Post-Acute Rhythm Diagnoses Rise After Infective Endocarditis
Patients who survived the first month after infective endocarditis still carried a measurable excess risk of new rhythm and conduction diagnoses over the next year. In a matched cohort drawn from the Protege database, the signal was not diffuse across time; it concentrated in the middle weeks after the acute infection, when lingering inflammation, valvular injury, periannular extension, and hemodynamic stress can still unsettle cardiac conduction. The pattern matters because it separates the immediate hospitalization phase from the period when outpatient surveillance is often less intense.
Among 6,810 patients with infective endocarditis who remained outcome-free through day 30, 383 went on to receive a composite diagnosis during days 31-365. In the control group, 1,150 of 37,463 bacteremia-matched patients had the same composite outcome over that interval. The adjusted hazard ratio for days 31-365 was 1.73, with the largest relative excess in days 31-90, when the hazard ratio reached 2.41. During days 91-365, the association persisted but weakened, with a hazard ratio of 1.36. The 1-year standardized risk was 6.98% after infective endocarditis versus 4.14% after bacteremia, an absolute difference of 2.84 percentage points.
The early burden was visible even before the post-acute analysis began. At the index date, 805 of 6,810 patients with infective endocarditis had a composite diagnosis, compared with 1,829 of 37,463 controls. Same-day diagnoses were handled separately, which helps keep the post-acute estimate focused on events that emerged after the first month rather than on arrhythmias already apparent during the index encounter.
Why the 31- to 90-Day Window Stands Out
The temporal shape of the results fits what clinicians see at the bedside. Infective endocarditis can injure the conduction system through local extension, abscess formation, valve destruction, embolic complications, and the systemic inflammatory response that follows severe infection. Those processes do not end when blood cultures clear or the admission closes. A vulnerable substrate can persist for weeks, and a previously silent conduction abnormality may become evident only after discharge, when activity increases, medications change, and the heart is asked to recover from a major infectious insult.
This is why the 31- to 90-day interval is clinically interesting. A hazard ratio of 2.41 in that period suggests that the excess is not simply a reflection of chronic comorbidity; it is temporally clustered in a window when healing, scar formation, residual inflammation, and delayed recognition are all plausible contributors. The later attenuation to 1.36 during days 91-365 suggests that the process may cool over time, even if some elevated risk remains.
The composite endpoint also captures more than atrial fibrillation. It included atrial fibrillation or flutter, supraventricular tachycardia, ventricular tachycardia or fibrillation, sick sinus syndrome, and atrioventricular block. Grouping these diagnoses together acknowledges that infective endocarditis can affect both atrial and ventricular electrical stability, as well as the conduction pathways that bridge the chambers. A single mechanism is unlikely to explain every diagnosis in the composite, but a shared infectious and structural substrate makes the cluster biologically credible.
Matched Controls and What the Comparison Can and Cannot Prove
The cohort design improved comparability by matching adults with a first infective endocarditis diagnosis to encounter-indexed bacteremia controls without endocarditis by sex and 10-year age strata, with up to a 5:1 ratio. Of 12,553 eligible patients with infective endocarditis, 9,599 were matched to 47,974 controls. After washout, 6,810 patients with infective endocarditis and 37,463 controls remained for the primary analysis. Models were adjusted for prespecified baseline factors, and the investigators also examined acute and annual intervals, landmarks, absolute risks, competing risk, overlap weighting, population controls, and a cataract control.
Those analytic choices strengthen internal consistency, but they do not turn an observational association into causation. Bacteremia is not a perfect stand-in for the clinical burden of infective endocarditis, and claims data cannot fully capture the severity of valve destruction, the extent of perivalvular invasion, or the exact rhythm-monitoring practices used after discharge. The preprint status also matters: the findings are new and have not yet undergone peer review.
Still, the consistency across the main analyses is hard to ignore. The Fine-Gray hazard ratio was 1.72, closely tracking the primary adjusted estimate. The 1-year absolute risk difference remained modest, which argues against overreading the findings as proof that every patient needs prolonged rhythm monitoring. The signal is real enough to warrant attention, but not so large that it automatically translates into blanket surveillance.
What Clinicians May Do With This Signal
The practical implication is narrower than universal monitoring and more focused than routine reassurance. Patients recovering from infective endocarditis may deserve heightened attention during the second and third months after the acute episode, especially if they had valve involvement, new murmurs, persistent symptoms, or any hint of conduction delay during the index admission. A targeted follow-up strategy during days 31-90 could include symptom review, pulse checks, electrocardiography when clinically indicated, and a low threshold to evaluate palpitations, presyncope, or unexplained fatigue.
The study does not justify a one-size-fits-all monitoring mandate. The absolute 1-year difference was 2.84 percentage points, which is meaningful at the population level but still modest in individual terms. What it does justify is a more deliberate look at rhythm surveillance after discharge, especially once the acute infection has passed and the patient is moving into the phase where delayed electrical complications are most likely to surface.
A focused prospective study could answer the question this preprint raises: whether selective rhythm surveillance in the 31- to 90-day window reduces missed diagnoses, emergency visits, or later complications. Until then, the clearest message is temporal, not sweeping—post-endocarditis rhythm risk is front-loaded after the first month, then gradually fades, rather than appearing as a uniform hazard spread evenly across the year.