Portal de Periódicos da CAPES

Linked HIV Epidemics in San Francisco

1996; Lippincott Williams & Wilkins; Volume: 11; Issue: 3 Linguagem: Inglês

10.1097/00042560-199603010-00016

2331-6993

Susan K. Service, Sally Blower,

HIV, Drug Use, Sexual Risk

To the Editor: In the early 1980s, a first wave of HIV infection swept through the gay community in San Francisco (1). Recently, a second wave of HIV infection has been reported in young gay men in this city (2). In this article, we assess the linkage between these two epidemic waves by analyzing sexual network data; sexual network and sexual mixing pattern data can be used to assess who has had sex with whom (3-6). Previously, we analyzed sexual network data collected from a large random sample of young gay men in San Francisco [these men participated in the San Francisco Young Men's Health Study (SFYMHS)] to determine which age groups have sex with which other age groups (7). We assessed the effects of age group mixing, because the HIV seroprevalence level is extremely age stratified in this community [42% of gay men >30 versus 18% of gay men <30 were infected by 1993 (2,7)]. We demonstrated that sexual partner selection (on the basis of age) is a significant risk factor for HIV infection in San Francisco (7); we used these sexual network data to explain the current seroprevalence patterns and to predict the rate of seroconversion in the gay community (7). In this letter we analyze these sexual network data further to address these questions: (a) To what degree are the two waves of HIV infection in the gay community linked? (b) Is the linkage the result of age preferences or the age distribution of the gay community? We used sexual network data collected in 1993 from the SFYMHS. The SFYMHS is a longitudinal cohort study of HIV transmission among young gay men; the study was begun in 1993 by a household-level random sample of 18 to 29-year-old single men in 21 census tracts in San Francisco (2). Both gay and straight men were interviewed. Gay respondents (n = 428) were asked to report both the number of sex partners experienced in the past 12 months and the age (in term of the dichotomous variable under or over 30 years of age) of each of their sex partners. We used these data to calculate the age distribution of partnerships reported by these young gay men; results are presented in Table 1A. Approximately half of the sexual partnerships of young gay men occurred with men 30 years old. Hence, Table 1A illustrates that the two waves of HIV infection in the gay community were tightly linked. We assessed whether the linkage was the result of age preferences by comparing the observed age distribution of sexual partnerships (Table 1A) with an expected age distribution of sexual partnerships derived under the assumption of no age preference [that is, assuming proportional mixing, with respect to age (3-6)]. If there is no age preference in the choice of sex partner, then partners are chosen only on the basis of the frequency with which the two age groups are sexually active in the community. Therefore, the expected age distribution of partnerships depends upon the age distribution of the gay community and on the average sexual activity level of gay men in each of the two age groups. Hence, the expected proportion of sexual partnerships (ej) in each of the two (j) age categories (under the assumption of no age preference) can be calculated by Eq. 1: where Cj is the average of male sex partners reported by gay men in age group j, Gj is the number of gay men in age group j, and j is the age group ( 30). To derive the expected distributions, we assumed that gay men in San Francisco were effectively closed populations with respect to their selection of sex partners; that is, we assumed that men picked few of their sex partners from outside of San Francisco. This assumption appears plausible for a large gay community in a large metropolitan area. Rubin et al. (8) have discussed the problems of estimating sexual mixing patterns in open populations. In 1993, gay men 30 reported in 1993 (C2) was calculated from the San Francisco Men's Health Study (SFMHS). The SFMHS is an ongoing longitudinal cohort study of HIV that was begun in 1984 by a household-based random probability sample of single men in 19 of the same census tracts in San Francisco as the SFYMHS; by 1993, the men in the SFMHS were all >30 years old (n = 466). From the SFMHS unpublished data, we calculated that C2 = 9.6. To calculate G1 and G2 [the numbers of gay men in the two age groups ( 30)], we used census data to obtain the number of men (straight and gay) in these 21 census tracts, and then we multiplied these estimates in the two age groups by the proportion of men that were gay. Census data revealed that 12,310 men were 30 in these 21 census tracts in 1990. The proportion in 1993 of the men 30 in these census tracts that are gay do not exist; therefore we used 1984 data to estimate this proportion and then derived upper and lower bounds for the estimate. In 1984, at the initiation of the SFMHS (which randomly sampled straight and gay men), 80% of the 1,043 men interviewed were gay. While 80% is a valid estimate for 1984, several factors (such as migration and AIDS-induced mortality) could alter the proportion of gay men by 1993; these factors could have either increased or decreased the portion of older gay men in this area. Therefore we used 0.80 as the average estimate of the proportion of men >30 that were gay, and we bracketed this estimate with upper and lower bounds. We used 0.40 (identical to the proportion of men 30 was estimated at 26,214 (0.80*32,891); upper and lower bounds for G2 were 32,891 and 13,156, respectively. The calculated expected partnership age distribution (under the assumption of no age preference) is shown in Table 1B. By comparing the observed partnership age distribution (Table 1A) with the expected partnership distribution derived under the assumption of no age preference (Table 1B), it may be seen that young gay men are in fact exhibiting a great deal of selectivity with respect to age in their selection of sex partners. Under the assumption of no age preference, only 17% [or 29% (using the lower bound estimate for G2) or 14% (using the upper bound estimate for G2)] of the sexual partnerships of young gay men would have occurred with men <30 years old (Table 1B); however, 51% of the partnerships actually occurred with men <30 years old (Table 1A). Hence, the probability of the selection of a sex partner 30 years old. Consequently, in spite of the age preferences, approximately half of the sexual partnerships of young gay men occurred with the older (and more heavily infected) men (Table 1A). Therefore, the age distribution of the gay community has ensured that the two waves of HIV infection are tightly linked. Because our analyses are based upon data collected from large random samples of gay men, our results have external validity and hence have significant implications for the transmission of HIV in San Francisco (7). A very high proportion of the young (<30) gay community (18%) is already infected with the virus (2). We have shown that age preferences occur and that, although these age preferences are protective, the age distribution in the gay community has ensured that the two waves of HIV infection have been tightly linked. Consequently, older gay men have been and continue to be a very important source of HIV infection of young gay men in San Francisco. Acknowledgment: We thank the respondents of the SFMHS and the SFYMHS for their participation and NIDA (grant 1R29DA08153) and NIAID (grant AI33831) for financial support. We also thank Dennis Osmond and Warren Winklestein for providing access to the data and for comments on earlier drafts of the manuscript. *Susan K. Service; †S. M. Blower *Neurogenetics Laboratory; Center for Neurobiology and Psychiatry; Department of Psychiatry; University of California-San Francisco; †Department of Epidemiology and Biostatistics; University of California-San Francisco; San Francisco General Hospital; San Francisco, California, U.S.A.REFERENCES 1. Winklestein W, Lyman DM, Padian N, et al. Sexual practices and risk of infection by the human immunodeficiency virus: the San Francisco Men's Health Study. JAMA 1987;257:321-5. Cited Here | PubMed | CrossRef 2. Osmond D, Page K, Wiley J, et al. Human immunodeficiency virus infection in homosexual/bisexual men, ages 18-29: the San Francisco Young Men's Health Study. Am J Public Health 1994;84:1933-7. Cited Here | PubMed | CrossRef 3. Blower SM, Hartel D, Dowlatabadi H, Anderson RM, May RM. Drugs, sex and HIV: a mathematical model for New York City. Philas Trans R Soc Lond B 1991;321:171-87. Cited Here 4. Blower SM. Behaviour changes and stabilization of seroprevalence levels, correlation or causation? J Acquir Immune Defic Syndr 1991;4:920-3. Cited Here | PubMed 5. Blower SM, McLean AR. Mixing ecology and epidemiology Proc R Soc Lond B 1991;245:187-92. Cited Here 6. Anderson RM, May RM. Infectious diseases of humans: dynamics and control. Oxford: Oxford University Press, 1991. Cited Here 7. Service SK, Blower SM. HIV transmission in sexual networks: an empirical analysis. Proc R Soc Lond B 1995;260:237-44. Cited Here | PubMed | CrossRef 8. Rubin G, Umbach D, Shyu S, Castillo-Chavez C. Using mark-recapture methodology to estimate the size of a population at risk for sexually transmitted diseases. Stat Med 1992;11:1533-49. Cited Here | PubMed | CrossRef © Lippincott-Raven Publishers.