Peritoneal Dialysis Use in Patients With Ascites: A Review

Abstract

Past few decades have seen steady increase in the prevalence of kidney failure (KF), needing kidney replacement therapy. Concomitantly, there has been progressive growth of heart failure and chronic liver disease, and many such patients develop ascites. Therefore, it is not uncommon to encounter patients with KF who concurrently have ascites. The presence of ascites adds many challenges in the management of KF. Poor hemodynamics make volume management difficult. Presence of coagulopathy, malnutrition and encephalopathy compounds the complexity of the management. Such patients do not tolerate hemodialysis well. However, several concerns have limited the use of peritoneal dialysis (PD), and hemodialysis remains the predominant dialysis modality in these patients. On the contrary, observational studies illustrate that PD provides hemodynamic stability and facilitates better volume management compared to hemodialysis. Moreover, PD obviates the need for therapeutic paracentesis by facilitating continuous drainage of ascites. PD potentially reduces hemorrhagic complications by avoiding routine anticoagulation use. Moreover, small studies suggest that outcomes such as peritonitis, and mechanical complications are comparable to that in KF-PD patients without ascites. PD does not affect transplant candidacy and these patients can successfully receive combined liver and kidney transplants. Hence, PD should be considered a viable dialysis option in KF patients with ascites

Introduction

Poor lifestyle behaviors such as smoking, unhealthy diet, inadequate physical activity and excessive alcohol-use have led to upsurge in chronic diseases like diabetes, hypertension, metabolic syndrome and obesity in the past few decades ^1–4. Consequently, there has been steady increase in prevalence of kidney failure (KF) ⁵. In parallel, there has been persistent growth of heart failure (HF) and chronic liver disease (CLD), particularly non-alcoholic fatty liver disease (NAFLD), as they share the aforesaid risk factors ^6–8. Many patients with CLD and advanced HF develop ascites during the course of their disease ^9–12. Therefore, it is not surprising to encounter patients with KF who concurrently have ascites.

The presence of ascites presents numerous challenges in the overall management of KF. These patients usually have poor hemodynamics that makes volume management very difficult ^13–15. Furthermore, presence of coagulopathy, malnutrition and encephalopathy compounds the complexity of the management. Such patients do not tolerate HD well ^16,17. Peritoneal dialysis (PD), by offering steady state treatment, may provide hemodynamic stability and better volume management. Furthermore, PD eliminates the need for LVP due to regular drainage of ascites fluid ^18–22. However, there are perceptions that use of PD is associated with higher risk of infections, excessive albumin loss and overall poor outcomes, and therefore, should not be offered in patients with KF and ascites ^18,19,23. In this review, we will discuss the challenges, outcomes and technical aspects of performing PD in patients with KF and ascites.

Kidney Failure and ascites

Epidemiology and outcomes

Most of the earlier reports describe nephrogenic ascites in HD patients, with variable incidence rates (0.7-20%) and poor survival (7-10.7 months) ^24,25. Nephrogenic ascites in contemporary ESKD cases is mainly observed in severely underdialyzed or malnourished HD patients and will not be discussed in this review ^25–27. The prevalence of ascites in KF due cirrhosis and other causes remains uncertain. Small studies from Asia have observed 4% – 6% incidence of cirrhosis in KF, but incidence of ascites is not reported in all studies ^28–31. Among 1069 adult KF patients from Korea, 4.1% had cirrhosis and 1.1% patients had ascites ²⁸. Presence of cirrhosis increases the overall mortality risk in KF patients. In a Taiwanese national cohort, cirrhosis was found to be an independent predictor of mortality in dialysis patients ²⁹. Likewise, in a Korean study, cirrhotic patients had significantly lower 1-,3-,5-year survival compared to the non-cirrhotic KF patients ²⁸.

Kidney replacement therapy in patients with ESKD and ascites

The presence of ascites in KF patients poses many challenges that make dialysis treatment arduous. While patient with ascites are overall hypervolemic, they have low effective arterial blood volume. Rapid removal of fluid from intravascular compartment during HD may, therefore, produce severe hypovolemia and hypotension ^13–15. This may preclude adequate ultrafiltration and result in insufficient fluid removal, progressive volume overload and worsening ascites. Furthermore, hemodynamic instability may necessitate early termination of HD, thereby compromising adequate solute clearance and predisposing uremic state. Such uremic state along with rapid osmotic shifts during HD treatment can alter cerebral water content and predispose patients to encephalopathy ^16,17. Moreover, patients with cirrhosis frequently have multiple abnormalities of hemostatic function that may increase the risk of bleeding as well as thrombosis ^32–34. Coexisting uremia may further exacerbate platelet dysfunction and increase the risk of bleeding complications. Use of anticoagulation in HD can aggravate the risks of gastrointestinal hemorrhage and excessive bleeding at the cannulation site.

Despite aforesaid challenges, HD remains the predominant dialysis modality among KF patients with ascites, and utilization of PD remains negligible. In a study using nationwide inpatient sample, less than 1% of ESKD patients with cirrhosis and ascites were initiated on PD ³⁵. There are several concerns namely, higher risks of infections or technique-failure, excessive albumin loss, poor candidacy for combined kidney-liver transplant and overall poor outcomes that raise doubts over the viability of PD in ESKD patients with ascites (Box 1) ^18,19,21,23.

Box 1: Potential advantages and disadvantages of peritoneal dialysis (PD) in patients with end-stage kidney disease and ascites.

Advantages:

• Better hemodynamic stability

• No need for anticoagulation

• Lower risk of hepatitis B and C transmission

• Continuous drainage of ascites-No need for serial therapeutic paracentesis

• Provision of caloric load

Disadvantages:

• Protein loss in dialysate effluent

• Risk of peritonitis

• Potential risks of peri-catheter leaks, hernia and other mechanical complications

• Need for stable home situation

• Inability to perform PD in the setting of physical or mental incapacity in the absence of assistance

While there may be potential drawbacks of PD such as albumin loss, inherent risk of peritonitis, need for stable home and psychosocial situation, there are many potential benefits that may outweigh the risks. PD provides slow and steady-state treatment that maintains hemodynamic stability and provides continuous drainage of ascites with daily exchanges. As PD does not require anticoagulation, it can potentially reduce the risk of hemorrhage. Additionally, PD may serve as a much needed source of calories in malnourished KF patients with ascites (Box 1) ³⁶. Therefore, PD should be considered a viable and perhaps the preferred dialysis option in KF patients with ascites.

Outcomes of PD in patients with ESKD and ascites

Impact of nutritional status

Patients with advanced cirrhosis and ascites are considerably malnourished due to inadequate dietary intake, impaired digestion and absorption, and altered metabolism of nutrients ³⁷. In addition, cirrhosis is associated with reduced synthesis of creatine, the precursor of creatinine. This may lead to lower serum creatinine levels in patients with cirrhosis, resulting in overestimation of GFR. Thus, the diagnosis of KF and initiation of dialysis can be delayed that may result in persistent uremic state and further worsening of the nutritional status ^38,39.

Hypoalbuminemia is a predictor of mortality in KF patients and an independent risk factor of peritonitis in PD ^40–42. Hence, given low baseline serum albumin in patients with KF and ascites, concerns have been raised that excessive loss of proteins in PD effluent may exacerbate hypoalbuminemia and result in poor clinical outcomes. However, most of the available data does not support this notion. One study comprising 11 KF patients with cirrhosis observed no change in serum albumin at 6 and 12 months after PD initiation ¹⁹. In contrast, small but non-significant reduction in serum albumin from the baseline was seen in two observational studies in patients with KF and cirrhosis and ascites with mean follow up of 4.5 and over 6 years respectively ^21,43. Interestingly, Selgas reported excessive protein losses (more than 30g/day) in the PD effluents initially but significantly decreased subsequently to 7-15 g/day among 8 KF patients with ascites ¹⁸. In parallel, serum albumin level increased after an initial drop in an inverse relation to peritoneal protein loss ¹⁸. Conceivably, increase in intra-abdominal pressure generated by PD fluid may counter portal pressure and thereby reduce the formation of ascites and ensuing protein losses ⁴⁴.

In summary, majority of the patients tolerate PD well with minimum change in serum albumin ^19,21,43. Protein losses are initially high but reduce with time ¹⁸. Even with daily loss of protein in peritoneal fluid, significant change in serum albumin from the baseline value is not observed ^18,19,21.

Infections and peritonitis

Cirrhosis facilitates translocation of gut microorganisms, particularly Escherichia coli and other gram-negative bacteria, into the peritoneum, which in concurrence with reduced function of peritoneal phagocytes and complement deficiency, cause spontaneous bacterial peritonitis (SBP) ^45–48. There are concerns that PD may compound the risk of peritonitis by adding the PD-catheter-related peritonitis risks to the inherent risk of SBP in these patients. However, most of the recent data suggests that this may not be the case (Table 1).

Table-1:

Experience and outcomes of peritoneal dialysis in patients with kidney failure and ascites

Study	Patients (n)	Follow-up	Peritonitis rate	Mechanical complications	PD technique failure
Marcus, et al49	9 (chronic liver disease and ascites)	3 months-8 years	1 episode/1.2 patient years	1 pericatheter leak-resolved with holding PD	2
Chow, et al44	25 PD patients with Hepatitis B cirrhosis, 36 patients with Hepatitis B and no cirrhosis	Mean follow-up 52 months	1 episode/19.2 patient-months in cirrhotic group, 1 episode/20-5 patient-moths in non-cirrhotic group	NA	NA
Bajo, et al23	5 (cirrhosis with ascites)	8-66.5 months	1 episode/24 patient months	4 abdominal hernias-surgically corrected	1
Huang, et al31	30 with cirrhosis (16 ascites), 60 non-cirrhotic PD patients	24-year experience	0.56 episode/year in cirrhotic group, 0.39 episodes/year in non-cirrhotic patients	Higher incidence of umbilical hernia in cirrhotic vs non cirrhotic patients (5 vs 1)
De Vecchi, et al19	21 PD patients with cirrhosis and 41 patients without cirrhosis	Jan 1985-Dec. 1999	Overall rate : 0.31 episode/year. No difference between the two groups	None	No difference between the two groups (6 in cirrhotic and 12 in non-cirrhotic group)
Lee, et al54	33 patients with cirrhosis (13 ascites) and 33 propensity matched control PD patients	Mean duration 46.1 months	1 episode/87.1 patient-month in cirrhotic and 1 episode/149 patient-month in control-non-significant difference	No difference between the two groups (6 in cirrhotic and 5 in control group)	No difference between the two groups (9 in cirrhotic and 5 in control group)
Selgas, et al18	8 cirrhosis with ascites	NA	1 episode/9-patient months (2.5 times higher than average incidence)	4 abdominal hernias-corrected surgically	NA
Jones, et al21	12 with liver cirrhosis (7 with ascites)	Mean follow up 54 months	0.2 episode/year	NA	None

While one small observational study of 11 patients did report 2.5 fold higher peritonitis rates in KF patients with cirrhosis and ascites, as compared to non-cirrhotic PD patients ¹⁸, several other studies have reported similar rates of peritonitis in cirrhotic and non-cirrhotic PD patients with ascites ^{19,21,31,44,49}. Jones reported peritonitis rate of 0.2 episodes per patient per year in 12 cirrhotic PD patients (58% with ascites), not different from that in non-cirrhotic PD patients ²¹. In a larger report, no significant difference in peritonitis rate was observed between 25 patients with hepatitis B cirrhosis and 36 hepatitis B patients without cirrhosis ²². Likewise, no difference in peritonitis-free survival was observed between 30 cirrhotic and 60 non-cirrhotic PD patients with 24 years of follow-up ³¹.

While rates of peritonitis may be similar between cirrhotic and non-cirrhotic PD patients, causative organisms may differ. One study reported gut-dwelling Streptococcus, Escherichia coli, and other gram-negative bacteria causing 43% of the cases of peritonitis in cirrhotic PD patients, compared to only 20% in non-cirrhotic patients ¹⁸. In contrast, another study reported more than 50% cases of peritonitis in cirrhotic patients from Gram-positive bacteria suggesting break in sterile technique ²¹. Notwithstanding the cause of peritonitis, the majority of cases were successfully treated with intraperitoneal antibiotics without causing technique failure ^18,21,49.

Patient survival and hemodynamic stability

While there is a general bias against using PD in patients with ESKD and ascites, several observational studies have successfully utilized PD in this population (Table 1). In a series of 9 KF patients with cirrhosis and ascites, survival up to 8 years was observed, with 6 of the 9 patients surviving more than 18 months with good control of uremic symptoms and volume status ⁴⁹. Similar observations were made in 5 and 8 KF patients with cirrhosis respectively, who demonstrated good hemodynamic tolerance with PD ^18,23. In fact, 3 of these patients were transferred from HD due to persistent hemodynamic instability ^18,23. In all studies, mortality was related to complications of cirrhosis and not from PD associated issues. In a more recent single-center observational study of 12 patients with KF and cirrhosis (58% with ascites), no deaths or PD technique-failure were observed over a mean follow-up of 4.5 years ²¹.

Not many studies have compared mortality between HD and PD among KF patients with ascites (Table 2). While a study from Korea reported similar mortality over 38 month follow-up between the two dialysis modalities in cirrhosis patients ²⁸, another observational study from China reported significantly lower all-cause mortality with PD as compared with HD in KF patients with cirrhosis with average follow-up of 6 years ⁴³. Likewise, among hospitalized individuals, there was a non-significant mortality difference between the KF-cirrhosis patients treated with PD and HD ³⁵. However, among the subgroup of KF patients with ascites, PD had a significantly lower in-hospital mortality compared with HD ³⁵.

Table 2:

Comparison of outcomes between hemodialysis (HD) and peritoneal dialysis (PD) in patients with kidney failure (KF) and liver cirrhosis and ascites

Study	Patients	Outcomes
Nader, et al35	26,135 cirrhotic patients with incident KF 25,686 on HD 449 (1.7%) on PD 1878 patients with ascites 18 (0.96%) on PD	No significant difference in in-hospital mortality between PD and HD among patients with KF and liver cirrhosis. Among sub-group of patients with ascites, significant lower in-hospital mortality (0 vs 26.67 p-0.03) with PD compared to HD Longer hospital stay among HD compared to PD patients (8.34 days vs 7.06 days, p<0.001) Higher in-hospital charges with HD ($74,501) compared to PD ($ 57,460). P<0.001.
Chou, et al43	Cohort 1: 85 PD and 340 HD patients with KF and cirrhosis. Cohort 2: 279 PD and 1116 HD patients with cirrhosis and KF Prevalence of ascites not available.	Lower mortality among PD patients compared to HD (Hazard ratio 0.48, 95% confidence interval 0.31-0.74, p<0.01). Average follow up 6 years Lower mortality among PD patients compared to HD (Hazard ratio 0.61, 95% confidence interval 0.47-0.79, p<0.01). Follow-up duration NA
Kim, et al28	44 KF patients with cirrhosis (33 on HD-11 with ascites, and 11 on PD-1 with ascites	No difference in mortality between HD and PD patients (p=0.562)
Chien, et al30	40 PD and 703 HD patients with KF and cirrhosis (prevalence of ascites not given)	Liver cirrhosis is an important predictor of mortality in ESKD but the effect on mortality was not different between HD and PD patients

Taken together, the available evidence, though based on small case series, indicates that PD is a viable option among patients with KF and cirrhosis with ascites, and has similar outcomes compared to cirrhotic KF patients with ascites undergoing HD.

Logistics of Peritoneal Dialysis Initiation in KF Patients with Ascites

Patient selection

PD should be considered in KF patients with ascites. Importantly, dialysis modality education should be provided to patients and family members, expeditiously if urgent-start dialysis is expected ⁵⁰. Home-evaluation should be conducted to determine if home conditions are safe and supportive of PD. In addition, the patient should be evaluated by multidisciplinary team including nephrologists, surgeon, hepatologists and other co-managing teams for any medical, psychosocial and surgical barriers to PD.

There are a few contraindications against use of PD in KF with ascites. Patients with poor home situation, who are not motivated, and those with malignant ascites are not candidates for PD ^50,51. Additionally, patients with recent abdominal surgery, acute bowel inflammation, and uncorrected hernias are not candidates for PD. Moreover, patients with acute liver failure, acute hepatitis, hepatorenal syndrome, severe chronic hepatitis with severe coagulopathy (prothrombin time greater than 3 seconds despite vitamin K administration, or platelet count of less than 50,000/dL) are also considered as contraindications for surgery ^34,52. Similarly, patients with hepatic encephalopathy are not PD candidates, unless option for assited PD, either at home or in a nursing facility is available.

In general, patients with ascites from cirrhosis are at higher risk for peri-operative complications such as bleeding, hypotension and liver inury from drugs used during anesthesia. Pre-operative planning in these patients should be meticulous and should take into consideration the severity of liver disease, type of surgery and the method of anesthesia, so as to allow patient-optimization and even consideration for alternative procedures such as percutaneous placement or bridge with HD until patient can be optimized for general anesthesia ^34,52,53.

PD catheter placement

Currently, there is no unanimity on optimal insertion technique in KF patients with ascites. While nephrologists placed PD catheter by percutaneous technique at one center ⁴⁹, open surgical technique was employed by others ^18,19,23,54 . Similarly, successful laparoscopic PD catheter placement has been described in this patient population ⁵⁵. Neither bleeding complications nor intestinal perforations have been reported to occur more often in patient with KF and ascites ¹⁸. Even in patients with prolonged coagulation times, successful percutaneous catheter placement has been performed with no bleeding complications ⁴⁹.

Overall, the risks of mechanical complications after PD catheter placement in KF with ascites have been quite low and can be successfully managed without causing technique failure (Table 1). De Vecchi observed no surgical or mechanical problems in 21 patients with KF and cirrhosis (12 patients with ascites) ¹⁹. One study reported higher incidence of umbilical hernia in cirrhotic KF patients, but no difference in catheter migration, leakage or hydrothorax when compared with non-cirrhotic KF patients ³¹. Marcus et al observed pericatheter leak in one patient that resolved with holding PD for 2 days ⁴⁹. Bajo et al noted development of abdominal hernia in 4 patients with KF and ascites that were appropriately corrected without discontinuation of PD (Table 1) ²³. Likewise, no significant difference in technical complications was observed between KF patients with cirrhosis (40% had ascites) and propensity matched non-cirrhotic KF patients (Table 1) ⁵⁴.

While current guidelines do not recommend a specific procedure for catheter placement in this special population, the choice of operator (surgeon, nephrologist or interventional radiologist) and the technique (open-surgical, laparoscopic or percutaneous) should be dictated by the center experience and the established mechanism for catheter placement at the individual site. Perioperative antibiotics should be administered to reduce the incidence of early peritonitis. The choice of prophylactic antibiotic should depend upon the spectrum of antibiotic resistance of the individual center ⁵⁶.

PD initiation

Unless there are medical reasons for urgent-start in hospital, the patient should be seen in the out-patient PD-center within 7 days or sooner after PD catheter placement. The initial evaluation should include assessment of volume status and uremic symptoms to determine the urgency of initiating PD. If there is no immediate need to start PD, patient can wait a 2-4 weeks for PD initiation to promote healing of the surgical site and reduce the risks of catheter complications ⁵⁷. However, even if PD initiation is delayed, the ascites fluid should be frequently drained and the catheter concomitantly flushed to ensure patency and monitor any leaks (Fig. 1). There is no unanimity on the amount of ascites fluid to be initially drained. Drainage of 5-6 L of ascites fluid has been described on the initial visit in patients with tense ascites ¹⁸. On subsequent visits, different strategies have been used to determine the volume of ascites fluid to be drained. Some centers have drained 10-20% extra fluid over the instilled fill volume (FV) with every exchange ^18,23 Others have drained 400-600 ml above the instilled FV till the ascites is completely drained ^19,54,58 Yet others have not specified the volume of ascites drained on each visit ²¹. Some centers have infused albumin while draining ascites, while others have not ^{18,19,21,23,54,58}.

Fig. 1:

Initiation of peritoneal dialysis in patients with end-stage kidney disease with ascites

There is no consensus on the initial dialysate FV to be instilled. While Seglas initiated with 2 L FV from the outset, De Vecchi started with 1000 ml FV and Lee with 500 ml with stepwise increase to 2 L ^18,19,54 . In general a typical starting FV is 500-750 ml. FV should be slowly increased every 3-4 days, as tolerated, while monitoring for catheter leaks and overfills, till the patient reaches a maximum prescribed FV, usually 1.5-2.5 L, typically in 2-4 weeks. All exchanges and drains should be performed in supine position till catheter site is completely healed (Fig. 1).

In summary, the amount of ascites fluid drained and fill-volume of the infused dialysate should be determined based upon the size and comfort of the patient. The tonicity of the dialysate, and number and frequency of PD exchanges are determined by the hemodynamic state, extent of symptoms, and volume status. In case urgent-start is needed, typically 2-5 exchanges over 4-6 hours a day are performed by the PD nurse, 2-5 days a week, and the PD prescription is adjusted frequently depending upon the clinical status of the patient. If urgent-start is not needed, PD flushes with concurrent ascites fluid drainage should be done frequently, typically 1-3 times a week, depending upon the clinical situation and patient schedule.

PD in KF with ascites and liver transplant

Simultaneous kidney and liver transplant (SLKT) is considered the preferred option in patients with concomitant liver and kidney failure needing transplant. Subjects with KF who receive liver transplant alone (LTA) and remain on dialysis have poor allograft and patient survival ⁵⁹ . On the other hand, compared to LTA, SLKT are more cost effective, get lower exposure to the risk of anesthesia, are exposed to one set of alloantigens, thereby reducing the risk of rejection of kidney allograft and provide better overall outcomes ^60,61.

There are perceptions that initiating PD in patients with KF and ascites will hamper their candidacy for liver transplant ²¹. However, data to support this dogma are lacking. In fact, in a study of 12 PD patients with KF and cirrhosis (7 with ascites) 3 patients underwent successful SLKT, 4 additional patients remained active on the transplant wait-list and 5 were deemed not transplant candidate due to comorbidities ²¹. On the same token, PD has been successfully performed in liver transplant recipients who later develop KF ⁶².

Clinical Vignette-Outcome and Summary

The patient eventually had PD catheter placed by the transplant surgeon. He did very well on PD. His volume status improved and blood pressure remained stable. He did not require any LVP while on PD. He had one episode of peritonitis caused by coagulase negative Staphylococcus, suggesting break in sterile technique as the root-cause rather than SBP. After about 3 years on PD, he received SLKT.

In summary, PD is a viable dialysis modality in patients with KF and ascites. It provides hemodynamic stability and facilitates better volume management compared to HD. Moreover, it provides continuous drainage of ascites thus mitigating need for LVP. PD potentially reduces hemorrhagic complications by avoiding routine anti coagulation use. Moreover, outcomes such as peritonitis, and mechanical complications are comparable to the control PD population. While there may be initial drop in serum albumin level, this is short-lasting and does not affect the overall outcomes. Furthermore, PD does not affect transplant candidacy and these patients can successfully receive SLKT.

Clinical Vignette.

A 65-year old male has history of hypertension, diabetes mellitus type-2, hepatitis-C related cirrhosis with ascites, and progressive chronic kidney disease (CKD) due diabetic nephropathy. In the past 3 months, he was hospitalized twice with episodes of spontaneous bacterial peritonitis (SBP) and hepatic encephalopathy. Lately he had required frequent large volume paracentesis (LVP). His Model for End-stage Liver Disease (MELD) score was 21. His CKD had progressed to stage-5 and he had been educated about renal replacement therapy options. He expressed preference for peritoneal dialysis (PD), and voiced interest in kidney-liver transplant. He was referred to surgery for placement of PD catheter, but was informed that “patient with ascites should not get PD catheter placed”. Consequently, he was initiated on in-center hemodialysis (HD). He did not tolerate ultrafiltration due to persistent hypotension and remained volume overloaded. He still needed frequent LVP to manage his ascites. Should this patient be reconsidered for PD?