1344387 - Mitral valve pathology predicts outcome after biventricular repair in patients with a hypoplastic left ventricle

Biventricular repair for children with hypoplastic left ventricles (HLV) may be preferred over single ventricle palliation, either due to non-candidacy for the Fontan operation, or to avoid long-term Fontan complications. However, surgical complexity carries a risk of reoperation or adverse outcome. Because prior knowledge of high-risk anatomic or physiologic features may guide decision-making, we studied preoperative CMR-based predictors of outcome in patients with HLV who underwent biventricular repair.

Methods:

In this retrospective study, all patients at a single center with CMR data and a left ventricular (LV) end-diastolic volume z-score < -2 prior to biventricular repair were included. Patients with intact ventricular septum and endocardial fibroelastosis were excluded. In addition to CMR ventricular volumes, ejection fraction, and flow data, anatomic features of the heart including mitral valve (MV) cross-sectional area at end-diastole were assessed (Figure 1). The composite outcome included time to death, transplant, biventricular takedown, heart failure admission, left atrial decompression, or reoperation.

Results:

Of 95 patients (median age 2.9 [IQR 0.9-4.0] years), 37 (39%) had R-dominant atrioventricular (AV) canal defects, 2 (2%) had transitional AV canal, and 56 (59%) had two separate AV valves. The mean LV end-diastolic volume index was 41 ± 9 ml/m², and 23 patients (24%) had a prior LV recruitment procedure. Fourteen patients (15%) had a MV orifice area z-score < -2.  Forty (42%) patients met at least one outcome measure (reoperation n=37, heart failure admission n=5, death n=5, left atrial decompression n=3, and biventricular take-down n=1) with a median time to composite outcome of 2.7 years. Re-operations included MV repair or replacement (21%), permanent pacemaker placement (12%), relief of sub-aortic obstruction (7%), and residual VSD closure (6%). MV orifice area z-score had a significant non-linear relationship with the primary outcome, with an increased risk of adverse outcome for z-scores < -2 but no association with z-scores in normal range (Figure 2). On multivariable analysis, seven independent risk factors were identified (c-statistic 0.72; Table 1). MV orifice z-score < -2 (HR 6.1, p< 0.01), larger 2-chamber MV annulus diameter (HR 1.3, p=0.04), and MV repair at the time of biventricular repair (HR 4.2, p< 0.01) were independently associated with the primary outcome. In addition, non-cardiac medical co-morbidity (HR 3.0, p=0.02), conal-septal ventricular septal defect (HR 14.2, p< 0.01), and transitional atrioventricular canal (HR 6.7, p=0.03) were independent predictors.

Conclusion:

MV orifice area, when z-score is < -2, is an independent risk factor for primary outcome when controlling for all other variables and is important to consider when evaluating for biventricular repair. Non-cardiac comorbidities, VSD morphology, and 2-chamber mitral valve annulus dimension are also independent risk factors for adverse outcome and should be considered.