부산대학교 도서관

Background:The hemodynamic response to mechanical ventilation (MV) may be influenced by left ventricular (LV) and right ventricular (RV) function, but data are limited to small preclinical and clinical series. In a cohort of cardiac intensive care unit (CICU) patients (pts) with LV or RV dysfunction, we evaluated the association between positive end-expiratory pressure (PEEP), duration of MV, and mortality.Methods:We included all pts in the MIMIC?III database admitted to the CICU receiving invasive MV during the first 24hrs. Pts were stratified by LV or RV dysfunction on echocardiography and average PEEP over 24hrs divided above (high) or below (low) the median cohort levels. Multivariable models were adjusted for demographics and clinical characteristics. The primary outcomes were 30-day all-cause mortality and ventilator free days (VFD).Results:Among the 308 pts included, 39% had LV, 26.6% had RV, and 44.5% had biventricular dysfunction. The median PEEP was 5.8 (IQR 5.0 ? 8.4) cmH20 and median plateau pressure was 19.7 (IQR 17-23) cmH2O. Unadjusted in-hospital mortality was not different between low and high PEEP (27.9% vs 29.9%, p = NS). After multivariable adjustment, differences were not significant but there was a trend towards lower mortality in pts with LV dysfunction receiving high PEEP (OR 0.14, 95% CI 0.02 ? 1.09, p=0.06; Figure A). Point estimates suggested potential harm in pts with RV dysfunction (OR 3.72 95% CI 0.55 ? 25.07, p=0.18). No differences in VFD between low and high PEEP were observed (22.5 vs. 18 days, p=NS). In pts with RV dysfunction, high PEEP was associated with prolonged MV, (lower VFD, OR 0.25 95% CI 0.11-0.56, p=0.001, figure B).Conclusions:In a large cohort of pts undergoing MV in the CICU, no differences in morality were observed by PEEP level. However high PEEP was associated with longer duration of MV in pts with RV dysfunction and possibly benefit in pts with LV dysfunction. Prospective studies are necessary to investigate these findings.