Dietary protein and risk of type 2 diabetes: findings from a registry-based cohort study and a meta-analysis of prospective cohort studies

Abstract Objectives Lifestyle intervention, especially dietary modification, has been the cornerstone in preventing type 2 diabetes (T2D). We aimed to investigate the associations of various protein intake exposures with the risk of incident T2D in adults with or without metabolic diseases. Methods We followed 29517 residents enrolled in the Shanghai Suburban Adult Cohort and Biobank (SSACB) without diabetes at baseline through the electronic information system. Cox proportional hazard models were used to evaluate the associations of various protein intake exposures with the risk of incident T2D, visualized by restricted cubic splines (RCS). Propensity-score matching and subgroup analysis were used to characterize the association between total protein and incident T2D by metabolic diseases. Meta-analysis further explored the association between protein intake and incident T2D in broader populations. Results In SSACB, 1511 (5.1%) participants developed T2D during a median follow-up period of 5.69 years. A U-shaped association between total protein and risk of incident T2D was found (protective range: 12.20-16.85 percentage energy (%E), cut-off point: 14.53%E). The U-shaped association (P-nonlinear < 0.001) remained in adults with hypertension with a narrower protective range (12.20–15.35%E), with a linear association in adults with NAFLD (HR per 1%E: 0.952, 95% CI: [0.910, 0.995]), whereas no significant association in adults with hyperlipidemia or central obesity. A negative association between plant protein and risk of incident T2D was also found in SSACB (HR per 1%E: 0.947, 95% CI: [0.900, 0.996]). In addition, the U-shaped association of total protein with the risk of incident T2D was reaffirmed in the dose-response meta-analysis (cut-off point: 15.10%E). Conclusion In SSACB, a U-shaped association between total protein intake and risk of incident T2D was found, which was reaffirmed in the dose-response meta-analysis, and differed by metabolic diseases, especially hypertension and NAFLD. Moreover, plant protein was inversely associated with the risk of incident T2D.