Steroids in Severe Alcohol-Associated Hepatitis: When a Relative Contraindication Isn't Absolute
A patient meets every severity threshold for corticosteroid therapy in severe alcohol-associated hepatitis — except that he was treated for spontaneous bacterial peritonitis three days ago. The pivot is whether a treated, source-controlled infection should function as the same hard stop as an active, uncontrolled one.
Walter K., a 52-year-old man, ran a small auto-body shop for over twenty years before his drinking cost him the business three years ago; his younger brother, who still works the counter at a hardware store nearby, is the one who brought him to the hospital when his eyes turned yellow and he stopped making sense some evenings. He drank heavily and daily for at least fifteen years, and this admission is his first for anything liver-related. He presented six days ago with worsening jaundice, low-grade fever, and diffuse abdominal discomfort; a diagnostic paracentesis grew a single organism, and he completed three days of targeted intravenous antibiotics for spontaneous bacterial peritonitis, with today's repeat paracentesis showing a clearing cell count and resolved fever. He has no prior liver disease diagnosis and no other chronic medical conditions.
His labs place him well inside the treatment window for severe disease: a Maddrey discriminant function of 38 and a MELD score of 26, both comfortably above the thresholds that define who benefits from corticosteroid therapy in STOPAH, the largest randomized trial of prednisolone for this diagnosis. What STOPAH's own enrollment criteria excluded, though, was uncontrolled infection — and whether his SBP, treated and now clinically resolving after three days, still counts as the risk that exclusion was written to keep out is the actual question in front of the team, not whether he needs treatment at all.
Whether treated infection changes the threshold
His numbers say treat, and I'd start prednisolone today. STOPAH's own benefit signal was strongest in the first 28 days, and every day we wait for a 'cleaner' picture is a day of ongoing severe disease he may not have to spare. Three days of source-appropriate antibiotics with a clearing paracentesis and resolved fever is not the same clinical state as the uncontrolled infection the trial actually excluded.
I understand the urgency, but STOPAH's exclusion criterion is written the way it's written because the trial investigators judged active infection too risky to include at all — not because they'd worked out where 'treated enough' begins. Immunosuppressing him 72 hours into treatment for a single-organism SBP, even a clearing one, is still immunosuppressing him inside the window most relapses or secondary infections would show up.
Calling this the same population STOPAH excluded assumes treated and untreated infection carry equivalent risk, and that's the piece the trial actually doesn't tell us either way — the exclusion criterion tells us the investigators didn't test it, not that the answer is no.
Neither of you is wrong about the uncertainty, which is exactly why I'd stop treating this as a single irreversible decision. Start prednisolone today, but commit now, in writing, to a hard reassessment at day 4 for fever, hemodynamic change, or rising white count — the same interval his Lille score reassessment would happen on anyway. An open-ended 'wait and see' rarely gets revisited on schedule once the days in between feel uneventful; a pre-committed checkpoint forces the infection question back onto the table regardless.
Prednisolone started at hospital day 6, with an explicit day-4 reassessment plan documented in the chart for fever, hemodynamic instability, or rising white count, alongside the already-planned day-7 Lille score check for treatment response.
All three services signed off on this specific plan; the infectious disease physician's residual concern was recorded directly rather than smoothed into consensus, so a future reader of the chart knows the caution was real, not overruled without a trace.