Twenty-two studies (n=177,260) were included: one randomised controlled trial (RCT); seven prospective cohort studies; 12 retrospective cohort studies; and two retrospective case-control studies. The sample size ranged from 53 to 69,168 patients. Approximately two thirds of the studies reported a statistically significant benefit with statin use.
Patients with sepsis or at risk of sepsis (nine studies): In one RCT there was no statistically significant difference between pravastatin and placebo in incidence of sepsis in patients with subarachnoid haemorrhage, although there was a reduction in sepsis-related mortality in the statin group (6.25% in the statin group versus 71.43% in the placebo group, p<0.001). Three cohort studies that used propensity-matched sub-cohorts reported a statistically significant association between use of statins and reduced sepsis-related hospitalisations and infection-related mortality. Two of five non-matched cohort studies reported decreased sepsis/infection rates with statin use, one reported no difference and two did not report this outcome. Satin use was associated with decreased overall 30-day or hospital mortality in one study, increased mortality in one study and no difference in two studies.
Patients with community acquired pneumonia (seven studies): In five studies (including four with propensity-matched cohorts and one case-control study), statin use was associated with decreased 30-day mortality; one study reported no difference. Two case-control studies reported decreased risk of pneumonia or fatal pneumonia with current statin use.
Patients with bacteraemia (three studies): In two cohort studies statin use was associated with decreased hospital mortality and bacteraemia-attributable mortality; in one study mortality was not lower with statins.
Prevention of infection in postoperative setting (three studies): In one cohort study of cardiac surgery patients, statin use was not associated with reduced incidence of serious infection and in one study was associated with reduced incidence of infection. In the single study that used propensity score analysis there was no benefit with statin use.
Seven potentially relevant RCTs were identified that were ongoing or not published at the time of the review.