Background: Non-steroidal anti-inflammatory drugs have been implicated in reports of liver injury. However, the precise risk of non-steroidal anti-inflammatory drugs for this rare complication is unknown.
Aim: To review systematically the published literature of population-based epidemiological studies reporting the incidence or comparative risk of non-steroidal anti-inflammatory drugs for liver injury resulting in clinically significant events, defined as hospitalization or death.
Data extraction: Duplicate extraction of the methodological quality, design, source, population, years studied, particular non-steroidal anti-inflammatory drugs studied, definitions, patient counts and follow-up, and the adjustment for confounders.
Results: Seven articles met inclusion criteria. The comparative risk of liver injury resulting in hospitalization for current non-steroidal anti-inflammatory drug users compared with past non-steroidal anti-inflammatory drug users ranged from 1.2 to 1.7, but none was statistically significant. The incidence of liver injury resulting in hospitalization ranged from 3.1 to 23.4/100,000 patient-years of current use of non-steroidal anti-inflammatory drugs, with an excess risk compared with past non-steroidal anti-inflammatory drugs users of 4.8-8.6/100,000 patient-years of exposure. There were zero deaths from liver injury associated with non-steroidal anti-inflammatory drugs use in over 396,392 patient-years of cumulative exposure.
Conclusion: These findings allow for the possibility of a small increase in the risk of clinically relevant hepatotoxicity with non-steroidal anti-inflammatory drugs use, but do not document that such a risk occurs.