Figueroa, J. P.*; Ward, E.*; Morris, J.†; Brathwaite, A. R.*; Peruga, A.‡; Blattner, W.§; Vermund, S. H.∥; Hayes, R.†
Human immunodeficiency virus (HIV) and human T-cell lymphotropic virus type I (HTLV-I) infections are sexually transmitted(1-3). Persons who attend sexually transmitted disease (STD) clinics appear to be at higher risk of HIV(4-7) and HTLV-I (8) infection than groups more representative of the general population because of a higher rate of partner exchange and risk behavior and because of an increased risk of transmission associated with other STDs(9,10).
A cross-sectional survey was conducted at the Comprehensive Health Centre (CHC), Kingston, Jamaica, from November 1990 through January 1991. Among 1001 heterosexuals presenting with a new STD complaint, HIV prevalence was 3.7% for men and 1.9% for women(11), and HTLV-I prevalence was 7.0% for men and 7.9% for women(8). Between 1986 and 1991, there was a tenfold rise in HIV seroprevalence among persons attending the CHC for a new STD complaint(11). In contrast, HTLV-I prevalence in 1986 was 3.3% for men and 9.0% for women(12).
A follow-up study was conducted among STD patients surveyed in 1990 and 1991 to estimate the incidence of HIV and HTLV-I infection and to identify risk factors for transmission.
A follow-up study to estimate HIV and HTLV-I incidence among STD clinic attenders was conducted at the CHC, the only facility offering specialized services for STDs in Kingston, Jamaica. In the cross-sectional study of November 1990 through January 1991, 1006 consecutive patients attending for a new STD complaint were studied. Two persons refused to participate, and a small but undetermined number of patients were missed during enrollment because of the large number presenting on certain days. Patients were informed individually of the study, gave written consent, had blood taken, and completed an oral 15-to 20-minute questionnaire concerning sociodemographic variables, medical and STD history, knowledge of AIDS, and sexual behavior. Each participant was given a physical examination, including pertinent genital or pelvic examinations. After counseling on HIV and STD risk reduction, participants were treated and given an appointment to return for their laboratory results and review of their condition.
Of the 1006 patients studied, 5 homosexual men were excluded, as were those who were HIV seropositive. A total of 970 heterosexual patients who were HIV negative were enrolled in the cohort. From January 1992 through July 1993, an attempt was made to reexamine all 970 persons initially enrolled. This was done in two ways. Persons returning to the STD clinic of their own accord during the follow-up period were recruited into the study (i.e., passive recruitment). Study subjects not returning to the STD clinic of their own accord were recruited through field visits to their homes and asked to attend CHC for the follow-up study (i.e., active recruitment). It was not appropriate to attempt to interview or take blood from subjects at their homes, because it would create too much curiosity within the household and among neighbors and could give rise to resentment or hostility.
A minimum of three field visits were made to each subject's home before considering the person as lost to follow-up, except for cases in which a false name or address was given by the STD patient. A follow-up questionnaire, including STD history since the initial study visit, a physical examination, and STD diagnostic tests were done and blood taken. These procedures were done at the CHC. Patients were treated if indicated, and all were counseled on HIV and STD prevention and condom use. STD diagnostic tests were performed according to standard protocols(13) and are described elsewhere (11). Patients were tested for HTLV-I using Cambridge Biotech enzyme-linked immunosorbent assay (ELISA; Cambridge Biotech. Worcester, MA, U.S.A.) and tested for HIV-1 and HIV-2 using the Abbott ELISA (Abbott, Chicago, IL, U.S.A.). Positive results were confirmed by Western blot(Cambridge Biotech). Bands corresponding to p19, rgp46I, and p21e needed to be present on Western Blot to confirm a test as HTLV-I positive.
To estimate incidence rates, the person years of observation were computed individually for each patient as the time between the cross-sectional survey and the time of their follow-up examination. Rates were estimated as numbers of seroconversions divided by person years of observation. Age-adjusted rate ratios for individual risk factors were calculated using Poisson regression with person-time as a rate multiplier(14).
Completeness of Follow-up
Of the 970 STD clinic attenders enrolled at baseline, 710 (73%) persons were reexamined at follow-up (Table 1). The median length of follow-up time was 552 days(range 118 to 1070 days). A significantly higher proportion of women than men were seen at follow-up (79% versus 67%, p < 0.0001). Eighty percent of persons had to be recruited actively.
The main reason for nonparticipation in the follow-up study was failure of study subjects to attend the clinic after being contacted, which accounted for 17.3% of the target population or 65% of those not recruited (Table 2). These persons were visited at home by field officers on at least three occasions. Most of these persons said that they would come into the clinic but did not. Apparently, 70 persons gave erroneous information concerning their addresses. Four persons had died (not of AIDS), and 15 had migrated. Only 3 persons explicitly refused to participate in the follow-up study.
There were two significant ways in which nonattenders differed from attenders with respect to risk factors identified at the initial cross-sectional survey. Men not attending for follow-up were significantly more likely to have 11 or more lifetime sex partners (p < 0.001) and to report bruising during sex (p = 0.02). There was no significant difference between male attenders and nonattenders with respect to age, self-perception of risk, sex with a prostitute, condom use, history of STD or genital ulcer, or positive serology by treponema-microhemagglutination assay (TP-MHA). There was no significant difference between female attenders and nonattenders.
There were differences between persons recruited passively and actively. Passively recruited men reported significantly more episodes of STDs since their initial study visit (p = 0.002), and they reported a higher number of sexual partners in the previous 4 weeks (p < 0.001). Significantly more of them had gonorrhea at the time of their follow-up visit (p < 0.001). They were also more likely to be HIV infected (4.3% versus 1.5%, p = 0.13) or TP-MHA seropositive (p = 0.087), although these differences did not achieve statistical significance. There was no difference in the proportion (3.5%) of men found to be VDRL reactive (≥2) or in the pattern of symptoms for those having a STD since their initial study visit. Condom use, sex with a prostitute since enrollment(12%), and bruising during sex (24% occasionally or more) were all reported similarly for men recruited passively or actively (data not shown).
Among women, those who were passively recruited were significantly more likely to have had an STD since enrollment (p = 0.001), perceive themselves to be at risk of AIDS (p = 0.007), and to be VDRL reactive (≥2; p= 0.03) and TP-MHA seropositive (p = 0.024). There was no significant difference between passively and actively recruited women with respect to the number of sex partners in the previous 4 weeks, condom use, ever being raped (7.7%), or ever accepting money for sex (4.3%)(data not shown).
Estimation of HIV and HTLV-I Incidence
Of 710 study subjects tested on follow-up, 8 were HIV positive: 7 of 333 men and 1 of 377 women. Men were significantly more likely to be HIV infected than women(age-adjusted rate ratio = 8.0, 95% CI = 1.03 to 365, p= 0.029). The ages of the HIV-positive men were spread among all age groups (range 18 to 45 years), and the single HIV-infected woman was 16 years of age. The total period of follow-up of study subjects was 1100.2 person years. The overall HIV incidence rate was estimated at 0.7 (CI = 0.3 to 1.4) per 100 person years(Table 3).
Of the 710 subjects seen at follow-up, 36 men and 38 women were known to be HTLV-I positive at enrollment. They were excluded from further HTLV-I testing, which was conducted on 270 men and 318 women. Of these, 27 men and 21 women were not retested because of insufficient sera or misplaced specimens. Four (1.5%) men and four (1.3%) women were confirmed HTLV-I positive on Western blot. Ten samples that tested positive on ELISA were indeterminate on Western blot and therefore classified as negative in keeping with current practice.
The HTLV-I incidence rate for men and women combined was estimated as 0.9 per 100 person years (CI = 0.4 to 1.7). There was no significant difference between the incidence rates for men (1.0 per 100 person years, CI= 0.3 to 2.5) and women (0.8 per 100 person years, CI= 0.2 to 2.0). There was an association between HTLV-I infection and an age of 30 years or older (chi-square test for trend, p < 0.01). One man was 25 years of age and three were 30 years; one woman was 30 years, two were 31 years, and one was 50 years of age.
Risk Factors for HIV Incidence
Because there was only one HIV incident case among the women, it was not possible to examine risk factors for HIV infection in women. Among men, there were significant associations between HIV infection and drinking alcohol before sex (half the time or more frequently, compared with occasionally or less), using cocaine, the total number of sex partners since the initial study visit, sex with a prostitute since enrollment, ever accepting money for sex, and increasing number of sex partners, on average, per month (Table 4). Having an STD since enrollment and bruising during sex (occasionally or more compared with once or never) were each associated with HIV infection. The identification of genital ulcers on follow-up examination was significantly associated with HIV. This analysis needs to be interpreted with caution in view of the small number of seroconverters.
There was no significant association between HIV incidence and perceived risk of HIV, number of sex partners in the previous 4 weeks, current STDs, laboratory diagnosis of active syphilis or gonorrhea or TP-MHA seropositivity, lack of circumcision, oral sex, anal sex, or condom use. It was not possible to test for independent effects in a logistic regression model because of the small number of HIV incident cases.
Risk factors for HTLV-I transmission could not be studied adequately because of the small number of incident HTLV-I infections. There was no statistically significant association between HTLV-I infection and any of the variables described for men or women or for both sexes combined. No incident cases had coinfection with HIV and HTLV-I.
The achievement of a follow-up rate of 73% for participants examined at enrollment after 2 years appears to be a fair level of recruitment, given the type of study population and the limited resources available. There appear to be two main groups of defaulters: those who gave false addresses and therefore could not be located and those who were contacted but did not attend the clinic. These two groups of defaulters represented 7.2% and 17.3% of the target population, respectively.
With a default rate of 27% of the study subjects, a question arises about how representative the follow-up participants of the study population were as a whole. This is not an easy question to answer. Those defaulters that gave false information are likely to be at greater risk for HIV. The other group does not appear to be at greater risk, but we cannot be sure. Because of passive recruitment of all study subjects returning to the clinic with an STD, we should have recruited many of those who continued to practice high-risk behaviors. However, we do not know how many persons got an STD and went for medical care elsewhere.
On balance, it appears that the study subjects are reasonably representative of the STD study population but that those excluded may have had a somewhat higher risk of HIV. The estimate of HIV incidence is likely to be a fair one, although possibly slightly underestimated. In any case, the HIV incidence rate is consistent with a steady, rather than steep, rise in infection among STD clinic attenders. This by no means should contribute to any form of complacency, because an acceleration in HIV incidence is feasible. Between 1990 and 1993, HIV seroprevalence increased from 0.2% to 0.4% among blood donors, from 0.1% to 0.14% among antenatal clinic attenders, and from 3.1% to 4.2% among STD clinic attenders(7).
Perhaps one of the most striking findings is the significant difference in HIV transmission rates between men and women. This is not surprising given the fact that men in this study were much more at risk than women. There appear to be very few prostitutes in this study, because only 16 (4.3%) women and 9 (2.7%) men at follow-up reported accepting money for sex, and only 3% of women at enrollment, compared with 48% of men, reported having 11 or more lifetime sex partners.
Among 531 female sex workers in Brazzaville, Democratic Republic of the Congo (formerly Kinshasa, Zaire), the HIV incidence was 8.0 per 100 person years(15). This contrasts with an HIV incidence of 1.28 per 100 person years among male army recruits in Thailand(16) and 0.027 among U.S. Army Reserve men (17). The HIV incidence of male STD attenders in Jamaica (1.2 per 100 person years) was similar to that of male army recruits in Thailand.
The finding of genital ulcers being associated with HIV transmission is consistent with the findings of other studies(6,15,18), as is the association of sex with a prostitute(4,19,20). The association between drinking alcohol before sex and HIV transmission probably results from alcohol serving as a marker for other high-risk behaviors. There was no significant association between HIV transmission and lack of circumcision. However, this was not a suitable population to test for this effect, because most of the men (91%) were not circumcised.
There were too few HIV incident cases to test for independent effects. This weakens the ability of the study to identify risk factors for HIV transmission. Given the small number of seroconverters, the results of the analysis need to be interpreted with caution. Patients attending the clinic with repeat STD infections appear to be at high risk of HIV infection and need to be targeted for preventive interventions.
Although the numbers are small and therefore must be interpreted cautiously, it is striking that the HTLV-I incidence rates are similar for men and women. Male-to-female sexual transmission of HTLV-I is considered to be far more efficient than female-to-male transmission, which is described as occurring rarely(3,21). Perhaps the higher risk behavior of the men and the presence of other STDs explain the similar incidence rates for the two sexes. The association between HTLV-I incidence and an age of 30 years or older supports the view that the age-related increase in HTLV-I seroprevalence reported in earlier studies(8,12,22) arises from an accumulation of seroconversions among adults and mainly results from sexual transmission.
The HTLV-I incidence rates in this study appear to be much higher than those found among Japanese residents older than 40 years of age in Kyushu (0.2 per 100 person years)(23) and similar to those for Japanese prostitutes in Fukuoka, Kyushu (0.8% per year)(24), although lower than rates among HTLV-I-discordant married couples(25). The overall incidence rates for HIV and HTLV-I were comparable; 0.7 per 100 person years for HIV and 0.9 for HTLV-I. The incidence rates were also similar for men: 1.4 per 100 person years for HIV and 1.0 for HTLV-I. However, there was a fourfold difference in incidence rates for women: 0.2 per 100 person years for HIV and 0.8 for HTLV-I. In making these comparisons, we need to be cautious because of the small numbers involved. We also need to take account of the default rate of 27%, because we do not know how many of these persons were HIV or HTLV-I positive. Even a few such cases could alter the results considerably.
The current understanding of the epidemiology of HIV and HTLV-I indicate a higher incidence rate for HIV than for HTLV-I should be expected, because HIV is thought to be more infectious than HTLV-I. This was not the case in this study, and the reasons are not clear. The prevalence of HTLV-I infection among STD clinic attenders in Kingston, Jamaica, is higher than that of HIV(11). This factor could conceivably contribute to the comparable incidence rates. The presumed lower transmission probability for HTLV-I may combine with a higher prevalence of HTLV-I in sexual partners to produce similar incidence rates for the two infections. As HIV prevalence rises, the HIV incidence rate can be expected to exceed the HTLV-I rate.
1. Piot P, Kreiss JK, Ndinya-Achola JO, et al. Heterosexual transmission of HIV [editorial review]. AIDS 1987;1:199-206.
2. Mann JM, Chin J. Editorial. N Engl J Med 1988;319:302-3.
3. Kajiyama W, Kashiwagi S, Ikematsu H, et al. Intrafamilial transmission of adult T-cell leukemia virus. J Infect Dis 1986;154:851-7.
4. Cameron DW, Simonsen JN, D'Costa LJ, et al. Female-to-male transmission of HIV-1: risk factors for seroconversion in men. Lancet 1989;2:403-7.
5. Cleghorn FR, Jack N, Murphy JR et al. HIV-1 prevalence and risk factors among sexually transmitted disease clinic attenders in Trinidad. AIDS 1995;9:389-94.
6. Otten MW, Zaidi AA, Peterman TA, et al. High rate of HIV seroconversion among patients attending urban STD clinics. AIDS 1994;8:549-53.
7. Figueroa JP, Brathwaite A, Ward E, et al. The HIV/AIDS epidemic in Jamaica. AIDS 1995;9:761-8.
8. Figueroa JP, Morris J, Brathwaite A, et al. Risk factors for HTLV-I among heterosexual STD clinic attenders. J Acquir Defic Syndr Hum Retrovirol 1995;9:81-8.
9. Aral SO. Heterosexual transmission of HIV: the role of other sexually transmitted infections and behaviour in its epidemiology, prevention and control. Annu Rev Public Health 1993;14-451-67.
10. Wasserheit JN. Epidemiological synergy, interrelationships between human immunodeficiency virus infection and other sexually transmitted diseases. Sex Transm Dis 1992;19:61-77.
11. Figueroa JP, Brathwaite A. Morris J, et al. Rising HIV-1 prevalence in STD clinic attenders in Jamaica: traumatic sex and genital ulcers as risk factors. J Acquir Immune Defic Syndr 1994;7:310-6.
12. Murphy EL, Figueroa JP, Gibbs WN, et al. Sexual transmission of human T-lymphotropic virus (HTLV-I). Ann Intern Med 1989;111:555-60.
13. Wentworth BB, Judson FN. eds. Laboratory methods for the diagnosis of sexually transmitted diseases. Washington, DC: American Public Health Association, 1987.
14.Egret reference manual. Seattle. WA: Statistics and Epidemiology Research Corporation, 1991.
15. Laga M, Alary M, Nzila N, et al. Condom promotion. STD treatment, and declining incidence of HIV-1 infection in Zairian sex workers. Lancet 1994;344:246-8.
16. Carr JK, Sirisopana N, Torugsa K, et al. Incidence of HIV-1 infection among young men in Thailand. J Acquir Immune Defic Syndr 1994;7:1270-5.
17. Cowan DN, Brundage JF, Pomerantz RS. The incidence of HIV infection among men in the US Army Reserve components, 1985-1991. AIDS 1994;8:505-11.
18. Plummer FA, Simonsen JN, Cameron DW, et al. Cofactors in male-female sexual transmission of human immunodeficiency virus type 1. J Infect Dis 1991;163:233-9.
19. Diaz T, Chu SY, Conti L, et al. Risk behaviors of persons with heterosexually acquired HIV in the US: results of a multistate surveillance project. J Acquir Immune Defic Syndr 1994;7:958-63.
20. Mastro TD, Satten GA, Nopkesorn T, et al. Probability of female-to-male transmission of HIV-1 in Thailand. Lancet 1994;343:204-7.
21. Stuver So, Tachibana N, Okayama A, et al. Heterosexual transmission of Human T-cell leukemia/lymphoma virus type I among married couples in southwestern Japan: an initial report from the Miyazaki cohort study. J Infect Dis 1993;167:57-63.
22. Murphy EL, Figueroa JP, Gibbs WN, et al. Human T-lymphotropic virus type I (HTLV-I) seroprevalence in Jamaica: 1. Demographic determinants. Am J Epidemiol 1991;133:1114-24.
23. Takezaki T, Tajima K, Komoda H, Imai J. Incidence of Human T-lymphotropic virus type I seroconversion after age 40 among Japanese residents in an area where the virus is endemic. J Infect Dis 1995;171:559-65.
24. Nakashima K, Kashiwagi S, Kajiyama W, et al. Sexual transmission of human T-lymphotropic virus type I among female prostitutes and among patients with STD in Fukuoka, Kyushu, Japan. Am J Epidemiol 1995;141:305-11.
25. Stuver SO, Mueller NE. Re: "Sexual transmission of human Tlymphotropic virus type I among female prostitutes and among patients with sexually transmitted diseases in Fukoka, Kyushu, Japan" [letter]. Am J Epidemiol 1995;142:1248-9.
© Lippincott-Raven Publishers.