Soluble CD163, a macrophage activation marker, is independently associated with fibrosis in patients with chronic viral hepatitis B and C

Authors


  • Potential conflict of interest: Nothing to report.

  • Supported by the NOVO Nordisk Foundation, the Danish Strategic Research Council (10-092797), and grants from the National Health and Medical Research Council of Australia (632630, 1049857, 1053206) and the Robert W. Storr Bequest to the Sydney Medical Foundation of the University of Sydney.

Abstract

Macrophages are involved in inflammation and liver fibrosis and soluble (s)CD163 is a specific marker of activated macrophages. We investigated associations between sCD163 and biochemical and histological parameters of inflammatory activity and fibrosis in 551 patients with chronic hepatitis C virus (HCV) and 203 patients with chronic hepatitis B virus (HBV) before antiviral treatment. Scheuer histological scores of activity and fibrosis were obtained. Clinical, biochemical, and metabolic parameters were recorded. We measured sCD163 by enzyme-linked immunosorbent assay (ELISA). Soluble CD163 was higher in patients with HCV compared to HBV (3.6 [interquartile range (IQR) 2.5-5.4] versus 2.4 [IQR 1.8-3.6] mg/L, P < 0.001). sCD163 was associated with fibrosis stages for both HCV (odds ratio [OR] 1.49, 95% confidence interval [CI]: 1.38-1.61) and HBV (OR 1.32, 95% CI: 1.17-1.49) patients, with highest levels in patients with advanced fibrosis and cirrhosis. sCD163 was a marker of fibrosis independent of other biochemical parameters and known risk factors. We created two novel sCD163-based fibrosis scores, CD163-HCV-FS and CD163-HBV-FS, which showed areas under the receiver operating characteristics curve (AUROC) of 0.79 (95% CI: 0.74-0.83) and 0.71 (95% CI: 0.62-0.79), respectively, for significant fibrosis. Compared to existing fibrosis scores, CD163-HCV-FS was significantly superior to the aspartate aminotransferase (AST) to platelet ratio index (APRI) for all fibrosis stages and to FIB-4 for significant fibrosis, but CD163-HBV-FS was not. Conclusion: sCD163 levels are increased in patients with chronic viral hepatitis, reflecting macrophage activation. Increased sCD163 is associated with the severity of disease and predicts fibrosis. A sCD163-based fibrosis score, CD163-HCV-FS, is superior to APRI and FIB-4 for the diagnosis of significant fibrosis in patients with HCV infection. (Hepatology 2014;60:521–530)

Abbreviations
ALP

alkaline phosphatase

ALT

alanine transaminase

ANOVA

analysis of variance

APRI

AST to platelet ratio index

AST

aspartate transaminase

AUROC

area under the receiver operating characteristics curve

BMI

body mass index

CD163-HBV-FS

sCD163-based fibrosis score for HBV patients

CD163-HCV-FS

sCD163-based fibrosis score for HCV patients

ELISA

enzyme-linked immunosorbent assay

GGT

gamma-glutamyltransferase

HBV

hepatitis B virus

HCV

hepatitis C virus

HOMA-IR

homeostatic model assessment of insulin resistance

IL-6

interleukin-6

INR

international normalized ratio

IQR

interquartile range

MELD

Model for Endstage Liver Disease

NPV

negative predictive value

PCR

polymerase chain reaction

PPV

positive predictive value

sCD163

soluble CD163

TGF-β

transforming growth factor-β

TLR

toll-like receptor

TNF-α

tumor necrosis factor-α

There is increasing evidence that macrophages are involved in inflammation and fibrosis in liver diseases. In the liver, macrophages are observed in close proximity with cells involved in fibrosis formation and produce a number of growth factors and proinflammatory substances involved in this process.[1, 2] The mechanism of action of monocytes and resident liver macrophages (Kupffer cells) in inflammation and fibrogenesis in chronic viral hepatitis is not fully understood, but studies suggest that macrophages function differently in the presence of hepatitis C (HCV) and hepatitis B (HBV) virus.[3-5] Further, a pathway involving activation of monocytes and macrophages by endotoxin with subsequent up-regulation of proinflammatory cytokines, such as tumor necrosis factor (TNF)-α and interleukin (IL)-6, has been suggested.[6] Moreover, activated Kupffer cells secrete transforming growth factor (TGF)-β, which activates hepatic stellate cells and induces fibrosis.[7] Thus, Kupffer cells play an important role in liver inflammation and fibrosis and macrophage-specific markers may function as specific fibrosis and/or inflammatory markers. Furthermore, future therapies may target Kupffer cells specifically.[8]

CD163 is a lineage-specific hemoglobin-haptoglobin scavenger receptor expressed exclusively on monocytes and macrophages and up-regulated in conditions with macrophage activation.[9-11] CD163 is shed from the macrophage surface into the circulation upon activation of cell surface Toll-like receptors (TLRs) and is found in the blood as soluble CD163 (sCD163).[12] sCD163 has been investigated as a biochemical marker of macrophage activation.[13]

We have shown close associations between sCD163 levels, liver dysfunction, and portal hypertension in patients with cirrhosis.[14, 15] In these reports, we primarily studied sCD163 levels in patients with alcoholic liver disease, but Kupffer cells are also known to be activated in chronic viral hepatitis. Hiraoka et al.[16, 17] studied cell surface and soluble CD163 in patients with acute and chronic hepatitis and found that the frequencies of macrophages expressing CD163 and levels of sCD163 were higher in acute viral hepatitis compared to chronic viral hepatitis. In line with this, we demonstrated very high sCD163 levels in patients with acute liver failure from viral hepatitis.[18]

We undertook this study in a large cohort of treatment-naïve chronic HCV- and HBV-infected patients with histological data for disease activity and fibrosis. We assessed macrophage activation by means of sCD163 and related our findings to biochemical and histological parameters of liver disease. We hypothesized that sCD163 levels were elevated in chronic viral hepatitis; that the levels correlated with the severity of liver disease; and that sCD163 might serve as an independent marker of advanced disease. To do this, we examined sCD163 in HCV and HBV patients for its association with fibrosis stage. Finally, new sCD163-based fibrosis scores, CD163-HCV-FS and CD163-HBV-FS, were developed and compared to the aspartate aminotransferase (AST) to platelet ratio index (APRI) and the FIB-4 score.

Materials and Methods

We performed a cross-sectional study in 551 patients with chronic HCV and 203 patients with chronic HBV infection who were referred to the Storr Liver Unit, Westmead Hospital, Westmead, Australia, between July 1991 and August 2010 for evaluation of chronic viral hepatitis. The diagnosis of chronic HCV infection was confirmed by the presence of anti-HCV antibodies (Monolisa anti-HCV; Sanofi Diagnostics Pasteur, Marnes-la-Coquette, France) and viral RNA as detected by polymerase chain reaction (PCR) (Amplicor HCV; Roche Diagnostics, Branchburg, NJ). Hepatitis C virus genotyping was performed with a second-generation reverse-hybridization line probe assay (Inno-Lipa HCV II; Innogenetics, Zwijndrecht, Belgium). The diagnosis of chronic hepatitis B was confirmed by the presence of hepatitis B surface antigen in the blood for more than 6 months, hepatitis B core antibodies and HBV-DNA detection by signal amplification hybridization microplate assay (Digene HBV Test using Hybrid Capture 2, Digene) with a lower limit of detection of 0.5 pg/mL (1.42 × 105 virus copies/mL), or by real-time PCR. Patients dually infected with HCV and HBV (n = 9), or coinfected with human immunodeficiency virus (n = 1) were excluded. None of the patients had antiviral treatment prior to inclusion.

Liver biopsy was performed as part of the workup for assessment of severity of steatosis, inflammation, and fibrosis. The stained biopsies were examined by experienced pathologists and scored according to the Scheuer scoring system.[19] Steatosis was graded as described by Brunt et al.[20] All biopsies had a minimum of 11 portal tracts, and inadequate biopsies were excluded. Consequently, histological data were missing in 38 patients with HCV and three patients with HBV infection. At the time of the liver biopsy, basic demographic and clinical data were obtained, including gender, age, ethnicity, height, weight, and waist circumference. Alcohol consumption was assessed by two separate interviews with the patient and close family members. Body mass index (BMI) was calculated from height and weight. At the same time, a fasting blood sample was drawn and routine biochemical tests were performed as described below. Additional blood samples were taken and frozen at −80°C for future research. All patients signed an informed consent form in accordance with the Helsinki Declaration; the acquisition, storage, and use of the blood samples was approved by the Sydney West Area Health Service Ethics Committee.

Biochemical Analyses

The concentrations of alanine transaminase (ALT), aspartate transaminase (AST), alkaline phosphatase (ALP), gamma-glutamyltransferase (GGT), total bilirubin, prothrombin time, international normalized ratio (INR), fasting glucose and insulin, serum albumin, hemoglobin, platelets, leukocytes, triglycerides, and cholesterol and its components were determined prior to the liver biopsy using standard assays and methods. Homeostatic model assessment of insulin resistance (HOMA-IR) was calculated from fasting glucose and insulin.[21] The Model for Endstage Liver Disease (MELD) score was calculated using bilirubin, creatinine, and INR.[22] The AST to platelet ratio index (APRI) was calculated according to the established formula: (AST(IU/L)/upper normal limit) × 100/platelets(109/L).[23] The FIB-4 index was calculated as follows: age(years) × AST(IU/L)/(platelets(109/L) × (ALT(IU/L))½).[24]

The plasma concentration of sCD163 was determined in duplicate in samples that had been frozen at −80°C by an in-house sandwich enzyme-linked immunosorbent assay (ELISA) using a BEP-2000 ELISA-analyzer (Dade Behring) as previously described.[25] Control samples and serum standards with concentrations that ranged from 6.25 to 200 μg/L were included in each run. The interassay coefficient of variation in the current project (n = 20) was 3.5-6.0% at a level of 1.31 mg/L and 6-10% at a level of 3.59 mg/L. The limit of detection (lowest standard) was 6.25 μg/L. Soluble CD163 is resistant to repeated freezing and thawing.[25]

Statistical Methods

One-way analysis of variance (ANOVA) was used for the comparison of multiple groups and Student t test to study differences of normally distributed variables between the groups. For the nonnormally distributed data, Kruskal-Wallis and Mann-Whitney tests, respectively, were used. The relationships between sCD163 and other continuous variables were analyzed by simple linear regression (after transformation using natural logarithm [log]) or Spearman's rank correlation. Spearman's rank test was used to study associations between sCD163 and histological scores. To study the differences in sCD163 between patients with HBV and HCV infection with the same histological scores of fibrosis and inflammation, we used two-way ANOVA with post-hoc t tests. To assess differences in proportions, χ2-test or Fisher's exact test were used.

We performed multiple ordered logistic regression analysis with Scheuer fibrosis score as the dependent and sCD163 as the explanatory variable in two different models. These models provided odds ratios (OR) for a given fibrosis stage corresponding to specific increases in sCD163. We chose to present the results corresponding to a 25% increase in sCD163 based on the actual median difference of 28% in sCD163 levels between patients who differed in Scheuer fibrosis score by 1 (Fig. 1C). In Model 1, we aimed to determine whether sCD163 was associated with the fibrosis score directly or through its relationship with inflammation scores and known fibrosis risk factors. Age, gender, BMI, ethnicity, alcohol consumption, and presence of genotype 1 (HCV only) were identified as risk factors for liver fibrosis in chronic viral hepatitis and included in Model 1. Scheuer scores for lobular and portal inflammation were also included.

Figure 1.

Soluble CD163 in histological scores of inflammatory activity and fibrosis in patients with HCV and HBV infection. (A) sCD163 and Scheuer lobular inflammation score (0-4). HCV: Spearman's rho = 0.31, P < 0.001; HBV: rho = 0.31, P < 0.001 (B) sCD163 and Scheuer portal inflammation score (0-4). HCV: rho = 0.39, P < 0.001; HBV: rho = 0.42, P < 0.001 (C) sCD163 and Scheuer fibrosis score (0-4). HCV: rho = 0.45, P < 0.001; HBV: rho = 0.32, P < 0.001. Boxes represent interquartile ranges with medians; whiskers show adjacent values (the highest value lower or equal to: 75% quartile +1.5 × interquartile range; the lowest value higher or equal to: 25% quartile −1.5 × interquartile range). Punctured lines represent reference interval (0.89-3.95 mg/L). *0.01 < P < 0.05 between patients with HCV and HBV infection; **P < 0.01 between patients with HCV and HBV infection.

In Model 2, our goal was to investigate whether sCD163 was a marker of fibrosis when adjusted for demographic, clinical, and biochemical parameters shown to be associated with fibrosis in previous studies.[26] Thus, we included age, gender, BMI, ethnicity, genotype 1, alcohol consumption, albumin, platelets, ALT, AST, INR, and HOMA-IR in Model 2. All continuous variables were logarithmically transformed. To identify candidate variables for the new sCD163-based fibrosis scores (CD163-HCV-FS and CD163-HBV-FS) we performed backward elimination based on the likelihood ratio test with significance limit of 0.1 including all variables in Model 2. Then all possible subsets of these candidate variables were examined using nonparametric receiver operating characteristics (ROC) analyses for the presence of liver cirrhosis (defined by Scheuer fibrosis score of 4), advanced (F≥3) and significant fibrosis (F≥2). The subsets of variables providing the highest areas under the ROC-curves (AUROCs) in patients with HCV and HBV infection were chosen for the new scores. These new scores were compared to APRI and FIB-4 using the test of equality of ROC areas. Sensitivity, specificity, positive and negative predictive values (PPV and NPV) were determined for appropriate cutoff values of CD613-HCV-FS and CD163-HBV-FS, based on the ROC-curves.

All data are expressed as medians with interquartile ranges (IQR) or proportions, and P ≤ 0.05 were considered statistically significant. STATA v. 12.0 (StataCorp) was used for data analysis.

Results

Patient Characteristics

Basic demographic, clinical, and biochemical data for the HCV and HBV patients are presented in Table 1. The two groups did not differ significantly in terms of gender, the majority of the patients being male. Patients with HCV infection were slightly, but significantly older than HBV patients. BMI and HOMA-IR were higher in HCV patients; however, with no significant difference in the prevalence of diabetes between the groups. There was a significant difference in the consumption of alcohol between HCV and HBV patients (Table 1). We observed higher ALT, AST, and GGT levels in HCV patients compared to HBV. Albumin was slightly but significantly lower in patients with HCV infection. INR was significantly lower in HCV patients, although the two groups had the same median values. A trend toward lower platelets in HCV patients was observed.

Table 1. Demographic, Clinical, and Biochemical Parameters in Patients With HCV and HBV Infection
Parameter

HCVn = 551

HBVn = 203

 
  1. Parameters are presented as medians (interquartile range) for continuous variables, and as total number (%) for categorical variables. Units and normal ranges are in parentheses.

  2. MELD, Model for Endstage Liver Disease; BMI, body mass index; ALT, alanine transaminase; AST, aspartate transaminase; ALP, alkaline phosphatase; GGT, gamma-glutamyltransferase; INR, international normalized ratio; HOMA-IR, homeostatic model assessment of insulin resistance; LDL, low-density lipoprotein; HDL, high-density lipoprotein.

Age (years)43 (36-49)41 (33-49)P = 0.05
Gender (Male : Female)348 (63%) : 203 (37%)130 (64%) : 73 (36%)P = 0.82
Alcohol consumption (n)  P = 0.005
< 10 g/day446 (82.6%)152 (74.9%)
10-19 g/day41 (7.6%)28 (13.8%)
20-39 g/day34 (6.3%)17 (8.3%)
40-59 g/day10 (1.8%)6 (3.0%)
≥60 g/day9 (1.7%)0 (%)
Missing record110
Fibrosis stage (n)  P = 0.001
No/mild fibrosis (F0-1)265 (51.7%)139 (69.5%)
Significant fibrosis or higher (F≥2)248 (48.3%)61 (30.5%)
Advanced fibrosis or higher (F≥3)99 (19.3%)21 (10.5%)
Cirrhosis (F4)45 (8.8%)4 (2.0%)
Missing histology383
MELD7.5 (6.4-8.4)8.5 (6.4-8.8)P = 0.55
BMI (kg/m2)26 (23-30)24 (21-27)P < 0.001
Diabetes (n)25 (4.6%)4 (2.2%)P = 0.15
ALT (IU/L, males <70 IU/L; females <45 IU/L)82 (52-144)49 (29-78)P < 0.001
AST (IU/L, <45 IU/L)62 (43-100)46 (37-62)P < 0.001
Bilirubin (μmol/L, 5-25 μmol/L)11 (8-14)10 (8-14)P = 0.31
ALP (IU/L, 35-105 IU/L)77 (66-96)80 (67-99)P = 0.35
GGT (IU/L, < 115 IU/L)50 (29-93)26 (19-47)P < 0.001
INR1 (0.9-1)1 (1-1.1)P < 0.001
Albumin (g/L, 36-48 g/L)43 (41-45)45 (43-48)P < 0.001
HOMA-IR2.22 (1.43-3.78)1.43 (0.67-2.35)P < 0.001
Cholesterol (mmol/L, <5 mmol/L)4.5 (3.9-5.1)4.9 (4.4-5.6)P < 0.001
LDL (mmol/L, <3 mmol/L)2.6 (2-3.2)2.9 (2.4-3.5)P < 0.001
HDL (mmol/L, >1.2 mmol/L)1.3 (1-1.6)1.3 (1.2-1.7)P = 0.01
Triglycerides (mmol/L, <2 mmol/L)0.98 (0.75-1.36)0.98 (0.72-1.37)P = 0.70
Hemoglobin (g/L, males 134-169 g/L; females 118-153 g/L)151 (140-160)146 (137-157)P = 0.007
Leucocytes (x 109/L, 3.5-10 x 109/L)6.8 (5.6-8.4)5.4 (4.6-6.5)P < 0.001
Platelets (x 109/L, 165-400 x 109/L)225 (186-276)220 (190-257)P = 0.07

Soluble CD163 levels were significantly higher in HCV patients (3.6(2.5-5.4) mg/L) compared to those with HBV infection (2.4(1.8-3.6) mg/L), P < 0.001.

Histological Scores of Activity and Fibrosis in HCV and HBV Patients

The patients with HCV infection had more advanced disease with higher scores for Scheuer fibrosis (P = 0.001) compared to HBV patients (Table 1). None of the patients with biopsy-verified HCV or HBV cirrhosis had decompensated cirrhosis.

In addition, HCV patients had higher scores for Scheuer portal inflammation (P < 0.001) and steatosis (P < 0.001). There was no significant difference in the score for Scheuer lobular inflammation between the two groups (P = 0.45). The full distribution of histological scores in patients with HCV and HBV infection is presented in Supporting Table 1.

Associations Between sCD163 and Demographic, Clinical, and Biochemical Parameters in HCV and HBV Patients

In patients with HCV infection, males had higher sCD163 levels (3.8(2.7-5.8) versus 3.4(2.2-4.7) mg/L; P = 0.009), while there was no significant difference in HBV patients (P = 0.92). Associations between sCD163 and demographic, clinical, and biochemical parameters are presented in Table 2. In HCV patients, sCD163 showed significant associations with more parameters than in patients with HBV infection, possibly reflecting the higher number of patients in this group. In both groups we observed significant associations with ALT, AST, and INR, and inverse associations with platelets and albumin. There was a significant association between sCD163 and HOMA-IR in patients with HCV infection, but not in HBV patients. Soluble CD163 did not show significant associations with HCV viral load or HBV DNA titer counts; it did not differ significantly between various HCV genotypes, while data on HBV genotypes were unavailable.

Table 2. Univariate Associations Between sCD163 and Demographic, Clinical, and Biochemical Parameters in Patients With HCV and HBV Infection
ParameterHCVHBV
rhoPrhoP
  1. Associations were statistically tested by Spearman's rank correlation.

  2. BMI, body mass index; ALT, alanine transaminase; AST, aspartate transaminase; INR, international normalized ratio; HOMA-IR, homeostatic model assessment of insulin resistance.

Age (years)0.24<0.001−0.010.93
BMI (kg/m2)0.22<0.0010.070.34
Waist circumference (cm)0.26<0.0010.100.15
ALT (IU/L)0.48<0.0010.53<0.001
AST (IU/L)0.61<0.0010.56<0.001
Albumin (g/L)−0.20<0.001−0.35<0.001
Platelets (x 109/L)−0.37<0.001−0.160.03
Hemoglobin (g/L)0.090.0460.040.57
INR0.21<0.0010.30<0.001
HOMA-IR0.34<0.0010.070.235
Cholesterol (mmol/L)−0.17<0.001−0.050.5
Triglycerides (mmol/L)0.16<0.0010.0030.97

Soluble sCD163 showed no significant association with alcohol consumption in patients with HCV or HBV infection.

Associations Between sCD163 and Histological Scores of Activity and Fibrosis in HCV and HBV Patients

Soluble CD163 increased in association with rising scores for histological fibrosis stage and inflammatory activity in patients with HCV and HBV infection (Fig. 1 A-C). Median sCD163 levels were higher in HCV compared to HBV patients with the same scores for inflammation and fibrosis, reaching statistical significance in a number of cases.

There was a weak association between sCD163 and steatosis score in patients with HCV (rho = 0.21; P < 0.001), but not HBV infection (P = 0.41).

Ordered Logistic Regression Analysis for Association Between sCD163 and Fibrosis in HCV and HBV Patients

We used ordered logistic regression analysis with Scheuer fibrosis score as the dependent and sCD163 as the explanatory variable in a univariate and two multiple models. In the univariate model, the OR for HCV patients was 1.49 (95% confidence interval [CI]: 1.38-1.61), P < 0.001. Thus, for two random HCV patients from the cohort, if one of them had 25% higher sCD163 compared to the other, this patient had 49% greater odds of, e.g., presence of fibrosis (F≥1) than the other patient. In patients with HBV infection, this OR was 1.32 (95% CI: 1.17-1.49), P < 0.001. It is a property of ordered logistic regression that this increase in odds is the same for significant fibrosis (F≥2), advanced fibrosis (F≥3), or liver cirrhosis (F = 4).

In multiple regression models we adjusted for risk factors and demographic, clinical, and biochemical parameters. Applying multiple models to the two hypothetical patients described above, the resulting ORs estimate the odds of, e.g., the presence of fibrosis (F≥1) for the patient with 25% higher sCD163, assuming that the two patients were completely alike in terms of the risk factors and parameters we adjusted for. In patients with HCV infection, sCD163 showed a significant independent association with fibrosis in both multiple models, whereas in HBV patients sCD163 was significantly associated with fibrosis in Model 2 (adjusted for the fibrosis risk factors and other parameters known to be associated with fibrosis from previous studies), but not in Model 1 (adjusted for the fibrosis risk factors and Scheuer inflammation scores) (Table 3).

Table 3. Univariate and Multiple Ordered Logistic Regression Models With sCD163 (Logarithmically Transformed) as the Explanatory Variable for Scheuer Fibrosis Score
 OR95% CIP
  1. Unadjusted and adjusted odds ratios (OR) with 95% confidence intervals (CI) are presented for each 25% increase in sCD163.

  2. Model 1. Age, gender, body mass index, ethnicity, alcohol consumption, presence of genotype 1 (hepatitis C only) and Scheuer lobular and portal inflammation scores included in the model.

  3. Model 2. Age, gender, body mass index, ethnicity, alcohol consumption, presence of genotype 1 (hepatitis C only), albumin, platelets, alanine transaminase, aspartate transaminase, international normalized ratio and homeostatic model assessment of insulin resistance included in the model.

Patients with hepatitis C infection
Unadjusted1.491.38-1.61<0.001
Model 11.281.17-1.41<0.001
Model 21.191.06-1.350.005
Patients with hepatitis B infection
Unadjusted1.321.17-1.49<0.001
Model 11.110.96-1.290.166
Model 21.311.05-1.640.018

Development of an sCD163-Based Predictive Fibrosis Score (CD163-HCV-FS) and Comparison to the APRI and FIB-4 Scores in HCV Patients

As described above, we performed backward elimination based on the likelihood ratio test with significance limit of 0.1 for the multiple ordered logistic regression analysis including parameters shown to be associated with fibrosis in previous studies (Model 2). In this analysis, the following variables were significantly associated with Scheuer fibrosis score: sCD163 (P = 0.005), age (P < 0.001), AST (P < 0.001), platelets (P = 0.007), HOMA-IR (P < 0.001), and INR (P < 0.001). We then performed new multiple ordered logistic regression analyses with Scheuer fibrosis score as the dependent variable and all possible subsets of the candidate variables above as the explanatory, and used the coefficients (β) from the regression equations to compute and examine all possible predictive models (data not shown). The two candidate models with the highest AUROCs are presented in Supporting Table 2. The presence of significant fibrosis (F ≥ 2) is usually used as a determinant for initiating antiviral therapy, and as the predictive model including sCD163, age, AST, HOMA-IR, and platelets had the highest AUROC for significant fibrosis, we chose this model as the novel sCD163-based fibrosis score in HCV patients (CD163-HCV-FS). We used simple coefficients that were close to the actual coefficients from the regression equation (Table 4); this simplified model provided exactly the same results as the one with the actual coefficients.

display math

Next, we compared the AUROCs of CD163-HCV-FS with those of APRI and FIB-4, presented in Table 5A. CD163-HCV-FS was significantly superior to both APRI and FIB-4 for the prediction of significant fibrosis and to APRI for advanced fibrosis and cirrhosis.

Table 4. Multiple Ordered Logistic Regression Analysis With Scheuer Fibrosis Score as the Dependent Variable in Patients With HCV and HBV Infection
VariableβP
  1. log, natural logarithm; sCD163, soluble CD163; AST, aspartate transaminase; HOMA-IR, homeostatic model assessment of insulin resistance; BMI, body mass index.

Patients with hepatitis C infection
log sCD1630.560.026
log Age1.67<0.001
log AST1.12<0.001
log Platelets−1.60<0.001
log HOMA-IR0.370.004
Patients with hepatitis B infection
log sCD1631.64<0.001
log BMI−1.870.063
Male gender0.770.030
Table 5. Areas Under the Receiver Operating Characteristics Curve (AUROCs) for Significant Fibrosis, Advanced Fibrosis, and Cirrhosis for the AST to Platelet Ratio Index (APRI), FIB-4, and the sCD163-Based Fibrosis Scores (CD163-HCV-FS and CD163-HBV-FS) in Patients With Chronic Viral Hepatitis
A   
 APRIFIB-4CD163-HCV-FS
F ≥ 20.74 (0.69-0.79)0.75 (0.70-0.79)0.79 (0.74-0.83)a, b
F ≥ 30.81 (0.76-0.86)0.82 (0.78-0.87)0.86 (0.82-0.90)a
F = 40.85 (0.79-0.91)0.89 (0.85-0.93)0.90 (0.85-0.94)a
B   
 APRIFIB-4CD163-HBV-FS
  1. A: Patients with HCV infection; B: Patients with HBV infection.

  2. Data are presented as AUROCs (95% confidence interval).

  3. a

    P < 0.01 compared to APRI;

  4. b

    P = 0.014 compared to FIB-4.

F ≥ F20.73 (0.65-0.81)0.67 (0.58-0.76)0.71 (0.62-0.79)
F ≥ F30.76 (0.63-0.88)0.72 (0.59-0.85)0.77 (0.67-0.88)
F = 40.79 (0.72-0.86)0.75 (0.39-1.0)0.82 (0.64-1.0)

We identified cutoff values for CD163-HCV-FS that provided the best discrimination for the presence or absence of significant fibrosis due to its importance for antiviral therapy. The cutoff for the absence of significant fibrosis was defined at CD163-HCV-FS = 1.55 (Marked 1 on Fig. 2A). The sensitivity was 90%, specificity 42%, PPV 59%, and NPV 82%. The cutoff for the presence of significant fibrosis was at CD163-HCV-FS = 3.50 (Marked 2 on Fig. 2A); the sensitivity was 34%, specificity 93%, PPV 82%, and NPV 60%, with the total prevalence of significant fibrosis in HCV patients at 48.3%.

Figure 2.

Receiver operating characteristics (ROC) analysis showing the predictive value of the sCD163-based fibrosis scores (CD163-HCV-FS and CD163-HBV-FS) for significant fibrosis (F≥2) in patients with HCV and HBV infection. (A) ROC curve for CD163-HCV-FS. Circles mark cutoff values of CD163-HCV-FS (1: cutoff at 1.55; 2: cutoff at 3.50). AUROC = 0.79 (95% CI: 0.74-0.83). (B) ROC curve for CD163-HBV-FS. Circles mark cutoff values of CD163-HBV-FS (1: cutoff at 5.0; 2: cutoff at 6.50). AUROC = 0.71 (95% CI: 0.62-0.79).

Development of an sCD163-Based Predictive Fibrosis Score (CD163-HBV-FS) and Comparison to the APRI and FIB-4 Scores in HBV Patients

We used the same approach for the development of the predictive score in HBV patients as described above. We applied backward elimination based on a likelihood ratio test with significance limit of 0.1 and identified sCD163, gender, and BMI as candidate variables (Table 4). By using simplified coefficients from the regression equation (using the actual coefficients did not alter the results), we computed a fibrosis score for HBV patients. However, this score had all negative values, and we modified it by adding 10. CD163-HBV-FS was calculated as follows:

display math

The new score had higher AUROCs than APRI and FIB-4 in the majority of comparisons, but the difference did not reach statistical significance (Table 5B). Analogous to HCV patients, we identified cutoff values for CD163-HBV-FS for the presence or absence of significant fibrosis, based on the ROC-curve (Fig. 2B). The cutoff for the absence of significant fibrosis was at CD163-HBV-FS = 5.0 (Marked 1 on Fig. 2B). The sensitivity was 89%, specificity 37%, PPV 38%, and NPV 88%. The cutoff for the presence of significant fibrosis was at CD163-HBV-FS = 6.50 (Marked 2 on Fig. 2B); the sensitivity was 30%, specificity 89%, PPV 54%, and NPV 74%, with the total prevalence of significant fibrosis in HBV patients at 30.5%.

Discussion

The important role of macrophages in inflammation and fibrosis in chronic liver diseases has increased the interest for systemic markers of macrophage activation. In this study, we measured levels of sCD163, a specific macrophage marker, in patients with chronic HCV and HBV infection. The main finding of the study was the progressive increase in sCD163 in association with the severity of liver disease. Moreover, the association between sCD163 and fibrosis was highly significant when adjusted for multiple biochemical and clinical parameters. Further, we computed new sCD163-based fibrosis scores for patients with HCV (CD163-HCV-FS) and HBV infection (CD163-HBV-FS); in HCV patients, the novel score was significantly superior to APRI for the prediction of all fibrosis stages and to FIB-4 for significant fibrosis.

The main strength of our study was the large number of well-classified patients with chronic viral hepatitis. All of these patients had undergone liver biopsy, which remains the gold standard for staging and grading of chronic viral hepatitis. The most significant limitation was the cross-sectional design of the study, which could challenge the interpretation of different associations regarding cause and effect. Another possible limitation was the lack of a validation cohort and a control group. However, we have previously measured sCD163 in a large cohort of healthy individuals with the same assay, which provided us with a reference interval (0.89-3.95 mg/L).[27]

Hiraoka et al.[16, 17] found increased surface-bound CD163 expression and soluble CD163 plasma levels in acute and chronic viral hepatitis. The authors also demonstrated that the liver cells expressing CD163 were Kupffer cells. Another study showed increased CD163 mRNA levels in the livers of patients with chronic HCV infection.[28] These studies support our findings of elevated sCD163 in chronic viral hepatitis.

Focusing on fibrosis as the most significant determinant of liver disease outcomes in chronic viral hepatitis, we showed that elevated sCD163 is associated with increased odds of fibrosis in an ordered logistic regression model. In addition, we demonstrated an association between sCD163 and fibrosis even after adjustment for numerous clinical and biochemical parameters and risk factors, including parameters often used as surrogate markers of disease severity, such as platelets, albumin, ALT, and AST. Thus, we provide strong evidence for macrophage activation in chronic viral hepatitis, with more profound activation with increasing disease severity.

In viral hepatitis, macrophages play an important role as the first line of defense, initiating an immunological response and interlinking the innate and adaptive immune systems for effective clearance of the virus with minimal hepatocyte damage.[29] However, viral proteins can interfere with this process, directly modulating the inflammatory response. Hepatitis C core and nonstructural proteins are capable of triggering TLR2-mediated proinflammatory macrophage activation, causing inflammation and liver damage.[4, 30, 31] Likewise, hepatitis B virus core antigen activates macrophages through TLR2 and shifts the macrophage phenotype towards secretion of proinflammatory cytokines.[32, 33] In line with this, an important role of macrophages in the process of fibrogenesis has been suggested.[1, 2]

On the other hand, several studies suggest a hypothesis of macrophage activation by circulating endotoxin, which is detected even in mild disease, but becomes more prominent with worsening fibrosis in chronic viral hepatitis.[28, 34-36] Sandler et al.[6] found high levels of endotoxin and evidence of endotoxin-mediated macrophage activation in chronically HBV- and HCV-infected patients with severe fibrosis, while patients without fibrosis had significantly lower degrees of endotoxemia and macrophage activation. In our cohort, patients with severe fibrosis had the highest levels of sCD163; interestingly, the patients who had milder disease still had elevated sCD163, suggesting macrophage activation even in moderate disease, albeit to a lower extent.

Supporting this hypothesis, activation of the TLR2-pathway by HCV-associated proteins seems to enhance susceptibility to endotoxin in Kupffer cells by interfering with TLR4-signaling.[3] Dolganiuc et al.[28] demonstrated preactivation of macrophages by HCV, which led to persistent and potent endotoxin-induced macrophage activation and TNF-α production. This enhanced susceptibility of macrophages to endotoxin might be responsible for the demonstrated high degree of macrophage activation, which was comparable with levels previously found in decompensated cirrhosis,[14, 18] despite the absence of clinical decompensation in our patients. As this mechanism has not been described in HBV, it may explain our finding of higher levels of sCD163 in HCV compared to HBV patients with the same grade and stage of disease. Moreover, we found that sCD163 was not significantly associated with fibrosis when adjusted for inflammatory scores in patients with HBV infection, which might also be an indication of less profound endotoxin-induced macrophage activation.

Owing to the cross-sectional design of this study, we can only speculate whether macrophage activation in our cohort as assessed by sCD163 levels is a primary event contributing to fibrogenesis, or a result of TLR4-mediated response to endotoxemia secondary to liver fibrosis and/or cirrhosis. Prospective studies of sCD163 in chronic viral hepatitis during antiviral treatment would help determine the relationships of cause and effect and elucidate therapeutic possibilities for intervention targeting Kupffer cells.

We developed new sCD163-based fibrosis scores for HCV (CD163-HCV-FS) and HBV (CD163-HBV-FS) patients. CD163-HCV-FS was superior to the existing scores APRI and FIB-4, mainly in predicting significant fibrosis. CD163-HBV-FS had higher AUROCs than APRI and FIB-4 in the majority of comparisons, but the differences did not reach statistical significance, probably due to the lower number of patients in this group. CD163-HBV-FS showed less accurate fibrosis prediction than CD163-HCV-FS, and a weaker association was noted between sCD163 and fibrosis in HBV patients. As discussed above, this may suggest that macrophage activation and function in HBV infection might differ from that in HCV infection, but mechanistic studies will be required to clarify the issue.

The new scores are simple and include parameters that are readily available; furthermore, they present the first models based on a marker of disease pathophysiology, i.e., macrophage activation. The scores need to be validated in independent cohorts, but are a promising tool for the noninvasive determination of fibrosis in chronic viral hepatitis.

In conclusion, we demonstrated that Kupffer cells are activated in chronic viral hepatitis, and that the degree of macrophage activation as assessed by sCD163 levels increases with disease severity and is independently associated with histological fibrosis stage. Moreover, a sCD163-based fibrosis score was superior to the existing scores APRI and FIB-4 for significant fibrosis in HCV patients, presenting a promising tool for the noninvasive diagnosis of fibrosis in these patients.

Acknowledgment

We thank Kirsten Bank Petersen, Department of Clinical Biochemistry, Aarhus University Hospital, for excellent technical assistance.

Ancillary