Resolvin D1 values in patients with alcohol-associated liver disease

Abstract

Background

Resolvin D1 (RvD1), a specialized pro-resolving lipid mediator, has been associated with liver injury in experimental models of alcohol-associated liver disease (ALD). This study aimed to evaluate the relationship between serum RvD1 levels and clinical phenotypes of ALD.

Methods

This case-control study included 82 patients with ALD and 48 healthy controls (HCs). The ALD patients were categorized into three subgroups: alcohol-associated steatotic liver (ASL), alcohol-associated hepatitis (AH), and alcohol-associated cirrhosis (AC). Serum RvD1 levels were measured using the enzyme-linked immunosorbent assay (ELISA) test. The Fibrosis-4 (Fib-4) index was used to assess liver fibrosis in all participants. For the AH and AC groups, the Model for End-stage Liver Disease Sodium (MELD-Na) scores were calculated. Child-Pugh classification was applied to the AC patients and Maddrey’s Discriminant Function was applied to the AH patients. An esophagogastroduodenoscopy was performed in the AH and AC patients to evaluate portal hypertension.

Results

The serum RvD1 levels were significantly lower in the AH and AC groups compared to the ASL and HC groups (p < 0.001), with no significant difference between the AH and AC groups or between the ASL groups and the HC group. In the AC group, patients with ascites and/or esophageal/gastric varices had lower serum RvD1 levels compared to those without (p = 0.013 and p = 0.004, respectively). Additionally, the serum RvD1 levels decreased with advancing Child-Pugh class (p = 0.004). The serum RvD1 levels were inversely correlated with the Fib-4 scores, the MELD-Na and Child-Pugh scores in both the AC group (p < 0.05 for all).

Conclusion

Serum RvD1 levels are associated with the severity of ALD and may represent a promising non-invasive biomarker. However, due to the absence of an external validation cohort, these findings should be interpreted as exploratory. Further validation in independent populations is warranted to confirm the clinical utility of RvD1 in ALD.

Introduction

Alcohol is a multisystem toxin and alcohol-associated liver disease (ALD) is a major cause of end-stage liver disease [1, 2]. The pathogenesis of ALD is not entirely understood and neutrophils have pathogenic significance in ALD pathogenesis [3–6]. Specialized pro-resolving mediators (SPMs) are synthesized from lipids through the action of macrophages and neutrophils, and the resolution process, which starts with inflammation, makes inflammation self-limited and prevents progression to the chronic phase [7, 8].

Resolvins, lipoxins, protectins, and marezins are SPMs that are synthesized from omega-3 polyunsaturated fatty acids (PUFA) [7]. Dysfunctions in the resolution process have been reported to relate to chronic inflammatory diseases, including inflammatory bowel disease, systemic lupus erythematosus, and multiple sclerosis [8, 9].

Resolvin D1 (RvD1) is a member of the resolvin family and has anti-inflammatory influences in various cells [10, 11]. SPMs exert protective influences via several mechanisms, including decreasing neutrophil recruitment, increasing phagocytosis, assembling epithelial cells, and inducing edema clearance [12]. Neutrophils are known to have a high capacity of RvD1 synthesis at the inflammation sites [12, 13].

The role of RvD1 in liver diseases, including metabolic dysfunction-associated steatotic liver disease (MASLD), acute liver injury, and hepatocellular carcinoma (HCC), was evaluated previously [11, 14, 15]. In a recent report, liver injury was alleviated by RvD1 in mice that were exposed to ethanol ingestion [13]. Related to marked liver inflammation and injury in the ethanol-fed mice, elevated levels of resolution molecules were noted [16].

There is a scarcity of data about the serum RvD1 levels in patients with ALD. Regarding the associations between RvD1 and the pathogenesis of ALD, we investigated whether serum RvD1 could serve as diagnostic and prognostic markers in ALD patients.

Materials and methods

Study population

Eighty-two patients with ALD and 48 healthy controls (HCs) who were admitted to our gastroenterology department between April 2023 and June 2024 were enrolled in the study. Ethical permission was received from the Local Ethics Committee (Approval Date: 12.04.2023 and Number: 51). The study protocol complied with the ethical principles of the Declaration of Helsinki (6th revision, 2008). Written informed consent was obtained from all participants.

Participants with any clinical conditions that could manipulate serum RvD1 levels, such as chronic and/or autoimmune inflammatory diseases, severe deterioration in organ functions, acute/chronic infections, any malignancies (including hepatocellular carcinoma), and those with acute or chronic hepatitis B (including inactive hepatitis B virus carriers) and C infection were excluded. Volunteers with diabetes mellitus, illicit drug use, any cause of chronic liver disease, or who had used non-steroidal anti-inflammatory drugs (NSAIDs) for at least 14 days were also excluded. ALD patients with clinical and laboratory findings of acute liver failure and hepatic encephalopathy were excluded as well. Although participants were asked to report all current medication use, the possibility of unreported or unconscious intake of medications such as proton pump inhibitors (PPIs) and antihyperlipidemic drugs cannot be completely ruled out, which may represent a potential confounding factor in the interpretation of serum omega-3 PUFA and RvD1 levels.

The patients who used excessive alcohol of > 60 g per day for men and > 40 g per day for women for at least five years were included in the study [1, 17]. Some patients with ALD were admitted to our department at the Institute of Alcohol and Substance Use Disorders Research and Treatment Center of the province for a gastroenterology consultation. The ALD patients were ingesting excessive amounts of daily ethanol, as it is defined below, without a cessation period when admitted to our department. The diagnosis of an alcohol use disorder was made by a psychiatrist after an interview, according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) [18]. The HCs did not use alcohol, no hepatic steatosis or hepatomegaly was reported in their ultrasonography (USG) scans, and their biochemistry tests were normal.

Data collection

The dose and duration of alcohol intake was noted for each patient with ALD. Body mass index (BMI), comorbidities, and medications were recorded for all the participants. Based on their disease history, the physical findings, their laboratory results, and the radiological results, the patients with ALD were classified into three sub-groups: alcohol-associated steatotic liver disease (ASL), alcohol-associated hepatitis (AH), and alcohol-associated cirrhosis (AC). A USG was applied to all the participants for the radiologic evaluation.

Alcohol-associated hepatitis (AH) was defined according to clinical and laboratory criteria. In this group, the participants’ alcohol intake was continuous for six months or more with < 60 days of abstinence before the onset of jaundice, serum aspartate transaminase (AST) values were higher than alanine aminotransferase (ALT) values (AST/ALT ratio > 1.5), serum total bilirubin > 3.0 mg/dL, and without any other cause [17]. Cirrhotic patients were not included in the AH group. AC was diagnosed according to the stigmata of cirrhosis in the participants’ physical examinations and from the results of the biochemical tests and the USG reports. The patients with ASL were non-cirrhotic and had no clinical or laboratory criteria for AH.

Assessment of disease severity in the patients with ALD

A complete blood count and biochemical tests including C-reactive protein (CRP) were obtained prior to the USG scan. In the USG scan, hepatic steatosis and ascites were examined in all the participants. The patients with AH and AC underwent an upper gastrointestinal endoscopy and they were evaluated for esophageal and/or gastric varices.

To predict liver fibrosis, the fibrosis-4 (FIB-4) index was used to assess liver fibrosis for all the volunteers [19]. In the patients with AH, Maddrey’s discriminant function (MDF) tests and the Model for End-stage Liver Disease Sodium (MELD-Na) scores were calculated [20, 21]. The Child-Pugh scores (CPSs) were obtained for the AC patients and they were grouped according to the Child-Pugh classification (CPC A, B, and C) [22]. The MELD-Na scores were also recorded for the cirrhotic patients.

Serum resolvin D1 measurement

The serum for RvD1 was extracted from the venous blood samples of the participants. The ALD patients continued to ingest excessive amounts of daily ethanol without a cessation period when admitted to our department. Just after admission of the participants, the venous blood samples were taken for biochemical evaluation and serum RvD1 measurement. After centrifuging for 10 min at 30 °C and 5000×g, the supernatant serum samples were stored in Eppendorf tubes at (−) 80 °C until analyses. For the measurement of RvD1, using a serum Enzyme-Linked Immunosorbent Assay (ELISA), the commercially available Human Resolvin D1 ELISA Kit (Bioassay Technology Laboratory, Cat. No. E 7450 Hu, Lot: 202311023) was used according to the instructions of the manufacturer (Intra-Assay: CV < 8%, Inter-Assay: CV < 10%) applying a microplate reader (Biotech Epoch 2 Microplate ELISA Reader, USA). Before the analysis, the reagents were stored at (−) 20 °C for two months. Three samples of known concentration were tested on one plate to assess intra-assay precision and three other samples were tested in separate assays to assess inter-assay precision. The standard curve range for the ELISA kit was 37.5–2400 ng/L and the sensitivity was 19.01 ng/L.

Statistical analysis

Statistical analyses were conducted using IBM SPSS Statistics version 25.0 (IBM Corp., Armonk, NY, USA). The normality of continuous variables was evaluated using both the Shapiro-Wilk and Kolmogorov-Smirnov tests. A variable was considered normally distributed only if both tests yielded p-values greater than 0.05. Normally distributed continuous variables were expressed as means ± standard deviations (SD), while non-normally distributed variables were presented as medians with interquartile ranges (IQR). Categorical variables were summarized using frequencies and percentages.

For comparisons of continuous variables among more than two groups, one-way ANOVA was used for normally distributed variables and the Kruskal–Wallis test for non-normally distributed variables. For two-group comparisons, the Mann–Whitney U test was applied for non-normally distributed variables. When overall group differences were statistically significant (p < 0.05), post-hoc pairwise analyses were conducted. Specifically:

• For normally distributed variables, Bonferroni correction was applied following ANOVA.

• For non-normally distributed variables, Mann–Whitney U tests with Bonferroni correction were performed after Kruskal–Wallis tests.

• For comparisons involving three diagnostic groups and a healthy control group (i.e., six pairwise comparisons), the Bonferroni-adjusted significance threshold was set at α = 0.05/6 = 0.0083.

• In the Child-Pugh classification (three groups: A, B, and C), three pairwise comparisons were performed, and the Bonferroni-adjusted α level was 0.05/3 = 0.0167.

Significant post-hoc differences are marked in Tables 1 and 2 with superscript letters (a–f), and only p-values that remained below the Bonferroni-adjusted thresholds were considered statistically significant.

Table 1.

Clinical, demographic, laboratory characteristics, and serum resolvin D1 levels in patients with alcohol-associated liver diseases and HCs

	Alcohol-associated steatotic liver (n = 22)	Alcohol-associated hepatitis (n = 25)	Alcohol-associated cirrhosis (n = 35)	Healthy controls (n = 48)	p	ⱡ Significant Differences (p < 0.0083)
Demographic and Anthropometric Characteristics
Male, n (%)	22 (100)	23 (92.0)	32 (91.4)	37 (77.1)	0.300 1
Age (year), mean ± SD	48.4 ± 9.5	48.7 ± 7.3	60.0 ± 7.0	53.4 ± 11.7	< 0.001 2,*	d,e ⱡ
BMI(kg/m²), mean ± SD	28.5 ± 2.6	30.8 ± 3.9	29.5 ± 3.8	25.1 ± 3.6	< 0.001 2,*	a, b, c ⱡ
Duration of alcohol intake (years), mean ± SD	25.5 ± 12.2	24.2 ± 8.9	35.8 ± 12.1	-	< 0.001 2,*	d, e ⱡ
Imaging and Clinical Findings
Presence of hepatic steatosis, n (%)	22 (100)	24 (96.0)	3 (8.6)	-	< 0.001 1,*	d, e
Presence of ascites, n (%)	0 (0)	9 (36.0)	18 (51.4)	-	< 0.001 1,*	e, f
Presence of esophageal and/or gastric varices, n (%)	-	4 (16.0)	22 (62.9)	-	< 0.001 1,*	d
Liver Disease Severity Scores
Fibrosis-4 index, median (IQR)	0.9 (0.7–1.6)	5.9 (3.3–9.1)	4.8 (2.7–7.4)	1.0 (0.7–1.2)	< 0.001 3,*	a,b,e,f ⱡ
MELD-Na score, median (IQR)	-	16.0 (12.5–22)	16.0 (10–21)	-	0.260 4
MDF, median (IQR)	-	16.0 (12.8–28.2)	-	-	-
Child-Pugh score, median (IQR)	-	-	7.0 (6–10)	-	N/A
Child-Pugh class, n (%)
A	-	-	11 (31.4)	-	N/A
B			13 (37.2)
C			11 (31.4)
Laboratory Parameters
Neutrophils (x103/µL), median (IQR)	4.2 (3.1–6.3)	6.5 (4.9–11.0)	3.4 (2.6-5.0)	4.6 (3.7–5.7)	< 0.001 3,*	a, b, d, f ⱡ
CRP (mg/L), median (IQR)	5 (1.3–11.6)	14.9 (7.5–25.8)	6.1 (2.2–21.1)	1.7 (0.8–3.3)	< 0.001 3,*	a, b, c, f ⱡ
Serum resolvin D1 (ng/L), median (IQR)	2341.1 (1491.5-2413.9)	475.8 (403.7-617.5)	468.2 (359.5-540.4)	2477.8 (1144.1-4139.2)	< 0.001 3,*	a, b, e,f ⱡ

Abbreviations: IQR Inter quartile range, SD Standard deviation, BMI Body mass index, MELD-Na Model for end-stage liver disease sodium, MDF Maddrey’s discriminant function, CRP C-reactive protein

Explanation:* Statistical significance was initially set at a p-value of < 0.05. Significant pairwise differences are indicated as follows: a = healthy controls vs. alcohol-associated cirrhosis, b = healthy controls vs. alcohol-associated hepatitis, c = healthy controls vs. alcohol-associated steatotic liver, d = alcohol-associated cirrhosis vs. alcohol-associated hepatitis, e = alcohol-associated cirrhosis vs. alcohol-associated steatotic liver, f = alcohol-associated hepatitis vs. alcohol-associated steatotic Liver. The methodology used to correct for multiple comparisons was as follows: 1: Chi-Square Test (for categorical variables; p-values adjusted using the Bonferroni method for comparisons involving more than two groups), 2: ANOVA Test (Post Hoc: Bonferroni correction), 3: Kruskal-Wallis Test (Post Hoc: Mann-Whitney U Test with Bonferroni correction), 4: Mann-Whitney U Test (for non-normally distributed variables). Multiple comparisons were corrected using the Bonferroni method. For group comparisons involving more than two groups, ANOVA or Kruskal-Wallis tests were followed by post hoc pairwise analyses. Specifically, post hoc t-tests and Mann-Whitney U tests were Bonferroni-adjusted to account for the number of pairwise comparisons (6 group pairs in total: cirrhosis vs. hepatitis, cirrhosis vs. steatosis, cirrhosis vs. control, hepatitis vs. steatosis, hepatitis vs. control, steatosis vs. control). Accordingly, the adjusted alpha level was 0.05/6 = 0.0083. Only p-values that remained significant after Bonferroni correction are marked with an asterisk (ⱡ). Specific pairwise differences are indicated by letters a, b, c, d, e, and f in Table 1

Table 2.

Serum resolvin D1 concentrations and clinical and laboratory variables in patients with alcohol-associated hepatitis and cirrhosis

			Serum Resolvin D1 (ng/L)
			n	Median	IQR		p
Alcohol-associated hepatitis (n = 25)	Hepatic steatosis	Present	24	482.6	402.3	631.2	N/A
		Absent	1	441
	Ascites	Present	9	412.7	305.9	609.9	0.169 1
		Absent	16	500.4	429.1	657.7
	Esophagus and/or gastric varices	Present	4	677.0	410.0	764.0	0.203 1
		Absent	21	464.5	403.7	575.1
Alcohol-associated cirrhosis (n = 35)	Hepatic steatosis	Present	3	451.0 - 521.5–1040.7			0.217 1
		Absent	32	458.2	355.1	539.2
	Ascites	Present	18	369.7	278.3	490.1	0.013 1,*
		Absent	17	521.5	422.5	668.8
	Esophagus and/or gastric varices	Present	22	369.7	303.6	511.6	0.004 1,*
		Absent	13	526.3	459.6	705.9
	Child-Pugh Classification	A a	11	535.5	448.2	870.5	0.004 2,*
		B	13	487.0	351.8	570.8
		C a	11	367.3	268.1	433.9

*Statistically significance level was lower than 0.05. N/A: There are not enough valid cases to perform the Mann-Whitney Test for Resolvin D1 (ng/L). 1: Mann-Whitney U test, 2: Kruskal-Wallis test (Post Hoc: Mann-Whitney U test with Bonferroni correction). Explanation: Multiple pairwise comparisons were conducted among Child-Pugh Classification (A, B, and C) for serum resolvin D1 levels (A vs B, A vs C, B vs C). To reduce the risk of Type I error due to multiple testing, a Bonferroni correction was applied, adjusting the significance threshold to 0.05/3 = 0.0167. After correction, the difference between Child-Pugh class A and C remained statistically significant (pa = 0.002 < 0.0167), while the difference between classes B and C (p = 0.040) was no longer significant. This approach minimizes the chance of false-positive findings due to multiple comparisons, ensuring more reliable results.

Explanation: Multiple pairwise comparisons were conducted among Child-Pugh Classification (A, B, and C) for serum resolvin D1 levels (A vs B, A vs C, B vs C). To reduce the risk of Type I error due to multiple testing, a Bonferroni correction was applied, adjusting the significance threshold to 0.05/3 = 0.0167. After correction, the difference between Child-Pugh class A and C remained statistically significant (pa = 0.002 < 0.0167), while the difference between classes B and C (p = 0.040) was no longer significant. This approach minimizes the chance of false-positive findings due to multiple comparisons, ensuring more reliable results.

Categorical variables were analyzed using the chi-square test, and where appropriate, post-hoc pairwise comparisons were conducted using adjusted standardized residuals with Bonferroni correction (adjusted p < 0.0083 for six comparisons).

Spearman’s rank correlation coefficient was used to assess the relationships between non-normally distributed continuous variables and serum resolvin D1 levels. To account for multiple correlation tests within each subgroup (alcohol-associated steatotic liver, alcohol-associated hepatit, and Alcohol-associated cirrhosis), Bonferroni correction was applied to the significance threshold as follows:

• Alcohol-associated steatotic Liver group: 4 correlation tests → α = 0.05/4 = 0.0125.

• Alcohol-associated hepatit group: 6 correlation tests → α = 0.05/6 = 0.0083.

• Alcohol-associated cirrhosis group: 6 correlation tests → α = 0.0083.

Only correlations with p-values below these adjusted thresholds were considered statistically significant.

Receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic performance of serum resolvin D1 levels. Throughout the study, a two-tailed p-value < 0.05 was considered statistically significant unless otherwise adjusted by Bonferroni correction to control the family-wise error rate due to multiple comparisons.

Statistical power analysis

To evaluate the statistical power of the study, a post-hoc power analysis was performed using the G*Power (3.1.9.7) software based on the actual sample sizes of the groups, which were 22, 25, 35, and 48 participants, respectively. A medium effect size (Cohen’s f = 0.25, approximately equivalent to Cohen’s d = 0.50) and a significance level of α = 0.05 were assumed. The calculated statistical power was approximately 99% (0.99), indicating a very high likelihood of detecting clinically meaningful differences in the study.

Results

Eighty-two patients with ALD (77 males, 5 females), including 22 ASL, 25 AH, and 35 AC, as well as 48 healthy controls (HC) (37 males, 11 females), participated in the study. Males dominated the study population. Table 1 presents the demographic, clinical, and histopathological characteristics of the study population. The median FIB-4 index values were similar between the AH and AC groups, as well as between the ASL and HC groups. The FIB-4 index differed significantly among groups (p < 0.001, Kruskal-Wallis test). Post hoc pairwise comparisons with Bonferroni correction revealed that FIB-4 values in the AH and AC groups were significantly higher than those in the ASL and HC groups (p < 0.0083). The median neutrophil count was the highest in the AH group and the median CRP value was the lowest in the HC group (Table 1).

In total, the median serum RvD1 levels were lower in the ALD group compared to the HC group (530.5 [409.9–1005.7] ng/L vs. 2477.8 [1144.1–4139.2] ng/L, respectively, p < 0.001) (Fig. 1). There was a statistically significant difference in serum RvD1 levels among the ALD subgroups and healthy controls (p < 0.001, Kruskal-Wallis test). Post hoc pairwise comparisons with Bonferroni correction (adjusted alpha = 0.0083) showed that median serum RvD1 concentrations in the AH (475.8 ng/L) and AC (468.2 ng/L) groups were significantly lower than those in the ASL (2341.1 ng/L) and HC (2477.8 ng/L) groups (p < 0.0083). No significant difference was observed between the AH and AC groups, nor between the ASL and HC groups (Table 1; Fig. 2).

Fig. 1.

Serum resolvin D1 levels in patients with alcohol-associated liver disease and HCs (p < 0.001)

Fig. 2.

Serum resolvin D1 levels in patients with alcohol-associated steatotic liver, alcohol-associated hepatitis, alcohol-associated cirrhosis and HCs

In the AH group, median serum RvD1 levels did not differ significantly according to the presence or absence of hepatic steatosis, ascites, or esophageal and/or gastric varices (all p > 0.05). Statistical analysis for steatosis was limited by the very low number of patients without steatosis (n = 1), preventing meaningful comparison. Similarly, in the AC group, median serum RvD1 levels were comparable between patients with and without hepatic steatosis (p = 0.217). However, patients with ascites showed significantly lower median serum RvD1 levels compared to those without ascites (369.7 ng/L vs. 521.5 ng/L, p = 0.013), and patients with esophageal and/or gastric varices also had significantly reduced serum RvD1 levels relative to those without varices (369.7 ng/L vs. 526.3 ng/L, p = 0.004). Furthermore, serum RvD1 concentrations significantly decreased with advancing Child-Pugh Classification (Class A: 535.5 ng/L, Class B: 487.0 ng/L, Class C: 367.3 ng/L; p = 0.004). Post hoc pairwise comparisons with Bonferroni correction (adjusted significance threshold α = 0.0167) revealed that the difference between Classes A and C remained statistically significant (p = 0.002), while the difference between Classes B and C (p = 0.040) did not meet the corrected threshold and was thus considered non-significant (Table 2).

No statistically significant correlations were found between serum RvD1 levels and BMI, duration of alcohol use, neutrophil count, or Fib-4 score in patients with alcohol-associated steatotic liver (all p > 0.05 after Bonferroni correction, adjusted α = 0.0125). Similarly, in the alcohol-associated hepatitis group, no significant correlations were detected between serum RvD1 levels and BMI, duration of alcohol use, neutrophil count, MELD-Na scores, or Maddrey’s discriminant function (MDF) test results after applying Bonferroni correction (adjusted α = 0.0083). Although a negative correlation between serum RvD1 levels and Fib-4 score was observed in this group (rho = (-) 0.478, p = 0.016), this did not remain significant after correction. In the alcohol-associated cirrhosis group, a negative correlation between serum RvD1 levels and duration of alcohol use was noted (rho = (-) 0.372, p = 0.028); however, this was not statistically significant after Bonferroni adjustment (adjusted α = 0.0083). Significant inverse correlations were identified between serum RvD1 levels and Fib-4 index (rho = (-) 0.484, p = 0.003), MELD-Na score (rho= (-) 0.685, p < 0.001), and Child-Pugh score (rho = (-) 0.605, p < 0.001), all of which remained significant after correction (p < adjusted α) (Table 3).

Table 3.

Correlations between the serum resolvin D1 levels and the clinical and laboratory variables of the patients with alcohol-associated liver diseases

Serum Resolvin D1 (ng/L)
Alcohol-associated steatotic liver (n = 22)	rho	p	Bonferroni-adjusted α	Significant After Correction?
BMI	(-) 0.173	0.442	0.0125	No
Duration of alcohol intake	0.072	0.749	0.0125	No
FIB-4 index	(-) 0.011	0.962	0.0125	No
Neutrophils	0.228	0.308	0.0125	No
Alcohol-associated hepatitis (n = 25)	rho	p	Bonferroni-adjusted α	Significant After Correction?
BMI	(-) 0.155	0.459	0.0083	No
Duration of alcohol intake	(-) 0.055	0.793	0.0083	No
FIB-4 index	(-) 0.478	0.016 *	0.0083	No
MELD-Na Score	(-) 0.327	0.111	0.0083	No
MDF	(-) 0.355	0.082	0.0083	No
Neutrophils	0.036	0.865	0.0083	No
Alcohol-associated cirrhosis (n = 35)	rho	p	Bonferroni-adjusted α	Significant After Correction?
BMI	(-) 0.170	0.330	0.0083	No
Duration of alcohol intake	(-) 0.372	0.028 *	0.0083	No
FIB-4 index	(-) 0.484	0.003 *	0.0083	Yes ⱡ
MELD-Na Score	(-) 0.685	< 0.001 *	0.0083	Yes ⱡ
Child-Pugh Score	(-) 0.605	< 0.001 *	0.0083	Yes ⱡ
Neutrophils	0.044	0.801	0.0083	No

Abbreviations: BMI Body Mass Index, MELD-Na Model for end-stage liver disease sodium, MDF Maddrey’s discriminant function

^*Significant at the level of 0.05. p: Spearman’s Correlation, rho: Spearman’s Correlation Coefficient. Bonferroni correction applied per group according to number of correlations tested: Alcohol-associated steatotic Liver: 4 tests → α = 0.05/4 = 0.0125, Alcohol-associated hepatitis: 6 tests → α = 0.05/6 ≈ 0.0083, Alcohol-associated cirrhosis: 6 tests → α = 0.05/6 ≈ 0.0083. ⱡ: Only correlations with p-values below the adjusted alpha are considered significant. Significant correlations after correction were found only in the cirrhosis group for FIB-4 index, MELD-Na score, and Child-Pugh score

ROC curve analysis was performed to evaluate the diagnostic ability of serum RvD1 concentrations to differentiate healthy controls (HCs) from patients with alcohol-associated liver disease (ALD). The analysis demonstrated a statistically significant discriminative power, with an area under the curve (AUC) of 0.784 (95% CI: 0.688–0.879, p < 0.001) (Fig. 3). A serum RvD1 concentration above 1044.93 ng/L was identified as the optimal cut-off value, predicting healthy status with a sensitivity of 79.2% and a specificity of 78%.

Fig. 3.

Receiver operating characteristics curve analyses of the serum resolvin D1 concentrations for the differentiation of HC group from patients with ALD (p < 0.001)

Discussion

The burden of ALD continues to grow. Although ethanol-related liver inflammation and injury can be steady in some patients, one-fifth of alcohol abusers can progress to end-stage liver disease [1]. The dose, duration, and being female are well-accepted risk factors in the progression of end-stage liver disease. However, it remains unclear how to predict which patients may develop severe disease [1–3].

Neutrophilic infiltration of the liver and peripheral neutrophilia have pathogenic significance and poorer clinical outcomes in ALD pathogenesis [4–6]. In our study, the median neutrophil count was the highest in the AH group, but it was similar in the other groups. For AH, neutrophilia, along with neutrophil accumulation in the liver parenchyma, is a unique entity, and it can mimic a febrile disease [5]. In the AH group, higher neutrophil counts along with elevated CRP values in our results were consistent with the results of previous reports [5, 6].

As a resolution molecule, RvD1 exhibited pathogenic significance in the liver disease models and in the limited number of clinical trials [11, 13–16]. In the MASLD patients, disease severity was associated with low levels of RvD1 [14]. In the patients with trauma and hemorrhagic shock, the plasma levels of RvD1 were lower than the levels for the HCs [15]. In ethanol- and lipopolysaccharide-challenged mice that mimicked human AH, their RvD1 concentrations were higher in their liver tissue and plasma compared to the control group [13]. In that study, the increase in RvD1 levels was declared as an endogenous adaptive response to ethanol-related hepatic inflammation. On the other hand, RvD1 was administered to the treatment group of ethanol- and lipopolysaccharide-challenged mice and hepatic inflammation was attenuated as it was confirmed histologically. Despite the RvD1 supplementation, attenuation, but not full regression, of the hepatitis was thought to be related to insufficient RvD1 elevation in the liver or to dysregulated interactions between RvD1 and its receptors as a result of ethanol ingestion [13]. In an alcohol-associated steatohepatitis model, ethanol-fed mice had altered hepatic concentrations of several PUFA hydroxy-metabolites, including eicosapentaenoic and docosahexaenoic acid, which are the precursors of SPMs. Additionally, the ethanol-fed mice had elevated SPMs associated with higher serum transaminase levels and enhanced hepatic steatosis and inflammation compared to saturated-fatty-acid-fed mice alone. Higher SPMs in the ethanol-fed mice were attributed to this group’s natural and adaptive immune responses to ethanol [16]. It was also stated that SPMs have anti-inflammatory features but cannot outweigh pro-inflammatory molecules, including oxylipins [2].

The failure of resolution and the low levels of SPMs in the tissue and circulation can lead to chronic inflammation and tissue damage [8, 9]. The growing body of evidence demonstrates the pathogenic significance of RvD1 in ALD pathogenesis [12, 16, 23]. Due to the anti-inflammatory and pro-resolving properties, RvD1 was declared to play a significant role in acute liver injury, liver ischemia/reperfusion injury, NAFLD, liver fibrosis, and liver cancer [11]. Regarding the pathophysiological issues, the scarcity of clinical research in humans was a disadvantage for us to establish a causal relationship between low serum RvD1 levels and severe ALD. Ethanol-related hepatic inflammation is a major pathogenic mechanism in ALD, but several pathways accompany the progression of the disease [3]. Failure of the resolution process via the increased consumption due to chronic inflammation and insufficient synthesis was declared to have pathophysiological significance in some chronic inflammatory diseases [8]. Thus, based on the experimental investigations and the limited number of clinical trials in ALD, increased consumption or insufficient synthesis may be major contributing factors for lower serum RvD1 concentrations in the severe ALD patients, and mostly in the AH and AC groups, who were continuing to ingest excessive amounts of daily ethanol without an abstinence period when admitted to our department [13, 23]. Decreased serum RvD1 levels might relate to the incapability of neutrophilic RvD1 production due to ongoing ethanol-related hepatic inflammation and injuries in the AH and AC patients. As noted earlier, dysfunctions between RvD1 and its receptors due to ethanol toxicity might be another factor of unresolved liver inflammation [13]. Although acute hepatic injury in the experimental ALD models leads to elevated RvD1 levels as an adaptive response, chronic ethanol ingestion might decrease RvD1 production in humans [13, 15, 16, 23]. In light of the current data, we hypothesized that altered serum RvD1 levels might relate to the phenotypes of ALD patients. Our results revealed that the median serum RvD1 levels were lower in the ALD patients (in total) compared to the HCs. The median serum RvD1 levels were similar between the AH and AC groups and between the ASL and HC groups, and they were lower in both the AH and AC groups. Despite the elevated neutrophil and CRP concentrations relating to inflammatory activity, our results revealed low serum RvD1 levels in the AH and AC groups.

Resolvins are omega-3 PUFA derivatives and a nutritional deficiency of omega-3 PUFA in alcoholic patients may also contribute to low serum RvD1 levels [24]. As a limitation, we did not assess traditional nutritional scores in either the patient or control groups. Given that omega-3 polyunsaturated fatty acids (PUFAs), the known precursors of RvD1, are influenced by dietary intake, the lack of nutritional data may represent a potential confounding factor. Measuring omega-3 PUFA levels and analyzing their correlation with RvD1 in both groups could have provided deeper insight into the observed decrease in RvD1 levels in ALD patients. This limitation is particularly relevant in severe ALD cases, where nutritional compromise is more likely. Acknowledging this, future studies should incorporate nutritional assessment to better understand the metabolic dynamics affecting RvD1 and to minimize potential bias in interpreting its levels. In our study, we systematically excluded individuals with reported comorbidities and chronic medication use to minimize confounding effects, as detailed in the Methods section. Specifically, participants using NSAIDs were excluded due to their known inhibitory effects on enzymes involved in the biosynthesis of resolvins, as previously described [15]. However, while participants reported no active medication use at the time of inclusion, the possibility of unreported or unrecognized use of medications—such as proton pump inhibitors (PPIs) and antihyperlipidemic drugs—cannot be entirely ruled out. PPIs, in particular, have been associated with long-term malabsorption of vitamin B12, magnesium, and potentially omega-3 polyunsaturated fatty acids (PUFAs), which are precursors of RvD1 [25]. These factors may have contributed to lower serum omega-3 PUFA and RvD1 levels and represent a potential source of residual confounding. We acknowledge this as a limitation of our study and recommend that future research includes systematic medication assessments to ensure more accurate interpretation of lipid mediator profiles.

In a recent study, divergent pro-inflammatory lipid mediators (PLMs) and SPMs, including RvD1, were measured in the plasma of AH patients using liquid chromatography with tandem mass spectrometry along with ELISA test [23]. A disrupted balance was detected between the PLMs and SPMs, and the PLM/SPM ratios were higher in the AH patients compared to HCs. An imbalance between the PLMs and SPMs values, including RvD1, was attributed to hepatic and systemic inflammation encountered in AH. The RvD1 levels of the AH patients had a trend towards lower values compared to the HCs. Additionally, the AH patients in the abstinence period had elevated SPMs. On the other hand, low levels of SPMs in the AH patients were ascribed to the impaired neutrophilic SPM synthesis [23]. The results of that study were parallel to ours, and the results demonstrated the low levels of circulating RvD1 concentrations in the AH patients. This finding suggests a potential link between RvD1 levels and the pathophysiology of AH, warranting further investigation into their role as biomarkers or therapeutic targets in managing this condition. Additionally, understanding the mechanisms behind these low concentrations could lead to novel treatment strategies that are aimed at restoring the RvD1 levels in affected patients.

Accumulating data exist about the pathogenic roles of neutrophils in AH [7, 23, 26]. In response to bacterial products and chemical agents (including ethanol), neutrophils secrete extracellular traps (NETs), including DNA fragments, histones, and bactericidal proteins, via a process called NETosis, and it was declared to contribute to liver injury in AH patients and experimental models [26]. NETosis leads to the transformation of the physiologically active high-density neutrophils (granulocytes) to low-density neutrophils, which have functional defects, including impaired synthesis of SPMs observed in AH patients [23, 26]. Deficient neutrophilic activity may be a contributing factor for decreased serum RvD1 levels in AH, and further investigations are needed to delineate the pathophysiological mechanisms that underlie low RvD1 values in severe ALD patients.

There was no correlation between the serum RvD1 levels and the duration of alcohol use or neutrophil counts in the ALD patients in our results. With respect to duration and dose of alcohol intake, alcohol abusers may not provide correct information and liver injury can exhibit inter-individual differences [1, 3]. However, neutrophilic increase in circulation and in the liver was reported with poorer clinical outcomes, with respect to RvD1 production, rather than the neutrophil counts, the pathogenic significance of neutrophils might exert an influence on liver injury [4–6, 23, 26].

As expected, the BMI of the ALD patients was higher compared to the HCs in our cohort. Obesity is associated with chronic low-grade inflammation, partly due to impaired resolution processes in adipose tissue [27]. In both obese humans and mice, reduced levels of SPMs and their precursors have been observed [27, 28]. Although RvD1 and RvD2 are key SPMs in adipose tissue homeostasis, we did not find a correlation between serum RvD1 levels and BMI in our cohort [27]. This finding suggests that decreased serum RvD1 levels in ALD may not be solely driven by increased adiposity but could involve other mechanisms such as neutrophil dysfunction, which warrants further investigation [23, 26, 27].

Our results revealed a positively significant association between the serum RvD1 levels, CPSs, and MELD-Na scores in the AC patients. Serum RvD1 might be a practical biomarker for disease severity of AC. There were inverse correlations between the serum RvD1 levels and the MELD-Na scores and MDF in the AH group. However, the correlations were not statistically significant. For the AC group, these findings might be a clue for the use of serum RvD1 in clinical practice. Since the study was a cross-sectional case-control study, we did not follow up with the ALD patients for the clinical outcomes. Thus, as a limitation, we could not make a constant conclusion about whether the magnitude of serum RvD1 change would be clinically meaningful or not. AH may present as an acute dysfunction of the liver in both ASL and AC patients. However, we did not include cirrhotic patients in the AH group and this may be a possible reason for the insignificant correlation between the MELD-Na scores and the serum RvD1 levels in these AH patients. The Fib-4 scores were negatively correlated to the serum RvD1 concentrations in the AH group. The Fib-4 index was reported to have a predictive value for hepatocellular carcinoma in the ALD patients and monitoring the serum RvD1 might be an adjunctive method for disease progression in the follow up of patients with AH [29].

In the AH group, there was no statistically significant difference in serum RvD1 levels between patients with or without hepatic steatosis or ascites. However, in the AC group, patients with ascites or esophageal and/or gastric varices showed significantly lower serum RvD1 levels compared to those without varices. The presence of ascites and varices are indicators of portal hypertension, with ascites also signifying hepatic decompensation [22]. Although patients with ascites in the AH group showed lower serum RvD1 levels, this difference was not statistically significant. Despite a significant inverse correlation between Child-Pugh scores and serum RvD1 levels, after Bonferroni correction, a significant difference in serum RvD1 levels was observed only between the Child-Pugh A and Child-Pugh C groups. Considering hepatic decompensation, a larger sample size may provide more robust findings regarding serum RvD1 concentrations. Although the CPC remains a widely used prognostic tool for AC, it includes subjective variables such as ascites and encephalopathy [22]. Therefore, if confirmed in future studies, serum RvD1 measurement could emerge as a valuable prognostic biomarker for AC patients. During clinical follow-up sessions for AC patients, changes in serum RvD1 levels could serve as an objective indicator of disease severity and its progression. Additionally, for both AH and AC, prospective evaluation with a larger sample size could determine whether changes in serum RvD1 magnitudes might be a useful biomarker in predicting hepatic decompensation and complications associated with portal hypertension.

A growing body of evidence demonstrates the therapeutic potential of SPMs [8, 10, 12]. Excessive oxidative stress can lead to hepatocyte apoptosis and cellular dysfunction. SPMs are involved in mitigating oxidative stress caused by viruses, alcohol, and other hepatotoxins. By increasing antioxidant molecules in various clinical situations, such as hepatitis, they can reduce the hepatotoxic effects of reactive oxygen species (ROS) [30]. For example, Maresin 1 has been shown to prevent liver damage by reducing ROS levels [31]. SPMs, which exhibit anti-inflammatory properties through various mechanisms, have also been reported to alleviate chronic liver disease-related pain, in addition to regulating neutrophil apoptosis and clearing chemokines at the site of inflammation [32]. In ALD models, the therapeutic efficacy of RvD1 and other SPMs has been evaluated, with protective effects of these molecules being documented in the literature [12, 15, 27]. Sordi et al. investigated the effects of RvD1 on organ injury and dysfunction caused by hemorrhagic shock in rats [15]. After intravenous treatment with RvD1, kidney and liver injury parameters were attenuated through inhibition of the nuclear factor kappa B (NF-κB) pathway, and by reducing the expression of tumor necrosis factor alpha, interleukin 1 beta, and interleukin 6 [15]. Lenabasum, an orally active drug that contains ajulemic acid, stimulates SPMs. Additionally, fish oil, rich in DHA and EPA, has demonstrated therapeutic efficacy [8]. Abstinence remains the primary goal in treating alcohol use disorders, and supplementing with RvD1 might represent a promising approach to ameliorate liver injury in ALD.

Study limitations

This study represents the first clinical investigation of serum RvD1 levels in patients with ALD and should be viewed as a preliminary step toward future research. However, several important limitations must be considered when interpreting the findings. The most significant limitation is the absence of an independent validation cohort. Without external validation in a separate population, the clinical applicability and generalizability of RvD1 as a biomarker for ALD remain uncertain. This limitation restricts our ability to make definitive claims about its diagnostic or prognostic utility in broader clinical settings. Future studies involving larger, multicenter cohorts are necessary to validate our findings and confirm the potential utility of RvD1 in clinical practice. Additionally, the relatively small sample size—particularly in the ASL subgroup—limits statistical power. As a single-center study, potential site-specific biases cannot be excluded. Liver biopsy, which could provide more definitive diagnostic confirmation and tissue-level RvD1 assessment, was not performed due to its invasiveness and ethical concerns. Moreover, the gender distribution in our cohort was skewed toward male patients, reflecting the higher prevalence of ALD in males. This imbalance limits our ability to explore potential sex-related differences in RvD1 levels, which may be relevant given known variations in alcohol metabolism, inflammatory responses, and SPMs synthesis between sexes. Larger studies with more balanced gender representation are needed to clarify these potential differences.

While our study has certain limitations, including the absence of a validation cohort and the cross-sectional design, a major strength of our study is the high statistical power (approximately 99%) achieved with the current sample sizes, which supports the reliability of our findings and reduces the risk of Type II error. Although a prospective power analysis was not feasible due to the lack of prior data, a post-hoc power analysis confirmed sufficient statistical power (1-β = 0.99), further validating the robustness of the observed results. These findings suggest that our study is sufficiently powered to detect clinically meaningful differences, enhancing the credibility of the results.

Conclusion

In patients with alcohol-associated cirrhosis (AC), serum RvD1 levels may reflect disease severity and could serve as a potential auxiliary biomarker during patient follow-up. However, the current findings are preliminary and should be interpreted with caution due to the absence of an independent validation cohort. Without replication in larger, external populations, the clinical applicability and generalizability of RvD1 as a diagnostic or prognostic marker in ALD cannot be confirmed at this stage. To better understand the relationship between serum RvD1 levels and disease characteristics in ALD—particularly in ASL patients—larger studies with more balanced representation are required. Additionally, considering genetic polymorphisms and ethnic and geographic differences in ethanol metabolism, multicenter studies involving diverse populations are essential to determine the broader clinical utility of serum RvD1. Future validation across laboratories will also be necessary before serum RvD1 can be considered a reliable, cost-effective tool for routine clinical use. While our results are promising and may guide future investigations, they primarily serve as a hypothesis-generating foundation rather than definitive evidence for clinical application. Further studies should also explore how RvD1 levels may contribute to the development of novel diagnostic strategies and therapeutic targets in ALD.

Authors’ contributions

Author Contributions: Conceptualization, İ.K.; methodology, İ.K.; investigation, İ.K.; resources, İ.K. and M.S.; data curation, İ.K.; writing—original draft preparation, İ.K.; writing—review and editing, S.B.; funding acquisition, İ.K., M.S. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request. Due to confidentiality and ethical restrictions, the data are not publicly available. However, they can be shared with the journal editor or reviewers upon request during the peer-review process.

Declarations

Competing interests

The authors declare no competing interests.