Cumulative institutional experience and the learning curve of balloon pulmonary angioplasty for chronic thromboembolic pulmonary hypertension
R Takano, K Kanaoka, Y Sumita, J Ueda, A Tsuji, K Tsujita, T Noguchi, Y Miyamoto, T OgoAbstract
Background
Balloon pulmonary angioplasty (BPA) is an established catheter-based treatment for chronic thromboembolic pulmonary hypertension. Since the safety of refined BPA techniques was reported in 2012, BPA has rapidly disseminated nationwide in Japan, with increasing adoption by centers with less experience. Although BPA is known to involve a learning curve, objective criteria defining institutional expertise based on cumulative procedural experience remain undefined.
Purpose
This study aimed to clarify institutional expertise in BPA based on cumulative procedural experience.
Methods
Using the Japanese Cardiovascular Disease All-Case Registry (2012–2022), we retrospectively analyzed 14,468 BPA procedures performed in 3,708 patients. Complications were defined as a composite of the following major adverse events within 30 days after the procedure: death, initiation of mechanical ventilation (including noninvasive positive pressure ventilation), initiation of extracorporeal membrane oxygenation, initiation of continuous hemodiafiltration, transfusion, cardiac tamponade, or surgery. Cumulative procedural experience for each procedure was defined as its chronological order number among BPA procedures performed at the same center. Cumulative procedural experience was categorized into tertiles (1–73, 74–324, ≥324 procedures) and evaluated using multivariable logistic regression. Nonlinear learning-curve effects were assessed using restricted cubic spline (RCS) modeling. A secondary analysis was performed among incident centers that initiated BPA in 2013 or later.
Results
Complication rates decreased stepwise with increasing cumulative procedural experience (low-experience 9.4%, intermediate-experience 5.4%, high-experience 2.6%; trend P < 0.001). This reduction was primarily driven by a lower incidence of mechanical ventilation, while transfusion also declined with increasing experience (trend P < 0.001 and P = 0.049, respectively). In multivariable analysis adjusting for age, sex, NYHA class, and procedural year, the high-experience group had significantly lower complication risk than the low-experience group (adjusted odds ratio 0.331, 95% CI 0.205–0.535). RCS analysis demonstrated a marked reduction in adjusted complication rate, with an elbow point at approximately 52 cumulative procedures, followed by a plateau around 264 procedures. Among incident centers (9,265 procedures), the elbow point was less distinct, whereas adjusted complication rate stabilized at low levels beyond approximately 250 cumulative procedures, consistent with findings in the overall cohort.
Conclusions
A cumulative procedural experience threshold of approximately 250 procedures may serve as a practical institutional benchmark for achieving low complication rates with BPA.FigureFor image description, please refer to the figure legend and surrounding text.