Incidence of Idiosyncratic Drug-Induced Liver Injury Caused by Prescription Drugs

Background The incidence of drug-induced liver injury (DILI) is not known for most prescription medications. We aimed to estimate the incidence of DILI for commonly prescribed outpatient drugs. Methods To establish a baseline estimate of DILI incidence, we used the estimated incidence (EI) of amoxicillin/clavulanate DILI from a previous population-based study in Iceland. This was combined with the multicenter prospective DILI Network (DILIN) cohort and the US population-based Medical Expenditure Panel Survey (MEPS). From 2005 to 2019, prescription drugs with at least five bona fide DILIN cases and data from at least 10 of the 15 years from MEPS during that timeframe were included. The EI for ‘drug A’ was calculated as follows: $$\text{EI}(\text{drug A})=\text{EI}\left(\text{AC}\right)\times \frac{\# \; \text{DILIN cases of drug A}}{\# \; \text{annual new prescriptions of drug A}}\times \frac{\# \; \text{annual new prescriptions of AC}}{\# \; \text{DILIN cases of AC}}$$ EI ( drug A ) = EI AC × # DILIN cases of drug A # annual new prescriptions of drug A × # annual new prescriptions of AC # DILIN cases of AC Results In total, 30 drugs met the inclusion criteria, of which 11 were antibiotics, 4 were antiepileptic drugs (AEDs), 4 were statins, and 11 were other drug types. The highest EI was seen with azathioprine and older AEDs, with one DILI case per 349–2329 new prescriptions. The EI of antibiotics ranged greatly, with the highest risk seen for minocycline, amoxicillin/clavulanate, and nitrofurantoin (approximately 1:1000–2400 new prescriptions), and lowest risk for clindamycin, doxycycline, azithromycin, and amoxicillin (approximately 1:40,000–170,000 new prescriptions). The EI for commonly prescribed statins was approximately 1:10,000–50,000. Important medication classes with > 5 million new prescriptions from 2005 to 2019 but fewer than five DILIN cases included β-blockers, thiazide diuretics, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, selective serotonin reuptake inhibitors, and metformin, which presumably have very low DILI incidence. Conclusions The highest EI was found for azathioprine, older antiepileptics, and minocycline. In contrast, many widely used drugs are rare causes of DILI. These findings may help clinicians better weigh potential benefits of medications against hepatotoxicity risk.