Timing Renal Replacement Therapy
The June 27 issue of New England Journal of Medicine featured a study from Australia and New Zealand regarding the timing of dialysis treatment. The standard of care became earlier initiation of dialysis in the last few years, even though the evidence to support it remained thin and contradictory.
Initiating Dialysis Early and Late (IDEAL; I want to make up clinical trial names for a living) randomly assigned adults with progressive chronic kidney disease and no living kidney donor to start dialysis early (estimated glomerular filtration rate 10-14 mL/min) or late (eGFR 5-7 mL/min). The late group could start earlier if deemed necessary by the treating physician without prior permission from the data coordinators. The primary outcome was death, with a variety of secondary outcomes examined as well.
2982 patients underwent screening for the study, and eventually 828 were randomized. 404 were assigned to the early group, although only 383 actually began dialysis. 424 were randomized to the late group, and 386 of them actually began dialysis. The groups were similar for all baseline characteristics, and the rate of death and all secondary endpoints showed no differences, including a quality of life instrument.
What was different? The early group began dialysis sooner, as planned, a median of 1.8 months (95% confidence interval 1.60-2.23) after randomization. The late group began dialysis 7.4 months (95% CI 6.23-8.27) after entering the study. The eGFR at initiation of dialysis averaged 12 mL/min in the early start group, while the late starters began at 9.8 mL/min. In the early starters, 18.6% had eGFR <10 mL/min at the start. In the late start group, the majority (76%) started dialysis with eGFR >7 mL/min, presumably due to uremic symptoms.
Why do we care about this study? First, dialysis carries risks and expenses. In any form, a dialysis patient has risks related to their access and procedures that could result in sickness or death. A year of dialysis is also quite expensive; an extra 5-6 months off of therapy represents substantial savings of about $30,000 per person in the United States, as shown in the US Renal Data Service graph to the lower right.
The authors and the media focused on the finding that earlier initiation of dialysis (before uremic symptoms) did not produce benefits; however, most patients assigned to a late start were unable to get to the goal range of eGFR. In other words, uremia becomes symptomatic between 10 and 12 mL/min of eGFR in most patients, something they taught me back in fellowship around 1990. Also, the difference between eGFR of 12 and 10 mL/min seems, well, trivial. They also do not report an analysis of the patients who waited until eGFR of 5-7 mL/min. I would love to know about outcomes in the 102 patients who were able to wait to this low range before starting therapy.
I would also like to point out that children were not included in this study. Patients in both groups averaged 60 years of age. Infants and children with chronic kidney disease often have issues with growth and development that make timing of dialysis somewhat different.
It is wonderful to have a prospective randomized study about such a basic but important clinical question.
So in an IDEAL world (heh heh, sorry), what should our approach to dialysis initiation be? I propose that planning for renal replacement therapy should begin when the eGFR nears 20 mL/min. At that point we know the end is near. Potential living kidney donors can be identified for preemptive surgery. The choice of dialysis modality can be discussed. Appropriate access can be obtained in a non-emergent manner. That way, when uremia becomes symptomatic, the patient and nephrologist are ready for action.