Has the Time Now Come to More Widely Accept Hemodiafiltration in the United States?

In this issue of JASN, the Spanish hemodiafiltration study is published.1 What new information has become available and what might be its relevance? Today, most patients with ESRD in the United States are treated with hemodialysis (HD). During low-flux HD, small uremic toxins are removed by diffusion, whereas larger solutes are retained within the body. When a high-flux dialyzer is used, the higher membrane permeability, and a certain amount of convection, increases the overall clearance of uremic molecules. Convection is created by ultrafiltration; in high-flux HD, the amount of ultrafiltration is uncontrollable, immeasurable, and unpredictable. The volume of ultrafiltration generally exceeds the desired weight loss, and back-filtration automatically compensates for the excess. In hemodiafiltration (HDF), diffusion and convective transport are also combined. With HDF, convective transport is obtained by filtering, through a high-flux dialyzer, amounts of plasma water considerably in excess of those required to achieve dry weight. Fluid balance is maintained by simultaneously infusing sterile substitution fluid directly into the patient’s bloodstream. The substitution fluid can be administered before (predilution), within (mid-dilution), or after (postdilution) the dialyzer. Clearance of middle- and large-molecular-weight substances is substantially greater during HDF than during high-flux HD. When HDF was introduced in the late 1970s, substitution fluid was supplied in bags, which made HDF an expensive and labor-intensive procedure and limited the magnitude of the infusion volumes and thus of convection. Today, sterile substitution fluid is prepared online, thus enabling production of large volumes of substitution fluid at an acceptable cost in everyday clinical practice.2 Increased interest in HDF was triggered by analysis of European Dialysis Outcomes and Practice Patterns Study (DOPPS) data.3 This retrospective study suggests that use of HDF confers a survival benefit compared with HD. The data also suggest that a certain minimum convection volume is necessary to obtain this benefit. These findings, combined with other observational studies suggesting a survival benefit, provided the rationale for prospective randomized trials comparing the effects of HDF with standard HD on relevant clinical endpoints.4 The Spanish HDF study reported here by Maduell et al.1 is the third randomized controlled trial to be published. It adds very relevant new information to the data already available from the two earlier studies.5,6 The first two studies found no difference between HDF and HD in the primary endpoint (all-cause mortality in CONTRAST [Convective Transport Study]5 and a composite of all-cause mortality and first cardiovascular events in the Turkish study6). In both studies, the convection volume actually achieved varied considerably. In both studies, secondary analysis suggests a dose-effect relationship, in that above a certain convection volume, application of HDF is associated with improved survival, even after extensive correction for confounding factors.7 In the study reported in this issue, the primary endpoint is also all-cause mortality. However, this study differs from the two earlier studies in that Maduell et al. paid specific attention to achieving higher convection volumes than previously.1 Their primary analysis showed survival benefit in the HDF group, whereas secondary analysis again showed a dose-effect relationship: that is, the patients in the highest two tertiles of achieved convection volumes were more likely to experience the benefit. So, the present study confirms and extends the suggestion of earlier studies. Altogether, several observational studies and two randomized controlled trials suggest,5,6 and one RCT clearly shows,1 that application of HDF could really be advantageous to our patients. Given that all three randomized trials suggest the presence of a dose-effect relationship, it is important to know more about the factors that control the amount of convection volume. This is discussed in more detail elsewhere.8,9 Are there any concerns with applying HDF? Although safety was not a specifically defined endpoint, the three trials do not indicate that HDF is unsafe. In particular, microbiological safety during HDF is a matter of concern because large volumes of online-produced fluid are directly infused into the patient. In the first two randomized trials, markers for inflammation did not differ between the HDF and the control groups.5,6 Furthermore, the CONTRAST investigators reported earlier that it is indeed possible to produce substitution fluid of adequate quality over a prolonged period.10 An often-heard concern is that HDF would be more expensive. Indeed, formal cost-effectiveness analysis of CONTRAST showed that HDF is approximately 3% more expensive than standard HD.11 The difference is mainly explained by more expensive disposables (based on 2009 price levels) and more frequent analysis of the quality of the dialysate/substitution fluid. Where do we go from here? Currently, HDF is mainly used in Europe and, to a lesser extent, in Asia and Canada; it is almost never used in the United States.12 The European Renal Association–European Dialysis and Transplant Association has instituted an official working group (EuDial [http://www.era-edta.org/eudial/ European_Dialysis_Working_Group.html]), which focuses mainly on HDF. This working group has produced one paper summarizing definitions, dose quantification, and safety; a second paper on clinical evidence will be released shortly.2 Several important aspects of HDF have not been addressed so far. The above-mentioned randomized trials used the traditional schedule of three treatments per week, thus addressing the question of whether HDF can provide a better outcome in that logistic and organizational infrastructure. It is unclear what value HDF could add to more frequent dialysis schedules.13 Furthermore, published data on HDF in children are very limited. In this group of patients, a totally different endpoint, growth acceleration, could be of great relevance. Finally, it is possible that specific subgroups of patients would especially benefit from HDF. Although the three available trials were unable to identify such groups, combining the individual data from the three studies might allow such groups to be defined. These and other issues are subjects to be addressed by EuDial. Has the time now come for HDF to be more widely accepted in the United States? Asking that question inevitably gives rise to a consideration of the barriers that have prevented its use up until today. In conversations with colleagues in the United States on this subject, several potentially relevant issues come up. First, until very recently, no Food and Drug Administration (FDA)–approved machine was capable of performing HDF. Is this because companies have not attempted to acquire approval? No. Rather, it seems that some years ago one of the major European dialysis companies initiated the process with one of their machines, which was widely used in Europe and other parts of the world. However, final approval was never obtained. Was this because the FDA process was too demanding? Colleagues in the United States tell me that the issue of fluid quality was of paramount importance or, perhaps more precisely, that there was doubt that optimal fluid quality could be guaranteed over a prolonged period. Fluid quality is indeed of great importance. However, it is important to realize that high-flux HD, which is widely accepted as first-choice therapy in the United States, is a form of low-dose HDF (as briefly explained above). So, strictly speaking, quality requirements for HDF should not differ from those for high-flux HD. Put another way, the same high-quality fluid levels defined for HDF should also be standard for high-flux HD.2 Thus, conceptually, it seems incorrect to demand stricter quality levels for HDF than for high-flux HD. Second, interest in HDF within the nephrology community in the United States seems to be low, with a few exceptions. It is assumed that the treatment is more troublesome, more costly in terms of both machines and disposables, and of little real additive value. However, many modern dialysis machines can deliver online HDF fairly easy. Indeed, no data are available on cost-effectiveness of HDF in the United States. Altogether, it seems reasonable to conclude that the results obtained with the three recent randomized controlled trials should be sufficient reason for the nephrology community, the regulatory authorities, and the large dialysis companies to sit down together and reconsider the present position of HDF in the United States. That could benefit patients with ESRD all over the globe. Disclosures P.J.B. received research funding for hemodiafiltration-related studies from the Dutch Kidney Foundation (Nierstichting Nederland), Gambro, Baxter, Roche, and Fresenius and speakers fees from Fresenius and Gambro.