Diagnostic accuracy of hemoconcentration for pulmonary edema as the cause of weaning failure.

Abstract Our objective was to assess the diagnostic accuracy of hemoconcentration for cardiogenic pulmonary edema (PE) as the cause of weaning failure, using left ventricular filling pressure elevation assessed by transthoracic echocardiography as the reference standard. This prospective observational study included 41 patients who failed their first spontaneous breathing trial of weaning from mechanical ventilation. They were divided into two groups, with and without PE by echocardiographic criteria. Hemoconcentration and other hemodynamic parameters were compared between the groups. The group (N.=21) with PE by echocardiographic criteria had a higher frequency of failure of the second spontaneous breathing trial (P=0.03) and a longer total weaning time (P=0.02) compared to the other group. The receiver-operating characteristics curve indicated that changes in plasma protein or hemoglobin concentration from initiation to completion of the second spontaneous breathing trial did not predict PE as the cause of failure (areas under the receiver-operating characteristics curve, 0.47±0.09 and 0.51±0.09, respectively). The only factor predicting failure due to PE was a positive fluid balance from intensive care unit admission to study inclusion (P=0.01). The increase in mean arterial blood pressure seemed suggestive of weaning failure due to cardiac causes. Compared to echocardiographic criteria for left ventricular filling pressure elevation, hemoconcentration assessed based on plasma protein and hemoglobin levels did not help to diagnose cardiogenic PE as the cause of weaning failure.