| |
Fatty Liver Common in HIV (42%), study finds
|
| |
| |
Magnetic Resonance Spectroscopy of Hepatic Lipid Content and Associated Risk Factors in HIV Infection.
JAIDS Journal of Acquired Immune Deficiency Syndromes. POST ACCEPTANCE, 23 August 2007
Hadigan, Colleen MD, MPH *; Liebau, James ANP ; Andersen, Rebecca BA ; Holalkere, Nagaraj-Setty MD ; Sahani, Dushyant V MD
From the *National Institutes of Health, National Institute of Allergy and
Infectious Diseases, Laboratory of Immunoregulation, Bethesda, MD; and
Program in Nutritional Metabolism and Department of Radiology,
Massachusetts General Hospital, Harvard Medical School, Boston, MA.
".....hepatic steatosis may be an important and clinically unrecognized factor influencing hepatic pathology and function in HIV.....Lemoine and colleagues14 obtained liver biopsies in a sample of 14 HIV-infected patients who were treated with HAART and experiencing unexplained elevations in transaminase levels. In this series, nonalcoholic steatohepatitis was the histopathologic finding in 57% of subjects.... Existing data, however, suggest a potential direct effect of antiretroviral medications on hepatic lipid production, handling, and storage that may contribute significantly to hepatic steatosis and subsequent liver injury....., in the current study, hepatic fat content was significantly associated with increased intramuscular lipid, measured by muscle attenuation on CT. Intramyocellular lipid content has been shown to be increased in and associated with insulin resistance in HIV-infected patients with lipodystrophy in several prior investigations.....abnormal deposition of lipid in muscle and in hepatocytes may be an important component of insulin resistance.... prior studies in which both MRS and liver biopsy were performed show excellent agreement between the 2 techniques. As in the general population without HIV infection, we identified a strong association between the size of the visceral fat compartment and hepatic fat content.37 In addition to the association with VAT, we also observed a strong independent relationship between hepatic fat content and insulin sensitivity and an increased prevalence of metabolic syndrome among
patients with steatosis......
....These data support the hypothesis that central adiposity may contribute to increased steatosis and that both VAT and steatosis are closely linked to insulin resistance in HIV-infected patients. In this preliminary study, we identified hepatic steatosis in 42% of HIV-infected patients, not selected for transaminase elevations, known liver disease, or metabolic complication. Further investigation is needed to more completely determine the prevalence, natural history, and clinical relevance of hepatic steatosis in the general HIV-infected
population in whom liver disease is increasingly recognized as a source of morbidity and who face decades of exposure to antiretroviral therapy...... Of those with steatosis, the range in hepatic fat content was 6.4% to 29%, with a mean (SEM) value of 14% (2)..... subjects with steatosis had significantly greater BMI, WHR, VAT area, and SAT area, and were more likely to meet criteria for metabolic syndrome (Table 1). In addition, thigh-muscle attenuation was significantly decreased, consistent with increased intramuscular lipid..... With respect to metabolic parameters, subjects with hepatic steatosis had significantly increased levels of ALT, triglyceride concentration, insulin, HOMA-IR, and decreased insulin sensitivity index......When the relationship of body composition and metabolic variables to MRS hepatic fat content was assessed by univariate analyses, ALT, HOMAIR
BMI, WHR, and VAT area were all significantly positively correlated with hepatic fat content (Table 2). There was a trend toward an association between hepatic fat and triglyceride concentration (P = 0.05). There were strong inverse correlations between hepatic fat content and thigh-muscle attenuation (P = 0.01) and ISI (P = 0.003). In a forward stepwise multivariate regression analysis including only variables identified as significant in univariate regression, HOMA-IR (P = 0.0004) and VAT area (P = 0.007) were the significant independent predictors of hepatic fat by MRS (whole model r2 = 0.63). The results of this analysis did not change when metabolic syndrome was included in the model.....
.....TABLE 2. Correlation of Hepatic Fat Content by MRS With
Metabolic Parameters (n = 33)
BMI, WHT, ALT, triglycerides (0.05), ISI, HOMA-IR, VAT, SAT, muscle attenuation all associated with liver fat."
Abstract:
Background: Liver-related illness is increasingly recognized as a source of morbidity in HIV-infected patients. Fatty infiltration of the liver is potentially an important consequence of HIV and treatment with antiretroviral (ARV) therapy.
Objective: The aim of the present study was to evaluate HIV-infected men and women for hepatic steatosis using noninvasive magnetic resonance spectroscopy (MRS) and to assess the relationship between liver fat content, insulin resistance, and other associated risk factors.
Methods: We examined 33 consecutively recruited HIV-infected adults without specific referral for liver disease. Subjects with alcohol abuse within 3 years or end-stage liver disease were excluded. The primary clinical measures were hepatic fat content measured by MRS, homeostasis model for assessment of insulin resistance (HOMA-IR), and body fat distribution assessed by cross-sectional computed tomography.
Results:
We identified hepatic steatosis (liver fat content >=5%) in 42% of subjects. Hepatic fat content was significantly correlated with HOMA-IR (r = 0.68, P < 0.0001) and increased visceral adiposity (r = 0.60, P < 0.001).
Subjects with steatosis had significantly increased body mass index and alanine aminotransferase and triglyceride levels, with lower muscle attenuation (ie, increased intramuscular fat) compared to subjects without steatosis. However, steatosis was not related to duration of HIV, ARV exposure, or HCV coinfection.
Conclusions: These data suggest that hepatic steatosis may be very common in HIV, not limited to those with HCV coinfection, and may play an important role in the metabolic profile among HIV-infected men and women.
A potentially important consequence of HIV infection and treatment with antiretroviral therapy may be fatty infiltration of the liver, or hepatic steatosis. Nonalcoholic fatty liver disease (NAFLD) is increasingly recognized in obesity
and in association with central fat accumulation in the HIV-seronegative population.1-4 Findings of increased visceral fat and decreased peripheral fat are well characterized in HIV-infected patients receiving highly active antiretroviral therapy (HAART),5,6 and such patients may therefore be at increased risk for steatosis. In addition, hepatic steatosis is closely associated with insulin resistance, diabetes, and hyperlipidemia, 7 metabolic abnormalities that are increasingly recognized in HIV-infected patients.8-10
Prior studies of select HIV-infected subgroups, such as men with lipodystrophy,11 hepatitis C virus (HCV)-coinfected patients,12,13 and patients with elevated transaminases,14 have identified significant hepatic steatosis by either 1H-magnetic resonance spectroscopy (MRS) or biopsy. For example,
Sutinen et al11 found that HIV-infected men with lipodystrophy had higher liver fat content in association with increased insulin levels compared to controls. In a review of 106 liver biopsies obtained in HIV/HCV-coinfected patients, Marks et al12 found steatosis present in 56% of samples, and the severity of steatosis correlated with the degree of fibrosis. Sulkowski and colleagues13 obtained similar results in a cohort of HIV/HCV-coinfected patients with 2 or more years of
antiretroviral therapy, of whom 40% demonstrated steatosis. In this study, use of the nucleoside reverse transcriptase inhibitor stavudine was associated with increased liver fat. Lemoine et al14 performed liver biopsies on 14 HIV-infected, HAART-treated patients with increased liver enzymes of unclear etiology and found nonalcoholic steatohepatitis (NASH) in 57% of subjects in association with increased expression of sterol regulatory element binding protein 1 (SREBP-1) and insulin resistance.
However, little is known about the prevalence of hepatic steatosis among the general population of HIV-infected patients and its potential association with antiretroviral therapy, fat distribution, and insulin resistance. To evaluate the extent of hepatic steatosis in HIV-infected patients and to assess its relationship to body fat distribution, insulin sensitivity, and other clinical characteristics, we performed 1H-magnetic resonance spectroscopy (MRS) among a group of men and women with HIV infection, irrespective of their antiretroviral status, who were not selected for hepatic coinfection, transaminase elevations, or lipodystrophy.
EXPERIMENTAL PROCEDURES
Subjects
Between November 2004 and March 2006, we recruited 33 HIV-infected men and women from Boston and the greater Boston region through advertisements in the multidisciplinary HIV clinics and community newspaper. Advertisements made no mention of liver disease, antiretroviral therapy, or lipodystrophy. Eligible subjects were HIV-infected, between the ages of 18 and 70 years, and not pregnant within the past 6 months or currently pregnant. Subjects were excluded if they had endstage liver disease, defined as a Model for End-stage Liver Disease (MELD) score >24 15; had previously diagnosed cirrhosis; were undergoing active treatment for viral hepatitis; had a hemoglobin level <9g/dL; reported active substance abuse within the past month or alcohol abuse within the past 3 years; used supraphysiologic testosterone therapy; used growth hormone, prednisone, or other anabolic agents currently or in the previous 6 months; had type 1 diabetes; or experienced a new opportunistic infection within the past 3 months. Due to potential effects on hepatic fat, current use of thiazolidinediones or metformin was exclusionary. If a participant was on antiretroviral therapy, the regimen had to be stable for at least 3 months before participation.
All clinical evaluations were completed in the General Clinical Research Center (GCRC) at the Massachusetts Institute of Technology (MIT) and in the department of Radiology at the Massachusetts General Hospital (MGH). Written informed consent was obtained from each subject before participation in accordance with MGH and MIT institutional review boards for research involving human subjects. Eligibility for the study was established at a screening visit, at which a detailed medical and antiretroviral medication history and physical exam were completed. When necessary, antiretroviral history was confirmed with the primary care provider. Screening blood tests included hemoglobin, creatinine,
bilirubin, and prothrombin time and a urine pregnancy test for all women. Current and past alcohol and substance abuse was assessed using the subset of questions from the Structured Clinical Interview for DSM-IV (SCID)16 used to
establish a diagnosis of substance abuse and dependence. Subjects were also asked to quantify their consumption of alcohol-containing beverages by answering standardized food-frequency questions used to assess alcohol intake in the past 12 months.17
Participants completed 2 subsequent study days for comprehensive metabolic testing, body composition evaluation, and 1H-MRS of liver to quantify hepatic fat content. Study visits were separated by 1 to 14 days and subjects were instructed to arrive in the morning after a 12-hour overnight fast for each visit. On day 1, the cross-sectional CT scan and magnetic resonance spectroscopy (MRS) radiographic tests were completed.
Cross-Sectional CT Scan
A 1-cm cross-sectional CT scan of the abdomen at L4 and a single slice image of the mid-femur were obtained to determine visceral adipose tissue (VAT), abdominal subcutaneous adipose tissue (SAT), and subcutaneous thigh fat area. Scan parameters for each image were standardized (144-cm table height, 80 kV, 70 mA, 2 seconds, 1-cm slice thickness, 48-cm field of view). Fat attenuation coefficients were set at -50 to -250 HU as described by Borkan et al.18 This study was also used to determine relative fat content of thigh muscle using the attenuation values for the hamstring muscles at mid-femur.
Magnetic Resonance Spectroscopy (MRS)
MRS examinations were performed on a 1.5-T whole-body system (GE Medical Systems, Milwaukee, WI) using a torso coil. For single-voxel 1H-MRS, a point-resolved singlevoxel spectroscopy (PRESS) localization sequence was used,
and the acquisition parameters were 3000 msec/30 msec (repetition time/echo time). PRESS spectra were acquired in all subjects in the supine position. A localization voxel of 20 mm3 was placed in the right hepatic lobe at spectroscopy, with care taken to avoid the large intrahepatic vessels. To
diminish tissue contamination from the adjacent structures and to maintain proper localization voxels position was selected by a trained radiologist experienced in liver MRS. Subjects were allowed to breathe in quiet, regular respiration without respiratory interruption during signal acquisition. The absolute
metabolite concentrations were obtained using the LCmodel (Version 6.1, S. Provencher, Oakland, CA) quantification algorithm, and the total hepatic triglyceride concentration/content (HTGC) was obtained by adding lipid concentration at 0.9, 1.3 and 2 ppm in the spectrum.
Blood Tests
On the second test day, subjects underwent the following blood tests: fasting determination of liver transaminases, alkaline phosphatase, and bilirubin levels; lipid profile (total, high-density lipoprotein [HDL], low-density lipoprotein [LDL], and triglycerides); glucose, insulin, and free fatty acids; hepatitis B and C viral serology; and CD4 T-cell count. In addition, a frequent sampled intravenous glucose tolerance test (FSIGTT) was completed. A bolus of 50% glucose
solution (0.5g/kg, with a maximal dose of 35g) was injected over 1 minute, and a single intravenous bolus of insulin (0.03 U/kg) was administered at 20 minutes.19,20 Blood samples were obtained at -15, -10, -5, -1, 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 19, 22, 24, 25, 27, 30, 40, 50, 60, 70, 90, 100, 120, 140, and 180 minutes to determine plasma glucose and insulin concentrations; results were analyzed using the MINMOD Millennium21 to generate an insulin sensitivity index.
Anthropometric Measurements
Standard anthropometric measurements were taken, including determination of height, weight, and waist and hip circumference using standard techniques. Subjects were categorized as having metabolic syndrome based on the updated National Cholesterol Education Program Adult Treatment Panel III (ATP III) criteria for metabolic syndrome. 22 Subjects were instructed to complete and return a 4-day food record that was reviewed in detail with the subject by a research dietician.
Bioassays
Levels of total cholesterol, HDL, triglyceride, glucose, creatinine, bilirubin, and alanine aminotransferase (ALT) were measured using standard techniques. Insulin levels were measured in serum using radioimmunoassay (Diagnostic
Products, Los Angeles, CA). Intra-assay and interassay coefficients of variation range from 3.1% to 9.3% and 4.9% to 10.0%, respectively. CD4 count was determined by flow cytometry (Becton Dickinson Biosciences, San Jose, CA).
Nonesterified fatty acids were measured with an in vitro enzymatic colorimetric assay kit (Wako Chemicals USA, Richmond, VA). The intra-assay coefficients of variation for fatty acids ranged from 1.1% to 2.7%. LDL was measured directly (Genzyme Diagnostics, Cambridge, MA). Hepatitis B surface antigen testing was performed on sera using enzyme immunoassay (EIA; Auszyme, Abbott Laboratories, Chicago, IL). Antibody testing for HCV was performed using EIA-2
(Abbott Laboratories, Chicago, IL). New aliquots from specimens reactive for anti-HCV were then subjected to testing for HCV RNA using qualitative reverse transcription polymerase chain reaction (PCR; HCV Amplicor 2, Roche Molecular Diagnostics, Branchburg, NJ) for confirmation. The sensitivity of this assay is 60 IU/mL.
Biostatistics
Liver fat content >.50 mg fat/g or 5% by wet weight is traditionally considered steatosis.23-25 Consistent with this definition of steatosis, Szczepaniak et al25 used MRS in a cohort of 345 nonobese, nondiabetic adults with low alcohol
intake and found that the 95th percentile of hepatic triglyceride content was 5.6% in this low-risk adult population. Therefore, we used values of >/=5% as the definition of steatosis. Subject characteristics with and without steatosis (ie, >/=5%) were compared using Student's t tests for continuous variables and
x2 statistics for categorical variables. Nonnormally distributed variables (insulin sensitivity index, triglycerides, and percent liver fat by MRS) were log-transformed for statistical analyses to approximate a normal distribution.
Univariate regression analyses were performed with percent liver fat as a continuous variable to assess potential characteristics associated with liver fat content, including age, duration of HIV, body mass index (BMI), waist-to-hip ratio
(WHR), ALT, triglyceride concentration, free fatty acids levels, insulin sensitivity index, homeostasis model for assessment of insulin resistance (HOMA-IR),26 SAT and VAT area, muscle attenuation, and self-reported alcohol intake. Metabolic syndrome, a noncontinuous variable, was evaluated by x2 statistic. Characteristics that were identified as statistically significant by univariate analysis (P <0.05) were entered in a multivariate forward stepwise logistic regression model to assess the relative strength of the associations to steatosis. All values are presented as mean (standard error) unless otherwise indicated, and a 2-level alpha of 0.05 was used to determine statistical significance. All statistical analyses were performed on SAS JMP software, Version 5.0.1.2 (SAS Institute, Cary, NC).
RESULTS
The clinical characteristics of the subjects are summarized in Table 1. Nine of the 33 (27%) subjects were female; overall participants were 55% white, 33% black, and 12% were mixed race and 12% were of Hispanic origin. The mean (SEM) duration of HIV infection was 11 (1) years and the mean (SEM) CD4 T cell count was 441 (45) cells/mm3. Five subjects were not currently on antiretroviral therapy. Seven subjects had HCV confirmed positive by PCR, and only 1 subject was positive for hepatitis B virus (HBV) surface antigen. Thirteen participants (39%) reported a past history of alcohol abuse or dependence, but none within the past 3 years.
Based on the definition of hepatic steatosis of >/=5% hepatic fat content by MRS, 14 of 33 (42%) of subjects had hepatic steatosis. Of those with steatosis, the range in hepatic fat content was 6.4% to 29%, with a mean (SEM) value of 14% (2).
When compared to participants without hepatic steatosis, subjects with steatosis had significantly greater BMI, WHR, VAT area, and SAT area, and were more likely to meet criteria for metabolic syndrome (Table 1). In addition, thigh-muscle attenuation was significantly decreased, consistent with increased intramuscular lipid, in HIV-infected patients with hepatic steatosis compared to those without steatosis (P =0.002). There was no difference between the 2 groups with respect to age, gender, duration of HIV, HCV coinfection, history of alcoholism (P = 0.3), or alcohol intake. The 1 subject with HBV infection did not have hepatic steatosis.
While type 2 diabetes was not an a priori exclusion criterion, no subject entered the study with a known diagnosis of diabetes. One participant met the criteria for type 2 diabetes at the time of participation based on a fasting blood glucose
>126 mg/dL, and this subject had steatosis. When steatosis was analyzed according to race, there was no difference in hepatic fat content or the percentage of subjects with steatosis between races or between Hispanic and non-Hispanic participants (all P values >0.4). In addition, there was no difference in hepatic fat content according to current protease inhibitor exposure or use of stavudine (P values 0.9 and 0.2, respectively). Similarly, 13 of 33 (39%) of participants were on a regimen containing ritonavir, but there was no difference in hepatic fat content (P = 0.8), percentage of subjects with steatosis (P = 0.7), or insulin resistance measured by HOMAIR (P = 0.7) between those on ritonavir and those not on ritonavir.
With respect to metabolic parameters, subjects with hepatic steatosis had significantly increased levels of ALT, triglyceride concentration, insulin, HOMA-IR, and decreased insulin sensitivity index (ISI; Table 1).
When the relationship of body composition and metabolic variables to MRS hepatic fat content was assessed by univariate analyses, ALT, HOMAIR
BMI, WHR, and VAT area were all significantly positively correlated with hepatic fat content (Table 2). There was a trend toward an association between hepatic fat and triglyceride concentration (P = 0.05). There were strong inverse correlations between hepatic fat content and thigh-muscle attenuation (P = 0.01) and ISI (P = 0.003). In a forward stepwise multivariate regression analysis including only variables identified as significant in univariate regression, HOMA-IR (P = 0.0004) and VAT area (P = 0.007) were the significant independent predictors of hepatic fat by MRS (whole model r2 = 0.63). The results of this analysis did not change when metabolic syndrome was included in the model.
DISCUSSION
Liver disease is one of the most common causes of non-AIDS-related deaths among HIV-infected patients with decreased CD4 cell count.27,28 In addition to the significant contribution of hepatitis C and B viral coinfection, hepatic steatosis may be an important and clinically unrecognized factor influencing hepatic pathology and function in HIV. In the era of chronic HIV infection and long-term management with antiretroviral therapy, HIV-infected patients may be exposed to a number of risk factors for hepatic steatosis, such
as drug toxicity, hepatic viral coinfection, dyslipidemia, diabetes mellitus, and abnormal patterns of fat distribution. In the present study we identified a high prevalence (42%) of hepatic steatosis among a sample of HIV-infected men and women not selected for specific risk factors related to steatosis.
Hepatic steatosis in this population was associated with traditional risk factors such as increasing BMI, central adiposity, insulin resistance, metabolic syndrome, and lowgrade transaminase elevations. These data indicate that hepatic steatosis may be common among HIV-infected patients and that it may play a significant role in the insulin resistance and metabolic disturbances seen in this population.
There have been a limited number of published reports evaluating hepatic steatosis in HIV-infected patients outside of the context of hepatitis C coinfection, where rates of steatosis are estimated to be as high as 40% to 56%.12,13 For example, Sutinen et al11 described liver fat content using proton spectroscopy in a sample of HIV-infected men with (n = 25) and without (n = 9) lipodystrophy compared to 35 HIV-negative healthy men. In this study, HIV-infected men with
lipodystrophy had higher liver fat content compared to both control groups. Fasting insulin level, but not visceral fat, was correlated with liver fat content in this study, and insulin sensitivity was not directly assessed. In the present study of patients not selected for lipodystrophy, we identified strong positive correlations between visceral abdominal adipose tissue area and hepatic fat content, and significant relationships between measures of insulin resistance (HOMA-IR and insulin sensitivity index) and degree of steatosis. We identified
an overall prevalence of metabolic syndrome of 27%, which is consistent with other studies of HIV-infected adults in the US.29 Not surprisingly, patients with steatosis were more likely to have a diagnosis of metabolic syndrome, but half of the subjects in this study with hepatic steatosis did not meet criteria for metabolic syndrome, suggesting that increased hepatic fat in HIV cannot be accounted for solely by the presence of a metabolic syndrome profile.
Lemoine and colleagues14 obtained liver biopsies in a sample of 14 HIV-infected patients who were treated with HAART and experiencing unexplained elevations in transaminase levels. In this series, nonalcoholic steatohepatitis was the histopathologic finding in 57% of subjects; hepatic expression of SREBP-1 was increased in HIV-infected patients with insulin resistance compared to non-HIV-infected controls with or without fatty liver. In vitro data using cultured hepatocytes support an effect of antiretroviral medications on the development of hepatic steatosis.30 Both protease inhibitors (indinavir and ritonavir) and the nucleoside reverse transcriptase inhibitor zidovudine were shown to significantly increase expression of SREBP-1 and SREBP-dependent genes related to fatty acid synthesis, a step believed to promote lipid accumulation in the liver. Further, exposure to stavudine was associated with an increased risk of steatosis among HIV/HCV-coinfected patients in 1 study.12,13 In the present study, 85% of subjects were on antiretroviral therapy. We were not able to detect any specific drug or drug class exposure associated with steatosis, but this may be due to the relatively small sample size of the study. Existing data, however, suggest a potential direct effect of antiretroviral medications on hepatic lipid production, handling, and storage that may contribute significantly to hepatic steatosis and subsequent liver injury.
Interestingly, in the current study, hepatic fat content was significantly associated with increased intramuscular lipid, measured by muscle attenuation on CT. Intramyocellular lipid content has been shown to be increased in and associated with insulin resistance in HIV-infected patients with lipodystrophy in several prior investigations.31-33 The relationship between intramuscular and intrahepatic fat has been evaluated in studies outside the context of HIV. For example, one study of patients with and without increased intrahepatic fat showed that intramyocellular lipid content (IMCL) of the soleus
muscle was significantly increased in subjects with steatosis compared to controls.34 However, another study using 1H MRS of the liver and muscle failed to demonstrate a correlation between intrahepatic lipid content and IMCL.35 In the setting of HIV infection, abnormal deposition of lipid in muscle and in hepatocytes may be an important component of insulin resistance that may be the result of antiretroviral toxicity14 or indirect effects of antiretroviral therapy on fat distribution.31,33
In the present study we do not provide data on other histopathologic findings such as fibrosis and inflammation because liver biopsies were not performed. Whereas liver biopsy and histopathologic grading remains the gold-standard
methodology for diagnosing hepatic steatosis, this is not a reasonable approach in a study designed to screen potential at-risk patients for increased hepatic fat. In addition, prior studies in which both MRS and liver biopsy were performed show excellent agreement between the 2 techniques.23,36 As in the general population without HIV infection, we identified a strong association between the size of the visceral fat compartment and hepatic fat content.37 In addition to the association with VAT, we also observed a strong independent relationship between hepatic fat content and insulin sensitivity and an increased prevalence of metabolic syndrome among
patients with steatosis.
These data support the hypothesis that central adiposity may contribute to increased steatosis and that both VAT and steatosis are closely linked to insulin resistance in HIV-infected patients. In this preliminary study, we identified hepatic steatosis in 42% of HIV-infected patients, not selected for transaminase elevations, known liver disease, or metabolic complication. Further investigation is needed to more completely determine the prevalence, natural history, and clinical relevance of hepatic steatosis in the general HIV-infected
population in whom liver disease is increasingly recognized as a source of morbidity and who face decades of exposure to antiretroviral therapy.
|
|
| |
| |
|
|
|
