A Fourth Recurrence, and One Fewer Option
The one agent aimed directly at his toxin burden left the US market in 2025. What is left has to be chosen against a heart that has only recently stopped decompensating.
Walter K., a 74-year-old retired postal supervisor, has been hospitalized twice this year for heart failure with reduced ejection fraction, most recently three months ago, and is now on his fourth episode of Clostridioides difficile infection in eight months — each one arriving while he is still recovering strength from the last. He lives alone since his wife passed last year, managed mostly by a home health aide who visits three times a week, and each recurrence has meant another hospitalization, another course of antibiotics, and another stretch of deconditioning that has made him noticeably frailer than he was at the start of the year. His current episode is being treated with oral vancomycin, and the recurrence-prevention question has reached the team in an awkward form: bezlotoxumab, the toxin-neutralizing monoclonal antibody whose trial population most closely resembles him, was discontinued by its sole manufacturer in January 2025 and has not been available in the United States since.
His most recent echocardiogram, six weeks ago, showed an ejection fraction of 30%, improved from 22% at his worst hospitalization but still reduced, and he has had no heart failure exacerbation or hospitalization since that admission — on optimized guideline-directed therapy that his cardiology team has kept stable rather than actively titrating during this acute infection. That 30% matters to a decision he can no longer actually be offered: in MODIFY I/II, serious heart-failure events occurred in 12.7% of bezlotoxumab recipients carrying a pre-existing heart-failure diagnosis against 4.8% on placebo, a signal that produced a labeled warning describing his own heart almost exactly. Had the drug still been marketed, his ejection fraction would have placed him squarely inside the subgroup that warning was written for — so the team's task now is not whether to accept that risk but what the agents still available actually buy him instead.
What the MODIFY trials (Wilcox et al., 2017) showed was a monoclonal antibody that neutralized circulating toxin B and lowered recurrence meaningfully across the full trial population, with the heart-failure signal isolated to the subgroup carrying that diagnosis at baseline rather than spread evenly across everyone who received it. None of that is retrievable for him now. What remains are the agents still on the market — fidaxomicin, which he has never been given, and the licensed live biotherapeutic products for recurrence prevention — and whether a man on his fourth episode in eight months, measurably frailer after each one, is better served by another narrow-spectrum antibiotic course or by an intervention aimed at the flora that keeps failing to reconstitute between them.
Infectious disease clinic, fourth episode this year
Before we reach for anything elaborate, I want to name what has never been tried: he has had four episodes and has never once received fidaxomicin. Louie et al. found lower recurrence with fidaxomicin than with vancomycin in the head-to-head trial, and every one of his courses has been vancomycin. A fifth recurrence isn’t a minor setback for him — it’s another hospitalization and another step down in function he may not get back — but the first move should be the drug with a recurrence advantage his chart shows he was simply never given.
I don’t dispute that fidaxomicin is the untried option, and I’d have no objection to starting it. What I won’t treat as an open question is what four failures actually tell us. Each course suppressed the organism and each one worked, right up until it stopped — that is a flora-reconstitution failure, not an inadequate-killing failure, and a narrower antibiotic addresses the second problem, not the first. The rebuttal to “try the better antibiotic first” is that we have now run that experiment four times, and van Nood et al. stopped their donor-stool trial early because the difference against continued vancomycin was too large to keep randomizing. He hasn’t run out of antibiotic options; his antibiotic options keep working and keep not lasting.
Neither of you has said anything I’d overrule, but you are both arguing about the agent and skipping the route, and for this man the route is where the cardiac risk actually sits. His EF is 30%, improved from 22%, no exacerbation in three months on stable GDMT — compensated, genuinely, but without margin to spare. Capsule delivery I have no concern about. Colonoscopic instillation means bowel preparation and sedation in a man whose last decompensation was this year, and that is a specific exposure, not a procedural footnote. The withdrawn antibody is exactly why I keep pressing: its heart-failure signal only became visible once somebody looked at the subgroup he belongs to. I would rather ask that question in advance this time. Choose the mechanism between you; let me choose how it is delivered.
Agreed: fidaxomicin this admission rather than a fourth vancomycin course, with the vancomycin stopped rather than overlapped, and a live biotherapeutic held in reserve — by capsule rather than colonoscopy if it is needed — with a weight and symptom check at two weeks.
Not agreed, and carried forward rather than smoothed over: whether fidaxomicin first is a reasonable sequence or a fifth iteration of an approach that has already failed him four times. The pharmacologist would have gone to flora restoration today and accepted the capsule route’s slower onset; the infectious disease physician would not skip a guideline-preferred drug the patient has demonstrably never been given. The cardiologist declined to arbitrate between them, having secured the one thing he came for, which was the route.