There is emerging evidence that proton pump inhibitors (PPIs), which are being used increasingly for upper gastrointestinal symptoms and pathology, are associated with cognitive decline, nephropathy and vasculopathy in the general population. To determine whether there is a similar safety signal in diabetes, we analysed data from the longitudinal community-based Fremantle Diabetes Study Phase II (FDS2). FDS2 includes 1,551 participants with type 2 diabetes (T2DM) followed through validated hospital morbidity/mortality databases. At baseline in 2008-2011, the T2DM participants were of mean±SD age 65.7±11.6 years, 51.9% were males and 342 (22.1%) were taking a PPI. During 3.6±1.1 years of follow-up, 141 (9.1%) died. In Cox proportional hazards modelling using age as the time-scale and with a range of other plausible independent variables included, baseline PPI use was associated with an increased risk of death (hazard ratio (95% CI) 1.52 (1.05 to 2.22); P=0.029). In 125 patients who were started on a PPI during follow-up, there was a mean difference (Δ) (95% CI) 9.5 (-2.7 to 21.7) mg/mmol increase in urinary albumin:creatinine (from a geometric mean (SD range) of 3.0 (0.9 to 10.3) to 3.8 (1.0 to 14.6) mg/mmol; P=0.031). In the subset of these patients (n=95) who were on a stable dose of ACE inhibitor (ACEI) or angiotensin receptor blocker (ARB) before and after PPI initiation, this difference was attenuated (Δ 11.4 (-4.4 to 27.2) mg/mmol or from 3.0 (0.8 to 11.1) to 3.7 (0.9 to 15.6) mg/mmol; P=0.084) but there was no increase in the 736 patients on stable ACEI/ARB therapy who did not start a PPI (Δ 2.7 (-0.9 to 6.20) mg/mmol or 2.8 (0.8 to 9.9) to 2.9 (0.7 to 12.1) mg/mmol; P=0.48). These data suggest that PPI use may be adverse in T2DM, increasing short-term mortality and worsening endothelial dysfunction as evidenced by increased urinary albumin excretion.