Herpes Zoster and Progression to AIDS in a Cohort of Individuals Who Seroconverted to Human Immunodeficiency Virus

Transcription

1 990 Herpes Zoster and Progression to AIDS in a Cohort of Individuals Who Seroconverted to Human Immunodeficiency Virus Maria Barbara Alliegro, Maria Dorrucci, Patrizio Pezzotti, Giovanni Rezza, Alessandro Sinicco, Mauro Barbanera, Francesco Castelli, Gloria Tarantini, Andrea Petrucci, and the Italian HIV Seroconversion Study* From Centro Operativo AIDS, Istituto Superiore di Sanitd, Rome; Infectious Diseases Clinic, University of Turin, Turin; Infectious Diseases Unit, Livorno Hospital, Livorno; Infectious Diseases Clinic, University of Brescia, Brescia; and Centro Anti-Venereo, Milan, Italy A prospective study, of 1,198 individuals who seroconverted to human immunodeficiency virus (HIV) was conducted to estimate the incidence and determinants of herpes zoster and to determine whether herpes zoster can accelerate the progression to AIDS. Herpes zoster was diagnosed for 48 individuals (4%). After adjusting for the CD4 cell count, individuals acquiring HIV infection through sexual contact were more likely to have herpes zoster than were injection drug users (relative hazard, 1.50). The crude relative hazard of AIDS for individuals who had herpes zoster compared with those without herpes zoster was 2.44; the adjusted relative hazard was After adjusting for the CD4 cell count, fever was the only specific characteristic of herpes zoster that was significantly associated with a more rapid progression to AIDS (relative hazard, 6.52). Data suggest that herpes zoster occurs more frequently in individuals acquiring HIV infection through sexual transmission. There was no evidence that herpes zoster per se is an independent cofactor of progression of HIV disease, although febrile episodes of herpes zoster may predict a faster progression to AIDS. The risk of herpes zoster has been found to be approximately nine times greater in HIV-infected individuals than in HIVseronegative individuals [1]. However, studies on the relationship between herpes zoster and the duration of HIV infection have produced conflicting results. Some studies have identified herpes zoster as an early manifestation of HIV infection [2-4], while other studies have found it to be a late manifestation [5]. Recent findings imply that the first episode of herpes zoster may occur at any time during the period of HIV infection at different CD4 T lymphocyte counts [6]. For this reason, it has been suggested that herpes zoster would not be a good indicator of imminent AIDS [7]. It is still not clear whether the risk of AIDS increases after an episode of herpes zoster. Some investigators found that herpes zoster was a predictor of a faster progression to AIDS [8, 9], while other investigators did not [10-14]. However, most of these studies had serious limitations: either they were HIV seroprevalence studies or the analyses were not adjusted for the CD4 cell count, which is the most important determinant of progression of HIV disease. Recently, after adjusting for the Received 27 February 1996; revised 24 June Financial support: Ministero della Sanita, Istituto Superiore di Sanita, AIDS Research Project (378). * Italian HIV Seroconversion Study members are listed at the end of the text. Reprints or correspondence: Dr. Maria B. Alliegro, Centro Operativo AIDS, Istituto Superiore di Sanita, Viale Regina Elena 299, 00161, Rome, Italy. Clinical Infectious Diseases 1996; 23: by The University of Chicago. All rights reserved /96/ $02.00 CD4 cell count, a study of homosexual men found that the occurrence of herpes zoster was not independently associated with a faster rate of progression to AIDS [15]. Using data for a large cohort of HIV-infected males and females with different exposure histories for whom dates of seroconversion were accurately estimated, we conducted a study of the incidence and determinants of herpes zoster. After adjusting for the CD4 cell count, we also evaluated if the risk of progression to AIDS is greater in HIV-infected individuals with herpes zoster than in those without herpes zoster. As a result of this evaluation, the prognostic value of different clinical manifestations of herpes zoster was also determined. Study Population and Methods We enrolled 1,198 HIV-seroconverted (a documented HIV test with negative results followed by.an HIV test with positive results within a maximum of 2 years) individuals from 16 outpatient facilities in Italy in the study. The seroconversion date was estimated as the midpoint in time between the last negative HIV test and the first positive HIV test. The study design, which has been previously described in detail [16, 17], is summarized below. All the study participants were requested to return approximately every 6 months for clinical and laboratory evaluation. To compensate for the lack of information for patients who had not undergone a medical examination in the last 2 years, cross-matching with the Italian National AIDS Surveillance Registry was performed. AIDS was defined in accordance with the 1987 case definition developed by the Centers for Disease

2 CID 1996;23 (November) Herpes Zoster and Progression to AIDS 991 Control and Prevention (CDC) and the World Health Organization [18]. HIV-related pathologies were defined according to the 1986 revised case definitions of the CDC [19]. When herpes zoster was reported, the following information was obtained on an additional form: clinical signs and symptoms accompanying the varicelliform eruption, duration of symptoms, site and number of dermatomes involved, CD4 cell count at the time of the diagnosis, antiviral chemotherapy, complications (pain associated with acute neuritis and postherpetic neuralgia), number of episodes, and whether symptoms were clinically observed or self-reported. Complete detailed clinical data were available for 80% of the individuals with herpes zoster. HIV serological status was ascertained through ELISAs; reactions in sera were confirmed by western blotting. CD4 + cell counts were determined by flow cytometry. Statistical Methods The first episode of herpes zoster and the diagnosis of AIDS were the endpoints of interest. To estimate the cumulative probability of progression to the endpoints, we used the Kaplan- Meier survival techniques [20]. Differences between curves were assessed by using the logrank test. The cumulative probability of progression to the endpoint of interest before reaching a given CD4 lymphocyte count was calculated by means of methods described elsewhere [21]. Crude and adjusted relative hazards of herpes zoster and AIDS were both estimated with use of models for Cox proportional hazards [20]. When the endpoint was herpes zoster, we defined herpes zoster free time as the interval from HIV seroconversion to the date of the first episode of herpes zoster (before AIDS had developed), date of death, or 30 June 1994 (the cutoff date). Possible covariates associated with the development of herpes zoster (age at seroconversion, gender, HIV transmission group, and CD4 cell count) were assessed by means of univariate and bivariate analyses. The CD4 cell count was entered in the models as a time-dependent covariate. When the endpoint was the diagnosis of AIDS, we defined AIDS-free time as the interval from HIV seroconversion to the date of the diagnosis of the first AIDS-defining illness, the date of death not due to AIDS, or 30 June The relative hazard of AIDS for patients who had herpes zoster compared with those who did not was estimated by using a univariate model and then adjusting for possible confounders (age at seroconversion, HIV-related pathologies other than herpes zoster before AIDS, and CD4 cell count). All the variables were considered as time-dependent covariates, except for age at seroconversion. Finally, we studied only individuals with herpes zoster for whom clinical data were complete. In this subanalysis, the date of the first episode of herpes zoster was time zero, and the AIDS-free time was calculated as the interval between the date of the diagnosis of herpes zoster and the date of the diagnosis of AIDS, the date of death not due to AIDS, or the cutoff date. To assess if different clinical manifestations of herpes zoster were associated with an increased risk of AIDS, relative hazards of AIDS associated with single clinical manifestations were estimated after adjusting for the CD4 cell count at the time of the diagnosis of herpes zoster. Results Progression to and Determinants and Characteristics of Herpes Zoster Of the 1,198 HIV-infected patients enrolled in the study, 48 had herpes zoster. The median age of the patients at the time of the diagnosis of herpes zoster was 27 years (range, years). Detailed clinical characteristics of herpes zoster were available for only 38 individuals (79.2%); 30 (78.9%) of these patients were treated with acyclovir during their first episode of herpes zoster. The detailed clinical characteristics were as follows: fever (yes, 31.5% of patients; no, 68.5%), number of dermatomes (1, 57.8%; >1, 42.2%), pain (mild, 44.7%; severe, 55.3%), length of episode (<12 days, 50%; 12 days, 50%), site (face and neck, 18.5%; rest of body, 81.5%), complications (yes, 10.5%; no, 89.5%), and number of episodes (1, 89.5%; >1, 10.5%). The diagnosis of herpes zoster was made directly by clinicians for all individuals but one, who self-reported symptoms of herpes zoster. The cumulative incidence of herpes zoster within 6 years of HIV seroconversion was 6.6% (95% CI, 4.5%-8.2%). The cumulative incidences of herpes zoster within 6 years of seroconversion that were stratified by age at seroconversion, sex, and HIV transmission category are shown in table 1. The crude and adjusted (for the CD4 cell count) relative hazards associated with these potential cofactors are listed in table 2. According to the.univariate model, individuals acquiring HIV infection through sexual transmission were significantly more likely to have herpes zoster than were those who were infected through injection drug use. The adjustment for the CD4 cell count tended to reduce the effect of each cofactor. At the time of the diagnosis of herpes zoster, the median CD4 cell count for the 48 individuals was 340/mm 3 (range, /mm3). The probability of being infected with herpes zoster at different CD4 cell counts is shown in figure 1A. The analogous relationship between the CD4 cell count and the risk of progression to AIDS is shown in figure 1B. At CD4-cell counts of >100/mm 3, the two curves are similar. The risk of AIDS increased more than the risk of herpes zoster. In fact, while the cumulative incidences of AIDS and herpes zoster at a CD4 cell count of 100/mm 3 were 13.2% (95% CI, 8.8%-17.5%) and 10.8% (95% CI, 7.2%-14.3%), respectively, the cumulative probabilities at a CD4 cell count of 1/mm3 were 100.0% (95% CI, 97.2%-100.0%) and 12.6% (95% CI, 8.3%-16.9%), respectively.

3 992 Alliegro et al. CID 1996;23 (November) A B 0.8 o 0.8 L. rsts L 0.4 L. 0.4 L a) -4 U 0.2 L I I I I I 1, CD4 lymphocyte count (imm 3) , CD4 lymphocyte count (/mm3) Figure 1. Kaplan-Meier estimates of cumulative probabilities of herpes zoster (A) and AIDS (B) developing at different CD4 cell counts in a cohort of HIV-infected individuals. Progression to AIDS During the study period, we observed 203 cases of AIDS. The Kaplan-Meier cumulative incidence of AIDS within 6 years of seroconversion was 18.7% (95% CI, 15.8%-21.5%). Seventeen patients with herpes zoster had AIDS. Of the 23 AIDS-defining illnesses in these 17 patients, the following were the most common: esophageal candidiasis, 7 cases (30.4%); wasting syndrome, 4 (17.4%); and Pneumocystis carinii pneumonia, 3 (13.0%). The frequency distribution of the diseases for individuals who had herpes zoster was not significantly different from that for those without herpes zoster. The relative hazard of AIDS for individuals who had herpes zoster compared with those without herpes zoster was 2.44 (95% CI, ). When we adjusted for the CD4 cell count, age at seroconversion, and presence of HIV-related pathologies before AIDS, the relative hazard was 1.08 (95% CI, ) (table 3). Similar results were obtained when the analyses were repeated with the detailed clinical information for only 38 patients. We estimated crude and adjusted relative hazards of AIDS for the 38 individuals for whom detailed clinical information was available by comparing different clinical characteristics of herpes zoster (reported above); time zero was defined as the date of the first episode of herpes zoster. The results of univariate and bivariate analyses are shown in table 4. Univariate analysis revealed that only the relative hazard of AIDS for febrile patients was significantly high (relative hazard, 3.27; 95% CI, ). After adjusting for the CD4 lymphocyte count at the time of the diagnosis of herpes zoster, the relative hazard associated with the presence of fever increased to 6.52 (95% CI, ); this finding suggests that fever was an independent covariate of the CD4 cell count measured at the time of the diagnosis of herpes zoster. Similar results were obtained when other possible confounding variables, such as age at the time of the diagnosis of herpes zoster and use of acyclovir (results not shown), were adjusted for. The effect was also confirmed after the analysis was repeated to compare patients with a febrile

4 CID 1996;23 (November) Herpes Zoster and Progression to AIDS 993 Table 1. Kaplan-Meier estimates of cumulative incidences of herpes zoster among HIV-infected patients within 6 years of seroconversion. Table 3. Adjusted relative hazards of AIDS for HIV-infected persons who had herpes zoster during follow-up. Covariate No. of patients (n = 1,198) Cumulative incidence (%) of herpes zoster Covariate Adjusted relative hazard Age (y) at seroconversion < ( ) ( ) > ( ) Sex Male ( ) Female ( ) HIV transmission category Injection drug use ( ) Sexual ( ) Unknown episode of herpes zoster with all the other participants with or without herpes zoster. The estimated relative hazard of AIDS for individuals with febrile episodes of zoster was 4.98 (95% CI, ); after adjusting for age of seroconversion, HIV-related pathologies before AIDS, and CD4 cell count, the relative hazard remained significantly high (relative hazard, 2.71; 95% CI, ). Discussion It is widely known that herpes zoster occurs more frequently in older persons than in other population groups [22] and in individuals with non-hiv-associated immunosuppression caused by conditions such as lymphoma, Hodgkin's disease, leukemia and other cancers, organ transplants, and chemotherapy. Herpes zoster has also been found to be associated with HIV-induced immunosuppression [8]; however, there is still uncertainty about the cofactors that increase the risk of herpes zoster in HIV-infected Table 2. Crude and adjusted relative hazards of herpes zoster for HIV-infected patients. Covariate Crude relative hazard Adjusted relative hazard* Sex: female vs. male 1.31 ( ) 1.26 ( ) HIV transmission: sexual vs. injection drug use 1.83 ( ) 1.50 ( ) Age at seroconversiont 1.29 ( ) 1.14 ( ) CD4 cell counts 1.29 ( ) NOTE. Individuals with unknown modes of HIV transmission were excluded from this analysis. * Adjusted for the CD4 cell count as a time-dependent covariate. t Values calculated for each increase in age of 10 years. $ Values calculated for each decrease in CD4 cell count of 100. Herpes zoster* 1.08 ( ) Age at seroconversiont 1.17 ( ) CD4 cell counes 1.75 ( ) HIV-related pathologies before AIDS* t 2.58 ( ) * Time-dependent covariates. t Values calculated for each increase in age of 10 years. s Values calculated for each decrease in CD4 cell count of 100. Other than herpes zoster. individuals and whether the occurrence of herpes zoster may accelerate the further course of HIV disease. We prospectively studied 1,198 HIV-infected persons for a median of 4 years after HIV seroconversion and found that the cumulative incidence of herpes zoster within 6 years of seroconversion was 6.6% (95% CI, 4.5%-8.2%). This estimate is lower than that reported in other studies [7, 15]. It has been suggested that herpes zoster may develop when CD4 cell counts are high; the explanation for this suggestion is possible impaired lymphocyte function [1]. We also observed that herpes zoster may occur at high CD4 cell counts; in particular, herpes zoster seems less specifically associated with the level of immunosuppression than with other "minor" opportunistic infections, such as oral candidiasis and hairy leukoplakia. In fact, we found that the risk of herpes zoster in our cohort increased with the decline of the CD4 cell count (table 2), but this risk increased at a slower pace than did the risk of AIDSdefining clinical conditions (table 4). Our data confirm that herpes zoster is not a good independent indicator of a higher risk of AIDS; even after adjusting for the Table 4. Crude and adjusted relative hazards of AIDS for HIVinfected patients with different clinical characteristics of herpes zoster. Characteristic Crude relative hazard Adjusted relative hazard* Fever: yes vs. no 3.27 ( ) 6.52 ( ) No. of dermatomes: >1 vs ( ) 2.50 ( ) Pain: severe vs. mild 1.78 ( ) 2.27 ( ) Length of episode:..^-12 d vs. <12 d 0.45 ( ) 0.41 ( ) Site: face and neck vs. other 0.86 ( ) 0.80 ( ) Treatment with acyclovir: yes vs. no 3.41 ( ) 3.09 ( ) CD4 cell counts 1.50 ( ) * Adjusted for the CD4 cell count at the time of the diagnosis of herpes zoster. t Values calculated for each decrease in CD4 cell count of 100.

5 994 Alliegro et al. CID 1996; 23 (November) CD4 cell count, the risk of AIDS is not modified significantly by the occurrence of herpes zoster. We found that the relative hazard of AIDS for individuals who had at least one episode of herpes zoster was 2.44; this value declined to 1.08 after adjusting for possible confounders, such as CD4 cell count, age at seroconversion, and presence of other HIV-related pathologies before AIDS. These findings are consistent with those of Moss et al. [10] who found that herpes zoster is not an independent predictor of AIDS and with those of a cohort study [15] that showed that herpes zoster is associated with a faster rate of progression to AIDS but that the effect is not independent of the CD4 cell count. We performed a specific analysis of the subgroup of patients with herpes zoster to study the prognostic value of specific clinical manifestations or clinical signs and symptoms of herpes zoster that were associated with progression to AIDS. Melbye et al. [9] suggested that painful zoster, zoster of cranial dermatomes, and repeated episodes of zoster were associated with a poorer outcome of HIV disease, but these researchers did not take into account the immunologic status of persons at the time of the diagnosis of herpes zoster. Our findings suggest that the risk of AIDS was not amplified by the severity of herpes zoster (e.g., a higher number of dermatomes). Only the presence of fever during the first episode was significantly associated with a faster progression to AIDS, even after adjusting for the CD4 cell count. Nevertheless, this finding merits further investigation because of the low statistical power of this subanalysis. There are several limitations in this observational study that must be considered. First, herpes zoster may have not been reported if it occurred between follow-up visits or if the episode was of brief duration and of mild severity. However, since clinicians were asked routinely about transient signs and symptoms, the probability of underreporting should have been low; it should also be mentioned that access to care and treatment is available to all population groups in Italy. Second, the analysis of the prognostic value of clinical characteristics of herpes zoster that were associated with progression to AIDS-defining clinical conditions did not include all the participants who had herpes zoster; only the 38 individuals for whom clinical data were complete (79.2% of 48 persons with herpes zoster) were included in the analysis. However, rates of disease progression did not change after excluding the 10 remaining individuals. For this reason, we are quite confident that the possibility of a selection bias can be excluded. Third, fever during herpes zoster could be a constitutional symptom rather than a manifestation of this mild opportunistic infection. However, the possibility of misclassification was low, since those patients with temperatures of >38 C for >1 month were considered as having an HIV-related disease (which we adjusted for in the analysis). In conclusion, herpes zoster occurs at CD4 cell counts higher than those associated with other minor opportunistic infections; nevertheless, its incidence increases with decreasing CD4 cell counts. In our study, we found a higher incidence of herpes zoster among individuals acquiring HIV infection through sexual transmission, even when adjusting for the CD4 cell count. Whether this finding could be due to a differential exposure to agents that may play a role in reactivating varicella-zoster virus will need to be evaluated further. We found that herpes zoster per se does not influence the clinical outcome for HIV-infected patients, although the presence of fever during the episode may modify the prognosis. References 1. Fessel WJ. Early symptoms of human immunodeficiency virus infection [invited commentary]. Am J Epidemiol 1995; 141: Panda S, Sarkar S, Mandal BK, et al. Epidemic of herpes zoster following HIV epidemic in Manipur, India. J Infect 1994; 28: Colebunders R, Mann JM, Francis H, et al. Herpes zoster in African patients: a clinical predictor of human immunodeficiency virus infection. J Infect Dis 1988; 157: Friedman-Kien AE, Lafleur FL, Gendler E, et al. Herpes zoster: a possible early clinical sign for development of acquired immunodeficiency syndrome in high-risk individuals. J Am Acad Dermatol 1986; 14: Melbye M, Biggar RJ, Ebbesen P, et al. Long-term seropositivity for human T-lymphotropic virus type III in homosexual men without the acquired immunodeficiency syndrome: development of immunologic and clinical abnormalities. A longitudinal study. Ann Intern Med 1986; 104: Holmberg SD, Buchbinder SP, Conley LJ, et al. The spectrum of medical conditions and symptoms before acquired immunodeficiency syndrome in homosexual and bisexual men infected with the human immunodeficiency virus. Am J Epidemiol 1995; 141: Buchbinder SP, Katz MH, Hessol NA, et al. Herpes zoster and human immunodeficiency virus infection. J Infect Dis 1992; 166: van Griensven GJP, de Vroome EMM, de Wolf F, Goudsmit J, Roos M, Coutinho RA. Risk factors for progression of human immunodeficiency virus (HIV) infection among seroconverted and seropositive homosexual men. Am J Epidemiol 1990; 132: Melbye M, Grossman RJ, Goedert JJ, Eyster ME, Biggar RI Risk of AIDS after herpes zoster. Lancet 1987; 1: Moss AR, Bacchetti P, Osmond D, et al. Seropositivity for HIV and the development of AIDS or AIDS related condition: three year follow-up of the San Francisco General Hospital cohort. BMJ 1988;296: Selwyn PA, Alcabes P, Hartel D, et al. Clinical manifestations and predictors of disease progression in drug users with human immunodeficiency virus infection. N Engl J Med 1992; 327: Glesby MJ, Moore RD, Chaisson RE, the Zidovudine Epidemiology Study Group. Herpes zoster in patients with advanced immunodeficiency virus infection treated with zidovudine. J Infect Dis 1993; 168: Rogues A-M, Dupon M, Ladner J, Ragnaud J-M, Pellegrin J-L, Dabis F, Groupe d'epidemiologie Clinique du SIDA en Aquitaine. Herpes zoster and human immunodeficiency virus infection: a cohort study of 101 coinfected patients [letter]. J Infect Dis, 1993; 168: Eyster ME, Rabkin CS, Hilgartner MW, et al. Human immunodeficiency virus-related conditions in children and adults with hemophilia: rates, relationship to CD4 counts, and predictive value. Blood 1993; 81: Veenstra J, Krol A, van Praag RME, et al. Herpes zoster, immunological deterioration and disease progression in HIV-1 infection. AIDS 1995; 9: Rezza G, Lazzarin A, Angarano G, et al. Risk of AIDS in HIV seroconverters: a comparison between intravenous drug users and homosexual males. Eur J Epidemiol 1990; 6:

6 CID 1996;23 (November) Herpes Zoster and Progression to AIDS The Italian Seroconversion Study. Disease progression and early predictors of AIDS in HIV-seroconverted injecting drug users. AIDS 1992; 6: Centers for Disease Control. Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome. MMWR 1987; 36(suppl 1S):3s-15s. 19. Centers for Disease Control. Recommendations for providing dialysis treatment to patients infected with human T-lymphotropic virus type III/lymphadenopathy-associated virus. MMWR 1986; 35: Lee ET. Statistical methods for survival data analysis. 2nd ed. New York: John Wiley & Sons, Phillips AN, Lee CA, Elford J, Janossy G, Kernoff PBA. The cumulative risk of AIDS as the CD4 lymphocyte count declines. J Acquir Immune Defic Syndr 1992; 5: Weksler ME. Immune senescence. Ann Neurol 1994; 35(suppl):S35-7. Italian HIV Seroconversion Study Members A. Cozzi Lepri (Istituto Superiore di Sanitd, Rome); G. Angarano and G. Buccoliero (University of Bari, Bari); M. Zaccarelli (Spallanzani Hospital, Rome); B. Salassa (Amedeo di Savoia Hospital, Turin); L. Tomasoni (University of Brescia, Brescia); E. Ricchi (University of Bologna, Bologna); S. Gafa (S. Maria Hospital, Reggio Emilia); P. Viale (Ospedale Civile, Piacenza); L. Ortona (Catholic University, Rome); U. Tirelli (Centro di Riferimento Oncologico, Aviano); A. Canessa (University of Genoa, Genoa); R. Pristerd (Regional Hospital, Bolzano); A. Lazzarin (San Raffaele Hospital, Milan); and F. Aiuti (First University of Rome, Rome).