Lactulose or Rifaximin: Is This Really a Treatment Failure?
A cirrhotic patient returns with his third hepatic encephalopathy episode in four months, still nominally on lactulose. Whether he has truly failed the drug, or simply couldn't tolerate taking it as prescribed, changes what happens next.
Robert A., a 58-year-old church custodian, has missed four of the last twelve Sundays — the days his job matters most, keeping the sanctuary and fellowship hall running through back-to-back services — because he couldn't be more than a few steps from a bathroom. He was diagnosed with NASH-related cirrhosis three years ago, Child-Pugh class B, and has had two hospitalizations for overt hepatic encephalopathy in the last four months, the most recent three weeks ago. He has been on lactulose since the first episode, titrated, on paper, to three soft stools daily. What his outpatient chart doesn't capture is that he has quietly been taking less than prescribed for the last six weeks, because the diarrhea it produces has been unpredictable enough to cost him, by his own count, two near-accidents on the job.
He arrives today with the same picture as his last two admissions — slowed speech, difficulty with serial subtraction, a flap on wrist extension his wife immediately recognized because she's seen it twice before — and an ammonia of 96, a number that diagnoses nothing by itself but sits close to where his last two admissions started and tracks his exam well enough to confirm what the exam already shows. His case sits at the center of a real, ongoing comparative-effectiveness question rather than a settled algorithm. The trial most often cited for adding rifaximin, Bass and colleagues' 2010 randomized study in the New England Journal of Medicine, enrolled patients with at least two overt HE episodes in the previous six months, and found that adding rifaximin to background lactulose — 91 percent of its patients were taking it — cut breakthrough episodes from 45.9 to 22.1 percent over six months. Two details of that design bear directly on him. The trial randomized patients who were in remission, and he is not in remission; he is mid-episode, in a bed, with a flap. And it never required documented lactulose failure to get in, which means the entry criterion he plainly meets is the episode count, not the failure his insurer wants proven. The open question isn't whether to switch him off lactulose. It's whether anyone can call six weeks of quiet under-dosing a failure of the drug at all.
On the wards, a recurrence that might not be one
He's had two overt HE admissions in four months and he's in his third episode right now. Bass and colleagues' entry bar in the 2010 NEJM trial was two episodes in six months; he has three in four. Whether he was on lactulose for one of those or all of them isn't what got patients into that trial, and rifaximin added on top of background lactulose cut breakthrough episodes by more than half. AASLD's own guidance recommends exactly this after a second episode. I don't think we need to relitigate whether his prior lactulose exposure was textbook-perfect before we act on data that clear. Every week we wait is a week he's at real risk of another confusional episode, and given what he's already told us about missed shifts, another admission costs him something beyond the medical picture.
I'm not disputing the trial, and you're right that its entry bar was episode count. That's exactly my problem with using it here. Clearing a trial's enrollment criteria isn't the same as having the clinical problem the drug is meant to solve. He told us himself — six weeks of under-dosing because the diarrhea has been unmanageable at work. If his ammonia climbed and his mental status declined because he wasn't reliably getting therapeutic lactulose exposure, that's an adherence barrier, not a demonstrated ceiling on what lactulose can do for him. And rifaximin gets expensive fast: his plan specifically requires documented lactulose failure before it'll approve it, which is a bar neither the trial nor his chart has actually cleared.
And "whether he was on lactulose isn't what got patients into that trial" cuts both ways. That trial also didn't enroll anyone mid-episode, and the 91 percent taking lactulose were taking a dose they could actually tolerate. He is neither of those things this morning.
You're both arguing about which drug, and I don't think that's actually the fork in the road. Bass's own trial added rifaximin to background lactulose rather than swapping one for the other, so this was never really an either-or question to begin with. Start both today. That sidesteps the "has he truly failed lactulose" argument without ignoring it, because he stays on lactulose either way. But if we send him home on the same regimen that produced six weeks of quiet non-adherence, we'll be back here with the same problem in another two months regardless of which drug label is on the chart. Split his lactulose dose smaller and more frequent instead of the standard TID, get him a concrete plan for what to do on days the diarrhea is worse before his shift, and loop in his employer's benefits office about what accommodation exists for a medical condition, since that's the actual barrier keeping him from the drug that already has the strongest evidence behind it.
Agreed: rifaximin started today alongside a restructured lactulose regimen — smaller, more frequent doses rather than the standard three-times-daily schedule that had been driving his non-adherence — with formal documentation of his two prior overt HE admissions submitted to support insurance approval.
Not agreed: whether his employer's accommodations process is something the medical team should actively help him navigate or something to hand off entirely to social work. The hospitalist raised it as directly relevant to whether this regimen actually holds once he's back at work; the hepatologist's view was that it's a real concern but outside what a hepatology consult can meaningfully act on beyond the referral already placed. Left for the primary team and social work to sort out before discharge, not resolved at the bedside.