By Peter Laird, MD
In 2001, the world of academic nephrology came alive in a harmonious chorus calling for a randomized and controlled trial to settle the issue on whether more frequent or longer duration dialysis improves outcomes. Out of an NIH consensus conference, the FHN was born and began to ring the virtues of its 10 year ordeal long before it was published. As we know, the short daily arm of the FHN showed significant improvements in the co-primary composite outcomes, but the nocturnal trial did not. Even with the "positive" outcome of the short daily trial, many nephrologists dismissed the findings because of the design with composite outcomes which some stated "didn't prove anything." The alleged "negative" nocturnal trial added only confusion and no closure on the issue of more frequent and longer duration dialysis. Even those in support of the FHN wished foremost for a RCT that evaluated mortality as the primary outcome:
Given these uncertainties, we do not have sufficient data on the effects of daily and nocturnal hemodialysis to advocate for its widespread use and underwrite its cost. In 2001, Chertow (19) argued that a RCT of frequent hemodialysis was necessary to answer the fundamental questions of whether these treatments improve outcomes and, if so, by how much. He urged that the primary outcome of such a trial should be mortality alone or mortality combined with a major morbid event.
Today, we are beginning to hear the same drum beat in the background pounding in rhythmic unison, R-C-T, R-C-T, R-C-T in the hallowed halls of the academic nephrology world. Today, the goal of a future RCT would be that of mortality, the holy grail itself of any randomized clinical trial. However, the reality of putting forth such a trial is daunting according to prior calculations of the number of patients needed to complete such a trial prior to the FHN.
We estimate that currently 55 centers in the United States and Canada are performing frequent hemodialysis five times per week or more. These centers are dialyzing a total of approximately 600 patients with high-frequency hemodialysis at home or in-center. Power analysis showed us that with 1 yr of follow-up, we would need to enroll >3500 patients in the daily study and >5000 in the nocturnal study to achieve 90% power to detect a relatively substantial 30% reduction in mortality. These numbers are severalfold higher than the total number of patients in North America who are receiving this treatment.
With the severe recruitment difficulties especially in the FHN nocturnal arm, the likelihood that we could assemble enough willing patients for even a small percentage of those needed in any such trial renders such an approach unattainable noting the inability to recruit even one hundred patients for the failed FHN nocturnal trial. Instead of acting upon the available evidence to date including two RCTs showing significant benefits and the recent observation trial confirming again the "killer dialysis weekend," the nephrology community appears paralyzed like a deer in headlights frozen motionless as the car approaches as they continue to ask for a randomized trial that NIH has voiced no wish to fund at this time.
There is no doubt that a randomized trial of session timing, length, and schedule is needed. This new study provides further justification (and “clinical equipoise” as noted by the authors) for such. In the meantime, we are mired in a delivery system driven by strict schedules and tight budgets. What do we tell our patients? How do we alter practice now within the confines of the current system? Here are some thoughts:
The authors cite several limitations to the study. First, the design was retrospective and relied on administrative codes to identify cardiovascular events. Also, accurate estimates of residual renal function were not available to the researchers, and cost data were not part of the study. "[G]iven that hospitalization is very costly in the dialysis population," the authors write, "it might be relevant to compare the upstream costs of an extra dialysis session every 2 weeks with the downstream costs associated with the use of other health care resources."
Despite the study's limitations, the authors conclude the results provide sufficient clinical rationale for conducting a controlled trial assessing how dialysis services are provided.
Essentially every article written on dialysis outcomes in the last decade that has revealed positive or negative findings has ended with the same mantra of we need a larger, better, clearer randomized and controlled trial that will answer all of the questions. Yet, for those of us already on dialysis, the likelihood that we will survive ten years waiting for such a study to be completed if they would ever find the funding in the first place and then be able to recruit 3500-5000 patients in the second place defies human logic. In the end analysis, calling for further study on a procedure going into it's sixth decade truly is a profession frozen in the lights of an oncoming car. Unfortunately, it won't be the nephrologists that suffer from their intransigent indecision, it will be patients now in need of their leadership and advocacy. Once again, it is likely that this medical specialty will once again fail its most needful patients.