An Evaluation of the Learning Curve in Pulmonary Endarterectomy Using Propensity Score Matching

Abstract

Abstract Objective Pulmonary endarterectomy (PEA) is the only causative, but demanding treatment of choice for chronic thromboembolic pulmonary hypertension (CTEPH). We analyzed our results with PEA to evaluate the learning curve. Methods Consecutive 499 patients who underwent PEA between 1995 and 2014 were divided into two groups according to the temporal order: early cohort (n = 200, December 1995–March 2006), and late cohort (n = 299, March 2006–December 2014). We assessed perioperative outcomes after PEA as compared between the early and the late cohort also in propensity-score–matched cohorts. Results Age at the surgery was older in the late cohort (p = 0.042). Preoperative mean pulmonary artery pressure (mPAP) was 46.8 ± 11.0 mm Hg in the early cohort and 43.5 ± 112.7 mm Hg in the late cohort (p = 0.0035). The in-hospital mortality in the early and late cohorts was 14.0% (28/200) and 4.7% (14/299), respectively (p = 0.00030). The duration of circulatory arrest (CA) became much shorter in the late cohort (42.0 ± 20.5 min in the early and 24.2 ± 11.6 min in the late cohort, respectively, p < .0001). In matched cohorts, the in-hospital mortality showed no significant difference (8.7% in the early cohort and 5.2% in the late cohort, < 0.0001). The CA duration, however, was still shorter in the late cohort (p <0.0001). Conclusions Over time, older patients have been accepted for surgery, more patients were operated for lesser severity of CTEPH. Duration of CA and mortality decreased even beyond the first 200 patients, indicating a long learning curve.