Summary Background High blood glucose concentration may increase risk of death and poor outcome after acute myocardial infarction. We did a systematic review and metaanalysis to assess the risk of in-hospital mortality or congestive heart failure after myocardial infarction in patients with and without diabetes who had stress hyperglycaemia on admission. Methods We did two searches of MEDLINE for Englishlanguage articles published from 1966 to October, 1998, a computerised search of Science Citation Index from 1980 to September, 1998, and manual searches of bibliographies. Two searchers identified all cohort studies or clinical trials reporting in-hospital mortality or rates of congestive heart failure after myocardial infarction in relation to glucose concentration on admission. We compared the relative risks of in-hospital mortality and congestive heart failure in hyperglycaemic and normoglycaemic patients with and without diabetes. Findings 14 articles describing 15 studies were identified. Patients without diabetes who had glucose concentrations more than or equal to range 6·1–8·0 mmol/L had a 3·9-fold (95% CI 2·9–5·4) higher risk of death than patients without diabetes who had lower glucose concentrations. Glucose concentrations higher than values in the range of 8·0–10·0 mmol/L on admission were associated with increased risk of congestive heart failure or cardiogenic shock in patients without diabetes. In patients with diabetes who had glucose concentrations more than or equal to range 10·0–11·0 mmol/L the risk of death was moderately increased (relative risk 1·7 [1·2–2·4]). Interpretation Stress hyperglycaemia with myocardial infarction is associated with an increased risk of in-hospital mortality in patients with and without diabetes; the risk of congestive heart failure or cardiogenic shock is also increased in patients without diabetes.