Oral Step-Down Antibiotics Against a Residual Abdominal Collection
A single patient, four days after percutaneous drainage of a periappendiceal abscess. The disagreement is whether a still-visible residual collection means source control isn't complete, or whether the route of antibiotic delivery never depended on that question in the first place.
Teresa A., a 39-year-old woman, has run the same neighborhood bakery for eleven years, opening at 5 a.m. most mornings herself, and joked with the surgical team that the only reason she noticed her abdominal pain at all was that it kept her from lifting the flour sacks she normally hauls without thinking. What she'd dismissed for two days as a bad stomach turned out, once she finally came to the emergency department, to be perforated appendicitis with a contained periappendiceal abscess; interventional radiology placed a percutaneous drain rather than proceeding to immediate surgery, given the degree of surrounding inflammation, and she was started on IV piperacillin-tazobactam for combined aerobic and anaerobic coverage.
Four days later, her clinical trajectory looks good by every measure that usually matters: she has been afebrile for thirty-six hours, her white count has normalized, she is eating a regular diet without nausea, and she is asking, reasonably, when she can go home. The complication is a repeat CT ordered before any discharge conversation, which shows the drain output has slowed to near nothing but a small residual fluid collection, roughly 2cm, is still visible around the drain tip — not large enough to prompt further intervention on its own, but present in a way that makes ‘source control achieved’ a genuinely debatable read rather than a settled one. The STOP-IT trial (Sawyer et al., New England Journal of Medicine, 2015) found that once source control is adequate, extending antibiotic therapy beyond a fixed four-day course after source control provides no measurable benefit over stopping early — a result the team keeps returning to, and one whose applicability here turns entirely on whether a persistent 2cm collection on imaging counts as the adequate source control the trial's own population was defined by, or as the kind of incomplete control the trial's conclusions were never meant to cover. The fluid aspirated through her drain on placement, and the culture drawn from it, grew mixed enteric flora typical of a perforated appendix, without any organism resistant to the amoxicillin-clavulanate now being considered for step-down — reassuring on the microbiology, even as the anatomy stays genuinely unsettled.
On the surgical ward, before the discharge conversation
Convert her to oral amoxicillin-clavulanate and plan for discharge. She meets every standard IV-to-oral conversion criterion — afebrile over 24 hours, tolerating a regular diet, clinically improving, white count normalized — and the STOP-IT trial found that once source control is adequate, extending IV therapy beyond a short fixed course adds nothing measurable. Her drain is nearly dry. Holding her longer as an inpatient exposes her to real hospital-acquired risk without a demonstrated benefit the trial evidence actually supports.
STOP-IT's whole conclusion rests on source control being adequate at enrollment — that's the population the trial defined itself by. A 2cm collection still visible on repeat imaging, drain output or not, is a real question about whether we're actually there yet. I've watched a ‘nearly dry’ drain and a small residual collection turn into a readmission for a recollected abscess more than once. I'd rather keep her IV and repeat imaging in another 48 hours to confirm the collection is actually resolving before calling source control complete.
I think you're both arguing about duration when part of this is actually a route question, and those are separable. Oral amoxicillin-clavulanate, properly dosed, achieves tissue and serum concentrations comparable to the IV formulation once GI absorption is confirmed adequate — which her tolerating a full diet already demonstrates.
So convert her to oral today; that part doesn't need to wait on the collection question. But don't discharge her, or set a fixed stop date, until the surgeon's repeat imaging call is answered — let the oral antibiotics run as an inpatient for the next 48 hours alongside that reassessment, rather than treating ‘convert to oral’ and ‘go home’ as the same decision.
Agreed same day: converted to oral amoxicillin-clavulanate, with discharge held pending a repeat CT in 48 hours rather than tied to a fixed antibiotic-course calendar date.
Not agreed: whether the STOP-IT trial's four-day-post-source-control framework should apply once route conversion happens, or whether a persistent visible collection resets that clock entirely regardless of route — the surgeon wants total duration counted from confirmed collection resolution; the infectious disease physician still expects the original short-course logic to hold once the 48-hour imaging comes back reassuring.