Total and specific potato intake and risk of type 2 diabetes: results from three US cohort studies and a substitution meta-analysis of prospective cohorts

Abstract

Objectives: To investigate the associations between total and individual potato intake and risk of type 2 diabetes (T2D), estimate the effect on T2D risk of replacing potatoes with whole grains and other major carbohydrate sources, and conduct a dose-response and substitution meta-analysis of prospective cohort studies.

Design: Prospective cohort study and dose-response meta-analysis of prospective cohort studies.

Setting: Individual participant data from Nurses' Health Study (1984-2020), Nurses' Health Study II (1991-2021), and Health Professionals Follow-up Study (1986-2018).

Participants: 205 107 men and women free of diabetes, cardiovascular disease, or cancer at baseline.

Main outcome measure: Incident type 2 diabetes.

Results: During 5 175 501 person years of follow-up, T2D was documented in 22 299 participants. After adjustment for updated body mass index and other diabetes related risk factors, higher intakes of total potatoes and French fries were associated with increased risk of T2D. For every increment of three servings weekly of total potato, the rate for T2D increased by 5% (hazard ratio 1.05, 95% confidence interval (CI) 1.02 to 1.08) and for every increment of three servings weekly of French fries the rate increased by 20% (1.20, 1.12 to 1.28). Intake of combined baked, boiled, or mashed potatoes was not significantly associated with T2D risk (pooled hazard ratio 1.01, 95% CI 0.98 to 1.05). In substitution analyses, replacing three servings weekly of potatoes with whole grains was estimated to lower T2D rates by 8% (95% CI 5% to 11%) for total potatoes, 4% (1% to 8%) for baked, boiled, or mashed potatoes, and 19% (14% to 25%) for French fries. In contrast, replacing total potatoes or baked, boiled, or mashed potatoes with white rice was associated with an increased risk of T2D. In a meta-analysis of 13 cohorts (587 081 participants and 43 471 diagnoses of T2D), the pooled hazard ratio for risk of T2D with each increment of three servings weekly of total potato was 1.03 (95% CI 1.02 to 1.05) and of fried potatoes was 1.16 (1.09 to 1.23). In substitution meta-analyses, replacing three servings weekly of total, non-fried, and fried potatoes with whole grains was estimated to lower the risk of T2D by 7% (95% CI 5% to 9%), 5% (3% to 7%), and 17% (12% to 22%), respectively.

Conclusions: Higher intake of French fries, but not combined baked, boiled, or mashed potatoes, was associated with a higher risk of T2D. The T2D risk linked to potato intake seemed to depend on the food being replaced: replacing potato with whole grains was associated with lower risk, whereas replacing with white rice was associated with increased risk.

Fig 1

Dose-response relations between intake of total potato; baked, boiled, or mashed potatoes; and French fries and incidence of type 2 diabetes. Data were combined from three cohorts (Nurses’ Health Study, Nurses’ Health Study II, and Health Professionals Follow-up Study; total n=205 107). Associations were evaluated using restricted cubic spline models with three knots placed at the 10th, 50th, and 90th centiles, adjusting for multiple covariates. The model was stratified by age (months), cohort, and calendar time (two year intervals), and adjusted for total energy intake, race/ethnicity (white adults, non-white adults), smoking status (never, past, and current (cigarettes/day): 1-14, >15-24, >24), alcohol intake (0, 0-4.9, 5-9.9, 10-14.9, 15-29.9, and >30 g/day), physical activity (<3, 3.0-8.9, 9.0-17.9, 18.0-26.9, ≥27 MET-h/week), multivitamin use, menopausal status and hormone use (Nurses’ Health Study or Nurses’ Health Study II), family history of type 2 diabetes, antihypertensive use, cholesterol lowering drug use, history of hypertension, socioeconomic status, time varying body mass index (<21, 21.0-22.9, 23.0-24.9, 25.0-26.9, 27.0-29.9, 30.0-32.9, 33.0-34.9, 35-39.9, ≥40), and dietary covariates intakes (including total red meat, poultry, fish, egg, total dairy, nuts and legumes, fruits, vegetables, sugar sweetened beverages, whole grain, and refined grain, and mutual adjustment for baked, boiled or mashed potatoes and for French fries. MET-h=metabolic equivalent tasks per hour

Fig 2

Associations between replacing potato intake with other foods and incidence of type 2 diabetes in the Nurses’ Health Study, Nurses’ Health Study II, and Health Professionals Follow-up Study (n=205 107). The effects of replacing three servings weekly of potatoes with alternative foods were assessed by simultaneously including both food intakes as continuous variables in the same multivariable Cox regression model. The model was stratified by age (months) and calendar time (two year intervals) and adjusted for race/ethnicity (white adults, non-white adults), smoking status (never, past, and current (cigarettes/day): 1-14, >15-24, >24), alcohol intake (0, 0-4.9, 5-9.9, 10-14.9, 15-29.9, >30 g/day), physical activity (<3, 3.0-8.9, 9.0-17.9, 18.0-26.9, ≥27 MET-h/week), multivitamin use, menopausal status and hormone use (Nurses’ Health Study or Nurses’ Health Study II), family history of type 2 diabetes, antihypertensive use, cholesterol lowering drug use, history of hypertension, socioeconomic status, time varying body mass index (<21, 21.0-22.9, 23.0-24.9, 25.0-26.9, 27.0-29.9, 30.0-32.9, 33.0-34.9, 35-39.9, ≥40), and dietary covariates intakes (including total red meat, poultry, fish, egg, total dairy, nuts and legumes, fruits, vegetables, sugar sweetened beverages, whole grain, and refined grain, and mutual adjustment for baked, boiled, or mashed potatoes and French fries) excluding the foods subject for substitutions. The models for brown rice and white rice did not include adjustments for whole grain and refined grain intakes, respectively. MET-h=metabolic equivalent tasks per hour

Fig 3

Estimated effects of replacing three servings weekly of different types of potatoes with whole grains on incidence of type 2 diabetes. Estimates are based on meta-analyzed data for potatoes and whole grains