Chloride-liberal fluids are associated with acute kidney injury after liver transplantation

Introduction Acute kidney injury (AKI) occurs frequently after liver transplantation and is associated with significant morbidity and mortality. Recent evidence has linked the predominant usage of ‘chloride-liberal’ intravenous fluids, such as 0.9% saline to the development of renal dysfunction in general critically ill patients. We compared the effects of perioperative fluid types on AKI in liver transplant recipients. Methods An observational analysis of liver transplant recipients over a 33-month period, between January 2010 and September 2013, was performed. Intensive care unit database and patient records were analyzed for determinants of early postoperative AKI. Univariate and multivariate regression analysis was carried out using a two-tailed P value less than 0.05 to establish significance. The institutional Research Ethics Committee approved the study methodology (RAC no. 2131 073). Results One hundred and fifty-eight liver transplants were performed, AKI developed in 57 (36.1%) patients: 39 (68.4%) fully recovered, 13 (22.8%) developed chronic renal failure and 10 (17.5%) required long-term hemodialysis. On univariate regression analysis, AKI was significantly associated with greater than 3,200 ml of chloride-liberal fluids infused within the first postoperative day (HR 5.9, 95% CI 2.64, 13.2, P <0.001), greater than 1,500 ml colloids received in the operating room (hazard ratio (HR) 1.97, 95% CI 1.01, 3.8, P = 0.046), vasopressor requirement for 48 hours posttransplant (HR 3.34, 95% CI 1.55, 7.21, P = 0.002), hyperchloremia at day 2 (HR 1.09, 95% CI 1.01, 1.18, P = 0.015) and preoperative model for end-stage liver disease (MELD) score (HR 1.08, 95% CI 1.03, 1.13, P <0.001). After stepwise multivariate regression, infusion of greater than 3,200 ml of chloride-liberal fluids (HR 6.25, 95% CI 2.69, 14.5, P <0.000) and preoperative MELD score (HR 1.08, 95% CI 1.02, 1.15, P = 0.004) remained significant predictors for AKI. Conclusions In a sample of liver transplant recipients, infusion of higher volumes of chloride-liberal fluids and preoperative status was associated with an increased risk for postoperative AKI.


Introduction
Acute kidney injury (AKI) occurs both frequently after liver transplantation, reportedly in 29 to 60% recipients [1][2][3] and, irrespective of severity, confers an increased risk of death [4]. This increase in risk of mortality extends from the early postoperative period (28 days) and up to one year after transplantation [1]. The National Institute of Diabetes and Digestive and Kidney Disease (NIDDK) long-term follow-up study ascribed a 2.66 hazard ratio (HR) directly attributable to renal dysfunction developing after liver transplantation [5].
Previously described risk factors for AKI in liver recipients are greater severity of illness pretransplant (higher model for end-stage liver disease (MELD) scores, intensive care unit (ICU) admission, and coagulopathy), vasopressors, and greater transfusions in the immediate perioperative period [5][6][7][8][9][10][11][12][13]. However, very little is known about the effects of intravenous fluid selection (chloride-liberal versus chloride-restrictive) or the effects of fluid balance on the risk of renal injury. An overall fluid overload state leads to renal congestion, compromised renal blood flow and reductions in glomerular filtration rate (GFR) [14]. In critically ill patients, a positive fluid balance has been associated with increased mortality [15,16] and poorer outcomes once AKI develops [17][18][19]. Recent evidence has highlighted the possible nephrotoxic effects of 'chloride-liberal' fluids (0.9% saline). Animal and human controlled studies have shown that infusions of chlorideliberal or solutions with supraphysiological chloride concentrations cause vasoconstriction of renal afferent arterioles, cortical hypoperfusion and decreased GFR [20][21][22]. In a recent pre-and postintervention study, restriction to chloride-restrictive fluids was associated with lower AKI and need for renal replacement therapy (RRT) as compared to chloride-liberal fluids [23]. Therefore, perioperative intravenous fluid selection and volume may prove to be a modifiable risk factor for the prevention of AKI. No similar data exists for liver transplant recipients. These patients clearly are at an increased risk of renal injury with mortal consequences. We hypothesized that AKI occurring in the early postoperative period after liver transplant may be associated with the use of chloride-liberal fluids and overall fluid overload causing renal congestion.

Material and methods
This study is reported following the STROBE statement checklist for observational studies [24]. All studies at our institution require ethical approval; the Office of Research Affairs (ORA) and ORA Research Ethics Committee approved the study (RAC no. 2131 073). Patient consent was waived by the Research Ethics Committee.

Study design and setting
This was an observational study of liver transplant recipients carried out at a tertiary care, university hospital over a 33-month period between January 2010 and September 2013.

Operational definitions
AKI was defined according to the risk, injury, failure, loss, end-stage renal failure (RIFLE) classification [25] of renal dysfunction, that is using both increases in creatinine from preoperative values and urine output measured as urine volume in milliliters/patient's baseline weight in kilograms/hour. Serum creatinine values were measured preoperatively and daily for up to the third postoperative day. Serum creatinine was measured using the COBAS Integra Creatinine plus ver. 2 assay (Roche Diagnostics Corp, Basel, Switzerland). This is an enzymatic method based on the determination of hydrogen peroxide after conversion of creatinine with the aid of creatininase, creatinase, and sarcosine oxidase. Patients were screened for the development of postoperative AKI on a daily basis until the third day.
Prolonged ICU stay was defined a priori as a cutoff value of the mean (or median for skewed data) ICU days after transplantation. Delayed weaning from mechanical ventilation was defined as <3 > days of invasive ventilation.

Participants
Consecutive adult liver transplant recipients within the specified study period were included. Patients undergoing multiorgan transplantation were excluded.

Variables
The primary outcome variable was the development of postoperative AKI. Other outcome variables studied were delayed weaning from mechanical ventilation, prolonged ICU stay (as defined above), ICU mortality, and 28-day mortality. Other variables collected were recipient demographic data, etiology of cirrhosis, comorbidities, posttransplant acute physiologic and chronic health evaluation II (APACHE II) scores, routine hematological, biochemical and organ dysfunction/physiological (AKI, vasopressors, RRT, mechanical ventilation) data, fluid balance, fluids and blood products received at admission to ICU and daily up to day 3 posttransplantation.

Data sources/measurement
In our ICU, patient data is routinely entered into an ICU database. Data entry is by a critical care nurse dedicated to the database. Patient data required for the study and missing from the database was extracted by the research team (AH, HS, BB, AN) from the patient's electronic medical records and laboratory computerized results.

Statistical analysis
Continuous data was tested for normality: measures of central tendency were compared as means ± standard deviations (SD) using the Student's t test for normally distributed variables and as medians (interquartile range, IQR) using the Mann-Whitney U test for skewed data. Categorical variables were compared using the chi-square test or the Fisher exact test for n <5. Fluid volumes were dealt with as continuous variables while fluid types were classified into either 'chloride-liberal' or 'chloride-restrictive' and correspondingly dealt with as continuous variables.
Logistic regression analysis was performed to determine the predictive ability of variables for AKI, prolonged ICU stay and prolonged mechanical ventilation. Univariate and multivariate techniques were used, and for multivariate regression, a backward mode with a threshold 0.15 was used for elimination. Multivariate associations were reported as odds ratios (OR) with 95% confidence intervals (CIs). A two-sided P value of <0.05 was considered as statistically significant. All analyses were carried out using IBM SPSS version 22.0 (IBM Corp, Armonk, NY, USA).
Outcomes AKI developed in 58 (36.4%) patients; classified as risk in 30 (52%), injury in 12 (21%) and failure in 16 (27%) patients. Ninety percent (53 patients) developed AKI within the first two postoperative days. All liver transplant patients, both the group that developed AKI and those that did not were on the calcineurin inhibitor, tacrolimus. Drug levels were routinely monitored and dosing adjusted accordingly. Thirty-nine (68.4%) patients recovered fully, 13 (22.8%) developed chronic renal failure and 10 (17.5%) required long-term hemodialysis. AKI was significantly associated with ICU mortality, P = 0.001 and 28-day mortality, P <0.001. Mean serum chloride levels on the second postoperative day were significantly greater in patients who developed AKI compared to those who did not; 114 ± 7.2 versus 112 ± 4.4, P = 0.01. The mean chloride levels on days one and three were not significantly different. There were no significant associations between mean serum chloride levels and the severity of renal failure (see Table 2).
Delayed weaning from mechanical ventilation was associated with higher volumes of chloride-liberal fluids, P = 0.02, higher colloid volumes, P = 0.015, blood products transfused, P = 0.017 and a cumulative positive fluid balance, P = 0.026. Higher pretransplant MELD scores, P = 0.001, male gender, P = 0.015, transplant for hepatocellular carcinoma, P = 0.014, crystalloid volume received in the first 72 hours, P = 0.034, need for vasopressors at 48 hours, P <0.001 and 72 hours, P = 0.031, AKI, P <0.001 and pleural effusion, P = 0.001 were significantly associated with a prolonged ICU admission. Drainage of effusion was significantly associated with a reduced ICU stay, P = 0.007 (see Tables 3 and 4).

Discussion
In this observational study, we found that liver transplant recipients were more likely to develop AKI if they received larger volumes of chloride-liberal (hyperchloremic) fluids. This association was significant, after adjusting for baseline variables, for both 5% albumin in 0.9% saline and only 0.9% saline infusions. Patients who developed AKI had significantly higher serum chloride  levels compared to transplant recipients that did not develop AKI. 'Normal' saline or 0.9% saline contains supraphysiological levels of chloride (154 mmol/L as compared to Hartmann's solution, Ringer's lactate or Plasma-Lyte 148, all of which contain chloride concentrations that are lower (94 to 111 mmol/L). Five percent albumin is available either as salt-poor or in sodium chloride (chloride concentration 128 mmol/L). Intravenous infusions of chloride-liberal fluids have been associated with hyperchloremia and metabolic acidosis when administered in large volumes [26].
Our results show a detrimental effect on renal function with use of chloride-liberal fluids in the immediate postoperative period (up to 48 hours). Support for our findings comes from animal studies that have demonstrated reductions in GFRs, renal arteriolar vasoconstriction [27], and human volunteer studies that have shown reduced renal cortical tissue perfusion, renal blood flow velocity after infusions of hyperchloremic solutions. [28]. In controlled trials, chloride-liberal fluids compared to chloride-poor fluids have been linked to longer time to micturition [21], lower urine output [29] and in a recent  observational study of over 31,000 postoperative patients, 0.9% saline compared to 'balanced' crystalloids increased the risk of acute renal failure requiring dialysis [30]. Yunos et al., in a pre-and postintervention study on 1,530 critically ill patients found that a chloride-restrictive fluid strategy resulted in a significant reduction in AKI, need for RRT and increase in creatinine as compared to a control group given chloride-liberal fluids [23]. Possible explanations for this renal 'toxicity' of chlorideliberal fluids come from animal studies that have demonstrated renal vasoconstriction [22] and thromboxane release after chloride infusions [20]. Chloride infusions increase delivery to the macula densa that stimulates glomerulotubular feedback leading to afferent arteriole constriction, mesangial contraction and resultant decrease in GFR [31].
Patients in our study received both 5% albumin in saline and 20% albumin in saline. Though it is possible that the observed renal dysfunction resulted from hyperoncotic albumin, the data on 20% albumin is so far inconclusive. A recent cohort study on 1,000 patients found a higher risk of renal injury and failure with the use of hyperoncotic albumin (OR 5.99) [32], however, a contradictory result was reported from two meta-analyses that concluded no harmful effects of hyperoncotic resuscitation [33,34]. The SAFE study that compared albumin and saline found no difference in adverse outcomes [35].
Fluid overload leads to organ dysfunction due to interstitial edema and visceromegaly. The limited accommodative capacity of the encapsulated kidney causes increased interstitial hydrostatic pressures with reduced renal perfusion and filtration [36]. Additionally, a fluid overload state contributes to third spacing, ascitic fluid accumulation and abdominal compartment syndrome [37]. Cumulative fluid overload has been linked to poor outcomes in all groups (pediatric, septic, postoperative) of critically ill patients with prolonged days on mechanical ventilation, ICU stay and mortality [16,17,[38][39][40][41][42][43]. Recovery of renal   function in patients on RRT is also determined by overall fluid balance [17,38,42]. In our study, a cumulative positive fluid balance increased the duration of stay in ICU. A limitation of our study is that the observational design does not establish a causal relationship of hyperchloremic fluid excess with the development of AKI in liver transplant recipients. These associations may be subject to bias from selection, confounding or random error. We attempted to control for confounders by using regression analysis. Another limitation is the external validity or generalizability of our results to other liver transplant recipients since we collected data only from a single institution.

Conclusions
In summary, large infusions of chloride-liberal fluids may predict a higher risk of AKI in liver transplant recipients. Our findings support the hypothesis that 'routine' intravenous fluids may not be routine and in themselves be associated with organ dysfunction. Our results can be used to build hypotheses for further controlled trials.

Key messages
Chloride-liberal fluids may cause renal dysfunction Large volumes (>3,200 ml) of chloride-liberal fluids infused in the first 24 hours after liver transplantation were associated with a higher risk of AKI.

Competing interests
The authors declare that they have no competing interests.
Authors' contributions AN participated in the design of the study, collected the data, analyzed and interpreted the data, drafted the manuscript, and revised the manuscript critically for important intellectual content. NS conceived the study, participated in the design of the study, made the figures and table, analyzed and interpreted the data, drafted the manuscript, and revised the manuscript critically for important intellectual content. AH, MJ, HAS and BB participated in the design of the study, collected the data, and participated in the coordination of the study. YS and DB participated in the design and coordination of the study and participated in the critical review of the final manuscript. All authors have given final approval of the version to be published.