Hepatic steatosis in HIV/hepatitis C coinfection: Prevalence and significance compared with hepatitis C monoinfection

Authors

  • Alexander Monto,

    Corresponding author
    1. Gastroenterology Section, Veterans Affairs Medical Center, University of California, San Francisco, CA
    2. Department of Medicine, University of California, San Francisco, CA
    • San Francisco Veterans Affairs Medical Center, 4150 Clement Street, #111B, San Francisco, CA 94121
    Search for more papers by this author
    • fax: 415-750-2196

  • Lorna M. Dove,

    1. Department of Medicine, Columbia University College of Physicians and Surgeons, New York, NY
    Search for more papers by this author
  • Alan Bostrom,

    1. Department of Epidemiology and Biostatistics, University of California, San Francisco, CA
    Search for more papers by this author
  • Sanjay Kakar,

    1. Department of Pathology, Veterans Affairs Medical Center, University of California, San Francisco, CA
    Search for more papers by this author
  • Phyllis C. Tien,

    1. Department of Medicine, University of California, San Francisco, CA
    2. Infectious Disease Section, Veterans Affairs Medical Center, San Francisco, CA
    Search for more papers by this author
  • Teresa L. Wright

    1. Gastroenterology Section, Veterans Affairs Medical Center, University of California, San Francisco, CA
    2. Department of Medicine, University of California, San Francisco, CA
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

Abstract

Liver disease in patients coinfected with HIV and hepatitis C virus (HCV) has received increasing attention in recent years. Steatosis is accepted as an important contributor to liver disease in patients with HCV, but despite coinfected patients having several reasons to have steatosis, the prevalence and significance of such changes has received scant attention. We examined steatosis in an unselected cohort of coinfected patients and compared its prevalence and predictors with findings in monoinfected patients, where these relationships have been established. We studied 92 coinfected and 372 monoinfected patients undergoing staging liver biopsy. Baseline characteristics of the two groups differed significantly, pointing at different contributors to steatosis in each. Histological inflammation and fibrosis were very similar in the two groups, but steatosis was less in coinfected patients. Steatosis had a univariate association with fibrosis in both groups, but retained a multivariate association only in monoinfected patients. Other multivariate predictors of steatosis in monoinfected patients were the accepted variables of elevated body mass index, male sex, and genotype 3a infection, as well as age. In coinfected patients, however, age was the only multivariate predictor. Undetectable HIV viral load was associated with steatosis in coinfected patients in univariate analysis, but highly active antiretroviral therapy or its individual components could not be initially linked to steatosis. In conclusion, steatosis is less common in HIV/HCV coinfected patients than similar HCV monoinfected patients, and predictors of steatosis differ between the two groups. (HEPATOLOGY 2005;42:310–316.)

Hepatic steatosis—collections of triglyceride within hepatocytes—has been found to be a common and important histological finding in patients chronically infected with hepatitis C virus (HCV).1–4 It remains unknown exactly what steatosis means to the liver, and whether it is directly associated with liver injury or is a marker, for example, of increased oxidative stress.5–6 Obesity and type 2 diabetes mellitus seem to contribute to steatosis in a variety of liver diseases, including nonalcoholic fatty liver disease, HCV, and alcoholic liver disease.1, 2, 7–8 HCV genotype 3a is also a specific cause of steatosis.2–4

A number of lines of evidence suggest that steatosis may play a role in the liver disease seen in patients coinfected with HIV and HCV. Hepatic steatosis is quite common in HIV-positive individuals9–10 and has been associated with medications used to treat HIV,11–12 as well as with HIV lipodystrophy syndrome13–14 and HIV-associated lactic acidosis.15 Steatosis is also common in HCV monoinfected patients, affecting approximately 40%. Hepatic fibrosis is believed to progress more quickly overall in coinfected than monoinfected patients,16–18 and because steatosis has an important association with fibrosis in monoinfected patients, it may also contribute to the accelerated fibrosis observed in coinfected patients. The prevalence and significance of steatosis in coinfected individuals are beginning to be examined; recent reports have found an association between steatosis and fibrosis in this patient group19, 20 and have also suggested that HIV medications may play a role in steatosis in coinfected patients.20

The objectives of this study were to compare steatosis in coinfected patients with that in monoinfected patients and to identify predictors of steatosis in each group. We began with the following hypotheses: (1) hepatic steatosis is more prevalent in coinfected than in HCV monoinfected patients and (2) steatosis is associated with the hepatic fibrosis seen in coinfected individuals. It also seemed likely that, because HIV-positive individuals can acquire steatosis in a variety of ways, the observed steatosis could be caused by predictors applicable to monoinfected patients as well as predictors unique to those with HIV infection. Thus, our third hypothesis was that steatosis in coinfected patients could result from elevated body mass index (BMI), genotype 3a infection, and type 2 diabetes (all associated with steatosis in HCV infection alone) as well as from antiretroviral therapy and/or HIV replication (potentially associated with steatosis in those with HIV infection). To test these hypotheses, we conducted a cross-sectional study of patients with HCV undergoing liver biopsy for staging of disease.

Abbreviations

HCV, hepatitis C virus; BMI, body mass index; HAART, highly active antiretroviral therapy.

Patients and Methods

Patients.

The population for this study was derived from liver or infectious disease clinics within the University of California–San Francisco hospital system. All patients had the risks and benefits of liver biopsy for staging of HCV disease before consideration of anti-HCV therapy explained to them, and all decided to undergo the procedure (sites: Veterans Affairs, Moffitt-Long, and San Francisco General Hospital). Liver biopsies performed (by T. L. W., A. M., and L. M. D.) between September 1997 and March 2004 were included. Of these, 288 were consecutive biopsies scored for steatosis prospectively, beginning in 2000; 176 biopsies performed before this time were scored for steatosis retrospectively, based solely on fairly random availability of tissue for re-review. The study was approved by local institutional review boards, and all patients signed written, informed consent. 311 patients were recruited from Veterans Affairs, 136 from Moffitt-Long, and 17 from San Francisco General Hospital, comprising a study group of 464.

Questionnaire.

Demographics and risk factors for HCV and HIV acquisition were recorded. Alcohol consumption was quantified by methods described previously.2 Beer, wine, and liquor consumption were quantified individually from patients' typical quantity and frequency, as well as their duration of use. These were summed and divided by the lifetime duration of alcohol intake to give average alcohol consumption in grams per day.

Diagnosis of HCVand HIV Infections.

All patients tested positive for specific HCV antibodies, had detectable serum HCV RNA as assessed via polymerase chain reaction–based methodology, and had liver histology compatible with chronic hepatitis C disease. Quantitative HCV viral load and genotyping were performed using standard methodologies and were available in the majority of patients. To allow comparison of different HCV RNA quantitative assays, we separated viremia into tertiles, as has been described previously (<33%, 33%-66%, >66%).16

HIV infection was documented by a positive HIV antibody test in the medical record. HIV viral load, CD4 counts, and HIV medication history were also obtained from the medical record.

Histology.

Liver histology was assessed by University of California–San Francisco staff pathologists. Biopsies were scored using the Batts-Ludwig scoring system for chronic hepatitis C, with single inflammation (0-4) and fibrosis (0-4) scores.21 Steatosis was scored according to an accepted scoring system22: 0, no steatosis; 1, <33% of hepatocytes with steatosis; 2, 33%-66% of hepatocytes affected; 3, >66% of hepatocytes affected. The steatosis observed was predominantly macrovesicular, which is consistent with previous studies.2, 22 The portion of the study cohort biopsies that were available for re-review (n = 191) were read by a single pathologist (S. K.) blinded to all clinical data, including HIV status. The remainder were scored by staff pathologists. Histological scores did not differ between those reviewed in blinded fashion and the cohort overall or coinfected or monoinfected patients individually. Progression of fibrosis was calculated by dividing fibrosis score by estimated duration of HCV infection in those patients who had an identified parenteral risk factor for infection.

Serum Assays.

Fasting serum triglycerides, serum alanine aminotransferase, HCV genotype and viral load, and (in HIV+ patients) CD4 count and HIV viral load were obtained proximate to the time of liver biopsy.

Variables Examined.

Clinical variables shown in Table 1 were recorded at the time of liver biopsy. BMI was calculated as weight (kilograms)/height (meters)2. Age at HCV infection and duration of HCV infection were estimated from first exposure to injection drug use or blood transfusion. The year of first injection drug use was used if both risk factors were present; if neither was present, these were not estimated. Questionnaire data were gathered prospectively; BMI and serum assays were obtained prospectively in most patients and by chart review in the remainder. All variables were available for more than 95% of patients (except as noted in the Table 1 footnote). Missing data did not differ significantly between monoinfected and coinfected patients.

Table 1. Patient Clinical and Demographic Variables
 HIV/HCV (n = 92)HCV (n = 372)P Value
  • *

    Available in 87%.

  • Available in 65%.

  • Estimated in patients with injection drug use or blood transfusion.

  • §

    Available in 90%.

Male sex92%87%.16
Age, yrs (mean ± SD)47 ± 749 ± 7.02
Ethnicity, %  .08
 African American3117 
 Asian American23 
 Caucasian6266 
 Latin American48 
 Other16 
BMI* (mean ± SD)25 ± 428 ± 5<.0001
Type 2 diabetes mellitus, %420.0005
Elevated ALT, %5271.0007
Serum triglycerides, mg/dL (mean ± SD)165 ± 91125 ± 65.0003
Risk factor for HCV acquisition, %  <.0001
 History of injection drug use7678 
 History of blood transfusion, no injection drug use114 
 No history of injection drug use nor blood transfusion238 
Age at HCV infection, yrs (mean ± SD)25 ± 925 ± 8.62
Duration of HCV infection, yrs (mean ± SD)22 ± 824 ± 8.03
Alcohol intake, g/d, median (95% CI)§15 (10-34)31 (24-40).03
HCV genotype  .05
 18068 
 2615 
 31115 
 Mixed/other32 
HCV viral load (% in upper third of all values)5632.0004

HIV-associated variables examined included estimated duration of HIV infection, current and nadir CD4 counts, current HIV viral load, and a complete HIV medication history. Highly active antiretroviral therapy (HAART) consisted of a combination of at least two nucleoside reverse-transcriptase inhibitors and either a protease inhibitor or a nonnucleoside reverse-transcriptase inhibitor.

Statistical Analysis.

Demographic and histological values and serum assays were compared across levels of steatosis. Data are expressed as percentages for categorical variables (e.g., sex, diabetes), and means and standard deviations for continuous variables (e.g., age, body mass index). Before fitting a continuous predictor variable as a linear term into a model, it was divided into quartiles, and its relationship to the outcome variable was examined to ensure that its modeling as a linear term in the model was appropriate.

Spearman rank correlations were used to assess the significance of associations between ordinal or continuous predictor variables and steatosis where possible. Nonparametric Mann-Whitney tests for dichotomous predictor variables and Kruskal-Wallis tests for multicategory predictors were also used where appropriate. The relationship between HIV medications and steatosis was assessed with Fisher exact tests and Cochran-Armitage trend tests. Variables are expressed as ordinal or continuous except as noted. The following HIV-associated values were used as dichotomization points because they approximated median values for the group: HIV duration of 7 years; current CD4 count of 400 cells/μL; HAART duration more than 24 months; individual HIV medication duration more than 24 months; HIV viral load undetectable. A CD4 nadir of less than 200 cells/μL was used because of its clinical significance to HIV disease progression.

The independent effect of variables significantly associated with steatosis in univariate analysis (P ≤ .10) was assessed via multivariate proportional odds models. One set of models was constructed using all variables significantly associated with steatosis; a second set was constructed using only clinical (i.e., nonhistological) predictors. Models were constructed for each group separately (HCV monoinfection or HCV/HIV coinfection) to look at predictors within groups, as well as for the entire cohort (HCV monoinfection plus HCV/HIV coinfection) to evaluate the independent effect of HIV infection on steatosis. Tests of proportional odds assumptions were met for all of the models described.

Results

Study Population.

The demographic and clinical characteristics of the study population (92 coinfected, 372 HCV-monoinfected) are shown in Table 1. Coinfected patients differed from monoinfected patients in that they were younger and had a lower BMI, less diabetes, less alanine aminotransferase elevation, higher serum triglyceride levels, and lower alcohol intake. Fewer coinfected patients had an identified parenteral exposure for HCV and/or HIV acquisition, and those who did were estimated to have a shorter duration of HCV infection. Coinfected patients were also more likely to be infected with HCV genotype 1, and had a higher HCV viral load, than their monoinfected counterparts. It should be noted that in the cohort overall and in coinfected and monoinfected patients individually, age at the time of liver biopsy and estimated duration of HCV infection were tightly linked (Spearman rank correlations [95% CI, P values]: cohort overall, 0.46 [0.38-0.54, P < .0001]; coinfected, 0.42 [0.21-0.60, P = .0002]; monoinfected, 0.47 [0.38-0.56, P < .0001]), making the effects of these variables difficult to separate from each other.

Coinfected patients were diagnosed with HIV on average 9 years prior to their liver biopsy (range <1-19), had a mean current CD4 T-cell count of 492 cells/μL, and had a mean CD4 nadir of 274 cells/μL. The mean current HIV viral load was 6,643 copies/mL in the 61% of patients who had a detectable HIV viral load, and 72% had been treated with HAART, for a mean duration of 26 months before liver biopsy. Patients with a current CD4 T-cell count of less than 400 cells/μL were also those who had a CD4 nadir of less than 200 cells/ μL (Fisher exact, P < .0001 for association between these two variables). Age did not correlate with BMI in coinfected patients (Pearson correlation 0.05; 95% CI −0.17-0.26; P = .68).

Histological Findings.

The cohort overall tended to have mild liver disease, with 50% and 51% of the two groups having stage 0-1 fibrosis (Table 2). Inflammation, fibrosis, and calculated rate of progression of fibrosis were similar between the two groups. Mean steatosis score was lower in coinfected than in monoinfected patients (P = .02), and the proportion of coinfected patients with no steatosis was higher (P = .02).

Table 2. Liver Histology
Finding and ScoreHIV/HCV (%)HCV (%)P Value
  • *

    Values are given as mean ± SD.

Inflammation*1.72 ± 0.841.72 ± 0.74.99
 096 
 12626 
 25158 
 31310 
 410 
Fibrosis*1.45 ± 1.261.53 ± 1.26.59
 03327 
 11724 
 22927 
 31413 
 479 
Fibrosis progression*0.137 ± 0.420.075 ± 0.11.98
Steatosis*0.50 ± 0.550.70 ± 0.69.02
 05241 
 14550 
 227 
 302 

Univariate Associations With Steatosis.

In monoinfected patients, variables associated with steatosis (P ≤ .05) included male sex, BMI, infection with HCV genotype 3a, and histological fibrosis (Table 3). In coinfected patients, none of these variables was associated with steatosis other than fibrosis score. Even coinfected patients whose BMI was in the upper quartile did not have higher steatosis scores than those in the lowest quartile (OR 0.75; 95% CI 0.20-2.91; P = .68). In coinfected patients, however, age and an undetectable HIV viral load were associated with steatosis.

Table 3. ORs for Univariate Associations With Steatosis, by Group
 Steatosis in HCVP ValueSteatosis in HIV/HCVP Value
Age1.01 (0.99-1.04).381.10 (1.02-1.18).01
Male sex1.83 (1.02-3.28).042.48 (0.45-13.54).30
BMI1.11 (1.07-1.16)<.00011.00 (0.90-1.11).93
Type 2 diabetes mellitus1.74 (0.80-3.78).162.78 (0.35-22.24).34
Duration of HCV1.02 (0.96-1.06).111.03 (0.97-1.09).32
HCV genotype    
 1 vs. other0.44 (0.28-0.68).00030.50 (0.18-1.39).18
 3a vs. other4.29 (2.29-8.03)<.00011.32 (0.34-5.19).69
HCV viral load    
 Upper 33% vs. lowest 33%1.18 (0.70-1.98).540.64 (0.21-1.97).43
Alcohol    
 >50 g/d0.98 (0.65-1.47).910.70 (0.23-2.18).54
 >80 g/d0.99 (0.63-1.56).981.42 (0.31-6.59).65
Histological inflammation    
 Grades 2-4 vs. grades 0-11.48 (0.97-2.25).072.07 (0.84-5.13).10
Histological fibrosis    
 Stages 2-4 vs. stages 0-12.07 (1.39-3.06).00042.48 (1.05-5.86).04
 Stages 3-4 vs. stages 0-21.87 (1.16-3.02).011.87 (0.65-5.40).25
In HIV-positive only    
 Duration of HIV >7 yrs vs. <7 yrs  1.40 (0.58-3.36).45
 Current CD4 < 400/μL vs. > 400/μL  1.56 (0.63-3.84).33
 Nadir CD4 < 200/μL vs. > 200/μL  1.34 (0.54-3.32).53
 HAART > 24 mos vs. none  2.00 (0.84-4.77).12
 Current HIV viral load undetectable vs. detectable  2.60 (0.99-6.79).05

HAART therapy for more than 24 months was associated with steatosis with an OR of 2.0, but this was not statistically significant (Table 3). In the 86 patients for whom complete HIV medication histories were available, the following were also not associated with steatosis: any history of HAART (P = .78); being in the upper 25% of HAART duration (P = .78); d4T, ddI, or ddC therapy for more than 24 months (P = .42); AZT therapy for more than 24 months (P = .11); d4T therapy for more than 24 months (P = .30); and protease inhibitor therapy for more than 24 months (P = .18).

Multivariate Associations With Steatosis.

In monoinfected patients, male sex, BMI, genotype 3a infection, and histological fibrosis were independently associated with steatosis when all variables with univariate association (P ≤ .10) were included in multivariate models (Table 4). In coinfected patients, only age was independently associated with steatosis in similarly contructed models.

Table 4. ORs for Multivariate Associations With Steatosis, by Group
VariableOR Steatosis in HCV (95% CI) (n = 323)P ValueVariableOR Steatosis in HIV/HCV (95% CI) (n = 85)P Value
  1. NOTE. Data include variables in Table 3 associated with steatosis in each group.

Male sex2.19 (1.02-4.73).05Age1.14 (1.04-1.26).006
BMI1.10 (1.05-1.14)<.0001HAART > 24 mo vs. none2.12 (0.72-6.26).17
Genotype 3a vs. other4.10 (2.11-7.97)<.0001HIV viral load detectable vs. undetectable0.45 (0.15-1.36).15
Inflammation 2-4 vs. 0-11.17 (0.70-1.96).55Inflammation 2-4 vs. 0-12.62 (0.63-10.83).18
Fibrosis 2-4 vs. 0-11.75 (1.08-2.84).02Fibrosis 2-4 vs. 0-11.02 (0.28-3.78).98

When multivariate models were constructed using clinical variables that were associated with steatosis in either monoinfected or coinfected individually (Table 5), age, BMI, and genotype 3a infection had an independent association with steatosis in monoinfected patients. In coinfected patients, as in the previous model, only age retained an independent association with steatosis.

Table 5. ORs for Multivariate Clinical Associations With Steatosis, by Group
VariableOR Steatosis in HCV (95% CI) (n = 323)P ValueOR Steatosis in HIV/HCV (95% CI) (n = 75)P Value
  1. NOTE. Data include clinical (i.e., nonhistological) variables associated with steatosis in either group.

Age1.07 (1.01-1.13).021.14 (1.04-1.26).008
Male sex0.92 (0.25-3.39).912.84 (0.15-54.50).49
BMI1.12 (1.05-1.20).0011.01 (0.88-1.16).92
Type 2 diabetes0.69 (0.27-1.73).422.29 (0.17-31.33).53
Genotype 3a vs. others6.62 (2.50-17.55)<.00011.11 (0.13-9.44).92
HIV viral load detectable vs. undetectable  0.46 (0.16-1.36).16

In a multivariate model of the entire cohort, using the clinical predictor variables in Table 5, with HIV serostatus substituting for HIV viral load detectable versus undetectable, steatosis was independently associated with genotype 3a infection (OR 4.74; 95% CI 2.07-10.89; P = .0002), BMI (OR 1.09; 95% CI 1.03-1.15; P = .005), and age (OR 1.08; 95% CI 1.03-1.13; P = .0006), but not with HIV serostatus (OR 0.87; 95% CI 0.47-1.61; P = .65).

Discussion

Before conducting this study, we anticipated that hepatic steatosis would be more common in coinfected than in HCV-monoinfected patients. However, in this large cohort of patients with chronic hepatitis C, we found that patients coinfected with HIV had somewhat less steatosis than those with HCV infection alone. Possible explanations for this somewhat surprising finding include the fact that, in this cohort, previously described steatosis risk factors (e.g., high BMI, type 2 diabetes mellitus, and genotype 3a infection)1, 2, 4 were less prevalent in coinfected than in monoinfected patients. In particular, the lower BMI (mean 25 vs. 28) in coinfected patients may have led to less steatosis, although the lack of association between BMI and steatosis in coinfected patients even in univariate analysis implies that BMI is not an important cause of steatosis in this group. Even those whose BMI was in the highest quartile did not have more steatosis than those in the lowest quartile. Genotype 3a HCV infection and type 2 diabetes were present in only a small number of coinfected patients in this study, making firm conclusions about the potential relationships between these variables and steatosis problematic. It should be borne in mind throughout these analyses that coinfected and monoinfected patients in this study differed in their characteristics, as would be expected, but that these differences make drawing firm conclusions about biological differences between the two groups difficult. In coinfected and monoinfected patients in this study, however, steatosis was associated with fibrosis in univariate analysis, as has been found in other cohorts,1, 2, 19, 20 but we did not find an independent association in our coinfected patients, as we did in our monoinfected patients. Whether steatosis plays an independent role in the fibrosis seen in coinfected patients, and the mechanisms by which this occurs, will need to continue to be investigated.

Another purpose of this study was to describe which variables predict steatosis in our coinfected patients. The most comprehensive study of this question to date found that Caucasian race, body weight above 86 kg, a history of hyperglycemia, and a history of stavudine (d4T) use were independently associated with steatosis.20 Other studies found histological inflammation23 to be the only multivariate predictor or did not find multivariate predictors.24 In monoinfected patients in this study, we found that the well-established predictors BMI and genotype 3a infection had independent associations with steatosis in all multivariate models. Male sex and age also had independent relationships with steatosis in some models. In coinfected patients, however, predictors of steatosis were different. The only HIV-specific variable associated with steatosis in univariate analysis was undetectable HIV viral load. The only variable independently associated in multivariate analysis was age. The mechanism by which an undetectable HIV viral load might contribute to steatosis is unclear. Having an undetectable HIV viral load has been associated with improved quality of life in HIV-positive patients,25 and this health state could contribute to steatosis through improved nutrition. It seems more likely, however, that undetectable HIV RNA is a surrogate for HAART use, and that it is HAART or different drugs used as part of HAART regimens that play a role in steatosis. Stavudine (d4T) and zidovudine (AZT) were being taken as a component of HAART by many of our patients, and both drugs have been clearly associated with steatosis in the setting of lactic acidosis.26, 27 Stavudine was also recently found to be specifically associated with steatosis in coinfected patients.20 When we analyzed the potential relationship between HAART duration and/or specific individual HIV medications with steatosis in our study, we were unable to demonstrate any such associations. However, it is possible that this is an artifact of relatively small numbers of patients taking these medications who were included. The observation that patients who had received more than 24 months of HAART had a twofold higher likelihood of having steatosis than those receiving shorter durations of HAART or no HAART at all supports this hypothesis.

Age was the variable independently associated with steatosis in coinfected patients in this study, and was also an independent clinical predictor of steatosis in monoinfected patients. The association between age and steatosis in monoinfected patients has been noted in several studies, one of which found the association only in those not infected with genotype 3, implying that age may emerge as a predictor of steatosis when the strong effects of virally mediated steatosis are removed.28 A second study found an independent association between age and steatosis in a multinational study of 3,068 monoinfected patients.29 The mechanism for this could be that older patients have increased susceptibility to insulin resistance and its associated metabolic consequences, including hepatic steatosis, making steatosis an age-dependent process. Age as a variable in coinfected patients may also reflect aspects of HIV infection, such as longer exposure to HIV drugs (duration of HAART therapy was associated with age) or longer duration of HIV infection (also tightly linked to age).

A potential explanation for our inability to demonstrate an association between BMI and hepatic steatosis in coinfected patients is that BMI may be an inaccurate measure of fat mass in HIV-positive individuals—because, for example, the HIV-related lipodystrophy syndrome is characterized by decreased peripheral and increased visceral distribution of fat.30 Future studies examining steatosis in coinfected patients should include measures of both peripheral and visceral fat. Whereas excess body fat, insulin resistance, and genotype 3a infection may ultimately contribute to steatosis in coinfected populations, in the current study, HIV-associated variables and age predominated.

Overall, this study comparing histology in unselected monoinfected and coinfected patients found less steatosis in coinfected patients. A striking finding was that clinical and demographic variables such as age, BMI, prevalence of type 2 diabetes, and alcohol use differed between these two groups. Coinfected patients may differ from monoinfected patients in the same community in important ways, which should be borne in mind when evaluating liver disease in these two groups. Predictors of steatosis in coinfected patients in this study also differed from those in monoinfected patients in all models examined. Whether interventions could be undertaken to ameliorate steatosis in coinfected patients and whether doing so would slow liver disease progression31 and/or improve response to HCV antiviral therapy32 will require further investigation.

Ancillary