Operative mortality and complication risk model for all major cardiovascular operations in Japan

Abstract

Background: The Japan Cardiovascular Surgery Database (JCVSD) is a nationwide benchmarking project to improve the quality of cardiovascular surgery in Japan. This study aimed to develop new JACVD risk models not only for operative mortality but also for each postoperative complication for coronary artery bypass grafting (CABG) operations, valve operations, and thoracic aortic operations.

Methods: We analyzed 24,704 isolated CABG operations, 26,137 valve operations, and 18,228 thoracic aortic operations. Risk models were developed for each operation for operative death, permanent stroke, renal failure, prolonged ventilation (>24 hours), deep sternal wound infection, and reoperation for bleeding. The population was divided into an 80% development sample and a 20% validation sample. The statistical model was constructed by multiple logistic regression analysis. Model discrimination was tested using the area under the receiver operating characteristic curve (C index).

Results: The 30-day mortality rates for isolated CABG, valve, and thoracic aortic operations were 1.5%, 2.5%, and 6.0%, respectively, and operative mortality rates were 2.4%, 3.8%, and 8.4%, respectively. The C indices for the end points of isolated CABG, valve, and aortic thoracic operations were 0.6358 for (deep sternal infection) to 0.8655 (operative mortality), 0.6114 (reoperation for bleeding) to 0.8319 (operative death), and 0.6311 (gastrointestinal complication) to 0.7591 (operative death), respectively.

Conclusions: These risk models increased the discriminatory power of former models. Thus, our models can be said to reflect the current state of Japan. With respect to major complications, useful feedback can now be provided through the Japan Cardiovascular Surgery Database Web-based system.