Frequent Nocturnal Hemodialysis—A Step Forward?

Dialysis treatments sustain the lives of nearly 400 000 US residents with end-stage renal disease (ESRD).¹ The mortality rate for dialysis patients is approximately 10 times that of the general population,^{2 - 3} and each year more than 20% of dialysis patients die.¹ These statistics have hardly changed in the past decade. Morbidity also remains high: dialysis patients frequently also have cardiovascular disease, anemia, bone disease, poor nutrition, inflammation, depression, and physical and cognitive impairment.

Even though ESRD is a worldwide problem, outcomes differ in various regions. For instance, survival for patients receiving dialysis is better in Japan and Europe than in the United States.⁴ Are ESRD patients in the United States older or sicker than their Japanese or European counterparts? Goodkin et al⁴ reported that when survival was adjusted for patient demographics and comorbidities, the mortality differences narrowed, but persisted. Might genetic factors or geographic differences explain the disparities? In a study of European dialysis patients, van Dijk et al⁵ found that 26% of mortality differences between northern and southern countries could be attributed to differences in general population mortality. Do differences in dialysis practices or medical care account for the disparity? Nearly a decade ago, McClellan et al⁶ showed that mortality is associated with facility-to-facility differences in the dose of hemodialysis. Does body size also play a role? Observational studies have suggested that dialysis dose and body mass index are strongly associated with survival in hemodialysis patients.^{7 - 8}

One reasonable hypothesis for the survival differences is that US patients have not received an adequate dialysis dose. Two prospective randomized controlled trials (RCTs) examined whether higher dialysis doses improved survival. The HEMO study enrolled 1846 patients in a 2 × 2 factorial design comparing high- vs low-dose and high- vs low-flux thrice-weekly dialysis.⁹ This 5½-year multicenter study demonstrated no change in survival for patients receiving the higher-dose dialysis. In the Mexico-based ADEMEX trial,^{10 - 11} higher peritoneal clearance rates failed to improve survival or quality of life (QoL) in peritoneal dialysis patients.

Retrospective analysis of data from the HEMO trial study, which compared different doses of dialysis within the confines of conventional thrice-weekly hemodialysis, showed that the higher dose delivered only 16% more urea removal than the lower dose.¹² This is expected because the efficiency of removal of small-molecule solute during hemodialysis declines during the course of any single dialysis treatment. Observational studies of more frequent hemodialysis, which increases small-molecule clearance by 50% to more than 100%, suggested improvements in cardiovascular and QoL measures.^{13 - 14} A systematic review of published observational studies of frequent nocturnal hemodialysis reported improved blood pressure control, anemia, and health-related QoL, with mixed results for changes in left ventricular hypertrophy and mineral metabolism.¹⁵

In this issue of JAMA, Culleton and colleagues¹⁶ report findings from a Canadian RCT comparing nocturnal hemodialysis performed 5 to 6 nights per week for a minimum of 6 hours per night vs conventional thrice-weekly hemodialysis. The primary outcome, change in left ventricular (LV) mass over 6 months of study, showed an impressive result. Among the 44 patients who had baseline magnetic resonance imaging, LV mass decreased a mean (SD) of 13.8 (23.0) g in the nocturnal hemodialysis group and increased 1.5 (24.0) g in the conventional dialysis group (P = .04). This improvement in left ventricular hypertrophy (LVH) was accompanied by reduced blood pressure and use of fewer antihypertensive medications, as well as lower parathyroid hormone levels and serum phosphate levels.

However, extracellular fluid volume and the effect of nocturnal hemodialysis on volume status were not measured. The authors acknowledge that they were unable to determine if improved LV mass and blood pressure were a result of differences in volume control. The primary QoL measure did not change, but unexpectedly, the QoL scores decreased among patients receiving standard thrice-weekly hemodialysis and remained essentially unchanged in nocturnal hemodialysis patients. It is difficult to understand this finding because study participants were stable, long-time dialysis patients (time receiving dialysis in the conventional hemodialysis group was a mean [SD] of 4.8 [3.8] years) studied over a 6-month interval. It is possible that this observation is spurious; the authors point out that the small sample size probably meant that the study was underpowered to detect clinically significant differences in QoL measures.

Why was LV mass chosen as the primary outcome rather than mortality? Six years ago, Chertow¹⁷ called for an RCT of frequent hemodialysis, urging that the primary outcome should be mortality alone, or mortality combined with a major morbid event. However, studies examining mortality would require large numbers of recruited participants. A power analysis performed for a similar study showed that enrollment of more than 5000 participants would be needed to achieve 90% power to detect a 30% reduction in mortality.¹⁸ Recruiting thousands of dialysis patients for such a study is not feasible.

Cardiology RCTs abound in which thousands of participants are enrolled and hard end points like mortality and hospitalization rates are examined. Why not do the same for kidney disease or dialysis patients? Himmelfarb¹⁹ recently discussed the rich epidemiologic investigations in nephrology, but a distinct paucity of interventional trials, perhaps fewer than in all other specialties of internal medicine. Dialysis patients have frequent and longitudinal medical follow-up, with a relatively small number of nephrologists, and so should be relatively easy to enroll in RCTs. However, enrolling such patients has proven difficult. Dialysis patients often have multisystem problems and spend many hours each week attached to dialysis machines, cyclers, or dialysate bags. These factors may make participation in RCTs more burdensome than for patients who are not undergoing dialysis.

Over the many years of recruitment in HEMO, fewer than 1900 patients were randomized into a study in which both the high- and low-dose groups continued receiving conventional thrice-weekly dialysis. Enrolling patients in a study in which the experimental group requires 5 to 6 overnight treatments each week, a virtual doubling of weekly dialysis days compared with conventional dialysis, is a formidable task. The study by Culleton et al¹⁶ enrolled only 52 patients, clearly insufficient to examine mortality. Nonetheless, changes in LV mass were significant. Cohort studies in dialysis patients suggest correlation between LVH and mortality²⁰ and improved survival when LVH regresses.²¹

Studies involving dialysis patients often have a brief window of opportunity to succeed. The Veterans Administration multicenter RCT comparing mortality and morbidity of hemodialysis patients vs intermittent peritoneal dialysis patients enrolled 114 participants and showed no difference in survival.²² However, by the end of the study intermittent peritoneal dialysis had largely been abandoned in favor of continuous ambulatory peritoneal dialysis. The nocturnal hemodialysis study by Culleton et al¹⁶ used dialysis machines that are similar to other standard hemodialysis machines used elsewhere. In the United States, many home dialysis programs are now using different dialysis machines that are small, easy to use at home, and are designed to deliver longer, slower-flow dialysis. However, the dialysis characteristics of these smaller machines are sufficiently different from conventional machines such that the current study results may not be transferable.

The RCT by Culleton et al¹⁶ is important for nephrology, clearly demonstrating reduced LVH with nocturnal hemodialysis. It would be interesting to see the effect of nocturnal hemodialysis on cardiac structure and function beyond the 6-month study period examined. While future studies may provide additional information,²³ the RCT by Culleton et al¹⁶ suggests that nocturnal hemodialysis may help improve the high morbidity and mortality of North American dialysis patients.