The role of fractional excretion of sodium and urinary osmolality in Acute Kidney Injury in liver transplant


 Objective

We analysed urinary osmolality and the fractional excretion of sodium (FeNa) in the perioperative period of liver transplant (LT) and their association with on renal impairment, dialysis and mortality.
Methods

We aimed to determine the pattern of elevation of urinary (U) osmolality and FeNa levels in the perioperative period of liver transplant and how these are associated with the development of acute kidney injury (AKI) according to the Kidney Disease Improving Global Outcomes- (KDIGO) criteria, AKI severity, differential diagnosis in acute tubular necrosis (ATN), need for renal replacement therapy (RRT) and mortality. We assessed the biomarkers in the perioperative period: pre-operative, after portal reperfusion (APR), and at 6, 18, 24 and 48 hours after LT.
Results

Of the 100 enrolled patients, 87 developed AKI in the first week after LT, with 59 considered KDIGO stages 2 and 3 as defined by severe AKI and 75 defined as ATN; 34 were dialyzed, and 21 died within 60 days after LT. The FeNa was also useful for differential diagnosis in ATN, but the values remained below 1%, with an increased median in poor outcomes: severe AKI, ATN, need-RRT and non-survival. For predicting need-RRT, FeNa achieved an AUC of 0,78 (CI 0,66–0,90). The APR U osmolality measurement showed differences in all outcomes (with p < 0,05), and high osmolality was revealed to be a renal protective factor and found to predict need for RRT and mortality with AUCs of 0,11 (CI 0,02–0,20) and 0,21 (CI 0,07–0,34), respectively.
Conclusion

The increase in FeNa reveals a loss of Na secretion capacity and even in liver disease patients it has been shown a tool that aided the differential diagnosis if the cutoff value was adjusted. Osmolarity demonstrated the maintenance of urine concentration capacity by nephrons. More large studies should confirm these results.


Introduction
Technological advances in the eld of medicine and diagnosis have built up over the decades. However, access to state-of-the-art technology is not the reality in many places, and even in advanced places there is a long way between research and clinical practice.
Then, utilizing available resources in clinical practice is a valuable tool. Urine one is available in most hospitals and health centers and it can be an important tool in diagnosing acute kidney injury (AKI) that's common in cirrhotic patients and major surgeries, such as liver transplantation.
Currently, the diagnosis is based on parameters of serum creatinine and urine output; unfortunately cirrhotic patients can present with oliguria or pollakiuria as an effect of diuretics, which does not re ect the glomerular ltration rate (GFR), and serum creatinine is commonly underestimated due to loss of muscle mass in cirrhotic patients. Neither are good parameters for assessing renal impairment 9,10 .
Urine 1 has numerous measures to inform data on the patient's renal function. Because it is available in clinical practice, it becomes an important tool analyzing as parameters of fraction excretion of sodium (FeNa) and urinary osmolarity, could be useful in the diagnosis of AKI, as well as in its differential diagnosis. Even knowing that cirrhotic patients have their Na balance totally altered due to increased renin secretion in portal hypertension states in chronic stages, the hypothesis that FeNa could be useful in AKI diagnostic remains unclear.
The idea FeNa is to assess the state of tubular resorption nephron, and is based on on the premise that intact tubules reabsorb sodium in the prerenal state while the injured tubules do not 5,6 . Hypoperfusion of nephrons suggests increased secretion of active renin and prorenin from the remaining nephrons 4 . The hypoperfusion, can affect the juxtamedular nephrons, more than other nephrons, hence the renal medulla is mainly responsible for maintaining the capacity of concentration of urine, lose your function turning the urine less concentrated, with low osmolarity.
This manuscript aimed to analyse the role of urinary osmolality, and FeNa in the perioperative period of liver transplant and their impact on renal impairment, the need for RRT and mortality.

Patients Characteristics
The patients' demographic and clinical characteristics are shown in Table 1 and the enrollment criteria are available in Fig. 1.The most common comorbidities were hypertension and diabetes. The main reasons for liver dysfunction were hepatitis C in 46 cases and alcoholic cirrhosis in 13, followed by cryptogenic cirrhosis in 11. We considered severe AKI in 59 patients to meet the criteria for stage 2/3 AKI.

FeNa
The median level of FeNa is demonstrated in Table 2, where the values generally remained below 1% and no differences were found between non-AKI and AKI patients. Regarding the severity of AKI, the level of FeNa was higher in severe AKI, and this difference achieved statistical signi cance 18

Urinary osmolality
The elevation of osmolality was not different among pre/ post operative period and non-AKI/ AKI. After PR, the median osmolality decreased in AKI patients and remained low during the follow-up period ( Table 3). Severe AKI reduced the osmolality in both groups APR, re ecting a dilution due to uid overload in the intra-operative phase. In the follow-up period 6 hours after LT, the group with no AKI/mild AKI

Discussion
Over the years, the differential diagnosis of AKI has been a challenge in clinical practice; FeNa has already been considered a differential diagnosis parameter of AKI, landmark studies have revealed a reference value of 1%, not found in recent studies. The osmolarity forgotten over the years, can reveal concentration parameters, being a surrogate of the function of the juxtamedullary nephron. Our study revealed that even in surgical patients, in the case of liver transplantation, adjusting the reference value, due for the patients' hypervolemia status, both FeNa and U osmolarity showed favorable results in the evaluation of the diagnosis of AKI and other outcomes.
The FeNa has con rmed the premise of AKI differential diagnosis pre-renal/haemodynamic AKI from ATN/persistent AKI. The increase in Na in urine re ects a reabsorption impairment of the tubules caused by more severe injuries, such as ATN, while in the pre-renal/haemodynamic injury, the remaining tubules can still reabsorb Na. The indication of pre-renal AKI occurs when FeNa is less than 1%, while values greater than 1% are consistent with ATN due to inappropriate sodium excretion in the setting of tubular damage 17 . Notwithstanding, fewer studies achieve this reference value, and in our study, the values remained below 1%, with increased median values found in poor outcomes: severe AKI, ATN, need for RRT and non-survival. The performance of FeNa in predicting the need for RRT achieved a high AUC of 0,78 (CI 0,66-0,90) 18 hours after LT.
The diagnosis of urinary concentration occurs when the maximum urine osmolality is < 600 mOsm/kg H²O. This manuscript shows osmolality median levels below 500 mOsm/kg H²O for all cohorts. The preoperative period before LT did not differ with respect to outcomes, and the median values remained at 450-516 mOsm/kg H²O. After portal reperfusion, there were differences in all outcomes. Notably, the reduction in capacity for urine concentration by the tubules and the osmolality appear to be a renal function surrogates in AKI, severe AKI and ATN. Moreover, the performance to predict a need for RRT was high, and the low AUC of 0,11 (CI 0,02-0,20) achieved 6 hours after LT revealed a renal protective effect of high osmolality. Additionally, 6 hours after LT, osmolality predicted 60-day mortality with a median of 354 in the survivor group vs 534 in the non-survivor group with a p 0,004 and an AUC 0,21 (CI 0,07-0,34). The multicentre study reported by Tabzadeh et al. with 2084 patients concluded that osmolality, in addition to GFR and albuminuria, is a useful tool for assessing non-glomerular damage in patients with CKD 18 .
Although our study had a limited number of patients due to several exclusion criteria and was not a multicentre study, it is still the largest cohort evaluating urinary osmolality and FeNa in perioperative liver transplantation periods. Diuretic use occurred in 42 patients in the rst week after LT, and these patients were not withdrawn from the analysis, even though it is known that diuretic use can affect the sodium excretion and osmolality. However, in clinical practice, patients use diuretics, and the favourable results in our study showed that even with their use, the outcomes could be determined. Frequent osmotic and electrolyte imbalances in liver transplant patients may be even more affected intraoperatively due to the abundant infusion of sodium-rich uids to replace blood loss. In this analysis, we used serum creatinine as a sole comparator in the performance of diagnosing AKI, not applying urine output criteria. Despite the KDIGO criteria 9 including urine output in AKI diagnosis, the International Club of Ascite did not recommend its use in cirrhotic patients due to oliguria caused by increased sodium retention, even with a relatively normal GFR. Conversely, increased urine output due to diuretic treatment does not re ect a reliable GFR 10 .

Conclusion
Despite an imbalance in Na and the regulatory system (SRAA, ADH), the differential diagnosis of ATN and the need for dialysis could be predicted with urinary osmolality and FeNa in the perioperative LT period. More large studies should con rm these results.

Materials And Methods
The

Patients
During a 24-month period from June 2013 through June 2015, a total of 189 liver transplants were performed; 139 recipients were eligible for enrolment, and 100 were enrolled in the study (Figure 1). Figure   1 shows reasons for non-enrolment. During this study period, all recipients older than 18 years old were screened. Patients were enrolled in the study after voluntary informed consent was obtained as per the guidelines of the Institution Ethics Committee. Exclusion criteria included the following: need for dialysis pre-operatively, second liver transplant, combined transplant, chronic kidney disease stage 5, or previous kidney transplant.
Patients were divided into six categories according to the aetiology of end-stage liver disease (ESLD): hepatitis B or C cirrhosis, alcoholic cirrhosis, cryptogenic cirrhosis, acute hepatitis and miscellaneous causes. The miscellaneous group included patients with non-alcoholic steatohepatitis (NASH), autoimmune hepatitis, haemochromatosis, Budd-Chiari syndrome, biliary atresia, Wilson disease, primary sclerosing cholangitis, polycystic disease, biliary cirrhosis and hepatocellular carcinoma.
We recorded baseline kidney function and comorbidity history from electronic medical records (EMR).
The following perioperative variables were included: main patient characteristics, clinical follow-up in the rst week after LT, data on the need for RRT, and outcomes.
The functional MELD was calculated based on serum bilirubin, international normalized ratio (INR) and serum creatinine using the formula score MELD = 0,957 x Log (creatinine mg/dl) + 0,378 x Log (bilirubin mg/dl) + 1,120 Blood and urine samples were collected simultaneously in the perioperative period of LT before induction of anaesthesia, after portal reperfusion, and at 6, 18, 24 and 48 hours after surgery. We recorded vital signs, process of care and lab results daily for 7 days after LT. Outcomes including maximum AKI severity stage, need for RRT in the rst week after LT and mortality were assessed in the intensive care unit (ICU), at hospital discharge and 60 days after enrolment.

Clinical Outcomes
The primary outcome was AKI development during the rst week of LT. Baseline renal function was de ned as the lowest value in the last three months and was used to assess renal recovery. AKI diagnosis was based on the Kidney Disease Improving Global Outcomes (KDIGO) 9 . We considered reference serum creatinine as the lowest value in the week before liver transplantation, and this value was used to determine AKI diagnosis. AKI stage was de ned according to KDIGO: stage 1 (1.5 -1.9 times sCr reference, or increase =>0,3 mg/dl until 48 hours), stage 2 (2.0-2.9 times sCr reference), stage 3 (3.0 times sCr reference or increase >4,0 mg/dl). Patients with no AKI or stage 1 AKI were categorized as the no/mild AKI group, whereas patients meeting the criteria for stage 2 or 3 AKI seven days after LT were identi ed as the severe AKI group. The acute tubular necrosis group was de ned as patients with KDIGO 2 or 3 who did not recover renal function in the rst week or required RRT. The calculation of fractional excretion of sodium is as follows: (FeNa or U/PNa ÷ U/P creatinine×100).

Statistical Analysis
Continuous variables are presented as the mean ± SD or median (25th -75th percentiles) and were compared using one-way ANOVA or the Kruskal-Wallis test according to the Gaussian distribution. Categorical variables are presented as absolute numbers and percentages and were compared by the Chisquare test. P values were two tailed, and p<0.05 was considered signi cant. Conventional receiver operating characteristic (ROC) curves were generated, and area under the curve (AUC) was used to assess the ability of continuous variables to distinguish the categorical state: AKI development, need for dialysis and non-survival.
The perioperative covariates were tested in univariate analysis for their impact on renal impairment (AKI, severe AKI and ATN), need for dialysis and mortality. Factors associated with the outcomes at p < 0.05 were used to construct a multivariate model in which the impact of each comorbidity or covariate was adjusted for that of all others. The model included variables urinary osmolality and, FeNa. SPSS (Statistical Package for Social Sciences) version 20 (Chicago, Illinois) was used for the statistical analysis.

Declarations
Competing interest No