Central Venous Pressure and Mean Airway Pressure Back for Outcome in Mechanically Ventilated Patients? Still More Consideration for Cardiopulmonary Interactions at the Bedside

    loading  Checking for direct PDF access through Ovid


Central venous pressure (CVP) has been for many years a topic of controversy in critically ill patients. Many studies have shown that either its measure was inaccurate to predict a successful volume infusion with at best only a 50% chance (1–4) or more useful only in its very low or very high values.
However, most CVP values are in clinical practice, within the middle values, preventing most of times a reliable answer.
Therefore for many years, CVP was not used any more in most centers.
However, since being a very important physiologic parameter in basic cardiovascular physiology (5), it stayed a useful parameter in some studies to be used in an algorithm (6).
However, in the last version of the Surviving Sepsis Campaign (7), CVP was definitely pulled off the traditional tools to assess hemodynamics because of its confirmed inaccuracy.
Therefore, it is still very interesting to observe the results of a study analyzing the outcome of a consecutive population of nonspecified mechanically ventilated (MV) patients from CVP values.
It is what Long et al (8) have performed in their study in a recent issue of Critical Care Medicine.
The second variable in this study tested was mean airway pressure (Paw), that is generally less used than plateau pressure and driving pressure to try to minimize the risks of MV. This parameters reflect in fact lungs compliance and the more elevated the worse is the lung disease especially in acute respiratory distress syndrome (ARDS).
Paw appears indeed to reflect more the actual mean alveolar pressure throughout the whole respiratory cycle than the plateau or even the driving pressure reflecting more the inspiratory phase.
Therefore, it is why Paw was tested in this study and was found also to be a significant parameter to predict the outcome, like CVP, where they were higher in the nonsurvivors group (9.82 and 9.64 cm H2O, respectively) compared to the survivors group.
These two parameters were found more predictive of the outcome in a multivariate regression including usual parameters: heart rate, mean arterial pressure, CVP, perfusion index, CO2 arteriovenous pressure gradient, lactates, arterial oxygen pressure, Pmean, driving pressure, positive end-expiratory pressure, dynamic compliance, Acute Physiology and Chronic Health Evaluation II, and Sequential Organ Failure Assessment scores.
The interpretation of the authors is that cardiopulmonary interactions seem somewhat not enough considered in MV patients because most attention is more driven by minimizing the deleterious effects of MV on the lung rather than on cardiovascular compartment.
The authors remind that first intrathoracic pressures decrease venous return and therefore cardiac output explaining why Paw appeared in this study the strongest parameter correlated to the outcome, through its negative cardiovascular effects.
Second, CVP, reflecting in part right ventricular pressure and therefore dysfunction, is elevated in most severe acute respiratory failure where appears an hypoxic vasoconstriction, pulmonary hypertension and elevated right ventricle (RV) afterload, and lower RV output and finally left ventricle output.
If these concepts are well known from all intensivists, the authors suggest, from their results, that a more important consideration to the whole process of cardiopulmonary interactions should be performed in MV patients. They remind also that recent studies have shown that ARDS patients were more often dying form circulatory failure than respiratory failure, underscoring the importance of careful and deep analysis of cardiopulmonary interactions in any patient under mechanical ventilation.
This study has some strengths: high number of patients, consecutively included, good statistics, and a correct interpretation of the results observed.
However, this study has two weaknesses: The first is the retrospective and monocentric methodology limiting the generalization of the results.
    loading  Loading Related Articles