Discussion
Pediatric patients who require ECLS are at increased risk for AKI. Important risk factors for AKI in this population include the indication for ECLS, underlying diagnosis, and patient age. Although fewer than 9% of ECLS neonates with predominately respiratory failure develop biochemical indicators of AKI, 22% of noninfant children who require ECLS for predominately cardiac failure develop clinically significant elevations in serum creatinine. Additional complications, such as clinically important hemorrhage, are common in patients on ECLS. Blood component replacement therapy, volume resuscitation, and generalized inflammatory response to the ECLS circuit contribute to excessive extravascular, third-space fluid volume. The impact of excessive total body fluid is compounded by concomitant renal dysfunction, necessitating renal replacement therapy in 26% of high-risk neonatal cardiac ECLS patients. The odds of developing renal failure increase by as much as 60% per day on ECLS. The presence of ECLS-related renal failure is associated with increased hospital mortality (odds ratio = 4.7) and reduced long-term, postdischarge survival.
Data from the Extracorporeal Life Support Organization Registry indicate that hemofiltration is performed in 18% of ECLS patients, whereas only 11% of patients develop biochemical evidence (elevated serum creatinine) of AKI. The use of CRRT is associated with improved negative fluid balance and better caloric intake in pediatric ECLS patients. Consequently, CRRT is used to achieve enhanced fluid management in fluid overloaded ECLS patients, irrespective of concomitant renal failure. Although renal failure that requires the use of CRRT while on ECLS is associated with reduced survival, it is unclear whether the use of CRRT independently increases mortality or plays a causative role in the development of chronic renal failure.
Ultrafiltration of blood is accomplished by generating a pressure gradient across the semipermeable hemofiltration membrane. Rate of ultrafiltration is governed by the magnitude of transmembrane pressure and the surface area and porosity of the membrane. Highly porous filters, as are used in CRRT, are capable of removing large volumes of fluid, which may have a negative impact on the hemodynamic status of critically ill patients. Free-flow CRRT systems lack integrated pressure monitoring capabilities and flow regulation, predisposing them to transmembrane pressure fluctuations and potentially inaccurate ultrafiltration volume. Furthermore, IV pumps were not designed for CRRT, raising questions regarding patient safety. The current generation of integrated CRRT devices are capable of monitoring transmembrane pressure and ultrafiltrate volume and automatically adjusting transmembrane pressure to achieve programmed targets. The majority of integrated CRRT devices employ a gravimetric ultrafiltration control system with sophisticated software to maintain appropriate fluid balance. However, the devices are designed for use principally in adult patients, with company-reported error ranges for ultrafiltrate fluid removal of up to ± 30 mL/hr. Inaccuracies in target ultrafiltration volume have a greater impact in pediatric patients than adults.
Unlike IV pump free-flow systems, integrated CRRT devices require precise coordination with the ECLS circuit. For instance, the software of the integrated CRRT system used in our institution interprets dynamic flow pressures that are lower than a set threshold in the blood access and/or blood return tubing as a blood line disconnection, resulting in an error condition and cessation of pumped blood flow to the device. Pressure readings in excess of a set threshold also result in an error condition and cessation of pump activity. Different approaches to achieve error-free ECLS-CRRT device integration have been reported, with considerable disagreement with respect to the safest and most reproducible methods. The device used in this study is programmed to discontinue CRRT blood flow when the circuit's integrated pressure monitors detect excessive or fluctuating line pressures. Adjustments of the stopcock connecting the access line to the ECLS circuit can be used to increase resistance and eliminate pressure-related alarms and interruptions in CRRT in our circuit configuration. In general, we do not recommend accessing the prepump limb of centrifugal ECLS circuits due to potentially increased risk of entrained air. None of the patients in this study experienced complications related to the prepump access site. Additionally, we have found that the Quadrox oxygenator is capable of filtering small amounts of entrained air with great efficiency. The final ECLS-CRRT configuration at an individual institution will vary considerably depending on the unique arrangement of the ECLS circuits and type of integrated CRRT being used. Each institution that provides ECLS and CRRT service must carefully weigh the potential risks and benefits of operating medical devices outside the manufacturer's specified variables of use.
Ongoing technological advancements create logistical challenges in performing longitudinal studies of complex therapies such as CRRT and ECLS. ECLS systems are becoming more compact and integrated, with an increasing number of centers using centrifugal pumps rather than roller pumps. Several integrated CRRT systems are currently available and each is subject to software updates that may affect accuracy and efficiency of performance. An important limitation of this study is that it examines the performance of only one integrated CRRT system. Another limitation is that the type of ECLS circuits used during the study were not uniform. Additional studies are needed to more clearly define potential interactions between ECLS and integrated CRRT systems.
This study demonstrates that an integrated CRRT device such as the Prismaflex system provides more accurate fluid management than IV pump-controlled free-flow CRRT systems. Although the mortality rates in the free-flow and integrated CRRT groups were similar, median duration of CRRT was shorter in the integrated CRRT group. Preferential use of integrated CRRT devices during ECLS appears to reduce the duration of CRRT and may reduce risks related to inaccurate ultrafiltration. Although the study is underpowered to determine the effect of mode of CRRT on clinical outcomes beyond accuracy of fluid management, increased accuracy facilitates earlier optimization of fluid balance and improved overall clinical status, which may permit earlier discontinuation of CRRT and ECLS, as observed in the integrated CRRT patients. Additional prospective studies are needed to better define criteria for CRRT in ECLS patients and to determine the long-term effects of fluid management in these high-risk patients.