All patients with valvular and congenital heart disease were excluded from the present analysis. A total of 378 patients were eliminated because of missing information on either pulmonary artery pressures (n — 323), ejection fraction (n = 49), or coronary arteriograms (n = 6). A sample of 1,371 patients with complete data on left-sided cardiac (unction as well as pulmonary artery pressures were thus available for survival analysis. The subsample of patients with right atrial pressure and pulmonary wedge pressure—and therefore estimates of pulmonary vascular resistance—was somewhat smaller (n = 1,027), and analyses requiring these variables were performed on a sample reduced by 344 persons. birth control yasmin
Diagnostic procedures were performed based on routine clinical indications and no attempt was made to standardize the clinical care provided. The performance of an echocardiogram and the right heart catheterization were considered standard procedures at our institution during the period of this study. It is possible, however, that some selection for more severe cases did occur. It was not possible in retrospect to identify potential selection biases, however. All prospective analyses were based on the cohort as defined at baseline, so the risk factor-outcome relationships themselves should not be biased.