- Top of page
Renal dysfunction of acute liver failure (ALF) may have distinct pathophysiological mechanisms to hepatorenal syndrome of cirrhosis. Yet, the impact of perioperative renal function on posttransplant renal outcomes in ALF patients specifically has not been established. The aims of this study were (1) to describe the incidence and risk factors for chronic renal dysfunction following liver transplantation for ALF and (2) to compare renal outcomes with age–sex-matched patients transplanted for chronic liver disease. This was a single-center study of 101 patients transplanted for ALF. Fifty-three-and-a-half percent had pretransplant acute kidney injury and 64.9% required perioperative renal replacement therapy. After transplantation the 5-year cumulative incidence of chronic kidney disease (eGFR <60 mL/min/1.73 m2) was 41.5%. There was no association between perioperative acute kidney injury (p = 0.288) or renal replacement therapy (p = 0.134) and chronic kidney disease. Instead, the independent predictors of chronic kidney disease were older age (p = 0.019), female gender (p = 0.049), hypertension (p = 0.031), cyclosporine (p = 0.027) and nonacetaminophen-induced ALF (p = 0.039). Despite marked differences in the perioperative clinical condition and survival of patients transplanted for ALF and chronic liver disease, renal outcomes were the same. In conclusion, in patients transplanted for ALF the severity of perioperative renal injury does not predict posttransplant chronic renal dysfunction.
Renal dysfunction is a common complication of acute liver failure (ALF) with two-thirds of patients manifesting acute kidney injury, and almost half requiring renal replacement therapy (1). Many have postulated that the pathogenesis is similar to the hepatorenal syndrome of cirrhosis (2,3). However, a growing body of evidence supports a systemic inflammatory response to ALF, and the systemic inflammatory response syndrome (SIRS) is an independent predictor of acute kidney injury in ALF patients (1,4,5). It follows that the renal dysfunction of sepsis may be a more accurate parallel than the hepatorenal syndrome (1). Additional factors that may contribute to renal dysfunction in ALF but are less likely in stable cirrhotic patients include hypovolemia, nephrotoxic drugs particularly acetaminophen, infection and disseminated intravascular coagulation (6–8).
Despite the contrasting perioperative clinical condition of patients transplanted for ALF and chronic liver disease (CLD), post-liver transplant renal outcomes have not been examined specifically in this group. Pretransplant glomerular filtration rate, pretransplant renal failure requiring renal replacement therapy and acute renal injury are consistent predictors of chronic renal dysfunction after elective liver transplantation (9,10). Given the greater baseline circulatory and neuro-humoral derangement of ALF it seems possible that the acute hemodynamic effects of the calcineurin inhibitors administered immediately following transplantation are exaggerated (11–14). On the other hand, the differing pathophysiological mechanisms could offer relative reno-protection and a reduced risk of chronic kidney disease.
The clarification of the impact of liver transplantation for ALF on posttransplant renal function has important implications for patient management. Chronic renal dysfunction is a major cause of patient morbidity and mortality and the minimization of renal injury has emerged as a priority for transplant physicians (9,15–17). Simultaneous liver–kidney transplantation is not an option in patients transplanted for ALF because of the medical urgency, but the identification of prognostic variables could help to determine those who may benefit from tailored renal sparing immunosuppressive regimens (18).
The aims of this study were first to describe the incidence and risk factors for chronic renal dysfunction following liver transplantation for ALF and second to compare renal outcome with an age–sex-matched group of patients transplanted for CLD.
- Top of page
This was a retrospective single-center study of consecutive patients who underwent super-urgent liver transplantation for ALF (UK Transplant Super Urgent Scheme Category 1–7) between December 1992 and July 2007 (19). Eight patients had inadequate documentation available and were excluded from the analysis. A further 1 patient was lost to follow-up. Therefore, the study cohort comprised 101 patients. The causes of ALF were acetaminophen (46 patients, 45.5%), seronegative hepatitis (27 patients, 26.7%), idiosyncratic drug reaction (11 patients, 10.9%), autoimmune hepatitis (7 patients, 6.9%), hepatitis B (4 patients, 4.0%), Budd-Chiari (3 patients, 3.0%), Wilsons disease (2 patients, 2.0%) and hepatitis A (1 patient, 1.0%).
ALF was defined as severe liver injury with hepatic encephalopathy in which the onset of encephalopathy was within 8 weeks for the first symptoms of illness, and in the absence of preexisting liver disease (20).
Data were collected on the following preoperative variables at the time of listing: age, gender, race, liver disease etiology, additional comorbidity, smoking status, international normalized ratio (INR), serum bilirubin, albumin, serum creatinine, serum sodium (hyponatremia; sodium <135 mmol/L) and presence of ascites (on ultrasound). SIRS was defined as ≥2 of temperature <36°C or >38°C, heart rate >90 beats per minute, white cell count <4 × 109/L or >12 × 109/L and PaCO2 < 4.3kPa at the time of admission (21). Documented perioperative variables were peak preoperative serum creatinine, preoperative renal replacement therapy, postoperative renal replacement therapy, inotropes (noradrenaline/adrenaline), bacterial sepsis and fungal sepsis. Immunosuppression was noted and calcineurin inhibitor trough levels at 1 week, 1 month and 12 months (a comparable 12-month value for the linear regression analysis was obtained for all patients regardless of calcineurin inhibitor by expressing the trough as relative to the median value). Renal function was recorded at 1 month, 6 months, 12 months and 2, 3, 4 and 5 years following transplantation. Patients still receiving renal replacement therapy at 1 month were given an arbitrary serum creatinine of 350 μmol/L and an estimated glomerular filtration rate (eGFR) of 15 mL/min/1.73 m2.
A patient was considered to have significant renal dysfunction preoperatively if they fulfilled the RIFLE criteria for acute kidney injury: peak serum creatinine ≥2 times the baseline level (22). The baseline serum creatinine was unavailable for most patients and was estimated as previously described (1,22). Following transplantation the main measure of renal function was eGFR, determined using the Modification of Diet in Renal Disease (MDRD) Study 4-variable equation (eGFR = 186 × creatinine (mg/dL)−1.154× age (years)−0.203× 1.210 (if black) × 0.742 (if female) (23). Chronic kidney disease was defined as eGFR < 60 mL/min/1.73 m2 on at least 2 occasions from 6 months posttransplant onwards: stage 3, stage 4 and stage 5 chronic kidney disease were defined as eGFR 30–59 mL/min/1.73 m2, 15–29 mL/min/1.73 m2 and <15 mL/min/1.73 m2 or on dialysis, respectively (24).
To examine whether the renal dysfunction of ALF has a different renal prognosis after transplantation to the renal dysfunction of CLD a control group of patients transplanted for CLD was identified. These patients were age-matched (to within 5 years) and sex-matched to the original cohort. The relatively young age of the patients transplanted for ALF meant that only 71 patients could be appropriately matched. The causes of CLD were primary biliary cirrhosis (18 patients, 25.4%), alcohol (10 patients, 14.1%), chronic active hepatitis (9 patients, 12.7%), sclerosing cholangitis (9 patients, 12.7%), cryptogenic cirrhosis (9 patients, 12.7%), hepatitis C (5 patients, 7.0%) and other (11 patients, 15.5%). Three patients (4.2%) were transplanted for hepatocellular carcinoma. None of the control patients had intrinsic renal disease prior to transplantation and no patient underwent combined liver–kidney transplantation.
Immunosuppression was similar for patients transplanted for ALF and for CLD, and consisted of a calcineurin inhibitor, azathioprine and prednisolone in most cases. Midway through the specified time period the unit policy for calcineurin inhibitor changed from cyclosporine to tacrolimus. Prednisolone was usually discontinued by 3 to 6 months posttransplant unless otherwise indicated. Deviation from the protocol occurred only in the setting of adverse event or graft rejection. Acute rejection was usually managed with 1 g of methyl-prednisolone intravenously for 3 days followed by reintroduction of oral steroids with or without increased dose of, or switch to, alternative calcineurin inhibitor. Chronic rejection was managed with the latter and in a small number of patients azathioprine was changed to mycophenolate. Interleukin (IL)-2 receptor antagonist induction therapy was not administered to any of the patients.
Cumulative incidence of chronic kidney disease was estimated using the Kaplan–Meier method. Survival was estimated using Kaplan–Meier plots with log-rank test for differences, and age-adjusted survival was determined using Cox proportional hazards analyses. Normally distributed continuous variables and nonparametric continuous variables were compared using the Student's t-test and Mann–Whitney test, respectively. Chi-squared analysis or Fisher's exact test were used for comparison of categorical data. A multivariate linear regression analysis was performed to explore the relationship between perioperative renal dysfunction and long-term renal function following transplantation. Clinically relevant factors were included simultaneously with 12-month eGFR as the dependent variable. Cox proportional hazards analysis was then used to identify variables predictive of chronic kidney disease by 5-years posttransplant. Three multivariate models were constructed with all clinically relevant factors entered simultaneously. Variables entered into Model 1 were age, gender, pretransplant diagnosed hypertension, category of ALF (acetaminophen-induced vs. nonacetaminophen-induced), SIRS, calcineurin inhibitor at time of hospital discharge and pretransplant acute kidney injury. In Models 2 and 3 acute kidney injury was replaced by the other measures of perioperative renal dysfunction, peak preoperative change in serum creatinine and immediate posttransplant renal replacement therapy, respectively. All three measures of perioperative renal dysfunction were not included in the same model because of collinearity. None of the multivariate models was adjusted for the presence of pretransplant diabetes mellitus secondary to small patient numbers. p < 0.05 was considered statistically significant unless otherwise stated. Data were analyzed using the SPSS 15 package (SPSS Inc., Chicago, IL, USA).
All values are expressed as mean and standard deviation (SD), and median and interquartile range (IQR) as appropriate.
- Top of page
In this large single-center study we have described for the first time the risk factors for chronic renal dysfunction following emergency liver transplantation for ALF. Importantly, we have shown that perioperative kidney injury does not appear to have negative consequences for long-term renal function in this population. Contrary to observations in CLD patients pretransplant acute kidney injury and renal replacement therapy were not associated with chronic kidney disease. Only failure to recover renal function, as evidenced by eGFR at 1-month posttransplant, was a predictive factor. Despite marked differences in the perioperative clinical condition of patients transplanted for ALF and CLD long-term renal outcome following transplantation was the same.
The rate of chronic kidney disease after transplantation for ALF was similar to that reported by Aberg et al., the single other publication of renal function in this setting (25). Although half of our patients fulfilled the criteria for acute kidney injury pretransplant, and more than 60% required renal replacement therapy during the immediate postoperative period, only 21% had an eGFR less than 60mL/min/1.73 m2 12 months thereafter. This dramatic renal recovery echoes clinical observations in spontaneous survivors of ALF. By 5-years posttransplant the cumulative incidence of chronic kidney disease was 42%.
The identical posttransplant renal function of ALF and CLD patients was unexpected when considering our current understanding of the underlying mechanisms (11–14,26). Based on the traditional hypothesis of hepatorenal syndrome one might predict that severe perioperative renal vasoconstriction would exacerbate calcineurin-inhibitor mediated kidney dysfunction (11–14,26). Cyclosporine and tacrolimus cause an initially hemodynamic dose-dependent renal impairment that is feasibly exaggerated in patients with greater baseline circulatory and neuro-humoral derangement (11–14). Our results support an alternative patho-physiological process underlying the renal injury that occurs in ALF.
We have previously demonstrated that SIRS predicts the development of acute kidney injury in patients with nonacetaminophen-induced ALF, a relationship that appears to be independent of the severity of liver injury (1). Consequently, we have postulated that the renal dysfunction of sepsis may be a more accurate parallel than the hepatorenal syndrome of cirrhosis (1). In fulminant hepatic failure the systemic inflammatory response may be the key mediator of renal impairment. Patients with subfulminant ALF are more likely to have clinically significant portal hypertension, and may develop ascites (27). Therefore, this group may share some of the hemodynamic and neuro-humoral features of hepatorenal syndrome (1). In sepsis, kidney injury may occur in the setting of preserved or even increased renal perfusion, which is in contrast to the intense renal vasoconstriction of hepatorenal syndrome (26,28). We propose that relative renal hyperemia may help to minimize the renal hemodynamic response to calcineurin inhibitors and explain the comparable long-term posttransplant renal function demonstrated by ALF patients (14).
Alternatively, the failure of perioperative renal dysfunction to impact on long-term posttransplant renal outcomes may reflect the duration of renal impairment. In patients transplanted for CLD renal dysfunction duration appears to be a key determinant of chronic renal impairment. Campbell et al. demonstrated that renal dysfunction duration of greater than 3.6 weeks pretransplant was an appropriate cut-off to identify patients at risk of renal insufficiency 12-months thereafter (29). In our cohort the renal injury, although more severe, was on the contrary short-lived.
In the nontransplant population acute kidney injury is a risk factor for chronic renal dysfunction. For example, in patients who undergo major vascular surgery the occurrence of perioperative acute kidney injury is associated with an increased risk of chronic kidney disease (30). Furthermore, patients requiring dialysis for acute kidney injury who are dialysis-independent at the time of hospital discharge are three times more likely to develop end-stage renal failure (31). Animal studies have confirmed that acute kidney injury can cause permanent structural kidney damage with progressive tubulo-interstitial fibrosis and long-term implications for renal function (32). Our failure to show a relationship between perioperative renal dysfunction and posttransplant chronic kidney disease is not in accordance with these observations. Acute kidney injury is an independent predictor of mortality in patients with ALF, and following transplantation for ALF (1). Yet, our findings suggest that beyond hospital discharge acute renal impairment, if short-lived, does not impact particularly on chronic renal function.
Acetaminophen as the cause of ALF was associated with a higher absolute eGFR at 12-months following transplantation and a reduced risk of chronic kidney disease. Acetaminophen is an independent predictor of acute kidney injury in patients with ALF and there are case reports of renal failure following acetaminophen overdose in the absence of significant hepatic injury (1,33,34). Animal models support a direct nephrotoxic effect although the mechanism remains unclear (33). It has been hypothesized that a locally produced metabolite induces proximal tubular cell necrosis while functional renal effects may also contribute (33,35,36). Our findings support the reversibility of acetaminophen-induced nephrotoxicity (37,38).
The study has some potential limitations that should be mentioned. First, baseline renal function was only available in a small number of ALF patients and it is possible that a proportion could have had undiagnosed intrinsic renal disease. The patients studied were of a relatively young age and it is assumed that premorbid renal function was normal. Second, nephrotoxic medications could have influenced the severity of perioperative renal dysfunction. Our unit avoids nephrotoxic drugs, yet this does not preclude exposure prior to transfer. Third, although our study consists of one of the largest single center cohorts of patients transplanted for ALF it remains possible that the relatively small numbers may have influenced our results.
With regards the CLD group, only 70% of the ALF patients could be matched because of the young age range. Furthermore, the pretransplant eGFR was only available at the time of listing and not immediately prior to transplantation. Pretransplant kidney function may, therefore, have been over represented in the CLD patients if there was a significant deterioration on the list. Nevertheless, no CLD patient required preoperative renal replacement therapy or reassessment for combined liver–kidney transplantation and, given the relatively short median waiting-list time of 52 days, it seems unlikely that this data would have influenced the results. The lack of pretransplant renal impairment in the control arm may also raise some concerns about its generalizability for a standard population of liver transplant recipients. This largely reflects the younger age of the patients. However, those with intrinsic renal disease or who received a simultaneous liver–kidney transplant were also deliberately excluded; we wished to examine whether the physiological differences between ALF and CLD would influence renal outcomes. Of course, it is well recognized that eGFR is not an accurate measure of renal function in patients listed for elective liver transplantation, tending to overestimate when the true GFR is reduced (39). Sixty percent of the CLD patients had ascites and one-third had hyponatraemia, indicating a high prevalence of portal hypertensive-related renal impairment (40). Finally, it is difficult to ensure retrospectively that ALF and CLD patients received similar immunosuppressive regimes. However, during the period studied our unit had a single protocol that was rarely deviated from with calcineurin inhibitor administration within 24 h of transplantation. The similar posttransplant calcineurin inhibitor trough levels support this claim.
The findings of our study have important implications for patient management. Patients who undergo liver transplantation for ALF should not be considered a high-risk group for developing chronic kidney disease even when perioperative acute renal impairment is severe. Consequently, we do not support the routine use of interleukin-2 receptor antagonists and delayed introduction of the calcineurin inhibitor in this setting (18). Renal sparing immunosuppression such as mycophenolate and reduced dose tacrolimus could be considered in select patients, for example older females transplanted for nonacetaminophen-induced ALF (41).
In conclusion, in this large single-center study of patients transplanted for ALF we have shown that the severity of perioperative renal dysfunction was not predictive of posttransplant chronic kidney disease. Despite greater perioperative physiological derangement in ALF patients when compared with an age–sex-matched cohort transplanted for CLD renal function following transplantation was the same.