CIRRHOTIC PATIENTS WITH ACUTE KIDNEY INJURY (AKI) HAVE HIGHER MORTALITY AFTER ABDOMINAL HERNIA SURGERY

ABSTRACT 
Background:
 The incidence of abdominal hernia in cirrhotic patients is as higher as 20%; in cases of major ascites the incidence may increase up to 40%. One of the main and most serious complications in cirrhotic postoperative period (PO) is acute kidney injury (AKI). 
Aim: To analyze the renal function of cirrhotic patients undergoing to hernia surgery and evaluate the factors related to AKI. 
Methods:
 Follow-up of 174 cirrhotic patients who underwent hernia surgery. Laboratory tests including the renal function were collected in the PO.AKI was defined based on the consensus of the ascite´s club. They were divided into two groups: with (AKI PO) and without AKI . 
Results:
 All 174 patients were enrolled and AKI occurred in 58 (34.9%). In the AKI PO group, 74.1% had emergency surgery, whereas in the group without AKI PO it was only 34.6%.In the group with AKI PO, 90.4% presented complications, whereas in the group without AKI PO they occurred only in 29.9%. Variables age, baseline MELD, baseline creatinine, creatinine in immediate postoperative (POI), AKI and the presence of ascites were statistically significant for survival. 
Conclusions:
 There is association between AKI PO and emergency surgery and, also, between AKI PO and complications after surgery. The factors related to higher occurrence were initial MELD, basal Cr, Cr POI. The patients with postoperative AKI had a higher rate of complications and higher mortality.


INTRODUCTION
Cirrhosis represents a late stage of progressive hepatic fibrosis characterized by distortion of the hepatic structure and the formation of regenerative nodules. It´s physiopathology involves progressive hepatic lesions resulting in portal hypertension and decompensation, including ascites, spontaneous bacterial peritonitis, hepatic encephalopathy, variceal bleeding, hepatorenal syndrome 12 , hepatocellular carcinoma 17 and colon dysfunction 6 . It is the 8 th most common cause of mortality in the United States and 13 th worldwide, with 45.6% increase in mortality. It is generally considered irreversible in its more advanced stages, for which the only definitive treatment is liver transplantation 7 .
One of the principal and most serious complications of cirrhosis is acute kidney injury (AKI) 11 . Its appearance at any phase of cirrhosis is associated with greater morbimortality. Often, its progression leads the patient to renal replacement therapy (dialysis), either temporarily or permanently. In the long term, kidney transplantation may be necessary 13 .Almost 20% of patients with cirrhosis will develop AKI over the course of hospitalization, with mortality rates as high as 50-90%(3). Over 60% of all cases of AKI in patients with cirrhosis are attributable to prerenal factors, such as hypovolemia, hypoperfusion, massive renal vasoconstriction by systemic arterial and splanchnic vasodilation 5 .
Many recent studies have assessed the incidence of AKI, its evolution and treatment in patients with cirrhosis 19 . There is a lack, however, in the literature assessing the incidence of AKI in patients with cirrhosis, either with or without ascites, who have been submitted to surgical procedures.
The objective of this study was to analyze the renal function of patients with cirrhosis submitted to a correction of abdominal wall hernias. Furthermore, it aimed to compare patients presenting acute kidney injury with others, to determine risk factors for its occurrence. The ones submitted to elective and emergency surgery were assessed and followed. The indications for emergency surgery were: hernial strangulation, painful irreducible hernia, skin perforation with extravasation of ascitic fluid or ischemic ulceration of the skin. These cases were submitted to emergency herniorrhaphy. The following demographic and clinical variables of the patients were assessed: gender, age, BMI (body mass index), etiology of cirrhosis, Child score, MELD score, patient's inclusion on the liver transplantation waiting list, ascites volume, diabetes, presence of TIPS(transjugular intrahepatic portosystemic shunt), types of hernia, and the presence of previous complications from hernias. The data were collected prior to surgery, in the pre-operative phase, or at hospitalization in emergency cases.

This study was approved by the Scientific Ethic
The laboratory tests were collected immediately before the surgery were: creatinine, INR (international normalized ratio), total bilirubin, sodium. Ten grams of human albumin were intravenously administered per liter of ascitic fluid removed during surgery. In cases of voluminous ascites in the postoperative phase, a relief parecentesis was carried out where there was tense ascites, and always with albumin expansion when the volume was over 5 l, in a dose of 10g of albumin per liter removed 10 .
Patients were divided into two groups: patients with AKI and without AKI after surgery.
Laboratorial tests were collected in the immediate postoperative phase, on the 1 st , 3 rd and 5 th postoperative days: creatinine, INR, total bilirubin and sodium. The patients not displaying complications were discharged on the 5 th postoperative day. In cases where there were significant alterations in results, or a clinical complication, the exams were carried out daily. Diuretics were not introduced in the early postoperative phase for these patients. It was done only at hospital discharge, or after the 5 th day without AKI signs. The patients who developed AKI received treatment with albumin, with initial doses of 10g administered three times per day 10,18 . Patients who showed no improvement of renal function after albumin treatment received terlipressin in initial doses of 0.5mg, four times per day.

Statistical analysis
Quantitative variables are presented as mean, median, standard-deviation, range, interquartile range. The qualitative variables are presented as absolute frequency and percentage. To verify the behavior of the variables in each group across time, ANOVA was used for repeated, non-parametric measures. To verify the differences between groups of quantitative variables, the Kolmogorov-Smirnov was first used to assess normality. When the supposition of normality was not rejected, the Student's t-test was used, if it was, the Wilcoxon-Mann-Whitney was used.
Fisher's exact test was used to assess the relation between qualitative variables.
Kaplan-Meier estimates were used to create the survival curves for each group, and the log-rank test to assess whether the variables were different. Cox's model was used to verify which quantitative variables were associated with mortality. The level of significance adopted was 5%.

RESULTS
Were included 174 patients who underwent procedures to correct hernias of the abdominal wall. The mean follow-up time was 2.58 years, with a median of 2.3 years (8-2241 days). The mean age was 53.71±11.66 years ( Table 1). In the population, 135 patients were male, corresponding to 77.6%. With regards to hernia type, 126 patients had umbilical hernias (72.4%). This was the most common hernia in the evaluated population, and 58 had inguinal hernias (33.3%). Of these, 14 (8%) were bilateral and 10 (6.0%) had incisional hernias.

Comparison between patients with and without postoperative AKI
A total of 58 patients (34.9%) had postoperative AKI. Table 2 shows the base creatinine and creatinine of each group in the immediate postoperative phase, and on the 1 st , 3 rd and 5 th postoperative days. The interaction was statistically significant. Creatinine behaved differently over time, depending on the group. In the group that had postoperative AKI, there was an increase in creatinine over time, while in the group that did not have postoperative AKI there was a small increase, followed by a decrease (after Cr3, Figure 1). CREAT=creatinine;POI=immediate postoperative; Anova for non-parametric repeated measures; red line patients with AKI; black line without AKI.

Comparison of groups (postoperative AKI)
The variables were tested in relation to the presence of postoperative AKI.
There was a significant difference between the groups in the following variables: initial MELD, base creatinine and immediate postoperative creatinine (Table 3). In all variables the postoperative AKI group had higher means than the group without postoperative AKI. There was no difference observed in age or BMI. There was no relation between etiology and postoperative AKI.
In the AKI group, 37.9% had moderate ascites, while in the group without AKI, only 15.7% did (p=0.0001), indicating an association between the degree of ascites and postoperative AKI.
In the AKI group, 69.1% had undergone paracentesis, while, in the group without AKI, only 38.1% had (p=0.0002). Moreover, in the AKI group, a larger number of patients underwent emergency surgery (Table 4). In patients with postoperative AKI, there was a larger proportion of individuals with postoperative complications ( Table 4).

Analysis of survival
The variables were analyzed in relation to patient survival. The presence of ascites, surgery, surgical complications and postoperative AKI were significantly related to survival (Table 5). The variables age, initial MELD, base creatinine and creatinine in the immediate postoperative phase were statistically significant, meaning that the risk ratio is different to 1. For each year of age, the risk of mortality increases by 1.0328 (or increases by 3.28%). In base creatinine, for each increase in unit, the risk of mortality increases by 1.514 (Table 6). Three different models for multivariate analysis were carried out using Cox models. The variables in each model were: model 1: age, initial MELD and base creatinine; model 2: age, initial MELD and IPO creatinine; model 3: age, initial MELD and creatinine delta (difference between IPO creatinine and base creatinine).
In model 1, the variable base creatinine was not statistically significantly related. The other variables had a risk ratio greater than 1, indicating that they were risk factors for mortality (Table 7). In model 2,immediate postoperative creatinine was not statistically significant. The other two factors again had a risk ratio greater than 1, indicating that they are risk factors for mortality (Table 7). In model 3,delta creatinine was not statistically significant. The other two factors again had a risk ratio greater than 1, indicating that they are risk factors for mortality (Table 7).  There is a lack of perioperative studies that show the assessment and evolution of the results of surgery in patients with cirrhosis and complications (9), and this study aimed to identify risk factors for AKI in this population.
This surgery is carried out more frequently today than in the past, in part due to the increased long term survival of patients with cirrhosis. Recent studies have concentrated on estimating perioperative risk in patients with liver disease 15 .
Hemodynamic instability in the perioperative period can worsen hepatic function in patients with cirrhosis 8 . The operative risk is correlated with the severity of liver disease and the type of surgical procedure. Detailed evaluation is necessary prior to elective surgery in these patients. The estimated perioperative mortality is unknown due to the retrospective nature and selection bias in published clinical studies. Child-Pugh classification, and particularly MELD score give reasonable estimates for perioperative mortality risk, but do not supplant the need for careful preoperative preparation and postoperative follow up, such as early detection of complications, which is essential to improve outcomes 2 .
The incidence of complications in this study was 44.8%, with ascites leakage through the operative wound, hematoma of the operative wound and skin dehiscence as the most common. This reflects the high rate of complications in this group of patients, as shown in the literature 2 .
The majority of hernias operated in this study were umbilical, representing 72.4% of procedures. This prevalence in patients with cirrhosis is also reflected in the literature 1 .
Patients with cirrhosis have a high incidence of abdominal wall hernias, and following surgical correction, have a higher morbimortality. The ideal surgical strategy, as well as time to correct abdominal wall hernias is controversial 14,16 . The advantage of having this knowledge would be to predict which patients are at elevated risk of developing postoperative AKI. This would allow a choice wherever possible, of better preparation before the operation, or attempting a prior intervention or treatment to prevent hernia's appearance.
A cohort study by Andraus et al. 1 evaluated 67 patients with cirrhosis submitted to hernial correction. The median MELD score was 14 (6-24). Emergency surgery was carried out in 34 patients due to hernial rupture (n=13), incarceration (n=10), strangulation (n=4) and skin necrosis or ulceration (n=7). Elective surgery was carried out in 27 cases. After a multivariate analysis, emergency surgery (OR 7.31; p=0.017) and Child Pugh C (OR 4.54; p=0.037) were risk factors for serious complications. Furthermore, emergency surgery was an independent risk factor for 30 days mortality (OR 10.83; p=0.028). The risk of morbidity and mortality is associated with emergency surgery in patients with advanced cirrhosis.
As mentioned, majority of patients in this study was Child B (53.4%), followed by Child A (25.3%) and C (19%) and the MELD score varied between 6 and 24.
Kidney dysfunction in cirrhosis is mainly related to the development of AKI, caused by acute hemodynamic alterations. It´s incidence increases according to the presence of other associated risk factors that can cause alteration to kidney function, such as diabetes, viral hepatitis, among others. AKI is defined as an increase in serum creatinine of 0.3mg/dl in <48 h or of 50% of base creatinine over the past three months, without defining a threshold of final serum creatinine. Stages 1, 2 and 3 of AKI are defined as 150%, 200% and 300% of base serum creatinine, respectively, which allows assessment of the progression of AKI 20 .
The main limitation of this study is being retrospective, meaning that it was not possible to verify results by comparing intervention with no intervention.
Abdominal wall hernia is a common complication for patients with cirrhosis.
Herniorrhaphy presents a higher morbimortality when compared with patients without cirrhosis. The presence of AKI in the postoperative is a frequent event and should be prevented when possible and treated adequately, in order to reduce mortality in this population.
Preoperative preparation with adequate volume expansion to replace ascites evacuation, careful hydration, an elective surgical approach, rigorous controls of hydration balance in the intra and postoperative phases are useful and effective measures to prevent AKI in patients with cirrhosis. As a consequence, a lower rate of morbimortality can be reached in this population when submitted to hernial correction