NAL-CI randomised 100 patients with compensated alcohol-associated cirrhosis and DSM-5 alcohol use disorder to naltrexone 50 mg daily or placebo for twelve weeks. Both arms received standardised psychosocial support. Groups were well matched, with mean MELD 12.6 against 12.7, Child-Pugh 5.9 against 6.2, and mean age around 43.
Point-prevalence abstinence at twelve weeks — no alcohol in the preceding four weeks — was 64% (32 of 50) with naltrexone and 22% (11 of 50) with placebo, an odds ratio of 10.86. Lapses at three months fell from 54% to 28%. Heavy-drinking relapse showed a trend at 12% against 28%, and abstinence maintained to six months favoured naltrexone at 22% against 8% without reaching significance. Craving scores were substantially lower on both obsessive and compulsive subscales.
The safety data are why this matters. Naltrexone's hepatotoxicity warning has kept it out of the population with most to gain from it. Here no patient developed hepatic decompensation attributable to the drug, and no AST or ALT elevation exceeded five times the upper limit of normal in either arm. Adverse events were comparable.
The boundary is explicit: compensated cirrhosis, mean MELD around 12. This says nothing about decompensated disease.
- Compensated cirrhosis only, mean MELD 12.6 — not decompensated disease
- Both arms received structured psychosocial support; naltrexone is an addition to it
- No decompensation attributable to the drug, no transaminase rise above 5× ULN
- Twelve-week primary endpoint; the six-month difference did not reach significance
- 100 patients, single trial — practice-changing in direction, not yet definitive
The statistics, in plain English
An odds ratio of 10.86 with an interval from 1.89 to 62.2 is enormous and enormously imprecise — with 100 patients the data are compatible with a doubling of the odds or a sixty-fold increase. The absolute figures are the more useful summary: 64% against 22%, a difference of 42 percentage points. The safety finding is a null in a small sample, so it excludes common harm and cannot exclude rare harm.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for gastroenterology & hepatology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free