Risk factors and a nomogram for bovine jugular vein conduit failure after right ventricular outflow tract reconstruction: a 10-year single-center cohort.

Background: Bovine jugular vein conduit (BJVC) is widely used for right ventricular outflow tract (RVOT) reconstruction, yet long-term durability varies and individualized risk tools remain limited. Methods: We conducted a single-center retrospective cohort of consecutive children undergoing primary BJVC implantation between 2011 and 2020. The primary endpoint was BJVC failure, defined as surgical or catheter-based reintervention for conduit dysfunction or infective endocarditis. Freedom from failure was summarized at 1, 3, 5, and 7 years. Candidate predictors comprised demographics, pre-operative echocardiography/laboratory data, operative metrics (conduit internal diameter, cardiopulmonary bypass and cross-clamp times), and pre-discharge residual RVOT gradient. Missing data (0%-11%) were handled using multiple imputation by chained equations. Predictors selected by LASSO were entered into multivariable Cox regression. Model performance was evaluated for discrimination (Harrell's C; time-dependent AUC) and calibration, with internal validity assessed by bootstrap optimism correction. A nomogram was constructed to provide individualized 1-, 3-, 5-, and 7-year estimates of freedom from failure. Results: Seventy-eight patients were included (median follow-up, 7.7 years); 29 conduit failures occurred (37.2%). All conduit reinterventions were surgical redo procedures; no patient underwent catheter-based balloon dilation of the BJVC prior to surgical reintervention. Independent risk factors were pre-discharge residual gradient ≥20 mmHg (HR 18.67; 95% CI 7.43-46.94), longer cardiopulmonary bypass time (per 10-minute increase HR 1.28; 95% CI 1.15-1.43), male sex (HR 2.25; 95% CI 1.11-4.58), age ≤1 year (HR 1.76; 95% CI 1.09-2.86), and conduit diameter ≤14 mm (HR 1.68; 95% CI 1.04-2.74). The model demonstrated good discrimination (Harrell's C-index 0.82) and acceptable calibration; bootstrap internal validation yielded similar performance, with time-dependent AUCs of 0.888, 0.850, 0.900 and 0.897 at 1, 3, 5, and 7 years, respectively. Freedom from failure at 1, 3, 5, and 7 years was 97.4%, 87.2%, 75.6%, and 62.8%, respectively. Conclusions: Using routinely available peri-operative variables, we developed an interpretable nomogram to estimate the risk of BJVC failure and to inform individualized surveillance and intervention planning. The model showed good internal performance but was derived in a single-center cohort with a modest number of events; prospective multicenter validation and, if necessary, recalibration are required before routine clinical implementation.