Abstract
Background: Findings on the relation between fruit consumption and the risk of type 2 diabetes mellitus (T2DM) have been inconsistent.
Objectives: We examined whether the consumption of total, temperate, subtropical, and tropical fruit is associated with T2DM risk and whether differences in the carbohydrate quality of fruit influence T2DM risk in Asians.
Design: We included 45,411 participants in the Singapore Chinese Health Study who were 45–74 y old and had no diabetes, cancer, or cardiovascular disease at recruitment (1993–1998). Fruit intake was assessed with the use of a validated food-frequency questionnaire. Physician-diagnosed incident T2DM cases were reported at follow-up 1 (1999–2004) and follow-up 2 (2006–2010) interviews. Cox proportional hazards regression was used to estimate HRs and 95% CIs of diabetes risk.
Results: In 494,741 person-years of follow-up, 5207 participants developed T2DM. After adjustment for lifestyle and dietary risk factors, high total fruit consumption was not consistently associated with lower T2DM risk [men: HR of 1.33 (95% CI: 1.04, 1.71) for ≥3 servings/d compared with <1 serving/wk (P-trend = 0.17); women: HR of 0.88 (95% CI: 0.71, 1.11) (P-trend = 0.008); P-interaction = 0.003]. The direct association in men was observed for higher–glycemic index (GI) fruit [HR: 1.51 (95% CI: 1.22, 1.86) for ≥1 serving/d compared with rarely consumed; P-trend = 0.001] but not for lower or moderate GI fruit. In women, the consumption of temperate fruit, but not of subtropical or tropical fruit, was associated with lower T2DM risk [HR: 0.79 (95% CI: 0.67, 0.92) for ≥1 serving/d compared with rarely; P-trend = 0.006].
Conclusions: The consumption of temperate fruit, such as apples, was associated with a lower risk of T2DM in women, whereas the consumption of higher-GI fruit, such as bananas, was associated with higher risk in men. The impact of fruit consumption on the risk of diabetes may differ by the type of fruit, which may reflect differences in the glycemic impact or phytochemical content.
Keywords: fruit, glycemic index, temperate, subtropical, tropical, type 2 diabetes mellitus
INTRODUCTION
Globally, ∼415 million people suffer from diabetes mellitus, and this number is expected to increase to 642 million by 2040 (1). To curb this trajectory of diabetes burden, concerted efforts for prevention are desirable. Lifestyle interventions with overall diet modifications have proven to be effective in reducing type 2 diabetes mellitus (T2DM)10 risk (2, 3). In addition, dietary fiber, magnesium, and phytochemicals (e.g., flavonoids), all of which are present in fruit (4, 5), have been suggested to improve insulin sensitivity and secretion (4, 5) thereby potentially reducing diabetes risk. Such findings have culminated in the recommendation of fruit consumption for the prevention of T2DM (6). However, in studies on the relation between fruit consumption and T2DM, results have been inconsistent (4, 5, 7–13). Previous studies (5, 7–13) have mostly focused on total fruit and broad categories such as citrus and noncitrus fruit. Few studies have systematically examined different types of individual fruit (4, 14), and limited data are available for tropical fruit that are generally higher in sugar contents and the glycemic index (GI) (4, 9). Individual fruits differ greatly in their GI, fiber, and phytochemical (e.g., flavonoids) contents. Dietary GI (15) and intakes of fruit fiber (16) and specific flavonoids (17, 18) have been postulated to affect risk of T2DM. Furthermore, few studies have examined fruit consumption and diabetes risk in Asian populations (5, 9).
Therefore, with the use of data from a large, prospective cohort study, we investigated whether the consumption of total fruit or different temperate, subtropical, and tropical fruit influenced T2DM risk in Asian men and women. Furthermore, we grouped individual fruit by the fruit’s GI value to determine whether differences in the quality of carbohydrates in fruit may influence T2DM risk. In addition, in view of increasing juice sales in Asia (19–21) and the potential dietary transition such sales could induce, we examined the modulation of T2DM risk with the substitution of whole-fruit intake for juice intake.
METHODS
Study design and population
We used data from the SCHS (Singapore Chinese Health Study) cohort, which is a population-based prospective investigation into the long-term influence of genetic and environmental factors on risks of chronic diseases. Details of the SCHS have been described elsewhere (22, 23). Briefly, 63,257 Chinese adults (aged 45–74 y) of Hokkien- and Cantonese-dialect groups were recruited between 1993 and 1998. Participants were citizens or permanent residents who were dwelling in state-built housing estates where 86% of the general population resided during the enrollment period. At recruitment, face-to-face interviews were conducted in participants’ homes by trained interviewers with the use of a structured questionnaire that ascertained information on demographics, height, weight, tobacco use, physical activity, habitual diet, and medical history. The first and second rounds of follow-up telephone interviews were performed during the periods of 1999–2004 and 2006–2010, respectively. All enrolled participants gave informed consent, and the SCHS was approved by the institutional review boards of the National University of Singapore and the University of Pittsburgh.
For the current study, individuals who had self-reported physician-diagnosed diabetes (n = 5696) or other chronic diseases such as cancer, heart disease, or stroke at baseline (n = 4108), individuals (402 men and 467 women) who had implausibly high daily energy intake (>3700 kcal for men; >3000 kcal for women) or low daily energy intake (<700 kcal for men; <600 kcal for women), and subjects without data for both follow-up interviews because of a nonresponse or death (n = 7173) were excluded. As a result, 45,411 individuals were available for the current analysis (Supplemental Figure 1).
Dietary assessment
At baseline, a 165-item semiquantitative food-frequency questionnaire (FFQ) was administered in person by a trained interviewer to ascertain participants’ habitual diets in the past 1 y (22). The FFQ was comprised of 14 items on fruit that provided information on temperate fruit (apples, pears, apricots and peaches, grapes, and persimmon), subtropical fruit (oranges and tangerines), and tropical fruit (banana, papaya, mango, pineapple, watermelon, honeydew melon, and cantaloupe) (24). Subjects were asked how often, on average, they consumed a standard serving size of each fruit. Eight frequency options were available, which ranged from never or hardly ever to ≥2 times/d. Three portion-size options, which corresponded to 0.5, 1, and ≥2 servings were available for each fruit. Reported intake frequencies were multiplied with portion sizes to derive the consumption of each fruit. Intake of fruit that was affected by seasonal availability (tangerines: 7 mo; apricots: 4 mo; peaches: 6 mo; and persimmon: 6 mo) was multiplied with the fraction of the annual availability of the fruit [e.g., intake for tangerines, which were available 7 mo/y, was multiplied by 0.587 (i.e., 7 divided by 12)] and fruit that had very low consumption in our population (i.e., apricots and peaches, persimmon, mango, pineapple, and cantaloupe) was included in the assessment of total fruit and fruit groups but was not studied individually. Percentages of contribution of individual fruit toward the various fruit groups are shown in Supplemental Table 1. Total juice intake was calculated by summing reported intakes of the FFQ items orange juice and other fruit and vegetable juices. Nine intake-frequency options were available for each juice item, which ranged from never or hardly ever to ≥6 times/d. These 2 juice items were summed to derive total juice consumption. Nutrient and total daily energy intakes were computed from the Singapore Food Composition Table that was developed in conjunction with the cohort (22). The FFQ has been validated previously with the use of two 24-h recalls in 810 cohort participants (22). Correlation coefficients between the FFQ and 24-h recalls for the fruit components fiber and vitamin C intakes ranged from 0.65 to 0.72 and from 0.63 to 0.67, respectively, for the 4 dialect sex groups.
Assessment of incident T2DM
During follow-up telephone interviews, diabetes was assessed with the use of the following question: “Have you been told by a doctor that you have diabetes (high blood sugar)?” If the answer was yes, subjects were asked: “Please also tell me the age at which you were first diagnosed?” Participants were classified as possessing incident diabetes if they reported having developed physician-diagnosed diabetes at any time between the baseline interview and the first (1999–2004) or second (2006–2010) follow-up telephone interviews.
Validation of a self-report of physician-diagnosed diabetes in this population was undertaken in a subpopulation (n = 1631) who reported incident diabetes at a follow-up 1 interview as described in detail previously (25). Briefly, the following 2 different methods were used to confirm self-reported diabetes: 1) linkage with a nationwide hospital-based discharge database and 2) the administration of a supplementary questionnaire about symptoms, diagnostic tests, and hypoglycemic therapy. On the basis of these data, the positive predictive value was 99% (25). We also measured glycated hemoglobin (HbA1c) concentrations in 2625 randomly selected participants who had answered no to the question on diabetes diagnosis. According to the latest HbA1c diagnostic criteria for diabetes (≥6.5%) (26), the negative predictive value was 94% (25).
Statistical analysis
The person-years for each subject were computed from the year of recruitment to the year of reported T2DM diagnosis or the year of the second follow-up interview for individuals who did not report a diabetes diagnosis. Fruit and juice consumption were categorized on the basis of thresholds that were used in the FFQ. Because of low numbers, some categories were combined to achieve a sufficient number of individuals and cases in each category.
HRs and 95% CIs of diabetes risk were estimated with the use of Cox proportional hazards regression models with adjustments for demographic, BMI (in kg/m2), lifestyle, and dietary risk factors. We fitted 3 models with different covariates to adjust for potential confounders. The first model was adjusted for age at the baseline interview (years), the year of the baseline interview (1993–1995 or 1996–1998), sex, dialect group (Hokkien or Cantonese), and total daily energy intake (kilocalories per day). In the second model, we further adjusted for moderate and vigorous physical activity (no moderate, vigorous, or strenuous activity; <4 h moderate activity/wk or <2 h vigorous or strenuous activity/wk; and ≥4 h moderate activity/wk or ≥2 h vigorous or strenuous activity/wk), education level (no formal education, primary school education, or secondary school or higher), cigarette smoking (never smoker, ex-smoker, and current smoker of 1–12 or ≥13 cigarettes/d), alcohol intake (0, <5, or ≥5 g/d), and BMI. In the third model, we further adjusted for total vegetable intake (grams per day), unsweetened soy intake (servings per day), saturated fat intake (percentage of kilocalories), dairy intake (grams per day), soft drink consumption (glasses per day), coffee intake (cups per day), combined black and green tea intake (cups per day), and fruit- and vegetable-juice intake (servings per day) and were mutually adjusted for fruit groups and individual fruit when appropriate. When we examined individual, climate-grouped or GI-grouped fruit, all other individual fruit or fruit groups were mutually adjusted in the multivariable models. We selected covariates for the multivariable models on the basis of the literature and exposures that were previously shown to be associated with T2DM risk in the SCHS to ensure minimal residual confounding while taking precautions to avoid overadjustment and multicollinearity. Moderate physical activities encompassed activities such as brisk walking, bowling, and bicycling on level ground, whereas vigorous and strenuous activities involved work-related activities such as moving heavy furniture and loading and unloading trucks and sports such as tennis, jogging, and swimming laps. The proportional hazards assumption was tested with the use of Schoenfeld’s residuals, and there was no evidence that the assumption was violated (P > 0.05 for all tests). Effect estimates per increment of 3 servings/wk and P-trend values were assessed by modeling continuous variables of fruit consumption with the truncation of outliers (beyond ±4 SDs of the mean).
To understand how the substitution of fruit intake for juice intake influenced diabetes risk, we modeled fruit and juice in the same multivariable model; HRs (which were calculated from the difference in regression coefficients between fruit and juice) and 95% CIs (which were calculated from the variance and covariance of the regression coefficients) of substitution effects were computed (27).
To investigate whether the associations between fruit intake and T2DM risk depended on GI values, individual fruit was categorized as lower-, moderate-, or higher-GI fruit (Supplemental Table 1). GI values [for all fruit except persimmon (28), tangerine (28) and honeydew melon (29)], with glucose as the reference, were obtained from the international GI database (http://www.glycemicindex.com/) (30). To classify (lower, moderate, or higher) fruit on the basis of GI values, all fruit was ranked in order of increasing GI values and categorized into 3 groups with the similar number of fruit in each group.
In supplementary analyses, the relevant models were stratified by sex and median BMI (<25.0 and ≥25.0) to assess a possible effect modification. Sensitivity analyses were performed with the exclusion of individuals who had reported diabetes within the first 4 y of follow-up to account for possible bias as a result of preclinical disease at recruitment (i.e., reverse causality).
All statistical analyses were performed with the use of STATA 13.1 software (StataCorp LP). All P values were 2-sided, and statistical significance was defined as P < 0.05.
RESULTS
In 494,741 person-years of follow-up from 1993 to 2010, we observed 5207 cases of incident T2DM, of which 2195 cases were in men (105 cases/10,000 person-years) and 3012 cases were in women (106 cases/10,000 person-years). Men and women with higher fruit intakes were more likely to be younger, from a Cantonese dialect group, more educated, nonsmokers, and physically active and to have a history of hypertension at baseline (Table 1, Supplemental Tables 2, 3). Higher fruit intake was also associated with higher intakes of vegetables, unsweetened soy, dairy, soft drinks, tea, and juice. Men with higher fruit intakes were also more likely to abstain from alcohol and to consume less coffee. Correlation coefficients between individual fruit were not high and ranged from 0.05 for apples and mango to 0.45 for watermelon and honeydew melon (data not shown).
TABLE 1.
Baseline characteristics in the SCHS in the highest compared with the lowest categories of total whole-fruit consumption1
| Total whole-fruit consumption |
||||
| Men (n = 19,409) |
Women (n = 26,002) |
|||
| Characteristic | <1 serving/wk | ≥3 servings/d | <1 serving/wk | ≥3 servings/d |
| Participants, n | 1194 | 1634 | 1545 | 1666 |
| Total fruit intake,2 servings/wk | 0.0 | 25.5 | 0.2 | 25.1 |
| Age, y | 57 (50–63)3 | 53 (48–59) | 59 (52–65) | 52 (48–58) |
| Dialect group, n (%) | ||||
| Cantonese | 381 (31.9) | 870 (53.2) | 543 (35.2) | 945 (56.7) |
| Hokkien | 813 (68.1) | 764 (46.8) | 1002 (64.9) | 721 (43.3) |
| Education, n (%) | ||||
| No formal education | 250 (20.9) | 88 (5.4) | 980 (63.4) | 377 (22.6) |
| Primary school education | 705 (59.1) | 691 (42.3) | 454 (29.4) | 677 (40.6) |
| Secondary school or higher | 239 (20.0) | 855 (52.3) | 111 (7.2) | 612 (36.7) |
| BMI, kg/m2 | 22.7 (20.3–23.3) | 23.1 (21.6–25.2) | 23.3 (21.3–23.8) | 23.2 (21.1–25.0) |
| Current smoker, n (%) | 749 (62.7) | 386 (23.6) | 254 (16.4) | 39 (2.3) |
| Any alcohol consumption, n (%) | 482 (40.4) | 537 (32.9) | 142 (9.2) | 193 (11.6) |
| Physical activity, n (%) | ||||
| No moderate or vigorous activity | 833 (69.8) | 724 (44.3) | 1323 (85.6) | 1058 (63.5) |
| <4 h moderate activity/wk and <2 h vigorous activity/wk | 144 (12.1) | 391 (23.9) | 114 (7.4) | 336 (20.2) |
| ≥4 h moderate activity/wk and ≥2 h vigorous activity/wk | 217 (18.2) | 519 (31.8) | 108 (7.0) | 272 (16.3) |
| Hypertension, n (%) | 183 (15.3) | 386 (23.6) | 270 (17.5) | 314 (18.9) |
| Total energy intake, kcal/d | 1421 (1122–1811) | 2142 (1728–2597) | 1075 (881–1346) | 1744 (1446–2074) |
| Total vegetable intake, g/1000 kcal | 56.9 (38.3–81.1) | 87.2 (63.9–120.4) | 79.5 (54.3–110.9) | 108.0 (80.6–147.5) |
| Unsweetened soy intake, servings/1000 kcal | 0.10 (0.04–0.20) | 0.14 (0.07–0.26) | 0.13 (0.06–0.23) | 0.17 (0.10–0.29) |
| Saturated fat intake,4 % of kcal | 8.6 ± 2.7 | 8.6 ± 2.5 | 8.9 ± 2.8 | 8.9 ± 2.5 |
| Dairy product intake, g/1000 kcal | 8.8 (1.6–28.2) | 15.3 (4.5–42.0) | 11.0 (1.4–31.7) | 21.0 (5.9–104.6) |
| Soft drink consumer, n (%) | 252 (21.1) | 515 (31.5) | 261 (16.9) | 412 (24.7) |
| Coffee drinker, n (%) | 1010 (84.6) | 1325 (81.1) | 1284 (83.1) | 1310 (78.6) |
| Green or black tea drinker, n (%) | 592 (49.6) | 1230 (75.3) | 497 (32.2) | 1074 (64.5) |
| Fruit- and vegetable-juice intake, n (%) | ||||
| Never or rarely | 991 (83.0) | 858 (52.5) | 1347 (87.2) | 905 (54.3) |
| <1 serving/wk | 79 (6.6) | 178 (10.9) | 100 (6.5) | 212 (12.7) |
| 1 serving/wk | 64 (5.4) | 252 (15.4) | 54 (3.5) | 205 (12.3) |
| ≥2 servings/wk | 60 (5.0) | 346 (21.2) | 44 (2.9) | 344 (20.7) |
SCHS, Singapore Chinese Health Study.
All values are medians.
Median; IQR in parentheses (all such values).
All values are means ± SDs.
Total whole-fruit consumption was significantly associated with higher T2DM risk (HR: 1.33; 95% CI: 1.04, 1.71) in men who consumed ≥3 servings fruit/d compared with men who consumed <1 serving fruit/wk after adjustment for demographic, BMI, lifestyle, and dietary factors (Table 2). However, the association between total fruit consumption and diabetes risk in men did not follow a clear dose-response relation (P-trend = 0.17). In women, a significant inverse trend was observed [≥3 servings/d compared with <1 serving/wk: HR of 0.88 (95% CI: 0.71, 1.11; P-trend = 0.008)] (Table 2).
TABLE 2.
Type 2 diabetes risk across categories of increasing total whole-fruit consumption stratified by sex in the SCHS1
| Total whole-fruit consumption |
|||||||||
| <1 serving/wk | 1 serving/wk | 2–3 servings/wk | 4–6 servings/wk | 1 serving/d | 2 servings/d | ≥3 servings/d | Per 3 servings/wk2 | P-trend | |
| All | |||||||||
| Intake, servings/wk | 0.13 | 1.5 | 3.0 | 5.5 | 9.6 | 16.6 | 25.3 | — | — |
| Cases/person-years, n | 288/29,373 | 247/23,218 | 675/57,797 | 1051/98,594 | 1832/178,751 | 726/70,869 | 388/36,139 | — | — |
| Model 1 | 1.00 (reference)4 | 1.09 (0.92, 1.29) | 1.21 (1.05, 1.38) | 1.10 (0.97, 1.25) | 1.06 (0.94, 1.20) | 1.06 (0.92, 1.22) | 1.11 (0.94, 1.29) | 0.99 (0.98, 1.01) | 0.43 |
| Model 2 | 1.00 (reference) | 1.09 (0.92, 1.29) | 1.15 (1.00, 1.32) | 1.11 (0.98, 1.27) | 1.06 (0.93, 1.20) | 1.07 (0.93, 1.24) | 1.07 (0.92, 1.26) | 0.99 (0.98, 1.01) | 0.34 |
| Model 3 | 1.00 (reference) | 1.10 (0.92, 1.30) | 1.15 (1.00, 1.32) | 1.11 (0.98, 1.27) | 1.06 (0.93, 1.21) | 1.08 (0.93, 1.25) | 1.08 (0.91, 1.27) | 0.99 (0.98, 1.01) | 0.37 |
| Men | |||||||||
| Intake, servings/wk | 0.0 | 1.5 | 3.0 | 5.5 | 9.7 | 16.7 | 25.5 | — | — |
| Cases/person-years, n | 107/12,608 | 80/9461 | 244/22,388 | 424/39,741 | 773/75,903 | 350/31,718 | 217/17,429 | — | — |
| Model 1 | 1.00 (reference) | 1.00 (0.75, 1.33) | 1.29 (1.03, 1.62) | 1.26 (1.02, 1.56) | 1.22 (0.99, 1.49) | 1.32 (1.06, 1.64) | 1.49 (1.17, 1.89) | 1.03 (1.01, 1.04) | 0.004 |
| Model 2 | 1.00 (reference) | 1.01 (0.76, 1.35) | 1.25 (0.99, 1.57) | 1.25 (1.01, 1.55) | 1.16 (0.94, 1.43) | 1.24 (0.99, 1.55) | 1.34 (1.05, 1.71) | 1.01 (1.00, 1.03) | 0.14 |
| Model 3 | 1.00 (reference) | 1.01 (0.75, 1.35) | 1.24 (0.99, 1.56) | 1.25 (1.00, 1.54) | 1.16 (0.94, 1.43) | 1.24 (0.99, 1.56) | 1.33 (1.04, 1.71) | 1.01 (0.99, 1.03) | 0.17 |
| Women | |||||||||
| Intake, servings/wk | 0.2 | 1.5 | 3.0 | 5.5 | 9.6 | 16.6 | 25.1 | — | — |
| Cases/person-years, n | 181/16,765 | 167/13,757 | 431/35,409 | 627/58,853 | 1059/102,848 | 376/39,151 | 171/18,710 | — | — |
| Model 1 | 1.00 (reference) | 1.14 (0.92, 1.40) | 1.15 (0.97, 1.37) | 1.01 (0.86, 1.20) | 0.98 (0.83, 1.15) | 0.91 (0.76, 1.09) | 0.86 (0.69, 1.07) | 0.97 (0.95, 0.99) | <0.0015 |
| Model 2 | 1.00 (reference) | 1.13 (0.91, 1.39) | 1.10 (0.92, 1.31) | 1.03 (0.87, 1.22) | 1.00 (0.85, 1.17) | 0.96 (0.80, 1.16) | 0.87 (0.70, 1.09) | 0.97 (0.96, 0.99) | 0.004 |
| Model 3 | 1.00 (reference) | 1.14 (0.92, 1.41) | 1.11 (0.93, 1.32) | 1.04 (0.88, 1.23) | 1.00 (0.85, 1.18) | 0.97 (0.81, 1.17) | 0.88 (0.71, 1.11) | 0.97 (0.96, 0.99) | 0.008 |
Model 1 was adjusted for age at baseline interview (years), sex, dialect group (Hokkien or Cantonese), year of baseline interview (1993–1995 or 1996–1998), and total daily energy intake (kilocalories per day). Model 2 was adjusted as for model 1 and for physical activity (no moderate, vigorous, or strenuous activity; <4 h moderate activity/wk or <2 h vigorous or strenuous activity/wk; and ≥4 h moderate activity/wk or ≥2 h vigorous or strenuous activity/wk), education (no formal education; primary school education; and secondary school, A levels, or university), smoking status (never smoker, ex-smoker, and current smoker of 1–12 or ≥13 cigarettes/d), alcohol intake (0, <5, and ≥5 g/d), and BMI (in kg/m2). Model 3 was adjusted as for model 2 and for total vegetable intake (grams per day), unsweetened soy intake (servings per day), saturated fat intake (percentage of kilocalories), dairy intake (grams per day), soft drink consumption (glasses per day), coffee intake (cups per day), black and green tea intake (cups per day), and fruit- and vegetable-juice intake (servings per day). P-interaction by sex for all models was significant (model 1: <0.001; model 2: 0.002; and model 3: 0.003). Data were analyzed with the use of Cox proportional hazards regression. Sex was not included as a covariate in sex-stratified models. SCHS, Singapore Chinese Health Study.
Estimated on the basis of every increment of 3 servings/wk.
Median (all such values).
HR; 95% CI in parentheses (all such values).
Significant after Bonferroni correction.
We also examined groups of fruit that were classified according to climate (temperate, subtropical, and tropical) and carbohydrate quality (GI). The consumption of temperate fruit was associated with lower risk in women but not in men (Table 3, Supplemental Table 4). In contrast, the consumption of tropical fruit (Table 3, Supplemental Table 4) or higher-GI fruit (Table 3, Supplemental Table 5) was associated with higher T2DM risk in men but not in women. In neither men nor women was intake of subtropical (citrus) fruit substantially associated with diabetes risk (Table 3, Supplemental Table 4).
TABLE 3.
Type 2 diabetes risk across categories of increasing temperate, subtropical, tropical, or glycemic index–grouped whole-fruit consumption by sex in the SCHS1
| Fruit intake |
|||||||||
| Never or rarely | <1 serving/wk | 1 serving/wk | 2–3 servings/wk | 4–6 servings/wk | ≥1 servings/d | Per 3 servings/wk2 | P-trend | P-interaction by sex | |
| Temperate fruit | |||||||||
| All | |||||||||
| Intake, servings/wk | 0.03 | 0.5 | 1.3 | 2.9 | 5.0 | 8.1 | — | — | — |
| Cases/person-years, n | 602/53,761 | 760/71,207 | 935/84,640 | 1497/142,465 | 716/69,344 | 697/73,324 | — | — | — |
| Multivariable model | 1.00 (reference)4 | 0.95 (0.85, 1.06) | 0.98 (0.88, 1.09) | 0.96 (0.87, 1.05) | 0.94 (0.83, 1.05) | 0.86 (0.77, 0.97) | 0.96 (0.93, 0.99) | 0.008 | 0.04 |
| Men | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.3 | 2.9 | 5.0 | 8.1 | — | — | — |
| Cases/person-years, n | 302/28,827 | 331/31,839 | 370/35,444 | 609/57,710 | 289/27,536 | 294/27,892 | — | — | — |
| Multivariable model | 1.00 (reference) | 0.99 (0.85, 1.16) | 1.02 (0.87, 1.19) | 1.03 (0.89, 1.19) | 0.99 (0.83, 1.17) | 0.97 (0.82, 1.16) | 0.98 (0.94, 1.03) | 0.43 | — |
| Women | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.4 | 3.0 | 5.1 | 8.1 | — | — | — |
| Cases/person-years, n | 300/24,934 | 429/39,368 | 565/49,196 | 888/84,755 | 427/41,808 | 403/45,432 | — | — | — |
| Multivariable model | 1.00 (reference) | 0.91 (0.78, 1.05) | 0.95 (0.82, 1.09) | 0.90 (0.79, 1.03) | 0.89 (0.76, 1.04) | 0.79 (0.67, 0.92) | 0.94 (0.91, 0.98) | 0.006 | — |
| Subtropical fruit | |||||||||
| All | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.1 | 2.5 | 5.1 | 7.1 | — | — | — |
| Cases/person-years, n | 1095/102,474 | 956/89,589 | 748/71,784 | 1362/129,097 | 435/42,551 | 611/59,246 | — | — | — |
| Multivariable model | 1.00 (reference) | 0.97 (0.89, 1.06) | 0.99 (0.90, 1.09) | 1.01 (0.93, 1.09) | 0.98 (0.88, 1.10) | 1.01 (0.90, 1.12) | 1.00 (0.97, 1.04) | 0.78 | 0.06 |
| Men | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.1 | 2.5 | 5.1 | 7.1 | — | — | — |
| Cases/person-years, n | 417/42,064 | 388/38,859 | 314/30,477 | 611/55,476 | 196/18,502 | 269/23,870 | — | — | — |
| Multivariable model | 1.00 (reference) | 0.96 (0.84, 1.11) | 1.03 (0.89, 1.20) | 1.07 (0.94, 1.22) | 1.02 (0.85, 1.21) | 1.07 (0.91, 1.26) | 1.02 (0.98, 1.07) | 0.31 | |
| Women | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.1 | 2.5 | 5.1 | 7.1 | — | — | — |
| Cases/person-years, n | 678/60,410 | 568/50,730 | 434/41,307 | 751/73,621 | 239/24,049 | 342/35,376 | — | — | — |
| Multivariable model | 1.00 (reference) | 0.99 (0.88, 1.11) | 0.97 (0.86, 1.09) | 0.97 (0.87, 1.08) | 0.96 (0.83, 1.12) | 0.96 (0.84, 1.11) | 0.99 (0.95, 1.03) | 0.53 | — |
| Tropical fruit | |||||||||
| All | |||||||||
| Intake, servings/wk | 0.0 | 0.6 | 1.4 | 2.8 | 5.0 | 10.0 | — | — | — |
| Cases/person-years, n | 413/40,372 | 658/63,014 | 952/88,157 | 1446/138,802 | 892/86,723 | 846/77,673 | — | — | — |
| Multivariable model | 1.00 (reference) | 1.02 (0.91, 1.16) | 1.05 (0.93, 1.18) | 1.05 (0.94, 1.17) | 1.01 (0.89, 1.14) | 1.08 (0.95, 1.22) | 1.02 (0.99, 1.04) | 0.15 | 0.10 |
| Men | |||||||||
| Intake, servings/wk | 0.0 | 0.6 | 1.4 | 2.8 | 5.0 | 10.1 | — | — | — |
| Cases/person-years, n | 131/15,256 | 212/20,811 | 329/32,641 | 595/56,601 | 440/41,522 | 488/42,417 | — | — | — |
| Multivariable model | 1.00 (reference) | 1.20 (0.97, 1.50) | 1.12 (0.91, 1.38) | 1.19 (0.98, 1.45) | 1.16 (0.95, 1.41) | 1.24 (1.01, 1.53) | 1.03 (1.00, 1.06) | 0.06 | — |
| Women | |||||||||
| Intake, servings/wk | 0.0 | 0.6 | 1.4 | 2.8 | 5.0 | 9.5 | — | — | — |
| Cases/person-years, n | 282/25,116 | 446/42,203 | 623/55,516 | 851/82,201 | 452/45,201 | 358/35,256 | — | — | — |
| Multivariable model | 1.00 (reference) | 0.95 (0.81, 1.10) | 1.02 (0.88, 1.17) | 0.98 (0.86, 1.13) | 0.94 (0.80, 1.10) | 0.99 (0.83, 1.17) | 1.00 (0.97, 1.04) | 0.99 | — |
| Lower–glycemic index fruit | |||||||||
| All | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.3 | 2.8 | 5.4 | 9.5 | — | — | — |
| Cases/person-years, n | 450/42,468 | 489/45,831 | 598/54,717 | 1161/108,143 | 1203/111,935 | 1306/131,647 | — | — | — |
| Multivariable model | 1.00 (reference) | 0.99 (0.87, 1.13) | 1.00 (0.89, 1.13) | 1.02 (0.91, 1.14) | 1.03 (0.92, 1.16) | 0.95 (0.85, 1.06) | 0.99 (0.97, 1.01) | 0.23 | 0.04 |
| Men | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.3 | 2.8 | 5.3 | 9.6 | — | — | — |
| Cases/person-years, n | 220/22,457 | 200/19,792 | 238/23,355 | 466/45,202 | 509/45,810 | 562/52,632 | — | — | — |
| Multivariable model | 1.00 (reference) | 0.99 (0.82, 1.20) | 1.02 (0.85, 1.23) | 1.04 (0.88, 1.23) | 1.11 (0.94, 1.31) | 1.03 (0.87, 1.22) | 1.01 (0.98, 1.04) | 0.51 | — |
| Women | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.3 | 2.8 | 5.4 | 9.5 | — | — | — |
| Cases/person-years, n | 230/20,011 | 289/26,039 | 360/31,362 | 695/62,941 | 694/66,125 | 744/79,015 | — | — | — |
| Multivariable model | 1.00 (reference) | 0.97 (0.81, 1.15) | 0.99 (0.83, 1.16) | 0.99 (0.85, 1.16) | 0.97 (0.83, 1.13) | 0.88 (0.75, 1.03) | 0.97 (0.94, 1.00) | 0.03 | — |
| Moderate–glycemic index fruit | |||||||||
| All | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.3 | 2.6 | 5.1 | 9.0 | — | — | — |
| Cases/person-years, n | 773/70,586 | 1511/137,957 | 1261/121,864 | 967/95,665 | 433/43,287 | 262/25,382 | — | — | — |
| Multivariable model | 1.00 (reference) | 1.00 (0.91, 1.09) | 0.96 (0.87, 1.05) | 0.92 (0.83, 1.02) | 0.89 (0.79, 1.01) | 0.91 (0.78, 1.06) | 0.95 (0.92, 0.99) | 0.02 | 0.76 |
| Men | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.3 | 2.6 | 5.1 | 8.8 | — | — | — |
| Cases/person-years, n | 290/29,770 | 595/53,718 | 520/51,243 | 465/43,275 | 198/19,551 | 127/11,691 | — | — | — |
| Multivariable model | 1.00 (reference) | 1.11 (0.96, 1.28) | 0.98 (0.85, 1.14) | 1.00 (0.85, 1.16) | 0.88 (0.73, 1.06) | 0.90 (0.71, 1.13) | 0.94 (0.88, 1.00) | 0.04 | — |
| Women | |||||||||
| Intake, servings/wk | 0.0 | 0.5 | 1.3 | 2.6 | 5.1 | 9.1 | — | — | — |
| Cases/person-years, n | 483/40,816 | 916/84,239 | 741/70,621 | 502/52,390 | 235/23,736 | 135/13,691 | — | — | — |
| Multivariable model | 1.00 (reference) | 0.94 (0.84, 1.05) | 0.95 (0.84, 1.07) | 0.88 (0.77, 1.00) | 0.91 (0.77, 1.08) | 0.92 (0.74, 1.13) | 0.97 (0.91, 1.02) | 0.22 | — |
| Higher–glycemic index fruit5 | |||||||||
| All | |||||||||
| Intake, servings/wk | 0.0 | 0.6 | 1.3 | 2.6 | 5.0 | 8.8 | — | — | — |
| Cases/person-years, n | 536/54,841 | 1010/94,602 | 1270/119,884 | 1308/126,513 | 609/56,967 | 474/41,934 | — | — | — |
| Multivariable model | 1.00 (reference) | 1.08 (0.97, 1.20) | 1.11 (1.01, 1.23) | 1.09 (0.99, 1.21) | 1.11 (0.98, 1.26) | 1.20 (1.04, 1.37) | 1.05 (1.01, 1.08) | 0.007 | 0.02 |
| Men | |||||||||
| Intake, servings/wk | 0.0 | 0.6 | 1.4 | 2.6 | 5.0 | 9.1 | — | — | — |
| Cases/person-years, n | 156/19,509 | 323/31,794 | 493/47,263 | 596/56,463 | 323/29,729 | 304/24,490 | — | — | — |
| Multivariable model | 1.00 (reference) | 1.22 (1.00, 1.48) | 1.30 (1.08, 1.56) | 1.30 (1.08, 1.56) | 1.28 (1.05, 1.56) | 1.51 (1.22, 1.86) | 1.08 (1.03, 1.13) | 0.0015 | — |
| Women | |||||||||
| Intake, servings/wk | 0.0 | 0.6 | 1.3 | 2.6 | 5.0 | 8.5 | — | — | — |
| Cases/person-years, n | 380/35,332 | 687/62,808 | 777/72,621 | 712/70,050 | 286/27,238 | 170/17,444 | — | — | — |
| Multivariable model | 1.00 (reference) | 1.02 (0.90, 1.16) | 1.04 (0.92, 1.18) | 1.01 (0.88, 1.15) | 1.05 (0.89, 1.23) | 0.99 (0.81, 1.20) | 1.01 (0.96, 1.06) | 0.80 | — |
Temperate fruit included apples, pears, apricots and peaches, grapes, and persimmon. Subtropical fruit included oranges and tangerines. Tropical fruit included bananas, papayas, mangos, pineapple, watermelon, honeydew melon, and cantaloupe. Lower–glycemic index fruit included apricots and peaches, apples, oranges, pears, and persimmon. Moderate–glycemic index fruit included tangerines, mango, papaya, and grapes. Higher–glycemic index fruit included bananas, honeydew melon, pineapple, cantaloupe, and watermelon. Data were analyzed with the use of Cox proportional hazards regression. Sex was not included as a covariate in sex-stratified models. Multivariable models were adjusted for age at baseline interview (years), sex, dialect group (Hokkien or Cantonese), year of baseline interview (1993–1995 or 1996–1998), total daily energy intake (kilocalories per day), physical activity (no moderate, vigorous, or strenuous activity; <4 h moderate activity/wk or <2 h vigorous or strenuous activity/wk; and ≥4 h moderate activity/wk or ≥2 h vigorous or strenuous activity/wk), education (no formal education; primary school education; and secondary school, A levels, or university), smoking status (never smoker, ex-smoker, and current smoker of 1–12 or ≥13 cigarettes/d), alcohol intake (0, <5, and ≥5 g/d), BMI (in kg/m2), total vegetable intake (grams per day), unsweetened soy intake (servings per day), saturated fat intake (percentage of kilocalories), dairy intake (grams per day), soft drink consumption (glasses per day), coffee intake (cups per day), black and green tea intake (cups per day), and fruit- and vegetable-juice intake (servings per day) and were mutually adjusted for climate-grouped or glycemic index–grouped fruit.
Estimated on the basis of every increment of 3 servings/wk.
Medians (all such values).
HR; 95% CI in parentheses (all such values).
Significant after Bonferroni correction.
Similarly, intake of individual fruit had different associations with T2DM risk in men and women. In men, banana intake was significantly associated with higher risk of T2DM (HR: 1.11; 95% CI: 1.04, 1.19) for every 3 servings consumed per week, whereas no association was observed in women (HR: 0.94; 95% CI: 0.87, 1.01) (Supplemental Table 6). In women, apple consumption was significantly associated with lower T2DM risk (HR: 0.90; 95% CI: 0.86, 0.95), whereas no association was observed in men (HR: 0.98; 95% CI: 0.92, 1.04) (Supplemental Table 7). Grape consumption was associated with lower T2DM risk (HR: 0.87; 95% CI: 0.76, 0.99; P-trend = 0.04) in men and women combined (Supplemental Table 7) with similar associations for men and women. Intake of other fruit was not significantly associated with risk of T2DM (Supplemental Tables 6, 7 and 8).
We further examined whether fiber or total carbohydrate intake could explain part of the association between fruit consumption and T2DM risk. Additional adjustment for total carbohydrate or fiber intake did not substantially change any of the observed direct associations in men or of the inverse association between grapes and T2DM risk in men and women (data not shown). Similarly, adjustment for total carbohydrate did not substantially change the observed inverse associations in women. In contrast, total fruit intake [≥3 servings/d compared with <1 serving/wk: HR of 1.19 (95% CI: 0.90, 1.57, P-trend = 0.67)] and temperate fruit intake [≥1 serving/d compared with never or rarely consumed: HR of 0.88 (95% CI: 0.73, 1.05, P-trend = 0.35)] were not associated with T2DM risk in women after adjustment for fiber. However, adjustment for fiber intake only partially explained the association between apple consumption [per 3 servings/wk: HR of 0.94 (95% CI: 0.88, 0.99, P-trend = 0.03)] and T2DM risk in women (data not shown).
Juice consumption was associated with a higher risk of T2DM [HR: 1.21 (95% CI: 1.04, 1.39; P-trend = 0.001)] for ≥4 servings/wk compared with never or rarely consumed for men and women combined, and this association was slightly attenuated after further adjustment for BMI (HR: 1.16; 95% CI: 1.00, 1.34; P-trend = 0.01) (Supplemental Table 9). The replacement of 3 servings juice/wk with 3 servings fruit/wk was associated with lower risk of T2DM (HR: 0.92; 95% CI: 0.86, 0.99) (Figure 1). Similarly, the substitution of juice with lower-GI fruit (HR: 0.92; 95% CI: 0.86, 0.99), grapes (HR: 0.81; 95% CI: 0.70, 0.94), or tangerines (HR: 0.84; 95% CI: 0.75, 0.95) was associated with lower T2DM risk without evidence of an effect modification by sex (Figure 1). The substitution of juice with moderate-GI fruit or papaya in men and with temperate fruit or apples in women was also significantly associated with lower T2DM risk.
FIGURE 1.
Multivariable adjusted HRs and 95% CIs (error bars) of type 2 diabetes risk of substituting 3 servings total fruit, climate-grouped fruit, glycemic index–grouped fruit, or specific individual fruit/wk for the same amount of fruit or vegetable juice in all subjects (n = 45,411), in men (n = 19,409), and in women (n = 26,002). All models were adjusted for age, sex, dialect, year of baseline interview, energy intake, physical activity, education, smoking status, alcohol intake, BMI, total vegetables intake, unsweetened soy intake, saturated fat intake, dairy intake, soft drink intake, coffee intake, and black and green tea intake and were mutually adjusted for the relevant fruit groups or individual fruit intake. Data were analyzed with the use of Cox proportional hazards regression. Sex was not included as a covariate in sex-stratified models. Substitution effects were estimated with the use of a previously described method (27).
DISCUSSION
In this large prospective cohort of Asian men and women, associations between fruit consumption and the risk of T2DM differed by the type of fruit and were partly modified by sex. In men, higher consumption of tropical fruit or fruit with a higher GI, such as bananas, was associated with higher T2DM risk. In women, higher consumption of temperate fruit, such as apples, was significantly associated with lower T2DM risk. Consumption of grapes was associated with lower risk of T2DM, and consumption of juice was associated with higher risk of T2DM, regardless of sex.
Previous studies on total fruit consumption and risk of T2DM have generated inconsistent results. In 2 recent meta-analyses (31, 32) that encompassed 10 prospective studies, total fruit intake was not associated with T2DM risk in men, whereby 1 (31) of the 2 meta-analyses reported a slightly lower risk in women [−8% (95% CI: −1% to −14%) for highest compared with lowest intakes]. Total fruit consumption was not associated with T2DM risk in Asian women or men in the Japan Public Health Center study (5) or in women in the Shanghai Women’s Health Study (9). Consistent with previous studies, our results do not support that high total fruit consumption substantially lowers risk of T2DM.
In a recent meta-analysis that included 17 prospective cohort studies (15), a higher dietary GI was associated with a higher T2DM risk. However, in contrast to our findings, differences in GI values of fruit did not influence T2DM risk in US men and women (4). This discrepancy could have been due to different consumption patterns between populations. For example, fruit may be more commonly consumed with other foods with a high GI and carbohydrate content such as white rice in Asian populations (thereby leading to greater glycemic spikes) as opposed to being consumed in isolation as snacks in Western populations. Possible explanations for detrimental effects of a high dietary GI are as follows: chronic stress on β cells to increase insulin output in response to greater postprandial glycemia; long-term exposure to elevated postprandial blood glucose concentrations, which could be toxic to β cells; and increased postprandial insulin secretion, which could contribute to a long-term hyperinsulinemic state that facilitates insulin resistance (15). In our study, the consumption of higher-glycemic fruit such as bananas was associated with higher T2DM risk in men but not in women. It is currently unclear what biological mechanisms may explain this possible interaction with sex. However, compared with men, higher first-phase insulin secretion (33) and greater insulin sensitivity (34) have been observed in women, which could limit the postprandial glycemic elevation and possibly reduce the impact of higher-GI fruit in women.
Our findings for individual fruit may be explained in part by their phytochemical and specific nutrient contents. Higher intakes of magnesium and fruit fiber have been associated with lower T2DM risk in recent meta-analyses of cohort studies (16, 35). In our study, the association between apple consumption, but not grape consumption, and T2DM in women was partly explained by fiber intake. However, it is also possible that other components of fruit that are closely associated with the fiber fraction are responsible for the putative beneficial effects of apples. Apples are rich in flavan-3-ols and flavonols and also contain substantial amounts of anthocyanidins; grapes are also rich in flavan-3-ols and anthocyanidins (36, 37). Several flavonoids or flavonoid-rich foods, including epigallocatechin-3-gallate (38), myricetin (39), and an anthocyanidin-rich bilberry extract (40), improved insulin sensitivity in animal models. These effects were mediated by changes in insulin and glucose translocation pathways rather than by general antioxidant effects (39, 40).
Findings from the current study and from previous studies suggest that the type of fruit consumed may be more relevant for T2DM risk than is total fruit consumption. In addition, modern types of fruit have been postulated to be higher in total sugars than are their wild counterparts (41). In line with our findings, the consumption of citrus fruit, such as oranges, was not associated with T2DM in several cohort studies (4, 5, 8, 9, 12, 42). We observed an inverse association between temperate fruit, such as apples, and T2DM risk in women. In agreement with these results, apple consumption was inversely associated with T2DM risk in women participating in the Women’s Health Study (43) and Nurses’ Health Study (NHS) I and II (4) although this association was not observed in the Iowa Women’s Health Study (42). The NHS I and II combined apples and pears, and our results suggest that apples, but not pears, may be responsible for this inverse association. In agreement with our findings in men, the consumption of apples and pears was not associated with T2DM risk in the HPFS (Health Professionals Follow-Up Study) in men (4). In contrast with our results in men, banana consumption was inversely associated with T2DM risk in men and women in the HPFS and NHS II, respectively, whereas no association was observed in the NHS I. The combination of grapes and raisins was associated with lower T2DM risk in the HPFS and NHS I and II. In our study, grape consumption showed a significant association with lower risk of T2DM. Our current study provides, for the first time to our knowledge, data on several types of individual fruit, such as pears, grapes, tangerines, papaya, and honeydew melon, that were not separately examined in relation to T2DM previously. The consumption of berries, such as blueberries, was also associated with a lower T2DM risk in several studies (4, 11, 14), but such berries are not commonly consumed in Singapore and were not assessed in the current study. Overall, our findings, in combination with results from previous studies, suggest that the consumption of apples (at least in women) and grapes may be associated with lower risk of T2DM. Further research is needed to establish whether the glycemic effects of carbohydrates or the phytochemical contents of fruit are more relevant for the impact of different types of fruit on glucose metabolism.
We showed that the substitution of most types of fruit, with the exception of higher-GI fruit, for juice was associated with lower T2DM risk in Asian men and women. Fruit and vegetable juices were combined in the assessment of juice intake in our study, because patrons do not make the distinction between the 2 types because fruit and vegetables are often mixed in juice products that are sold in this population. Our results were generally in agreement with a US study (4) that reported the consumption of most types of whole fruit over juice was beneficial in reducing T2DM risk. During the juicing process, fiber and phytochemicals are reduced or removed from whole fruit (25), and naturally occurring sugars in fruit are concentrated in juice, thereby increasing the glycemic load per serving. In an intervention study, the consumption of juice led to greater insulin and glucose responses than were shown with whole-fruit consumption (44). Sales of juices are increasing in Asia (19–21). Thus, an understanding of how the replacement of juices with specific types of whole fruit may affect T2DM risk is of public health relevance in Asia.
To our knowledge, this is the first large prospective study with a comprehensive examination of different types of fruit in relation to T2DM risk in an Asian population. Several limitations of our study also need to be acknowledged. Random measurement error in the estimation of fruit intake could have attenuated the associations toward null. While appropriate to describe juice intake habits in our population, the inclusion of vegetable juice with fruit juice does not allow us to distinguish the health effects between the 2 types of juices. Although we adjusted for a wide range of potential confounders, we cannot exclude the possibility of residual confounding due to unmeasured or imperfectly measured covariates as it could have contributed to our observed associations. Because multiple fruit and fruit groups were examined in our study, the possibility of obtaining a false positive result should be considered. After correcting for multiple testing with the use of the Bonferroni approach (for 17 tests), the direct associations for banana and higher-GI fruit in men and the inverse association for apple intake in women remained significant. Given the large size of our study, blood glucose screening to identify undiagnosed T2DM in all participants was not feasible. However, validation of self-reported incident T2DM through several strategies (hospital discharge records, supplementary questionnaires and HbA1c) indicated that the diabetes classification in our study was reasonably accurate. Temporality bias was limited as the exclusion of cases within 4 y of follow-up did not change the observed associations substantially. Our findings emanate from an ethnic Chinese population and may therefore not be generalizable to other ethnic groups.
In conclusion, our findings suggest that the association between fruit consumption and T2DM risk differs by the type of fruit. The high consumption of tropical or higher-GI fruit, such as bananas, is associated with higher T2DM risk in men, whereas the high consumption of temperate fruit, such as apples, is associated with lower T2DM risk in women. Differences in associations may be partly explained by variations in GI values in the different types of individual fruit but could also reflect the differences in phytochemical contents between types of fruit. Furthermore, choosing lower-GI whole fruit, grapes, or tangerines instead of juice is associated with lower T2DM risk. Our findings support recommendations to promote the consumption of specific types of whole fruit, especially fruit with lower or moderate GI indexes such as apples and grapes, instead of juice as a public health strategy for reducing T2DM risk.
Acknowledgments
We thank Siew-Hong Low of the National University of Singapore for overseeing the fieldwork of the SCHS and Renwei Wang for the development and maintenance of the cohort-study database. We also thank Mimi C Yu for being the founding and long-standing principal investigator of the SCHS.
The authors’ responsibilities were as follows—DJA: conceived the study, performed the statistical analysis and interpretation of the data, drafted the initial manuscript, contributed to discussions about the results, and critically revised and edited the final manuscript; LMB: contributed to discussions about the results and reviewed and edited the final manuscript; W-PK and J-MY: supervised the data collection, contributed to discussions about the results, and reviewed and edited the final manuscript; RMvD: conceived the study, contributed to discussions about the results, critically revised and edited the final manuscript, was the guarantor of the work, had full access to all of the data in the study, and took responsibility for the integrity of the data and the accuracy of the data analysis; and all authors: read and approved the final manuscript. None of the authors reported a conflict of interest related to the study.
Footnotes
Abbreviations used: FFQ, food frequency questionnaire; GI, glycemic index; HbA1c, glycated hemoglobin; HPFS, Health Professionals Follow-Up Study; NHS, Nurses’ Health Study; SCHS, Singapore Chinese Health Study; T2DM, type 2 diabetes mellitus.
REFERENCES
- 1.International Diabetes Federation (IDF). IDF diabetes atlas. 7th ed Brussels (Belgium): International Diabetes Federation; 2015. [Google Scholar]
- 2.Gillies CL, Abrams KR, Lambert PC, Cooper NJ, Sutton AJ, Hsu RT, Khunti K. Pharmacological and lifestyle interventions to prevent or delay type 2 diabetes in people with impaired glucose tolerance: systematic review and meta-analysis. BMJ 2007;334:299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Carter P, Gray LJ, Troughton J, Khunti K, Davies MJ. Fruit and vegetable intake and incidence of type 2 diabetes mellitus: systematic review and meta-analysis. BMJ 2010;341:c4229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Muraki I, Imamura F, Manson JE, Hu FB, Willett WC, van Dam RM, Sun Q. Fruit consumption and risk of type 2 diabetes: results from three prospective longitudinal cohort studies. BMJ 2013;347:f5001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kurotani K, Nanri A, Goto A, Mizoue T, Noda M, Kato M, Inoue M, Tsugane S, Japan Public Health Center-based Prospective Study Group. Vegetable and fruit intake and risk of type 2 diabetes: Japan Public Health Center-based Prospective Study. Br J Nutr 2013;109:709–17. [DOI] [PubMed] [Google Scholar]
- 6.Franz MJ, Bantle JP, Beebe CA, Brunzell JD, Chiasson JL, Garg A, Holzmeister LA, Hoogwerf B, Mayer-Davis E, Mooradian AD, et al. Evidence-based nutrition principles and recommendations for the treatment and prevention of diabetes and related complications. Diabetes Care 2003;26(Suppl 1):S51–61. [DOI] [PubMed] [Google Scholar]
- 7.Cooper AJ, Sharp SJ, Lentjes MA, Luben RN, Khaw KT, Wareham NJ, Forouhi NG. A prospective study of the association between quantity and variety of fruit and vegetable intake and incident type 2 diabetes. Diabetes Care 2012;35:1293–300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Cooper AJ, Forouhi NG, Ye Z, Buijsse B, Arriola L, Balkau B, Barricarte A, Beulens JW, Boeing H, Buchner FL, et al. Fruit and vegetable intake and type 2 diabetes: EPIC-InterAct prospective study and meta-analysis. Eur J Clin Nutr 2012;66:1082–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Villegas R, Shu XO, Gao YT, Yang G, Elasy T, Li H, Zheng W. Vegetable but not fruit consumption reduces the risk of type 2 diabetes in Chinese women. J Nutr 2008;138:574–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bazzano LA, Li TY, Joshipura KJ, Hu FB. Intake of fruit, vegetables, and fruit juices and risk of diabetes in women. Diabetes Care 2008;31:1311–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Montonen J, Jarvinen R, Heliovaara M, Reunanen A, Aromaa A, Knekt P. Food consumption and the incidence of type II diabetes mellitus. Eur J Clin Nutr 2005;59:441–8. [DOI] [PubMed] [Google Scholar]
- 12.Liu S, Serdula M, Janket SJ, Cook NR, Sesso HD, Willett WC, Manson JE, Buring JE. A prospective study of fruit and vegetable intake and the risk of type 2 diabetes in women. Diabetes Care 2004;27:2993–6. [DOI] [PubMed] [Google Scholar]
- 13.Meyer KA, Kushi LH, Jacobs DR Jr, Slavin J, Sellers TA, Folsom AR. Carbohydrates, dietary fiber, and incident type 2 diabetes in older women. Am J Clin Nutr 2000;71:921–30. [DOI] [PubMed] [Google Scholar]
- 14.Knekt P, Kumpulainen J, Jarvinen R, Rissanen H, Heliovaara M, Reunanen A, Hakulinen T, Aromaa A. Flavonoid intake and risk of chronic diseases. Am J Clin Nutr 2002;76:560–8. [DOI] [PubMed] [Google Scholar]
- 15.Bhupathiraju SN, Tobias DK, Malik VS, Pan A, Hruby A, Manson JE, Willett WC, Hu FB. Glycemic index, glycemic load, and risk of type 2 diabetes: results from 3 large US cohorts and an updated meta-analysis. Am J Clin Nutr 2014;100:218–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Yao B, Fang H, Xu W, Yan Y, Xu H, Liu Y, Mo M, Zhang H, Zhao Y. Dietary fiber intake and risk of type 2 diabetes: a dose-response analysis of prospective studies. Eur J Epidemiol 2014;29:79–88. [DOI] [PubMed] [Google Scholar]
- 17.Wedick NM, Pan A, Cassidy A, Rimm EB, Sampson L, Rosner B, Willett W, Hu FB, Sun Q, van Dam RM. Dietary flavonoid intakes and risk of type 2 diabetes in US men and women. Am J Clin Nutr 2012;95:925–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Zamora-Ros R, Forouhi NG, Sharp SJ, Gonzalez CA, Buijsse B, Guevara M, van der Schouw YT, Amiano P, Boeing H, Bredsdorff L, et al. The association between dietary flavonoid and lignan intakes and incident type 2 diabetes in European populations: the EPIC-InterAct study. Diabetes Care 2013;36:3961–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Juice in Japan. Euromonitor International; 2014. [Internet]. [cited 2014 Sep 30]. Available from: http://www.euromonitor.com/juice-in-japan/report.
- 20. Juice in India. Euromonitor International; 2014. [Internet]. [cited 2014 Sep 30]. Available from: http://www.euromonitor.com/juice-in-india/report.
- 21. Juice in China. Euromonitor International; 2014. [Internet]. [cited 2014 Sep 30]. Available from: http://www.euromonitor.com/juice-in-china/report.
- 22.Hankin JH, Stram DO, Arakawa K, Park S, Low SH, Lee HP, Yu MC. Singapore Chinese Health Study: development, validation, and calibration of the quantitative food frequency questionnaire. Nutr Cancer 2001;39:187–95. [DOI] [PubMed] [Google Scholar]
- 23.Yuan JM, Stram DO, Arakawa K, Lee HP, Yu MC. Dietary cryptoxanthin and reduced risk of lung cancer: the Singapore Chinese Health Study. Cancer Epidemiol Biomarkers Prev 2003;12:890–8. [PubMed] [Google Scholar]
- 24.James JB, Ngarmsak T. Processing of fresh-cut tropical fruits and vegetables: a technical guide. Bangkok (Thailand): FAO of the UN; 2010. [Google Scholar]
- 25.Odegaard AO, Koh WP, Arakawa K, Yu MC, Pereira MA. Soft drink and juice consumption and risk of physician-diagnosed incident type 2 diabetes: the Singapore Chinese Health Study. Am J Epidemiol 2010;171:701–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.International Expert Committee. International Expert Committee report on the role of the A1C assay in the diagnosis of diabetes. Diabetes Care 2009;32:1327–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Bernstein AM, Sun Q, Hu FB, Stampfer MJ, Manson JE, Willett WC. Major dietary protein sources and risk of coronary heart disease in women. Circulation 2010;122:876–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ryu J-H, Yim J-E, Suk W-H, Lee H, Ahn H, Kim Y-S, Park C-S, Choue R. Sugar composition and glycemic indices of frequently consumed fruits in Korea. Korean J Nutr 2012;45:192–200. [Google Scholar]
- 29.Bao J, de Jong V, Atkinson F, Petocz P, Brand-Miller JC. Food insulin index: physiologic basis for predicting insulin demand evoked by composite meals. Am J Clin Nutr 2009;90:986–92. [DOI] [PubMed] [Google Scholar]
- 30.Atkinson FS, Foster-Powell K, Brand-Miller JC. International tables of glycemic index and glycemic load values: 2008. Diabetes Care 2008;31:2281–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Li S, Miao S, Huang Y, Liu Z, Tian H, Yin X, Tang W, Steffen LM, Xi B. Fruit intake decreases risk of incident type 2 diabetes: an updated meta-analysis. Endocrine 2015;48:454–60. [DOI] [PubMed] [Google Scholar]
- 32.Li M, Fan Y, Zhang X, Hou W, Tang Z. Fruit and vegetable intake and risk of type 2 diabetes mellitus: meta-analysis of prospective cohort studies. BMJ Open 2014;4:e005497. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Flanagan DE, Holt RI, Owens PC, Cockington RJ, Moore VM, Robinson JS, Godsland IF, Phillips DI. Gender differences in the insulin-like growth factor axis response to a glucose load. Acta Physiol (Oxf) 2006;187:371–8. [DOI] [PubMed] [Google Scholar]
- 34.Mauvais-Jarvis F. Sex differences in metabolic homeostasis, diabetes, and obesity. Biol Sex Differ 2015;6:14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Larsson SC, Wolk A. Magnesium intake and risk of type 2 diabetes: a meta-analysis. J Intern Med 2007;262:208–14. [DOI] [PubMed] [Google Scholar]
- 36.van Dam RM, Naidoo N, Landberg R. Dietary flavonoids and the development of type 2 diabetes and cardiovascular diseases: review of recent findings. Curr Opin Lipidol 2013;24:25–33. [DOI] [PubMed] [Google Scholar]
- 37.Beecher GR. Overview of dietary flavonoids: nomenclature, occurrence and intake. J Nutr 2003;133:3248S–54S. [DOI] [PubMed] [Google Scholar]
- 38.Chen YK, Cheung C, Reuhl KR, Liu AB, Lee MJ, Lu YP, Yang CS. Effects of green tea polyphenol (-)-epigallocatechin-3-gallate on newly developed high-fat/Western-style diet-induced obesity and metabolic syndrome in mice. J Agric Food Chem 2011;59:11862–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Liu IM, Tzeng TF, Liou SS, Lan TW. Myricetin, a naturally occurring flavonol, ameliorates insulin resistance induced by a high-fructose diet in rats. Life Sci 2007;81:1479–88. [DOI] [PubMed] [Google Scholar]
- 40.Takikawa M, Inoue S, Horio F, Tsuda T. Dietary anthocyanin-rich bilberry extract ameliorates hyperglycemia and insulin sensitivity via activation of AMP-activated protein kinase in diabetic mice. J Nutr 2010;140:527–33. [DOI] [PubMed] [Google Scholar]
- 41.Brand-Miller JC, Holt SH. Australian aboriginal plant foods: a consideration of their nutritional composition and health implications. Nutr Res Rev 1998;11:5–23. [DOI] [PubMed] [Google Scholar]
- 42.Nettleton JA, Harnack LJ, Scrafford CG, Mink PJ, Barraj LM, Jacobs DR Jr. Dietary flavonoids and flavonoid-rich foods are not associated with risk of type 2 diabetes in postmenopausal women. J Nutr 2006;136:3039–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Song Y, Manson JE, Buring JE, Sesso HD, Liu S. Associations of dietary flavonoids with risk of type 2 diabetes, and markers of insulin resistance and systemic inflammation in women: a prospective study and cross-sectional analysis. J Am Coll Nutr 2005;24:376–84. [DOI] [PubMed] [Google Scholar]
- 44.Bolton RP, Heaton KW, Burroughs LF. The role of dietary fiber in satiety, glucose, and insulin: studies with fruit and fruit juice. Am J Clin Nutr 1981;34:211–7. [DOI] [PubMed] [Google Scholar]