SURGICAL REPAIR OF COMPLETE ATRIOVENTRICULAR-CANAL DEFECTS IN INFANCY- 20-YEAR TRENDS

Case histories of 301 patients with complete atrioventricular canal de fect presenting to our institution in infancy between January 1972 and January 1992 were reviewed with the purpose of identifying the factor s responsible for the observed improvement in perioperative mortality over this time period. A retrospective analysis of hospital records ex amined 46 patient-related, morphologic, procedure-related, and postope rative variables for associations with perioperative death and reopera tion. Operative mortality decreased significantly over the period of t he study from 25% before 1976 to 3% after 1987 (p < 0.0001). A number of the 46 variables examined showed trends over time that were similar to that for mortality. Palliative procedures decreased over time. Reo peration for most residual lesions also decreased to the degree that t hey were essentially eliminated in recent years. The exception to this was reoperation for postoperative left atrioventricular valve regurgi tation, which also decreased but remained at 7% in recent years. Both technical and support-related procedural variables showed no trends ov er time, with the exception of the performance of left atrioventricula r valve annuloplasty, which increased over time. Closure of the left-s ided cleft was performed in 61% of the patients, with no trend over ti me. Annuloplasty and cleft closure were not associated with less posto perative left atrioventricular valve regurgitation, fewer reoperations , or lower mortality. Multivariate logistic regression analysis identi fied only earlier year of operation, the presence of double-orifice le ft atrioventricular valve, and postoperative residual regurgitation of the left atrioventricular valve as risk factors for death. Experience -related improvements in technical precision achieved over time best a ccount for the reduction in the rate of reoperation for most types of residual lesions and also for the reduction in mortality.