Home hemodialysis: Present state of the evidence


  • Mitchell H. Rosner MD

    1. Associate Professor of Medicine, Department of Medicine, Division of Nephrology at the University of Virginia Health System in Charlottesville, Virginia
    Search for more papers by this author

Hemodialysis began in the 1960s as a means to treat patients with acute renal failure, but evolved into a treatment for those with end-stage renal disease (ESRD) following the development of the Scribner/Quinton shunt, which permitted repetitive dialysis to be performed over prolonged periods of time.

Initial hemodialysis sessions were long; the first dialysis session on the first chronic hemodialysis patient, Clyde Shields, lasted 76 hours. Obviously, such prolonged sessions were impractical and treatment times and frequency were examined to determine the best balance between maintenance of health, patient convenience, and cost.

As the demand for therapy increased in the 1970s, coincident with the passage of legislation funding hemodialysis, dialysis time was shortened, sometimes to 4 hours. This was a practice supported by the National Cooperative Dialysis Study (NCDS), which demonstrated that the time-averaged concentration of urea was the most important determinant of patient morbidity or withdrawal from the study.1 While time on dialysis was also studied, longer times were not associated with improvement in the rate of hospitalization, withdrawal from the study, or death (p = .06). Following release of the NCDS study, clinicians focused on attaining adequate small molecule clearance and lowered treatment time as increasingly larger, more permeable, high efficiency dialyzers were developed that could achieve solute clearance goals (typically a target Kt/V of 1.2 or greater) in shorter times.2

The role of increased dialysis dose with a Kt/V > 1.2 or the use of high flux dialyzers was studied in the HEMO trial.3 This trial showed that a high (single pool Kt/V of 1.71, equilibrated Kt/V of 1.53) compared with a standard dose of Kt/V (single pool Kt/V of 1.32, equilibrated Kt/V of 1.16) or use of a lowflux compared with a high-flux membrane made no clear difference on patient outcomes. These disappointing results suggested that the high mortality associated with hemodialysis could not be impacted upon with relatively small increases in solute clearance. It also further questioned the rationale of focusing on Kt/V as a measure of dialysis adequacy.

The mortality rate for hemodialysis patients in the U.S. is high at approximately 24.4 deaths/100 patient-years.4 The high mortality rate has persisted despite progressive increases in Kt/V over the past two decades, from 1.11 in 1991 to 1.52 in 2002.5 In addition, despite this increase in Kt/V, the complications of hypertension,6 malnutrition,7 congestive heart failure,8 and bone and mineral disorders9 have remained unacceptably high. These outcome data clearly indicate that dialysis adequacy cannot be simply measured in terms of Kt/V.

Given these persistently poor outcomes in patients on chronic dialysis, there is ongoing interest about whether modifications of standard hemodialysis regimens, such as increasing the time, dose, and/or frequency of hemodialysis, can improve outcomes. Many believe that at least part of this morbidity can be attributed to the non-physiologic nature of the conventional thrice-weekly hemodialysis schedule. The intermittency of thrice-weekly hemodialysis permits large fluctuations in the levels of uremic toxins and in extracellular volume. Such imbalances may be particularly hazardous in patients with underlying cardiomyopathy, cardiac arrhythmias, and coronary disease. These comorbidities are especially prevalent in patients with diabetes mellitus, the number-one cause of ESRD in the U.S.4 These drawbacks in current conventional dialysis, as well as the high mortality rate of this population, have inspired clinicians to determine whether more frequent renal replacement therapy can achieve better body homeostasis and improved elimination of toxins and, as a result, better outcomes.

A number of alternative dialysis strategies such as short daily hemodialysis (SDHD) and long nocturnal hemodialysis (NHD) are being actively investigated (terminology is defined in TableI). TableII lists the typical treatment parameters for each modality, including SDHD, using the most prevalent system in the U.S. at this time, the NxStage System One Machine. At this time, the National Institutes of Health (NIH) is sponsoring two studies of daily dialysis, one short daily (2–3 hours, 5 or more days weekly) and the other long nightly (6–8 hours, 5 or more days weekly).10 The results from these studies will not be known for a number of years.

Table I. Terminology.
  • Conventional hemodialysis (CHD): intermittent hemodialysis (IHD) performed in a dialysis center for 3 to 5 hours per session, 3 times weekly

  • Quotidian dialysis: daily (or more than 3 times weekly) hemodialysis treatments that can be performed as:

    • Nocturnal hemodialysis (NHD): performed while a patient sleeps for sessions lasting as long as 8 to 9 hours

    • Short daily hemodialysis (SDHD): performed daily but with a shortened duration of 2 to 3 hours

  • Long intermittent dialysis (LIHD): includes either nocturnal or hemeral (daytime) sessions that are long (6 to 9 hrs) but performed 3 sessions weekly

Table II. Parameters for hemodialysis modalities.
ModalityTreatments per weekTreatment Time (hrs)Blood Flow Rate (mL/min)Dialysate Flow Rate (mL/min)
  1. Abbreviations: CHD, conventional hemodialysis; NHD, nocturnal hemodialysis; SDHD, short daily hemodialysis.

NxStage SDHD5–62–3400130

While these studies are being performed, there is a movement in the U.S. (as well as the world) to increase access to more frequent dialysis for many of their patients who may benefit from such therapies. This movement is fueled by a series of studies demonstrating benefit, in terms of specific clinical parameters, when the intensity of dialysis therapy is increased (either by more frequent sessions, longer sessions, or both). While these studies are of variable size and quality, there is a remarkable consistency in outcomes that demonstrates that more frequent dialysis provides superior control of many parameters and offers hope for better outcomes.

Effects on Solute Removal

Given the differences between more frequent and intermittent therapies, it is important to have a measure of dialysis dose that can be interpreted across different modalities. The measure most often cited is the standardized Kt/V (stdKt/V).11 As formulated by Gotch, this value is calculated based upon the mid-week pre-dialysis blood urea (BUN) level. Using this method, equivalent doses of dialysis can be prescribed for different modalities and treatment frequencies. Using this measure, conventional hemodialysis (CHD) and peritoneal dialysis (PD) have a weekly stdKt/V of roughly 1.7 to 2.0 (corresponds to a single session intermittent HD Kt/V of 1.2). Nocturnal hemodialysis typically achieves weekly stdKt/V of 4 to 5 (with a single session Kt/V of 1.8–2.5). Finally, SDHD typically targets a weekly stdKt/V of 2.0 (with a corresponding single session Kt/V of 0.5).

An important difference with more frequent dialysis is its effects on the removal of middle molecular weight molecules (such as β-2 microglobulin) that are not accounted for by urea kinetics. With both increased frequency of dialysis as well as longer sessions, removal of these potential toxins is increased. For example, the weekly dialysate β-2 microglobulin mass clearance increased from 127 to 585 mg when the patient was switched from CHD to NHD.12 Furthermore, removal of protein-bound substances such as indole-3-acetic acid and acid indoxy sulfate are increased on SDHD as compared with CHD.13

Clinical Benefit of More Frequent Hemodialysis

Numerous small trials have investigated the effects of more frequent (either SDHD or NHD) on clinical outcomes ranging from cardiovascular effects, nutritional effects, to impact on quality of life.

Cardiovascular Effects

Several studies (including 1 randomized controlled trial) have assessed the changes in cardiovascular parameters associated with more intensive dialysis therapies. The effects studied have included important surrogate outcome measures that in separate studies have been associated with improved mortality (i.e., improvement in cardiovascular mortality associated with improved blood pressure control or regression of left ventricular hypertrophy). No studies to date have investigated the effects of more intensive dialysis on cardiovascular mortality per se.

Multiple investigations performed in patients undergoing both SDHD as well as NHD have clearly demonstrated that blood pressure goals are more readily met by patients receiving dialysis through these modalities over those on CHD.14–19 While the mechanism of improved blood pressure control is uncertain, a study in patients undergoing SDHD showed improved control in extracellular fluid volume20 and a study in NHD patients demonstrated a decrease in peripheral vascular resistance and lower levels of circulating catecholamines.21

Four separate studies have demonstrated a reduction in left ventricular mass index (LVMI) as measured by either echocardiography or magnetic resonance imaging.14, 17, 18, 22 These studies all consist of CHD patients that were converted to either SDHD or NHD and then followed prospectively.

It is hoped that these beneficial effects on blood pressure, volume, and LV mass will translate into important improvements in mortality in these patients as has been true with other therapies that achieve these goals (such as anti-hypertensive medications).

Nutritional Effects

Two studies have demonstrated increases in appetite, weight gain, and increases in muscle mass when patients are converted to daily dialysis.23, 24 A study of nitrogen kinetics in patients on NHD revealed that, despite the large amounts of amino acids lost in the dialysate, there was no decline in total body nitrogen.25 However, studies looking at serum albumin levels have been conflicting, with several studies showing an improvement in serum albumin levels and others showing no effect.15

Mechanistically, the improvement in nutritional parameters may be secondary to improved appetite and the ability to liberalize the diet when patients are switched to either SDHD or NHD.26–28 There may also be an effect of more intensive dialysis to decrease the inflammatory milieu associated with ESRD as levels of interleukin (IL)-6 and C-reactive protein have been shown to decrease in one study of patients undergoing daily hemodialysis.29

Mineral Metabolism

Daily hemodialysis is associated with significant improvements in net phosphate removal. With NHD, phosphate removal is approximately twice that of CHD and many, if not the majority, of patients no longer require phosphate binders or dietary phosphorus restriction. In fact, many patients require supplementation of phosphate in the dialysate.14, 16, 30, 31 With SDHD, serum phosphate levels tend to fall when the daily sessions are longer than 2 hours and most of these patients still require phosphate binders.26, 32 Long-term studies on bone histology or fracture rates are not available.

Hematologic Effects

There has been conflicting data on whether intensification of hemodialysis is associated with either increases in serum hemoglobin or increased responsiveness to erythropoietin (EPO). For SDHD, one study has demonstrated a small (3%) rise in hematocrit with stable EPO dosages when patients were switched from CHD to SDHD.33 Another study demonstrated that there was a fall in EPO dosages by 45% along with a rise in serum hemoglobin levels in patients switched from CHD to SDHD.26 For NHD, one study has demonstrated both a fall in EPO dosages and a rise in hemoglobin levels,34 while another study could not demonstrate either change when patients were switched from CHD to NHD.14 Further work in this area is needed to better understand the effects of more frequent dialysis.

Effects on Sleep

It has been proposed that the high prevalence of sleep disorders in ESRD may reflect suboptimal dialysis and may impact quality of life as well as cardiovascular mortality.35 The data regarding the effects of more intensive hemodialysis on sleep disorders are limited to studies in NHD patients. The greatest effects of NHD on sleep disorders have been seen in patients with sleep apnea. After conversion from CHD to NHD, there was a significant reduction in the apneahypopnea index.36, 37

Quality of Life

Several studies have examined the changes in quality of life when patients make the switch from CHD to more intensive therapy.14, 23, 38–42 In this regard, only NHD patients have been studied. Despite receiving more dialysis with its attendant time and labor demands, the majority of studies have reported improved cognition, psychomotor efficiency, as well as improved quality of life parameters using several different survey instruments (such as the Beck Depression Index, SF-36, and Sickness Impact Profile)14, 33, 38, 40–42 One study utilizing the EuroQoL-5D index did not demonstrate a difference in overall quality of life between CHD and NHD patients.39 Interestingly, a cross-sectional study revealed that patients undergoing NHD did not perceive their therapy as being more intrusive than a similar cohort undergoing PD.40

It should be pointed out that these studies may be influenced by modality-selection bias, as healthier patients with better baseline quality of life may opt for NHD at a higher rate than CHD.

Hospitalization Rates

Only a few studies have looked at the difference in hospitalization rates between CHD and more intensive therapy. In one study, high-comorbidity patients with ESRD who were converted from CHD to SDHD while maintaining the same total weekly dialysis time were studied prospectively over 72 months. SDHD was associated with a significant 34% decrease in hospitalization days (note that there was no increase in vascular access hospitalizations).26 In another study, 32 NHD patients were studied 1 year before and 2 years after conversion to NHD and compared with 42 CHD patients (matched for age, dialysis vintage, and controlled for comorbidities) during the same time period.43 While hospitalization rates were stable for the CHD group, the group that was converted to NHD experienced a fall in dialysis or cardiovascular admission rates from 0.50 ± 0.15 to 0.17 ± 06 admission per patient year (p = .04). Cardiovascular disease (37%) was the principal cause for hospitalization in the control population. In comparison, vascular access-related admission was the primary cause of admission for the NHD cohort (56%).

Impact on Survival

Survival data for modern cohorts of patients receiving intensive dialysis is limited and, where available, has been criticized for the possibility of modality selection bias (i.e., those patients choosing intensive therapies, especially those choosing home therapies may be healthier).44 One study of 415 SDHD patients (combined American and European cohort) demonstrated a standardized mortality ratio of 0.34 (95% confidence interval [CI]: 0.20–0.54) as compared with a comparable United States Renal Data System (USRDS) cohort matched for age, sex, race, and type of renal disease.45

Survival data for NHD patients is limited to studies only available in abstract form. One of these studies compared the out-comes of 591 home hemodialysis patients (predominantly NHD) with conventional facility-based hemodialysis between 1996 and 2005.46 This study revealed a hazard ratio of 0.53 (95% CI: 0.45–0.61) for home-based therapy versus conventional therapy. Another study from Turkey prospectively followed 224 prevalent patients that were undergoing thrice-weekly 4 hour hemodialysis and were converted to 8 hour, thrice weekly in-center NHD.47 These patients were compared with age-, sex-, diabetes status-, and vintage-matched patients who continued on conventional hemodialysis. After 1 year of follow-up, there was an adjusted relative risk of death of 0.22 (95% CI: 0.06–0.76) that favored NHD.

Most recently, intensive dialysis out-comes have been compared with those of renal transplantation, the “gold-standard” of renal replacement therapies.48 There is one comparison of SDHD with transplantation that includes 274 daily dialysis patients compared with the 2004 USRDS deceased donor recipient survival statistics. In this comparison, the Kaplan-Meier survival curves are nearly identical.45 This analysis is limited by the lack of patient-specific matching and multivariable adjustments for patient characteristics.

The data comparing NHD with kidney transplantation is limited to one study of 171 Canadian NHD patients between 1994 and 2006 compared with living and deceased donor transplants in the USRDS.49 In this study, multivariate survival analysis was performed (adjusting for age, sex, body mass index, era, vintage of dialysis, history of cancer, ischemic heart disease, and peripheral vascular disease). This study demonstrated the superiority of living donor transplantation over NHD (hazard ratio [HR]: 0.39, 95% CI: 0.23–0.67). However, there was no survival advantage of deceased-donor transplantation over NHD (HR: 0.77, 95% CI: 0.46–1.30). More data regarding survival outcomes will be available from the NIH trial when completed.

Cost Effectiveness

While studies examining cost effectiveness of more frequent dialysis are limited, some generalizations can be made.50, 51 The expected reductions in staffing costs are seen when patients move to the home environment, but are balanced by higher costs for dialysis supplies needed to perform more frequent treatments. In some programs, the reductions in overhead associated with a reduced physical presence in the dialysis center may be offset by the costs of transportation and servicing the equipment needed to maintain a program spread over a large geographic area. The net effect is that much of the cost savings from performing dialysis in the home is lost when the frequency of dialysis is increased. Despite this, the studies to date suggest that quotidian home hemodialysis is less expensive than conventional in-center hemodialysis.

It is also important to realize that cost analysis such as these presume that patient outcomes with a new intervention are as good or better than with conventional care. If, however, the emerging data on the benefits of more intensive hemodialysis are integrated into this analysis, then more intensive dialysis may be considered a “dominant” therapy in that it is both less expensive and more effective than conventional in-center hemodialysis.


Given the benefits described, there are sub-groups of patients with ESRD that may be particularly good candidates for more intensive hemodialysis. These include:

  • Patients with poor quality of life on current renal replacement modalities;

  • Patients who want to work during the daytime;

  • Patients who need or want to liberalize their diet;

  • Patients who have disabling intradialytic or interdialytic complications;

    • Unstable blood pressure during dialysis

    • Severe cramping during dialysis

    • Uncontrollable hypertension

    • Impaired left ventricular function or congestive heart failure

    • Persistent hyperphosphatemia

    • Calciphylaxis

  • Severe sleep apnea;

  • Patients who wish to stay with a home therapy after transitioning from peritoneal dialysis;

  • Patients who may not be candidates for kidney transplantation;

  • Patients who have difficulty in controlling uremic symptoms:

    • Obese patients

    • Patients with poor vascular access blood flow.

More than these patients, having the ability to tailor a hemodialysis therapy to meet their unique needs may be critically important to ensure good outcomes and better quality of life.

More than 100 abstracts and peer-reviewed journal articles have demonstrated clear and consistent benefits of more intensive hemodialysis on numerous biochemical, physical, and psychological functions. This is the case whether the therapy is performed as SDHD or as NHD (although the evidence base is more robust with NHD). These benefits include improvements in cardiovascular outcomes, bone and mineral metabolism, nutrition, sleep, and quality of life. Furthermore, very limited data supports improved hospitalization rates and overall mortality as compared to CHD. At this time, the NIH frequent dialysis study is comparing these more intensive dialysis strategies in a prospective manner. It is critical that properly controlled clinical trials be designed to support conclusions based upon observational or cohort studies. The results of this important trial will provide important comparative information to allow clinicians to offer a wide spectrum of renal replacement therapies to meet the goals of their patients.