Comparison of invasive and noninvasive assessment of aortic stenosis severity in the elderly.

BACKGROUND Aortic valve area (AVA) in aortic stenosis (AS) can be assessed noninvasively or invasively, typically with similar results. These techniques have not been validated in elderly patients, where common assumptions make them most prone to error. Accurate assessment of AVA is crucial to determine which patients are appropriate candidates for aortic valve replacement. METHODS AND RESULTS Fifty elderly patients (mean 86 years, 46% female) referred for cardiac catheterization to evaluate AS also underwent transthoracic echocardiography within 24 hours. To minimize assumptions all patients had 3-dimensional echocardiography (Echo-3D), and at catheterization using directly measured oxygen consumption (Cath-mVo(2)) and thermodilution cardiac output (Cath-TD). Correlation between Cath-mVo(2) and Echo-3D AVA was poor (r=0.41). Cath-TD AVA had a moderate correlation with Echo-3D AVA (r=0.59). Cath-mVo(2) (AVA=0.69 cm(2)) and Cath-TD (AVA=0.66 cm(2)) underestimated AVA compared with Echo-3D (AVA=0.76 cm(2;) P=0.08 for comparison with Cath-mVo(2); P=0.001 for Cath-TD). Compared with Echo-3D, the sensitivity and specificity for determining critical disease (AVA <0.8 cm(2)) were 81% and 42% for Cath-mVo(2), and 97% and 53% for Cath-TD. The only independent predictor of the difference between noninvasive and invasive AVA was stroke volume index (P<0.01). Resistance, a less flow-dependent measure, showed a stronger correlation between Echo-3D and Cath-mVo(2) (r=0.69), and Echo-3D and Cath-TD (r=0.77). CONCLUSIONS Standard techniques of AVA assessment for AS show poor correlation in elderly patients, with frequent misclassification of critical AS. Less flow-dependent measures, such as resistance, should be considered to ensure that only appropriate patients are treated with aortic valve replacement.