Hemodialysis Prescription for Incident Patients: Twice Seems Nice, But Is It Incremental?

During the next decade, about one million people in the United States are expected to start dialysis therapy.¹ The majority of these people will be treated with a fixed-dose (single-pool Kt/V_urea [spKt/V_urea] ≥ 1.2) thrice-weekly hemodialysis regimen irrespective of whether they are just starting dialysis therapy (incident) or have been receiving dialysis for some time (prevalent) and without consideration for residual kidney function (RKF). Although the regulatory agencies might consider this hemodialysis regimen as “standard of care” and “adequate,” it is by no means perfect.^2–4 Therefore, questions remain. What is the optimal regimen for incident hemodialysis patients? How did we get to the current fixed-dose thrice-weekly paradigm? How do we move forward?

The path to thrice-weekly hemodialysis was gradual, guided by dialysis prescribed to control patient symptoms, peripheral neuropathy in particular.⁵ In the 1960s, sessions took place once weekly, lasting 20 to 24 hours. Due to recurrent uremic symptoms, frequency was increased to twice weekly for 16 to 23 hours per session, and later to thrice weekly for 8 to 10 hours per session.⁶ In 1973, the Seattle thrice-weekly 8- to 10-hour per session hemodialysis program was the norm when the Medicare End-Stage Renal Disease Program legislation was passed. Subsequently, attempts to individualize hemodialysis prescriptions focused only on prevalent patients dialyzed thrice weekly. The National Cooperative Dialysis Study (NCDS; 1978–1981) led to the eventual development of Kt/V_urea as a patient-specific metric.^7,8 The Hemodialysis (HEMO) Study (1995–2001) failed to demonstrate a survival benefit from a higher spKt/V_urea (1.7) compared to standard dose (1.3) delivered thrice weekly.⁹ Thus, the thrice-weekly schedule stems from observations during the 1960s and the per-session dialysis dose (spKt/V_urea) from the NCDS and HEMO Study. However, it is important to note that the NCDS and HEMO Study findings are only generalizable to prevalent patients without RKF because incident patients and patients with RKF were explicitly excluded from both trials.

The optimal regimen for patients initiating hemodialysis therapy is not known. It is plausible that the routine practice of fixed-dose thrice-weekly hemodialysis in incident patients with substantial RKF may be harmful, contributing to accelerated loss of this residual function.^10,11 Loss of RKF is linked to decreased survival,^12,13 likely from poorer uremic solute clearance,¹⁴ volume and blood pressure control,^15,16 higher erythropoietin requirements,¹² more inflammation,¹² and higher left ventricular mass.¹⁷ RKF is routinely incorporated into peritoneal dialysis dose calculations. A similar approach to prescribing hemodialysis for incident patients is to individualize the prescription to the patient’s small-solute (urea) clearance needs, incrementally increasing hemodialysis clearance (dose and/or frequency) as RKF declines.¹⁸ This incremental dialysis approach is supported by opinion-based US and European guidelines,^19,20 but its safety and efficacy in incident hemodialysis patients remains undefined.

In this issue of AJKD, Obi et al²¹ report results of a retrospective cohort study of incident in-center hemodialysis patients comparing twice-weekly hemodialysis prescribed to 351 patients to thrice-weekly hemodialysis prescribed to a matched cohort of 8,068 patients. Of note, only 0.3% (351/103,703) of all incident patients were treated with twice-weekly hemodialysis, limiting the generalizability of the findings. As displayed in Fig 2A of the report, the slope of decline in urinary urea clearance (KRU) for the thrice-weekly group from the first to the second patient-quarter was ~ 1 mL/min/1.73 m², quite a bit steeper than the decline of ~ 0.25 mL/min/1.73 m² for the twice-weekly group. However, there was significant variability, with wide confidence intervals. After year 1, the risk for death with twice-weekly hemodialysis was statistically similar to thrice-weekly hemodialysis (P = 0.3), but with wide confidence intervals ranging from 11% lower to 38% higher risk for death. In subgroup analyses, patients on a twice-weekly regimen with a baseline KRU ≤ 3 mL/min/1.73 m² (n = 73) had a trend toward a higher risk for death compared with patients with KRU > 3 mL/min/1.73 m². The authors concluded that twice-weekly hemodialysis may be safe in incident hemodialysis patients with substantial RKF at baseline; thus, they called for a clinical trial to examine the safety and efficacy of twice-weekly hemodialysis.

Overall, the study is thoughtfully designed within the confines of an administrative data set. However, the key question is whether twice-weekly hemodialysis, as prescribed in routine care, was truly incremental hemodialysis that is adjusting dialysis clearance to supplement RKF. Or did a fixed-dose twice-weekly regimen simply supplant the fixed-dose thrice-weekly schedule without consideration of RKF, perhaps driven by factors such as patients’ wishes or health? Looking at Table S2 in the report, dialysis frequency, duration, and treatment time in the twice-weekly group remained generally similar from patient-quarters 1 to 4. Results for per-session spKt/V_urea are not reported. Of the 351 patients in the twice-weekly group, repeat urine measurements were available for only 24%, 63%, 30%, and 42% of patients in patient-quarters 2 to 5, respectively. The absence of increasing dialysis frequency or treatment time and infrequent urine monitoring implies that these patients were simply on a fixed-dose twice-weekly regimen and not an individualized incremental hemodialysis schedule. It might also explain why only 30% of the patients with KRU ≤ 3 mL/min/1.73 m² met the NKF-KDOQI (National Kidney Foundation–Kidney Disease Outcomes Quality Initiative) adequacy goal. Hence, although it is possible to conclude from this study that a fixed-dose twice-weekly regimen can only be safely undertaken in patients with KRU > 3 mL/min/1.73 m², this conclusion cannot be extended to incremental dialysis, which may be safe in patients with lower RKF levels.

Other studies of fixed-dose twice-weekly hemodialysis, generally prescribed in a resource-constrained environment, also suggest better preservation of RKF, as well as less intradialytic hypotension and similar nutritional indexes and survival.^22–33 Reports of incremental hemodialysis are far fewer, but suggest that this approach may be safe in patients with KRU as low as 1 mL/min¹⁵ and may improve survival in incident dialysis patients.³⁴ The results of the study by Obi et al are generally consistent with these prior studies. Additionally, less frequent dialysis may reduce access complications and risk for infections, potentially improving quality of life and reducing cost.

The 2 approaches, fixed-dose twice weekly and incremental hemodialysis, are not mutually exclusive. Extensive work on urea kinetics over the past few decades, including the development of the dialysis “equivalent renal urea clearance (EKR_C)” by Casino and Lopez³⁵ and weekly standard Kt/V_urea proposed by Gotch³⁶ and refined by Daugirdas et al,³⁷ provide the framework for combining dialysis and residual urea clearances. Figure 1, adapted from an approach described by Keshaviah et al,³⁸ illustrates this concept for incremental hemodialysis. The mathematical complexity of these approaches has limited their widespread application; however, increasing availability of computers at the point of care may simplify their implementation.

Figure 1.

Incremental hemodialysis prescription with adjustment of hemodialysis dose based on residual kidney function. Exponential decline in residual urea clearance over time is shown as a solid line and light gray shaded area. Incremental hemodialysis clearance (broken line and dark gray area) is calculated to achieve a total (residual + dialysis) standard Kt/V_urea (stdKt/V_UREA) of 2.1. Dialysis frequency can be increased from once weekly to thrice weekly, as needed, to achieve the target total stdKt/V_urea. Per-treatment single-pool Kt/V_urea (spKt/V_UREA) can be calculated to achieve the target dialysis stdKt/V_urea at each time point. Modeling assumptions include total-body water of 40 L, minimum treatment time per hemodialysis session of 2.5 hours, and ultrafiltration volume of 1 L per treatment. Note: This is a hypothetical scenario; individual patients’ hemodialysis regimens must also consider uremic symptoms, hyperkalemia, and volume management, among other parameters.

A key factor limiting the feasibility of incremental hemodialysis is the requirement for frequent and reliable assessment of RKF by timed urine collections, which are unreliable and burdensome.¹⁹ Glomerular filtration rate measurement using plasma clearance of iohexol can be considered for assessing RKF, but it is also cumbersome, is not feasible on a large scale, and may give erroneous results due to nonrenal iohexol clearance.³⁹ We recently developed and validated equations that estimate KRU in dialysis patients from serum β-trace protein (BTP), β₂-microglobulin, or cystatin C concentration without requiring urine collection.⁴⁰ These equations (available at www.kidneymodels.org/rkf/) are valid for patients with self-reported urine volume of at least 1 cup (250 mL) per day. In particular, the BTP equation seems attractive given that BTP remains in a steady state during the interdialytic interval.⁴⁰ Routine estimation of RKF using serum markers may simplify incremental hemodialysis prescription, allowing frequent monitoring of RKF without increasing patient burden.

However, we must also consider the limitations of an incremental dialysis approach.⁴¹ Refractory volume overload is often an indication for dialysis initiation, but volume status in dialysis patients is difficult to assess and there is no gold standard for volume assessment.⁴² Volume overload may require more frequent or longer dialysis sessions. Additionally, the native kidneys’ removal of low-molecular-weight proteins and protein-bound solutes by tubular clearance and removal of sequestered solutes by virtue of the slow continuous clearance are not matched by dialysis. Therefore, we must be cognizant that although combining urea clearance from dialysis and RKF is mathematically feasible, the nature of solute clearance provided by dialysis and RKF is fundamentally different.

We agree with the authors’ call for a clinical trial. We would like to stress that the only way to improve the care of more than a million patients who will start hemodialysis therapy in the next decade is by challenging the existing paradigm and striving for evidence-based changes in clinical practice. The equipoise between the use of incremental hemodialysis or usual-care fixed-dose thrice-weekly hemodialysis for incident patients can only be answered by a clinical trial. Such a trial could have a cluster-randomized pragmatic non-inferiority design, using RKF estimation instead of urine collections to enhance the feasibility of implementation and safety of the participants. The potential benefits and economic savings make the reasons to pursue such an endeavor evident.