Surgical management of complete atrioventricular septal defect: associations with surgical technique, age, and trisomy 21

Abstract

Objectives: We sought to evaluate the contemporary results after repair of a complete atrioventricular septal defect and to determine the factors associated with suboptimal outcomes.

Methods: The demographic, procedural, and outcome data were obtained within 1 and 6 months after repair of a complete atrioventricular septal defect in 120 children in a multicenter observational study from June 2004 to 2006.

Results: The median age at surgery was 3.7 months (range, 9 days to 1.1 years). The type of surgical repair was a single patch (18%), double patch (72%), and a single atrial septal defect patch with primary ventricular septal defect closure (10%). The incidence of residual septal defects and the degree of left atrioventricular valve regurgitation (LAVVR) did not differ by repair type. The median interval of intensive care stay were 4 days, ventilation use 2 days, and total hospitalization 8 days. All were independent of the presence of trisomy 21 (80% of the cohort). The in-hospital mortality rate was 2.5% (3/120). The overall 6-month mortality rate was 4% (5/120). The presence of associated anomalies and younger age at surgery were independently associated with a longer hospital stay. The age at repair was not associated with residual ventricular septal defect or moderate or greater LAVVR at 6 months. Moderate or greater LAVVR occurred in 22% at 6 months, and the strongest predictor for this was moderate or greater LAVVR at 1 month (odds ratio, 6.9; 95% confidence interval, 2.2-21.7; P < .001).

Conclusions: The outcomes after repair of complete atrioventricular septal defect did not differ by repair type or the presence of trisomy 21. An earlier age at surgery was associated with increased resource use but had no association with the incidence of residual ventricular septal defect or significant LAVVR.

Trial registration: ClinicalTrials.gov NCT00113698.

Figure 1

Piecewise regression fit for ventilator duration, ICU stay, and hospital stay by age at surgery. The solid line indicates the predicted values from the regression versus age at surgery. Each outcome was negatively correlated with age at surgery from birth to about 2.5 months, but did not vary by age after 2.5 months.

Figure 2

Box plots of ventilator duration, and ICU stay, and hospital stay by age group and number of concurrent surgeries. P-values are from linear regression modeling of log-transformed outcome. The asterisk denotes the group mean. The lower, middle, and upper edges of the boxes represent the 25^th, 50^th, and 75^th percentiles, respectively.