Despite the fact that no new clinical outcome studies comparing intermittent hemodialysis (IHD) and continuous renal replacement therapy (CRRT) for acute kidney injury (AKI) have been published in the past year, two meta-analyses addressing the topic (Bagshaw et al, Crit Care Med 2008; 36: 610–7, and Pannu et al, JAMA 2008; 299: 793–805) have been published recently. With respect to randomized controlled trials (RCTs), there was a substantial overlap between the studies considered in the analysis by Bagshaw et al and those considered in the analysis by Pannu et al. Although neither metaanalysis showed a benefit for either modality with respect to mortality or renal recovery, the two publications offered vastly different conclusions. Bagshaw et al concluded it is impossible to make any definitive recommendations about dialysis modality choice in AKI because previous studies were not adequately powered and failed to standardize for treatment dose. On the other hand, because the metaanalysis of Pannu et al demonstrated equivalent patient outcomes, and in light of the lower costs of IHD, they suggested that alternate-day hemodialysis should become the preferred therapy in many critically ill patients. As the clinical practice recommendations made by Pannu and colleagues have very important implications, we believe their analysis should be critically assessed. In this review, the weaknesses of the RCTs considered in the meta-analysis by Pannu et al are presented. Furthermore, the assumption by Pannu et al that IHD is associated with lower costs than CRRT is challenged, as they did not consider adequately both the short-term and long-term costs associated with the dialytic management of AKI patients. Based on our critical analysis, we believe the AKI dialytic treatment approach recommended by the JAMA investigators (Pannu et al) is not supported by the aggregate of the available clinical outcome data and, therefore, remains highly controversial.
We would like to join with others in the AKI field by strongly recommending that investigators and other clinicians stop trying to make conclusive determinations about dialysis modalities when robust supportive data simply are not available. Instead of additional intermodality comparisons, the focus of future clinical research should be toward generating high-quality data on intramodality interventions, such as treatment dose and timing of treatment initiation. In this regard, at least for CRRT, we anxiously await the results of the ongoing RCTs evaluating the effect of CRRT dose on patient outcome.
Restricted accessAnnouncementFirst published March, 2008pp. 197-197
