| |
HCV/HIV Coinfection Higher Death Rates Higher in VA Study
|
| |
| |
The American Journal of Gastroenterology
Volume 101 Page 760 - April 2006
Nathan A. Merriman1, Steven B. Porter1, Colleen M. Brensinger2, K. Rajender Reddy1, and Kyong-Mi Chang1,3
1Division of Gastroenterology, Department of Medicine, University of Pennsylvania, 2Center for Clinical Epidemiology and Biostatistics, 3Philadelphia Veterans Affairs Medical Center, Philadelphia, Pennsylvania
"....Our findings confirm previous observations of increased morbidity and mortality in HCV/HIV coinfected patients.....HIV coinfection in HCV-infected patients results in differential outcomes both virologically (e.g., increased HCV persistence and higher HCV viral load) and clinically (e.g., reduced response to antiviral therapy and accelerated liver disease progression). Similarly, HBV/HIV coinfection is associated with increased liver disease progression. In this study, we examined the impact of viral coinfections on HCV and HIV-associated outcomes among a U.S. veteran population.... seeking care at the VA Medical Center in Philadelphia......with high prevalence of African Americans, HBV exposure, injection drug use, and alcohol use disorder in the era of highly active antiretroviral therapy. We found an increased frequency of liver dysfunction and all cause mortality in HCV/HIV-coinfected patients independent of CD4 count, antiretroviral usage or HCV/HIV titers but significantly associated with race....."
".....HCV/HIV coinfection was associated with greater mortality compared to HCV or HIV monoinfection.....more than twice as many patients in the coinfected group expired between 2000 and 2002 compared to the monoinfection groups (11% HCV/HIV vs 4% HCV vs 5% HIV). This trend continued into the 3rd yr between 2000 and 2003 (17% HCV/HIV vs 6% HCV vs 9% HIV)...."
"......HCV/HIV-coinfected patients displayed significantly greater abnormalities in albumin level, INR, and platelet count compared to HCV-monoinfected patients. Therefore, more of the HCV/HIV-coinfected patients displayed evidence of clinically significant liver disease (i.e., at least one abnormality in standard liver function parameters) compared to HCV-monoinfected patients...."
"Racial Difference in Mortality Among U.S. Veterans with HCV/HIV Coinfection"
ABSTRACT
OBJECTIVES: This study was performed to examine the impact of viral coinfections and race on clinical and virological outcome of hepatitis C virus (HCV) infection.
METHODS: Three groups of patients (265 HCV/HIV coinfected, 251 HCV monoinfected, 227 HIV monoinfected) were identified between 2000 and 2002 from the computerized patient record system at the Philadelphia VA Medical Center and analyzed for clinical and virological parameters.
RESULTS: HCV/HIV coinfection was associated with higher frequency of liver function abnormalities (37%vs 21%vs 20%; p< 0.0003) and greater mortality (17% vs 6% vs 9% over 3 yr period, p= 0.0003, p= 0.027) compared to HCV or HIV monoinfection, respectively.
However, HCV/HIV coinfection was not associated with worsened HIV-related parameters (CD4 count, HIV titer, and use of antiretroviral therapy) or increased HCV titers compared to HIV or HCV monoinfection in our population, respectively. Interestingly, mortality among HCV/HIV coinfected patients was significantly greater in white than in black patients (31%vs 15%, p= 0.011). This racial disparity in mortality was not apparent in the monoinfected groups and not explained by HBV coinfection or history of alcohol use disorder.
CONCLUSIONS: We conclude that HCV/HIV coinfection is associated with worsened liver disease and higher mortality than HCV- or HIV-monoinfection without directly influencing CD4 count and HCV or HIV titers. Furthermore, we demonstrated a racial disparity in survival of HCV/HIV-coinfected patients that needs further investigation.
INTRODUCTION
Hepatitis C virus (HCV) is a highly persistent virus causing chronic necroinflammatory liver disease that can progress to cirrhosis and liver cancer over many years. The natural history and treatment responsiveness of HCV infection is determined by various host, viral, and environmental factors (1). One relevant factor is coinfection with human immunodeficiency virus (HIV), another highly persistent virus with a similar route of transmission as HCV. Indeed, over a third of HIV-infected patients are coinfected with HCV (2) and HCV coinfection has emerged as a major comorbid factor in HIV-infected patients with their improved survival on highly active antiretroviral therapy (3-5). For example, liver fibrosis may be accelerated in HCV-infected patients coinfected with HIV in association with low CD4 count and heavy alcohol use (6, 7). Furthermore, antiretroviral therapy may be more hepatotoxic in HCV-infected patients due to direct drug toxicity or immune reconstitution (8, 9). As for the overall survival, there are studies that show a greater progression to AIDS and death with HCV/HIV coinfection (10-12) as well as those that do not (13, 14). Similarly, an inverse relationship between CD4 count and HCV RNA was observed in some (15-17) but not all studies (18-21).
In this study, we examined the clinical and virological outcomes of HCV and/or HIV-infected U.S. veterans identified in the Philadelphia Veterans Affairs Medical Center (PVAMC) in 2000-2002, further exploring the contribution of hepatitis B virus (HBV) exposure and race in their outcomes. Our results show that HCV/HIV coinfection is associated with worsened overall morbidity and mortality but without significant changes in antiretroviral therapy, HIV, or HCV viral titers or CD4 counts. This worsened outcome was not due to increased HBV coinfection in HCV/HIV coinfected patients although chronic HBV infection was independently associated with increased mortality. Finally, we identified an unexpected racial difference in mortality with higher mortality in HCV/HIV coinfected white patients compared to black patients.
AUTHOR DISCUSSION
HCV is a parenterally transmitted viral pathogen causing persistent infection and chronic liver disease. The outcome of HCV infection is determined by the virus-specific T-cell response as well as host, viral, or environmental factors that may modify this response (23, 24). In particular, African-American ethnicity has been associated with less fibrosis and liver function abnormalities in the setting of more robust (albeit dysfunctional) antiviral CD4 T-cell response (24) while alcohol may accelerate liver disease progression via direct hepatotoxic and immunomodulatory effect. Furthermore, HIV coinfection in HCV-infected patients results in differential outcomes both virologically (e.g., increased HCV persistence and higher HCV viral load) and clinically (e.g., reduced response to antiviral therapy and accelerated liver disease progression) (16, 25-31). Similarly, HBV/HIV coinfection is associated with increased liver disease progression (32). In this study, we examined the impact of viral coinfections on HCV and HIV-associated outcomes among a U.S. veteran population with high prevalence of African Americans, HBV exposure, injection drug use, and alcohol use disorder in the era of highly active antiretroviral therapy. We found an increased frequency of liver dysfunction and all cause mortality in HCV/HIV-coinfected patients independent of CD4 count, antiretroviral usage or HCV/HIV titers but significantly associated with race.
We found a twofold greater mortality among HCV/HIV coinfected white patients compared to black patients (31%vs 15%, p= 0.011). Such difference was not apparent for HCV- or HIV-monoinfected patients, suggesting a finding specific to HCV/HIV coinfection. HCV/HIV-coinfected white patients also tended to die at a younger age compared to black patients. These results suggested a racial difference in the outcome of HCV/HIV coinfection among our veteran population, perhaps consistent with better liver-associated laboratory parameters and reduced histological progression in HCV-monoinfected African American patients (33, 34). The increased and earlier death among HCV/HIV-coinfected white patients was not associated with differences in antiretroviral usage, CD4 count, HIV viral load, HBV coinfection, or alcohol use. It is unlikely to be based on HCV therapy since patients with history of IFNƒ¿-based antiviral therapy were excluded in this study. However, we cannot rule out underlying genetic or sociodemographic differences between the ethnic groups that could not be addressed in the current study.
As for the effect of HCV on HIV-associated parameters, HCV/HIV coinfection was not associated with lower CD4 count or higher HIV viral load, consistent with several previous reports (13, 14) and suggesting that HCV does not directly influence HIV replication or CD4 T cell survival. Furthermore, contrary to reports suggesting that HCV/HIV coinfection may negatively impact antiretroviral therapy (9, 35-37), we found a comparable frequency and pattern of antiretroviral therapy utilization and similar efficacy in the HCV/HIV-coinfected and HIV-monoinfected patients without worsening liver function tests. These results suggest that HCV does not directly influence HIV replication or CD4 T cell survival. Nonetheless, the strong inverse relationship between HIV viral load and CD4 count was diminished in HCV/HIV-coinfected patients not on antiretroviral therapy as shown in Figure 1. Although the mechanism for this change is not clear, one potential area for exploration may be a subtle T-cell suppressive effect recently reported in HCV persistence (38).
As for another hepatotropic viral infection, HBV coinfection with detectable HBsAg was significantly greater among HIV-monoinfected patients and significantly less for HCV infected patients with and without HIV coinfection. Thus, HBV may be cleared more readily in the setting of HCV infection. Indeed, HBV replication and gene expression can be completely terminated by inflammatory cytokines (e.g., IFNƒÁ, TNFƒ¿, or IL-12) induced by unrelated hepatotropic viral or parasitic pathogens in murine model of HBV replication (39). Nonetheless, the association between HBsAg and increased mortality as well as previous studies suggesting a more fulminant outcomes of acute viral hepatitis in HCV-infected patients (40, 41) support the continued need for HBV vaccination (as well as hepatitis A).
A number of limitations in our study must be acknowledged. First, liver histology was not available for most patients in our study due to the limited number of veterans referred for liver biopsy at the PVAMC. Furthermore, we could not determine if the increased mortality was specifically due to liver disease. Yet, the available liver function and mortality data remain clearly relevant for the overall outcome of the patients in each group. Second, there are obvious limitations with the cross-sectional study design (e.g., unknown duration of HIV and HCV infection, survivorship bias). Nonetheless, patients in all three groups were similar in age while mortality was monitored over a 3-yr period. Third, while alcohol use disorder was not associated with mortality in our study, this can be missed without a direct patient survey.
In conclusion, we examined the clinical, immunological, and virological consequences of HCV and HIV coinfection among U.S. veterans seeking care at the VA Medical Center in Philadelphia. Our findings confirm previous observations of increased morbidity and mortality in HCV/HIV coinfected patients, while identifying an unexpected racial disparity in morbidity and mortality among HCV/HIV-coinfected patients. Our results also suggest that HCV coinfection does not directly impact HIV titers, CD4 count and anti-retroviral therapy while it is negatively associated with chronic HBV infection. Future studies examining the pathogenetic as well as sociodemographic factors contributing to these findings are needed.
RESULTS
General Demographic and Clinical Parameters in the Patient Groups
A total of 265 HCV/HIV-coinfected subjects, 227 HIV-monoinfected subjects, and 251 HCV-monoinfected subjects were studied. As shown in Table 1, patients in all three groups were predominantly males aged between 40 and 60 yr old. Black Americans were highly represented in all groups, although more so in the two HIV+ groups (HCV/HIV-coinfection and HIV-monoinfection) compared to HCV-monoinfection group (84%vs 81%vs 62%, p< 0.0003). Documented alcohol use disorder was significantly greater among HCV-monoinfected patients (63%) than in HCV/HIV-coinfected (51%) or HIV-monoinfected patients (36%). Nonetheless, the liver function parameters (e.g., albumin, total bilirubin, INR, and platelet count) were mostly normal on average in all three groups (Table 2), with the exception of the transaminase levels which were greater in HCV-infected than in HCV-uninfected patients. Thus, these 743 mostly middle-aged U.S. male veterans with defined demographic, clinical, and virological characteristics provided the basis for our subsequent analysis.
Effect of HIV on HCV-Associated Liver Disease and Viral Titer
To determine the effect of HIV on HCV-associated liver disease, we further examined the laboratory measures of liver function in HCV-infected patients with and without HIV coinfection. As shown in Table 2, HCV/HIV-coinfected patients displayed significantly greater abnormalities in albumin level (p< 0.0003), INR (p< 0.0003), and platelet count (p= 0.001) compared to HCV-monoinfected patients. Therefore, more of the HCV/HIV-coinfected patients displayed evidence of clinically significant liver disease (i.e., at least one abnormality in standard liver function parameters) compared to HCV-monoinfected patients (37%vs 21%, p< 0.0003). Although HIV infection can independently influence these laboratory parameters, HIV-monoinfected patients were less likely to have such liver function abnormalities compared to HCV/HIV-coinfected patients (20%vs 37%, p< 0.0003), suggesting an increased liver-related morbidity due to HCV/HIV coinfection.
This worsened liver function in HCV/HIV-coinfected patients was not associated with greater alcohol consumption since alcohol use disorder was actually less frequent among coinfected than HCV-monoinfected patients as shown in Table 1 (51%vs 63%p= 0.018). Also, while genotype 1 infection was more prevalent among HCV/HIV coinfected than in HCV-monoinfected patients as shown in Table 2 (95%vs 82%, p= 0.003), there was no genotypic difference in liver function abnormalities among HCV-infected patients (data not shown). Furthermore, contrary to previous reports (16, 17), there was no difference in median HCV RNA titer between HCV-monoinfected and HCV/HIV-coinfected patient groups (Table 2). While 11 patients in the HCV/HIV-coinfection group were HCV RNA-negative, exclusion of these HCV RNA-negative patients did not affect our findings on liver function abnormalities and HCV RNA titers. In multivariable analysis controlling for patient age, race, ETOH use, HBcAb, and HBsAg status, HCV- and HIV-monoinfected patients had significantly lower odds of abnormal liver function parameters compared to HCV/HIV-coinfected patients (OR = 0.43, 95% CI 0.28-0.66 for HCV group, p= 0.0001; and OR = 0.34, 95% CI 0.22-0.55, p< 0.0001). Thus, HIV coinfection was associated with increased liver function abnormalities but not increased HCV titer in HCV-infected patients.
Effect of HCV on HIV-Associated Outcomes
The effect of HCV-coinfection on HIV was evaluated by comparing the HIV viral load and CD4 count in HCV/HIV-coinfected and HIV-monoinfected patients. All HIV+ patients had been tested for CD4 count and HIV viral load as part of our study design. We did not find significant differences in HIV-associated outcomes between HCV/HIV-coinfected and HIV-monoinfected groups. For example, as shown in Table 3, most HIV+ patients were on antiretroviral therapy irrespective of HCV status (78% and 80%, p= 0.51), usually combining multiple antiretrovirals as expected. HIV+ patients on antiretroviral therapy displayed higher serum albumin levels (mean 3.9 vs 3.6, p< 0.0001), platelet counts (mean 217 vs 204, p= 0.044), and lower mortality (16%vs 6%, p= 0.0009) compared to untreated patients, suggesting a more favorable outcome associated with antiretroviral therapy. Effective virus suppression (HIV viral load <50 copies/mL) was observed in about a third of patients in both groups (30%vs 36%, p= 0.21). Among patients with quantifiable HIV viremia, there was no significant difference in mean HIV titer with and without HCV-coinfection (4.8 vs 5.4 X 104 copies/mL, p= 0.36). There was also no significant difference in CD4 count between HCV/HIV-coinfected and HIV-monoinfected patients (360 vs 415, p= 0.19). Among HIV-viremic patients not on antiretroviral therapy, CD4 count was negatively correlated with HIV viral load, although significantly more so for HIV-monoinfected than in HCV/HIV-coinfected group (R: -0.57 vs-0.09, p= 0.011) (Fig.1A). By contrast, HCV RNA titer showed a positive, rather than a negative correlation with CD4 count (Fig. 1B), suggesting that HCV replication is not increased by HIV-associated loss of CD4 T cells. Taken together, these results suggest that HIV-associated outcomes are not markedly influenced by HCV coinfection.
Increased Morbidity and Mortality in HCV/HIV-Coinfected Patients
We then evaluated the all cause mortality relative to clinical and demographic characteristics in the patient groups. HCV/HIV coinfection was associated with greater mortality compared to HCV or HIV monoinfection. As shown in Table 4, more than twice as many patients in the coinfected group expired between 2000 and 2002 compared to the monoinfection groups (11% HCV/HIV vs 4% HCV vs 5% HIV, p= 0.0009). This trend continued into the 3rd yr between 2000 and 2003 (17% HCV/HIV vs 6% HCV vs 9% HIV, p< 0.009). The expired patients tended to be older than the survivors across the three groups (mean age 51 vs 48 yr, p= 0.0032). While the cause of death could not be obtained in our study, the expired patients were more likely to have abnormal liver function parameters compared to the survivors in all three groups (Table 4). Mortality did not differ between persons with and without alcohol use disorder (11%vs 12%, p= 0.56). The expired HIV+ patients also displayed a lower CD4 count (198 vs 415, p= 0.0001) and higher HIV titer (107,495 vs 24,905 copies/mL, p= 0.0001) compared to the survivors, consistent with a more advanced HIV-associated immune deficiency. In multivariable logistic regression analysis adjusting for patient age, race, and HBcAb/HBsAg status, both monoinfected groups had significantly lower odds of mortality than the HCV/HIV coinfected group (OR = 0.29, 95% CI 0.16-0.56 for HCV group, p= 0.0003; OR = 0.31, 95% CI 0.15-0.61 for HIV group, p= 0.0006). For every 10 yr increase in age, the odds of mortality increased by 1.82 (95% CI 1.33-2.48, p= 0.0002) while odds for liver function abnormalities increased by 1.32 (95% CI 1.05-1.66, p= 0.02). Albumin level had the greatest impact on mortality across all three groups in that patients with albumin levels below 3.5 had were four times more likely to die compared to those with normal albumin levels (p< 0.0001).
Racial Disparity in Morbidity and Mortality Among HCV/HIV Coinfected Patients
Interestingly, there was a significant racial disparity in morbidity and mortality among the HCV/HIV-coinfected patients. As shown in Table 4, twice as many white patients expired over the 3 yr period compared to black patients (31%vs 15%p= 0.011). The median age at death was also younger in white than black patients with HCV/HIV coinfection (46 vs 52 yr old, p= 0.033). HCV/HIV-coinfected white patients were more likely to exhibit abnormal platelet count or total bilirubin levels compared to coinfected black patients (Fig. 2), despite a lower prevalence of alcohol use disorder compared to coinfected black patients (37%vs 53%, p= 0.05). While HBsAg-positive patients had significantly higher odds of mortality in multivariable logistic regression analysis (OR = 3.24, 95% CI 1.16-9.06, p= 0.03), concurrent HBV infection did not explain the racial difference in mortality since detection of HBsAg was actually greater among HIV monoinfected patients than HCV/HIV-coinfected patients (8%vs 3%, p= 0.033) or HCV monoinfected patients (1%, p= 0.429) (Table 5). HCV/HIV-coinfected patients were more likely to test positive for anti-HBc than HCV- or HIV-monoinfected patients (87% HCV/HIV vs 68% HCV vs 60% HIV) including for isolated anti-HBc (48% HCV/HIV vs 37% HCV vs 17% HIV), raising the possibility of occult HBV infection. However, the distribution of isolated anti-HBc was equally among expired and surviving HCV/HIV-coinfected patients (% anti-HBc+: 50% among expired vs 48% among survivors). Taken together, these results suggest an increased morbidity and mortality among HCV/HIV-coinfected patients with an unexpected ethnic disparity independent of alcohol, HBV coinfection.
Study methods
Study Subjects
This was a cross-sectional study of HCV-monoinfected, HIV-monoinfected, and HCV/HIV-coinfected patients utilizing the computerized patient record system (CPRS) at the PVAMC performed with the approval of Internal Review Board (IRB) at the PVAMC. Three groups of subjects were identified on the basis of virological status: (1) HIV-monoinfected; (2) HCV/HIV-coinfected; (3) HCV-monoinfected. HIV+ patients in Group 1 and 2 were initially identified on the basis of available CD4 T-cell count performed between February 2000 and February 2002. Since all HIV-infected patients followed at the PVAMC are tested regularly for their CD4 T-cell count, this approach ensured that we identified all potentially HIV-infected patients followed at PVAMC for subsequent confirmation of HIV-infection based on chart review for prior HIV antibody and viral titers. This resulted in a total of 525 subjects of which 492 were confirmed to be HIV antibody (Ab) positive. Of these HIV positive subjects, 265 subjects were HCV/HIV coinfected (HCV Ab+, HIV Ab+), and 227 were confirmed to be HIV monoinfected (HCV Ab-, HIV Ab+). HCV monoinfected subjects (HCV Ab+, HIV Ab-) were randomly selected from 991 individuals who tested positive for HCV RNA by Roche Cobas Amplicor v2.0 and/or Monitor v2.0 at the PVAMC during the same time period and a negative HIV Ab test within 1 yr of HCV RNA testing. Exclusion criteria included unknown HIV or HCV Ab status and known past or current interferon or ribavirin therapy.
Data Abstraction
Demographic information (age, sex), clinical liver function parameters (AST, ALT, total bilirubin, albumin, platelets, PT/INR or prothrombin time/international normalized ratio), alphafetoprotein (AFP), CD4 count, and virological parameters (HCV and HIV viral load by Roche COBAS Amplicor, HCV genotype, HBsAg, HBsAb, HBcAb) were obtained for each subject from the CPRS database, selecting the values closest to the time of study entry, defined by the dates of CD4 count (for all HIV+ subjects) and HCV RNA (for HCV monoinfected subjects). Race/ethnicity was determined by CPRS database based on self-report and/or health-care providers' assessment. All values were obtained within 6 months of study entry.
All cause mortality was compiled during the initial 2 yr study period (February 2000 to February 2002), and extended through February 2003, providing a mortality rate over a 3 yr period. The specific cause of death was not obtainable from the CPRS database. Since liver biopsy was not performed in most patients, clinically significant liver disease was defined by the presence of one or more clinical laboratory parameters including albumin ≦3.5 g/dL, INR ≥1.25, total bilirubin ≥2.0 mg/dL, or platelets ≦130,000 based on the respective normal ranges.
Determination of Antiretroviral Therapy
The antiretroviral therapy status of all HIV+ patients was determined through the PVAMC computerized pharmacy record, which maintains an accurate log of the type and frequency of all antiretroviral prescriptions provided to HIV+ patients receiving care at the PVAMC. We compiled lists of individual antiretroviral medications as well as their medication class (i.e., protease inhibitors (PI), non-nucleoside reverse transcriptase inhibitors (NNRTI), and nucleoside reverse transcriptase inhibitors (NRTI)). HIV+ patients were then designated as being treated with antiretrovirals based upon computerized medication refill records demonstrating compliance with two or more antiretroviral medications during the study period.
Determination of Alcohol Use Disorder
History of alcohol use disorder and/or alcohol abuse was determined in all subjects by thorough chart review as previously described (22), based on one or more of the following parameters: (1) treatment at the VA alcohol addiction treatment center; (2) explicit physician designation of alcohol abuse in the patient's clinic notes; (3) positive CAGE questionnaire; and (4) assignment of alcohol abuse to the patient's computerized medical record active problem list.
Statistical Analysis
Pearson's X2 or Fisher's exact tests were used to test for associations between study group status and demographic or clinical parameters. The Kruskal-Wallis test was used to compare AST, ALT, albumin, total bilirubin, PT, INR, HIV viral load, and CD4 counts between the three study groups. When a significant difference was found, all possible pairwise comparisons between groups were examined with the p values adjusted for multiple comparisons by Bonferroni adjustments, and hence control the type I error rate. Mann-Whitney U tests were used to compare various outcomes between patients with and without alcohol use disorder and to compare antiretroviral therapy users versus nonusers. Pearson correlation coefficients assessed the strength of the linear association between HIV and HCV titer relative to CD4 count. Finally, multivariable logistic regression models were used to test for differences in the odds of mortality and cirrhosis between the three groups, adjusting for patient age, race, ETOH use, HBcAb, and HbsAg status. Results of the models are presented as odds ratios (OR) and 95% confidence intervals (CI). All analyses were performed using SAS version 8.1 (SAS Institute, Cary, NC). Statistical significance was defined by p values <0.05.
|
|
| |
| |
|
|
|
