In this study we hypothesized that MV with high Vt during LPS-induced peritonitis increases coronary VCAM-1 expression and cardiac edema and thereby aggravates LPS-induced myocardial dysfunction. In contrast to our hypothesis, MV with high Vtin vivo was associated with improved myocardial contractility ex vivo.
We used a LPS-induced peritonitis model as this model is reproducible  and targets, amongst others, the endothelium . Indeed, LPS-induced peritonitis increased pulmonary vascular permeability, as suggested by pulmonary edema. In addition, Vt was set at 19 ml/kg to induce VILI, in accordance with the literature [22, 23]. MV induced a further Vt-dose dependent increase in pulmonary edema and an increase in intra-alveolar cell count, despite unaltered PaO2/FiO2 ratios . These characteristics of our model comply with early indicators of VILI . Furthermore, LPS treatment decreased MAP and maintained CO and CVP without an increased heart rate, which indicates LPS-induced systemic vasodilatation and myocardial dysfunction, as cardiac work for a given preload decreased, and is consistent with previous observations . As MAP decreased after LPS treatment, these rats received more fluids but this was not associated with an increase in cardiac edema. Finally, treatment with LPS characteristically reduced myocardial function ex vivo in line with previous observations [21, 24]
MV with increasing Vt decreased cardiac output in vivo, however, it increased myocardial contractile function ex vivo as evidenced by an attenuated LPS-induced decrease in LV systolic pressure and + dP/dtmax. Diastolic function, measured as -dP/dtmin and tau, was not affected by MV. As coronary flow ex vivo was not affected by MV, the observed changes in myocardial function ex vivo after MV are not due to flow differences. We observed correlations between LV contractile and relaxation parameters and cardiac edema. As edema can hamper myocardial function , these correlations suggest that decreased cardiac edema was, at least in part, responsible for better myocardial contractile function in rats treated with LPS and MV with high Vt compared to rats treated with LPS and MV with low Vt.
Formation of edema depends on microvascular permeability through endothelial activation, and hydrostatic and osmotic pressure differences. VCAM-1 was chosen as parameter for endothelial activation since its coronary expression was shown to be upregulated 4 hours after LPS treatment [6, 24]. Our study confirmed this, but the phenomenon was not affected by MV. The lack of effect of MV on VCAM-1 expression indicates that it is unlikely that the difference in edema among the modes of MV was caused by differences in microvascular permeability but rather resulted from differences in filtration pressure. We may speculate that a lower transmural coronary venous pressure during MV with high Vt was partly responsible for the decrease in edema, particularly during increased permeability in LPS-treated rats. Transmural pressure of the coronary veins, which is a driving force for cardiac edema, can be estimated from CVP-(mean airway pressure*pressure transmission). CVP increased by 1 mmHg during MV with high Vt. Mean airway pressure increased during high Vt ventilation by about 3.7 mmHg (5 cm H2O) as mean airway pressure depends on PEEP, which was similar between the two ventilated groups and Pplat, which was higher during MV with high Vt. Airway pressure transmission is normally about 50%, but can decrease to 30% in diseased lungs due to increased stiffness . These pressure differences result in a lower transmural pressure of the coronary veins during MV with high Vt, even in the case of a decreased airway pressure transmission, and this may have partially prevented edema formation. In any case, the hydrostatic and colloid osmotic pressure forces in the Langendorff-perfused heart set up were similar across groups, so that development of group differences in determinants of cardiac edema formation ex vivo can be excluded .
Although it is generally accepted that hydrostatic edema affects diastolic function , a decrease in myocardial edema after MV with high Vt did not attenuate LPS-induced diastolic dysfunction in our study. Hence, the decrease in diastolic function during LPS-induced peritonitis was only partly caused by increased permeability edema, thereby perhaps rendering diastolic function, in contrast to systolic function, relatively insensitive to attenuation of edema during MV with high Vt. In fact, most studies indicating edema-induced diastolic dysfunction were performed in otherwise healthy animals subjected to hydrostatic rather than increased permeability cardiac edema formation [8, 27].
Cardiac output in vivo decreased with increasing Vt. In contrast, myocardial function ex vivo increased when animals were ventilated with increasing Vt. As HR did not change with increasing Vtin vivo and as myocardial function ex vivo was measured independent of changes in loading and, our data suggest that changes in cardiac output with high Vtin vivo were most likely caused by changes in loading conditions and not by changes in myocardial contractile function. Moreover, Gurkan et al. found increased inflammation in livers and kidneys but not in hearts of mice after acid aspiration and ventilation with high Vt . Indeed, patients on MV often die from multiple organ failure and renal dysfunction may be more prevalent than cardiac dysfunction .
Limitations of our study include the absence of an intervention to show a direct relation between less myocardial edema and increased contractility in vivo. At the time the study was designed, a control group with injurious ventilation but unchanged airway pressure was not included. Moreover, in this study we measured myocardial function ex vivo as this provides the opportunity to measure function independent from differences in coronary flow, heart rate, loading and endocrine control. However, isolation of the heart from the whole animal may also be considered as a limitation as this takes the study further away from clinical relevance .
In conclusion, MV with high Vt lead to mild VILI but did not induce or aggravate myocardial dysfunction caused by LPS-induced peritonitis. Instead, LPS-induced myocardial dysfunction was attenuated in a Vt-dependent manner. Attenuation was most likely due to a decrease in cardiac edema as a result of a fall in transmural coronary venous pressure in face of increased endothelial activation and permeability, associated with coronary inflammation. This contrasts with the injurious effect of VILI on other remote organs like the kidneys .