CELENTANO, DAVID D. ScD, MHS*,†; NELSON, KENRAD E. MD†; SUPRASERT, SOMBOON RN§; EIUMTRAKUL, SAKOL MD∥; KUNTOLBUTRA, SURINDA MS†; BEYRER, CHRIS MD, MPH†; ZENILMAN, JONATHAN MD†; WRIGHT, NICHOLAS H. MD, MPH¶; THEETRANONT, CHOTI MD#
INTEREST IN THE EPIDEMIOLOGY of sexually transmitted diseases (STDs) in the developing world has recently been increased since the hypothesis has been advanced that STDs may be important cofactors for human immunodeficiency virus (HIV) infection. Both ulcerative (particularly chancroid) and nonulcerative STDs have been implicated as risk factors for the sexual transmission of HIV in the developing and the developed world.1–4 However, it is not apparent whether STDs are important risk factors for the transmission of HIV, markers of high‐risk sexual exposures, or represent confounding due to similar sexual behaviors associated with the incidence of each type of infection, or whether all three of these associations may occur.5 Wasserheit6 has reviewed the various methodologic issues in this “epidemiologic synergy:” HIV infection is itself an STD; STDs may be proxy indicators for high‐risk sexual behavior (hence, behavior itself must be analytically controlled for in the analysis associations between STDs and HIV); many STD patients are multiply infected, and the link with HIV‐1 is frequently not clear; STDs may facilitate HIV transmission in some circumstances or may be markers for immunosuppression; and STDs and individuals who acquire them are heterogenous. Therefore, it may be analytically difficult to associate specific STDs with the HIV infection.
There are numerous reports in the literature that support the association between HIV infection and several different STDs.6 However, when the reports are limited to those that include multivariate analysis in prospective studies (or nested case‐control studies in cohorts), the data are more limited. There is evidence from prospective studies, which have controlled for sexual behavior, that genital ulcer disease is associated with moderately higher HIV‐1 infection among both men and women.4,7,8 The evidence linking syphilis9,10 and chlamydial infection4,8 to HIV is less conclusive. The association between HIV incidence and genital herpes10–12 or gonorrhea4,9,10,12 is inconsistent. Insufficient evidence exists to determine whether trichomoniasis and hepatitis B virus infection are risk factors for HIV‐1 infection.6 Because most bacterial STDs are curable, their identification and control may be an important part of the strategy for HIV‐1 prevention.4 Epidemiologic studies of the association of HIV and other STDs in populations from Southeast Asia are somewhat limited.13
Because STD control is an important component of the acquired immune deficiency syndrome control program in Thailand, understanding the epidemiology of STDs may offer important insights for HIV/acquired immune deficiency syndrome prevention activities. As part of a prospective study of the epidemiology of HIV‐1 infection in northern Thailand, we evaluated the incidence of STDs in young men over a 2‐year period and have analyzed epidemiologic risk factors for incident STD infection in this population.
Subjects and Methods
Young men (n = 2,417) who were 19 to 23 years old and had been conscripted by lottery into the Royal Thai Army and Royal Thai Air Force from six uppernorth provinces were enrolled in a longitudinal study of HIV infection in May 1991 (n = 1,236) and November 1991 (n = 1,181). Details of the methodology have been reported elsewhere.14 Briefly, participants were evaluated by interview and serology at 6‐ month intervals between their induction and discharge from the military 2 years later. Highly educated men, skilled workers, and those in poor health often are deferred from military service. Sexual orientation, current drug use, and HIV infection are not grounds for exclusion from military service. Two hundred eighty‐nine (12.0%) of the conscripts had antibody to HIV‐1, and 2,128 (88.0%) were seronegative at enrollment. The men were interviewed on at least 1 semiannual follow‐up visit at 1 of 13 bases in northern Thailand. Men on border patrol or stationed in the capital were not available for routine testing, and some conscripts missed some follow‐up visits because of remote assignments. The average length of follow‐up for this report was 22 months.
Within a few weeks of induction, conscripts completed face‐to‐face interviews conducted by trained fifth‐year male medical students from Chiang Mai University. They were supervised and monitored by experienced University field staff. Beginning with the second follow‐up visit, interviews were performed by trained paramedics with the same supervisory team. All interviews were field edited and reviewed for quality control. Supervisory staff noted no difference in the quality of the data between the two groups of interviewers. The questionnaire asked about socioeconomic background (education and marital status); lifetime and recent sexual history (wives, girlfriends, and commercial sex workers [CSWs]); use of condoms with different sexual partners; medical history, including information concerning hospitalizations, illnesses, and blood transfusion and donation; alcohol and drug (licit and illicit) use; and STD history. Because the men reported infrequent use of most illicit drugs, we combined all illicit drug use into one category for analysis.
Sexually Transmitted Diseases Determination
We determined the incidence of sexually transmitted diseases at interval visits by physician or other medical practitioner diagnosis or by symptom reports, including gonorrhea, syphilis, nongonococcal urethritis, chancroid, genital warts, and herpes. The STD history was determined by a clinical description (signs and symptoms) of the infection, supplemented by pictures, from an STD atlas, asked for by the medical interviewer. Descriptive terms were used for infections such as chancroid and chlamydia, which do not have precise definitions in Thai. Chancroid history was elicited by asking if subjects experienced “a painful genital ulcer,” and chlamydia was assessed by asking subjects about “penile discharge without pus.” Thus, subjects were asked about specific diagnoses (e.g., “have you ever had gonorrhea?”) and about clinical symptoms (e.g., “have you ever had a penile discharge with pus?”), and responses were combined in the analysis. The diagnoses included syphilis, gonorrhea, herpes, and genital warts. The clinical symptoms included penile discharge with and without pus, dysuria, inguinal lymphadenopathy, genital ulcers with and without pain, and vesicular ulcers. Confirmation by culture for the reported STDs was not performed. The time parameter for each interview included lifetime history at baseline, and interval (6‐month) experience at each subsequent follow‐up. Because genital herpes reports may represent new infections or recurrent episodes and warts may be inadequately treated or recurrent, we did not include them in prospective analysis. However, they were included in analyses that controlled for lifetime STD history. Interviews were conducted confidentially, and data were removed from the military bases after completion of the interview. Research protocols were approved by institutional review boards of the Royal Thai Army, Chiang Mai University, and The Johns Hopkins University.
The data were analyzed using chi‐square, Mantel‐Haenzel chi‐square test for trend, and relative risk ratios with 95% confidence intervals to examine the associations of risk factors and incident STD infections. We used the Generalized Estimating Equation15 technique to determine the adjusted odds ratios associated with new infection in multivariate analysis. This method allows time‐dependent measures from longitudinal studies to be incorporated into the estimation procedure, thereby accounting for autocorrelation among responses. All risk factors that attained significance were incorporated initially into logistic regression, and a backward elimination process was used to construct a reduced model. The Statistical Analysis Software, version 6.04 (SAS Institute, Cary, NC), and a Statistical Analysis Software macro program for GEE (Generalized Estimating Equations, Karim MR, The Johns Hopkins University) were used for the data analysis. A total of 2,417 persons were observed for 3,574 person‐years of follow‐up.
A total of 609 individuals reported incident STDs of all types in the cohort, a rate of 17.04 per 100 person‐years (Table 1). Gonorrhea was the leading STD syndrome (313 cases), with an incidence rate of 8.8 per 100 person‐years. Chancroid was the second most common STD syndrome reported (290 cases), with an incidence rate of 8.1 per 100 person‐years. Nongonococcal urethritis was reported (139 cases), with an incidence rate of 3.9 per 100 person‐years. Syphilis was reported less frequently (66 cases), with an incidence rate of 1.8 per 100 person‐years. The STD rates declined significantly over time, dropping from a rate of 21.1 per 100 person‐years in the first 6 months to 10.4 per 100 person‐years by the fourth follow‐up interval. This declining trend in all STDs also was seen for several of the four STDs assessed. The most precipitous decline occurred between the third and fourth follow‐up visits. Because the incidence of each of the four STDs was relatively low, we combined them for subsequent analyses.
Risk Factors for Infection of Sexually Transmitted Disease
Among the demographic risk factors for STD infection assessed (Table 2), having only compulsory primary school education (reported by 70% of the conscripts) carried with it a modest increase in risk. Being single was marginally associated with incident STDs, although most (69.4%) men at age 21 were not married. Among substances used by these men, cigarette smoking, frequent alcohol use, and illicit drug use (primarily by the oral or inhalation route) were each associated with an increased risk of acquiring an STD. Among sexual risk factors, reporting sexual intercourse with a wife was associated with a decreased risk of STDs, whereas sexual intercourse with girlfriends had a slight increased risk. A threefold increased risk of STD infection was noted for men who reported any CSW visits during military service.
We explored a variety of factors associated with visits to CSWs in relation to STD incidence, as CSW intercourse was the most common sexual exposure in this cohort (reported by 73.4% of the men). There was a significant trend (Mantel‐Haenzel chi‐square = 147.3, P < 0.001) for increased STD infection incidence associated with higher frequency of visits to CSWs. Risk of STD infection did not vary by whether commercial sex visits occurred when the respondent was alone or in a group, although those who reported sex with a CSW had a fivefold risk of STDs compared with men who refrained from visiting CSWs. Drinking at the time of brothel visits was not associated with an increased risk of STD infection compared with nondrinkers. Finally, the cost of the most recent commercial sex visit showed no association with STD infection, although there was a trend for higher STD incidence with lower price.
Because type of sexual partner is related to risk of STD infection, we determined the independent and joint effects of multiple sexual partners (Table 3). For men who reported no CSW visits, sex with a wife was not a risk for STDs compared with men who were unmarried. The STD incidence is identical among married and unmarried men who reported visits to CSWs (12.8 per 100 person‐years for each). Among men who reported not visiting brothels, STD incidence was twofold higher among men who reported sex with girlfriends compared with men who did not report sex with partners. Among men who reported visiting CSWs, having sex with a girlfriend was not an additional risk for STD infection.
Condom Use and Sexually Transmitted Disease Infection
The efficacy of condoms in preventing STD infection was assessed independently for each type of sexual partner reported by the conscripts. Condom use was relatively infrequent with wives (4.9%) and girlfriends (7.8%), and STD infection rates were not associated with frequency of condom use with these partners. The most frequent sex partner reported was a CSW. The largest proportion of brothel visitors (72%) reported using a condom every time they had sex; their STD rate was 2.3 times that of men with no brothel visits. Men who reported inconsistent condom use (18%) and men who reported never using a condom in brothels (10%) had significantly increased reports of incident STDs compared with men who did not go to brothels (Table 2).
A series of multivariate models were used to examine the effect of all risk factors that were significant (P < 0.05; two‐tailed) in the bivariate analysis. We then reduced these models until only risk factors that remained significant were retained (with the exception of sex with wives or girlfriends). Because STD history before conscription was a strong predictor of incident STDs during military service, we stratified the sample on the multivariate analysis by the prior history of an STD.
Overall, risk factors for STD incidence were similar among men with and without a prior STD history before conscription (Table 4). Any CSW visit was associated with an increased risk for incident STDs after adjustment for all other factors included in the model. However, risk associated with CSW encounters also was a function of condom use. Compared with men who reported no CSW visits, consistent condom users had a nearly twofold risk of STD infection. Men who reported inconsistent condom use had a rate of STDs 4 to 5 times higher than that of nonusers of CSWs. Men who reported never using condoms had the highest incidence rates of STDs, regardless of their preconscription STD history, although not appreciably higher than that of inconsistent condom users. Finally, men who visited CSWs frequently (monthly or more often) had a threefold increase in rate of incident STDs if there was a prior STD history; there was no influence of CSW visit frequency on incident STDs among men with no STD history.
In both models, sex with a wife was not associated with incident STDs, and only a slight increase was noted in STD rates among men with no prior STD history who reported sex with girlfriends. Substance use also was associated with more frequent reports of STDs. Illicit drug use (primarily amphetamines and inhalants) was associated with a nearly twofold increase in the incidence of STDs, regardless of prior STD history. Acquisition of STD infections was more common among those men who smoked cigarettes and who had low education if they reported no prior STD history. Neither of these factors was significant among those men with a history of STD.
After controlling for prior history of STDs (both samples combined), visits to CSWs, lack of consistent condom use, and frequency of CSW visits were each associated with increased rates of incident STDs, as was illicit drug use, cigarette smoking, and low education. Reports of sex with wives or girlfriends were not associated with the rate of incident STDs when CSW visits were controlled for in the analysis.
Neisseria gonorrhoeae and chancroid were the leading STDs experienced by these young men. Gonorrhea is the most frequent STD treated at governmental STD clinics in northern Thailand.16 Nongonococcal urethritis is intermediate in frequency, and relatively low rates of syphilis and genital warts are seen. Nongonococcal urethritis, probably representing Chlamydia trachomatis, is also common and often may be inappropriately managed in this setting, where self‐treatment is quite common. Self‐reports of nongonococcal urethritis also may be underreported, especially in the absence of confirmatory cultures.
The incidence of STDs in this cohort, both overall and for specific infections, declined nearly 60% over the 2 years of observation. This may reflect decreasing risk, improvements in condom use, increased self‐treatment, or declines in self‐reports. One might hypothesize that reduced STD exposure might be associated with a parallel decline in the rate of incident HIV infection if STDs are an important cofactor for HIV transmission. However, we observed divergent trends in the incidence of STDs (significant decline) and HIV (no trend in incidence) in this population. A substantial decline in the rates of STDs treated at government‐run clinics in Thailand during the past few years has been reported.16 It is likely that this decline in STDs reflects in part the “100% Condom Campaign” directed at commercial sex establishments. However, part of the decline also could relate to self‐treatment or prophylaxis with antibiotics in response to STD symptoms that are never documented in official statistics17 or differences in the populations of potential clients with STDs seeking care at government clinics. If a substantial part of the decline is not because of the increased and effective use of condoms but rather because of differences in health care‐seeking behavior or the use of antibiotics, then one might see a reduction in STD rates, whereas HIV‐1 incidence remains stable. Also, differences in the incubation period of bacterial STDs and HIV might lead to differences in the temporal trends of incidence after an intervention. Increasing HIV‐1 prevalence or incidence among CSWs infected with HIV also could serve to maintain high rates of HIV transmission despite decreases in the incidence of other STDs.
There are several limitations to these data that must be considered. First, we were unable to obtain confirmation by culture of the reported STDs. Although we know of no motivations to underreport or overreport theses events, there may be biases in self‐reports of specific syndromes. Infections that are frequently asymptomatic (e.g., chlamydia) may be underenumerated using this method. This would lead to an underestimation of the true incidence of STDs in this population. Self‐medication in the presence of initial symptoms also might lead to reduced reports of acute infections, especially if they occurred early in the period of recall. There is also the potential for problems with the specificity of particular diagnoses (e.g., chlamydia may be reported as gonorrhea or chancroid may be reported when the patient has herpes). We attempted to overcome this potential bias by grouping all reported infections for analyses of incident STDs. Second, the change in interviewers from civilian medical students to military paramedics also may have affected the quality of the data. Onsite supervision and intensive training of both teams were instituted to overcome this potential bias. Third, whereas the overall response rate was quite high, 10% to 20% of the subjects could not be interviewed at each data collection period because of duty assignments or other reasons. Although the method used maximizes the use of available data, there are missing visits that could affect the incidence rates reported.
Substance use was associated with STD incidence in both bivariate and multivariate analysis. Although alcohol use was a significant predictor of STD infections in bivariate analysis, it was not significant in multivariate models. Drug use has been associated with “disinhibition” and more frequent high‐risk behaviors in a variety of studies of sexual behavior,18–20 reflecting increased exposure to high‐risk partners and/or to diminished condom use, especially in brothels.21,22 Both drinking and drug use are common among subjects in our population, who are predominately of rural, agrarian backgrounds and who have little formal education. The control of both drugs and alcohol consumption before sexual activity might reduce the incidence of STDs and HIV infection in this population.
The principal exposure that results in STDs among these young men is sexual intercourse with CSWs. In bivariate analysis, there was a dose‐response association between frequency of CSW visits and STD infection. In addition, frequent CSW visits was a significant predictor of STDs in multivariate analysis. Brothel patronage is widespread among young Thai men, and frequency of visits did not decrease substantially during their period of military service.23 However, reported condom use did increase over time. Unprotected intercourse with lower cost brothel‐based CSWs (who usually charge approximately US $4, or approximately 1 day's wage for a laborer or farmer) is associated with a high risk for STD exposure.
Condom use appears to be associated with reduced risk of STDs among men who frequent brothels. Although rates of condom use among men with wives and girlfriends are low, condoms may be used for family planning purposes or if the man fears exposing his partner to an STD (because of his brothel patronage). With CSWs, condoms appear to offer significant protection from STDs, although they are only protective with consistent use.
Sexual intercourse with wives was not associated with STD acquisition by these young men. However, sex with a girlfriend was associated with an increased incidence of STDs among men who had not visited a CSW. Condoms are infrequently used during sex with wives or girlfriends, but their use with these partners was not associated with incident STDs. However, wives and other regular partners of these men who have previously visited CSWs are at risk for STDs. Regardless of marital status, men who went to brothels were at higher risk of acquiring STDs.
Statistical models of the heterosexual spread of HIV are based on assumptions concerning STD rates and their distribution in populations.24,25 The “core group” hypothesis suggests that a relatively small group of highly infected individuals, such as CSWs and injection drug users, expose multiple partners to STDs. In populations such as young men in Thailand where there is a high degree of exposure to these core groups (through frequent visits to brothel‐based CSWs), HIV has spread quite rapidly, as was seen for HIV infection among men attending STD clinics in northern Thailand.13 Thai women who are exposed to men with STDs acquired through patronage of CSWs may become infected. Over time, despite serial monogamy (as is generally the practice among Thai women), there will be a growing equalization in the numbers of men and women who have STDs and HIV. In Thailand, the male:female ratio of acquired immune deficiency syndrome cases has dropped from 17:1 in 1986 to 3:1 in 1992.26 Anderson27 has shown that the rate of partner change (rather than multiple partners, per se) is one of the critical components in projecting the growth rate of HIV epidemics.
The STD control efforts have frequently been focused on the detection and treatment of the core group. However, once the STD epidemic has moved into the larger, general population, as has now occurred in Thailand, this approach becomes less effective. Women who are not involved in any perceived risk activities (e.g., multiple partners, injection drug use) may be at risk of STD infection solely from their partner's behavior. Intervention programs focused solely on risk groups could miss the general population at risk for acquiring STDs; in developing countries, female CSW visits among the general population of young men remain common and carry with them high risk.
The Thai STD control program is based on aggressive detection and management of STDs in a integrated and decentralized system of STD clinics. This program has been met with some success in lowering the incidence of STDs.16 Brothel‐based CSWs often receive periodic STD examinations in STD clinics operated by the Ministry of Public Health, and quarterly HIV tests are performed. The MOPH designed the “100% Condom Campaign” in northern Thailand, an effort to improve condom use in the brothel setting. Reminders to use condoms are displayed in brothels throughout the north, and owner‐operators have been instructed to provide condoms (provided by the MOPH) to customers at no charge. Enforcement of this program to control STD and HIV transmission is critically important. However, as the epidemic of HIV increases, control efforts will have to involve the general public as well.
1. Stamm W, Handsfield H, Rompalo AM, et al. The association between genital ulcer disease and acquisition of HIV infection in homosexual men. JAMA 1988; 260:1429–1433.
2. Simonsen JN, Plummer FA, Ngugi EN, et al. HIV infection among lower socioeconomic strata prostitutes in Nairobi. AIDS 1990; 4:139–144.
3. Greenblatt RM, Lukehart SA, Plummer FA, et al. Genital ulceration as a risk factor for human immunodeficiency virus infection. AIDS 1988; 2:47–50.
4. Laga M, Manoka A, Kivuvu M, et al. Non-ulcerative sexually transmitted diseases as risk factors for HIV-1 transmission in women: results from a cohort study. AIDS 1993; 7:95–102.
5. Mertens TE, Hayes RJ, Smith PG. Epidemiologic methods to study the interaction between HIV infection and other sexually transmitted diseases. AIDS 1990; 4:57–65.
6. Wasserheit JN. Epidemiological synergy: interrelationships between human immunodeficiency virus infection and other sexually transmitted diseases. Sex Trans Dis 1992; 19:61–77.
7. Cameron DW, Lourdes JD, Gregory MM, et al. Female to male transmission of human immunodeficiency type 1: risk factors for seroconversion in men. Lancet 1989; 2:403–407.
8. Plummer FA, Simonsen JN, Cameron DW, et al. Co-factors in male-female transmission of HIV. J Infect Dis 1991; 163:233–239.
9. Darrow WW, Echenberg DF, Jaffe HW, et al. Risk factors for human immunodeficiency virus (HIV) infections in homosexual men. Am J Public Health 1987; 77:479–483.
10. VanRaden M, Kaslow R, Kingsley L, et al. The role of ulcerative genital diseases in promoting acquisition of HIV-1 by homosexual men. Fifth International AIDS Conference; 1989; Montreal, Canada. Abstract no. Th.A.O.17.
11. Holmberg SD, Stewart JA, Gerber AR, et al. Prior herpes simplex virus type 2 infection as a risk factor for HIV infection. JAMA 1988; 259:1048–1050.
12. Kingsley LA, Armstrong J, Rahman A, et al. No association between herpes simplex type-2 seropositivity or anogenital lesions and HIV seroconversion among heterosexual men. J Acq Imm Def Syn 1990; 3:773–339.
13. Weniger BG, Limpakarnjanarat K, Ungchusak K, et.al. the epidemiology of HIV infection and AIDS in Thailand. AIDS 1991; 5(suppl 2):S71–S85.
14. Nelson KE, Celentano DD, Suprasert S, et al. Risk factors for human immunodeficiency virus infection among young adult males in northern Thailand. JAMA 1993; 270:955–960.
15. Liang K-Y, Zeger S. Longitudinal data analysis using generalized linear models. Biometrika 1986:73:13–22.
16. Hanenberg RS, Rojanapithayakorn W, Kunasol P, Sokal DC. Impact of Thailand's HIV-control programme as indicated by the decline of sexually transmitted diseases. Lancet 1994; 344:243–245.
17. Kamboonruang C, Beyrer C, Natpratan C, et al. HIV infection and self-treatment for sexually transmitted diseases among northern Thai men. Sex Trans Dis 1996; in press.
18. Zenilman JM, Hook EW, Shepherd M, Smith P, Rompalo AM, Celentano DD. Alcohol and other substance use in STD clinic patients: Relationships with STDs and prevalent HIV infection. Sex Trans Dis 1994; 21:220–225.
19. Weinstock HS, Liondan C, Bolan G, Kegeles SM, Hearst N. Factors associated with condom use in a high-risk heterosexual population. Sex Trans Dis 1993; 20:14–20.
20. Clapper RL, Lipsitt LP. A retrospective study of risk-taking and alcohol-mediated unprotected intercourse. J Substance Abuse 1991; 391–96.
21. Celentano DD, Akarasewi P, Sussman L, et al. HIV-1 infection among lower class commercial sex workers in Chiang Mai, Thailand. AIDS 1994; 8:533–537.
22. Celentano DD, Nelson KE, Suprasert S, et al. Behavioral and sociodemographic risks for frequent visits to commercial sex workers among northern Thai men. AIDS 1993; 7:1647–51.
23. Celentano DD, Nelson K, Suprasert S, et al. Risk factors for HIV-1 seroconversion among young men in northern Thailand. JAMA 1996; 275;122–127.
24. Laga M. Epidemiology of HIV in type I and type II countries. Abstracts of the 6th International Conference on AIDS; June 20–24, 1990; San Francisco, California.
25. Bongaarts J. Modeling the spread of HIV and the demographic impacts of AIDS in Africa. Center for Policy Studies. Working Paper No. 140. New York: Population Council, 1988.
26. Dunlop D. The cost of AIDS in Asia. AIDS in Asia Newsletter. No. 2. Washington, DC: World Bank, 1994.
27. Anderson RM. Mathematical and statistical studies of the epidemiology of HIV. AIDS 1989; 3:333–346.