Long-term outcomes of truncus arteriosus repair: A modulated renewal competing risks analysis

Abstract

Objective: In this study, we sought to identify independent risk factors for mortality and reintervention after early surgical correction of truncus arteriosus using a novel statistical method.

Methods: Patients undergoing neonatal/infant truncus arteriosus repair between January 1984 and December 2018 were reviewed retrospectively. An innovative statistical strategy was applied integrating competing risks analysis with modulated renewal for time-to-event modeling.

Results: A total of 204 patients were included in the study. Mortality occurred in 32 patients (15%). Smaller right ventricle to pulmonary artery conduit size and truncal valve insufficiency at birth were significantly associated with overall mortality (right ventricle to pulmonary artery conduit size: hazard ratio, 1.34; 95% confidence interval, 1.08-1.66, P = .008; truncal valve insufficiency: hazard ratio, 2.5; 95% confidence interval, 1.13-5.53, P = .024). truncal valve insufficiency at birth, truncal valve intervention at index repair, and number of cusps (4 vs 3) were associated with truncal valve reoperations (truncal valve insufficiency: hazard ratio, 2.38; 95%, confidence interval, 1.13-5.01, P = .02; cusp number: hazard ratio, 6.62; 95% confidence interval, 2.54-17.3, P < .001). Right ventricle to pulmonary artery conduit size 11 mm or less was associated with a higher risk of early catheter-based reintervention (hazard ratio, 1.54; 95% confidence interval, 1.04-2.28, P = .03) and reoperation (hazard ratio, 1.96; 95% confidence interval, 1.33-2.89, P = .001) on the right ventricle to pulmonary artery conduit.

Conclusions: Smaller right ventricle to pulmonary artery conduit size and truncal valve insufficiency at birth were associated with overall mortality after truncus arteriosus repair. Quadricuspid truncal valve, the presence of truncal valve insufficiency at the time of diagnosis, and truncal valve intervention at index repair were associated with an increased risk of reoperation. The size of the right ventricle to pulmonary artery conduit at index surgery is the single most important factor for early reoperation and catheter-based reintervention on the conduit.

Keywords: competing risks; long-term outcomes; modulated renewal; risk factors; statistical analysis; truncus arteriosus.

FIGURE E1.

Consort diagram of the study (early mortality, N = 11; late mortality, N = 21; reinterventions, N = 149).

FIGURE E2.

Number of events in the cohort based on the number of repeated events to accompany the cumulative incidence function. TV, Truncal valve; RV-PA, right ventricle to pulmonary artery.

FIGURE E3.

Visual assessment of partial Cox-Snell residuals for reoperation on TV (A), catheter-based reintervention on RV-PA conduit (B), and reoperation on RV-PA conduit (C). RV-PA, Right ventricle to pulmonary artery.

FIGURE 1.

Kaplan–Meier 1-, 5-, 10-, and 20-year survivals. Shaded area depicts 95% CIs. The numbers at risk in each group are shown at the bottom. The curve has been truncated for time points at which patients at risk were less than 10. CI, Confidence interval.

FIGURE 2.

Cumulative incidence for reoperation on TVover 30 years of follow-up according to number of TV cusps (A) and TV insufficiency (B). Forest plots indicating adjusted HR and 95% CI of the statistically significant risk factors identified by multivariable modulated renewal competing risks regression analysis for reoperation on TV (C). Covariates included in the model for reoperation on RV-PA conduit were TV insufficiency, TV stenosis, and number of TV cusps. HR, Hazard ratio; CI, confidence interval. *P<.05 (confidence limits for cumulative incidence estimates and number of patients at risk are shown in Table E5).

FIGURE 3.

A novel analysis for risks of mortality and reintervention after TA repair. RV-PA, Right ventricle to pulmonary artery; TV, truncal valve.

FIGURE 4.

Cumulative incidence for catheter-based reintervention on RV-PA conduit over 35 years of follow-up according to RV-PA conduit size (A). Cumulative incidence for reoperation on RV-PA conduit over 30 years of follow-up according to RV-PA conduit size (B) and year of surgery (C). Forest plots indicating adjusted HR and 95% CI of the statistically significant risk factors identified by multivariable modulated renewal competing risks regression analysis for catheter-based reintervention (D) and reoperation on RV-PA conduit (E). Covariates included in the model for reoperation on RV-PA conduit were RV-PA conduit size and year of surgery. Only RV-PA conduit size was included in the model for catheter-based reintervention on RV-PA conduit. RV-PA, Right ventricle-pulmonary artery; HR, hazard ratio; CI, confidence interval. *P<.05 (confidence limits for Cumulative Incidence Estimates and number of patients at risk are shown in Table E5).

VIDEO 1.

Dr Meena Nathan, MD, FRCS, MPH, presents the rationale and importance of the study. Dr David Zurakowski, MS, PhD, describes the innovation of the statistical model used. Video available at: https://www.jtcvs.org/article/S0022-5223(21)00252-X/fulltext.

Webcast

You can watch a Webcast of this AATS meeting presentation by going to: https://aats.blob.core.windows.net/media/20AM/Presentations/Long-Term%20Outcomes%20of%20Truncus%20Arteri.mp4.