Possible Endocarditis With Negative Cultures and a Non-Diagnostic Echo: How Long to Wait?
He meets a real clinical picture for endocarditis, but neither of the two major diagnostic criteria has been confirmed yet — and his occupation points toward organisms that need their own specific tests before starting empiric antibiotics blindly.
J.M., a 45-year-old man who works as a large-animal veterinarian, has spent three weeks feeling progressively unwell in a way he initially blamed on a long calving season — low-grade fevers, drenching night sweats, a fatigue heavy enough that he's had to hand off farm calls to a colleague, something he says he's never done in twelve years of practice. He came in this week when his primary care physician noticed splinter hemorrhages under two fingernails and a new regurgitant murmur that wasn't on his exam a year ago.
Three sets of blood cultures, drawn appropriately before any antibiotics were given, have come back negative. His transthoracic echocardiogram was technically limited by body habitus and didn't clearly show a vegetation either way — not reassuring, not diagnostic, an equivocal study rather than a clean negative one. By modified Duke criteria, he currently sits in 'possible' rather than 'definite' infective endocarditis: a compatible clinical picture and peripheral stigmata, but neither of the major criteria has been met. It is worth being exact about what the second major criterion is, because it is broader than most students remember — it is evidence of endocardial involvement, which is satisfied either by a positive echocardiogram or by objectively documented new valvular regurgitation. His murmur is new, but a technically limited study has not actually documented the regurgitation, and a changed murmur on auscultation alone does not meet the criterion. His occupation adds a specific differential most blood-culture panels don't reliably catch: Coxiella burnetii, the organism causing Q fever, and Bartonella species are both classic causes of culture-negative endocarditis, and both are diagnosed by serology or specialized PCR rather than routine culture. Coxiella is the one his occupation points to directly — cattle, sheep, and goats are the reservoir, and calving season is the classic exposure. Bartonella's usual vectors are cats and body lice rather than livestock, which makes it the second question to ask rather than the first. The question in front of his team is whether starting broad empiric antibiotic therapy now, before those specific tests return, is the right call for a man who is clinically unwell and meets a real if not yet definite clinical picture — or whether doing so risks suppressing the growth or serologic response that would otherwise nail down the actual diagnosis.
On the wards, awaiting the TEE
His occupational exposure changes the differential meaningfully — Bartonella and Coxiella are both well-documented causes of culture-negative endocarditis in people with significant livestock contact, and neither shows up on standard blood culture panels. I want those serologies and the TEE back before committing to broad empiric antibiotics, because starting treatment now could suppress a serologic response or delay the specific diagnosis these organisms actually need to guide long-term therapy correctly.
If he were acutely unstable — hemodynamically compromised, in septic shock — I would not be arguing for any delay at all; empiric broad-spectrum therapy would start immediately regardless of what it costs diagnostically. The case for waiting rests specifically on his current clinical stability.
I agree he's stable enough to tolerate a short diagnostic window today, but I want that window to be genuinely short and defined, not open-ended. A three-week symptomatic course with embolic phenomena and a new murmur is a real, evolving picture, and the TEE result specifically could convert this from 'possible' to 'definite' IE quickly if it shows a vegetation the transthoracic study missed. I'd want the TEE prioritized for today or tomorrow, not scheduled routinely, precisely so this diagnostic window doesn't drift longer than it needs to.
That's fair, and I'm not proposing an indefinite wait either — my point is that even a same-day TEE and rapid serology turnaround is worth protecting, given what starting antibiotics blind could cost us diagnostically for what may be a straightforward Q fever or Bartonella endocarditis once properly identified.
The practical plan is to set a hard boundary on how long the diagnostic window stays open, not to leave 'stable enough to wait' undefined. I'd hold empiric antibiotics for 24 to 48 hours specifically to get the TEE result and allow serologies time to return, with explicit criteria for starting empiric therapy immediately regardless of pending results — new hemodynamic instability, a new embolic event, or evidence of heart failure. If the TEE confirms a vegetation or he meets one of those triggers, empiric coverage for the most likely organisms, including doxycycline for possible Coxiella, starts without waiting on serology confirmation — noting that if Q fever endocarditis is ultimately confirmed, doxycycline alone is not the treatment; it is paired with hydroxychloroquine for at least 18 months.
Agreed: TEE prioritized for same-day or next-day, serologies expedited, and a defined 24-to-48-hour window before starting empiric antibiotics — with explicit triggers that start treatment immediately regardless of pending results.
Empiric broad-spectrum therapy, including doxycycline for possible Coxiella given his exposure, starts immediately without waiting on serology confirmation.
Organism-specific therapy waits for serology results a short time longer, with the diagnostic window reassessed rather than extended indefinitely by default.
His farm and livestock exposure history will be documented explicitly in his chart, since it's the detail most likely to change which organisms actually get tested for.