Prediction of postoperative performance in aortic valve disease

Israel Mirsky, Claudia Henschke, Otto M. Hess, Hans P. Krayenbuehl

Research output: Contribution to journalArticlepeer-review

26 Scopus citations


A new direct method has been developed for predicting postoperative performance in patients undergoing aortic valve replacement. Employing micromanometry and cineangiography, a number of conventional hemodynamic and angiographic variables, including the peak value of the first derivative of ventricular pressure divided by ventricular pressure (dP/dt/P)max were evaluated in 171 patients studied preoperatively and in 44 patients studied pre- and postoperatively with an additional 14 patients serving as control subjects. Normal contractile state relations (dP/dt/P)max versus end-diastolic pressure (over a range of 15 mm Hg or less to more than 15 mm Hg) were derived from patients whose preoperative ejection fraction and peak wall'stress were equal to or more than control mean - 2 standard deviations. Postoperative function was predicted to be abnormal (ejection fraction less than control mean - 2 standard deviations) if preoperative values of (dP/dt/P)max and enddiastolic pressure fell below the 95 percent confidence bands of these contractile state relations. The method accurately predicted postoperative function in 40 of 44 patients with a sensitivity of 100 percent. This result was confirmed by a discriminant function analysis (based on preoperative ejection fraction, end-diastolic pressure and [dP/dt/P]max) that yielded correct classifications in 42 of 44 patients. These studies indicate that the preoperative contractile state of the myocardium is the major determinant of postoperative performance in aortic valve disease.

Original languageEnglish
Pages (from-to)295-303
Number of pages9
JournalAmerican Journal of Cardiology
Issue number2
StatePublished - Aug 1981
Externally publishedYes


Dive into the research topics of 'Prediction of postoperative performance in aortic valve disease'. Together they form a unique fingerprint.

Cite this