Prediction of FR is critical in fluid therapy after cardiac surgery, especially in children. Significant changes in ventricular systolic and diastolic functions after cardiac surgery often lead to difficulties in predicting FR [15,16]. Here, we demonstrated that PLR related SV changes (ΔSV-PLR) as evaluated by bedside ultrasound could effectively predict FR in intubated children undergoing congenital heart surgery procedure. Our data showed that the optimal ΔSV-PLR threshold for predicting FR was 13%.
PLR is a simple, noninvasive and repeatable method to change the cardiac preload [8,9,17]. Studies have shown that hemodynamic changes such as SV and aortic blood flow can be observed after 30 sec of the elevation of lower limbs [18]. Our results showed that PLR can cause significant changes in hemodynamic parameters such as MAP, CVP, CO and SV (from base 1 to base 2), and all these parameters can also be restored to baseline level when the body position is returned to the Base 1 position (base 3 vs base 1). This preliminary observation indicated that PLR could be used as a reversible fluid challenge.
In recent years, bedside ultrasonography has been considered as a noninvasive, real-time, convenient, low-cost, and repeatable tool for monitoring hemodynamics [10,11,19-21]. Previous studies have confirmed that echocardiography was highly correlated with PICCO in CO and SV, and PLR combined with non-invasive ultrasound has significant advantages in evaluating FR, which can be determined by monitoring SV and aortic blood flow [22-24]. The use of ΔSV-PLR to predict FR is based on the beneficial effects of cardiac preload on left ventricle function [25], which is not affected by changes in intrathoracic pressure, myocardial compliance, mechanical ventilation, or drug use. The best cut-off value for ΔSV in predicting FR fluctuates from 7% to 20%, indicating a large variation among different studies [22-24]. Our study demonstrated that an increase of more than 13% in the ΔSV-PLR can predict the FR in children after cardiac surgery, with a sensitivity of 81%, a specificity of 86%, and the AUC of 0.879.
Whereas CVP, MAP, and HR were relatively easy to monitor, we did not observe any correlation with ΔSV-VE. Our data showed that ΔCO-PLR could predict FR, with an optimal threshold of 8%, sensitivity of 81% and specificity of 71%. But the correlation between ΔCO-PLR and ΔSV-VE was not comparable to that of the ΔSV-PLR, which may be due to the influence of the HR on CO measurement. Consistent with the previous findings, our study found that ΔSV-PLR can better predict FR compared with the other PLR-Δs [26].
There are some limitations in this study. First, echocardiographic measurement errors may have occurred, even though the same expert obtained all echocardiographic data. Second, because of the different types of congenital heart diseases and the diverse changes in cardiac structure after operation, our results cannot be extrapolated to other types of congenital heart diseases and to children of all ages, because leg elevation in newborns or infants obviously does not have the same volume effect as leg elevation in older children. Third, based on previous studies, we defined a 10% increase in SV with rapid fluid loading as FR-positive. Whether the 10% cut off is the appropriate threshold needs further research. Another limitation is that it is also a small sample, non-blinded study, and the results may not be applicable to other centers. Finally, someone is a fluid "responder" does not mean that they need a fluid bolus. If a healthy normally hydrated person is given a fluid bolus, their stroke volume will also increase, however, that does not mean that they need fluid. This is probably why the increases in MAP overall are very modest in our study even in responders. Therefore, further investigations of high-quality are needed to confirm our findings.