Up to six weeks of antibiotics for left-sided infective endocarditis rests largely on expert consensus rather than trial data. POET II tested whether clinical response could be used to stop earlier. This international open-label trial randomised 508 clinically stable adults with endocarditis caused by Staphylococcus aureus, Enterococcus faecalis or streptococci. Everyone had already received at least the prespecified 2 to 4 weeks of therapy and met stabilisation criteria before randomisation. The tailored group then stopped; the standard group completed 4 to 6 weeks in total.
The efficacy endpoint was days alive without antibiotics for endocarditis or bacteraemia over six months. Tailored therapy gave a median 183 days against 169, a difference of 13 days (95% CI 12 to 13, p<0.001). The safety composite of death, unplanned cardiac surgery or symptomatic embolism occurred in 21 patients (8.2%) against 27 (10.7%), a difference of -2.4 percentage points (95% CI -7.7 to 2.7), meeting the prespecified non-inferiority margin of 7.5 points.
The number that decides how you use this is the relapse rate. Relapse of bacteraemia or endocarditis occurred in 13 patients (5.1%) on tailored therapy against 4 (1.6%) on standard therapy, p=0.04. That is roughly a tripling, on a secondary endpoint, in a trial not powered for it. Stopping early does not appear to kill people or send them to theatre, but it does let more infections come back — and a relapse in endocarditis is a serious event even when it is survived.
So the practice change is conditional. In a stabilised patient who has completed the front-loaded 2 to 4 weeks, with a susceptible organism and a clean clinical response, stopping is now a defensible choice rather than a deviation. It becomes indefensible without follow-up attached. In Indian practice, where prolonged inpatient intravenous therapy is expensive and outpatient parenteral services are patchy, this trial is genuinely useful — but it is only useful if the patient can be seen again and cultured, and that has to be arranged before the cannula comes out.
- Only consider stopping after the full 2 to 4 weeks of front-loaded therapy and documented clinical stabilisation.
- Restrict this to the organisms studied — S aureus, E faecalis and streptococci — with confirmed susceptibility.
- Quote the relapse figure honestly to the patient: roughly 5% against 1.6%.
- Arrange follow-up review and blood cultures before discharge, not afterwards.
- Do not extend this to prosthetic valve endocarditis, uncontrolled infection or unresolved complications.
The statistics, in plain English
Two different questions were asked with two different statistical standards. The efficacy question was superiority, and 13 more antibiotic-free days with an interval of 12 to 13 is a precise result — though it is a measure of exposure avoided, not of the patient being better. The safety question was non-inferiority with a margin of 7.5 percentage points, meaning the shorter strategy had to be shown not to increase serious events by more than that much; it passed, but the interval still reaches 2.7 points in the direction of harm. Relapse was a secondary endpoint, so its p value of 0.04 was not adjusted for multiple comparisons — treat it as a real warning rather than a precisely measured risk.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for infectious diseases, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free