Eight Days or Fifteen: Duration of Therapy for Pseudomonas Ventilator-Associated Pneumonia
A single patient, improving on day six of therapy for Pseudomonas ventilator-associated pneumonia. The disagreement is whether to trust the trial that shortened this diagnosis's standard duration, or the same trial's own subgroup finding that argues against applying it to her specific organism.
Georgia P., a 73-year-old woman, ran an alterations shop out of her home for nearly forty years, taking in wedding dresses and prom gowns from three generations of the same families in her town, and her daughter still brings photographs of the shop's old treadle machine to prop against the bed rail during visits, hoping something familiar will reach her. She was admitted three weeks ago after a large ischemic stroke left her with significant dysphagia, and despite careful swallow precautions she aspirated during an early attempt at oral intake, leading to a prolonged ventilator course complicated eight days ago by a new fever, purulent secretions, and a chest X-ray infiltrate consistent with ventilator-associated pneumonia. Bronchoalveolar lavage grew Pseudomonas aeruginosa, susceptible to the cefepime she was started on empirically, and she has been on that regimen since.
Six days into therapy, she looks meaningfully better: her secretions have thinned and decreased, her oxygen requirement has come down, and her white count is trending toward normal. That improvement is exactly what would ordinarily support stopping at eight total days rather than continuing further — the PneumA trial (Chastre et al., JAMA, 2003) established that an 8-day course of therapy for ventilator-associated pneumonia was non-inferior to 15 days for the trial population overall, with fewer resistant organisms emerging in the shorter-course group. What complicates applying that headline result to Georgia specifically is that PneumA stratified its randomization by causative organism, and in the non-fermenting gram-negative subgroup — Pseudomonas foremost among them, which is precisely what grew from her own lavage — pulmonary infection recurred in 40.6 percent of the 8-day arm against 25.4 percent of the 15-day arm. The qualifier that usually gets dropped is that the formal interaction test for recurrence was not significant, at P equal to .16, so the subgroup is a flag rather than a finding. The trial that established the short-course standard is also the trial that, in its own data, flagged the one pathogen category where that standard may not hold — and nearly twenty years later that flag still hasn't been resolved.
In the ICU, day six of therapy
Plan to stop at 8 days, per the standard she'll reach in two more days. The PneumA trial established 8 days as non-inferior to 15 for ventilator-associated pneumonia generally, with the added benefit of fewer resistant organisms emerging in the shorter-course arm — real advantages in a patient who's already spent three weeks in this ICU accumulating exposure to broad-spectrum agents. She's improving clinically on every measure that matters; extending exposure past the point the evidence supports adds cost without a demonstrated benefit for the average patient in her situation.
She isn't the average patient the headline result describes — she's growing Pseudomonas, and PneumA stratified randomization by organism precisely because that was expected to matter. Recurrence in that subgroup ran 40.6 against 25.4 percent. I'll concede the interaction test didn't reach significance, so this isn't settled by PneumA alone. But it isn't only PneumA: Bouglé and colleagues ran the dedicated randomized trial in Pseudomonas VAP and published it in Intensive Care Medicine in 2022, and although slow recruitment stopped it early and left it underpowered, the short-duration arm again showed more recurrence. Two attempts, same direction, neither conclusive. I'd extend her course closer to 15 days.
I don't think we have to choose between the headline result and the subgroup signal — both are real, and a stratified subgroup that has now pointed the same way in two separate randomized attempts isn't something I'd wave off, even with neither one reaching significance on its own.
But reflexively reverting to the pre-PneumA default of 15 days ignores how well she's actually responding clinically, which the subgroup data doesn't override either. Extend to day 10 or 11, and repeat inflammatory markers and clinical exam at day 8 to decide whether to stop then or continue the extra few days — let her own trajectory at the point of decision settle it, rather than pre-committing to either endpoint today.
Agreed same day: continue cefepime past day 8 with a formal reassessment at that point to decide between stopping and extending to roughly day 10–11.
Not agreed: how much weight a stratified subgroup signal that has twice failed to reach significance should carry against her own strong clinical trajectory going forward — the intensivist remains more inclined to stop at day 8 if the reassessment looks as good as today's trend suggests it will; the infectious disease physician would extend regardless of the day-8 numbers, given the specific organism involved.