BACKGROUND: Long-term statin treatment reduces the frequency of cardiovascular events, but safety and efficacy in patients with abnormal liver tests is unclear. We assessed whether statin therapy is safe and effective for these patients through post-hoc analysis of the Greek Atorvastatin and Coronary Heart Disease Evaluation (GREACE) study population.METHODS: GREACE was a prospective, intention-to-treat study that randomly assigned by a computer-generated randomisation list 1600 patients with coronary heart disease (aged <75 years, with serum concentrations of LDL cholesterol >2·6 mmol/L and triglycerides <4·5 mmol/L) at the Hippokration University Hospital, Thessaloniki, Greece to receive statin or usual care, which could include statins. The primary outcome of our post-hoc analysis was risk reduction for first recurrent cardiovascular event in patients treated with a statin who had moderately abnormal liver tests (defined as serum alanine aminotransferase or aspartate aminotransferase concentrations of less than three times the upper limit of normal) compared with patients with abnormal liver tests who did not receive a statin. This risk reduction was compared with that for patients treated (or not) with statin and normal liver tests.FINDINGS: Of 437 patients with moderately abnormal liver tests at baseline, which were possibly associated with non-alcoholic fatty liver disease, 227 who were treated with a statin (mainly atorvastatin 24 mg per day) had substantial improvement in liver tests (p<0·0001) whereas 210 not treated with a statin had further increases of liver enzyme concentrations. Cardiovascular events occurred in 22 (10%) of 227 patients with abnormal liver tests who received statin (3·2 events per 100 patient-years) and 63 (30%) of 210 patients with abnormal liver tests who did not receive statin (10·0 events per 100 patient-years; 68% relative risk reduction, p<0·0001). This cardiovascular disease benefit was greater (p=0·0074) than it was in patients with normal liver tests (90 [14%] events in 653 patients receiving a statin [4·6 per 100 patient-years] vs 117 [23%] in 510 patients not receiving a statin [7·6 per 100 patient-years]; 39% relative risk reduction, p<0·0001). Seven (<1%) of 880 participants who received a statin discontinued statin treatment because of liver-related adverse effects (transaminase concentrations more than three-times the upper limit of normal).INTERPRETATION: Statin treatment is safe and can improve liver tests and reduce cardiovascular morbidity in patients with mild-to-moderately abnormal liver tests that are potentially attributable to non-alcoholic fatty liver disease.FUNDING: None.