[Coronary angiography by 16-slice computed tomography prior to valvular surgery]

Multislice computed tomography is an excellent technique for the detection of significant coronary artery lesions. Our purpose was to assess whether computed tomography could replace routine invasive coronariography before valvular surgery.

We studied 106 consecutive patients (mean age: 67 [10]): 76% aortic valvular disease (62% stenosis, 14% regurgitation), 20% mitral valvular disease (4% stenosis, 16% regurgitation), and 4% mitro-aortic disease. Non-invasive studies were performed by helical computed tomography. Eighty-four percent of patients were in sinus rhythm (40% using beta-blockers, 32% nitrates). Findings from both techniques were analyzed according to a predetermined segmented anatomical model of the coronary artery (a total of 1802 segments).

The incidence of coronary artery disease in these patients was 30%. Using computed tomography, 96.8% of segments could be evaluated and 3.2% could not. Calcium score ranged from 0 to 7572 (median: 182). In the per patient analysis, computed tomography showed a sensitivity of 95%, specificity 94%, positive predictive value 84%, and negative predictive value 98%.

Computed tomography is an excellent technique for ruling out coronary lesions prior to valvular surgery, making an invasive study unnecessary if the quality of the study is good and the result is negative.