Continuous Renal Replacement Therapy in the Treatment of Acute Renal Failure: Critical Assessment Is Required

Evidence from Randomized, Clinical Trials

Several randomized, clinical trials have compared the outcomes of CRRT and IHD in critically ill patients with ARF, although, until recently, none has been powered adequately (reviewed in reference [13]). Mehta et al. (14) compared CRRT and IHD in 166 critically ill patients with severe AKI, with the finding of a significantly higher ICU mortality rate in patients who were randomly assigned to CRRT (60 versus 42%; P = 0.02). However, despite randomization, patients who were assigned to CRRT were found subsequently to be more likely to have had liver failure and had a higher overall severity of illness, as determined by Acute Physiology, Age, Chronic Health Evaluation III (APACHE III) score. After adjustment for these factors, the increased risk that was attributed to CRRT no longer was statistically significant (odds ratio 1.6; 95% confidence interval [CI] 0.7 to 3.3). However, within each tertile of severity of illness, randomization to CRRT was associated with a trend toward higher rather than lower risk (Figure 1). In a meta-analysis that encompassed six published randomized, controlled trials that compared IHD with CRRT through 2002, the relative risk for mortality with IHD was 0.96 (95% CI 0.85 to 1.08; P = 0.50) (13). In a recently reported randomized trial, 125 patients were randomly assigned to CRRT (venovenous hemodiafiltration) or IHD from a single-center hospital ICU. In hospital, mortality rates did not differ by treatment assignment (47 versus 51%, venovenous hemodiafiltration versus IDH; P = 0.72) (15). Therefore, published data from randomized trials do not support the contention that CRRT is a superior therapy.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Relationship between severity of illness and mortality by renal replacement therapy (RRT) modality in a randomized, clinical trial. □, intermittent hemodialysis (IHD); , continuous RRT. Adapted from reference (14), with permission.

More recently, Vinsonneau et al. (16) from the Hemodiaf Study Group reported the results of the largest, best powered, prospective, randomized, multicenter study reported to date comparing the results of CRRT with IHD. A total of 360 critically ill patients were randomly assigned. In an intention-to-treat analysis, there was no difference in the primary end point of 60-d survival (32% in the IHD group versus 33% in the CRRT group; Figure 2). Of interest, the authors noted that there was an unexpected progressive and significant increase in survival rates in the IHD group over time (relative risk 0.67 per year; 95% CI 0.56 to 0.80; P < 0.001). This suggests that there may have been a learning curve for optimizing IHD therapy in this study environment. Perhaps if IHD therapy had been optimized from the onset of the trial, there would have been at least a trend to a better outcome with IHD compared with CRRT (as was seen in the Mehta trial). Thus, the weight of available evidence, including the most recent evidence from randomized, clinical trials, provides no convincing evidence to support superiority of CRRT over IHD.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Comparison of survival characteristics between continuous venovenous hemodiafiltration (CVVHDF) and IHD in a randomized, clinical trial. Reprinted from reference (16), with permission.

Evidence from Observational Studies

A number of observational cohort studies also previously compared outcomes with CRRT or IHD in critically ill patients with severe AKI. The early published literature that compared results of CRRT and IHD seem to confirm significant superiority of CRRT on the basis of overall patient survival (reviewed in reference [13]). However, these studies generally were retrospective and often used noncontemporaneous controls for assessment of results in the IHD group. Furthermore, IHD often was delivered using less biocompatible cellulosic membranes, which may have exacerbated the systemic inflammatory process and reduced the likelihood of renal functional recovery (17). Several of these reported studies used acetate-based dialysate, further increasing the likelihood that intradialytic hypotension would develop with IHD. In a meta-analysis that included several of these studies, the overall relative risk for mortality with the provision of IHD versus CRRT was 1.00 (95% CI 0.92 to 1.08). However, it is instructive perhaps that of the five studies performed within the past decade, two revealed significantly higher adjusted risk for mortality with the provision of CRRT, and two additional studies displayed a trend toward a higher relative risk for mortality with CRRT (Figure 3). Although overall these data suggest that there may be equipoise in comparing CRRT with IHD, recent trends actually favor IHD.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Meta-analysis of the relationship between RRT modality and mortality. Recent studies trend toward improved outcome with IHD. Reprinted from reference (13), with permission.

Recently, the Program to Improve Care in Acute Kidney Disease (PICARD) Group extended the previous work in this area comparing dialysis modalities by incorporating into an analysis for the first time multiple sites, multivariable regression analysis, and the propensity score approach to address residual confounding (18). In this multicenter, prospective, observational cohort study in critically ill patients with AKI, the provision of CRRT in comparison with IHD was associated with a significantly higher relative risk for mortality. The higher relative risk for mortality using CRRT was noted in fitted models that were adjusted for covariates only, propensity score only, and a combination of covariates plus the propensity score. Furthermore, the mortality risks were nominally higher, with assignment to CRRT across tertiles reflecting mortality risk (although this did not reach statistical significance in the middle tertile). These data corroborate other recent observational studies that provide no evidence for a survival benefit afforded by CRRT and suggest that under some circumstances, the use of CRRT may have been associated with harm.

Claims for better outcomes from CRRT always have been based on the concept that CRRT can provide more hemodynamic stability, better control of circulating volume, increased convective clearance, and the ability to provide nutritional support, BP support, and other obligate fluids more fully. Because available recent data do not support the hypothesis that CRRT will provide better outcomes than IHD, it may be worthwhile to reassess this concept critically. In the past two decades, a number of technical advances have allowed for safer delivery of IHD therapy. These include the introduction of volumetric control, the use of lower dialysate temperature to minimize intradialytic hypotension, the use of more compatible dialysate, and the widespread use of more biocompatible high-flux synthetic dialysis membranes. Several reports have indicated that it now is possible to deliver IHD therapy with improved hemodynamic tolerance even in critically ill patients. In the only randomized, crossover study reported to date, the hemodynamic response to IHD and continuous hemofiltration were similar (19). Schortgen et al. (20) also demonstrated that implementation of clinical practice guidelines that are designed to improve hemodynamic tolerance can lessen hemodynamic instability and may improve outcome in patients with ARF. Therefore, the original rationale that led to the development of CRRT techniques may have dissipated over time.

Potential Risks Associated with CRRT

Although much attention has been focused on perceived benefits of CRRT compared with IHD, comparatively less attention has been focused on the potential for increased risks with CRRT therapy. Because CRRT is a continuous extracorporeal therapy, there frequently is a requirement for prolonged anticoagulation therapy, which may increase bleeding risk. Clotting of the extracorporeal circuit occurs frequently with CRRT, which may increase blood loss and exacerbate anemia. Recently, several safety concerns have been raised with a CRRT delivery device (21). In some circumstances, CRRT is performed by personnel who are less experienced in the procedure than trained nurses who perform IHD. The use of CRRT may enhance removal of amino acids, vitamins, small peptide hormones, catecholamines, and other solutes with beneficial function in critically ill patients.

Reassessment of the Utility of the Pulmonary Artery Catheter: Analogous to CRRT?

An excellent example of a process of critical reassessment is the use of the pulmonary artery catheter, introduced as a novel technology for use in critically ill patients in the 1970s (22). The introduction of right heart catheterization in 1929 by Forssmann, followed by the work of Cournand and Richards in the 1940s, provided tremendous advances in understanding cardiopulmonary physiology, cardiovascular hemodynamics, and gas exchange. In 1956, Forssman, Cournand, and Richards shared the Nobel Prize for this seminal work. By 1970, the Swann-Ganz balloon-tipped flow-directed pulmonary artery catheter further brought cardiopulmonary physiology to the bedside, achieving near instantaneous widespread use in the ICU environment (23). Acceptance of the utility of the Swann-Ganz catheter in improving patient outcomes went unquestioned until a now-famous editorial entitled “The Cult of the Swann-Ganz Catheter: Overuse and Abuse of Pulmonary Flow Catheters” was published by Robin in 1985 (24). Although highly controversial at the time it was written, it stimulated the development of large, prospective, observational cohort studies and eventually randomized, clinical trials. A decade after this editorial was published, a large cohort study suggested that the use of pulmonary artery catheters in ICU settings actually might increase morbidity and mortality (25). In the past decade, three studies now have reported no benefit to the use of pulmonary artery catheters in patients who have congestive heart failure or undergo high-risk surgery and in the management of acute respiratory distress syndrome (26–28). Although there still is a clinical role for the use of pulmonary artery catheters (particularly in the management of primary pulmonary hypertension), the range of clinical scenarios in which the use of these catheters is beneficial has been circumscribed narrowly compared with its original use. In the case of pulmonary artery catheters, three decades of clinical use and careful research has evolved the use of these catheters from unbridled enthusiasm to a more careful, sober, evidence-based recognition of risk-benefit ratio in diverse clinical settings (22).

There are other examples in which technologies, devices, and drugs that have been introduced into medical practice on the basis of a sound rationale subsequently were demonstrated through clinical practice and research not to be associated with benefit and perhaps even to be associated with harm. Recent examples include liberal use of blood transfusions in the ICU (29), the use of certain antiarrhythmic agents for sudden death prevention (30), the use of cyclooxygenase-2 selective nonsteroidal anti-inflammatory agents (31–34), and the use of selected inotropes and vasodilators in the treatment of congestive heart failure (35). When examining the results of recent observational studies and clinical trials, there is no convincing evidence to support superiority of CRRT over IHD in the treatment of critically ill patients with ARF.

Disclosures

None.