Eight Months Later, a New Fever and an Old Culture
A single patient, hours into a second septic-appearing febrile UTI. The disagreement isn't about whether he needs broad coverage tonight — it's about how much weight an eight-month-old resistant culture should carry against a fresh blood culture that hasn't resulted yet.
T.M., a 68-year-old man who spent thirty years driving long-haul routes, has spent his retirement trying to reset a sleep schedule that never fully adjusted — so when the shaking chills woke his wife at three in the morning, he was already awake enough to know something was wrong before the fever hit 103.1. Eight months ago he was hospitalized for a febrile UTI that grew an ESBL-producing E. coli, treated then with meropenem after an initial cephalosporin failed on susceptibility testing; he recovered fully and nothing since has suggested a structural or functional reason for the recurrence beyond his age and an enlarged prostate he has managed conservatively. Tonight's presentation looks similar but sicker: flank pain, a blood pressure of 96/58 that has not responded to the first liter of fluid, and a lactate of 2.9. He has not been catheterized, hospitalized, or on antibiotics in the eight months between episodes, which matters directly to what happens next — a resistant organism eight months in the past is a real prior, not a permanent label, and the question the team is actually arguing is how much weight that one prior culture should carry against a fresh blood culture that has not yet resulted.
The MERINO trial (Harris et al., JAMA 2018) is the reason piperacillin-tazobactam is not simply reached for as the obvious carbapenem-sparing choice: in bacteremic patients with ceftriaxone-resistant E. coli or Klebsiella, it was stopped early after 30-day mortality ran three times higher on piperacillin-tazobactam than meropenem — 12.3% versus 3.7%, a number needed to harm under twelve. Whether that finding transfers cleanly onto a febrile UTI not yet proven bacteremic is the question, and T.M.'s own interval history is what answers most of it: MERINO's patients had a resistant organism in hand at randomization, while his has had eight months, no catheter, no hospitalization, and no antibiotic exposure to persist through. The trial describes a resistance state; his chart describes a resistance history, and those are not the same variable.
In the emergency department, before cultures are back
Meropenem, now. He has an ESBL organism on record, a lactate climbing toward 3, and a blood pressure that a full liter of fluid hasn't fixed. If this is bacteremic again tonight, undertreating it is the harder mistake to walk back than one dose of a broad-spectrum drug.
I want to name exactly what MERINO showed before we extrapolate it further than it goes. Harris et al. randomized patients who were already bacteremic with a confirmed ceftriaxone-resistant E. coli or Klebsiella — piperacillin-tazobactam lost badly there, 12.3% versus 3.7% mortality at 30 days. T.M. doesn't have a confirmed resistant organism tonight; he has one from eight months ago, with a clean interval since. We now have cefepime-zidebactam, approved this May for complicated UTI including pyelonephritis on the strength of ENHANCE-1, which randomized against meropenem itself and met its composite cure endpoint — and zidebactam isn't only a beta-lactamase inhibitor, it binds PBP2 directly, which is where the ESBL activity actually comes from. It's a genuine option that didn't exist for his admission eight months ago.
Defaulting to meropenem because of a culture that old is treating a resistance pattern as a permanent label rather than a snapshot, and that habit is exactly what drives the carbapenem-resistance problem MERINO was trying to help us avoid in the first place.
Both of you are arguing about which drug is right for the organism, and I think tonight the organism isn't the fastest-moving variable — his physiology is. A lactate that hasn't responded to fluids and a pressure staying soft after a liter is enough on its own to justify covering broadly for the next 48 hours, independent of how old that ESBL culture is.
I'm not dismissing the population-mismatch point — it's a real reason to de-escalate fast once cultures result, not a reason to start narrow tonight while he's still unstable.
Agreed: meropenem started empirically given the hemodynamic picture, blood and urine cultures already sent, reassessment at 48 hours or sooner if susceptibilities return first.
Not agreed, and carried forward explicitly:
The stewardship pharmacist's plan runs as originally proposed — de-escalate off meropenem as soon as he's hemodynamically stable.
The eight-month-old ESBL culture turns out to have been exactly the snapshot the pharmacist argued it was, and the critical care physician's caution — reasonable at the time — will have cost one avoidable carbapenem exposure, a cost everyone at the bedside accepted as the price of treating tonight's instability first.