BACKGROUND: The proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitor evolocumab reduced LDL cholesterol and cardiovascular events in the FOURIER trial. In this prespecified analysis of FOURIER, we investigated the efficacy and safety of evolocumab by diabetes status and the effect of evolocumab on glycaemia and risk of developing diabetes. METHODS: FOURIER was a randomised trial of evolocumab (140 mg every 2 weeks or 420 mg once per month) versus placebo in 27 564 patients with atherosclerotic disease who were on statin therapy, followed up for a median of 2·2 years. In this prespecified analysis, we investigated the effect of evolocumab on cardiovascular events by diabetes status at baseline, defined on the basis of patient history, clinical events committee review of medical records, or baseline HbA of 6·5% (48 mmol/mol) or greater or fasting plasma glucose (FPG) of 7·0 mmol/L or greater. The primary endpoint was a composite of cardiovascular death, myocardial infarction, stroke, hospital admission for unstable angina, or coronary revascularisation. The key secondary endpoint was a composite of cardiovascular death, myocardial infarction, or stroke. We also assessed the effect of evolocumab on glycaemia, and on the risk of new-onset diabetes among patients without diabetes at baseline. HbA was measured at baseline then every 24 weeks and FPG was measured at baseline, week 12, week 24, and every 24 weeks thereafter, and potential cases of new-onset diabetes were adjudicated centrally. In a post-hoc analysis, we also investigated the effects on glycaemia and diabetes risk in patients with prediabetes (HbA 5·7-6·4% [39-46 mmol/mol] or FPG 5·6-6·9 mmol/L) at baseline. FOURIER is registered with ClinicalTrials.gov, number NCT01764633. FINDINGS: At study baseline, 11 031 patients (40%) had diabetes and 16 533 (60%) did not have diabetes (of whom 10 344 had prediabetes and 6189 had normoglycaemia). Evolocumab significantly reduced cardiovascular outcomes consistently in patients with and without diabetes at baseline. For the primary composite endpoint, the hazard ratios (HRs) were 0·83 (95% CI 0·75-0·93; p=0·0008) for patients with diabetes and 0·87 (0·79-0·96; p=0·0052) for patients without diabetes (p=0·60). For the key secondary endpoint, the HRs were 0·82 (0·72-0·93; p=0·0021) for those with diabetes and 0·78 (0·69-0·89; p=0·0002) for those without diabetes (p=0·65). Evolocumab did not increase the risk of new-onset diabetes in patients without diabetes at baseline (HR 1·05, 0·94-1·17), including in those with prediabetes (HR 1·00, 0·89-1·13). Levels of HbA and FPG were similar between the evolocumab and placebo groups over time in patients with diabetes, prediabetes, or normoglycaemia. Among patients with diabetes at baseline, the proportions of patients with adverse events were 78·5% (4327 of 5513 patients) in the evolocumab group and 78·3% (4307 of 5502 patients) in the placebo group; among patients without diabetes at baseline, the proportions with adverse events were 76·8% (6337 of 8256 patients) in the evolocumab group and 76·8% (6337 of 8254 patients) in the placebo group. INTERPRETATION: PCSK9 inhibition with evolocumab significantly reduced cardiovascular risk in patients with and without diabetes. Evolocumab did not increase the risk of new-onset diabetes, nor did it worsen glycaemia. These data suggest evolocumab use in patients with atherosclerotic disease is efficacious and safe in patients with and without diabetes. FUNDING: Amgen.