Liver Transplantation for Acute Liver Failure in Presence of Acute Kidney Injury

Abstract

Acute liver failure (ALF) is a catastrophic illness, which is associated with high mortality in absence of liver transplantation. ALF is associated with multisystem involvement including acute kidney injury (AKI). AKI worsens the already poor prognosis of ALF. There is limited literature on impact of AKI on outcomes of liver transplantation (LT). The use of continuous renal replacement therapy (CRRT) may have a role in transplant-free survival or bridging to LT. Although results suggest a somewhat lower survival in patients with ALF and AKI, LT is a life-saving option and should not be deferred in absence of other contraindications. In the current review, we discuss impact of AKI on transplant-free survival, possible role of CRRT, and role of LT in patients with ALF associated with AKI.

Acute liver failure (ALF) is characterized by sudden and dramatic liver dysfunction that clinically manifests as jaundice, coagulopathy, and hepatic encephalopathy. ALF is potentially associated with dysfunction in other organs also including acute kidney injury. ALF is associated with poor prognosis. Although survival has increased without liver transplantation (LT) also in recent years, there is a wide gap between survival with or without LT. Bernal et al. reported survival of 2095 patients with ALF. The survival without LT was 48% in 2004–2008 period, which is considerably better from 17% survival in initial period of study. The survival rate after LT was 86% for same time period (2004–2008).¹ Both deceased donor liver transplantation (DDLT) and living donor liver transplantation (LDLT) are associated with good survival rates.² ALF is uncommon disease; hence, it is difficult to have prospective studies, and most of data are retrospective. In addition, it is not possible to do randomized controlled trials to look for impact of acute kidney injury (AKI) on LT as one arm will not get LT, which is the only life-saving option in very sick patients. In the current review, we discuss prevalence and impact of AKI in patients with ALF, current literature on role of continuous renal replacement therapy (CRRT), and outcomes of LT in presence of AKI.

Pathophysiology of AKI in ALF

AKI is a frequent complication of ALF and is associated with worse prognosis with or without transplantation.^{3, 4, 5, 6, 7, 8} Multiple factors contribute to AKI in patients with ALF that include direct toxic effect of an ingested agent (e.g. acetaminophen), volume depletion, systemic hypoperfusion, sepsis, inflammation, tissue hypoxia, cytokines, damage-associated molecular patterns, and hepatorenal syndrome like pathogenesis in some cases.⁹ Circulating cytokines and damage-associated molecular patterns from hepatocyte death lead to inflammation and necrotic tubular cell death. The kidney is highly susceptible to ischemic injury that results in vasoconstriction, endothelial injury, and activation of inflammatory processes. Sepsis leads to generation of inflammatory mediators derived from pathogens and activated immune cells (known as pathogen or damage-associated molecular patterns), which also mediate host cellular injury. These mediators also affect epithelial and parenchymal cells in addition to immune system cells through pattern recognition receptors. Thus, cells in kidney are affected by these mediators in addition to microcirculatory dysfunction.^{10, 11}

Cross-talk between organs is broadly defined as endogenous homeostatic signaling between vital organs by physiologic and molecular mechanisms (inflammatory cytokines, endothelial injury, and oxidative stress). Kidney is affected by liver, heart, and brain dysfunction, and vice versa is also true,^{12, 13} and there is suggestion of kidney–brain cross-talk that may contribute to acute kidney injury in acute liver failure. While acute kidney injury can lead/contribute to encephalopathy by electrolyte imbalance, toxins accumulation, and effect of inflammation/cytokines; brain dysfunction also contributes to kidney dysfunction by altered sympathetic and parasympathetic outflow to kidneys and by inflammatory cytokines.^{13, 14, 15, 16}

Prevalence of AKI in Patients With ALF

Although, etiology profile of ALF is different in East and West, studies have shown that AKI is common despite different etiology profile. Table 1 shows various criteria to define AKI.^{15, 16, 17} While viral etiologies are a common cause of ALF in the East, drug-induced liver injuries are more common in the West.¹⁸ Acharya et al. conducted a randomized controlled trial on l-Ornithine l-Aspartate (n = 93) versus placebo (n = 92) for patients with ALF. The study did not include very sick patients or patients with baseline renal injury (defined as urine output <400 mL/day and/or creatinine >1.5 mg/dL). Thus, after exclusion of very sick patients or patients with baseline AKI, still 21% patients developed AKI as defined by urine output <400 mL/day and/or creatinine >1.5 mg/dL. The mortality rate was 33.3% patients in placebo arm and 42.4% in l-Ornithine l-Aspartate arm despite exclusion of very sick patients in beginning.¹⁹ Although, the primary aim of the study was to access effect of l-Ornithine l-Aspartate on outcomes, the study provides prospective data from India regarding incidence of AKI in ALF.

Table 1.

Various Criteria to Define AKI (Based on References 15–17).

Criteria	RIFLE	AKIN	KDIGO
Diagnostic criteria		Increase in serum creatinine of ≥0.3 mg/dL or ≥50% rise from baseline within 48 h OR Urine output of <0.5 mL/kg/hour for >6 h	Increase in serum creatinine of ≥0.3 mg/dL within 48 h or 1.5 times of baseline within prior 7 days OR Urine output of <0.5 mL/kg/hour for >6 h
Staging
Risk (RIFLE) or stage 1 (AKIN/KDIGO)	Increase in serum creatinine x 1.5 times or GFR decrease >25% OR Urine output of <0.5 mL/kg/hour for 6 h	Increase in serum creatinine of ≥0.3 mg/dL or increase to ≥150–200% OR Urine output of <0.5 mL/kg/hour for >6 h	Increase in serum creatinine of ≥0.3 mg/dL or 1.5 to 1.9 times OR Urine output of <0.5 mL/kg/hour for 6–12 h
Injury (RIFLE) or stage 2 (AKIN/KDIGO)	Increase in serum creatinine of x 2 times or GFR decrease >50% OR Urine output of <0.5 mL/kg/hour for 12 h	Increase in serum creatinine of >2–3 fold from baseline OR Urine output of <0.5 mL/kg/hour for >12 h	Increase in serum creatinine to 2 to 2.9 times of baseline OR Urine output of <0.5 mL/kg/hour for 12–24 h
Failure (RIFLE) or stage 3 (AKIN/KDIGO)	Increase in serum creatinine x 3 times or GFR decrease >75% OR Increase in serum creatinine to >4.0 mg/dL (acute rise >0.5 mg/dL) OR Urine output of <0.3 mL/kg/hour for x 24 h or anuria for x 12 h OR Initiation of renal replacement therapy	Increase in serum creatinine of >3 fold from baseline OR Increase in serum creatinine to ≥4.0 mg/dL with an acute increase of at least 0.5 mg/dL OR Urine output of <0.3 mL/kg/hour for >24 h or anuria for >12 h OR Initiation of renal replacement therapy	Increase in serum creatinine > 3 times OR to ≥4.0 mg/dL OR Urine output of <0.3 mL/kg/hour for ≥24 h or anuria for ≥12 h OR Initiation of renal replacement therapy
Loss	Persistent acute renal failure (complete loss of kidney function> 4 weeks)
End stage (RIFLE)	Need for renal replacement therapy for >3 months

AKI: acute kidney injury, GFR: glomerular filtration rate.

In a retrospective analysis of 1604 patients with ALF (ALF study group data from 1998 to 2010), 70% of patients had AKI as defined by Acute Kidney Injury Network criteria, 30% received renal replacement therapy (RRT). The authors noted that patients in AKI group had higher international normalized ratio, higher grade of encephalopathy, and need of vasopressors. Patients with acetaminophen-induced ALF had more AKI but had better prognosis than patients with other etiology of ALF.⁵ The studies on ALF and AKI have not reported incidence of AKI alone in absence of other organ failures. The patients with ALF have high INR and may need mechanical ventilation due to advanced encephalopathy in absence of other (extrahepatic) organ failures, thus need of ionotropes should be considered as true organ failure in addition to kidney failure. It appears that AKI in absence of vasopressors is common. Tujio et al. reported that 24.6% patients with AKI stages 1 and 2, and 54.4% on RRT required vasopressors. This study also had patients with shock, a cause of ALF, which required vasopressors.⁵ In another study, 44% of patients in RRT group were not on vasopressors.²⁰ Knight et al. reported sepsis only in 9% of patients in RRT group.²¹ Thus, AKI without sepsis or need of vasopressors occurs in a significant number of patients.

Prognosis of AKI in Patients With ALF

Dhiman et al. showed that renal impairment (defined as reduced urine output of <400 mL in 24 h, with serum creatinine of >1.5 mg/dL) was one of six independent predictors of mortality with an odds ratio (OR) 7.54 (95% confidence interval [CI] 2.53–22.49) in a series of 144 ALF. The mortality was 63.9%. The positive predictive value of serum creatinine >.5 mg/dL was 80% for mortality (48/60 patients died).³ In a study of 2280 adult patients (Organ Procurement and Transplantation Network), data from 2002 to 2012 with ALF listed for LT, 56% had renal dysfunction defined as estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 M² by Chronic Kidney Disease Epidemiology Collaboration creatinine 2009 equation, mortality increased with increasing grade of renal failure.⁶ The authors noted that 6-week survival probabilities on waiting list were 71%, 59%, 56%, 59%, and 42% with stages of 1, 2, 3, 4, and 5 renal dysfunction, respectively. It should be noted that survival probabilities were almost similar in stage 2, 3, and 4 AKI, while it was significantly lower for stage 5 AKI. Similarly, the hazard ratio for mortality as compared with eGFR >90 mL/min was 1.42 for eGFR >60 to <90 ml/min that increased to 2.77 for eGFR <15 mL/min.⁶ Cardoso et al. analyzed US-ALF group data, 314 out of 1186 patients underwent RRT on day 1, 21-day transplant-free survival was 47% in RRT versus 65% in non-RRT group (P < 0.001).²⁰ Table 2 shows important studies highlighting prevalence and impact of AKI in patients with ALF.

Table 2.

AKI in ALF and Impact on Survival.

Author (year)	N, design	Comments
Dhiman RK,4 2007	144 Retrospective	Creatinine≥1.5 mg/dL one of important factors on multivariate analysis, OR 7.54 (95% CI 2.53–22.49), P ≤ 0.0001 48/60 (80%) with creatinine >1.5 mg/dL died
Tujios SR,52015	1604, Retrospective, 1998–2010 data	70% AKI, 30% RRT, transplant-free survival-stage 3 AKI or RRT, 50.2% acetaminophen group, 56.8% shock group, 19.1% other etiologies, total 37.3%
Urrunaga NH,62016	2280 Retrospective, OPTN data from 2002 to 2012	56% had eGFR<60, 6-week survival probabilities in patients with ALF on the liver transplant waiting list were 71%, 59%, 56%, 59%, and 42% with renal dysfunction stages of 1, 2, 3, 4, and 5
Cardoso FS,20 2017	1186, Retrospective, US-ALF group	314 required RRT on day 1, 21 transplant-free survival in 47% versus 65% non-RRT, P < 0.001

eGFR: estimated glomerular filtration rate, OR: odds ratio, AKI: acute kidney injury, RRT: renal replacement therapy; CI: confidence interval.

Specific Therapies for AKI in Patients With ALF

Careful monitoring of hemodynamic parameters is needed with assessment of fluid status. Intravenous contrast studies should be avoided as use of contrast is associated with nephrotoxicity. The use of nephrotoxic medications should be avoided or minimized. The use of N-acetylcysteine may be beneficial in patients with both acetaminophen and non-acetaminophen–induced ALF regarding survival, there are no data to support a renoprotective effect. The mean arterial pressure should be maintained above 75 mm Hg, and norepinephrine should be used as needed.⁹ As infection, systemic inflammatory response is associated with AKI, attempts should be made to early identification and timely treatment of infection.

RRT for AKI in ALF

RRT can remove small water-soluble solutes and toxins from plasma including ammonia. However, this improvement remains modest, and re-accumulation will occur unless liver function improves.²² RRT can be used as intermittent or continuous. CRRT offers several theoretical advantages over intermittent renal replacement therapy (IRRT). CRRT allows more hemodynamic stability, thus allowing more adequate fluid/metabolites removal. Avoidance of hypotensive episodes during CRRT may help in better renal recovery. The prolonged dialysis duration in CRRT should promote higher solute removal because of better mobilization of solutes from extra-plasma compartments. CRRT may help in cytokines removal also. IRRT is easily available and easy to use, because of less duration and is preferred for the acute treatment of life-threatening electrolyte abnormalities or metabolic acidosis.²³ CRRT is more costly because of cost of fluids involved. The ideal timing/dose of CRRT is not defined at present in patients with ALF. CRRT is feasible in intraoperative period also.^{24, 25} CRRT has been shown to be better than IRRT. The beneficial mechanisms of CRRT as compared with IRRT may include slow solute clearance and avoidance of hypotension. If urea is cleared from plasma at faster rate than its movement across cerebral tissues, it will cause an osmotic gradient and increase of intracranial hypertension by moving water back into brain. Patients with ALF are also more susceptible to hypotension.²² Ammonia levels are associated with cerebral edema and survival,²⁶ and CRRT may help in spontaneous recovery or bridging to LT by decreasing ammonia levels.^{6, 7, 8}

In a prospective randomized controlled study, Davenport et al. showed that intermittent form of dialysis was associated with reduction of cardiac index, mean arterial pressure, and tissue oxygen uptake and increase of intracranial pressure. There was no significant change in these parameters in continuous mode of treatment,²⁷ and 2 retrospective studies have looked at role of CRRT in patients with ALF recently.^{20, 28} Deep et al. analyzed 136 pediatric ALF, 45 of these received CRRT before transplantation or recovery. Of these, 19 were successfully bridged to LT and 7 spontaneously recovered. The survival was significantly better in CRRT group when compared with no CRRT group.²⁸ In another multicenter retrospective study from the United States, 340 ALF patients had serial ammonia levels. Sixty-one patients (18%) were on CRRT, while 59 (17%) were on intermittent RRT, and 220 (65%) received no RRT in first 2 days. The authors noted a significant reduction of Ammonia levels between CRRT group and no RRT group. Although, ammonia reduction was not statistically different in CRRT and IRRT groups, the use of CRRT was associated with reduction (odds ratio [OR] 0.47, 95% CI 0.26–0.82) in 21-day transplant-free all-cause mortality while IRRT was associated with increased mortality (OR, 1.68, 95%CI 1.04–2.72).²⁰

Table 3 summarizes data of CRRT in ALF.^{20, 28, 29} European Association for the Study of the Liver guidelines on management of ALF (2017) have added CRRT in the management of these patients.³⁰ The guidelines state that “CRRT should always be undertaken as opposed to intermittent hemodialysis (evidence level III, grade of recommendation 1), and early RRT should be considered for persistent hyperammonemia, control of hypernatremia, and other metabolic abnormalities, fluid balance, and potentially temperature control (evidence level III, grade of recommendation 1).” The use of CRRT may be considered early in countries like India where LT is not available at all centers and cadaveric transplantation is uncommon. There are no good randomized studies on artificial liver support in patients with ALF and majority of uncontrolled studies have not shown a survival benefit.³¹

Table 3.

Role of CRRT in Patients With ALF.

Author (year)	N	Results	Comments
Deep,28 2016	136 total 45 CCRT	Survival significantly better in CRRT group (26/45) versus no CRRT group	Pediatric patients, 19/24 survived after LT and 7/21 survived without LT in CRRT group, CRRT had a significantly increased chance of survival in non-LT pediatric acute liver failure (HR, 4; 95, P = 0.006) Survivors had significant decrease of ammonia and lactate levels
Cardoso FS,20 2017	61 CRRT 59 IRRT 220 no RRT	CRRT was associated with reduction (OR 0.47) in 21-day transplant free mortality, IRRT associated with increased mortality (OR, 1.68)	Median ammonia decreased by 38%, 23%, and 19% with CRRT, IRRT, and no RRT, difference significant (P = 0.007) between CRRT and no RRT
Kenkre,29 2017 (published in abstract form)	47 CRRT 1 SLED	36.4% mean ammonia reduction in 48 h	24/27 LT survived, 11 survived without LT, 10 died

LT: liver transplantation, CRRT: continuous renal replacement therapy, OR: odds ratio, HR: hazards ratio, IRRT: intermittent renal replacement therapy; ALF: acute liver failure.

A multicenter randomized controlled French study by Saliba et al. compared MARS with conventional treatment (n = 53) to conventional treatment (n = 49). The 6-month survival not significantly different between two groups, however, the delay from randomization to liver transplantation was only 16.2 h (median) that precluded meaningful evaluation.³²

LT for ALF and AKI

As ALF is an uncommon disease, and patients with multiorgan failure are generally not taken up for LT, there are limited data on outcomes of LT in presence of ALF with AKI. Randomized studies are not available and cannot be done, as control group cannot be denied of a potential life-saving LT. Most of data are available from cadaveric LT setting, as LDLT centers may not take patients with high risk of adverse outcome for LT because of risk to donors, as reflected by near normal creatinine values in some LDLT studies including a nation-wide survey from Japan.^{33, 34, 35} Barshes et al. analyzed data of LT for ALF from the UNOS database, United States. The study group comprised of a modeling group (n = 972) and a cross-validation group (n = 486). The authors found that serum creatinine >2.0 mg/dL was independent factor for post LT mortality, hazards ratio 1.43 (95% CI 1.05–1.95, P = 0.022).⁷ Other important factors included history of life support, recipient age >50 years, and recipient BMI >30 kg/M².⁷ Two other retrospective analysis also showed worse survival after LT in patients with renal failure in pretransplant period.^{8, 25} LDLT studies have also shown that pretransplant renal dysfunction is associated with increased risk.^{36, 37} Table 4 summarizes outcomes of LT for patients with ALF and AKI. Knight et al. studied effect of pretransplant RRT on outcomes of LT for ALF.²¹ The authors compared outcomes of LT for 336 patients without RRT to 389 patients with RRT. While indications for RRT may be several, patients in RRT group had serum creatinine of 157 (117–237 μmol/L). The 3-year patient and graft survival in RRT group was 77.7% and 72.6%, which was somewhat inferior when compared with 85.1% and 79.4% in non-RRT group, P values being significant for both. Patients with serum creatinine greater than 175 μmol/L in non-RRT had a significant risk of graft failure.²¹ Leithead et al. studied 101 patients, 53.5% fulfilled criteria of acute kidney injury, and majority of these underwent RRT before transplantation. Sixty-three patients of ALF received RRT after transplantation also. In most of patients, renal function (eGFR) recovered to values equal to patients transplanted with cirrhosis with better eGFR and no RRT at baseline. The authors identified older age, female gender, pretransplant-diagnosed hypertension, and cyclosporine as immunosuppressive therapy (and not pretransplant AKI) to be predictors of chronic kidney disease (CKD) after transplantation. Patients transplanted for acetaminophen-related ALF were at lower risk of CKD in long-term follow-up than patients transplanted for other etiologies of ALF. The log rank P value for survival of patients with or without AKI was 0.061.⁸ The recovery of renal function should be related to absence of CKD and short duration of renal injury in these patients. It should be noted that the data of LT in ALF patients with AKI are mainly available from western world where acetaminophen is main etiology of ALF,³⁷ and DDLT is done. It is known that patients with acetaminophen have better recovery of renal functions and have better outcomes than other etiologies.^{38, 39} As stated above, experience is limited for LDLT as many centers may not accept patients with ALF on RRT for LDLT because of slight inferior outcomes in presence of risk to living donors. Kim et al. analyzed outcomes of 30 LDLT for ALF, 11 needed RRT, and 6 of 7 mortalities occurred in patients on RRT.³⁶ Jin et al. also showed renal function as significant factor on multivariate analysis in a predominantly LDLT series.³⁷ More data are needed regarding outcomes of AKI in setting of LDLT and non-acetaminophen etiology. It is not possible to make algorithm based on current literature. We believe that patients without increasing or on high doses of vasopressors and no active sepsis should be considered for LT in presence of AKI. Decision to transplant in presence of RRT is difficult in LDLT setting.

Table 4.

AKI in ALF and Impact on Survival After Liver Transplantation.

Author (year)	N, design	Comments
Barshes NR7Transplantation. 2006	1457 Retrospective 1988–2003 UNOS database	Creatinine >2 mg/dL, HR 1.43 (95% CI 1.05–1.95, P = 0.022), 4 variables important: (1) history of life support; (2) recipient age >50 years; (3) recipient BMI >30 kg/m2; and (4) serum creatinine >2.0 mg/dL
Leithead JA,82011	101, retrospective	preOP AKI had higher mortality postOP, age-adjusted HR 2.09; 95% CI 1.01–4.34, P = 0.048), requirement of postOP RRT: higher mortality (age-adjusted HR 6.22; 95% CI 2.01–19.26, P = 0.002)
Knight SR212016	N = 725, RRT 389, no RRT 336, Retrospective	3-yr patient and graft survival for patients receiving RRT were 77.7% and 72.6% compared with 85.1% and 79.4% for those not requiring RRT (P < 0.001 and P = 0.009. Patients with a preoperative Creatinine >120 μmol/L (1.35 mg/dL) in the absence of RRT had a similar probability
Kim TS,36 2017	30 LDLT, Retrospective	11 on RRT, 7 died including 6 on RRT, RRT predicted mortality (HR 15.1 95% CI 1.80–126.5, P = 0.012

AKI: acute kidney injury, ALF: acute liver failure; HR: hazard ratio, LDLT; living donor liver transplantation, RRT: renal replacement therapy.

AKI is a common problem in ALF. AKI is associated with worse prognosis in patients with ALF. CCRT should be used early in presence of electrolyte imbalance/metabolic abnormalities, persistent hyperammonemia, and to maintain fluid balance. The utility and of CRRT in patients not fulfilling abovementioned criteria needs further evaluation. Also timing, dose, and duration of CRRT need further data. As duration of AKI is short in these patients, there is potential for recovery post-LT. LT, although associated with somewhat lower survival in presence of AKI, should not be deferred in patients with ALF and AKI, in absence of other contraindications in cadaveric transplantation setting, and same cannot be said for living donor liver transplantation in absence of good quality data. Outcomes of living donor liver transplantation for patients with ALF and AKI are not well described, and more data are needed in this setting.

Conflicts of interest

The authors have none to declare.