Participants in the highest versus the lowest quintiles of dietary diabetes risk-reduction score were less likely to be male, be married or living as married, be current smokers, and have a history of hypertension but more likely to be single or multivitamin supplement users, have lower body mass index and energy intake from diet and had higher educational level, alcohol consumption, physical activity level, and Healthy Eating Index 2015 (Table 1). In addition, compared with participants in the lowest quintile of dietary diabetes risk-reduction score, those in the highest quintile had lower glycemic index and lower intakes of trans-fatty acids, sugar-sweetened beverages, red and processed meat, and saturated fatty acids but higher intakes of cereal fiber, nuts, coffee, fruits, and vegetables.
Dietary Diabetes Risk-reduction Score and All-cause and Cause-specific Mortality
During 1,174,401.6 person-years of follow-up, we observed a total of 17,532 all-cause deaths, of which 4,809 (27.4%) were attributable to cardiovascular disease and 5,719 (32.6%) to cancer (Table 2). The mean follow-up was 13.6 (standard deviation, 3.2) years. The crude death rates per 10,000 person-years were 149.28, 40.95, and 48.70 for mortality from all causes, cardiovascular disease, and cancer, respectively, which were obviously lower than those from the National Institutes of Health–AARP study, a contemporary US cohort study involving 521,120 participants (176.99, 53.03, and 62.66 deaths per 10,000 person-years for mortality from all causes, cardiovascular disease, and cancer, respectively). In the fully adjusting model, participants in the highest (5th) vs. the lowest (1st) quintiles of dietary diabetes risk-reduction score were found to be at lower risks of death from all causes (HR = 0.76, 95% CI: 0.72, 0.80; P for trend< 0.001; absolute risk difference = −81.94, 95% CI: −93.76, −71.12), cardiovascular disease (HR = 0.73, 95% CI: 0.66, 0.81; P for trend < 0.001; absolute risk difference = −17.82, 95% CI: −24.81, −11.30), and cancer (HR = 0.85, 95% CI: 0.78, 0.94; P for trend < 0.001; absolute risk difference = −9.92, 95% CI: −15.86, −3.59) (Table 2 and Web Table 4). We obtained similar results when repeating the above-mentioned Cox regression analyses in participants with complete data (Web Table 5) and using the modified dietary diabetes risk-reduction score (Web Table 6).
Interestingly, subgroup analyses found that the inverse association with cardiovascular mortality was more pronounced in women than in men (P for interaction = 0.024), whereas the inverse association with cancer mortality was more pronounced among men than women (P for interaction = 0.032) (Table 3 and Web Table 7). Moreover, the inverse associations with all-cause (P for interaction = 0.023) and cancer (P for interaction = 0.023) mortality were more pronounced among participants with heavy alcohol consumption than those with no, light, or moderate alcohol consumption. In addition, the inverse association with cancer mortality was more pronounced among current or past smokers than never smokers (P for interaction = 0.002). No significant interaction effect was found for the remaining stratification factors (all P for interaction > 0.05).
The initial associations of dietary diabetes risk-reduction score with risks of death from all causes, cardiovascular disease, and cancer did not change materially in a large range of sensitivity analyses (Web Table 8).
Associations by Each Component of Type 2 Diabetes–Prevention Diet
Comparing quintile 5 to quintile 1, higher intake of cereal fiber (HR = 0.79, 95% CI: 0.74, 0.85; P for trend < 0.001), nuts (HR = 0.82, 95% CI: 0.78, 0.86; P for trend < 0.001), or coffee (HR = 0.88, 95% CI: 0.84, 0.93; P for trend< 0.001) was found to be associated with a lower risk of all-cause mortality, whereas higher intake of sugar-sweetened beverages (HR = 1.00, 95% CI: 0.95, 1.05; P for trend = 0.022) was found to be associated with a higher risk of all-cause mortality (Table 4); moreover, an inverse association was found for the ratio of polyunsaturated to saturated fatty acids and all-cause mortality (HR = 0.84, 95% CI: 0.80, 0.89; P for trend< 0.001). A marginally significant positive association was found for red and processed meat consumption and all-cause mortality (HR = 1.01, 95% CI: 0.95, 1.07; P for trend = 0.052). Similar results were obtained for cardiovascular and/or cancer mortality. No significant associations with mortality from all causes, cardiovascular disease, and cancer were found for glycemic index and trans-fatty acid intake.
Am J Epidemiol. 2022;191(3):472-486. © 2022 Oxford University Press