Acute Cholangitis After Source Control: How Long to Keep the Antibiotics Going
Acute cholangitis, cleared endoscopically and clinically improving by hospital day three — and whether the antibiotic course a procedure note actually documents, not a fixed number of days, should decide when to stop.
Marcus J., 52, works nights as a hospital security guard — a job that means he's walked past the ED entrance hundreds of times before finally walking through it himself, two days into fever and a right-upper-quadrant pain sharp enough that he couldn't finish his rounds. Blood cultures grew E. coli; imaging and labs pointed to acute cholangitis from a stone lodged in his common bile duct, moderate severity by the Tokyo Guidelines criteria on two of their five counts — a white count of 14,800 and a temperature of 39.2°C — with no organ dysfunction to push him to Grade III. Jaundice and fever are what made the diagnosis; those two numbers are what set the grade, and only the second pair licenses the early drainage he got. ERCP on his first hospital day achieved stone extraction and sphincterotomy, with a completion cholangiogram read as clean. Empiric piperacillin-tazobactam started at admission has done its job: he's been afebrile since hospital day two, and today, day three, his blood culture sensitivities came back — pan-susceptible, including to ceftriaxone, a far narrower drug than the one he's currently on.
Two separate questions are actually in front of the team, not one. The first — whether to de-escalate from piperacillin-tazobactam to ceftriaxone now that the organism's susceptibilities are known — isn't seriously contested by anyone. The second, how long the antibiotic course should run in total, is where the real disagreement sits. Srinu's randomized trial in moderate-to-severe cholangitis found four days of antibiotics after successful drainage non-inferior to eight; the larger COBRA trial went further, stopping at a single day after adequate ERCP drainage with no loss of clinical cure. Both stand against a longer-standing practice, and the Tokyo Guidelines' own four-to-seven-day recommendation, of continuing well past the point of source control out of caution about relapse. What the ERCP note actually documents about how completely his duct was cleared turns out to matter more to this decision than either number does on its own. COBRA required fever resolution within twenty-four hours of ERCP before it would randomize anyone, and Marcus was afebrile by hospital day two — so the question is not whether he resembles those patients but whether his duct was cleared as completely as theirs were, which is a fact already written down rather than one anybody in the room has to estimate.
On rounds, hospital day three
De-escalate today — his cultures grew pan-susceptible E. coli, and continuing piperacillin-tazobactam once a narrower, equally effective agent is available is unnecessary broad-spectrum exposure. On duration, he's had adequate source control (stone extraction, sphincterotomy) and has been afebrile for over 48 hours; Srinu randomized exactly this situation — cholangitis with confirmed adequate drainage — and four days held up against eight. COBRA then stopped at one day and still matched four-to-seven on clinical cure, 95.1 percent against 93.7. What the longer arm bought in COBRA wasn't fewer relapses; it was twice the antibiotic-related adverse events, 16.6 percent against 8.3.
Doesn't dispute de-escalating the antibiotic itself — that part isn't in question. Duration is where he'd rather be conservative. Cholangitis relapses when biliary stasis persists undetected, and the cost of getting duration wrong isn't symmetric: two extra days of an oral, well-tolerated cephalosporin is a minor inconvenience, a bounce-back admission for recurrent cholangitis is not.
Calling the short-course data definitive skips over how new some of it still is — plenty of practicing clinicians, and some guidelines, haven't caught up to it yet, and "the trial found no benefit to extending" isn't the same as "extending causes harm."
There's a way to settle this that doesn't require picking a side on the general question. The short-course evidence both of them are half-arguing about was generated in patients with confirmed complete biliary clearance — COBRA excluded anyone whose drainage wasn't adequate, and Srinu did the same. Check today's actual ERCP note. If it documents complete stone extraction with no residual filling defect, the ID physician's four-to-five-day plan is exactly what that evidence supports and there's no real relapse risk it's leaving on the table. If the note hedges at all — a possible residual fragment, incomplete visualization of the distal duct — that's a different patient than the trial studied, and the hospitalist's longer, cautious course is the right call for him specifically, not as a general insurance policy.
The ERCP note was pulled during rounds and documented complete stone extraction with a clean completion cholangiogram, no residual filling defect. On that basis, the team agreed to the shorter four-to-five-day total course with the de-escalated ceftriaxone, discharge planned for tomorrow with a call-back instruction for any recurrent fever.
Not fully agreed: the hospitalist still wants a scheduled telehealth check-in at day 7 specifically to ask about fever, rather than relying only on a call-back-if-symptomatic instruction — a small residual disagreement about how much the shorter course still deserves active follow-up versus passive safety-netting.