OBJECTIVES:Low levels of vitamin D are associated with a higher mortality in cirrhotic patients, but the role of this deficiency is still unknown. The purpose of this study was to assess the levels of vitamin D in cirrhotic patients with and without bacterial infection.METHODS:25-hydroxy (25-OH) vitamin D was assessed by immunoassay in 88 patients hospitalized in our hepatology unit.RESULTS:The causes of cirrhosis were mainly alcohol (70%), hepatitis C (10%), or both (9%). Infections (n=38) mainly included bacteriemia (21%), urinary tract infections (24%), and spontaneous bacterial peritonitis (29%). A severe deficiency in vitamin D (<10 ng/ml) was observed in 56.8% of patients. Infections were more frequent in patients with a severe deficiency compared with the others (54 vs. 29%, P=0.02). A severe deficiency in vitamin D was a predictive factor of infection (odds ratio=5.44 (1.35–21.97), P=0.017) independently of the Child–Pugh score (odds ratio=2.09 (1.47–2.97) P=0.00004) and the C-reactive protein level (odds ratio=1.03 (1.002–1.052), P=0.03) in a logistic regression also including the alanine amino transferase (not significant). By a Cox regression analysis, only the presence of an infection was significantly associated with mortality (relative risk=3.24 (1.20–8.76), P=0.02) in a model also associating the Child–Pugh score (not significant) and the presence of a severe deficiency in vitamin D (not significant).CONCLUSIONS:Low levels of 25-OH vitamin D were independently associated with bacterial infections in cirrhotic patients. The impact of 25-OH vitamin D supplementation on the infection rate and death of cirrhotic patients should be assessed in randomized trials.