Evaluation of the pulse-contour method of determining stroke volume in man.

one patients with ischemic heart disease and 21 patients with mitral valve disease were subjected to a variety of hemodynamic interventions. These included exercise, isoproterenol infusion, and practolol, morphine or pentazocine administration. The pulse-contour estimations, using three different formulae derived by Warner, Kouchoukos, and Herd, were compared with indicator-dilution outputs. The pressurepulse-contour computations were performed by a small on-line cardiac catheterization laboratory computer. A comparison of the results of the two methods for determining stroke volume yielded correlation coefficients ranging from 0.59 to 0.84. The better performing Warner formula yielded a coefficient of variation of about 20%. The type of hemodynamic interventions employed did not significantly affect the results using the pulse-contour method. This method yielded nearly as reliable results with the catheter positioned in the axillary artery as in the central aorta. Good results were obtained using the pulse-contour method in patients with atrial fibrillation as compared to those with sinus rhythm. Although the correlation of the pulse-contour and indicator-dilution stroke volumes is high, the coefficient of variation is such that small changes in stroke volume cannot be accurately assessed by the pulse-contour method. However, the simplicity and rapidity of this method compared to determination of cardiac output by Fick or indicator-dilution methods makes it a potentially useful adjunct for monitoring critically ill patients.