Renal Replacement Therapy in Cirrhosis: A Contemporary Review

Abstract

Acute kidney injury is a common complication of decompensated cirrhosis, frequently requires hospitalization, and carries a high short-term mortality. This population experiences several characteristic types of acute kidney injury: hypovolemic-mediated (prerenal), ischemic/nephrotoxic-mediated (acute-tubular necrosis), and hepatorenal syndrome. Prerenal acute kidney injury is treated with volume resuscitation. Acute-tubular necrosis is treated by optimizing perfusion pressure and discontinuing the offending agent. Hepatorenal syndrome, a unique physiology of decreased effective arterial circulation leading to renal vasoconstriction and ultimately acute kidney injury, is treated with plasma expansion with albumin and splanchnic vasoconstrictors such as terlipressin or norepinephrine. Common acute stressors such as bleeding, infection, and volume depletion often contribute to multifactorial acute kidney injury. Even with optimal medical management, many clinicians are faced with the challenge of initiating renal replacement therapy in these patients. This article reviews the epidemiology, indications, and complex considerations of renal replacement therapy for acute kidney injury in decompensated cirrhosis.

EPIDEMIOLOGY OF RRT IN CIRRHOTIC PATIENTS WITH AKIE

Renal replacement therapy (RRT) should be considered in patients whose kidney injury does not respond to medical management and continues to have worsening kidney function and complications of severe azotemia, volume overload, or electrolyte/acid-base disturbances. Overall, indications for RRT are largely similar between patients with cirrhosis and the general population.^1,2 Typical uremic symptoms of poor appetite and pruritus may be difficult to distinguish from symptoms of hepatic encephalopathy and hyperammonemia.³ Both hemodialysis (HD) and continuous renal replacement therapy (CRRT) have been used for acute kidney injury (AKI) in the setting of cirrhosis. CRRT is better tolerated by critically ill patients (particularly in those with cirrhosis and low effective circulating volume) due to less hemodynamic instability and slower correction of metabolic derangements.⁴ Compared to intermittent hemodialysis (IHD), CRRT is associated with higher mortality given that it is only employed in the critically ill, and observational analyses thus carry this indication bias.^5,6 1 population-based retrospective cohort of 722 adult patients with cirrhosis and AKI in Canada showed the cumulative incidence of liberation from RRT at 1, 3, 6, and 12 months to be 3%, 22%, 25%, and 26%, respectively.⁷ Higher Model for End-Stage Liver Disease (MELD)-Na, acute-on-chronic liver failure, and sepsis were associated with lower rates of recovery, whereas those on a liver transplant (LT) list and those admitted to a teaching hospital were more likely to achieve recovery. In another study, Allegretti and colleagues showed that 15% of patients not listed for LT were alive at 6 months following initiation of RRT.⁸ 1 nationwide cohort study from 2006 to 2012 identified in-hospital mortality at 41% overall in cirrhotic patients with AKI requiring RRT as opposed to 7% in those who did not require RRT.⁹ Table 1 summarizes several key studies that focus on the use of RRT in cirrhosis.

Table 1.

Selected Studies Reporting Utilization and Outcomes of Patients With Cirrhosis Who Require Renal Replacement Therapy

Source	Received RRT/Total Patients (n)	Liver Transplant Status (n)	Liver Function at Admission	RRT Duration (Days)	Type of RRT	Survival	Liberation From RRT	Comments
Feldkamp10 and colleagues, 2020	54/149 LT recipients	149 LT recipients	MELD 18	NR	NR	1 y survival: -RRT: 46.4% -No RRT: 92.2%	NR	Survival significantly lower with RRT after 28 d, 1 y, 3 y
Allegretti8 and colleagues, 2018	472/472 cirrhosis + RRT	L = 131 (48% received LT) NL = 341	MELD L = 36 NL = 34	NR	CRRT = 70% IHD = 30%	6 mo survival: -L = 48% -NL = 15%	Among 6 mo survivors: -L (no LT) = 38% -NL = 78%	No association between AKI etiology + mortality. LT status, MELD, indicators of critical illness associated w/survival.
Sharma11 and colleagues, 2013	2112/2112 L	2112 LT recipients	MELD 39	11 d	NR	6 mo survival after LT: 79%	After LT 91.1%	Significant association between nonrecovery and advanced age, diabetes, re-LT, longer pre-LT RRT duration
Northup12 and colleagues, 2010	1041/1041 L	1041 LT recipients	MELD 34 (recovered) vs 36 (no recovery)	15 d (recovered) 36 d (no recovery)	NR	1 y survival: −99% (recovered) −38% (no recovery)	After LT: 67.9% RRT <30 d: 70.8% RRT >90 d: 11.5%	Duration of pre-LT RRT is highly associated with recovery post LT.
Wong4 and colleagues, 2005	102/102 L	31% received LT	MELD 38.9 APACHE II 28.7	6d	CRRT = 65% IHD = 35%	Survived to LT: 32% Survived to discharge, no LT: 4% 1 y mortality after LT: 30%	NR	RRT Justifiable for LT candidates experiencing AKI considering. Postoperative LT mortality increased in those who received pre-LT RRT.
Gonwa13 and colleagues, 2001	1985-1995: 86/1037 LT recipients	1037 LT recipients	NR	NR	NR	1 y survival: -Pre-LT RRT only or Pre/Post LT RRT: 89.5% -Post-LTRRT only: 55% -No RRT: 90.6%	NR	The need for RRT has increased along with waiting time in LT patients. Patients needing RRT post-LT have lower survival at 1 y.
	1996-1999: 62/498 LT recipients	498 LT recipients	CP Score RRT: 10.8 No RRT: 8.8	NR	CRRT = 81% IHD = 19%	1 y survival: -RRT: 54.5% -No RRT: 88.6%	16% developed ESRD
Keller5 and colleagues, 1995	38/107 liver failure and AKI	NR	CP Class C 74/107	NR	NR	1 y mortality rate: -RRT pts: 50% -Pts with compensated renal failure: 40% -Pts not receiving RRT for poor prognosis: 87%	Renal recovery in those who survived and received RRT: 74%	Association between higher mortality rate and thrombocytopenia, encephalopathy, and malignancy but not HRS
Saraiva14 and colleagues, 2020	66/66 with ACLF + CRRT	L = 5 LT = 1	MELD 29 SOFA 14	3.7 d	CRRT = 100%	89.4% patients died in the hospital	NR	ACLF and AKI requiring CRRT has poor survival. SOFA, CLIF-SOFA are good prognostic tools.
Wang15 and colleagues, 2022	722/722 cirrhosis + RRT	L = 56 LT = 19	MELD 28	NR	CRRT = 20% IHD = 63% Both = 17%	1 y mortality: 68%	At 3 mo: -L: 50% -NL: 25%	Recovery off RRT unlikely after 3 mo.

ACLF, acute on chronic liver failure; AKI, acute kidney injury; CP, Child Pugh; CRRT, continuous renal replacement therapy; ESRD, end stage renal disease; IHD, intermittent hemodialysis; L, listed patients; LT, liver transplant; MELD, Model for End-Stage Liver Disease; NL, not listed patients; NR, not reported; SOFA, sequential organ failure assessment; RRT, renal replacement therapy; RF, renal function.

Of note, peritoneal dialysis (PD) has been examined in very small studies for patients with AKI and cirrhosis. PD has less hemodynamic consequence compared to IHD and therefore may be associated with less intradialytic hypotension. Many of these patients have concurrent ascites, which may also be drained by the PD catheter. However, there may also be an increased risk of infection and hypoalbuminemia (via albumin loss through the dialysate) in a population already at risk for both.^16-19 There is also a need for specific procedural logistics to initiate PD in highly morbid patients with cirrhosis, which may not be available in all healthcare systems. Overall, PD is not commonly performed in patients with cirrhosis and AKI, but more studies are needed to further evaluate its value in this susceptible, acutely ill population.

GENERAL CONSIDERATIONS FOR RRT IN CIRRHOTIC PATIENTS WITH AKI

There are special considerations when evaluating the need for RRT in cirrhotic patients with AKI. Patients with decompensated cirrhosis commonly have low blood pressure (due to decreased effective circulating volume), which can be further complicated by intradialytic hypotension, a complication that is associated with increased mortality in patients with HD at large.²⁰ As previously stated, patients with cirrhosis are at risk for hyperammonemia and hepatic encephalopathy. Of note, RRT can clear ammonia and can be considered an adjunct therapy to first-line oral agents such as lactulose and rifaximin, though the role of RRT in ammonia removal in adult populations to prevent complications like cerebral edema is not standard of care as halting ammonia production is more critically important than transient clearance.²¹

Anticoagulation during CRRT remains a challenge for patients with cirrhosis. In the general population, regional citrate anticoagulation (RCA) is recommended as first-line anticoagulation for CRRT.²² RCA reduces the risk of bleeding compared to heparin and prolongs the lifespan of the RRT circuit,²³ but it has not been well studied in cirrhosis. Acutely ill patients with cirrhosis are often not offered RCA due to concerns for acute citrate toxicity/accumulation leading to hypocalcemia and metabolic acidosis, as citrate is metabolized in the liver.^24,25 However, more recent meta-analysis²⁶ and one recent trial²⁷ of RCA in CRRT suggest that RCA’s filter-prolonging effect is seen in patients with liver disease, though rates of citrate toxicity and hypocalcemia are increased. While these observational studies and small trials suggest citrate may be a consideration in some circumstances, further study in advanced decompensated patients with cirrhosis is needed.^26-28

RRT IN TRANSPLANT CANDIDATES

The decision to start RRT in nontransplant candidates remains a challenging issue for clinicians, patients, and caregivers (Fig 1). Short-term prognosis remains unfavorable, with mortality rates exceeding 60-80% at 6 months,⁸ and only ~25% of patients being alive and off-dialysis at 1 year.¹⁵ We therefore advocate for early involvement in palliative care consultation, consistent with societal guidelines, both for symptom management and to assist in serious illness discussions.^29,30 Regardless of palliative care involvement, we also recommend multidisciplinary discussions about the role of RRT in the overall management of the acute illness and whenever important changes in the clinical status of the patient are noted, but in particular, prior to the start of RRT, if possible. Despite poor prognosis, a time-limited trial of RRT may be appropriate for many patients while looking for signs of renal recovery or improvement in liver function, as bridge-to-candidacy or as bridge-to-decision (allowing time for families to process the patient’s prognosis and/or arrange for final wishes). When considering RRT in this population, we must also acknowledge that while short-term prognosis is poor over a period of weeks or months even with RRT support, it is often markedly shorter without it. This should be both acknowledged and thoughtfully framed when discussing options with patients and their caregivers. Fig 2 offers a potential approach to the decision-making process for initiating RRT in acutely ill patients with cirrhosis and AKI.

Figure 1.

Considerations when initiating renal replacement therapy for acute kidney injury in patients with cirrhosis. AKI, acute kidney injury; CKD, chronic kidney disease; MELD, Model for End-Stage Liver Disease. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Figure 2.

Practical approach to a patient with cirrhosis and severe acute kidney injury. AKI, acute kidney injury; RRT, renal replacement therapy; KDIGO, Kidney Disease Improving Global Outcomes. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

In the past, the decision to start RRT was influenced by the cause of AKI. This is because hepatorenal syndrome (HRS) was believed to have the worst prognosis compared to AKI due to other causes,^5,31 leading to the conclusion that initiating RRT was not warranted in nonlisted patients with HRS. However, contemporary studies have shown that the cause of AKI is not a significant factor associated with mortality once starting RRT and may not be associated with renal recovery post-transplant.^14,32 In contrast, the status of critical illness severity, prognostic scores such as MELD, age, and transplant listing status⁸ are more relevant factors associated with mortality that can assist with risk stratification. In particular, critical illness parameters such as mechanical ventilation, need for vasopressors, and evidence of multisystem organ failure have been strongly associated with exceedingly high mortality in several studies,^8,33,34 suggesting that initiating RRT should be based on global acute illness severity and traditional prognostic models, like the MELD score.

RRT could be considered in nonlisted patients for the possibility of change in transplant listing status. In a small study (n = 47 patients), Lenhart and colleagues³² found that 20% of patients with alcohol-associated liver disease who required RRT survived to 6 months of sobriety and became eligible for LT re-evaluation. Among them, 40% received a simultaneous liver-kidney (SLK) transplant, and 60% recovered renal function and did not require a LT.

RRT IN NONTRANSPLANT CANDIDATES

RRT should be used as a bridge-to-transplant in listed patients with cirrhosis and AKI who fail medical management; up to 50% of listed patients on RRT survive the LT.^4,8 This outcome is remarkable given the overall high mortality rate observed in cirrhotic patients with AKI requiring RRT.

Two-thirds of patients with HRS-AKI undergoing pre-LT RRT recovered renal function after the transplant, with time on RRT being the main predictor of renal recovery, along with advanced age and diabetes.¹² If liberation from dialysis occurs, it usually happens within 6 months of LT.¹¹ In one study, the duration of pre-LT RRT was associated with renal recovery after LT, with every additional day of RRT beyond 14 days adding a 3-6% risk of nonrecovery.³⁵ Thus, current United Network for Organ Sharing/ Organ Procurement and Transplantation Network guidelines recommend SLK transplantation if patients have been on RRT for >6 weeks and a “safety net” evaluation for future kidney transplants for those with moderate to severe CKD after LT alone.³⁶ Additional studies using other biomarkers of kidney function, such as cystatin C, functional imaging of the kidneys, realtime GFR measurements, or kidney biopsy may further assist in the risk-stratification of SLK transplantation in this patient population.

POST-LIVER TRANSPLANT RRT

Up to 60% of LT recipients develop AKI with or without the need for RRT.^37,35 Factors that contribute to postoperative AKI include extended vena cava cross-clamping time, immunosuppression protocols that involve the use of calcineurin inhibitors, perioperative hypotension, massive blood transfusions, and severe volume depletion.^39-41

Post-LT AKI is a complication with long-lasting consequences and is a significant independent risk factor for death during the first year following LT, with the highest mortality occurring in patients receiving CRRT compared to IHD (42% vs 25%).^13,15,38,42 Preoperative LT serum creatinine is a strong predictor of the development of postoperative AKI requiring RRT.¹⁰ A Center for Medicare and Medicaid study of 2112 LT patients who required RRT within 90 days prior to transplantation showed that the incidence of renal nonrecovery, defined as dependence on dialysis or listing/receipt of kidney transplant was 8.9% within 6 months.¹¹ Longer preoperative RRT duration, advanced age, and diabetes were significantly associated with an increased risk of renal nonrecovery.

Monitoring hemodynamics to maintain adequate perfusion pressure, keeping an adequate fluid balance according to the phases of resuscitation, preventing increases in central venous pressure (high values have been associated with a greater risk of postoperative AKI),⁴³ avoiding unnecessary blood transfusions, and avoiding nephrotoxic drugs during the perioperative period are essential measures that could help in reducing post-LT AKI. RRT plays a vital role in supporting these patients through the postoperative period when medical management is insufficient.

UNANSWERED QUESTIONS

Limitations of current data include the retrospective nature and limited generalizability of many studies, as well as the heterogeneity of RRT practices in different parts of the world. As most literature is based on data from North America, larger international databases are needed to corroborate this evidence and address broader geographic variations. Further, better clinical sub-phenotyping of patients with cirrhosis through the use of novel biomarkers of kidney disease and/or artificial intelligence leveraging electronic health record data is needed to assist with risk classification and enable more informed and personalized clinical decisions. The use of the “best case/worst case” decision aid, a tool generated from the surgical literature to facilitate understanding of potential postoperative courses, may be useful in contextualizing decisions about starting RRT in critically ill patients with cirrhosis and AKI.⁴⁴

The value-based implementation of drugs such as terlipressin, which is newly available in the United States, may impact RRT rates and outcomes in this population.⁴⁵ Other forms of extracorporeal liver support devices, including plasma exchange, single-pass albumin dialysis, and the use of adsorption columns have been studied for the treatment of liver-related complications such as hepatic encephalopathy, but not specifically for the support of AKI.^46,47 Of note, these blood purification technologies are not used consistently across different centers and may not be available in all regions.

SUMMARY AND FUTURE DIRECTIONS

Patients with cirrhosis who suffer from AKI requiring RRT have a poor prognosis. When approaching the decision to start RRT, the patient’s transplant status, acute illness parameters, comorbidities, and overall baseline functional status and goals of care should be considered. Palliative care should be involved early, with the goal of eliciting patients’ values and improving the shared decision-making process. All stakeholders in the patients’ care, including hepatology, nephrology, primary medicine/intensivist, and transplant teams, should be involved and adopt a consistent and unified approach in these difficult serious illness discussions. RRT should be used in listed patients as a bridge to transplant. A time-limited, goal-specific trial of RRT may be considered for nonlisted patients. The cause of AKI is not a specific factor associated with mortality or renal recovery and therefore should not be overrated. Further study of best RRT practices in this susceptible population is direly needed.

CLINICAL SUMMARY.

• In patients with decompensated cirrhosis, kidney replacement therapy is used as a bridge therapy for patients listed for liver transplant.

• Short-term outcomes are poor for those patients with decompensated cirrhosis who are not listed for liver transplant and require kidney replacement therapy, regardless of the cause of acute kidney injury.

• We suggest a multidisciplinary and patient-centric approach when considering candidates for kidney replacement therapy in this population.