The future of peritoneal dialysis: PD in 2010 and beyond


  • John Burkart MD

    1. Professor of Medicine/Nephrology in the Department of Nephrology at Wake Forest University Medical Center in Winston Salem, North Carolina
    2. Chief Medical Officer of Health Systems Management in Tifton, Georgia
    Search for more papers by this author


Peritoneal dialysis (PD) has been used to treat patients with stage V chronic kidney disease since 1976. The number of patients on PD varies from country to country and was only about 8% of the U.S. population by the end of 2009, despite the fact that medical outcome data would seem to favor more utilization of PD. For instance, data from the United States Renal Data Systems suggests that the relative risk of death for patients on PD versus center hemodialysis (CHD) has been improving, tending to favor those on PD for longer periods of time. In addition, infectious complications have been markedly reduced. In these economic times it is important to note that PD is less costly than CHD, which has influenced anticipated changes in government reimbursement—such as the bundling of dialysis-related services in the U.S. These facts should stimulate a renewed interest in home therapies, especially PD. However, for this therapy to be more successful, the PD community must better educate its members (in the U.S., PD education during fellowship is often lacking, many current PD units are small and the number of well-trained PD nurses may not be able to optimally meet an increase in demand). Peritoneal dialysis will face many challenges, such as: frequent hemodialysis schedules that may be more and more appealing to patients and providers; the aging of the end-stage renal disease (ESRD) population; regulatory issues; lack of infrastructure at some units; and problems related to long-term use of PD. This article will review some of these challenges.

Peritoneal dialysis (PD) has been used as a chronic treatment for patients with end-stage renal disease (ESRD) since 1976.1 In the United States, peak utilization occurred in 1993 when approximately 14.9% of patients were on PD. Since then, while the total number of patients on PD has been relatively stable, the increasing numbers of patients receiving dialytic care for ESRD has lead to a decline in the percentage of patients utilizing PD, to less than 8% in 2008.

This is markedly different from the experience in other developed regions such as Mexico, Canada, New Zealand, Australia, and Hong Kong.2 This article will explore the reasons for this, as well as the complex web of financial, cultural, and medical practice differences between countries to consider what the future is for PD in the U.S. considering patient outcomes, utilization, and technological advances.

Since becoming an attending nephrologist in 1984, I have treated many patients with PD, some for more than 15 years. At our university, I am the medical director of both a thriving center hemodialysis (CHD) unit and a robust home unit (offering both PD and home hemodialysis [HHD]). I think that all modalities are equally good but, when possible, a patient should go home. However, based on my experiences and my review of the literature, I am biased towards home dialysis, particularly PD. I believe that PD is currently underutilized and that, over time, patient outcomes will improve even further, and that this therapy clearly has a future.

State of the Art of Renal Replacement Therapies in 2010

USRDS Observational Data—First-Year Mortality Concerns

Although there have been numerous technological advances in dialysis (most in hemodialysis [HD]-such as high flux dialyzers, online clearance, sodium variable, ultrafiltration (UF) profiling, newer synthetic membranes, etc.), there has only been minimal improvement in overall patient survival. What is especially alarming is the high mortality experienced during the first and second years on renal replacement therapy (RRT).

These observations are best illustrated by data from the United States Renal Data Systems (USRDS), the national database that captures clinical data from all treated patients with ESRD in the U.S. whose treatments are paid for by Medicare and Medicaid from day 90 onward.3 According to the USRDS data, the adjusted first-year mortality is approximately 25%, with a peak in the third month of dialysis.4 This high early mortality is not unique to the U.S. and is seen throughout the world, as noted in the Dialysis Outcomes and Practice Patterns data.5

This excessive rate of death is in part related to infections as a result of temporary vascular access. These infections cannot only cause morbidity and mortality, but may also initiate the catheter-infection-inflammation-cardiovascular disease pathobiologic pathway. Relentless inflammation is known to be pro-atherogenic and its persistence modulates vascular and nutritional factors important in predicting patient survival. Patients on thrice-weekly CHD tend to be getting shorter and shorter treatment sessions due to the improved ability to minimize hypotension (sodium variable, UF profiling, etc.) and better small molecule solute removal (high flux dialyzers), but as a result are intermittently volume overloaded and may be chronically hypertensive leading to left ventricular hypertrophy.6 They also are at risk for repeated episodes of transient hypotension (cardiac stunning) leading to intra myocardiac fibrosis, which can predispose to ventricular arrhythmias and sudden death.7 In short, we have to re-examine how we do dialysis.8 Peritoneal dialysis is certainly a possible solution for many of these problems.

Observational Long-term Patient Survival Data

Data from the 2008 USRDS report shows no significant difference in the overall 60-month survival between PD and CHD for the population as a whole.9 Additionally, analysis of these data showed that the relative risk of death for PD verses CHD changes over time. In general, patients starting renal replacement therapy on PD have a survival advantage, but over time this advantage becomes less robust. How quickly this change happens differs depending on the cause of ESRD, dialysis vintage, gender, country of treatment, and presence or absence of comorbidity.10, 11

It is important to note that while the first-year mortality for CHD patients has been minimally improving, there has been a very substantial improvement in first and second year survival for PD patients. Indeed recent USRDS data suggests that for the patients who started PD between 1996 and 2000, superior PD survival compared with HD was maintained for approximately 36 months.12 These observations are not confined to the United States. There are similar data from Canada,13 Northern Europe,14, 15 and from Australia and New Zealand.16 Again, these observational long-term survival data suggest that in the future we should consider using PD more often.

Economic/Cost Considerations

Numerous databases have shown that a typical PD patient is less costly than a typical CHD patient. According to the 2005 USRDS report, in 2001 the cost of a patient on PD to CMS was approximately $13,900 less than for a CHD patient, and the corresponding savings for PD versus CHD in 2002 and 2003 were approximately $15,600 and $17,200/patient/year, respectively.17 Since 2003, this difference has continued to grow, so that in 2006 the USRDS reported that the average cost for a PD patient was $53,327 while that for a CHD patient was $71,889.

Much of the excess cost for CHD is related to vascular access issues and the higher use of “injectable” medications such as intravenous iron, vitamin D receptor activators, and erythropoietin stimulating agents. It is anticipated that payment for dialysis-related services in the U.S. will transition from an “as used” payment system to a “bundled” (for the overall delivery of dialysis-related care) payment starting in 2011 (Congressional Bill HR 6331). This would favor use of PD.

One analysis of potential cost savings in the U.S. suggested the following: If PD usage were to decrease from the current rate of 8% to only 5%, Medicare spending for dialysis would increase by an additional $401 million over the next 5 years, whereas if PD utilization would increase from 8% to 15%, Medicare costs could decrease as much as $1.1 billion over 5 years.18

Healthcare officials in other countries have taken different approaches. For instance, as a result of observed differences in medical outcome data and financial cost, Hong Kong has initiated a PD first policy.19 As of 2007, 80% of patients were on PD and 20% were on CHD. Peritoneal dialysis costs were 40% those of CHD ($13,000 versus $30,000 US/year). Using this approach the reported technique survival for PD in Hong Kong at that time was 75% at 8 years.

In summary, if one reviews the state of the art of RRT in 2010, we see that there continues to be a high mortality rate, especially in the early years on dialysis. This mortality is in part due to an increased sepsis and infectious risk associated with temporary tunneled vascular catheters and an accelerated cardiovascular death risk. There is also the spiraling cost to the healthcare system for dialysis services. Increased PD use is a potential fix for many of these problems.

Peritoneal Dialysis Today

As reviewed by Isaac Teitelbaum in this issue and, as stated above, overall patient survival on PD has been improving, and for most patients during the first years of RRT is expected to be as good or better than it would be for a similar patient on CHD.20 These improvements largely occurred before widespread use of newer dialysis solutions. Changes in clinical practice such as, but not limited to: innovations in connection technologies, catheter design,21 and the use of prophylactic antibiotics22 have resulted in marked overall improvement in catheter function and decreased peritonitis and exit site infection rates.23 As a result, recent vintages of PD cohorts have shown a reduction in patients who transfer to CHD because of infection-related complications.24 In fact, the ANZDATA analysis acknowledges that active infection prevention and treatment policies were not as vigorous as they were in other countries, a possible explanation for the finding that the advantage for PD over CHD was not as robust as in other countries. Those investigators report a high incidence of deaths due to fungal peritonitis and high infection-related complication rates.25 Furthermore they state that use of standardized infection policies such as the use of prophylactic antibiotics was not common in 70% of the units.

Improved efforts to minimize glucose exposure, optimize ultrafiltration, and attain euvolemia have been underway and perhaps is an additional reason for the improved PD outcomes seen over the past 10 years. Multiple retrospective cohort observational studies suggest that if patients receive a transplant while on PD they are less likely to have delayed graft function, and therefore are likely to have longer graft survival. A more recent study from the U.S. suggested that patients transplanted from PD (or if at least 50% of their pre-transplant RRT time was on PD) had a 6% lower risk of death and a 3% lower risk of graft failure than those transplanted from CHD.26 Current trends in practice patterns and improvement in patient outcomes should favor increased use of PD.

Probable Challenges for PD in the Future

Infrastructure Issues and Increased Utilization

The likelihood that economic forces may stimulate an increased interest in and utilization of PD may itself be a challenge for PD. During the past 10 years, most home dialysis units had a median of about 10 patients.27 As a result most units have appropriately allotted enough nursing personal to home training commensurate with the number of patients on PD. This would imply that the number of trained PD nurses would be markedly less than those trained to do CHD. If there is a tremendous demand for PD, there may not be enough trained staff to continue to support the therapy as well as it has been supported in the past. The PD community and dialysis providers will need to meet that challenge with increased educational activities for nursing personal. One way to circumvent that problem would be to “centralize” home training units. For example, three small units in one city could combine into one larger central unit. This would likely improve training, reduce nursing call, and facilitate problem solving. A benefit of that growth would be that as unit size and experience increase there will be a more robust home infrastructure for nurses and physicians to refine their skills. In addition, there is the potential for better continuous quality improvement (CQI) programs which may even further improve patient outcomes.28

Physician Education

One important reason that PD may be underutilized is that the physicians are not knowledgeable about it, and therefore are not comfortable with it. Two surveys of nephrology fellows have suggested that their training is suboptimal.29, 30 When asked, only 32% of fellows stated that they had attended an outpatient PD clinic; 52% stated that their PD rotation was less than four weeks in duration; 53% attended fewer than 10 half-day PD clinics; 24% of fellows had never initiated PD in a patient; 57% initiated PD on fewer than 5 patients; and 38% felt their training was inadequate.

During the American Society of Nephrology-sponsored Board Review Course and Update held during August of 2009, a survey of fellows attending in preparation for initial certification was conducted. One of the questions asked was, “In what area of nephrology do you feel your fellowship was most deficient?” A total of 30% stated it was most deficient in PD, while another 21% stated PD was the second most deficient area of their training.

Given these data, one challenge for the future is to educate the renal replacement community and the physician in all aspects of PD. The International Society for Peritoneal Dialysis and the North American Chapter of that society plan to continue to lead the way with education.

The ESRD Population

It is true that the “burden of therapy” is related to the procedure, the device, the frequency of use, and so forth. But how the burden is overcome is related to ease of use, center tolerance, and patient characteristics. The ESRD population in the U.S. and other countries is becoming older and more likely to have diabetes and other comorbidities. However, it does not mean a patient should not do PD.

It may mean that we need to be more careful about how PD is prescribed. Hence, physician and nursing education, centralization of PD centers, and so forth, may be important. As more patients become nursing-home residents we may have to find a clinically appropriate way to make that venue a safe and financially feasible place to do “home” dialysis. It will be a future challenge for PD, but I believe that this can be done.

Frequent Home Hemodialysis Therapies

Increasingly, there is more and more interest in frequent hemodialysis (FHD) schedules, such as short daily and daily nocturnal. This is appropriate and most cohort studies show many potential benefits for patients.31, 32

We currently await the results of the randomized NIH FHD trials. If the results are positive, the question is, Why? These therapies often have higher small solute clearances (Kt/Vs), better weekly middle molecule removal, and better patient volume, and blood pressure control.

What will happen to PD if these studies are positive? To date, I do not feel that the resurgence of HHD has hindered PD. In the Wellbound experience, of the 98 new HHD patients during the years 2004 to 2006, 82% (n = 80 patients) transitioned from CHD; 8% (n 8) were incident patients; 1% (n = 2) had a failed transplant; and only 8% (n = 8) transitioned from PD.33 Of the more than 2,000 patients undergoing HHD with the NxStage System One, 87% came from CHD, 6% from PD, 4% from failed renal transplant, and only 3% were incident patients.34 I feel the therapies are complimentary and should not be competitive.

Growth in Indications for PD

If we find that FHD is better, but recognize that perhaps only 5% to 10% of our current U.S. dialysis population can or will do frequent home hemodialysis (FHHD), how do we as nephrologists offer a similar treatment to patients who have to do CHD?

One option is to offer frequent in-center HD therapies. That is a model that would be hard to do, given our current reimbursement constraints. But, what might work is to combine PD and HD. There are limited retrospective observational data reporting outcomes of patients who have been treated with this approach and the results are favorable.35, 36 In my opinion, the fact that the patient has to have two accesses is not an issue. In general, PD catheters are complication-free access. When a patient is going to transition from PD to HD (CHD or HHD) one could combine therapies. The PD catheter is already there and can facilitate easy travel PD. In addition, if the vascular access clots, the PD catheter is available for treatment.

The major problem is reimbursement. Consider doing two PD exchanges a day along with three weekly CHDs. This should augment solute clearance, phosphorus removal and, if done correctly, volume removal.

Finally, what about patients with decompensated congestive heart failure? These patients actually have a failure of the kidney to maintain normal volume homeostasis. Two studies have suggested that the use of PD ultrafiltration as a supplemental treatment to standard care may be benefical.37, 38 Both of these areas are potentials for PD growth in the future.

The Long-term PD Patient

As we become better at prescribing PD and managing its complications, more and more patients may be on the therapy for extended periods of times. Historically, these “long-term” PD patients have been found to have changes in their peritoneal membrane that make volume control more difficult, can lead to ultrafiltration failure, and in some, ultimately to encapsulating peritoneal sclerosis (EPS).

Initial peritoneal membrane changes are not life threatening. They tend to cause an increase in transport of solute both to and from the peritoneal cavity. As a result of the increased glucose uptake, UF becomes difficult, but can usually be managed by a change in prescription and occasional “resting” the peritoneal membrane with transient HD. In some patients these changes progress and may cause EPS which was first described in Japan39 and Australia.40 Unfortunately, the historic literature is confusing, in part because the condition is rare, and there have been a lack of well-defined criteria for diagnosis, randomized intervention trials, and minimal epidemiologic data. The incidence appears to vary across the world.

The International Society for Peritoneal Dialysis (ISPD) has recently published a position paper to try and reach a consensus on diagnosis, predictors of risk, and intervention.41 Their conclusions were that: EPS is a potential, but rare, complication of long-term usage (usually greater than 3–5 years) which remains infrequent; that there are no specific predictors of who will develop EPS; many of the more severe complications of EPS actually occur after stopping PD; and, overall prognosis may be more related to underlying age and comorbidities. They conclude that there is no single optimal length of time for a patient to be on PD and acknowledge that many patients choose to do PD because of quality-of-life issues, while recognizing that PD is life saving and other alternatives—such as no dialysis, HD, and transplant—have their own inherent risks.

In practice, many PD patients are reluctant to transfer to HD, even when clinically indicated, because of perceived adverse effects on their quality of life. In the future, we need to continue to gather data and hopefully be able to better identify patients at risk. We all need to learn to be comfortable with transitions from PD to other therapies. It is common to transition a patient from CHD, FHHD, or PD to transplant and from transplant back to one of these therapies. During these transitions there are risks, but overall there are benefits so we need to embrace transplant (and transitions) in the appropriate patient. We should also do this for PD.

New Solutions/Regulatory Issues

Unfortunately, the Food and Drug Administration (FDA) treats PD solutions as a “drug,” and HD solutions as a “device.” As a result, it is difficult to get new PD solutions approved in the U.S. unless a specific indication or “superiority” to standard glucose-containing fluids can be identified.

As of April 2010, there are only two types of dialysis solutions available in the U.S.: standard glucose-containing solutions, and solutions containing icodextrin as an alternative osmotic agent. Other countries have additional choices, and those PD fluids have theoretical benefits for long-term clinical outcomes that may further prolong the survival advantage associated with doing PD. Additional studies are needed to document what benefit, if any, these solutions will have on certain clinical outcomes, such as minimizing peritoneal membrane changes and the occasional resultant EPS. A current and future problem for PD growth is how the FDA looks at these potential innovations. I think we need to consider some changes to that process.

The Future Starts Now

The future of PD starts now.42 Many interventions that have documented improvements in patient outcome are not widely used. Consider antibiotic prophylaxis, proper catheter implantation technique, peritonitis treatment protocols, blood pressure, and volume control. A nephrologist can do a better job with these things right now, and certainly should in the future.

The most common reasons for transfer to HD continue to be related to catheter problems and infectious complications. In reports from centers where there is a dedicated PD implantation team, catheter-related problems are minimal and seldom a reason for transfer.43 Unfortunately, this expertise is not available at all centers. Multiple studies have shown that there is a reduction in both exit-site infections and peritonitis when prophylactic topical antibiotics are applied to the exit site during daily exit site care,44 a practice not universally followed.

It is known that the relative risk of survival between PD and HD changes over time. One possible reason for this is that the PD patients' prescription is not always changed to compensate for the loss of residual kidney function (and the associated water and sodium losses, etc.). One prescription does not fit all.45 The potential ultrafiltration volume/dwell varies based on dwell time, transport type, instilled volume, and osmotic agent used in that instilled dialysis fluid.

Optimization of drain volume and the sodium content in that drain is dependent on the physician being aware of these factors and adjusting the prescription to be able to accomplish these patient specific goals.46 If the prescribing physician is not offered free access to all available options, his or her ability to maximize blood pressure control while, at the same time minimizing glucose exposure and optimizing patient quality-of-life issues may be suboptimal. If true, this would represent another “subtle,” but real, impediment to being able to prescribe PD in a “user friendly” environment for both the patient and the doctor. This is an intervention we could potentially do now.


Peritoneal dialysis is an excellent form of RRT, and is currently underutilized in the U.S. It is not for all patients, but certainly should be considered more often than it is. There have been remarkable improvements in clinical outcomes for patients on PD over the past decade that should encourage its increased use, especially given the financial implications of proposed changes in dialysis reimbursement. The future of PD is bright, but there are continuing challenges such as infrastructure issues, regulatory issues, education, and comfort with transitioning between therapies that need to be addressed along the way.


The author has received grant support from the NIH, and unrestricted educational grants from Baxter, NxStage, Genzyme, Davita, AMAG, Abbott, Amgen. He currently sits on advisory boards for Baxter, NxStage, Genzyme, and Amgen, and receives honorarium from Baxter, Fresenius, and the American Society of Nephrology.