Aims/hypothesis Despite the condition's high prevalence, the influence of hyperglycaemia on clinical outcomes in noncritical-care inpatients with infections remains ill defined. In this study, we analysed associations of glucose levels at admission and during initial inpatient treatment with the inflammatory response and clinical outcome in communityacquired pneumonia (CAP) patients. Methods This secondary observational analysis included 880 confirmed CAP patients. We used severity-adjusted multivariate regression models to investigate associations of initial and 96 h mean glucose levels with serially measured biomarker levels over 7 days (C-reactive protein [CRP], procalcitonin, white blood cell count [WBC], pro-adrenomedullin [ProADM]) and adverse clinical course (death and intensivecare unit admission). Results In the 724 non-diabetic patients (82.3% of the study population), moderate or severe hyperglycaemia (glucose 6-11 mmol/l and >11 mmol/l, respectively) was associated with increased risk for adverse clinical course (adjusted OR [95% CI] 1.4 [0.8, 2.4] and 3.0 [1.1, 8.0], respectively) and with higher CRP, WBC and ProADM levels over 7 days (p <0.05, ANOVA, all days). In diabetic patients (n =156), no similar associations were found for initial hyperglycaemia, although mean 96 h glucose levels ≥9 mmol/l were associated with adverse clinical course (adjusted OR 5.4 [1.1, 25.8]; p =0.03). No effect modification by insulin treatment was detected (interaction terms p >0.2 for all analyses). Conclusions/interpretation Initial hyperglycaemia in nondiabetic CAP patients, and prolonged hyperglycaemia in diabetic or non-diabetic CAP patients, are associated with a more pronounced inflammatory response and CAP-related adverse clinical outcome. Optimal glucose targets for insulin treatment of hyperglycaemia in non-critical-care settings should be defined.