My paper 20 years later: Effect of positive end-expiratory pressure on right ventricular function in humans

In 1992, we published a report on the effect of positive end-expiratory pressure (PEEP) on right ventricular (RV) function in humans. We measured RV volumes and pressures and pericardial pressure (Ppc) as PEEP was increased from zero to 15 cm H20 in 12 patients after thoracotomy, using a pulmonary arterial catheter equipped with a rapid responding thermistor that allowed measurement of RV ejection fraction (RVef), while Ppc was measured via a pericardial balloon catheter. RV end-diastolic volume (EDV) was estimated as the ratio of stroke volume (SV) to RVef, whereas RV end-systolic volume (ESV) were estimated as RV EDV-SV. PEEP increased Ppc and Pra, but RVef unaltered. There was no relation between either RV filling pressure (Pra-Ppc) and EDV or the change in RV filling pressure and EDV, although EDV varied significantly with PEEP (p < 0.05). The relations between EDV and both SV and RVef were weak (r = 0.54 and 0.55, respectively). RVef varied inversely with ESV (r = -0.77), although it showed no relation to transmural peak pulmonary artery pressure (r = 0.28). However, both absolute and relative changes in EDV corresponded closely with respective ESV values (r = 0.94). We concluded that EDV varies independently of changes in filling pressure and that changes in ESV occur independently of changes in ejection pressure. These data can be explained by assuming that the RV shape changes can dissociate changes in RV EDV from changes in RV wall stress (preload). Thus, changes in RV EDV may or may not alter SV but should proportionately change ESV to a degree dependent on election pressure and contractility. Subsequent studies confirmed our findings which can be summarized as 1) RV filling is independent of Pra; thus central venous pressure cannot be used to estimate RV preload; and 2) for cardiac output to increase by the Starling mechanism the RV must dilate increasing RV ESV. Since the pericardium limits absolute biventricular volume, there is a finite limit to which cardiac output can increase by the Starling mechanism defined not by left ventricular contractility but by RV function. And 3) if fluid loading causes Pra to increase without increasing cardiac output, then resuscitation should stop as the patient is going into acute cor pulmonale. These truths help bedside clinicians understand the echocardiographic and hemodynamic signatures of both RV failure and volume responsiveness.