Adipocyte Fatty Acid–Binding Protein as a Predictor of Outcome in Alcohol-induced Acute-On-Chronic Liver Failure

Abstract

Background

Alcohol is the leading cause of acute-on-chronic liver failure (ACLF). Several severity scores predict the outcome of ACLF. However, there is a lack of simple biomarkers in predicting the outcome of these sick patients. Fatty acid–binding proteins (FABPs) are small cytosolic proteins that play a major role in lipid metabolism, energy homeostasis, and inflammation, but, have not been investigated in alcohol-induced ACLF (A-ACLF).

Objectives

The primary objective was to assess the correlation between serum adipocyte-FABP (A-FABP) and liver-FABP (L-FABP) levels on mortality at day 90. Secondary objectives were to compare the levels between controls and A-ACLF, correlate L-FABP, and A-FABP levels on the development of organ failure/sepsis at day 90.

Methods

In this prospective observational pilot study, we included patients with A-ACLF and age-matched healthy controls. FABP's were analyzed by enzyme-linked immunosorbent assay method. The patients were followed up for 90 days.

Results

Twenty-five patients with A-ACLF (mean age: 40years; mean model for end-stage liver disease NA: 29.8; median Modified Maddrey's discriminant function [mDF]: 95) and 12 controls (mean age: 36.83yrs) were included in the study. A-FABP and L-FABP levels were significantly high in patients with A-ACLF than controls. Forty-four percent of patients with A-ACLF developed sepsis, 48% developed organ failure, and 44% expired by day 90. On multivariate Cox regression analysis, A-FABP (hazard ratio [HR]: 1.27 [1.08–1.5]; P = 0.003), Asian Pacific Association for the Study of Liver ACLF research consortium score (HR: 3.3[1.15–9.54]; P = 0.02), L-FABP (HR: 0.69 [0.52–0.91]; P = 0.009), and serum protein levels (HR: 0.03 [0.003–0.36]; P = 0.005) predicted mortality. A-FABP (1.17 [1.07–1.29]; P = 0.001), and serum bilirubin (1.05 [0.99–1.12]; P = 0.06) predicted development of organ failure, and only mDF (HR: 1.04 [1.01–1.07]; P = 0.009) predicted the development of sepsis on multivariate analysis. Fifteen patients received steroid therapy, of which 13.34% were nonresponders.

Conclusions

In a selected group of patients with A-ACLF, A-FABP is highly sensitive at predicting mortality and outcome. If validated in a large, diverse sample, A-FABP can be used as a simple biomarker for prognostication in A-ACLF.

Alcohol is the leading cause of acute-on-chronic liver failure (ACLF).¹ Although ACLF has varying definitions, the end point of each is high short-term mortality.² Several severity scores aid in predicting the mortality of patients with ACLF. A systematic review upheld age, hepatic encephalopathy, the model for end-stage liver disease (MELD) score, total bilirubin, and international normalized ratio (INR) as promising candidates.³ However, there are no effective biomarkers to predict the development of complications of alcohol-induced ACLF such as acute kidney injury (AKI) sepsis, or mortality beforehand.^4,5 Fatty acid–binding proteins (FABPs) are small (14–15 kDa) cytosolic proteins that bind to unesterified long-chain fatty acids (LCFAs) with nanomolar affinity.⁶ They are intracellular lipid chaperones that coordinate lipid responses in the cell, thereby playing a major role in metabolic and inflammatory pathways.⁷ Of the nine human isoforms, most notable are liver type FABP (L-FABP) (FABP1), heart FABP (FABP3), and adipocyte type FABP (A-FABP) (FABP4). FABPs aid in predicting outcomes in patients with acetaminophen-induced acute liver failure (L-FABP), decompensated cirrhosis (A-FABP), and hepatocellular carcinoma (L-FABP).8, 9, 10 However, there are no studies carried out on FABPs in ACLF. We evaluated the role of FABPs in predicting the outcomes of alcohol-induced ACLF (A-ACLF). We hypothesized that FABPs, especially A type and L type, maybe elevated beforehand in ACLF, where hepatic and systemic inflammation plays a crucial role.

Patients and methods

The study was a prospective observational pilot study conducted at the Asian Institute of Gastroenterology hospital from January 2019 to April 2019. Twenty-five consecutive patients with A-ACLF attending the outpatient department or admitted to the hospital during the study period were recruited. Patients with active sepsis, active ongoing gastrointestinal bleed, renal failure (serum creatinine >1.5 mg/dl), diabetes mellitus, coronary artery disease, pregnant patients, patients with shock requiring vasopressor support, and patients who refused to participate were excluded. The institutional review board approved the study vide letter number AIG/AHF/IRB: 5/A/2019 dated 31/01/2019. Written informed consent was obtained from each patient, and the study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the institution's human research committee. After taking informed consent, baseline demographics, including clinical and biochemical data and 4 ml blood, were collected. Bodyweight was corrected for ascites, and pedal edema and corrected body mass index (BMI) was calculated. Five, 10, and 15% were deduced for mild, moderate, and severe ascites, respectively. An additional 5% was deduced from bodyweight if the patient had bilateral pedal edema.¹¹ The patients were followed up for 90 days to observe the outcome and complications. Controls were the staff of the hospital who underwent routine health check-ups and were consuming alcohol in safe limits, i.e., no more than 40 g/week (or 2 drinks for women and 3 drinks for men [each containing 10 g of alcohol] per day).

Blood samples were stored at −80 °C until analyzed. FABPs were analyzed by an enzyme-linked immunosorbent assay (ELISA) method (Elabscience, India PVT. LTD). Briefly, the procedure was: 100 μL of the sample was added to each well and incubated for 90 min at 37 °C. A 100 μL of Biotinylated detection antibody was then added and incubated for 60 min at 37 °C. The fluid was aspirated and washed three times, followed by the addition of 100 μL of horseradish peroxidase conjugate. After 30 min of incubation and another five washes, 90 μL of substrate reagent was added. Finally, 50 μL of stop solution was added and read at 450 nm immediately.

Objectives

The primary objective was to study the correlation between serum L-FABP and A-FABP levels on mortality at day 90 in patients with A-ACLF. The secondary objective was to compare the levels of serum L-FABP/A-FABP in patients with A-ACLF and healthy controls, as well as to assess the correlation of A-/L-FABP levels on the development of complications, i.e., organ failure/sepsis at day 90 in patients with A-ACLF and finally to determine the steroid response at day 7.

Definitions

Severe alcoholic hepatitis was defined as alcoholic hepatitis with Modified Maddrey's Discriminant Function (mDF) score ≥32 &/or MELD ≥21.¹² ACLF was defined as per Asian Pacific Association for the Study of Liver (APASL) criteria as an acute hepatic insult manifesting as jaundice (serum bilirubin ≥5 mg/dl) and coagulopathy (INR ≥1.5) complicated within four weeks by clinical ascites and/or encephalopathy in a patient with previously diagnosed or undiagnosed chronic liver disease/cirrhosis and is associated with high 28-day mortality.¹³ All the patients received the standard of care for alcoholic hepatitis (N-acetyl cysteine, intravenous ceftriaxone, with or without albumin infusion). In the absence of active sepsis and organ failure, steroids were initiated for severe alcoholic hepatitis. Day-7 Lille's score was calculated to determine the response to steroids. Lille's score<0.45 was considered as steroid response.

Since the included patients should have (based on APASL criteria for ACLF) bilirubin > 5 mg% and INR >1.5 (liver injury), extrahepatic organ failures were defined as per EASL criteria (European Association for the Study of the Liver).¹⁴ Renal failure: serum creatinine ≥2 mg/dl. Cerebral failure: grade 3 or 4 hepatic encephalopathy according to West-Haven criteria or need for mechanical ventilation due to encephalopathy. Circulatory failure: patients requiring vasopressors for maintaining mean arterial pressure >65 mm Hg. Respiratory failure: PaO2/FiO2 ≤200 or SpO2/FiO2 ≤214 or patient requires mechanical ventilation, not due to hepatic encephalopathy. Sepsis was defined as per SEPSIS-3 criteria.¹⁵

Statistical methods

The data were entered in Microsoft excel sheet and analyzed using Statistical Package for Social Sciences, version 25.0 (IBM Corp Ltd, Armonk, NY). Descriptive statistics are expressed as mean (standard deviation) or median (minimum, maximum) for parametric or nonparametric continuous data, respectively, and number (%) for categorical data. The comparison of means between the two groups was made using t-test. The categorical data were compared with Pearson Chi-square test (or Fisher's exact test when required). The various predictors were derived using stepwise multivariate Cox regression analysis involving parameters that have P < 0.1 on univariate Cox regression analysis. Further to find the different cutoff points for sepsis/mortality, receiver operating characteristic curve analysis was also carried out along with Youden's index calculation. All statistical tests with P < 0.05 were considered significant.

Results

Fifty-four patients with A-ACLF were screened, and of them, 29 were excluded. Thirty percent of patients had sepsis, 20.34% had AKI at baseline, and one patient had diabetes mellitus, and these subsets were excluded. The mean age was 40 ± 9.94years, and all were men with a mean corrected BMI of 22.81 ± 2.5 kg/m² (Table 1). The mean MELD NA was 29.8 ± 5.92. Fifteen patients received steroids, of whom two did not respond to steroids. Twelve healthy controls (doctors) consented for the study.

Table 1.

Baseline Characteristics of Included Patients and Controls.

Parameter	ACLF (n = 25)	Controls (n = 12)
Age	40 ± 9.94	36.83 ± 6.83
Males, n (%)	25 (100%)	11 (91.66%)
Alcohol intake (g/week)	504 ± 32.72	33.34 ± 11.54
BMI (kg/m2)	22.81 ± 2.5	24.05 ± 2.8
Hemoglobin (g/dL)	10.09 ± 1.55	13.3 ± 1.4
TLC (x 103/l)	11.57 ± 5.18	6.82 ± 2.28
Total bilirubin (mg/dl)	23.46 ± 14.22	1.13 ± 0.48
AST (IU/L)	156 ± 77.54	37.36 ± 12.25
ALT (IU/L)	58.36 ± 31.51	34.91 ± 13.43
Total protein (g/dL)	6.57 ± 0.92	7.53 ± 0.35
Albumin (g/dL)	2.91 ± 0.57	4.31 ± 0.38
A-FABP∗ (ng/ml)	16.8 (4.86,38.6)	1.86 (0.94,8.55)
L-FABP∗ (ng/ml)	14.93 (1.2,26.37)	1.5 (0.44,8.33)
MELD NA	29.8 ± 5.92	NA
mDF∗	95 (46.5346.2)	NA
AARC score	8.48 ± 1.68	NA
CLIF-SOFA score	9.32 ± 1.34	NA

All the values are expressed as mean ± standard deviation except that marked ∗ which are expressed as median (Minimum, Maximum). BMI, body mass index; TLC, Total leukocyte count; AST, aspartate transaminase; ALT, alanine transaminase; A-FABP, adipocyte type fatty acid–binding protein; L-FABP, liver type fatty acid–binding protein; INR, International normalized ratio; MELD NA, model for end-stage liver disease-sodium; mDF, Modified Maddrey's discriminant function; AARC, APASL ACLF research consortium; CLIF-SOFA, chronic liver failure-sequential organ failure assessment score; NA, not applicable.

Correlation of Serum L-FABP and A-FABP Levels on Mortality at Day 90 in Patients with A-ACLF

A total of eleven patients (44%) succumbed by day 90. A-FABP was high, and L-FABP was low in patients who succumbed (Figure 1). On univariate analysis, total leukocyte count, A-FABP, L-FABP, serum protein, albumin, bilirubin, INR, and severity scores predicted mortality (Table 2). On multivariate analysis, only A-FABP and APASL ACLF research consortium (AARC) score increased the risk, whereas L-FABP and serum protein levels decreased the risk of death. A-FABP>19.7 ng/ml, AARC score >8.5, mDF>92.8, CLIF-SOFA score >9.5, and MELD NA >29 had high area under the receiver operating characteristic curve (AUROC) for predicting mortality at 90-days (Figure 2).

Figure 1.

A-FABP and L-FABP levels (ng/ml) in alive and dead cases. A-FABP was high (p < 0.001), and L-FABP was low in patients who expired (p = 0.09). A-FABP, adipocye type fatty acid–binding protein; L-FABP, liver type fatty acid–binding protein.

Table 2.

Baseline Characteristics of Alive and Dead Patients and Predictors of 90-Day Mortality on Univariate and Multivariate Stepwise Cox Regression Analysis.

Parameter	Univariate HR (95% CI)	P	Multivariate HR (95%CI)	p
A-FABP (ng/ml)	1.19 (1.08–1.31)	<0.001	1.27 (1.08–1.5)	0.003
L-FABP (ng/ml)	0.9 (0.81–1.006)	0.06	0.69 (0.52–0.91)	0.009
Age (years)	1.01 (0.95–1.06)	0.72
BMI (kg/m2)	1.03 (0.82–1.27)	0.79
Hemoglobin (g/dL)	0.78 (0.5–1.2)	0.26
TLC (x 109/L)	1.12 (1.01–1.25)	0.02
Platelets (x 109/L)	0.99 (0.98–1.004)	0.21
Total bilirubin (mg/dl)	1.04 (1.01–1.08)	0.01
AST (IU/L)	1.01 (0.99–1.02)	0.16
ALT (IU/L)	0.97 (0.94–1.01)	0.22
Total protein (g/dL)	0.33 (0.15–0.7)	0.004	0.03 (0.003–0.36)	0.005
Albumin (g/dL)	0.06 (0.01–0.3)	<0.001
INR	2.61 (1.28–5.35)	0.008
Creatinine (mg/dl)	12.16 (0.98–150)	0.06
Sodium (mmol/L)	0.78 (0.69–0.89)	<0.001
MELD NA score	1.17 (1.04–1.32)	0.007
mDF score	1.02 (1.01–1.03)	0.001
AARC score	2.02 (1.28–3.19)	0.002	3.3 (1.15–9.54)	0.02
CLIF-SOFA score	2.68 (1.49–4.82)	0.001

HR, hazard ratio; A-FABP, adipocyte type fatty acid–binding protein; L-FABP, liver type fatty acid–binding protein; BMI, body mass index; TLC, total leukocyte count; AST, aspartate transaminase; ALT, alanine transaminase; INR, International normalized ratio; MELD NA, model for end-stage liver disease-sodium; mDF, Modified Maddrey's discriminant function; AARC, APASL ACLF research consortium; CLIF-SOFA, chronic liver failure-sequential organ failure assessment score. Variables included for multivariate analysis: albumin, MELD NA, CLIF-SOFA, mDF, TLC, A-FBAP, L-FABP, AARC score, total protein.

Figure 2.

Receiver operating characteristic curve (ROC curve) of different models for mortality prediction along with Youden's index. Footnotes for figure 2: A-FABP, adipocyte fatty acid–binding protein; L-FABP, Liver type fatty acid–binding protein; mDF, Modified Maddrey's discriminant function; AARC, APASL ACLF research consortium; CLIF-SOFA,chronic liver failure-Sequential organ failure assessment score; MELD NA, Model for end-stage liver disease-sodium; ACLF, acute-on-chronic liver failure.

Serum L-FABP/A-FABP in Patients with A-ACLF and Healthy Controls

Of the twelve controls, eleven were men. The mean age and BMI were similar among cases and controls. Alcohol consumption was significantly high in patients with ACLF (504 ± 32.72 g/week) than controls (33.34 ± 11.54 g/week). Both A-FABP and L-FABP levels were elevated in patients with A-ACLF than healthy controls (Figure 3, Table 1).

Figure 3.

A-FABP and L-FABP levels (ng/ml) in cases and controls. A-FABP and L-FABP both were high in cases than controls (p < 0.001). A-FABP, adipocyte type fatty acid–binding protein; L-FABP, liver type fatty acid–binding protein;

Correlation of A-FABP and L-FABP levels on the development of complications, i.e., sepsis and organ failure at day 90 in patients with A-ACLF

Twelve patients developed organ failure, and A-FABP was significantly high (26.48 ± 5.21 ng/ml vs. 11.01 ± 3.69 ng/ml; P < 0.001) in patients who developed organ failure. The most common organ failure was kidney failure (58.33%), followed by both circulatory and respiratory failure (25%) and cerebral failure (16.67%). L-FABP levels were similar in both the groups (13.48 ± 7.23 ng/ml vs. 15.44 ± 3.91 ng/ml; P = 0.4). On multivariate Cox regression analysis, A-FABP (1.17 [1.07–1.29; P = 0.001], and serum total bilirubin (1.05 [0.99–1.12]; P = 0.06) at baseline predicted development of organ failure (Supplementary Table 1).

Similarly, eleven patients who developed sepsis had high A-FABP (23.2 ± 8.43 ng/ml) compared with 14 patients with no sepsis (14.7 ± 7.84; P = 0.01), whereas the L-FABP was similar in both the groups (14.09 ± 6.54 vs. 14.81 ± 5.16; P = 0.76). The source of sepsis was pneumonia in 72.73%, urinary tract infection in 18.18%, and the source was unknown in 9.09% of patients. On univariate Cox regression analysis, A-FABP: (1.19[1.02–1.4]; P = 0.02), hemoglobin: (0.61 [0.36–1.04]; P = 0.07), INR: (3.18 [1.14–8.84]; P = 0.02), CLIF-SOFA score: (2.99 [1.13–7.89]; P = 0.02), and mDF: (1.04 [1.01–1.07]; P = 0.009) predicted sepsis, whereas only mDF (1.04 [1.01–1.07]; P = 0.009) predicted sepsis on multivariate analysis (Supplementary Table 2). mDF >83.8 predicted sepsis with a sensitivity of 91.7%, specificity of 76.9%, and AUROC of 0.86 (0.7–1), P = 0.002.

Response to Steroids

Fifteen of the 25 patients received steroids, and of them, two did not respond to steroids. Owing to the development of organ failure and sepsis, ten patients could not be started on steroids. Patients who were started on steroids had lower A-FABP levels, lower total leukocyte counts, lower NL ratio, MELD NA, mDF, and AARC/CLIF-C score (Table 3). A-FABP was significantly high in nonresponders (28.79 ± 3.83 ng/ml) than responders (11.01 ± 3.69 ng/ml; P < 0.001). However, L-FABP was similar in both the groups (nonresponders:17.44 ± 0.7 vs. responders: 15.44 ± 3.91ng/ml; P = 0.5).

Table 3.

Baseline Characteristics of Patients Who Received Steroids and Those Who Did Not.

Parameter	Steroids (n=15)	No steroids (n=10)	P
A-FABP (ng/ml)	13.38 ± 7.2	26.02 ± 5.49	<0.001
L-FABP (ng/ml)	15.7 ± 3.69	12.69 ± 7.73	0.2
Age (years)	39 ± 10.48	41.5 ± 9.41	0.54
BMI (kg/m2)	24.64 ± 3.76	25.27 ± 2.69	0.65
Hemoglobin (g/dL)	10.56 ± 1.2	9.39 ± 1.8	0.06
TLC (x 109/L)	9.53 ± 2.76	14.64 ± 6.51	0.01
Platelets (x 109/L)	180.8 ± 80.12	130.2 ± 39.1	0.07
NL ratio	3.82 ± 1.11	8.88 ± 4.67	<0.001
Total bilirubin (mg/dl)	17.76 ± 13.77	31.95 ± 10.48	0.01
AST (IU/L)	150 ± 77.27	165 ± 81.23	0.64
ALT (IU/L)	58.7 ± 34.95	58.13 ± 30.28	0.96
Total protein (g/dL)	6.87 ± 0.81	6.12 ± 0.91	0.04
Albumin (g/dL)	3.18 ± 0.5	2.51 ± 0.4	0.002
INR	2.2 ± 0.61	2.98 ± 1.22	0.04
Creatinine (mg/dl)	0.87 ± 0.22	1.12 ± 0.13	0.005
Sodium (mmol/L)	135.2 ± 2.85	127.7 ± 3.94	<0.001
Potassium (mmol/L)	3.49 ± 0.59	3.94 ± 0.36	0.04
Duration of hospitalization (days)	11.13 ± 5.37	17.5 ± 3.95	0.004
MELD NA	27.13 ± 5.7	33.8 ± 3.64	0.003
mDF	81.96 ± 35.22	150.5 ± 78.04	0.006
AARC score	7.66 ± 1.49	9.7 ± 1.15	0.001
CLIF-C score	8.66 ± 1.11	10.3 ± 1.05	0.001

All the values are expressed as mean ± standard deviation. A-FABP, adipocyte type fatty acid–binding protein; L-FABP, liver type fatty acid–binding protein; BMI, body mass index; TLC, total leukocyte count; NL ratio, neutrophil-to-lymphocyte ratio; AST, aspartate transaminase; ALT, alanine transaminase; INR, international normalized ratio; MELD NA, model for end-stage liver disease-sodium; mDF, Modified Maddrey's discriminant function; AARC, APASL ACLF research consortium; CLIF-SOFA, chronic liver failure-sequential organ failure assessment score.

Discussion

This unique study highlighted the importance of FABPs in the prediction of the outcome of alcohol-induced ACLF. The salient features noted in the study were as follows: (a) A-FABP and L-FABP are significantly higher in patients with A-ACLF than controls; (b) A-FABP predicts mortality, organ failure, and sepsis in patients with A-ACLF.

Free fatty acids regulate hormone action and can contribute to metabolic regulation and disease.¹⁶ However, free fatty acids are relatively water-insoluble and have the potential for toxicity.¹⁷ Hence FABPs were discovered, which can bind and buffer the LCFAs and play a crucial role as mediators in various metabolic and biological activities.¹⁷ A-FABP (FABP4) is one of the most abundant proteins in mature adipocytes and adipose tissue. A-FABP mRNA is expressed in the kidney and heart. A-FABP suppresses adipose tissue lipogenesis and promotes lipolysis, with direct effects on the composition of the circulating free fatty acid pool. A-FABP is also implicated in modulating eicosanoid balance by affecting both cyclooxygenase-2 activity and leukotriene A4 stability. Hence A-FABP can influence macrophage function and adipose tissue inflammation. A-FABP opposes peroxisome proliferator-activated receptor-γ (PPARγ) action by regulating ligand availability.¹⁷ Consequently, A-FABP acts on multiple integrated pathways to regulate lipid metabolism and inflammation, impair insulin action, promote glucose production, and contribute to the pathogenesis of metabolic diseases. A-FABP correlates with complications of cirrhosis and is a marker of poor survival.¹⁰ A-FABP is also a marker of nonalcoholic fatty liver disease (NAFLD).^18,19

A-FABP is an adipokine secreted from adipocytes and macrophages. Liver resident macrophages are activated by pathogen-associated molecular pattern (PAMP) and damage-associated molecular pattern (DAMP) in ACLF and are central to the pathogenesis of ACLF.²⁰ Higher recruitment of macrophages leads to compensatory anti-inflammatory response syndrome and immune paresis, thereby leading to sepsis and mortality.^21,22

L-FABP (FABP1) is the first discovered unique member of the sizeable FABP family. L-FABP is an endogenous antioxidant protein and is expressed primarily in the liver and proximal tubular epithelial cells in the kidney.⁸ Urine L-FABP levels are a marker of acute kidney injury and chronic kidney disease.^23,24 L-FABP plays a central role in cell-specific modulation of hepatic lipid metabolism, and downregulation of L-FABP promotes fibrogenesis.²⁵ L-FABP binds ligands (LCFA, fatty acid synthesis inhibitors, fibrates) in the cytosol for cotransport into nuclei and activates nuclear PPARα and is thereby involved in LCFA uptake, transport, and metabolism. Serum L-FABP levels are potential markers of NAFLD and correlate with liver injury in hepatitis C and vascular endothelial growth factor in hepatocellular carcinoma.^8,26

The baseline values of A-FABP are lower than the previous study, which included decompensated cirrhosis with complications.¹⁰ The same study also reported that A-FABP>37.2 ng/ml on multivariate Cox regression analysis predicted 90-day transplant-free survival.¹⁰ The possible reason for us getting a lower cut off value (19.7 ng/ml) is because we excluded patients with sepsis/AKI at baseline, who were included in the previous study.

L-FABP levels were similar to previous studies.⁹ However, we noted that patients with complications or those who expired had lower L-FABP levels, which is ambiguous to the earlier studies, which reported that higher levels predict poorer outcomes in cirrhotics and acute liver failure patients.^8,9 Although this cannot be justified, a plausible explanation is that lower L-FABP levels are associated with a higher degree of fibrosis.²⁵ Further, a failing liver may not be able to produce L-FABP.

There are specific biomarkers such as soluble CD163 and neutrophil gelatinase-associated lipocalin, which have been previously shown to be independently associated with short-term mortality in hepatitis B virus (HBV)–associated ACLF.²⁷ A quantitative proteomic study identified six novel HBV-ACLF biomarker candidates which might provide basic information for the development of HBV-ACLF.²⁸ Some of the biomarkers that have been demonstrated to be useful in alcoholic hepatitis include cytokeratin 18 and its fragments, micro RNA 192, miRNA 30a, long noncoding RNAs, and extracellular vesicles.29, 30, 31, 32 The major advantage of FABPs is the ease of measuring through the ELISA method. FABP analysis is also economical.

Our study had certain limitations. The major limitation was the sample size. It may be premature to justify the results based on this small sample size. In addition, we did not perform the interleukin or endotoxin assay, which would have made the study more robust. Third, we included only male patients, but A-FABP is known to be high in women, and the results may vary in women.³³ The reason for this is the higher prevalence of alcoholism among men in our country. We included only alcohol as an etiology to maintain uniformity in the study.

In conclusion, A-FABP may be a better biomarker which aids in predicting the outcome of alcohol-induced ACLF. If validated in a large, diverse sample, A-FABP can be used as a simple biomarker for prognostication in alcohol-induced ACLF.

Author contributions

A.V.K. and P.N.R. contributed to study concept and design; A.V.K., M.S., and S.T.R. contributed to data collection; V.S. and S.M. contributed to biochemical laboratory support; A.V.K. contributed to statistical analysis; A.V.K., P.K., D.N.R.; and PNR contributed to compilation and critical revision. All members approved the final draft.

Grant support

None.

Financial disclosures

None.

Ethics and IRB approval

Yes.

Article guarantor

Dr. Anand V Kulkarni.

Conflicts of interest

The authors have none to declare.

Appendix A. Supplementary data

The following is/are the Supplementary data to this article: