Empiric Antibiotics for Endocarditis When the Cultures Stay Negative
A single patient whose clinical picture is convincing for endocarditis and whose blood cultures won't confirm it, with an occupational history that shifts the differential toward organisms routine coverage doesn't treat.
V.H., a 66-year-old woman, spent thirty-four years as a large-animal veterinarian before retiring, decades of direct hands-on contact with livestock during exactly the kind of exposures — assisting births, handling placental tissue — that her medical team would come to view as clinically significant rather than simply biographical. She presented with weeks of low-grade fevers, fatigue, and a new murmur her primary care physician caught on routine exam.
Echocardiography confirmed a vegetation on her native mitral valve, and her inflammatory markers are elevated, with a clinical picture that reads as endocarditis by any reasonable assessment. What hasn't followed the expected pattern is her blood cultures: three separate sets, drawn correctly with extended incubation, have all come back negative — culture-negative endocarditis, a genuine diagnostic problem rather than a reassuring result.
Standard practice for culture-negative endocarditis is broad empiric coverage targeting the organisms most commonly responsible — viridans streptococci, staphylococci, enterococci — while further testing continues. Her occupational history complicates that default, though, in a specific and clinically real way: decades of direct livestock exposure, especially around birthing, is the classic epidemiologic picture for Coxiella burnetii, the organism causing Q fever, and also raises meaningful suspicion for Bartonella species — neither of which standard empiric endocarditis coverage reliably treats, and both of which require entirely different confirmatory testing (specific serologies and PCR) that takes time to return. Her most recent direct livestock contact was roughly three months ago — not an incubation interval in the usual sense, since Q fever endocarditis is a chronic manifestation that surfaces months to years after the initial infection, which is exactly the interval in play here — a detail her infectious disease team flagged as specifically relevant once cultures came back negative.
Treating broadly while the diagnosis is still unsettled
I'd start standard broad empiric therapy now, covering the organisms most commonly responsible for endocarditis, given the clinical urgency of an active presentation. In parallel, I'd send specific serologic and PCR testing for Coxiella and Bartonella rather than waiting for those results before starting anything — the standard coverage buys time and treats the more statistically likely possibilities while the more specific testing is pending.
I'd add empiric doxycycline now rather than waiting for serology, specifically because of her occupational history. Standard empiric coverage doesn't reliably treat Coxiella or Bartonella if either turns out to be the actual organism, and her decades of direct livestock exposure meaningfully raises the pretest probability of exactly those organisms — high enough, in my view, that empiric coverage for them belongs in the initial regimen, not held until confirmatory testing returns.
I'd underline that her exposure history isn't background color here — it's specifically livestock, specifically involving birthing-related contact, which is the exact epidemiologic picture associated with Coxiella. That's a load-bearing clinical detail, not an incidental biographical fact, and I think it genuinely earns the broader empiric coverage being proposed.
Started on standard broad empiric endocarditis coverage plus empiric doxycycline, given her occupational exposure history. Coxiella and Bartonella serology and PCR sent, with the regimen to be narrowed once results return.
Therapy is narrowed to the organism-specific regimen, and standard coverage is discontinued.
Standard coverage continues as the primary regimen, and the doxycycline addition is reassessed once that specific concern has been reasonably excluded.
Nobody at the table knew which organism was actually responsible, and the plan didn't pretend otherwise — it treated broadly across the genuinely most plausible possibilities her history raised, rather than waiting for a diagnostic certainty that culture-negative disease may not offer quickly.