Letter from the Editor

Antibacterials are frequently associated with idiosyncratic drug‐induced liver injury (DILI). The objective of this study was to estimate the risk of macrolides and amoxicillin/clavulanate (AMC) on DILI. We conducted a systematic review (SR) and meta‐analysis (MA) with studies retrieved from PubMed, Cochrane Library Plus, Web of Knowledge, clinicaltrials.gov, Livertox and Toxline (1980–2014). We searched for macrolides, AMC and MeSH and synonym terms for DILI. We included all study designs except case reports/series, all population ages and studies with a placebo/non‐user comparator. We summarized the evidence with a random‐effects MA. Quality of the studies was appraised with a checklist developed for SR of adverse effects. Heterogeneity and publication bias were assessed with different exploratory tools. We finally included 10 (two randomized clinical trials, six case–control, one cohort and one case–population studies) and 9 (case–population excluded) articles in the SR and MA, respectively. The overall summary relative risk of DILI for macrolides was 2.85 [95% confidence interval (CI) 1.81–4.47], p < 0.0001, I² = 57%. Three studies were perceived to be missing in the area of low statistical significance. Year of study and selected exposure window partly explained the variability between studies. For AMC, the risk of DILI was 9.38 (95% CI 0.65–135.41) p = 0.3, I2 = 95%. In conclusion, although spontaneous reports and case series have long established an association between macrolides and AMC with acute liver injury, these SR and MA have assessed the magnitude of this association. The low incidence of DILI and the therapeutic place of these antibiotics might tilt the balance in favour of their benefits.