Can frequent hemodialysis be too frequent?

Many nephrologists believe that more hemodialysis, whether in the form of longer treatments, more frequent treatments, or both, is better than less. Emerging evidence largely supports this view, with most, although not all studies, showing biochemical, cardiovascular, quality of life, or survival benefits with short daily or slow nocturnal hemodialysis. Much of the focus of studies thus far has been on outcomes expected to improve with intensive dialysis such as BP and phosphorous control. In this issue of JASN, Suri and colleagues provide detailed analyses of vascular access outcomes from the two Frequent Hemodialysis Network (FHN) trials, the largest randomized controlled trials of frequent maintenance hemodialysis. For both trials, vascular access events were prespecified safety outcomes because of concerns that potential benefits of more intensive dialysis might come at the cost of an increase in vascular access complications. The findings of these analyses suggest that the concerns were warranted. The FHN trials randomly assigned 245 patients to incenter hemodialysis at a frequency of six times per week or three times per week in the “daily” trial, and 87 patients to nocturnal hemodialysis six times per week or to conventional hemodialysis three times per week, performed at home for most participants, in the “nocturnal” trial. As previously reported, a benefit of frequent dialysis on the two coprimary outcomes of death or change in left ventricular mass and death or change in self-reported physical health was found in the daily trial but not in the smaller nocturnal trial. For both trials, the primary vascular access outcome was the time to first access event (access repair, access loss, or access-related hospitalization). As reported by Suri and colleagues, in the daily trial, the hazard ratio (HR) for a first access event was 1.76 (95% confidence interval [95% CI], 1.17, 2.79; P50.02) with dialysis six times per week compared with dialysis three times per week. Although not statistically significant, a trend toward increased risk of an access event with dialysis six times per week was also evident in the nocturnal trial (HR, 1.81; 95%CI, 0.94, 3.48; P50.08). When arteriovenous (AV) accesses (fistulas and grafts) and tunneled dialysis catheters were analyzed separately, in both the daily and nocturnal trials the risk of an AV access event was greater with more frequent dialysis, with a HR of 1.90 (95% CI, 1.11–3.25; P50.02) in the daily study and a HR of 3.23 (95% CI 1.07–10.35; P50.04) in the nocturnal trial. In the daily trial, there was also a statistically nonsignificant trend toward a greater risk of a first catheter event (HR, 2.70; 95% CI, 0.71–10.2; P50.14) with frequent dialysis. Attempts to analyze grafts and fistulas separately were limited by small numbers, but the authors report a 2.2-fold increase in the rate of interventions for grafts with dialysis six times per week in the daily trial, without an increased rate for fistulas. In contrast, in the nocturnal trial, a trend toward a higher rate of fistula, but not graft, interventions was seen in the nocturnal group receiving dialysis six times per week. When the components of the composite vascular access outcome were evaluated individually, it was apparent that in both trials, the higher risks for AVaccess events with frequent dialysis were driven by access interventions (angioplasty, stenting, thrombectomy, and surgical revision) rather than losses (access abandonment or removal). It is notable that one patient in each treatment group of the daily trial (0.8%) died from AVaccess hemorrhage and one patient in the nocturnal trial (nocturnal arm) died from a catheter-associated air embolism; these risks of hemodialysis are probably underappreciated. How convinced should we be from the findings of Suri et al. that frequent dialysis increases the risk of vascular access complications? Other than access type, no information about the vascular accesses is provided, leaving open the possibility of imbalances between treatment arms in vascular access characteristics such as access age or number of previous repairs that might underlie differences in outcomes. As the authors point out, it is unlikely that the findings of increased risk are due simply to increased surveillance for access complications among the patients undergoing dialysis six times per week. In the frequent dialysis arm of the daily trial, thrombectomies comprised a large proportion of the access events; access thrombosis is an outcome that should not increase as a result of more intensive observation. Because vascular access outcomes were among many secondary outcomes for these trials, it is also possible that any individual statistically significant relationship is a chance Published online ahead of print. Publication date available at www.jasn.org.