2	41.2	82.7	85.0	48.4	38.5	84.6	82.5
HRT current users (%)	46.2	39.7	—	—	46.2	40.4	—	—
Race
White (%)	89.9	86.6	92.7	91.2	92.3	83.4	93.9	89.3
African American (%)	3.6	9.1	1.7	4.5	2.7	10.5	1.4	5.2
Physical activity, ≥5 times/wk (%)	19.8	13.3	23.6	18.2	13.5	19.7	18.2	23.8
College graduate or postgraduate (%)	35.2	24.4	52.0	35.1	30.8	30.0	46.6	42.7
Current smoker or quit <1 y ago (%)	17.5	22.9	11.0	18.9	23.0	14.8	16.7	11.5
History of hypertension (%)	21.0	19.6	23.8	18.7	21.0	19.9	22.8	19.9
History of cholesterolemia (%)	33.2	31.0	27.1	28.7	33.2	30.6	27.3	28.1
Daily use of aspirin in the past 12 mo (%)	9.1	9.0	14.3	11.2	8.9	9.4	13.1	12.8
Dietary intake
Energy (kcal/d)	1453 ± 3.59	1694 ± 4.20	1973 ± 4.52	2173 ± 4.48	1555 ± 3.84	1602 ± 4.08	2184 ± 4.78	2011 ± 4.26
Total fructose (g/1000 kcal)	22.0 ± 0.06	36.0 ± 0.07	18.4 ± 0.05	34.8 ± 0.05	14.9 ± 0.02	42.4 ± 0.06	12.8 ± 0.02	39.7 ± 0.05
Added fructose (g/1000 kcal)	11.4 ± 0.05	29.2 ± 0.07	10.5 ± 0.03	29.6 ± 0.05	9.0 ± 0.02	30.5 ± 0.08	8.3 ± 0.01	30.8 ± 0.06
Sucrose (g/1000 kcal)	16.3 ± 0.04	36.2 ± 0.06	14.1 ± 0.03	34.1 ± 0.05	14.6 ± 0.03	34.8 ± 0.06	12.5 ± 0.02	33.2 ± 0.05
Added sucrose (g/1000 kcal)	9.5 ± 0.03	31.7 ± 0.06	9.0 ± 0.02	30.5 ± 0.05	10.5 ± 0.03	27.4 ± 0.07	9.4 ± 0.02	27.5 ± 0.05
Dietary fiber (g/1000 kcal)	12.7 ± 0.03	9.2 ± 0.02	11.1 ± 0.02	8.3 ± 0.01	10.2 ± 0.02	11.9 ± 0.03	9.1 ± 0.01	10.5 ± 0.02
Total fat (g/1000 kcal)	33.0 ± 0.06	32.8 ± 0.05	32.0 ± 0.05	33.8 ± 0.04	39.2 ± 0.05	26.3 ± 0.04	37.0 ± 0.05	28.3 ± 0.03
Saturated fat (g/1000 kcal)	9.9 ± 0.02	10.6 ± 0.02	9.8 ± 0.02	11.0 ± 0.02	11.8 ± 0.02	8.4 ± 0.02	11.4 ± 0.02	9.1 ± 0.01
Polyunsaturated fat (g/1000 kcal)	8.2 ± 0.02	7.5 ± 0.01	7.5 ± 0.01	7.2 ± 0.01	9.5 ± 0.02	6.1 ± 0.01	8.4 ± 0.01	6.2 ± 0.01
Vegetables (cup equivalents/1000 kcal)	1.5 ± 0.005	0.9 0.003)	1.2 ± 0.003	0.8 ± 0.002	1.3 ± 0.004	1.2 ± 0.004	1.0 ± 0.002	0.9 ± 0.003
Fruit (cup equivalents/1000 kcal)	1.5 ± 0.006	1.0 ± 0.004	1.2 ± 0.005	0.8 ± 0.003	0.6 ± 0.002	2.1 ± 0.007	0.5 ± 0.002	1.7 ± 0.005
Red meat (g/1000 kcal)	30.7 ± 0.13	27.7 ± 0.10	38.6 ± 0.12	37.4 ± 0.10	40.3 ± 0.13	19.1 ± 0.08	49.2 ± 0.13	27.1 ± 0.08
Alcohol (g)	13.2 ± 0.17	3.2 ± 0.05	39.7 ± 0.34	8.0 ± 0.09	14.1 ± 0.17	2.6 ± 0.04	44.9 ± 0.36	6.0 ± 0.07

¹P-trend < 0.01 for all variables except first-degree relative with cancer (men: P = 0.31; women: P = 0.11) and for daily use of aspirin in women (P = 0.80) across intake of added sugars and for daily use of aspirin in men (P = 0.54) across intake of total sugars, as estimated by using Cochran-Armitage tests for categorical variables, and for total fat intake in women (P = 0.14) across intake of added sugars, as estimated by the t test for slope in generalized linear models for continuous variables. HRT, hormone-replacement therapy; Q, quintile.

²Mean ± SEM (all such values).

In women, total sugars, fructose, and sucrose intakes were inversely associated, whereas added sugar intake was positively associated with all-cause mortality in age- and energy-adjusted models (see Supplemental Table 1 under “Supplemental data” in the online issue). After control for potential confounders in multivariable models (Table 2), total sugars (HR_{Q5 vs Q1}: 1.13; 95% CI: 1.06, 1.20; P-trend < 0.0001), total fructose (HR_{Q5 vs Q1}: 1.10; 95% CI: 1.04, 1.17; P-trend < 0.0001), and added fructose (HR_{Q5 vs Q1}: 1.07; 95% CI: 1.01, 1.13; P-trend < 0.005) intakes became positively associated with all-cause mortality, whereas we found no association between intake of added sugars and all-cause mortality. Smoking was the strongest confounder in these analyses (data not shown). In men, the trend of associations in age- and energy-adjusted models was similar to the one found in women (see Supplemental Table 1 under “Supplemental data” in the online issue), whereas in the fully adjusted models we observed a modest increase in the risk of all-cause mortality with a high fructose intake (HR_{Q5 vs Q1}: 1.06; 95% CI: 1.01, 1.10; P-trend < 0.002) and an inverse association between other-cause mortality and intake of added sugars (P-trend = 0.04), sucrose (P-trend = 0.03), and added sucrose (P-trend = 0.006) (Table 2). In both women and men, there was a suggestion for an increased risk of CVD mortality with high total sugars and fructose intake (P-trend = 0.08–0.09). None of the intakes of individual sugars or added sugars were associated with cancer mortality, in either women or men. The investigation of consumption of sugars by source showed that the positive association with mortality risk was confined to only sugars from beverages (Table 3). In women, all sugars from beverages were positively associated with risk of all-cause, CVD, and other-cause mortality, whereas in men, a high intake of fructose from beverages was associated with a significantly increased risk of all-cause and CVD mortality risk. Sugars from solid foods were mostly inversely associated with mortality risk, in both women and men. Added fructose and added sucrose generated risk estimates almost identical to those for fructose and sucrose, respectively (data not shown). Mutual adjustment of the models for sugars from solid foods and beverages only slightly attenuated both the positive and inverse associations (data not shown).

TABLE 2.

Multivariable HRs (95% CIs) for all-cause and cause-specific mortality by quintile of intake of sugars in 147,381 women and 206,370 men in the NIH-AARP Diet and Health Study¹

	Multivariable HRs (95% CIs) by quintile2 of intake
	Q2	Q3	Q4	Q5	P-trend
Women
All-cause (no. of cases =15,062)3
Total sugars	0.99 (0.94, 1.04)	1.05 (1.00, 1.11)	1.05 (0.99, 1.11)	1.13 (1.06, 1.20)	<0.0001
Added sugars	0.96 (0.91, 1.01)	0.93 (0.89, 0.98)	0.97 (0.92, 1.02)	1.00 (0.95, 1.06)	0.30
Fructose	0.96 (0.91, 1.01)	1.00 (0.95, 1.05)	1.03 (0.97, 1.09)	1.10 (1.04, 1.17)	<0.0001
Added fructose	0.98 (0.93, 1.03)	1.02 (0.97, 1.07)	1.01 (0.96, 1.07)	1.07 (1.01, 1.13)	0.005
Sucrose	0.97 (0.92, 1.02)	0.95 (0.90, 1.00)	0.97 (0.92, 1.03)	1.00 (0.94, 1.05)	0.89
Added sucrose	0.99 (0.94, 1.05)	0.99 (0.94, 1.05)	0.98 (0.92, 1.03)	1.00 (0.94, 1.05)	0.86
Cancer (no. of cases = 6318)4
Total sugars	1.00 (0.92, 1.08)	0.99 (0.91, 1.08)	1.02 (0.93, 1.11)	1.04 (0.94, 1.15)	0.40
Added sugars	0.94 (0.86, 1.01)	0.93 (0.86, 1.01)	0.94 (0.87, 1.02)	0.99 (0.91, 1.08)	0.63
Fructose	0.95 (0.88, 1.03)	0.99 (0.91, 1.07)	0.98 (0.90, 1.07)	1.03 (0.94, 1.13)	0.40
Added fructose	1.00 (0.92, 1.08)	0.97 (0.89, 1.05)	0.96 (0.89, 1.05)	0.99 (0.90, 1.08)	0.71
Sucrose	0.98 (0.90, 1.06)	0.95 (0.88, 1.03)	0.98 (0.90, 1.07)	1.00 (0.91, 1.08)	0.90
Added sucrose	0.97 (0.90, 1.05)	0.98 (0.90, 1.06)	0.97 (0.89, 1.06)	0.97 (0.89, 1.06)	0.59
CVD (no. of cases = 3406)5
Total sugars	0.91 (0.82, 1.02)	0.97 (0.86, 1.08)	0.97 (0.86, 1.09)	1.10 (0.96, 1.25)	0.09
Added sugars	0.94 (0.84, 1.04)	0.82 (0.73, 0.92)	0.94 (0.84, 1.05)	0.96 (0.86, 1.08)	0.94
Fructose	0.90 (0.81, 1.00)	0.89 (0.79, 0.99)	0.97 (0.86, 1.08)	1.07 (0.95, 1.21)	0.08
Added fructose	0.90 (0.81, 1.00)	0.93 (0.83, 1.03)	0.97 (0.87, 1.08)	1.05 (0.93, 1.18)	0.12
Sucrose	0.96 (0.87, 1.07)	0.86 (0.77, 0.96)	0.88 (0.79, 0.98)	0.95 (0.85, 1.06)	0.36
Added sucrose	0.99 (0.89, 1.10)	1.00 (0.89, 1.11)	0.88 (0.79, 0.99)	0.97 (0.86, 1.09)	0.33
Other causes (no. of cases = 4461)1
Total sugars	0.95 (0.86, 1.04)	1.05 (0.96, 1.15)	0.98 (0.89, 1.08)	1.04 (0.95, 1.15)	0.29
Added sugars	0.96 (0.87, 1.06)	0.97 (0.88, 1.07)	0.99 (0.90, 1.09)	0.97 (0.88, 1.08)	0.86
Fructose	0.97 (0.89, 1.07)	0.98 (0.89, 1.07)	1.00 (0.91, 1.10)	1.05 (0.96, 1.16)	0.16
Added fructose	1.00 (0.91, 1.10)	1.11 (1.01, 1.22)	1.02 (0.92, 1.12)	1.10 (1.00, 1.22)	0.05
Sucrose	0.95 (0.86, 1.04)	0.97 (0.88, 1.06)	0.95 (0.87, 1.05)	0.91 (0.83, 1.00)	0.09
Added sucrose	1.01 (0.91, 1.11)	0.97 (0.88, 1.06)	1.01 (0.92, 1.11)	0.96 (0.87, 1.06)	0.47
Men
All-cause (no. of cases = 28,617)3
Total sugars	0.93 (0.90, 0.97)	0.94 (0.90, 0.98)	0.97 (0.93, 1.01)	1.03 (0.99, 1.08)	0.03
Added sugars	0.94 (0.90, 0.97)	0.94 (0.90, 0.97)	0.92 (0.89, 0.96)	0.95 (0.91, 0.99)	0.19
Fructose	0.95 (0.92, 0.99)	0.97 (0.94, 1.01)	0.97 (0.93, 1.01)	1.06 (1.01, 1.10)	0.002
Added fructose	0.96 (0.92, 0.99)	0.94 (0.90, 0.97)	0.94 (0.91, 0.98)	1.02 (0.97, 1.06)	0.24
Sucrose	0.95 (0.92, 0.99)	0.95 (0.91, 0.98)	0.94 (0.91, 0.98)	0.98 (0.94, 1.02)	0.72
Added sucrose	0.96 (0.93, 1.00)	0.93 (0.89, 0.97)	0.93 (0.89, 0.96)	0.97 (0.93, 1.01)	0.24
Cancer (no. of cases = 11,412)4
Total sugars	0.96 (0.90, 1.01)	0.94 (0.88, 1.00)	0.96 (0.90, 1.03)	0.97 (0.90, 1.04)	0.51
Added sugars	0.94 (0.89, 1.00)	0.96 (0.90, 1.02)	0.98 (0.92, 1.04)	0.97 (0.91, 1.03)	0.87
Fructose	0.96 (0.90, 1.01)	0.98 (0.93, 1.04)	0.97 (0.91, 1.03)	1.00 (0.93, 1.07)	0.80
Added fructose	0.96 (0.90, 1.01)	0.96 (0.90, 1.02)	0.94 (0.89, 1.00)	0.98 (0.92, 1.05)	0.24
Sucrose	0.95 (0.90, 1.01)	0.95 (0.89, 1.01)	0.94 (0.89, 1.00)	1.01 (0.95, 1.08)	0.45
Added sucrose	0.94 (0.89, 1.00)	0.93 (0.88, 0.99)	0.91 (0.85, 0.96)	1.01 (0.95, 1.08)	0.53
CVD (no. of cases = 7488)5
Total sugars	0.97 (0.90, 1.04)	0.96 (0.89, 1.04)	0.95 (0.88–1.03)	1.08 (0.99, 1.18)	0.08
Added sugars	0.91 (0.85, 0.98)	0.87 (0.82, 0.95)	0.87 (0.81, 0.94)	0.91 (0.84, 0.98)	0.07
Fructose	0.97 (0.90, 1.04)	0.98 (0.91, 1.06)	0.94 (0.87, 1.01)	1.08 (1.00, 1.18)	0.08
Added fructose	0.97 (0.90, 1.04)	0.91 (0.84, 0.98)	0.95 (0.88, 1.02)	1.01 (0.94, 1.10)	0.62
Sucrose	0.96 (0.89, 1.03)	0.93 (0.86, 1.00)	0.91 (0.85, 0.99)	0.93 (0.86, 1.00)	0.06
Added sucrose	1.01 (0.94, 1.08)	0.93 (0.86, 1.00)	0.92 (0.85, 0.99)	0.93 (0.86, 1.01)	0.02
Other causes (no. of cases = 8161)1
Total sugars	0.85 (0.79, 0.91)	0.88 (0.82, 0.94)	0.92 (0.85, 0.98)	0.94 (0.88, 1.01)	0.71
Added sugars	0.92 (0.86, 0.99)	0.93 (0.86, 0.99)	0.88 (0.82, 0.94)	0.91 (0.84, 0.98)	0.04
Fructose	0.90 (0.84, 0.96)	0.89 (0.83, 0.95)	0.92 (0.85, 0.98)	0.95 (0.88, 1.02)	0.50
Added fructose	0.92 (0.86, 0.98)	0.89 (0.83, 0.95)	0.88 (0.82, 0.95)	0.95 (0.88, 1.02)	0.33
Sucrose	0.93 (0.87, 1.00)	0.92 (0.86, 0.99)	0.90 (0.84, 0.97)	0.92 (0.86, 0.99)	0.03
Added sucrose	0.92 (0.86, 0.99)	0.89 (0.83, 0.96)	0.91 (0.85, 0.98)	0.89 (0.83, 0.95)	0.006

¹Cox proportional hazards regression model adjusted for age, BMI (in kg/m²; <18.5, ≥18.5 to <25, ≥25 to <30, ≥30 to <35, ≥35, or missing), marital status (married/living as married or widowed/divorced/separated/never married/unknown), smoking (never smoker, quit ≥10 y ago, quit 5–9 y ago, quit 1–4 y ago, quit <1 y ago ≤20 cigarettes/d, quit <1 y ago >20 cigarettes/d, current ≤20 cigarettes/d, current 20–40 cigarettes/d, current >40 cigarettes/d, or missing), race (white, black, Hispanic/Asian/Pacific Islander/American Indian/Alaskan native, or unknown/missing), education (less than high school, high school graduate, some college, college graduate, or unknown/missing), physical activity (never/rarely, 1–3 times/mo, 1–2 times/wk, 3–4 times/wk, ≥5 times/wk, or unknown/missing), and intake of energy (continuous), vegetables (quintiles), and alcohol (quintiles). CVD, cardiovascular disease; Q, quintile.

²Medians by quintile of intake for women: 38.5, 51.5, 61.3, 72.3, and 91.1 g/1000 kcal total sugars; 10.1, 15.1, 20.6, 28.6, and 45.4 g/1000 kcal added sugars; 14.8, 20.4, 25.0, 30.3, and 40.4 g/1000 kcal total fructose; 7.4, 11.4, 15.2, 20, and 30.0 g/1000 kcal added fructose; 13.6, 18.6, 22.8, 27.9, and 37.3 g/1000 kcal sucrose; and 7.9, 12.2, 16.1, 21.5, and 31.8 g/1000 kcal added sucrose. Medians by quintile of intake for men: 33.5, 45.7, 55.2, 65.9, and 84.7 g/1000 kcal total sugars; 9.2, 14.7, 21, 29.4, and 47.0 g/1000 kcal added sugars; 12.7, 18.1, 22.5, 27.8, and 37.8 g/1000 kcal total fructose; 7.3, 11.3, 15.1, 20.0, and 29.9 g/1000 kcal added fructose; 11.8, 16.8, 21.1, 26.2, and 35.4 g/1000 kcal sucrose; and 7.5, 11.8, 15.9, 21.3, and 31.0 g/1000 kcal added sucrose. First quintile is the reference group.

³Additionally adjusted for family history of cancer (yes or no/unknown), intake of saturated fat (quintiles), intake of polyunsaturated fat (quintiles), intake of red meat (quintiles), history of hypertension (no, yes, or unknown/missing), history of hypercholesterolemia (no, yes, or unknown/missing), and use of aspirin (no, monthly, weekly, daily, or missing).

⁴Additionally adjusted for family history of cancer and intakes of total fat (quintiles) and red meat (quintiles).

⁵Additionally adjusted for intakes of saturated and polyunsaturated fats, history of hypertension, history of hypercholesterolemia, and use of aspirin.

TABLE 3.

Multivariable HRs (95% CIs) for all-cause and cause-specific mortality for quintile 5 compared with quintile 1 of intake of sugars from solid foods and beverages in 147,381 women and 206,370 men in the NIH-AARP Diet and Health Study¹

	Women				Men
	Sugars from solid foods		Sugars from beverages		Sugars from solid foods		Sugars from beverages
	HRQ5 vs Q1	P-trend	HRQ5 vs Q1	P-trend	HRQ5 vs Q1	P-trend	HRQ5 vs Q1	P-trend
All-cause2
Total sugars	0.99 (0.94, 1.06)	0.85	1.05 (1.00, 1.11)	0.0008	1.01 (0.97, 1.06)	0.38	0.99 (0.95, 1.03)	0.38
Added sugars	0.89 (0.83, 0.93)	<0.0001	1.09 (1.03, 1.15)	<0.0001	0.88 (0.84, 0.91)	<0.0001	1.00 (0.96, 1.04)	0.34
Fructose	0.96 (0.90, 1.01)	0.10	1.09 (1.03, 1.15)	<0.0001	0.95 (0.91, 0.99)	0.18	1.05 (1.01, 1.09)	0.007
Sucrose	0.91 (0.86, 0.96)	<0.0001	1.07 (1.02, 1.13)	0.0007	0.92 (0.88, 0.95)	0.0004	1.02 (0.98, 1.06)	0.12
Cancer3
Total sugars	0.94 (0.85, 1.03)	0.37	1.04 (0.96, 1.13)	0.11	0.92 (0.86,0.99)	0.03	1.00 (0.94, 1.07)	0.17
Added sugars	0.91 (0.84, 0.99)	0.03	1.02 (0.94, 1.10)	0.15	0.92 (0.87, 0.98)	0.15	1.01 (0.95, 1.07)	0.43
Fructose	0.97 (0.89, 1.07)	0.53	0.99 (0.92, 1.08)	0.48	0.94 (0.88, 1.01)	0.22	0.99 (0.93, 1.06)	0.90
Sucrose	0.94 (0.86, 1.02)	0.06	1.03 (0.95, 1.11)	0.11	0.92 (0.86, 0.98)	0.05	1.02 (0.97, 1.09)	0.07
CVD4
Total sugars	0.96 (0.85, 1.08)	0.46	1.07 (0.96, 1.20)	0.03	1.01 (0.93, 1.10)	0.89	1.00 (0.93, 1.08)	0.48
Added sugars	0.81 (0.73, 0.91)	<0.0001	1.13 (1.01, 1.26)	0.02	0.78 (0.72, 0.85)	<0.0001	1.01 (0.94, 1.09)	0.30
Fructose	0.87 (0.77, 0.99)	0.005	1.14 (1.02, 1.28)	0.002	0.88 (0.81, 0.95)	0.002	1.13 (1.05, 1.22)	0.001
Sucrose	0.80 (0.71, 0.89)	<0.0001	1.08 (0.97, 1.21)	0.09	0.82 (0.76, 0.89)	<0.0001	1.05 (0.97, 1.13)	0.38
Other causes1
Total sugars	0.92 (0.84, 1.01)	0.08	1.08 (0.98, 1.19)	0.008	0.94 (0.87, 1.02)	0.41	0.96 (0.89, 1.03)	0.39
Added sugars	0.80 (0.73, 0.88)	<0.0001	1.14 (1.03, 1.25)	0.002	0.86 (0.80, 0.92)	0.009	0.95 (0.89, 1.03)	0.31
Fructose	0.84 (0.76, 0.93)	0.0009	1.17 (1.06, 1.29)	0.0007	0.91 (0.84, 0.98)	0.07	0.99 (0.92, 1.06)	0.89
Sucrose	0.84 (0.76, 0.92)	0.0001	1.12 (1.02, 1.24)	0.03	0.92 (0.86, 0.99)	0.08	0.96 (0.90, 1.03)	0.23

¹Cox proportional hazards regression model adjusted for age, BMI (in kg/m²; <18.5, ≥18.5 to <25, ≥25 to <30, ≥30 to <35, ≥35, or missing), marital status (married/living as married or widowed/divorced/separated/never married/unknown), smoking (never smoker, quit ≥10 y ago, quit 5–9 y ago, quit 1–4 y ago, quit <1 y ago ≤20 cigarettes/d, quit <1 y ago >20 cigarettes/d, current ≤20 cigarettes/d, current 20–40 cigarettes/d, current >40 cigarettes/d, or missing), race (white, black, Hispanic/Asian/Pacific Islander/American Indian/Alaskan native, or unknown/missing), education (less than high school, high school graduate, some college, college graduate, or unknown/missing), physical activity (never/rarely, 1–3 times/mo, 1–2 times/wk, 3–4 times/wk, ≥5 times/wk, or unknown/missing), and intake of energy (continuous), vegetables (quintiles), and alcohol (quintiles). Median intakes for quintiles 1 and 5 for women: from solid foods (26.1 and 71.3 g/1000 kcal total sugars, 6.7 and 26.0 g/1000 kcal added sugars, 8.7 and 25.5 g/1000 kcal total fructose, and 8.7 and 26.0 g/1000 kcal sucrose) and from beverages (0.02 and 53.1 g/1000 kcal total sugars, 0.08 and 26.9 g/1000 kcal added sugars, 0.9 and 22.7 g/1000 kcal total fructose, and 0.5 and 19.5 g/1000 kcal sucrose). Median intakes for quintiles 1 and 5 for men: from solid foods (20.3 and 60.9 g/1000 kcal total sugars, 5.8 and 23.4 g/1000 kcal added sugars, 6.7 and 20.7 g/1000 kcal total fructose, and 6.9 and 21.5 g/1000 kcal sucrose) and from beverages (0.06 and 62.2 g/1000 kcal total sugars, 0.12 and 30.9 g/1000 kcal added sugars, 1.5 and 23.5 g/1000 kcal total fructose, and 0.8 and 20.7 g/1000 kcal sucrose). CVD, cardiovascular disease; Q, quintile.

²Additionally adjusted for family history of cancer (yes or no/unknown), intake of saturated fat (quintiles), intake of polyunsaturated fat (quintiles), intake of red meat (quintiles), history of hypertension (no, yes, or unknown/missing), history of hypercholesterolemia (no, yes, or unknown/missing), and use of aspirin (no, monthly, weekly, daily, or missing).

³Additionally adjusted for family history of cancer and intakes of total fat (quintiles) and red meat (quintiles).

⁴Additionally adjusted for intakes of saturated and polyunsaturated fats, and history of hypertension, history of hypercholesterolemia, and use of aspirin.

In never smokers, we found an increased risk of all-cause mortality with a high intake of total sugars in both women (HR_{Q5 vs Q1}: 1.11; 95% CI: 0.99, 1.25; P-trend = 0.05) and men (HR_{Q5 vs Q1}: 1.13; 95% CI: 1.02, 1.25; P-trend = 0.006) and with high fructose (HR_{Q5 vs Q1}: 1.13; 95% CI: 1.03, 1.25; P-trend = 0.002) and added fructose (HR_{Q5 vs Q1}: 1.09; 95% CI: 1.00, 1.20; P-trend = 0.02) intake in men only (Figure 1). In men who never smoked, we also observed a positive association with cancer mortality across all investigated sugars, except total sugars, and an inverse association between added sugars and CVD mortality (P-trend = 0.02) (Table 4). Among ever smokers, we detected a statistically significant positive association in women between all-cause mortality and intake of total sugars (HR_{Q5 vs Q1}: 1.15; 95% CI: 1.07, 1.24; P-trend = 0.0002), fructose (HR_{Q5 vs Q1}: 1.10; 95% CI: 1.02, 1.18; P-trend = 0.0007), and added fructose (HR_{Q5 vs Q1}: 1.08; 95% CI: 1.00, 1.16; P-trend = 0.006) (Figure 1) and between CVD mortality and intakes of total sugars (P-trend = 0.003) and fructose (P-trend = 0.005) (Table 3).

FIGURE 1.

HRs and 95% CIs for all-cause mortality for quintile 5 compared with quintile 1 of sugar intake by smoking status in women (A) and men (B) in the NIH-AARP Diet and Health Study in fully adjusted models. A: Women (never smokers: n = 65,593, n = 4488 deaths; ever smokers: n = 76,308, n = 9868 deaths). B: Men (never smokers: n = 64,443, n = 5698 deaths; ever smokers: n = 133,553, n = 21,551 deaths). The dots indicate HRs, and the horizontal lines indicate 95% CIs. ¹Cox proportional hazards regression model adjusted for age, BMI (in kg/m²; <18.5, ≥18.5 to <25, ≥25 to <30, ≥30 to <35, ≥35, or missing), marital status (married/living as married or widowed/divorced/separated/never married/unknown), family history of cancer (yes or no/unknown), smoking (only in ever smokers: quit ≥10 y ago, quit 5–9 y ago, quit 1–4 y ago, quit <1 y ago ≤20 cigarettes/d, quit <1 y ago >20 cigarettes/d, current ≤20 cigarettes/d, current 20–40 cigarettes/d, current >40 cigarettes /d, or missing), race (white, black, Hispanic/Asian/Pacific Islander/American Indian/Alaskan native, or unknown/missing), education (less than high school, high school graduate, some college, college graduate, or unknown/missing), physical activity (never/rarely, 1–3 times/mo, 1–2 times/wk, 3–4 times/wk, ≥5 times/wk, or unknown/missing), history of hypertension (no, yes, or unknown/missing), hypercholesterolemia (no, yes, or unknown/missing), use of aspirin (no, monthly, weekly, daily, or missing), and intake of energy (continuous), saturated fat (quintiles), polyunsaturated fat (quintiles), red meat (quintiles), vegetables (quintiles), and alcohol (quintiles). ²P-trend ≤ 0.05. ³P-trend < 0.01.

TABLE 4.

HRs (95% CIs) for cause-specific mortality for quintile 5 compared with quintile 1 of intake of sugars by smoking status in women and men in the NIH-AARP Diet and Health Study¹

	Never smokers				Ever smokers
	No. of cases	Q5 vs Q1	HR (95% CI)	P-trend	No. of cases	Q5 vs Q1	HR (95% CI)	P-trend
Women
Cancer2	1789				4262
Total sugars (g/1000 kcal)		<48.4 vs >81.9	1.08 (0.89, 1.31)	0.40		<43.8 vs >77.6	1.03 (0.92, 1.16)	0.55
Added sugars (g/1000 kcal)		<13.4 vs >34.0	1.00 (0.84, 1.18)	0.83		<12.2 vs >34.9	0.99 (0.89, 1.10)	0.70
Fructose (g/1000 kcal)		<19.1 vs >35.2	1.02 (0.85, 1.22)	0.76		<17.0 vs >33.0	1.02 (0.91, 1.13)	0.49
Added fructose (g/1000 kcal)		<10.2 vs >24.2	1.01 (0.85, 1.20)	0.53		<9.1 vs >23.1	1.00 (0.90, 1.11)	0.83
Sucrose (g/1000 kcal)		<17.4 vs >31.8	0.99 (0.84, 1.17)	0.86		<15.7 vs >31.2	0.97 (0.88, 1.08)	0.74
Added sucrose (g/1000 kcal)		<10.7 vs >25.5	1.04 (0.88, 1.23)	0.95		<9.9 vs >25.5	0.94 (0.85, 1.05)	0.65
CVD3	1082				2139
Total sugars			0.93 (0.74, 1.17)	0.52			1.25 (1.06, 1.47)	0.003
Added sugars			1.04 (0.85, 1.28)	0.99			0.92 (0.79, 1.07)	0.89
Fructose			0.92 (0.74, 1.14)	0.84			1.18 (1.01, 1.37)	0.005
Added fructose			1.08 (0.88, 1.34)	0.44			1.05 (0.91, 1.22)	0.14
Sucrose			0.84 (0.69, 1.03)	0.08			1.03 (0.89, 1.18)	0.59
Added sucrose			0.90 (0.73, 1.11)	0.07			1.00 (0.86, 1.16)	0.97
Other causes	1275				2976
Total sugars			1.18 (0.98, 1.42)	0.02			0.98 (0.87, 1.10)	0.70
Added sugars			1.08 (0.89, 1.30)	0.12			0.98 (0.86, 1.11)	0.53
Fructose			1.18 (0.98, 1.42)	0.02			0.97 (0.86, 1.09)	0.84
Added fructose			1.17 (0.97, 1.41)	0.09			1.08 (0.96, 1.22)	0.16
Sucrose			1.12 (0.94, 1.35)	0.28			0.92 (0.82, 1.03)	0.09
Added sucrose			1.01 (0.83, 1.21)	0.56			0.99 (0.87, 1.11)	0.53
Men
Cancer2	2029				8886
Total sugars (g/1000 kcal)		<43.8 vs >76.4	1.12 (0.94, 1.34)	0.16		<38.8 vs >71.3	0.95 (0.88, 1.03)	0.36
Added sugars (g/1000 kcal)		<12.6 vs >35.3	1.14 (0.97, 1.33)	0.04		<11.8 vs >36.1	0.94 (0.87, 1.01)	0.29
Fructose (g/1000 kcal)		<17.2 vs >33.1	1.16 (0.98, 1.38)	0.02		<15.0 vs >30.7	0.98 (0.91, 1.06)	0.54
Added fructose (g/1000 kcal)		<10.2 vs >24.6	1.15 (0.98, 1.35)	0.01		<9.0 vs >23.1	0.94 (0.88, 1.02)	0.13
Sucrose (g/1000 kcal)		<15.8 vs >29.9	1.16 (0.99, 1.35)	0.02		<14.0 vs >29.6	0.99 (0.92, 1.06)	0.78
Added sucrose (g/1000 kcal)		<10.5 vs >24.9	1.22 (1.04, 1.42)	0.005		<9.5 vs >25.3	0.98 (0.91, 1.05)	0.65
CVD3	1655				5440
Total sugars			1.10 (0.92, 1.33)	0.21			1.09 (0.99, 1.21)	0.15
Added sugars			0.81 (0.68, 0.95)	0.02			0.95 (0.86, 1.04)	0.61
Fructose			1.06 (0.89, 1.27)	0.30			1.07 (0.97, 1.17)	0.27
Added fructose			1.04 (0.88, 1.23)	0.70			1.00 (0.91, 1.10)	0.76
Sucrose			0.88 (0.74, 1.04)	0.09			0.95 (0.87, 1.04)	0.21
Added sucrose			0.90 (0.76, 1.07)	0.11			0.94 (0.85, 1.03)	0.09
Other causes	1624				6156
Total sugars			1.03 (0.88, 1.21)	0.27			0.93 (0.85, 1.01)	0.47
Added sugars			0.86 (0.72, 1.02)	0.11			0.93 (0.85, 1.01)	0.11
Fructose			1.07 (0.91, 1.26)	0.32			0.93 (0.86, 1.01)	0.32
Added fructose			0.97 (0.82, 1.14)	0.66			0.92 (0.85, 1.00)	0.21
Sucrose			0.89 (0.76, 1.06)	0.18			0.90 (0.83, 0.98)	0.04
Added sucrose			0.83 (0.70, 0.98)	0.06			0.91 (0.84, 0.99)	0.05

¹Cox proportional hazards regression model adjusted for age, BMI (in kg/m²; <18.5, ≥18.5 to <25, ≥25 to <30, ≥30 to <35, ≥35, or missing), marital status (married/living as married or widowed/divorced/separated/never married/unknown), smoking (never smoker, quit ≥10 y ago, quit 5–9 y ago, quit 1–4 y ago, quit <1 y ago ≤20 cigarettes/d, quit <1 y ago >20 cigarettes/d, current ≤20 cigarettes/d, current 20–40 cigarettes/d, current >40 cigarettes/d, or missing), race (white, black, Hispanic/Asian/Pacific Islander/American Indian/Alaskan native, or unknown/missing), education (less than high school, high school graduate, some college, college graduate, or unknown/missing), physical activity (never/rarely, 1–3 times/mo, 1–2 times/wk, 3–4 times/wk, ≥5 times/wk, or unknown/missing), and intake of energy (continuous), vegetables (quintiles), and alcohol (quintiles). CVD, cardiovascular disease; Q, quintile.

²Additionally adjusted for family history of cancer and intakes of total fat (quintiles) and red meat (quintiles).

³Additionally adjusted for intakes of saturated (quintiles) and polyunsaturated (quintiles) fats, history of hypertension (no, yes, or unknown/missing), history of hypercholesterolemia (no, yes, or unknown/missing) and use of aspirin (no, monthly, weekly, daily, or missing).

When we conducted a stratified analysis by alcohol intake (see Supplemental Table 2 under “Supplemental data” in the online issue), we observed a positive association between intakes of total sugars, fructose, and added fructose and all-cause and CVD mortality in both women and men with low alcohol consumption. Contrary to the finding in low consumers, intake of sugars was inversely associated with the risk of all-cause, cancer, and other cause mortality in men with a high alcohol consumption. The intake of sugars was higher in low than in high alcohol consumers in both men and women (see footnote in Supplemental Table 2 under “Supplemental data” in the online issue).

In sensitivity analyses (data not shown), exclusion of energy intake from the multivariable models did not appreciably change the findings and only slightly strengthened the existing associations. Furthermore, exclusion of energy underreporters from the analysis and of BMI from the multivariable models or including dietary fiber as a confounder did not appreciably change any of our findings. We also found no effect of hormone-replacement therapy use on any of the risk estimates in women.

DISCUSSION

In this large US prospective study, intake of total fructose, but not of added sugars, was associated with a modest increase in the risk of all-cause mortality in both men and women. All investigated sugars from beverages, including added sugars, were positively associated with the risk of all-cause, CVD, and other-cause mortality in women, whereas only fructose from beverages was positively associated with the risk of all-cause and CVD mortality risk in men. Previous studies that examined sugar-sweetened beverages (SSBs) and glycemic index (GI) or glycemic load (GL) as proxies for intake of sugars, in relation to mortality, generally found no association. A prospective study that examined the effect of SSBs on mortality in adults aged ≥59 y (43) found no increased risk of all-cause mortality in participants consuming at least one can of SSB per week compared with never consumers. Of 4 prospective studies on GI and GL and mortality (44–47), 3 found no association of GI and GL with total or CVD mortality (44, 45, 47), whereas one cohort reported an almost 2-fold increase in risk from stroke death in participants in the highest versus the lowest tertile of GI but no association between GL and coronary heart disease–related death (46). In contrast with these studies, our study—the first study to investigate total, added sugars, sucrose, and fructose in association with mortality—found that fructose, which is a low-GI carbohydrate, was related to a significantly increased risk of all-cause mortality. Moreover, fructose from beverages was found to be positively related with all-cause and CVD mortality in both women and men. The positive association between fructose and mortality risk became stronger in men who never smoked. Furthermore, in never-smoking men, all investigated sugars, except for total sugars, were positively associated with cancer mortality. In earlier analyses of this cohort, total sugars, fructose (27), and GL (48) were found to be weakly inversely associated with total cancer incidence in men. Yet, in the current analysis, we had a 5-y longer follow-up and we investigated cancer mortality rather than incidence. There was also a suggestion of an increased risk of CVD mortality with high intakes of total sugars and fructose, which became statistically significant and stronger in low alcohol consumers. Participants with a low alcohol intake tended to consume more sugars, which suggested that there may be a threshold effect of sugars in relation to mortality. This threshold effect was observed to some extent in the main analyses, where the HR estimate increased steeply in the fifth quintile. Although we carefully controlled for alcohol consumption, we acknowledge that some residual confounding may still exist, given that nondrinkers and low drinkers are at greater risk of all-cause and CVD mortality than are moderate drinkers, with a characteristic J-shaped association (49).

Few mechanisms explain the link between intake of sugars and mortality. With regard to cancer, postprandial hyperglycemia induced by diets high in sugars triggers insulin and insulin-like growth factor I synthesis, which may enhance tumor development through promoting cell proliferation and inhibiting apoptosis, which stimulates synthesis of sex steroids (50) or vascular endothelial growth factor (8). Of all the investigated sugars, fructose and added fructose were more often associated with an increased risk of all-cause, cancer, and CVD mortality in our analyses. As a potent reducing sugar (51), fructose can produce AGEs, which may be involved in the development and progression of cancer and CVD (9). Furthermore, AGEs can trigger oxidative damage and inflammation, cause changes in the arterial wall structure, and facilitate LDL deposition (52). Low-density lipoproteins may get glycated and, as such, may become more susceptible to oxidation and promote atheroma formation (52). A high fructose intake, through unregulated fructose phosphorylation and accumulation of AMP—a precursor of uric acid (53)—promotes uric acid formation, which can then lead to insulin resistance and hyperinsulinemia (54). Fructose involvement in de novo lipogenesis in the liver and associated hypertriglyceridemia may further explain the association between sugars and CVD (55). Uric acid, induced by high fructose consumption (53), has been implicated in the disruption of endothelial function, which leads to atherosclerosis (56).

High fructose consumption increased the risk of pancreatic cancer in some (57–59), but not all, other cohort studies (60, 61). One (62) of 3 (62, 63) cohorts observed an increased risk of colorectal cancer with high fructose consumption. Prospective evidence is more conclusive with regard to type 2 diabetes, where 2 (64, 65) of 3 (64–66) studies reported an increased risk with high fructose intake but more equivocal with regard to fructose and risk of hypertension (22, 67). Few prospective studies have reported on sugars as nutrients in relation to incidence of CVD (68–70), but none found a statistically significant association with fructose consumption. Prospective epidemiologic evidence on SSBs in relation to type 2 diabetes (19, 71, 72) and coronary heart disease incidence (28, 73) has shown positive associations, whereas findings in relation to cancer, pancreatic cancer in particular, have been largely inconsistent (74).

We observed contrasting effect of sugars in solid foods compared with beverages on mortality risk. Whereas sugars from beverages were positively associated with mortality risk, sugars from solid foods were found to be protective. Sugars from beverages may have a different metabolic effect than sugars from solid foods. Being readily available for absorption and being rapidly metabolized (31), they induce sharp increases in glucose and insulin concentrations (75). Added fructose was more often associated with mortality risk than was added sugars. Besides sugars added at the table or used as ingredients in processed or prepared foods and drinks (definition we used to create added sugars variable) (36), added fructose additionally included total fructose (fructose + fructose in sucrose) from fruit juice and half of total fructose from apple sauce and dried fruit. The contribution of sugars from fruit juice in our population was rather high, where almost half of sugars from fruit came from fruit juice. Because the intake of added sugars among our participants was relatively low, our study may not have had a wide enough range to study its effect on mortality. The median percentage of energy intake from added sugars in our population was 8.3% in both women and men, which conformed with the Dietary Guidelines for Americans, 2010 (76), which recommends that total intake of discretionary calories from added sugars and solid fats be limited to 5–15% of energy intake per day. The intake of added sugars in our population was lower than the energy intakes of 13.3% and 11.8% measured in a representative sample of US men and women aged ≥55 y during 1999–2000 and 2007–2008, respectively (77). Yet, in women, but not in men, added sugars from beverages were associated with an increased mortality risk, which may have been because of hormonal and biological differences and because of different levels of dietary misreporting between the sexes. Furthermore, we observed an inverse association between added sugars, sucrose, and added sucrose and other-cause mortality in men. Because other-cause mortality included deaths from various diseases and conditions, further studies on death from specific diseases are needed. We are nonetheless unsure what might have caused the inverse association between sugars from solid foods and mortality risk.

A major strength of our study was the large sample size and its prospective design, which limits the possibility of recall bias and reverse causality. We had a large number of deaths, which allowed for a stratified analysis by smoking and alcohol. Although we controlled for many potential confounders, creating disease-specific models, we cannot rule out the possibility of residual confounding, particularly from unmeasured or unknown risk factors. Furthermore, there was a probability that some of the findings may have occurred because of chance, because we investigated the effect of 6 dietary exposures in relation to 4 outcomes. An inherent major limitation of all prospective studies is the use of self-reported dietary intake, which is commonly associated with measurement error. A recent biomarker-based study detected substantial measurement error in food-frequency questionnaire–based estimates of total sugars, which showed that a certain level of attenuation may be expected in a disease model with self-reported sugars measured with error (78). Added sugars may be associated with an even greater error than total sugars because of selective misreporting and a tendency among participants to underreport foods perceived to be unhealthy (79). Therefore, measurement error in our analyses may have attenuated the true RR estimates for total sugars, fructose, and added fructose; in the case of added sugars, it may have attenuated the association toward the null.

In summary, in this large prospective study, we found that fructose, but not added sugars consumption, was related to an increased risk of death from all-causes mortality, and fructose from beverages was positively associated with risk of death from all-causes and CVD mortality in both men and women. In addition, we found no consistent evidence of an association between any of the investigated sugars and risk of death from cancer. More studies are needed to replicate our findings and explore the relation of sugars to other morbidity outcomes.