Prevalence and Correlates of HIV Infection among Female Sex Workers in Two Mexico-U.S. Border Cities

Abstract

Background

We examined HIV prevalence and correlates among female sex workers (FSWs) in Tijuana and Ciudad Juarez, two large Mexico-U.S. border cities.

Methods

FSWs aged ≥18 years underwent interviews and testing for HIV, syphilis, gonorrhea and Chlamydia. Logistic regression identified factors associated with HIV infection.

Results

Of 924 FSWs, prevalence of HIV, gonorrhea, Chlamydia and syphilis titers >1:8 was 6%, 6.4%, 13% and 14.2%. Factors independently associated with HIV were injecting cocaine (OR= 2.96), smoking/snorting/inhaling methamphetamine (OR=3.32) and having syphilis titers >1:8 (OR= 4.16).

Conclusions

Culturally appropriate interventions are needed to identify and treat ulcerative STIs and reduce HIV risks associated with stimulants among FSWs in the Mexico-U.S. border region.

Introduction

Tijuana and Ciudad Juarez are Mexican cities bordering San Diego, California and El Paso, Texas, respectively. The states of Baja California and Chihuahua, where these cities are located, are second and fourteenth among Mexico's 32 states in terms of cumulative AIDS incidence [1]. Commercial sex is quasi-legal in the Zonas Rojas [red light districts] in most Mexican cities. There are an estimated 9000 female sex workers (FSWs) in Tijuana and 4000 in Ciudad Juarez [2]. Until relatively recently, HIV prevalence among FSWs in Mexico was low, although STI prevalence among Mexican FSWs has been high for decades [3].

Between 2004 and 2005, we studied 155 FSWs in Tijuana and 140 FSWs in Ciudad Juarez and reported an HIV seroprevalence of 4.8% and 4.9%, respectively [2]. During the same period, a study of more than one thousand pregnant women at Tijuana General Hospital reported that HIV prevalence was 1.1% [4], which was suggestive of an emerging epidemic. Herein, we describe HIV prevalence and correlates among a larger sample of FSWs in Tijuana and Ciudad Juarez through January, 2006.

Methods

Study Population

Between January 2004 and January 2006, outreach workers from municipal and community health clinics recruited 924 FSWs in Tijuana and Ciudad Juarez into a behavioral intervention study to increase condom use [2]. Eligibility requirements included age ≥18 years, having traded sex for drugs, money, or goods and reporting unprotected vaginal sex with ≥1 client in the past two months. As the outcome for the behavioral intervention was HIV incidence, women were excluded if they reported knowingly having tested HIV-positive prior to enrollment.

Data Collection

Participants underwent interviewer-administered surveys in Spanish, a blood draw, and cervical swab for which they were compensated $30 US. Interview topics included sociodemographics, sexual and substance use behaviors, working conditions, social influence, mood, self-esteem, social cognitive factors, and self-reported STIs.

HIV antibody was detected using the “Determine”^® rapid HIV antibody test (Abbott Pharmaceuticals, Boston, MA); reactive samples were confirmed by EIA and Western Blot. The rapid plasma reagin (RPR) test was used to detect syphilis antibody (Macro-Vue, Becton Dickenson, Cockeysville, MD, USA); reactive samples were confirmed by Treponema pallidum hemagglutinin assay (TPHA) (Fujirebio, Wilmington, DE, USA). Neisseria gonorrhea and Chlamydia trachomatis were detected from cervical swabs using the Aptima^® Combo-2 collection device (Genprobe, San Diego, CA). Pre- and post-test counseling was conducted; women testing positive were referred to municipal health clinics for free medical care. STI tests and confirmatory HIV tests were conducted at either the San Diego County Health Department (for Tijuana samples) or the El Paso County Health department (for Cd. Juarez samples). Study protocols were reviewed and approved by Institutional Review Boards in San Diego, Tijuana, and Ciudad Juarez.

Statistical Analysis

Statistical analyses compared HIV-positive and HIV-negative FSWs using Wilcoxon Rank Sum or Fisher's Exact tests, where appropriate. Univariate and multivariate logistic regressions were performed to identify factors associated with HIV-positive serostatus. In multivariate regressions, all variables attaining <10% significance were considered from most to least significant; likelihood ratio tests were used to compare nested models, using a 5% significance level for variable inclusion. Pearson and Hosmer and Lemeshow tests were conducted to assess goodness-of-fit. Although site-specific models were examined, results are presented for the overall sample due to the relatively small sample of HIV-positive cases and because site-specific associations were generally similar.

Results

924 eligible FSW were enrolled (474 in Tijuana; 450 in Ciudad Juarez), of whom 55 (6%) tested HIV-positive. Only two false-positive rapid HIV tests were observed. Median age was 32 (Interquartile Range [IQR]: 26, 39), one fifth spoke English (19.7%), and most had minimal education. Most women were married or living common-law; 93.7% had children (Table 1). Compared to HIV-negative FSWs, HIV-positive FSWs were significantly more likely to live in Tijuana (Table 2), to have lived in the study site for longer durations and to live in rented rooms. Over half of HIV-positive FSWs reported a steady partner who was an injection drug user (IDU), compared to a quarter of HIV-negative FSWs (OR=3.59). There was no significant differences between groups in amounts earned for protected and unprotected sex, and proportions describing themselves as a street worker.

Table 1. Characteristics of FSWs with and without HIV infection in Tijuana and Ciudad Juarez.

Baseline Characteristics	HIV Positive N=55	HIV Negative N=869	Total N=924	P-value
Demographics
Median age (IQR)*	32 (27, 38)	33 (26, 40)	32 (26, 39)	0.59
Speaks English	14 (25.5%)	167 (19.3%)	181 (19.7%)	0.29
Born in state where they live	26 (47.3%)	333 (38.5%)	359 (39.0%)	0.20
Highest year of school completed (IQR)	6 (5, 8)	6 (4, 8)	6 (4, 8)	0.61
Married or common-law	12 (22.2%)	210 (24.3%)	222 (24.2%)	0.87
Worked in the sex trade for more than 4 years	31 (58.5%)	400 (46.4%)	431 (47.1%)	0.09
Had children	51 (92.7%)	811 (93.8%)	862 (93.7%)	0.77
Lived in own home	4 (7.3%)	209 (24.1%)	213 (23.1%)	0.003
Lived in Tijuana (vs. Cd. Juarez)	36 (65.5%)	438 (50.4%)	474 (51.3%)	0.04
Years Lived in Study Location	20 (10, 28)	13 (5, 25)	13 (6, 26)	0.02
Social Influence
Steady Partner was an IDU	9 (56.3%)	73 (26.4%)	82 (28.0%)	0.02
Earned ≤ $30 US on average for vaginal sex with a condom	12 (23.1%)	211 (24.6%)	223 (24.5%)	0.87
Earned ≤ $30 US on average for vaginal sex without a condom	22 (40.0%)	366 (42.2%)	388 (42.1%)	0.78
Fellow FSWs injected drugs	28 (50.9%)	362 (41.9%)	390 (42.4%)	0.21
Risk Behaviors
Total # clients past month	17 (6, 36)	30 (10, 66)	28 (10, 64)	0.008
# of non-regular clients	10 (2, 31)	24 (5, 57)	23 (5, 56)	0.005
Mean # unprotected vaginal sex acts with clients (last mo)	4 (2, 10)	6, (3, 12)	6 (3, 12)	0.23
Often or always used condoms for vaginal sex with clients	21 (38.2%)	378 (43.6%)	399 (43.3%)	0.48
Mean # unprotected anal sex acts with clients (last mo)	4 (2, 8)	2 (1, 6)	2 (1, 6)	0.43
Often or always used drugs before or during vaginal sex	20 (36.4%)	111 (12.8%)	131 (14.2%)	<0.001
Had at least one IDU sex partner last month	20 (36.4%)	174 (20.4%)	194 (21.4%)	0.01
Had male clients who have injected drugs	24 (43.6%)	268 (30.9%)	292 (31.7%)	0.05
Ever injected drugs	23 (41.8%)	143 (16.5%)	166 (18.0%)	<0.001
Injected drugs last month	14 (25.5%)	100 (11.5%)	114 (12.3%)	0.005
Injected heroin past month	19 (35.2%)	126 (14.5%)	145 (15.7%)	<0.001
Injected cocaine past month	8 (14.5%)	52 (6.0%)	60 (6.5%)	0.02
Injected speedball past month	11 (20.0%)	56 (6.5%)	67 (7.3%)	0.001
Smoked, inhaled, or snorted methamphetamine past month	33 (60.0%)	225 (25.9%)	258 (27.9%)	<0.001
Ever shared needles or injection equipment	18 (32.7%)	114 (13.1%)	132 (14.3%)	<0.001
HIV/STIs
Syphilis titer ≥1:8	22 (41.5%)	108 (12.5%)	130 (14.2%)	<0.001
Chlamydia	16 (36.4%)	88 (11.7%)	104 (13.0%)	<0.001
Gonorrhea	11 (25.0%)	40 (5.3%)	51 (6.4%)	<0.001

Table 2. Factors associated with FSWs with and without HIV infection in Tijuana and Ciudad Juarez.

Baseline Characteristics	Univariate OR, 95% CI	Adjusted Odds Ratio, 95% CI*
Age (per year increase)	0.99 (0.96, 1.02)	—
Speaks English	1.43 (0.76, 2.68)	—
Born in the state where they now live	1.43 (0.83, 2.47)	—
Highest year of school completed (per year increase)	1.00 (0.91, 1.09)	—
Married or common-law	0.89 (0.46, 1.72)	—
Worked in the sex trade for more than 4 years	1.63 (0.93, 2.86)	—
Had children	0.85 (0.30, 2.44)	—
Lived in own home	0.25 (0.09, 0.69)	—
Lived in Tijuana (vs. Cd. Juarez)	1.86 (1.05, 3.30)	—
Years lived in study location (per year increase)	1.02 (1.00, 1.04)	—
Social Influence
Steady Partner is an IDU	3.59 (1.29,10.00)	—
Earned ≤ $30 US on average for vaginal sex with a condom	0.92 (0.47, 1.79)	—
Earned ≤ $30 US on average for vaginal sex without a condom	0.91 (0.52, 1.59)	—
Fellow FSWs inject drugs	1.44 (0.83, 2.48)	—
Risk Behaviors
Total number of clients past month	0.99 (0.98, 1.00)	—
# of non-regular clients	0.99 (0.98, 1.00)	—
Often or always used condoms for vaginal sex with clients	0.80 (0.46, 1.40)	—
Mean # unprotected anal sex acts with clients (past month)	1.02 (0.99, 1.05)	—
Often or always used drugs before or during vaginal sex	3.89 (2.17, 6.97)	—
Had at least one IDU sex partner last month	2.23 (1.25, 3.95)	—
Has male clients who currently use drugs	1.2 (0.63, 2.27)	—
Has male clients who have injected drugs	1.73 (0.99, 3.00)	—
Ever injected drugs	3.65 (2.07, 6.42)	—
Injected drugs last month	2.63 (1.38, 4.99)	—
Injected heroin past month	3.20 (1.78, 5.77)	—
Injected cocaine past month	2.67 (1.20, 5.95)	2.96 (1.29, 6.80)
Injected speedball past month	3.63 (1.78, 7.40)	—
Smoked, inhaled, or snorted methamphetamine past month	4.29 (2.45, 7.52)	3.32 (1.85, 5.95)
Have ever shared needles or injection equipment	3.22 (1.77, 5.85)	—
HIV/STIs
Syphilis titer ≥1:8	4.95 (2.76, 8.86)	4.16 (2.28, 7.59)
Chlamydia	4.32 (2.25, 8.31)	—
Gonorrhea	5.95 (2.80,12.63)	—

^*Variables in the multivariate model are adjusted for all other variables in the model.

In terms of sexual and drug use behaviors, HIV-positive FSWs had significantly fewer clients in the past month than did HIV-negative FSWs (17 vs. 30, p=0.008: OR=0.99). However, this difference was confined to non-regular (10 vs. 24, p=0.005: OR=0.99) as opposed to regular clients (median 4 vs. 4, p=0.73). In the past month, HIV-positive FSWs were more likely than HIV-negative FSWs to have male clients who injected drugs (43.6% vs. 30.9%, p=0.05: OR=1.73), to have injected drugs themselves (25.5% vs 11.5%, p=0.005: OR=3.65), to have used any illicit drugs, and to have often/always used drugs before or during sex (36.4% vs 12.8%, p<0.001: OR=3.89).

HIV-positive FSWs were more likely to have syphilis titers ≥ 1:8 (41.5% vs. 12.5%, p<0.001: OR=4.95), and test positive for Chlamydia (36.4% vs. 5.3%, p<0.001: OR=4.32), and gonorrhea (25% vs. 5.3%, p<0.001: OR=5.95). High STI prevalence contrasted sharply with FSWs' self-reported STI histories, which were markedly lower and did not differ between groups.

In the final multivariate model (Table 2), three factors were independently associated with HIV-positive serostatus. HIV-positive FSWs had an almost three times greater odds of having injected cocaine in the past month (OR=2.96, CI 1.29, 6.80), having snorted, smoked or inhaled methamphetamine in the past month (OR=3.32, CI 1.85, 5.95), and having a syphilis titer > 1:8 (OR=4.16: CI 2.28, 7.59). These associations were essentially unchanged after adjusting for city of residence.

Discussion

In this large study of FSWs in two U.S.-Mexico border cities, HIV prevalence has increased significantly in the last decade. Applying UNAIDS criteria, the HIV epidemics in both cities appear to have transitioned from low-level to concentrated, since overall HIV prevalence among FSWs is now >5% [5]. FSWs who injected cocaine, smoked, inhaled or snorted methamphetamine, and whose syphilis titers were consistent with active infection were at least three times more likely to be HIV-infected compared to other FSWs. These findings have important implications for cross-border HIV transmission, since both cities adjoin busy ports of entry to the U.S. and are destinations for large numbers of sex tourists.

FSWs who used stimulants had greater odds of HIV infection, even if they did not inject. Injection drug use and sex work have been shown to overlap considerably in various settings [6-9]. An elevated risk of HIV among injectors is not surprising; however, the independent association we observed between HIV infection and non-injection methamphetamine use was unexpected. Methamphetamine use has been closely associated with incident HIV infection among MSM [10], but few reports have examined this association among heterosexuals. In one such study, St. Petersburg, Russia, an independent association was reported between HIV infection and methamphetamine use among heterosexual—and predominantly male—IDUs [11]. Use of stimulants, including methamphetamine and cocaine, has been associated with high risk sexual behavior in FSWs and other populations [12]. A study of pregnant women in Tijuana found that HIV was independently associated with methamphetamine use, even after adjusting for injection drug use [4], which mirrors our findings.

Although we did not observe an independent association between HIV infection and unprotected sex, methamphetamine use may be a proxy measure of high risk sexual behaviors. Since methamphetamine production and use has risen dramatically in Mexico following crackdowns on its production in the U.S. [13], it is important to identify whether the mechanism(s) responsible are directly or indirectly drug-related so that appropriate interventions can be developed. Volkow and colleagues suggested that stimulants induce sexual desire, which in turn could account for an increased frequency of high risk sexual behaviors [14]. On the other hand, stimulant use could be a marker of sensation-seeking or other personal attributes which are associated with higher levels of unprotected sex [10].

The finding that FSWs who inject cocaine were at elevated risk of HIV infection is also cause for concern. These women may acquire HIV through sharing of needles or other injection paraphernalia, since cocaine injection is associated with frequent, more chaotic injection behavior. Alternatively, cocaine injection could be a marker of high risk sexual behaviors, such as exchanging sex for drugs, which can compromise women's ability to negotiate safer sex. In our sample, HIV prevalence among FSWs who inject drugs was 12% compared to 4% among FSWs who had never injected drugs [9], which underscores the need for targeted interventions that reduce both sexual and injection risks in this vulnerable subgroup. Although subsidized drug abuse treatment exists in these Mexican cities, there are limited treatment options for persons addicted to stimulants, as in other countries.

While many variables were significantly associated with HIV serostatus in univariate analyses, some such as Chlamydia and Gonorrhea were not independently associated with HIV infection after controlling for other factors (e.g. syphilis). Indeed, high syphilis titers were independently associated with a four-fold odds of HIV infection, which is not surprising since it is an important cofactors of HIV transmission. At enrollment, few women reported being aware of having an infectious STI, which could suggest that the majority of these infections were chronic and untreated or acute, incident infections. Although free treatment for syphilis and other STIs is available in Mexico under the country's universal health system, limited resources compromise the ability to conduct surveillance and treatment for high risk populations. Recently, Mexico's federal and state Ministries of Health provided at least two mobile HIV/STI health clinics to most Mexican states in an effort to reach high risk populations, an important first step towards a more active case-finding approach. Given the clear and immediate need for improved surveillance and aggressive STI treatment in both cities, binational funding strategies to improve financing and delivery of these services should be undertaken.

Since HIV-positive FSWs were more likely to report that their regular partners and clients injected drugs, there is a concomitant need to develop interventions to reduce high risk behaviors among these male partners. We recently reported that more than two thirds of FSWs in Tijuana and Ciudad Juarez reported having U.S. clients and that this subgroup tended to engage in higher risk behaviors [15], which underscores the necessity for a binational response to preventive interventions among men paying for sex.

The cross-sectional nature of this analysis limited our ability to draw causal inferences. Due to small cell sizes, we may have been under-powered to detect some associations (e.g., unprotected anal sex). Like most studies of sexual behavior, ours is potentially subject to socially desirable responding. Some women may have been aware of having an STI, but were unwilling to report it, fearing that this would impact their ability to practice sex work. However, less than half reported consistent condom use for vaginal sex with clients, and frequency of reported drug use was also high, suggesting that reporting bias may have been minimal. Since subjects were recruited through convenience sampling and eligibility criteria were designed to target high-risk women in need of intervention, our estimates of HIV/STI prevalence should not be generalized beyond the present sample. Indeed, HIV prevalence may be under-estimated, since women who were knowingly HIV-infected were excluded.

Although FSWs in Tijuana and Ciudad Juarez who use stimulants and have high syphilis titers are at increased risk of HIV infection, these risk factors are amenable to interventions. Our data emphasize the need for interventions to identify and treat ulcerative STIs and to reduce the risks associated with injection and non-injection use of stimulants. Multifaceted interventions need to be developed in a binational context to curtail the growing HIV epidemic among Mexican FSWs in these border communities.