Naltrexone vs. Acamprosate for Alcohol Use Disorder: Organ-Function-Driven Selection
A man newly diagnosed with compensated cirrhosis wants to start medication for alcohol use disorder today. The two first-line options split cleanly along organ clearance — but the drug that’s cleanest for his liver isn’t the one with the stronger trial result.
At his cousin's funeral eighteen months ago, R.M. learned his father had cirrhosis — the family had never discussed it, and hearing it secondhand across a reception hall was what first made him start paying attention to his own drinking. He is fifty-two, has run the same auto-body shop for twenty-six years, the last eight of them alone since his business partner retired, and has lived in the apartment above the shop since his divorce finalized fourteen months ago — close enough to work that he stopped noticing how rarely he left the block some weeks. He was previously healthy apart from well-controlled hypertension on lisinopril, with no diabetes and no liver disease of his own, and no family history he knew of before that funeral.
The diagnosis he's carrying now came from a physical he'd been putting off, ordered mostly to reassure him. Instead it returned AST 118 and ALT 96, and an abdominal ultrasound read as a coarsened, echogenic liver with early nodularity — findings consistent with compensated cirrhosis, Child-Pugh class A. Asked directly, he admitted to ten to twelve drinks most evenings for roughly two years, a pattern that escalated steadily after the divorce and that he hadn't said out loud to anyone before his doctor asked. He has no ascites, no encephalopathy, and no variceal history on endoscopy, and his synthetic function — albumin, INR, bilirubin — remains normal. The histologic direction is already set, though, and it is set specifically by continued drinking, not by the disease existing in the abstract.
This is his first real attempt at treatment. He has never tried counseling, never tried a formal program, only years of assuming he could manage it alone above the shop where no one was watching. What changed, he says, wasn't the labs so much as picturing his father's cirrhosis becoming his own — not an abstraction anymore but a specific, familiar shape. He has told his doctor plainly that he is done handling it quietly on his own.
Choosing a first medication with an already-injured liver
Start with acamprosate. It isn't hepatically metabolized at all — it's absorbed unchanged and cleared renally, so his AST and ALT are simply irrelevant to how the drug behaves in his body, which is not something I can say about naltrexone. Naltrexone's label carries a real hepatotoxicity caution, dose-related and mostly described at doses well above the standard 50 mg — but he already has two elevated transaminases and imaging consistent with early cirrhosis, and I don't want to be the reason his next set of labs is ambiguous between drug effect and disease progression.
His renal function is completely normal, so acamprosate's own dosing constraint — it needs reduction below an eGFR of 50 and is contraindicated under 30 — doesn't apply to him at all. This is about as clean a renal-favors-this-drug case as exists.
I want to name the complication directly: COMBINE, the largest US trial of these two drugs head-to-head, found naltrexone plus structured medical management produced a real improvement in percent days abstinent — acamprosate, in that same trial, did not separate from placebo. If we pick acamprosate purely on the renal-versus-hepatic clearance logic, we're picking the drug that underperformed in the most relevant real-world US trial, on a patient who has never tried any pharmacotherapy and may not get many more real attempts.
The hepatotoxicity signal you're citing is real, but it's overwhelmingly drawn from trials using doses several times higher than what we'd prescribe him — the FDA's own labeled caution is about supratherapeutic exposure, not the standard 50 mg dose in a patient with compensated, Child-Pugh A disease. Continued drinking is doing far more liver damage right now than 50 mg of naltrexone plausibly would.
You're both right about something the other is underweighting, and I don't think this needs to be resolved as a single irreversible choice. Start acamprosate now — it removes any hepatic question entirely, and he can begin today without waiting on anything. But COMBINE's effect size is real and shouldn't just be set aside because the pharmacology looks cleaner on the acamprosate side; and COMBINE mostly enrolled patients without significant liver disease, so its result about acamprosate's weaker average performance may not even generalize cleanly to him.
If his transaminases hold steady or improve over the first month on acamprosate alone and cravings remain the limiting factor, low-dose naltrexone with baseline-and-monthly LFT monitoring becomes a reasonable add, not a reach. If his liver panel worsens on acamprosate alone, that's itself informative about how much runway we actually have before naltrexone's caution stops being theoretical.
Agreed: acamprosate started immediately, baseline hepatic panel drawn today, repeat panel in four weeks, addiction medicine follow-up scheduled for the same interval to reassess craving control and decide on naltrexone.
Not agreed, and carried forward rather than smoothed over: how much weight COMBINE's efficacy difference should carry against a patient this trial didn't really enroll. The addiction psychiatrist still reads the trial as the single best available guide to which drug is more likely to work and would add naltrexone readily at the first sign acamprosate isn't enough on its own. The hepatologist remains genuinely reluctant to introduce any hepatically-flagged drug into an already-scarring liver even with normal early labs, and wants at least two clean monthly panels, not one, before agreeing to it. Nothing about today's regimen forces that disagreement to resolve before it needs to.