Implications of an audible third heart sound in evaluating cardiac function.

We prospectively compared auscultatory findings of third heart sounds with radionuclide ventriculographic analysis of systolic and diastolic function. Cardiac auscultation was done to detect an S3 in patients referred for radionuclide ventriculographic analysis of ventricular function. Of 49 adult men with the diagnosis of chronic, nonvalvular heart failure who were referred for the evaluation of heart failure, 22 (45%) at the time of the ventriculography had an S3 present on examination. For the entire study group, the radionuclide ventriculography-derived ejection fraction was 33% +/- 19.5 (mean +/- SD) with a range of 6% to 74%. The peak ejection rate was 2.05 +/- 1.09 end-diastolic volume per second with a range of 0.30 to 4.56. The peak filling rate was 1.97 +/- 1.07 end-diastolic volume per second with a range of 0.44 to 3.94, and the time to peak filling rate was 0.18 +/- 0.11 per second with a range of 0.05 to 0.61. The presence of an S3 was associated with a reduced ejection fraction and also with impaired diastolic function as determined by the peak filling rate. The sensitivity and specificity for the S3 in detecting abnormal systolic function (ejection fraction < 50%) were 51% and 90%, respectively, with a positive predictive value of 95% and a negative predictive value of only 32%. For an ejection fraction of less than 30%, the S3 had a sensitivity and specificity of 78% and 88%. The presence of an S3 was highly predictive of an abnormal ejection fraction. The absence of an S3, however, is not uncommon in patients with a mildly impaired ejection fraction.