Acute Uncomplicated Diverticulitis: Treating a Habit, Not a Fever
A single low-risk, immunocompetent patient with a first episode of acute uncomplicated diverticulitis, testing whether routine antibiotics still earn their place against a decade of trial evidence favoring observation alone.
Walter B., a 58-year-old man who works as an accountant, spent the morning of tax-extension deadline day doubled over at his desk with left lower quadrant pain sharp enough that his office manager drove him to the emergency department herself rather than let him finish the afternoon. He has no prior history of diverticulitis, no immunosuppressing condition, no diabetes, and takes only a daily multivitamin. His CT shows sigmoid diverticulosis with localized fat stranding and a mildly thickened bowel wall — uncomplicated diverticulitis, no abscess, no perforation, no free air, no evidence of fistula — and his exam matches: localized tenderness without rebound or guarding, low-grade fever at 100.4, WBC mildly elevated at 12.4, and he is tolerating oral intake without difficulty.
The reflexive move for a first diverticulitis episode has long been a course of ciprofloxacin and metronidazole, and that's still what most patients arrive expecting once the CT confirms the diagnosis. But the actual randomized evidence behind that reflex has not held up well. AVOD, a Scandinavian multicenter trial, and DIABOLO, its Dutch successor, both randomized patients with CT-confirmed uncomplicated diverticulitis to antibiotics versus observation and found no difference in recovery time, complication rate, or recurrence between the two arms — a genuinely surprising result to clinicians trained to treat any confirmed intra-abdominal infection with antibiotics by default, and consistent enough across both trials that ACP, AGA, and ASCRS guidance has since shifted toward selective rather than routine antibiotic use in exactly this population. Walter's CT and exam match the trials' own inclusion criteria closely: localized findings, no abscess, no systemic sepsis physiology, tolerating oral intake. The genuine question isn't whether antibiotics ever help in diverticulitis — complicated disease is a different conversation entirely — it's whether his disease, read honestly against those criteria, is the uncomplicated kind the trials were actually testing.
Emergency department, reviewing the CT
I'd manage this with observation and analgesia, no antibiotics. His CT is a close match to what AVOD and DIABOLO actually enrolled — localized findings, no abscess, no perforation, no systemic sepsis physiology — and both trials found no difference in outcomes with or without antibiotics in exactly this presentation. This isn't an experimental position at this point; it's what ACP, AGA, and ASCRS guidance already recommends for low-risk, immunocompetent patients.
I know the trial data, and I still find myself reaching for the prescription pad on a first presentation like this. Part of it is genuine uncertainty at the bedside — a CT read as uncomplicated this morning can still evolve, and if I'm wrong about which direction this goes, undertreating feels like the costlier mistake to make. Part of it is just that most patients leave the ED expecting an antibiotic prescription for a confirmed infection, and 'no antibiotics needed' is a harder conversation than writing the script.
Both of those are real considerations, I'm not dismissing them — but neither one is actually an argument against the trial data itself. They're arguments about discomfort with applying it, which is a different problem.
The comfort question and the categorization question are worth separating cleanly. If Walter's CT genuinely doesn't meet the trials' inclusion criteria — if there were an abscess, or free air, or he looked systemically unwell — withholding antibiotics wouldn't be applying this evidence, it would be misapplying it to a population the trials never tested. But his imaging and exam do match: localized, no abscess, tolerating oral intake, stable vitals apart from a low-grade fever. Given that match, observation with close follow-up and an explicit return-precautions conversation is the evidence-concordant choice, not a shortcut. The discomfort with the conversation is real, but it's a communication problem to solve, not a reason to treat a patient who doesn't need it.
Agreed: discharged home on acetaminophen and a clear liquid-to-regular diet advance, explicit return precautions for fever above 101.5, worsening pain, or inability to tolerate oral intake, and a scheduled 48-hour phone follow-up rather than a silent wait-and-see. The primary care physician's discomfort with the conversation wasn't dismissed — the discharge instructions were written out explicitly enough that Walter left with a clear understanding of why no prescription was being sent, not just that one wasn't.