Long-term outcomes after heart valve surgery in Brazil: a nationwide cohort of 63,426 patients with longitudinal multidatabase linkage and socioeconomic profiling

Background Heart valve disease represents a major health burden in Brazil, where rheumatic, degenerative, and congenital etiologies coexist amid pronounced socioeconomic inequality. We developed a high-throughput framework integrating 8 billion SUS (Brazilian Public Health System) records to reconstruct complete patient trajectories and characterize long-term outcomes following valve surgery. Methods Five SUS databases (SIA, SIH, SIM, RHC, IVS Atlas) were unified into continuous individual-level longitudinal records. Each patient’s clinical history was reconstructed from birth to death or censoring by linking the outpatient system (SIA) to inpatient records (SIH) through a hierarchical deterministic algorithm, attaching each hospitalization to the patient’s existing trajectory. Mortality was ascertained via a SIM or SIH. A proof-of-life algorithm corrected right-censoring bias by confirming survival whenever post-cutoff activity was recorded. The analysis included patients who underwent at least one valve surgery between 2018 and 2024. Results This analysis encompassed 63,426 patients, with a mean age at surgery of 56.9 ± 15.2 years and a mean follow-up duration of 253.9 ± 691.5 days. Analysis of patient flow between centers showed that São Paulo was the largest receiver, and in 2020, the number of surgeries decreased by 34%, possibly due to the COVID-19 pandemic. Overall in-hospital mortality was 8.6%, with the highest rates in congenital cases (11.3%), and five-year mortality reached 9.1%. Reoperation had nearly double the mortality compared to primary surgery (17.9% vs 9.3%), with each additional reoperation associated with increased odds of death (OR 1.46; 95% CI 1.02–2.11; p=0.039), whereas longer inter-surgical intervals were protective (HR 0.73 per year; 95% CI 0.70–0.76; p<0.001). Higher socioeconomic vulnerability was independently associated with higher mortality (HR 1.33; 95% CI 1.17–1.52; p<0.001), and rheumatic disease demonstrated a higher risk of death compared to degenerative disease (HR 3.23; 95% CI 2.99–3.5; p<0.001). The most frequent ambulatory drug was Atorvastatin, and the main ambulatory exams were echocardiogram (ECO), TP monitoring, and ECG. The main causes of death were Ebstein's anomaly (congenital), mitral stenosis (rheumatic), and aortic stenosis (degenerative). In the rheumatic group, the only valvopathy with a significant difference compared to mitral stenosis was mitral regurgitation (HR 0.85; p=0.02), and patients with rheumatic mitral stenosis undergoing TP monitoring had a lower risk of death compared to those who did not (HR 0.35; 95% CI 0.21–0.59; p<0.001). Conclusion Our study revealed that long-term outcomes after valve surgery are shaped by clinical complexity, reinterventions, and social inequality, offering a comprehensive, scalable framework to understand valve disease burden and guide surveillance in LMICs.