| |
Liver Steatosis Common in HIV+
|
| |
| |
Histological Findings and Clinical Characteristics Associated with Hepatic Steatosis in Patients Coinfected with HIV and Hepatitis C Virus
The Journal of Infectious Diseases Dec 1, 2005;192:1943-1949
Kristen M. Marks,1 Lydia M. Petrovic,2 Andrew H. Talal,3 Melissa P. Murray,2 Roy M. Gulick,1 and Marshall J. Glesby1
1Division of International Medicine and Infectious Diseases, 2Department of Pathology, and 3Center for the Study of Hepatitis C, Weill Medical College of Cornell University, New York, New York
"......our results suggest that hepatic steatosis is common in patients coinfected with HIV and HCV and that it is associated with more advanced fibrosis .......The high prevalence of steatosis observed in the present study (56%) is consistent with other studies in patients coinfected with HCV and HIV (steatosis frequencies of 40%-67%) and HCV-monoinfected patients (30%-70%). The prevalence of nonalcoholic fatty liver disease in the general population is estimated at 10%-15% among persons of normal weight, which suggests that the frequency of steatosis is significantly increased in the setting of HIV and HCV coinfection...."
ABSTRACT
Background. Hepatic steatosis, a common histological finding in hepatitis C virus (HCV)infected patients, is associated with severity of fibrosis. The prevalence and significance of steatosis in patients coinfected with human immunodeficiency virus (HIV) and HCV are not well characterized.
Methods. To determine the prevalence and severity of steatosis, a single pathologist evaluated liver-biopsy samples from 106 patients coinfected with HIV and HCV but without hepatitis B infection (negative results for hepatitis B surface antigen) for findings associated with steatosis or steatohepatitis and viral hepatitis. Medical records were reviewed retrospectively to elucidate risk factors for steatosis.
Results.
--Steatosis was present in 56% of biopsy samples, with moderate to severe grades in 9%. Severity of steatosis was associated with fibrosis (odds ratio [OR], 1.84 [95% confidence interval {CI}, 1.06-3.20]; P = .03) but not with necroinflammation.
--In multivariate analysis, the severity of steatosis was associated with lower levels of high-density lipoprotein cholesterol (OR, 0.71 per 10-mg/dL increase [95% CI, 0.52-0.95]; P = .02), higher body-mass index (OR, 1.30 per kg/m2 increase [95% CI, 1.13-1.49]; P < .001), and the presence of lipodystrophy (OR, 3.82 [95% CI, 1.13-12.88]; P = .03).
--There was a trend toward an association between the severity of steatosis and fibrosis in multivariate analysis (OR, 1.69 [95% CI, 0.91-3.16]; P = .10).
Conclusions. In patients coinfected with HIV and HCV, hepatic steatosis is common and associated with more-advanced fibrosis. Lower levels of high-density lipoprotein cholesterol, higher body-mass index, and lipodystrophy are potentially modifiable risk factors associated with the severity of steatosis.
BACKGROUND
In recent years, death rates from opportunistic infections have decreased in HIV-infected persons treated with potent antiretroviral therapy [1]. However, 30% of the estimated 900,000 HIV-infected patients in the United States are coinfected with hepatitis C virus (HCV), which remains a substantial source of morbidity and mortality [2, 3]. Patients with HIV and HCV coinfection have an accelerated course of progression to cirrhosis and end-stage liver disease, compared with that in patients with HCV infection alone [47]. As a consequence, end-stage liver disease resulting from HCV infection is now a common cause of death in HIV-infected persons in developed countries [810]. Because the treatment of HCV in the setting of HIV coinfection often does not result in a sustained virological response, it is important to identify the risk factors affecting the progression of liver disease that are potentially modifiable.
The reasons for the increased risk of progression of liver fibrosis in coinfected patients are not well understood. Studies have found associations with advancing age, alcohol consumption, and low CD4+ cell count [11, 12]. Hepatic steatosis, a frequent histopathologic finding on liver biopsy in the setting of HCV infection, has been identified as an additional risk factor for the progression of fibrosis in HIV-uninfected patients with HCV infection. Its prevalence, severity, and impact on liver disease in the setting of HIV infection have not been adequately studied.
Risk factors that have been elucidated for steatosis in HIV-uninfected patients with HCV infection include viral factors, such as HCV genotype 3, and host factors, such as an elevated body-mass index (BMI), visceral obesity, and possibly alcohol use, insulin resistance, and lipid abnormalities [1321]. Antiretroviral therapy may predispose HIV-infected patients toward the development of some of these risk factors. In particular, visceral obesity, which has been associated epidemiologically with the use of protease inhibitors (PIs), may be a component of the lipodystrophy syndrome seen in patients with HIV infection. Certain HIV PIs also appear to cause insulin resistance and hyperlipidemia [2224]. Furthermore, hepatic steatosis in association with hyperlactatemia or lactic acidosis is a rare complication of therapy with nucleoside reverse-transcriptase inhibitors, likely as a result of mitochondrial toxicity [25]. Taken together, these observations suggest that patients with HIV and HCV coinfection could be predisposed toward developing hepatic steatosis, which could adversely affect fibrosis.
To study the relationships among HIV and HCV disease status, antiretroviral therapy, and metabolic complications associated with HIV and its therapy, we conducted a retrospective study of patients with HIV and HCV coinfection attending an urban HIV clinic. The objectives of the study were to determine whether the presence of hepatic steatosis is associated with more-advanced histological findings on liver biopsy and to identify potential risk factors for steatosis in patients with HIV and HCV coinfection.
DISCUSSION
The high prevalence of steatosis observed in the present study (56%) is consistent with other studies in patients coinfected with HCV and HIV (steatosis frequencies of 40%-67%) [29-31] and HCV-monoinfected patients (30%-70%) [13-15, 17-20, 32]. The prevalence of nonalcoholic fatty liver disease in the general population is estimated at 10%-15% among persons of normal weight [33], which suggests that the frequency of steatosis is significantly increased in the setting of HIV and HCV coinfection. Whether this increase is related to HCV infection, HIV infection or treatment, or other causes is not yet known. We observed moderate-to-severe steatosis in only 7% of biopsy specimens from coinfected patients, which is similar to the 4%5% prevalence described in a large study of patients from North America who were monoinfected with HCV, where HCV genotype 1 predominates [34]. Other surveys of HCV-monoinfected patients, however, have found higher prevalences of severe steatosis of 8%23% [13-15, 17-20, 32]. The reasons for these discrepant findings are uncertain, but regional variations in genotype distribution may contribute.
We found that the severity of steatosis on liver biopsy was associated with more-advanced fibrosis, whereas necroinflammation was not. In multivariate analysis, this association between steatosis and fibrosis only approached statistical significance (P = .10). Because, for unknown reasons, a lower prevalence of steatosis has been observed in the setting of cirrhosis in studies of HCV-monoinfected patients and with nonalcoholic fatty liver disease [19, 32, 35, 36], we repeated the multivariate analysis excluding the 7 patients with cirrhosis. This yielded a statistically significant association between steatosis and fibrosis stage (data not shown; OR, 2.14; P = .025). Multiple studies of HCV-monoinfected patients have documented a similar association between steatosis and fibrosis [13, 17-19, 37]. Proposed mechanisms by which steatosis may increase fibrosis in the setting of HCV infection are similar to those proposed for nonalcoholic fatty liver disease. In the setting of steatosis, the addition of oxidative stress from mitochondrial toxicity or viral inflammatory response may lead to lipid peroxidation and the release of proinflammatory and profibrotic cytokines [38]. Other possible contributors include cytokine release due to hyperinsulinemia and/or effects of the adipocyte-derived hormone leptin. It is not known whether the etiology of steatosis affects the association with fibrosis.
We identified plausible risk factors for steatosis in coinfected patients, including increasing BMI, decreasing HDL cholesterol, and the presence of lipodystrophy. Elevated BMI is a well-established risk factor for nonalcoholic fatty liver disease as well as for HCV-associated steatosis in HCV-monoinfected patients, particularly of genotype 1. The association between BMI and steatosis is thought to be due to insulin resistance and hyperinsulinemia [38]. Because of our retrospective design, we could not evaluate insulin resistance; however, neither random serum glucose level determinations nor the presence of diabetes was found to be associated with severity of steatosis.
We also found that lower HDL cholesterol levels were associated with increasing severity of steatosis; however, total cholesterol and non-HDL cholesterol levels were not. The use of lipid-lowering medications is unlikely to be a confounding factor, because only 3 patients in our study were being treated for hypercholesterolemia. Dyslipidemia, particularly increased total cholesterol and triglyceride levels, is a known risk factor for nonalcoholic fatty liver disease, although its contribution to HCV-related steatosis has not been well studied.
The presence of lipodystrophy, defined here as lipoatrophy and/or lipoaccumulation, was associated with steatosis in multivariate analysis. Although our assessment of lipodystrophy was limited, there is biological plausibility for the association with steatosis. Visceral obesity, along with increased BMI, has been shown to be a risk factor for steatosis in HCV-monoinfected patients. The association of lipoatrophy with insulin resistance and/or mitochondrial toxicity offers an additional potential biological mechanism for the development of steatosis. In support of this contention, a recent small study showed that steatosis was found on liver-biopsy samples from 5 of 9 HIV-infected patients with lipodystrophy and abnormal transaminase levels (in the absence of viral hepatitis) [39].
Although lipodystrophy was associated with steatosis in our study, particular antiretroviral therapies were not. In general, steatosis was less advanced in patients receiving antiretroviral therapy; however, this was significant only for NNRTI use in univariate analyses. This inverse association may be explained by certain features of NNRTIs: they are not known to cause insulin resistance and are frequently associated with significant elevations in HDL cholesterol levels [40]. Another preliminary study in coinfected patients showed a similar protective effect of NNRTI use on steatosis in univariate analyses [31]. We found no association between steatosis and PI or stavudine use, in contrast to a study of 113 patients coinfected with HIV and HCV that found an association with prior stavudine use [30].
The 4 patients infected with HCV genotype 3 in our study had steatosis (2 mild and 2 moderate to severe), and genotype 3 was strongly associated with steatosis in univariate analyses, although it did not remain significant in multivariate analysis, perhaps because of limited statistical power. This association has been demonstrated clearly in HCV-monoinfected patients; it is thought to result from a direct viral effect of HCV [14].
Although we did not observe an association between necroinflammation and steatosis, necroinflammation was strongly associated with fibrosis, as has been consistently seen in studies of HCV-monoinfected patients. Other risk factors for fibrosis seen in multivariate analysis included race and the presence of lipodystrophy. The association between African American race and reduced fibrosis does not appear to be due to confounding factors, because African American patients had baseline characteristics similar to those of white and Hispanic patients (data not shown). Furthermore, similar findings were described in a study of HCV-monoinfected patients [41]. In contrast, the impact of lipodystrophy on fibrosis has not been previously reported, and whether it may affect fibrosis through mechanisms other than insulin resistance or steatosis needs further study.
The limitations of our study include the inability to consistently obtain fasting measurements of lipids and glucose and the nonstandardized assessments of the lipodystrophy and metabolic syndromes. The generalizability of our results may be limited by a selection bias, because the study consisted of patients from a single center who were referred for the evaluation of HCV-related liver disease and underwent liver biopsy. However, our patient population is diverse, with significant numbers of women, African Americans, and Hispanics represented.
In summary, our results suggest that hepatic steatosis is common in patients coinfected with HIV and HCV and that it is associated with more advanced fibrosis. Of note, we found that steatosis was associated with potentially modifiable risk factors that are prevalent in HIV-infected patients, such as low HDL cholesterol levels and lipodystrophy, as well as traditional risk factors, such as BMI. Further investigation into the mechanism, impact, and reversibility of steatosis in patients coinfected with HIV and HCV is warranted, given that steatosis may worsen the progression of HCV-related fibrosis, increase the risk of hepatocellular carcinoma [42], and worsen the treatment prognosis of HCV infection [43].
RESULTS
Of 1279 patients identified as being positive for HCV antibody and/or RNA, 143 had liver biopsies done between January 1998 and June 2003. Thirty-seven patients were excluded for the following reasons: fine needle aspirate only (3), positive for hepatitis B surface antigen (6), negative for HCV RNA (2), interferon therapy predating biopsy (9), incomplete medical record (7), and unavailable liver-biopsy slides (10); this left 106 evaluable patients. Table 1 summarizes characteristics of patients at the time of liver biopsy and histological findings.
The distribution of macrovesicular steatosis was as follows: grade 0 (none), 44%; grade 1 (mild), 47%; grade 2 (moderate), 7%; grade 3 (severe), 2%. Only 16 biopsy samples showed microvesicular steatosis: 15 showed grade 1, and 1 showed grade 3. With the exception of 1 patient (grade 1), all patients with microvesicular steatosis had very similar grades of macrovesicular steatosis.
Relationship between steatosis and histological features of HCV infection and steatohepatitis .
Fibrosis stage was associated with the severity of steatosis (OR, 1.84 [95% CI, 1.06-3.20]; P = .031). There was no association between necroinflammatory activity and the severity of steatosis (OR, 0.89 [95% CI, 0.49-1.62]; P = .71). We observed a trend between the severity of steatosis and increased pericellular fibrosis, a pattern of fibrosis that is known to be associated with steatohepatitis (OR, 1.49 [95% CI, 0.96-3.07]; P = .07). The severity of steatosis was not associated with albumin, total bilirubin, transaminase, -fetoprotein, or HCV RNA levels.
Association between clinical characteristics and the severity of steatosis.
In univariate analyses, HCV genotype 3 was the strongest predictor of severity of steatosis (OR, 14.9 [95% CI, 2.03-109.71]; P = .0008). There was no association between duration (in months) of particular types of antiretroviral therapy (e.g., PIs and stavudine) and steatosis. The only significant association between steatosis and antiretroviral therapy was a negative association between use of nonnucleoside reverse-transcriptase inhibitors (NNRTIs) at the time of biopsy and severity of steatosis, which was significant in univariate analyses only. In multivariate analysis, greater BMI, the presence of lipodystrophy (lipoatrophy or lipoaccumulation), and lower levels of high-density lipoprotein (HDL) cholesterol were significantly associated with severity of steatosis, whereas genotype 3 no longer was significantly associated with steatosis.
Association of clinical characteristics with stage of fibrosis.
In multivariate analysis, increasing hepatitis necroinflammatory activity and lipodystrophy each was associated with more severe fibrosis. In contrast, patients of African American descent had less-severe fibrosis than did patients of other races/ethnicities. The severity of steatosis, which was associated with fibrosis in univariate analyses, remained positively associated with severity of fibrosis in multivariate analysis, although this finding was no longer statistically significant (P = .10). The history or duration of antiretroviral therapies was not associated with fibrosis in univariate or multivariate analyses (data not shown).
PATIENTS AND METHODS
Study population. The present study included patients coinfected with HIV and HCV from the 2 HIV clinics (the Centers for Special Studies) of the New York Presbyterian Hospital Weill Cornell Center, which serve a diverse population of about 1800 HIV-infected patients. An electronic database contains information on about 8000 patients, including past and present clinic patients. Patients with documented HIV and HCV infection who underwent liver biopsy between January 1998 and June 2003 who had electronic medical records and liver-biopsy specimens available were included in the study. Patients who underwent fine needle aspirate only or who had an HCV RNA level below the assay limit of detection, history of treatment of HCV infection before liver biopsy, diagnosis of hepatoma, or hepatitis B virus infection (surface antigen or DNA positive) were excluded. The indication for liver biopsy for all remaining subjects was the evaluation of HCV disease.
Histological evaluation.
An experienced liver pathologist (L.M.P.) who was unaware of the clinical characteristics of patients evaluated and graded the liver-biopsy samples with the hepatitis scoring system described by Scheuer [26] and the steatosis and steatohepatitis scoring system described by Brunt et al. [27]. The hepatitis scoring system included necroinflammatory grade (04) and fibrosis stage (04). Steatosis was graded on the basis of the percentage of hepatocytes with macrovesicular steatosis: grade 0, no steatosis; grade 1, 1%-33%; grade 2, 34%-66%; grade 3, >66%. In addition, histological features that are characteristic of steatohepatitis (and not viral hepatitis) were assessed and graded in the following manner, as described by Brunt et al. [27]: hepatocellular ballooning and disarray (none, mild, or marked), intra-acinar (lobular) inflammation (graded 0-3 on the basis of inflammatory foci per 20~ field), and pericellular fibrosis (graded 0-3 on the basis of percentage of zone 3 foci involved).
Variables examined.
Medical records were reviewed retrospectively by one of us (K.M.M.) to ascertain demographic and clinical characteristics and laboratory findings during the 12 months preceding liver biopsy. Demographic variables included age, sex, and race/ethnicity. BMI was calculable for most patients by available height and weight [weight in kilograms/(height in meters)2]. HIV- and HCV-associated variables included risk factors for transmission, years since HIV diagnosis, AIDS diagnosis, CD4+ cell count, and HIV RNA level at time of biopsy (with a limit of detection of <400 or <50 copies/mL as available), nadir CD4+ cell count, HCV genotype, and HCV RNA level. Information on antiretroviral regimen(s) and duration was collected. Lipodystrophy, defined as the presence of lipoatrophy or lipoaccumulation, was based on documentation in the medical record as reported by either the patient or the provider. Alcohol use was elucidated from patients' medical records as well as from a questionnaire completed at the time of initial evaluation in the hepatology clinic. It was quantified as none, mild to moderate, or heavy and was qualified as past or present use. Lipid and glucose levels did not necessarily reflect fasting specimens. Clinical characteristics potentially related to steatosis, such as diabetes mellitus, were also ascertained. The study was approved by the Institutional Review Board of the Committee on Human Rights in Research at Weill Cornell Medical College (New York, NY).
Statistical analysis.
The hepatitis inflammatory grade and fibrosis stage as well as other histological findings were compared across the level of steatosis with Spearman's rank correlation coefficient and by ordinal logistic regression. Because few patients had advanced grades of steatosis, grades 2 and 3 were combined for statistical analysis. This resulted in 3 categories of steatosis: no steatosis (grade 0), mild steatosis (grade 1), and moderate to severe steatosis (grade 2-3). An exploratory analysis was done to identify clinical characteristics associated with steatosis and fibrosis using the nonparametric Mann-Whitney U test (for dichotomous predictor variables), Spearman's rank correlation coefficient (for ordinal or continuous predictor variables), and the Kruskal-Wallis test (for multicategory predictor variables).
Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated by ordinal logistic regression with steatosis or fibrosis categories as the response (dependent) variables. These ORs approximated the risk of having a steatosis or fibrosis grade that is 1 category higher as a result of the defined increase in a continuous independent variable or a change from "no" to "yes" in a dichotomous independent variable. Multivariate analysis was done by ordinal logistic regression with forward, stepwise regression; the entry criterion into the model was P <.2. All covariates used in the models met the assumption of parallel regression [28]. HIV RNA levels below the level of detection (<400 or <50 copies/mL) were assigned values of 400 and 50, respectively, before log transformation. Histological scores that fell between 2 scores were uniformly rounded to the higher value; for example, fibrosis stage 12 was recorded as 2. Data were analyzed with STATA software (version 7.0; STATA).
|
|
| |
| |
|
|
|
