Postoperative kidney injury does not decrease survival after liver transplantation

PURPOSE: To explore the effect of acute kidney injury (AKI) on long-term survival after conventional orthotopic liver transplantation (OLT) without venovenous bypass (VVB). METHODS: A retrospective cohort study was carried out on 153 patients with end-stage liver diseases transplanted by the Department of General Surgery and Liver Transplantation of the University of Pernambuco, from August, 1999 to December, 2009. The Kaplan– Meier survival estimates and log-rank test were applied to explore the association between AKI and long-term patient survival, and multivariate analyses were applied to control the effect of other variables. RESULTS: Over the 12.8-year follow-up, 58.8% patients were alive with a median follow-up of 4.5-year. Patient 1-, 2-, 3and 5-year survival were 74.5%, 70.6%, 67.9% and 60.1%; respectively. Early postoperative mortality was poorer amongst patients who developed AKI (5.4% vs. 20%, p=0.010), but long-term 5-year survival did not significantly differed between groups (51.4% vs. 65.3%; p=0.077). After multivariate analyses, AKI was not significantly related to long-term survival and only the intraoperative transfusion of red blood cells was significantly related to this outcome (non-adjusted Exp[b]=1.072; p=0.045). CONCLUSION: The occurrence of postoperative acute kidney injury did not independently decrease patient survival after orthotopic liver transplantation without venovenous bypass in this data from northeast Brazil.


Introduction
The conventional technique of orthotopic liver transplantation (OLT) has been pointed as associated with increased rate of postoperative acute kidney injury (AKI) [1][2][3] ; an overall major and frequent complication after OLT [4][5][6][7] .Following these patients, we have not found serious complications inherent to this approach 8,9 ; however, whether this complication may contribute to independently decrease post-transplant survival remains unclear.
Previously, exploring predictors of AKI in a cohort of our patients who underwent conventional OLT without VVB, we observed a trend of lower short-term survival after this approach, but this difference did not reach the level of significance (six months survival of 71.6% vs. 82.2%;p=0.07) 10 .In this current study, we present our analysis with regard to effect of AKI on long-term survival using the same cohort of patients.

Methods
A retrospective cohort study was carried out on patients transplanted by the Department of General Surgery and Liver Transplantation of the Oswaldo Cruz University Hospital, University of Pernambuco, from August, 1999 to December, 2009.This study was approved by our Research on Human Beings Ethics Committee complying with current ethical guidelines.
We limited our analysis to adults and adolescent patients (>16 years) who underwent deceased donor OLT using conventional technique.For patients who had undergone re-transplantations, data were collected from the first procedure only.We also excluded patients transplanted due to fulminant hepatic failure, recipients of modified techniques (i.e.: split-liver, sequential and venovenous transposition) or liver-kidney transplants, as well as patients with incomplete data in their medical records or deceased during surgical procedure.

Conceptual model
MELD score was calculated using laboratory results collected immediately prior to the LT with no adjustments for malignancy or other ''special'' conditions used to prioritize these patients on the waiting list.Serum markers were used to confirm viral hepatitis diagnosis.The preoperative diagnosis of hepatocellular carcinoma (HCC) was based on the Barcelona-2000 conference diagnostic criteria 11 and fits with Milan criteria to select patients for liver transplantation 12 .Donor livers were recovered using standard procurement techniques.Collins and Belzer UW (University of Wisconsin) solutions were used for perfusion procedures until 2007, when the latter was replaced by Celsior solution.The same surgical team performed all OLT using conventional (retrohepaticcaval resection) techniques without VVB based on surgeons' preference according to clinical and surgical parameters.After transplantation, tacrolimus, mycophenolate (sodium or mofetil) and prednisone were used as immunosuppressive treatment, with no major changes in the protocols applied.All patients were followed up routinely or when it was necessary because some clinical demand and the medical record were taken up to May, 2012.
We considered postreperfusion syndrome (PRS) a decrease in mean arterial pressure greater than 30% below the baseline for a minimum of one minute during the first five minutes of the reperfusion 13 and classified marginal grafts based on Briceño's score 14 .AKI was defined as a serum creatinine level of ≥1.5mg/dl or urinary output of <500ml/24h within the first three days following OLT, according to RIFLE criteria 15 .Early postoperative mortality were accessed in the first 30-days after OLT as we had published before 8,9 .

Analytic approach
Descriptive statistics were summarized as medians (ranges) or frequencies (percentages) and analyzed using Mann-Whitney U-test and chi-square tests, including Yates's correction or Fischer's exact test as appropriated.The Kaplan-Meier survival estimates and log-rank test were applied to explore the association between AKI and long-term patient survival, and multivariate analyses were also applied to control the effect of other variables.
In these settings, the association of each demographic variable with survival was first assessed using univariate Cox's proportionalhazards models or log-rank test.Then, variables whose association showed p-value <0.10 were tested in a multivariate Cox's model considering a p-value <0.05 as statistically significant.We additionally adjusted our multivariate analyses for temporal variables such as transplantation period (three consecutive periods including 53 patients) and allocation criteria (chronologic vs.

Discussion
The conventional technique of OLT without VVB is considered simpler and has been safely performed with recognized swiftness by our team even in patients suffering of advanced liver disease.Amongst our patients underwent transplantation by this approach, we have reported AKI rates of 39.2% 10 , according to RIFLE criteria 15 .In line with our findings, Tinti et al. 15 reported AKI occurred in 37.5% of patients, using these same criteria.Similarly, when comparing conventional vs. piggy-back techniques, Cabezuelo et al. 1 found AKI rates of 39% associated to the former.However, these rates may reach 50% in some Centers, where it has been linked to poor survival outcomes after OLT 4 .
Accordingly, even in those who had normal renal function before surgery, postoperative AKI may occur in 30% of patients receiving OLT 5 .

Several variables have been pointed as risk factors
or predictors of AKI.Often, patient/liver disease severity (i.e.: as measured by MELD score or Child-Pugh class) 7,15,16 , hypoalbuminemia 15,17 , preoperative decreased renal function 7,17 , hemocomponents transfusion 1,5,18 , postreperfusion syndrome 5,19 and surgical technique 1,2 have been highlighted in the most current literature.Some preoperative co-morbidity such as hypertension, obesity (i.e.: increased BMI) and diabetes mellitus have also been related to outcomes after OLT when the renal dysfunction is explored in the settings of OLT 4,7,20 .Using this same cohort, we previous presented that patient/liver disease severity (i.e.: as measured by Child-Pugh) was the main factor related to AKI after conventional OLT without VVB 10 .In this settings, MELD score was not found as a predictor of AKI 10 , which probably result from the poor accuracy of this score (i.e.: as a continuous variable) in predicting post-OLT outcomes in low degrees of hepatic dysfunction (i.e.: low median MELD scores) 8,21 .Thus, we hypothesized that higher potential of Child-Pugh classification to point patients at increased risk to develop postoperative AKI possibly results from adding some clinical parameters, such as ascites and encephalopathy, that comprehensively best reflect the complex and dynamic interaction between liver and kidney dysfunction in patients suffering of end-stage liver disease.
Regarding the effect of surgical technique on posttransplant renal function, several authors demonstrate that the piggy-back technique significantly reduces the probability of AKI after OLT [1][2][3]22 . Howver, the benefit of this surgical approach appears to be not fully achieved when it is used with VVB 3 and there is currently no clear evidence to recommend or the use of piggy-back method of OLT in general, according to a recent systematic review from Cochrane Database 23 .At our Department, we have performed OLT without VVB using either conventional or piggy-back techniques 9,24 .Recently, we retrospectively reviewed our data in order to compare these different approaches and no significantly differences were noted in relation to kidney function 9 .
In our experience, AKI appears mainly related to patient severity than to the surgical technique itself.In other words, we clearly observed that AKI was associated to variables linked to disease severity (see Tables 1 and 2) that may serve to identify those sickest patients, for whom poorer outcomes are usually expected.
In this present study, we avoided dichotomizing most of the continuous variables in order to minimize some loss in the statistical power of our analysis and the occurrence of residual confounding factors 25,26 .Similarly, we applied multivariate analysis to control the effect of each variable for the others and also adjusted these tests for temporal variables such as allocation criteria and transplantation periods that have influenced survival outcomes at our Department, as previously reported 8,21 .Herein, separate regression models were first fit to each group and the loglikelihoods for those models were summed up.This log-likelihood was then compared to that of the overall model (collapsed across groups).Accordingly, AKI did not significantly correlated with long-term survival and only the intraoperative transfusion of red blood cells appeared to significantly influence this outcome.
Although the need for blood transfusion may act only as a marker of disease severity of the patients undergoing liver transplantation 18 , lower post-transplantation survival 8,21,[27][28][29][30][31] mainly stems from transfusion-related immunomodulation (TRIM syndrome) and changes of its components resulting from the storage process 32 , which may increase the risk of nosocomial infections, acute lung injury and development of late autoimmune diseases 33 .
In agreement with some of our previous reports 8,21 , this study revealed the trend of poorer survival associated to AKI probably resulted from the higher volume of intraoperative red blood cell transfusion that accompanies this complication, suggesting that transfusion of blood components might be independently able to decrease patient survival after OLT without VVB.
Our study has several strengths.In the first place, our own database has been prospectively maintained and continuously updated.Secondly, we have accumulated considerable experience with the management of marginal grafts 34,35 and use of conventional technique without VVB 8,9,24 .Additionally, to our best knowledge, this is the longest series in the Brazilian literature discussing a single-center experience in AKI after OLT without VVB and its consequences for long-term patient survival.

Conclusions
The occurrence of postoperative acute kidney injury (AKI) did not independently decrease patient survival after orthotopic liver transplantation (OLT) without venovenous bypass (VVB) in this data from northeast Brazil.Although being a common postoperative complication in these settings, our study suggests that AKI might work just as a marker of greater disease severity and its related poorer survival appears mainly related to patient severity than to occurrence of AKI itself.

FIGURE 2 -
FIGURE 2 -Kaplan-Meier analysis of patients stratified according to development of acute kidney injury.Log-rank test did not show statistically significant difference between patient survival at 5-years (51.4% vs. 65.3%;p=0.077).

TABLE 1 -
Baseline characteristics and descriptive statistics of recipient-related variables.
1Comparisons between groups using the Mann-Witney U test or chisquare tests.

TABLE 2 -
Baseline characteristics and descriptive statistics of Center-related variables.
1Comparisons between groups using the Mann-Witney U test or chisquare tests.4According to Briceño score for marginal liver graft.