Preoperative blood glucose concentrations and postoperative outcomes after elective non-cardiac surgery: an observational study

BACKGROUND: The association between preoperative blood glucose (BG) concentration and outcomes after non-cardiac surgery and the impact of the diabetes diagnosis status remain unclear. We tested two hypotheses: that preoperative BG is related to surgical outcomes; and that this relationship depends on the diabetes diagnosis status of the patient. METHODS: We retrospectively analysed data on 61 536 consecutive elective non-cardiac surgery patients treated at our tertiary care facility. Logistic regression models were used to test the hypotheses before and after adjustment for baseline patient characteristics. Our primary outcome was a composite of in-hospital serious complications and mortality. A second primary outcome was 1 yr mortality. RESULTS: The crude incidence of the composite in-hospital outcome was significantly related to preoperative BG (P<0.001), but not after covariable adjustment (P=0.40). This relationship did not significantly differ between patients with and without diagnosed diabetes (P=0.09). One year mortality was significantly related to preoperative BG, both univariably (P<0.001) and after covariable-adjustment (P<0.001). Patients with diagnosed diabetes and preoperative euglycaemia generally had worse 1 yr mortality than those without diabetes at the same BG {e.g. odds ratio (OR) [95% confidence interval (CI)] of 1.27 (1.06, 1.53) at 6 mmol litre-1 (108 mg dl-1), P=0.003}. Conversely, hyperglycaemic patients with diagnosed diabetes displayed a significantly lower 1 yr mortality than hyperglycaemic patients without diabetes [OR (95% CI) of 0.58 (0.44, 0.77) at 12 mmol litre-1 (216 mg dl-1), P