Myocardial injury after non-cardiac surgery (MINS) is the most common perioperative cardiovascular complication and is independently associated with 30-day mortality. We aimed to assess the association between preoperative glucose concentration and postoperative MINS and mortality. The VISION study is a prospective cohort study done at 12 centres in eight countries. Patients aged 45 years or older who required at least one overnight hospital admission for non-cardiac surgery were enrolled from Aug 6, 2007, to Jan 11, 2011. In the GlucoVISION analysis, we assessed the relations between preoperative casual or fasting glucose concentration and MINS within 3 days after surgery using logistic regression, and 30-day mortality using Cox proportional regression, in people with and without diabetes. 11 954 patients were included in this analysis, of whom 2809 (23%) had diabetes. Within the first three postoperative days, MINS occurred in 813 (7%) patients. 249 (2%) patients died by day 30. More patients with diabetes had MINS (odds ratio [OR] 1·98 [95% CI 1·70-2·30]; p<0·0001), and died (OR 1·41 [1·08-1·86]; p=0·016) than did patients without diabetes. Casual glucose concentrations were associated with MINS in all patients (adjusted OR 1·06 [1·04-1·09] per 1 mmol/L increment in glucose; p=0·0003), and with death in patients without diabetes (adjusted hazard ratio [HR] 1·13 [95% CI 1·05-1·23] per mmol/L; p=0·002). We noted a progressive relation between unadjusted fasting glucose concentration and both MINS (OR 1·14 [1·08-1·20] per mmol/L; p<0·0001), driven by the effect in the subgroup without previous diabetes (pinteraction=0·025), and 30-day mortality (HR 1·10 [1·02-1·19] per mmol/L; p=0·013). For patients without diabetes, casual glucose of more than 6·86 mmol/L and fasting glucose of more than 6·41 mmol/L predicted MINS (OR 1·71 [1·36-2·15]; p<0·0001, and OR 2·71 [1·85-3·98]; p<0·0001, respectively). For patients with diabetes, only casual glucose concentration more than 7·92 mmol/L predicted MINS (OR 1·47 [1·10-1·96]; p=0·0096). Preoperative glucose concentration, particularly casual glucose concentration, predicts risk for postoperative cardiovascular outcomes, especially in patients without diabetes. Full funding sources listed at the end of the paper (see Acknowledgments).