| |
The effect of hepatitis C on progression to AIDS before and after highly active antiretroviral therapy
|
| |
| |
AIDS: Volume 18(17) 19 November 2004
Dorrucci, Mariaa; Valdarchi, Catiaa; Suligoi, Barbaraa; Zaccarelli, Maurob; Sinicco, Alessandroc; Giuliani, Massimod; Vlahov, Davide; Pezzotti, Patrizioa; Rezza, Giovannia; and the Italian HIV Seroconversion Study
From the aCentro Operativo AIDS, Dipartimento di Malattie Infettive, Parassitarie e Immunomediate, Istituto Superiore di Sanità, Rome, the bIstituto Nazionale di Malattie Infettive, Lazzaro Spallanzani, IRCCS, Rome, the cOspedale Amedeo di Savoia, Torino, the dIstituto Dermatologico S. Maria e S. Gallicano (IRCCS), Rome, Italy and the eNew York Academy of Medicine, New York; Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA. *See Appendix.
Abstract
Objectives: To assess the effect of infection with hepatitis C virus (HCV) on the progression of human immunodeficiency virus (HIV) disease, before and after the introduction of highly active antiretroviral therapy (HAART).
Methods: We used data from a multi-centre prospective study of HIV seroconverters. Survival analyses were performed to compare the progression to AIDS by HCV serostatus in the period before HAART (i.e. June 1991-May 1996) and in the HAART era (i.e. June 1996-June 2001), controlling for duration of HIV infection.
Results: Among the 1052 persons enrolled, 595 (56.6%) were co-infected; the median follow-up time was 9.7 years. Adjusting for demographic variables (age at HIV seroconversion and gender), HCV infection had no effect on the progression to AIDS in the pre-HAART era [relative hazard (RH) = 0.84; 95% confidence interval (CI), 0.63-1.11], whereas it increased the risk in the HAART era (RH = 1.77; 95% CI, 1.15-2.73). In the HAART era, the proportion of person-time spent on HAART out of the total time at risk was significantly lower among co-infected persons (30 versus 40% for non-co-infected persons; P-value = 0.001); no significant difference was found for dual-therapy (29 versus 25%, respectively; P-value = 0.205); a significant difference was found for mono-therapy (15 versus 8%, respectively; P-value < 0.001).
Conclusions: HCV infection was not a determinant of HIV disease progression in the pre-HAART era, whereas since the introduction of HAART, co-infected individuals seem to have had a faster disease progression. This may in part be explained by differences in person-time spent on different antiretroviral regimens.
Author Discussion
To the best of our knowledge, this is the first study to consider the duration of HIV infection in evaluating the effect of HIV-HCV co-infection on HIV disease progression, allowing an unbiased comparison of co-infected and non-co-infected individuals to be performed before and after the introduction of HAART. When comparing disease progression in the two groups, we found an effect of calendar period, which, as in other studies, was used as a surrogate of the use of HAART. Whereas we found no association between co-infection and disease progression before HAART, we did observe an association in the HAART era. The one other study that controlled for the major determinants of HIV progression and considered both the pre-HAART period and the HAART era found no association in either period. However, the study was performed on a prevalent cohort (i.e. no accurate estimate of the duration of HIV infection was available) and with a short follow-up, especially in the HAART era. Other studies on co-infection did not take into account both periods, and most studies performed before HAART showed no association between co-infection and HIV disease progression.
The results obtained in the HAART era are conflicting. In two European studies, the probability of progressing to AIDS was significantly higher for co-infected persons, although only persons undergoing HAART were considered and neither study was able to totally attribute the increased probability of progressing to AIDS to a reduced responsiveness to HAART. These results are in contrast with those of Sulkowsky et al., who found no difference in the risk of death or of developing AIDS when comparing co-infected and non-co-infected persons in the HAART era.
Our finding of a lower proportion of person-time spent on HAART for co-infected individuals is consistent with the shorter duration of HAART for co-infected individuals reported by Sulkowsky et al., who, however, reported that the differences in the management of patients in terms of HAART did not influence the progression to AIDS. By contrast, in the Swiss cohort, when adjusting for type of HAART, the differences in survival between co-infected and non-co-infected persons remained.
In our study, in the HAART era, the proportion of person-time spent on dual-therapy was similar between co-infected and non-co-infected individuals, whereas the proportion spent on mono-therapy was higher for co-infected individuals. Interestingly, among co-infected individuals, the proportion of person-time spent on dual-therapy was similar to that spent on HAART. Although to the best of our knowledge no data are available for comparisons, we can speculate that the hepatotoxicity of HAART may have discouraged clinicians from prescribing it for persons already suffering from liver damage. Moreover, according to international guidelines, HCV infection should be treated before HIV infection induces a strong immunosuppression, which would delay the start of HAART.
Before drawing conclusions, several limitations should be considered. We assumed that HCV seroconversion pre-dated HIV seroconversion, which may have caused us to underestimate the association between HCV positivity and the risk of AIDS. In other words, it is possible that the effect of HCV seroconversion occurring during the course of HIV disease could be associated with a poorer prognosis. However, most of the patients routinely tested for HCV were HCV-positive when the first test was performed, supporting the validity of our assumption. In addition, a US study has shown that most short-term IDUs became infected with HCV before becoming infected with HIV.
Furthermore, most of the co-infected individuals were IDUs, whereas men who have sex with men represented most of the HCV-negative individuals, which may have affected the association between HCV positivity and the risk of AIDS in the HAART era. For example, the shorter time on HAART for the co-infected group may reflect the tendency of IDUs to have a lower compliance with drug therapy. In other words, the association may have been due to confounding, as opposed to a biological impact of HCV, which is supported by the observation that there was no difference in the pre-HAART period. Finally, we were not able to determine whether the effect observed in the era of HAART was actually due to a poorer adherence to HAART of IDUs, as reported in several studies, and/or to the hepatotoxicity due to antiretroviral therapy.
The results of this study confirm that HCV was not a determinant of progression of HIV disease in the pre-HAART era. However, co-infection appears to have had an effect since the introduction of HAART, although it remains to be determined whether this effect is directly due to HCV co-infection (i.e. a different response to HAART) or to differences in HAART use.
Introduction
Since the introduction of highly active antiretroviral therapy (HAART) and the consequent increases in survival, hepatitis C virus (HCV) has become an emerging problem among persons infected with HIV. In both the United States and Western Europe, an estimated 30% of HIV-infected persons are co-infected with HCV. Co-infection is quite common among transfusion recipients, and prevalence rates as high as 70-90% have been reported among injecting drug users (IDU). Co-infection is apparently uncommon among persons with sexually transmitted HIV infection.
The effect of HCV infection on the progression of HIV disease remains controversial, as does the effect of HAART on this relationship. Most studies conducted before HAART found that HCV had no effect on HIV disease progression, although one study reported a more rapid progression among co-infected persons. The results obtained since the introduction of HAART are contradictory. According to the Swiss Cohort Study, HCV infection accelerates HIV disease progression and death, whereas a US study found no difference in the risk of developing a new AIDS-defining illness between co-infected and non-co-infected persons. However, both studies were based on prevalent cohorts (i.e. the date of HIV seroconversion was unknown) and probably suffered from potential biases, given that, since HAART was introduced, the effect of co-infection could be confounded by the survival bias introduced by persons who have survived without developing AIDS long enough to receive HAART.
We assessed the clinical progression of HIV disease in the Italian Seroconversion Study before and since the introduction of HAART, comparing persons who were co-infected with HCV to those who were HCV-negative, controlling for the duration of HIV infection.
Materials and methods
Study population
The Italian Seroconversion Study is an open, multi-centre cohort of HIV-seroconverters. Participants are recruited in 18 clinical centres located throughout Italy. The main inclusion criterion is a documented negative HIV test result followed by a positive result within 24 months; the date of seroconversion is estimated as the midpoint in time between the dates of the last negative test and the first positive test. We restricted the present analysis to individuals with known HCV serostatus.
Data collection
Clinical and laboratory data were collected approximately every 6 months. For persons lost to follow-up, we performed cross-matching with Italy's National AIDS Registry to determine whether the person had developed AIDS and to identify the AIDS-defining disease. Vital status was also checked with the census bureau of the town of residence and/or birth.
RESULTS
General characteristics of the study population
The study included 1052 persons (942 in the pre-HAART period and 795 in the HAART era). The date of the test for HCV was known for 46% of the study participants.
In both periods, the co-infected participants were younger (P < 0.001) and more likely to be IDUs (P < 0.001). The median CD4 cell count at baseline (i.e. measured within 1 year of entering the period being considered) was similar between the two groups (P = 0.290 in the pre-HAART period; P = 0.536 in the HAART era). In the pre-HAART period, the proportion of persons who developed AIDS was higher, although not significantly (P = 0.180), among the HCV-negative participants, whereas in the HAART era, it was significantly higher among the co-infected participants (P = 0.009).
In the pre-HAART period, the proportion of person-time spent on mono-therapy (out of the total time at risk) was 33% for the co-infected participants and 36% for the other group (P-value = 0.566); the proportion of person-time on dual therapy was 2 and 3%, respectively (P-value = 0.766). In the HAART era, the proportion of person-time spent was 16 and 8% (P-value < 0.001), respectively, for mono-therapy; 29 and 26% (P-value = 0.205) for dual therapy; and 31 and 41% (P = 0.001) for HAART.
Progression to AIDS
Since the co-infected participants were younger than the non-co-infected participants, we estimated the AIDS hazard functions adjusting for age at HIV-seroconversion (i.e. considering 30-year-olds). In the pre-HAART period, the co-infected participants had a slower progression to AIDS, although not significantly so (P = 0.278). In the HAART era, although the hazard of AIDS greatly decreased for both groups, progression was faster among the co-infected participants (P = 0.033).
In the pre-HAART period, when applying the multiple Cox model including the effects of HCV, age at seroconversion, and gender, the relative hazard (RH) of AIDS was 0.84 [95% confidence interval (CI), 0.63-1.11] for co-infected versus non-co-infected individuals, whereas in the HAART era it was 1.77 (95% CI, 1.15-2.73). Of the variables included in the model, before HAART, only age at HIV seroconversion was associated with a faster progression to AIDS, with a higher risk for persons over 30 (RH = 1.62; 95% CI, 1.19-2.19), whereas in the HAART era no association was found for age (RH = 1.01; 95% CI, 0.64-1.59).
|
|
| |
| |
|
|
|
