Selective Digestive Decontamination After ARDS: Does a Dutch Mortality Trial Apply Here?
A single patient, four days into mechanical ventilation for ARDS, in an ICU weighing whether a Dutch mortality trial for selective digestive decontamination applies to a unit with a very different resistance landscape — and to a patient whose own prior culture already carries a resistant organism.
J.K., a 54-year-old man, has spent seventeen years as the maintenance supervisor at a regional distribution center, the kind of job where the building runs invisibly when he's doing it right and loudly when he isn't — and lately he's been doing more of it than his shift technically requires, back at his desk within two weeks of his wife's death last year because, as he told a coworker, an empty kitchen at home was worse than a full inbox at work. Five days ago a cough he'd been ignoring turned into fever and real shortness of breath; a chest X-ray showed multilobar community-acquired pneumonia, and within thirty-six hours he had progressed to ARDS requiring intubation. He has type 2 diabetes on insulin and, buried in a discharge summary from a hospitalization eight months ago, a stool culture positive for ESBL-producing E. coli — asymptomatic colonization at the time, never treated, never re-tested since.
He is now on hospital day 4 of mechanical ventilation, and the ICU is weighing whether to start him, and the rest of the unit, on selective digestive decontamination — topical polymyxin, tobramycin, and amphotericin B applied to the oropharynx and stomach, plus a four-day course of intravenous cefotaxime, which is the part that makes it decontamination of the whole digestive tract rather than the mouth alone. De Smet and colleagues' cluster-randomized crossover trial across thirteen Dutch ICUs found that bundle reduced 28-day mortality in more than five thousand patients, an adjusted odds ratio of 0.83 against standard care. What that trial doesn't answer directly is whether the result travels: it was run in a health system with one of the lowest rates of MRSA, ESBL-producing organisms, and vancomycin-resistant enterococci in the developed world, and this unit's own most recent antibiogram shows real, non-trivial ESBL and VRE prevalence — the exact organism classes SDD's critics worry the regimen selects for over time. J.K.'s own chart adds a second, more specific problem the population-level argument doesn't capture: he is already colonized with an ESBL organism, which means SDD wouldn't be introducing a hypothetical resistance risk in him specifically, it would be adding antibiotic pressure directly onto flora that has already demonstrated it can adapt.
ICU rounds, hospital day 4
De Smet's trial randomized more than five thousand ICU patients by unit and found selective digestive decontamination reduced 28-day mortality — not just VAP rates, actual mortality, in a trial large enough and long enough that I don't think we get to wave it off. He's exactly the kind of patient that trial was built for: ventilated, ARDS, real risk of a secondary infection killing him before his lungs recover.
I'm not disputing the mortality number — it's real. But I'd be careful with the word replicated. The second big Dutch trial, Oostdijk's, compared SDD head-to-head against oropharyngeal decontamination rather than against standard care, so it never retested the question de Smet answered; and SuDDICU, run across nineteen Australian ICUs, didn't reach a statistically significant mortality reduction at all. But de Smet's units had some of the lowest baseline MRSA and ESBL prevalence in Europe. Our own antibiogram from this quarter shows meaningfully higher ESBL Enterobacterales and VRE rates than anything in that trial's setting. What that means practically: the organisms SDD is designed to suppress in the gut are already less controllable here than they were in the population where the mortality benefit was measured, and the resistance-selection cost the Dutch trials didn't have much room to show is exactly the cost a higher-baseline-resistance unit would pay first.
Whichever way the unit-wide policy question lands — and I don't think either of you is wrong about the general argument — J.K. himself already has an ESBL organism in his chart from eight months ago. Starting him on SDD wouldn't be exposing a clean patient to a hypothetical resistance risk; it would be adding topical antibiotic pressure directly onto flora we already know can adapt.
That's not a reason to rule out SDD for this whole ICU. It's a reason to hold it for him specifically until the unit actually resolves which way the general policy goes.
Agreed for J.K. specifically: hold selective digestive decontamination given his documented ESBL colonization, and continue the standard non-antimicrobial VAP-prevention bundle.
Not agreed, and explicitly left open rather than smoothed over: whether the ICU as a whole should adopt universal selective digestive decontamination. The intensivist points to de Smet's mortality data as too strong to ignore; the stewardship physician wants the unit's own resistance surveillance data formally reviewed by the pharmacy and therapeutics committee before any unit-wide change. The question was referred upward rather than decided at the bedside tonight.