Study that followed 29,517 residents without diabetes at baseline showed a U-shaped association between total protein intake and risk of incident type 2 diabetes, which was reaffirmed by a dose-response meta-analysis. Moreover, plant protein was inversely associated with the risk.