| |
HCV coinfection increases mortality in HIV patients
|
| |
| |
NEW YORK (Reuters Health) - Despite highly active antiretroviral therapy (HAART), it appears that hepatitis C virus (HCV) coinfection significantly increases mortality in patients with HIV infection, researchers report in the November 15th issue of Clinical Infectious Diseases. These findings are in contrast with study results obtained in pre-HAART era.
"The impact of HCV coinfection on HIV survival is controversial due to conflicting data from several large cohort studies," co-author Dr. Jodie L. Guest of the Atlanta VA Medical Center told Reuters Health.
To investigate further, Dr. Guest and colleagues "used our large VA cohort to study the impact of coinfection on survival and found that in the era of HAART and with longer follow-up, coinfection did adversely impact the survival of our patients after controlling for important confounding variables."
The researchers found that in 970 HIV-infected patients seen between January 1997 and May 2001, the prevalence of HCV coinfection was 31.6%. Co-infection was significantly more likely in patients who were older, black or injection drug users.
The results of multivariate analysis indicated that survival from the time of AIDS diagnosis was significantly shorter for coinfected patients (hazard ratio 1.84), "as was time from HIV diagnosis to death (HR 2.47)," the researchers report.
Altogether, Dr. Guest concluded, "it is likely that the varying results seen in the literature are due to cohort and methodological differences."
Clin Infect Dis 2004;39:1507-1513.
Hepatitis C Virus Coinfection Increases Mortality in HIV-Infected Patients in the Highly Active Antiretroviral Therapy Era: Data from the HIV Atlanta VA Cohort Study
CID Nov 15, 2004
Katie B. Anderson,1 Jodie L. Guest,1,3 and David Rimland2,3
1Rollins School of Public Health at Emory University and 2School of Medicine, Emory University, and 3Atlanta Veterans Affairs Medical Center, Atlanta, Georgia
Background. We compared survival among patients coinfected with human immunodeficiency virus (HIV) and hepatitis C virus (HCV) with that among patients infected solely with HIV.
Methods. Descriptive, bivariate, and survival analyses were conducted using data for all HIV-positive patients who were seen during the period of January 1997 through May 2001 in the HIV Atlanta VA Cohort Study (HAVACS) and who had been tested for HCV antibody since 1992 (n = 970).
Results. The prevalence of HCV coinfection was 31.6%, and coinfected patients were significantly more likely to be older, black, and injection drug users. In multivariate analysis, the duration of survival from the time of diagnosis of acquired immunodeficiency syndrome (AIDS) was significantly shortened for HIV-HCV--coinfected patients (hazard ratio [HR], 1.84; 95% confidence interval [CI], 1.09--3.10), as was time from HIV diagnosis to death (HR, 2.47; 95% CI, 1.26--4.82). Recovery of CD4+ cell count from the time of initiation of HAART did not differ significantly by coinfection status.
A total of 73.1% of patients were receiving HAART at some point during follow-up, and 66.6% had AIDS diagnosed. Conclusions. HCV coinfection is common in this HIV-infected population and negatively affects survival from the time of both HIV and AIDS diagnoses, although this is apparently not associated with a difference in CD4+ cell recovery while receiving HAART. These findings differ from those of a previous study that was conducted in this cohort in the pre-HAART era, which found no association between HIV-HCV coinfection and HIV disease progression.
INTRODUCTION
HAART has been associated with a dramatic decrease in mortality caused by HIV infection, but it has complicated clinical management of HIV-infected individuals, which now must include the treatment of an array of chronic diseases. The impact of coinfection with hepatitis C virus (HCV) is a focus of particular concern.
The prevalence of HCV coinfection in HIV-infected individuals is high, ranging from 10% to 40%, and it can be up to 80% in highly exposed cohorts. This high rate of coinfection is likely associated with shared transmission routes, particularly injection drug use and the receipt of blood products or transfusions. The long latency period prior to HCV disease may account for the lack of considerable liver disease in HIV-infected individuals before the introduction of HAART. With improvements in treatment and therapy, HCV infection will continue to be a serious and highly prevalent disease in these longer-living individuals.
Numerous studies document the negative effects of HIV coinfection on the course of HCV disease. The effect of HCV coinfection on HIV disease progression has been extensively investigated, although study findings differ widely in both the magnitude and nature of the reported associations. Several studies have found significantly increased rates of progression to death or of AIDS-defining events in HCV-coinfected patients, whereas others have found no such effect.
The present study is an extension of an earlier evaluation of the same cohort (the HIV Atlanta VA Cohort Study [HAVACS]), which found no association between HCV coinfection and clinical progression of HIV disease. The previous HAVACS study, like many earlier studies, primarily used data from the pre-HAART era. We speculated that many factors, including the hepatotoxicity of HAART regimens and increased survival durations for patients receiving this therapy, may yield different findings.
METHODS
Study population. The study population consisted of all HIV-infected patients who were tested for HCV infection at the Atlanta Veterans Affairs Medical Center (Georgia) since 1992 and who were seen from January 1997 through October 2001. HAART became available in this cohort in mid-1996. This retrospective, clinic-based cohort study included clinical and demographic information collected from January 1982 through October 2001.
HCV infection status was determined using a second-generation EIA. Eighty-eight percent of all patients were tested for HCV antibody during the study period, and a patient was considered to be HCV infected from the date of a positive test result forward. A minority of patients (n = 58) had HCV RNA levels determined, and 91% of these tests yielded positive results. History of hepatitis B virus (HBV) infection was defined as a positive test result either for hepatitis B surface antigen or for antibody to surface antigen.
CD4+ cell counts and HIV load data were extracted for 3 time points: within 6 months of HCV test date, 6 months after HAART was started, and at the last visit before October 2001. CD4+ cell counts were measured by standard 3-color flow cytometry using FACscan (Becton Dickinson), and HIV-RNA loads were determined by PCR (Amplicor RT-PCR).
Date of AIDS diagnosis was defined as the date the patient's condition met either the 1987 or 1993 Centers for Disease Control and Prevention (CDC) definition of AIDS. Antiretroviral (ARV) use was categorized as: (1) HAART for >1 month, (2) history of use of ARVs but not use of HAART, and (3) no history of ARV therapy or of receipt of HAART for <1 month.
The patient's risk factor for HIV infection, race, and age at the time of HCV testing were included in the analysis. The risk factor was defined as injection drug use, compared with other risk factors, and race was defined as black versus white race, with Hispanic ethnicity classified as black or white race as appropriate. No other race groups were represented.
RESULTS
Cohort characteristics. A total of 970 subjects were eligible for inclusion in the study. The prevalence of HCV infection in the cohort was 31.6%. A total of 73.1% of patients were receiving HAART at some point during follow-up, and 66.6% had AIDS diagnosed. Being a man who has sex with men (MSM) was the most prevalent risk factor for HIV infection (48.0%); 22.4% of subjects had injection drug use listed as the primary risk factor. The cohort was predominantly black (73.2%), with a median age of 41.7 years at the date of HCV testing.
HCV-infected patients were more likely to be older, to be of black race, to be intravenous drug users, and to have never received HAART (P < .0001 for all). They were less likely to have tested positive for hepatitis B antigen (P = .043). HCV-infected and HCV-uninfected patients experienced similar recoveries in the CD4+ cell count after initiating HAART (both for those who received short-term HAART and those who received long-term HAART). The HIV load was not found to be significantly different at date of HCV testing. Furthermore, because CD4+ cell counts were more routinely determined in this population and were found to be highly colinear with HIV load, HIV load data were not incorporated into any multivariate models.
Multivariate survival analyses. HIV-HCV--coinfected patients experienced shorter durations of survival from the date of diagnosis of HIV infection than did HIV-infected/HCV-uninfected patients (hazard ratio [HR], 2.47; 95% CI, 1.26--4.82). The use of HAART was independently associated with an improved duration of survival (HR, 0.25; 95% CI, 0.16--0.39), whereas a CD4+ cell count of <50 cells/mm3 at the time of HCV testing was associated with a decreased duration of survival (HR, 3.24; 95% CI, 2.01--5.22). Patients with AIDS and older patients had significantly decreased survival durations during follow-up (HR, 4.98 [95% CI, 2.68--9.26] and 1.79 [95% CI, 1.14--2.82], respectively). Patients coinfected with HCV and HBV had better survival durations (borderline significance). Although HBV-infected persons were less likely to be injection drug users overall, there was no difference in the frequency of injection drug use between HCV-infected/HBV-uninfected individuals and HCV-HBV--coinfected individuals.
HIV-HCV--coinfected patients also experienced significantly decreased durations of survival from the time of AIDS diagnosis (HR, 1.84; 95% CI, 1.09--3.10). The use of HAART improved survival duration (HR, 0.26; 95% CI, 0.16--0.42), whereas a CD4+ cell count of <50 cells/mm3 at the time of HCV testing was associated with a decreased survival duration (HR, 3.39; 95% CI, 2.15--5.35). There was no difference between HCV-infected and HCV-uninfected individuals with regard to risk to progression to AIDS from HIV diagnosis (data not shown).
Immunological response to HAART. HCV infection was not associated with an altered CD4+ cell response to HAART (data not shown). All patients experienced a similar recovery in the CD4+ cell count 6 months after initiation of HAART, regardless of HCV infection status, although injection drug use was independently associated with a decreased CD4+ cell count recovery. The long-term model that considered CD4+ cell count recovery from initiation of HAART to time of the last visit also found no difference.
DISCUSSION
This study addresses the impact of HIV-HCV coinfection on the progression of HIV disease and presents 4 clinically relevant findings. First, the prevalence of HCV coinfection (31.6%) in this cohort is high and is in accordance with the previous HAVACS study. Second, coinfected patients were less likely to have received HAART than were HCV-uninfected patients. Third, coinfection with HCV results in a significantly decreased duration of survival from the times of diagnoses of HIV infection and AIDS. Fourth, HCV coinfection does not affect short-term or long-term recovery of the CD4+ cell count associated with HAART.
These findings demonstrate a shift in the effect of HCV coinfection from the pre-HAART era to the HAART era. A previous evaluation of survival in the HAVACS cohort found that HCV infection had no effect on progression of HIV infection, as did several other longitudinal studies that included data from prior to 1997. Our present findings are supported by other large studies conducted in the HAART era, notably the Swiss Cohort Study, which observed decreased durations of survival from initiation of HAART in HIV-HCV--coinfected patients. In contrast to Greub et al., however, our data suggest that this difference in the durations of survival is not attributable to a compromised CD4+ cell response to treatment. Other recent studies have found no effect of HCV coinfection on survival; Sulkowski et al., for example, reported no significant difference in survival from time of entry into the cohort. It is unclear why the reported impact of HCV coinfection continues to vary, although cohort differences and variability in the time intervals employed for modeling may be important factors.
Possible reasons for the observed difference between the pre-HAART era and the HAART era include the following: (1) the observed decrease in the duration of survival is because of the introduction of HAART and treatment-associated hepatotoxicity; (2) a larger proportion of patients were tested for HCV in the HAART era, correcting a selection bias (88% vs. 53% previously); and (3) a longer duration of follow-up time allowed for the development of significant hepatic disease. The latter is unlikely, because there was no significant change in mortality rates for liver disease after 1997 (D.R., unpublished data).
This study has several strengths. The HAVACS cohort is large (n = 970) and includes a long follow-up period, and routine HCV testing has been done since 1992. Patients who use the Atlanta Veterans Affairs Medical Center's services receive the majority of their care within the system, decreasing the number of losses to follow-up and standardizing clinical data. Finally, the use of date of AIDS diagnosis—a clinically defined time point—as a starting point for survival analysis adds strength to these findings.
One limitation of this study may be its lack of generalizability, because the HAVACS cohort consists of veterans who are primarily black and male. The subset of patients coinfected with HCV is further distinguished from the general population, because they are more likely than HIV-infected/HCV-uninfected patients to be injection drug users. The growing epidemic of HIV-HCV coinfection in the developing world is primarily attributable to injection drug use, however, and we believe that this cohort provides an accurate reflection of current demographic and clinical trends regarding this dual infection. This study is also limited by the unavailability of important data regarding liver disease, including HCV genotype, alcohol use, and clinical data, such as aspartate aminotransferase and alanine aminotransferase levels, which were not routinely and reliably available for most patients. However, only 20 patients had liver disease--related deaths in this period (5 HCV-uninfected patients and 15 HCV-infected patients). Although efforts were made to control for important confounders, it is possible that some variables remain unavoidably biased. The patient's risk factor for HIV infection was self-reported, and injection drug users may be less likely than others to report their risk factor. It is possible that the 1-month cutoff for dichotomized HAART use disguises differences in the duration of and adherence to HAART therapy between HCV-infected and HCV-uninfected individuals, which could also be linked to a decrease in the survival duration. The median duration of HAART use was slightly lower for HCV-infected patients than for HCV-uninfected patients (593 vs. 741 days), but the causality of this difference cannot be determined; it may be that the decreased survival of HCV-infected patients is driving the decreased use of HAART, or vice versa.
The nature of clinical progression in HIV-HCV--coinfected individuals must be better understood, such that safe and effective treatment and preventive measures may be employed. Although HIV-HCV--coinfected patients were less likely than those infected only with HIV to receive HAART, their CD4+ cell count recovery with treatment was as strong as that observed for HCV-uninfected patients. This suggests that HIV-HCV--coinfected patients may be treated successfully for HIV infection. The high prevalence of HCV coinfection in this cohort and the accelerated progression to death underscore the importance of this secondary infection, now and into the future.
|
|
| |
| |
|
|
|
