Abstract
Background
Accurate noninvasive biomarkers of fibrotic progression are important for hepatitis C virus (HCV) management, but commonly used modalities may have decreased efficacy in human immunodeficiency virus (HIV)/HCV-coinfected persons. The enhanced liver fibrosis (ELF) index is a highly sensitive noninvasive marker of hepatic fibrosis that has had limited assessment in the HIV/HCV population. We compared ELF index performance to FIB4 and aspartate to platelet ratio index (APRI) at different stages of liver fibrosis as determined by liver histology, and validated the efficacy of the three noninvasive biomarkers in HIV/HCV-coinfected versus HCV-monoinfected.
Methods
The ELF index was determined in 147 HIV/HCV-coinfected and 98 HCV-monoinfected persons using commercial ELISA assays for the component elements of the index. Area under the receiver-operator curve was used to validate ELF and to compare its performance to liver histology as well as to other noninvasive biomarkers of liver fibrosis, FIB4, and APRI.
Results
The ELF index increased with histological stage of liver fibrosis and exhibited a linear relationship with Metavir score in all subjects. ELF performance was comparable between HIV/HCV and HCV with advanced liver fibrosis/cirrhosis. In the HIV/HCV cohort ELF cutoffs of 8.45 and 9.23 predicted mild and moderate fibrosis with 85% sensitivity, whereas the ELF cutoff of 9.8 had the highest specificity for advanced fibrosis and the cutoff of 10.4 was 99% specific for cirrhosis. ELF performance was superior to FIB4 and APRI in all subjects regardless of HIV status.
Conclusions
ELF index demonstrated excellent characteristics toward accurate prediction of liver fibrosis and cirrhosis with superior performance to APRI and FIB4 in HIV/HCV coinfection. Applying this noninvasive biomarker index for diagnosis of liver fibrosis and progression in HIV/HCV is warranted.
Keywords: ELF index, FIB4, APRI, HIV/HCV, liver fibrosis
ELF-index performance in Human Immunodeficiency Virus (HIV)/hepatitis C virus liver fibrosis/cirrhosis prediction is accurate and superior to other noninvasive biomarkers of liver fibrosis, FIB4, and APRI. The ELF index is sensitive and can be used in the clinic for liver fibrosis diagnosis and progression in persons living with HIV.
Human immunodeficiency virus (HIV) coinfection with hepatitis C virus (HCV) is associated with a higher rate of progression toward hepatic fibrosis [1] and advanced liver disease [2] as compared to HCV monoinfection [3]. Hence, HCV-associated liver disease is a major source of morbidity and mortality in HIV-positive patients [4, 5] and accurate diagnosis of liver fibrosis is important in the treatment and management of HIV/HCV coinfection. The gold standard method for diagnosis of liver fibrosis is liver biopsy with pathologist-based tissue histology classification. However, noninvasive modalities have replaced liver biopsy in clinical practice due to known procedural risks and cost. Despite the fact that these noninvasive biomarkers have been reported to perform suboptimally in the HIV/HCV population, they are used in lieu of other options [6, 7].
The enhanced liver fibrosis (ELF) index is a direct noninvasive blood test that combines the expression of extracellular matrix proteins: Hyaluronic acid (HA), tissue inhibitor of matrix metalloproteinase (TIMP)-1, and procollagen III amino-terminal pro-peptide (PIIINP) in a mathematical algorithm. Rosenberg and coauthors were the first to describe the ELF index score that can predict liver fibrosis and cirrhosis [8]. Later, the algorithm was effectively simplified to eliminate the influence of age on the ELF index measurement [9]. The ELF index has been since validated in patients with hepatitis C, hepatitis B, nonalcoholic steatohepatitis (NASH), and other chronic liver diseases [9–17]. However, limited studies have evaluated the noninvasive ELF index in persons living with HIV (PLWH) using liver histology. In a large cohort, we compared the performance of ELF index to the fibrosis index based on 4 factors (FIB4) and the aspartate to platelet ratio index (APRI) at different stages of liver fibrosis as determined by liver histology and sought to validate the efficacy of the 3 noninvasive biomarkers in HIV/HCV coinfection as compared to HCV monoinfection.
METHODS
Study Design
This study is a retrospective multicenter study of repository specimens and data. Samples were derived from research study participants and patients receiving clinical care at the University of Cincinnati (1995 to 2012) and 154 participants enrolled in the HEPRES/HEPRO cohorts from the University of Maryland/National Institutes of Health (2009 to 2017) repository collections. All subjects signed a written consent for use of samples and clinical data in accordance with Institutional Review Board guidance and approval. Patient samples were included if they were infected with HIV and/or HCV and had a biopsy report with serum samples collected at the time of the biopsy and stored at −80˚C. Patient samples were excluded if the record showed other liver pathology or if the biopsy was inadequate.
Liver Histology
Liver biopsies were obtained by a hepatologist or interventional radiologist. An adequate biopsy specimen was typically defined as one containing at least 11 portal areas and not fragmented. Formalin-fixed paraffin-embedded liver biopsy sections were stained with hematoxylin and eosin, Masson trichrome, and for reticulin. Interpretation of the biopsies was performed by independent highly trained hepatopathologists. Scoring was performed by either the Ishak scoring system or the Metavir scoring system. For the purpose of this analysis, all Ishak stage 4–6 cirrhosis were reclassified as Metavir 4. For clinical cases in our repository that did not have formal scoring performed, descriptive text was converted to the Metavir score [18]. For these specimens fibrosis was scored as follows: none = F0; mild portal fibrosis = F1; periportal fibrosis = F2; bridging fibrosis = F3; and cirrhosis = F4.
Enhanced Liver Fibrosis Index Measurement
The ELF index was validated using serum from patients with HCV mono (n = 98) and HIV/HCV coinfection (n = 147), who underwent liver biopsy; representing all Metavir fibrosis stages (F0–F4). Measures of hepatic fibrosis were determined by calculation of the ELF index using the following simplified equation: ELF index score = [−7.412 + [(ln [HA] × 0.681) + (ln [PIIINP] × 0.775) + (ln [TIMP1] × 0.494)] + 10] [9]. ELF index components were quantified using commercial assays: Human PIIINP ELISA Kit (Cloud-Clone Corp), Human HA Quantikine ELISA kit (R&D Systems), and Human TIMP-1 ELISA (Abcam). Data were measured in duplicate, and averages were calculated.
FIB4 and APRI Calculation
FIB4 and APRI were calculated in a subset of subjects (n = 206) where data were available. The APRI scores were calculated according to the equation: [(AST/ULN [upper limit of normal] of AST) ×100] / platelets (109/L)] [19], and FIB4 was determined according to the equation [(age × AST) / (platelets [109/L] × ALT [1/2])] [20].
Statistical Analysis
Data were analyzed using GraphPad Prism version 6.00 for Windows, (GraphPad Software, La Jolla, California, USA; www.graphpad.com) and MedCalc Statistical Software version 18.11 (MedCalc Software bvba, Ostend, Belgium; https://www.medcalc.org; 2018). A P value of ≤.05 with a 2-tailed hypothesis was used. The Kruskal-Wallis analysis of variance was used to compare the ELF index outcomes according to the histological Metavir scores of fibrosis and the Spearman rank correlation to assess the relationship between ELF index scores and histological stage of fibrosis. Tukey multiple comparisons tested for significant differences and confidence intervals between group means (F0–F4) of ELF index.
The performance (diagnostic accuracy) of the ELF index in the identification of fibrosis stage defined by the Metavir staging of liver histology was evaluated by calculating the area under receiver operating characteristic curve (AUROC). Different levels of fibrosis were assessed as follows: mild fibrosis (F ≥1), moderate fibrosis (F ≥2), advanced fibrosis (F ≥3), and cirrhosis (F = 4). Sensitivity (proportion of true positives, correctly identified by the ELF index), specificity (proportion of true negatives correctly identified by the ELF index), positive predictive value (PPV) (proportion of individuals with true liver fibrosis by ELF scores) and negative predictive value (NPV) (proportion of individuals who truly do not have liver fibrosis by ELF scores) were derived for various cutoff points. We used previously published ELF cutoffs as a reference [21, 22] for comparison. Additionally, we derived population-specific cutoff values for each liver fibrosis stage within the HIV/HCV and HCV patient subsets. Statistically significant differences between the AUROC curves of ELF, FIB4, and APRI tests were calculated using the methods implemented in the DeLong test at different Metavir scores. We used the AUROC curve to examine the difference in ELF index and other noninvasive biomarkers performance between HIV/HCV coinfection and HCV monoinfection.
RESULTS
Population Characteristics
We evaluated the association of ELF index with liver disease histological biopsy scores in a cohort of 147 subjects coinfected with HIV and HCV, and 98 subjects monoinfected with HCV. In total, 71% were male and 29% were white with an average age of 51 years. Forty-three (18%) subjects had no histological evidence of liver fibrosis (F0). In total, 136 (56%) had mild to moderate fibrosis (F1–2), and 66 (30%) patients had advanced fibrosis/cirrhosis (F3–4) (Table 1).
Table 1.
Baseline Characteristics of Patients
| HIV/HCV | HCV | Total | P | |
|---|---|---|---|---|
| N | 147 | 98 | 245 | |
| Male sex, n (%) | 116 (79) | 59 (60) | 175 (71) | .0390b |
| White race, n (%) | 43 (29) | 28 (29) | 71 (29) | nsb |
| Age (year)a | 49 (23–75) | 54 (21–71) | 51 (21–75) | <.0001c |
| HIV viral load, copies/mLa | 631.7 (39–37717) | NA | … | |
| HCV viral load, IU/mLa | 5.8 (0–7.7) | 5.2 (1.04–7.93) | … | .0322c |
| AST level, IU/mLa | 65 (12–367) | 46 (12–217) | … | nsc |
| ALT level, IU/mLa | 67 (10–323) | 53 (10–208) | … | nsc |
| CD4 count, cells/µLa | 628 (192–1453) | 981 (431–1866) | … | <.0001c |
| ELF indexa | 9.2 (6.9–12.3) | 9.3 (7.9–11.1) | … | nsc |
| Metavir stage | ||||
| F0 | 25 | 18 | 43 | |
| F1 | 69 | 49 | 118 | |
| F2 | 18 | 0 | 18 | |
| F3 | 19 | 17 | 36 | |
| F4 | 16 | 14 | 30 |
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotranferase; ELF, enhanced liver fibrosis; HCV, hepatitis C virus; HIV, human immunodeficiency virus; ns, not significant.
aAverage (range).
bχ 2 test.
cUnpaired t-test.
Comparison of ELF Index Performance to Metavir Histological Stages
In all patients (n = 245) the ELF index scores displayed a near linear relationship with significant correlation to Metavir fibrosis stages (r = 0.8018, P < .0001, 95% confidence interval [CI] .70–.87). Group means of the ELF index at different histological stages of fibrosis (F0–F4) differed significantly (P ≤ .05) (Figure 1). A linear regression analysis curve fit displayed an average 0.5 unit increase in ELF per change of one Metavir histological stage of fibrosis from F0 to F3. This change increased to 0.8 ELF units as the Metavir stages changed from F3 to F4. These results validate the ELF index as a prognostic marker of liver fibrosis progression (Figure 2). We then evaluated the performance of the ELF index in all patients by AUROC analysis. The ELF index showed excellent prediction of moderate liver fibrosis (≥F2) (AUROC = 0.94; 95% CI .90 to .97), advanced liver fibrosis (≥F3) (AUROC = 0.97; 95% CI .95 to .99), and cirrhosis (F4) (AUROC = 0.98; 95% CI = .92 to .99), but slightly subordinate performance in mild fibrosis (AUROC = 0.86, 95% CI = .81 to .89).
Figure 1.
Significant difference between group means (F0–F4) of ELF index with 95% CI using Tukey multiple comparisons test. Abbreviations: CI, confidence interval; ELF, enhanced liver fibrosis.
Figure 2.
Linear regression analysis of ELF index, FIB4, and APRI against histological Metavir fibrosis stages in all patients. A, Superior correlation of increased Metavir fibrosis stages with ELF index (r = 0.80) than (B) FIB4 (r = 0.50) and (C) APRI (r = 0.44). Abbreviations: APRI, aspartate to platelet ratio index; ELF, enhanced liver fibrosis.
Clinical Diagnostic Performance of ELF Index for Liver Fibrosis/Cirrhosis
To detect lower stages of liver fibrosis (any to moderate) we considered 85% sensitivity for the ELF index cutoff threshold with the limitation of missing the diagnosis in 15% (false negative) of patients to be a clinically relevant measure. Accordingly, the ELF index cutoff threshold for mild fibrosis was 8.54 when all patients were considered. Among patients with HIV/HCV a cutoff threshold of 8.45 versus 8.72 in patients with HCV infection was optimal. In moderate fibrosis, the ELF index cutoff of 9.4 had the sensitivity to detect the disease in 86% of all patients. In HIV/HCV coinfected persons, a cutoff of 9.23 could detect moderate fibrosis in 85% of patients. This cutoff could not be evaluated in those with HCV alone because of insufficient representation. The ELF cutoff of 9.8 had the highest specificity (~ 98%) for detecting advanced fibrosis in all patients as well as in both HCV monoinfected and HIV/HCV coinfected groups. However, the sensitivity of this cutoff was only ~72%. The ELF cutoff threshold of 9.6 had a better sensitivity and specificity performance for detecting advanced fibrosis in all patients and within the subgroups. In the cirrhotic group, the cutoff of ≥9.8 had a high sensitivity (90%) and specificity (89%) but with lower PPV (around 50%), which indicated that 50% of cirrhotic patients could be misdiagnosed. To limit the probability of misdiagnosing patients with cirrhosis we increased specificity to 99%. At this level of specificity the cutoff threshold of 10.4 in all patients had a sensitivity of detecting cirrhosis in 63–64% of patients with high NPV (~95%) and PPV (~90%). The higher cutoff of >11.34 had 100% specificity for detection of cirrhosis but with the chance of missing the diagnosis in 70% of patients with cirrhosis. The cutoff of <7.7 could exclude fibrosis in all patients but with very low specificity (Tables 2 and 3). These results were comparable to the previously published ELF index cutoffs [21, 22].
Table 2.
ELF Index Performance for Identification of Liver Fibrosis/Cirrhosis in All Subjects
| ELF Index Cutoff | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | |
|---|---|---|---|---|---|
| Any/mild ≥F1 AUC = 0.86 | 7.76 | 100 | 9 | 84 | 80 |
| 7.91 | 99 | 19 | 85 | 73 | |
| 8.54 | 86 | 63 | 92 | 49 | |
| 8.89 | 72 | 88 | 97 | 40 | |
| Moderate ≥F2 AUC = 0.94 | 9.29 | 91 | 84 | 75 | 94 |
| 9.4 | 86 | 88 | 78 | 92 | |
| 9.46 | 82 | 89 | 80 | 91 | |
| 9.63 | 74 | 97 | 93 | 88 | |
| 9.78 | 61 | 99 | 98 | 83 | |
| Advanced ≥F3 AUC = 0.97 | 9.49 | 96 | 91 | 79 | 98 |
| 9.62 | 91 | 95 | 87 | 97 | |
| 9.63 | 89 | 96 | 88 | 96 | |
| 9.8 | 73 | 98 | 94 | 91 | |
| Cirrhosis F4 AUC = 0.98 | 9.8 | 90 | 89 | 53 | 99 |
| 10.12 | 73 | 95 | 69 | 96 | |
| 10.45 | 63 | 99 | 86 | 95 | |
| 11.34 | 23 | 100 | 100 | 90 |
Abbreviations: AUC, area under the curve; ELF, enhanced liver fibrosis; NPV, negative predictive value; PPV, positive predictive value.
Table 3.
ELF Index Performance for Identification of Liver Fibrosis/Cirrhosis in HIV/HCV Versus HCV Patients
| HIV/HCV | HCV | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ELF Index Cutoff | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | ELF Index Cutoff | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | |
| Any/mild fibrosis ≥F1 | 7.42 | 100 | 12 | 85 | 100 | 7.91 | 100 | 11 | 83 | 100 |
| 7.86 | 98 | 24 | 86 | 67 | 8.3 | 99 | 39 | 88 | 88 | |
| 8.45 | 85 | 60 | 91 | 46 | 8.72 | 85 | 67 | 92 | 50 | |
| 8.52 | 81 | 68 | 93 | 43 | 8.85 | 81 | 89 | 97 | 52 | |
| 8.85 | 68 | 84 | 95 | 35 | 8.91 | 76 | 89 | 97 | 46 | |
| 8.89 | 66 | 88 | 96 | 34 | 9.05 | 67 | 100 | 100 | 41 | |
| 9.09 | 56 | 100 | 100 | 32 | … | … | … | … | … | |
| Advanced fibrosis ≥F3 | 9.48 | 97 | 89 | 74 | 99 | 9.49 | 100 | 90 | 81 | 100 |
| 9.62 | 89 | 96 | 86 | 96 | 9.62 | 97 | 94 | 88 | 98 | |
| 9.81 | 71 | 98 | 93 | 92 | 9.8 | 73 | 99 | 96 | 89 | |
| Cirrhosis F4 | 9.49 | 100 | 79 | 37 | 100 | 9.69 | 100 | 82 | 48 | 100 |
| 9.81 | 81 | 89 | 48 | 98 | 9.8 | 93 | 88 | 57 | 99 | |
| 10.18 | 69 | 98 | 79 | 96 | 10.17 | 71 | 98 | 83 | 95 | |
| 10.45 | 63 | 99 | 80 | 94 | 10.41 | 64 | 99 | 90 | 94 | |
| 11.34 | 38 | 100 | 100 | 93 | 11.14 | 7 | 100 | 100 | 87 |
Abbreviations: ELF, enhanced liver fibrosis; HCV, hepatitis C virus; HIV, human immunodeficiency virus; NPV: negative predictive value; PPV, positive predictive value.
Comparison of ELF Performance to FIB4 and APRI in All Patients
FIB4 and APRI were calculated in all subjects where platelet count was available (n = 206). However, this limitation did not affect the demographic distribution of the study subjects. All 3 markers demonstrated a positive relation to Metavir fibrosis stage with higher association for ELF index (r = 0.80, P < .001) than FIB4 (r = 0.5, P < .001) and APRI (r = 0.44, P < .001). The ELF index had the best linear relationship with Metavir histological stages as compared to FIB4 and APRI (Figure 2). Figure 3 and Table 4 demonstrate that ELF index was a better biomarker than FIB4 and APRI in identifying any liver histological fibrosis or cirrhosis as determined by different Metavir histological stages. Pairwise comparison of AUROC curves displayed superior performance of ELF index versus FIB4 and APRI at mild fibrosis ≥F1 (AUROC = 0.85 vs 0.73, P = .010 and 0.74, P = .030), moderate fibrosis ≥F2 (AUROC = 0.91 vs 0.77, P = .0001 and 0.75, P = .0001), advanced fibrosis ≥F3 (AUROC = 0.976 vs 0.77, P < .0001 and 0.73, P < .0001), and cirrhosis F4 (AUROC = 0.979 vs 0.87, P = .003 and 0.84, P = .001).
Figure 3.
Pairwise comparison of AUROC for ELF, FIB4, and APRI against Metavir liver fibrosis stages in all patients. Higher performance of ELF index than FIB4 and APRI scores at (A) mild fibrosis ≥F1, (B) moderate ≥F2, (C) advanced ≥F3 fibrosis, and (D) cirrhosis = F4. Abbreviations: APRI, aspartate to platelet ratio index; ELF, enhanced liver fibrosis.
Table 4.
ELF Index, FIB4, and APRI Performance in All Patients
| Variable | ELF Index | FIB4 | APRI | ELF ~ FIB4 | ELF ~ APRI | FIB4 ~ APRI | |
|---|---|---|---|---|---|---|---|
| n = 206 | n = 206 | n = 206 | P | P | P | ||
| Any/mild fibrosis ≥F1 | AUC | 0.85 | 0.73 | 0.74 | |||
| SEa | 0.027 | 0.041 | 0.042 | .010 | .030 | .695 | |
| 95% CIb | .80–.90 | .67–.79 | .68–.80 | ||||
| Moderate fibrosis ≥F2 | AUC | 0.91 | 0.77 | 0.75 | .0001 | .0001 | .399 |
| SEa | 0.025 | 0.0337 | 0.0349 | ||||
| 95% CIb | .86–.94 | .71–.83 | .69–.81 | ||||
| Advanced fibrosis ≥F3 | AUC | 0.976 | 0.77 | 0.73 | .0001 | .0001 | .054 |
| SEa | 0.010 | 0.037 | 0.041 | ||||
| 95% CIb | .95–.99 | .71–.83 | .66–.79 | ||||
| Cirrhosis F4 | AUC | 0.98 | 087 | 084 | .003 | .001 | .146 |
| SEa | 0.012 | 0.044 | 0.050 | ||||
| 95% CIb | .95–.99 | .82–.91 | .78–.89 |
Abbreviations: APRI, aspartate to platelet ratio index; AUC, area under the curve; CI, confidence interval; ELF, enhanced liver fibrosis; SE, standard error.
aDeLong et al 1988.
bBinomial exact.
Comparison of ELF Index, FIB4, and APRI Performance in HIV/HCV Versus HCV
ELF performed better than FIB4 and APRI in both HIV/HCV coinfection and HCV monoinfection at the Metavir histological stages tested. FIB4 and APRI performance were suboptimal in HIV/HCV with advanced fibrosis (AUROC = 0.71 and 0.60) versus (0.83 and 0.84) in HCV; and cirrhosis (AUROC = 0.83 and 0.79) versus (AUROC = 0.94 and 0.90) in HCV, respectively. However, ELF performance was almost comparable between HIV/HCV and HCV with advanced fibrosis (AUROC = 0.97 vs 0.98) and liver cirrhosis (AUROC = 0.98 vs 0.9) (see Figure 4 and Table 5).
Figure 4.
Pairwise comparison of ELF index, FIB4, and APRI performance in HIV/HCV versus HCV. ELF performed superior to FIB4 and APRI in both HIV/HCV coinfection (A, B, and C), and HCV monoinfection (D, E, and F) at mild fibrosis, advanced fibrosis, and cirrhosis. Suboptimal performance of FIB4 and APRI is detected in HIV/HCV when compared to HCV monoinfection at advanced fibrosis (B, E) and cirrhosis (C, F). ELF index performance was almost comparable between HIV/HCV and HCV with advanced fibrosis (B, E) and liver cirrhosis (C, F). Abbreviations: APRI, aspartate to platelet ratio index; ELF, enhanced liver fibrosis; HCV, hepatitis C virus; HIV, human immunodeficiency virus.
Table 5.
ELF Index, FIB4, and APRI Performance in HIV/HCV Versus HCV Infection
| Variable | HIV/HCV | HCV | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n = 147 | P | n = 98 | P | ||||||||||
| ELF Index | FIB4 | APRI | ELF ~ FIB4 | ELF ~ APRI | FIB4 ~ APRI | ELF Index | FIB4 | APRI | ELF ~ FIB4 | ELF ~ APRI | FIB4 ~ APRI | ||
| Any/mild fibrosis ≥F1 | AUC | 0.82 | 0.72 | 0.74 | .162 | .327 | .557 | 0.91 | 0.73 | 0.71 | .000 | .001 | .110 |
| SEa | 0.04 | 0.06 | 0.06 | 0.033 | 0.056 | 0.06 | |||||||
| 95% CIb | .74–.89 | .63–.79 | .65–.82 | .83–.96 | .63–.82 | .60 –.80 | |||||||
| Advanced fibrosis ≥F3 | AUC | 0.97 | 0.71 | 0.599 | <.000 | <.000 | .008 | 0.98 | 0.83 | 0.84 | .001 | .002 | .074 |
| SEa | 0.016 | 0.056 | 0.063 | 0.010 | 0.049 | 0.048 | |||||||
| 95% CIb | .92–.99 | .62–.79 | .51– .69 | .93–.100 | .73–.90 | .75–.91 | |||||||
| Cirrhosis F4 | AUC | 0.98 | 0.83 | 0.79 | .014 | .0125 | .216 | 0.99 | 0.94 | 0.90 | .62 | .55 | .23 |
| SEa | 0.015 | 0.073 | 0.085 | 0.011 | 0.025 | 0.046 | |||||||
| 95% CIb | .94– .100 | .75–.89 | .71–.86 | .93–.100 | .87–.98 | .82–.96 |
Abbreviations: APRI, aspartate to platelet ratio index; AUC, area under the curve; CI, confidence interval; ELF, enhanced liver fibrosis; HCV, hepatitis C virus; HIV, human immunodeficiency virus; SE, standard error.
aDeLong et al 1988.
bBinomial exact.
Discussion
Liver disease is a major cause of morbidity and mortality among HIV/HCV infected patients [23–25]. Accurate characterization of disease stage drives clinical decision making. In this study, we evaluated the performance of three noninvasive fibrosis biomarker tests (ELF index, APRI, and FIB4) against the gold standard liver biopsy staging in a cohort of HIV/HCV infected and HCV monoinfected participants. The ELF index demonstrated significant association with stages of liver Metavir histological scores. The test can reliably classify nonfibrosis (F0) from any (≥F1), significant (≥F2), advanced fibrosis (≥F3), and liver cirrhosis (F4) with high sensitivity and specificity. The ELF index outperformed 2 noninvasive blood tests, APRI and FIB4, in diagnosis of significant/advanced liver fibrosis and cirrhosis with higher AUROC. APRI and FIB4 are measures of liver injury and portal hypertension. In contrast, the ELF index actually reflects serum extracellular matrix (ECM) turnover by measuring the homeostasis of three biomarkers of extracellular matrixes, HA, TIMP1, and PIIINP. This index reflects the core process of hepatic fibrosis. The original ELF score described by Rosenberg et al, included markers involved in the synthesis and degradation of the liver extracellular matrix (HA, TIMP1, and PIIINP) in addition to age [8]. Parkes et al [9, 10] later defined a derived simplified ELF equation that omits the age component. This omission of age added more clinical utility to the test. An automated, population-based ELF index test will not require entering demographic data for tested patients. Simplification of the algorithm by removing the age component maintained good diagnostic performance in the same cohort and in the validation cohort with mixed etiology of chronic liver diseases [11] and in NAFLD patients [16]. These published results led us to use the ELF index simplified equation in our cohort of patients.
Additionally, we found that the ELF index progressed in a linear manner with histological Metavir stage, which indicates its value as a reliable prognostic marker for fibrosis progression or regression. We previously employed the ELF index in a small cohort (n = 80) [26] of HCV/HIV infected subjects and followed it in another longitudinal cohort and showed that subjects with CCR5 delta-32 mutation or exposed to a blockade agent decreased their ELF scores over time, as compared to controls [27]. Other reports showed that the ELF index can predict disease progression in patients with chronic liver diseases [10, 13, 22, 28] and predict clinical outcomes such as transplant-free survival in primary sclerosing cholangitis [14], as well as mortality in HIV/HCV-coinfected women from the Women’s Interagency HIV Study [29].
Suboptimal performance of noninvasive biomarkers (such as APRI and FIB4) has been reported in HIV/HCV coinfection [6, 7]. In our study, we confirmed a suboptimal performance of APRI and FIB4 in the coinfected population versus HCV monoinfection. However, the ELF index demonstrated excellent performance in identifying advanced fibrosis and cirrhosis in both cohorts. Furthermore, we wanted to elucidate the clinical utility of the ELF index as a diagnostic marker, so we derived data-specific cutoff points for inclusion and exclusion of different stages of liver fibrosis (F1–F4). We find our cutoff values are consistent with previously published data [9, 21, 22].The accuracy of the ELF index in the midranges may make it the best noninvasive biomarker for evaluation of disease progression over time in PLWH.
Some reports suggest that different liver disease etiology (ie, HCV, HBV, and NASH) may have an effect on the ELF index thresholds used for diagnosis [30]. A limitation of our study was the lack of other liver disease etiologies as a primary cause of liver disease.
The ELF index has been validated as a marker for liver fibrosis in patients with hepatitis C [9–11], hepatitis B [12], primary biliary cirrhosis [13], primary sclerosing cholangitis [14], methotrexate-induced liver injury [15], nonalcoholic fatty liver disease [16], and alcoholic liver disease [17]. Others have studied ELF performance as a diagnostic and prognostic test in HIV/HCV coinfection in comparison to noninvasive biomarkers but not to the gold standard of histological staging [28,29,31]. Lichtinghagen et al have shown that the difference in ELF index score reference values were significantly higher for men versus women [21]. In our study, we found no difference in ELF index scores between male (mean ± standard deviation [SD] = 9.23 ± 0.94) and female patients (mean ± SD = 9.1 ± 0.80) (P = .3105) regardless of age or fibrosis stage. We included both sexes, male (n = 175) and female (n = 70), so this study reflects the performance of ELF index in both sexes infected with HIV and/or HCV in comparison to the Women’s Interagency HIV Study (WIHS) that studied ELF performance in HIV/HCV-coinfected women [28,29,31].
To our knowledge, this is the largest study to test the performance of ELF index as guided by Metavir histological stage in HIV/HCV coinfected persons. We also compared the performance of ELF to other noninvasive (FIB4 and APRI) markers of liver fibrosis, which are commonly employed in clinical practice and found them lacking.
In total, 105 subjects with HIV/HCV coinfection from the Swiss HIV Cohort Study compared transient elastography (TE), noninvasive blood biomarkers APRI, FIB 4, Fibrotest, and ELF index only for moderate fibrosis and cirrhosis as classified by histological Metavir stage [32]. The reported results describe a lower performance of ELF at ≥F2 (AUROC = 0.77; 95% CI .67 to .86), and F4 (AUROC = 0.82; 95% CI .69 to .95). In this study we have better performance for ELF at significant fibrosis (≥F2) (AUROC = 0.94; 95% CI .90 to .97), and cirrhosis (F4) (AUROC = 0.98; 95% CI = .92 to .99) in HIV/HCV patients. Contrary to our findings, the authors also reported that ELF performance in the same cohort was equal to FIB4 in significant fibrosis but inferior to FIB4 and APRI in liver cirrhosis.
Zeremski and Talal described the ideal marker of fibrosis in the HIV/HCV population or any population as one that demonstrates high sensitivity and specificity, readily availability, safe, inexpensive, reproducible, and able to follow disease progression [33]. In this study, we found ELF index appears to meet those characteristics of an ideal biomarker of liver fibrosis in HIV/HCV patients. ELF index is particularly important in the HIV/HCV population where the rate of liver fibrosis progression is high and where other noninvasive biomarkers have suboptimal performance.
Notes
Author contributions. E. A. A. contributed to the study concept and design, validation of method, acquisition of data, analysis and interpretation of data, drafting of the manuscript, and critical revision of the manuscript. K. E. S. contributed to the study concept and design; obtained funding, critical revision of the manuscript, study supervision. S. K. contributed by providing samples, obtained funding, critical revision of the manuscript. S. D. R. contributed to critical revision of the manuscript, technical and administrative support.
Disclaimer. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Financial support. This research is supported in part by a grant from the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under Award Number R01 AI065256-06-A1 awarded to K. E. S.
Potential conflicts of interest. S. K. reports grants from Gilead, Merck Inc, and Arbutus, and advisory board fees from Merck, outside the submitted work. K. E. S. reports grants from AbbVie, BMS, and Gilead, Inovio; and personal fees from Inovio, Uniqure, MedPace, and Watermark, outside the submitted work. All other authors report no conflicting financial arrangements related to this manuscript. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
- 1. Bräu N, Salvatore M, Ríos-Bedoya CF, et al. Slower fibrosis progression in HIV/HCV-coinfected patients with successful HIV suppression using antiretroviral therapy. J Hepatol 2006; 44:47–55. [DOI] [PubMed] [Google Scholar]
- 2. Klein MB, Althoff KN, Jing Y, et al. ; North American AIDS Cohort Collaboration on Research and Design of IeDEA; North American AIDS Cohort Collaboration on Research and Design (NA-ACCORD) of IeDEA . Risk of end-stage liver disease in HIV-viral hepatitis coinfected persons in North America from the early to modern antiretroviral therapy eras. Clin Infect Dis 2016; 63:1160–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Graham CS, Baden LR, Yu E, et al. Influence of human immunodeficiency virus infection on the course of hepatitis C virus infection: a meta-analysis. Clin Infect Dis 2001; 33:562–9. [DOI] [PubMed] [Google Scholar]
- 4. Smith C, Sabin CA, Lundgren JD, et al. Factors associated with specific causes of death amongst HIV-positive individuals in the D:A:D Study. Aids 2010; 24:1537–48. [DOI] [PubMed] [Google Scholar]
- 5. Weber R, Sabin CA, Friis-Møller N, et al. Liver-related deaths in persons infected with the human immunodeficiency virus: the D:A:D Study. Arch Intern Med 2006; 166:1632–41. [DOI] [PubMed] [Google Scholar]
- 6. Shire NJ, Rao MB, Succop P, et al. Improving noninvasive methods of assessing liver fibrosis in patients with hepatitis C virus/human immunodeficiency virus co-infection. Clin Gastroenterol Hepatol 2009; 7:471–80, 80.e1-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Mehta SH, Buckle GC. Assessment of liver disease (noninvasive methods). Curr Opin HIV AIDS 2011; 6:465–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Rosenberg WM, Voelker M, Thiel R, et al. ; European Liver Fibrosis Group . Serum markers detect the presence of liver fibrosis: a cohort study. Gastroenterology 2004; 127:1704–13. [DOI] [PubMed] [Google Scholar]
- 9. Parkes J, Guha IN, Roderick P, et al. Enhanced Liver Fibrosis (ELF) test accurately identifies liver fibrosis in patients with chronic hepatitis C. J Viral Hepat 2011; 18:23–31. [DOI] [PubMed] [Google Scholar]
- 10. Parkes J, Roderick P, Harris S, et al. Enhanced liver fibrosis test can predict clinical outcomes in patients with chronic liver disease. Gut 2010; 59:1245–51. [DOI] [PubMed] [Google Scholar]
- 11. Fernandes FF, Ferraz ML, Andrade LE, et al. Enhanced liver fibrosis panel as a predictor of liver fibrosis in chronic hepatitis C patients. J Clin Gastroenterol 2015; 49:235–41. [DOI] [PubMed] [Google Scholar]
- 12. Trembling PM, Lampertico P, Parkes J, et al. Performance of Enhanced Liver Fibrosis test and comparison with transient elastography in the identification of liver fibrosis in patients with chronic hepatitis B infection. J Viral Hepat 2014; 21:430–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Mayo MJ, Parkes J, Adams-Huet B, et al. Prediction of clinical outcomes in primary biliary cirrhosis by serum enhanced liver fibrosis assay. Hepatology 2008; 48:1549–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. de Vries EMG, Färkkilä M, Milkiewicz P, et al. Enhanced liver fibrosis test predicts transplant-free survival in primary sclerosing cholangitis, a multi-centre study. Liver Int 2017; 37:1554–61. [DOI] [PubMed] [Google Scholar]
- 15. Martyn-Simmons CL, Rosenberg WM, Cross R, Wong T, Smith CH, Barker JN. Validity of noninvasive markers of methotrexate-induced hepatotoxicity: a retrospective cohort study. Br J Dermatol 2014; 171:267–73. [DOI] [PubMed] [Google Scholar]
- 16. Guha IN, Parkes J, Roderick P, et al. Noninvasive markers of fibrosis in nonalcoholic fatty liver disease: validating the European liver fibrosis panel and exploring simple markers. Hepatology 2008; 47:455–60. [DOI] [PubMed] [Google Scholar]
- 17. Thiele M, Madsen BS, Hansen JF, Detlefsen S, Antonsen S, Krag A. Accuracy of the enhanced liver fibrosis test vs fibrotest, elastography, and indirect markers in detection of advanced fibrosis in patients with alcoholic liver disease. Gastroenterology 2018; 154:1369–79. [DOI] [PubMed] [Google Scholar]
- 18. Bedossa P, Poynard T. An algorithm for the grading of activity in chronic hepatitis C. The METAVIR Cooperative Study Group. Hepatology 1996; 24:289–93. [DOI] [PubMed] [Google Scholar]
- 19. Lin ZH, Xin YN, Dong QJ, et al. Performance of the aspartate aminotransferase-to-platelet ratio index for the staging of hepatitis C-related fibrosis: an updated meta-analysis. Hepatology 2011; 53:726–36. [DOI] [PubMed] [Google Scholar]
- 20. Sterling RK, Lissen E, Clumeck N, et al. ; APRICOT Clinical Investigators . Development of a simple noninvasive index to predict significant fibrosis in patients with HIV/HCV coinfection. Hepatology 2006; 43:1317–25. [DOI] [PubMed] [Google Scholar]
- 21. Lichtinghagen R, Pietsch D, Bantel H, Manns MP, Brand K, Bahr MJ. The enhanced liver fibrosis (ELF) score: normal values, influence factors and proposed cut-off values. J Hepatol 2013; 59:236–42. [DOI] [PubMed] [Google Scholar]
- 22. Day JW, Rosenberg WM. The enhanced liver fibrosis (ELF) test in diagnosis and management of liver fibrosis. Br J Hosp Med (Lond) 2018; 79:694–9. [DOI] [PubMed] [Google Scholar]
- 23. Monga HK, Rodriguez-Barradas MC, Breaux K, et al. Hepatitis C virus infection-related morbidity and mortality among patients with human immunodeficiency virus infection. Clin Infect Dis 2001; 33:240–7. [DOI] [PubMed] [Google Scholar]
- 24. Kitahata MM, Gange SJ, Abraham AG, et al. ; NA-ACCORD Investigators . Effect of early versus deferred antiretroviral therapy for HIV on survival. N Engl J Med 2009; 360:1815–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Smith CJ, Ryom L, Weber R, et al. Trends in underlying causes of death in people with HIV from 1999 to 2011 (D:A:D): a multicohort collaboration. Lancet 2014; 384:241–8. [DOI] [PubMed] [Google Scholar]
- 26. Abdel-hameed EA, Rouster SD, Sherman KE. Tu1541–validation of non-invasive biomarkers of liver disease in a cohort of HIV/HCV coinfection. Gastroenterology 2019; 156:S-1051. [Google Scholar]
- 27. Sherman KE, Abdel-Hameed E, Rouster SD, et al. Improvement in hepatic fibrosis biomarkers associated with chemokine receptor inactivation through mutation or therapeutic blockade. Clin Infect Dis 2019; 68:1911–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. French AL, Hotton A, Young M, et al. Isolated hepatitis B core antibody status is not associated with accelerated liver disease progression in HIV/hepatitis C coinfection. J Acquir Immune Defic Syndr 2016; 72:274–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Peters MG, Bacchetti P, Boylan R, et al. Enhanced liver fibrosis marker as a noninvasive predictor of mortality in HIV/hepatitis C virus-coinfected women from a multicenter study of women with or at risk for HIV. AIDS 2016; 30:723–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Sebastiani G, Castera L, Halfon P, et al. The impact of liver disease aetiology and the stages of hepatic fibrosis on the performance of non-invasive fibrosis biomarkers: an international study of 2411 cases. Aliment Pharmacol Ther 2011; 34:1202–16. [DOI] [PubMed] [Google Scholar]
- 31. Swanson S, Ma Y, Scherzer R, et al. Association of HIV, hepatitis C virus, and liver fibrosis severity with the enhanced liver fibrosis score. J Infect Dis 2016; 213:1079–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Schmid P, Bregenzer A, Huber M, et al. ; Swiss HIV Cohort Study . Progression of liver fibrosis in HIV/HCV co-infection: a comparison between non-invasive assessment methods and liver biopsy. PLoS One 2015; 10:e0138838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Zeremski M, Talal AH. Noninvasive markers of hepatic fibrosis: are they ready for prime time in the management of HIV/HCV co-infected patients? J Hepatol 2005; 43:2–5. [DOI] [PubMed] [Google Scholar]