Introduction

Sex workers often have a disproportionately high prevalence of HIV infection and they, along with their clients, are considered a core group contributing to the transmission of HIV in many countries [1, 2]. Several countries with low HIV prevalence have shown rapid increases in HIV infection among female sex workers (FSW) well before similar increases are seen in the general population [3]. HIV infection among FSW is most often attributed to sexual risk behaviors through unprotected sexual intercourse with multiple partners [2]. Only recently has the overlap between sex work and injection drug use been linked to growing HIV epidemics, but this is generally described in Eastern Europe and Asian countries [4].

The Republic of Mauritius (2010 pop. 1.3 million), an island nation located in the Indian Ocean some 2,000 km off the Kenyan coast is considered part of sub-Saharan Africa. A middle income country, Mauritius provides residents a high standard of education and health care, largely supported by growth in the banking, manufacturing and tourism industries. Despite the favorable economic and social climate, Mauritius is experiencing an HIV and Hepatitis C (HCV) epidemic principally concentrated in its large population of people who inject drugs (PWID). Whereas in 2002 injection drug use accounted for 14 % of all new HIV infections in Mauritius, this percentage increased dramatically to 73 % in 2010 (Personal communication, Ministry of Health and Quality of Life). A recent study of 511 PWID [5] using respondent driven sampling (RDS) found HIV prevalence to be 47.4 %, Hepatitis C (HCV) to be 97.3 % and HCV among HIV seropositive participants to be 89 %. Compared to males who injected drugs, females who injected drugs were found to be four times more likely to be HIV seropositive. A rapid assessment conducted in 2004 in Mauritius estimated that 74 % of FSW inject drugs and that 25 % share needles [6].

Although much of the HIV epidemic in Mauritius is largely concentrated in PWID, there have been indications that heterosexual HIV transmission is increasing. HIV prevalence in the general population is just under 1 % (0.97 %; around 12,000 among 15–49 year-olds) [7]. Among all cases of HIV in the country, heterosexual HIV transmission has increased from 5.9 % in 2005 to 21.7 % in 2008 and the majority of infected females in 2009 in Mauritius have been identified as housewives (44 %) or FSW (41 %) [7, 8].

HIV and other infections prevalence and associated risk behaviors of FSW vary widely in the region. One study conducted in 2007 in Zanzibar using RDS, found HIV prevalence to be 10.8 %, syphilis to be 1.3 %, Hepatitis B (HBV) to be 5.1 % and HCV to be 1.9 % [9]. In other studies of FSW conducted in Eastern Africa (Tanzania, Uganda and Kenya) HIV prevalence was found to range from 19 % to 68 % [10–13] and among the few studies conducted, consistent condom use with clients among FSW ranged from roughly 20 % [14] to 45 % in Uganda, Kenya and Zanzibar [9, 15].

Given the lack of information about HIV related risk practices and prevalence among FSW in Mauritius, the government of Mauritius, conducted a biological-behavioral surveillance survey (BBSS) using respondent-driven sampling (RDS) in 2010.

Methods

Study Data Collection

Eligible persons were females who reported having vaginal, anal or oral sex in the last 6 months with a male in exchange for money or gifts, aged 15 years and older, living in Mauritius and in possession of a valid referral coupon. The small, semi-autonomous sister island of Rodrigues was excluded from the study because of the apparent lack of social network ties of FSW between the islands. Sampling using RDS [16, 17] began with five ‘seeds’ (initial recruits), identified through key contacts and selected to reflect diversity including: geographic residence throughout the island, age, type of sex work (e.g., street based, hotel/bar based, etc.), levels of risk behaviors, age, marital status and education. After the seeds provided informed consent and completed the study requirements (face-to-face interview in Creole, pre-test counseling and a venous blood draw for HIV, HBV, HCV and syphilis testing), they were given three recruitment coupons each to recruit the first wave of participants into the study. Subsequent waves of participants (who completed the study requirements) also received up to three recruitment coupons to recruit additional participants until the calculated sample size of 299 was reached. All participants received 500 Mauritian Rupees (~USD 17.50) for completing the study and 200 Mauritian Rupees (~USD 7.00) for each eligible recruit who consented to participate in the study. Participants were provided their test results, together with post-test counseling, 2 weeks after their enrollment and those who had positive results were referred for treatment and/or for further management.

Biological Testing

Seven milliliters of venous blood, collected from participants by venipuncture, were transferred daily from the study sites to the Virology Department, Central Laboratory, in Port Louis for testing. HIV P24 antigen and HIV1 and HIV2 antibodies were detected using an enzyme immunoassay-Genscreen Ultra HIV Ag-Ab (BIORAD, France). Reactive specimens were confirmed by Western Blot Assay using HIV Blot 2.2 MP Diagnostics (Singapore). HBV surface antigen (HBsAg) was detected using MONOLISA HBsAg Ultra (BIORAD, France) and antibodies to HCV were detected using Murex anti-HCV (ABBOTT, South Africa). Syphilis infection was tested with IMMUTREP TPHA (Omega Diagnostics, UK), a hemagglutination test for treponema pallidum antibodies (IgG and IgM) in serum. Reactive TPHA specimens were tested by IMMUTREP carbon antigen (Omega Diagnostics, UK).

Data Analysis

Estimates and 95 % confidence intervals (CI) were calculated using the multiplicity estimate [18] in RDS Analysis Tool (RDSAT) version 6.0 (Tables 1, 2, 3, 4, 5). Bi and Multivariate logistic regression were performed to identify statistically significant correlates of HIV infection among women who reported not having HIV positive status (Table 6) and injection drug use in the past 3 months (Table 7) using RDSAT-generated weights. Women who reported having HIV positive status from a prior HIV test were excluded to account for the possibility that they practice safer sex practices than FSW who do not know their status and to better assess recent risk behaviors related to condom use. Variables associated with HIV infection or injection drug use at p < 0.2 in the bivariate analysis were included in the initial multivariate model. Variables that remained associated with either dependent variable at a p < 0.05 significance level or were considered important confounders were retained in the model. Adjusted odds ratios (AOR) and 95 % CI were estimated and presented in the final model.

Table 1 Weighted estimators and confidence intervals (CI) for demographic variables among female sex workers, Mauritius, 2010
Full size table
Table 2 Weighted estimators and confidence intervals (CI) for sexual risk variables for female sex workers, Mauritius, 2010
Full size table
Table 3 Weighted estimators and confidence intervals (CI) for substance use and associated services for female sex workers, Mauritius, 2010
Full size table
Table 4 Weighted estimators and confidence intervals (CI) for STI signs and symptoms and HIV testing for female sex workers, Mauritius, 2010
Full size table
Table 5 Prevalence of HIV, HBV, HCV, and Syphilis among FSWs in Mauritius, 2010
Full size table
Table 6 Correlates of HIV sero prevalence among female sex workers who do not know of their positive HIV status, Mauritius, 2010: odds ratios (OR) and adjusted ORs (AOR) with confidence intervals (CI)
Full size table
Table 7 Correlates of injection drug use in the past 3 months among female sex workers, Mauritius, 2010: odds ratios (OR) and adjusted ORs (AOR) with confidence intervals (CI)
Full size table

Ethical Consideration

This study was reviewed and approved by an ethical committee of the Mauritius Ministry of Health. To ensure confidentiality, participants’ coupons, questionnaires, specimens, and test results were identified using a unique study identification number and no personal identifying information was collected.

Results

The study was completed during 2 weeks in August 2010. All five seeds recruited additional participants. The final sample comprised a maximum chain length of eight waves and 299 participants; 210 (including three seeds) enrolled at the Port Louis study location and the remainder (including two seeds) enrolled in the Curepipe study location. Participants were recruited across the two interview location sites, indicating that FSW social networks formed one complete geographic network component, an important assumption in RDS. All variables presented reached convergence in advance of the maximum number of waves, usually by the second or third wave. No participants refused biologic testing.

Demographic Variables

The median age of FSW was 31 years (range: 16–56 years), just under half (48.1 %) reported completing primary education, and 44 % reported being single. All FSW reported receiving money (rather than gifts) in exchange for sex and the vast majority reported that their sole source of income was through sex work (81.9 %); among the 18.1 % who had sources of income in addition to sex work, the primary type of work performed was as a maid or cleaner (60.7 %). Sixty nine percent reported providing support to a median number of three (range: 1–9 persons) other persons.

Sexual History

The median age reported for first exchanging sex for money was 18 years (range: 13–50 years). Just over 40 % of FSW reported having 3 + clients on their last day of work (median: 2, range: 1–15) and more than 60 % reported having 11 + clients in the past 3 months (median: 20, range: 1–300). The majority reported having one non-paying partner in the past 3 months (median: 3, range: 1–25); 86.8 % reported using a condom at last sex with a client whereas only 38.5 % reported using a condom at last sex with a non-paying partner. Just under 60 % of FSW reported always using condoms with either paying or non-paying partners in the past 1 month.

Alcohol and Drug Use Behaviors and Access to Services

The majority of FSW reported drinking alcohol (73.5 %) but almost half (48.2 %) of them reported doing so less than one time a week. In the past 3 months 28.3 % of FSW reported using non-injection drugs. Almost 40 % reported injecting drugs sometime in their lives and 30.5 % of all FSW reported doing so in the previous 3 months. Among FSW who ever injected drugs, 82.5 % did so in the past 3 months and among those, 60 % reported injecting drugs at least once a day. A large proportion reported injecting brown heroin (45.3 %), followed by Subutex® (25.7 %), a narcotic often prescribed for opiate treatment but being obtained and used illegally in Mauritius, and white heroin—La Blanche—(25.6 %) in the past 3 months. Among FSW who ever injected drugs, 17.5 % reported sharing a needle at last injection and 83 % reported being aware of services for PWID. Among those aware of services for PWID, 26.6 % were currently using the government sponsored needle exchange program and 34.7 % were in the methadone treatment program. However, 38.7 % were not involved in any program.

STI Signs and Symptoms and HIV Testing

Almost 30 % of FSW reported experiencing signs/symptoms of a STI in the past 12 months. Sixty percent reported having ever had an HIV test, most of whom did so within the last 6 months and, among those 66.7 % received their results at last testing of which 21.9 % received positive test results.

HIV, Hepatitis, and Syphilis Status

HIV sero prevalence among FSW was 28.9 %; HCV anti-bodies were detected among 43.8 %; and, syphilis anti-bodies were detected among 5.1 %. HBV was not detected. Most (88.2 %) FSW with positive HIV sero status also had anti-bodies to HCV.

HIV Risk

FSW living in common with a partner, reporting injected drugs ever and in the past 3 months and having anti-bodies to HCV had statistically significantly higher odds of being HIV seropositive. Needle sharing could not be assessed in regression analysis as the number of HIV positive FSW who reported sharing needles was too few. In the multivariate model controlling for age and marital status, FSW who were ≥45 years of age, used a condom at last sex with a client, and ever injected drugs had statistically significantly higher odds of being HIV seropositive. FSW who always used condoms in the past 3 months with a paying partner had statistically significant lower odds of being HIV seropositive.

Injection Drug Use Risk

Given the high correlation of injection drug use and HIV, a separate regression analysis was conducted to understand correlates of injection drug use on sociodemographic, sexual and disease variables. In the bivariate analysis, FSW between the ages of 25 and 44 years, ≥2 paying and non-paying partners in the past 3 months, ever tested for HIV and positive for HIV, HCV or Syphilis had statistically significantly higher odds of injecting drugs in the previous 3 months. In the multivariate model controlling for age, FSW who where HIV positive had statistically significantly higher odds and those who used a condom at last sex with a non-paying partner had statistically significantly lower odds of injecting drugs in the past 3 months.

Discussion

The results of this study indicate that just under 30 % of FSW in Mauritius are infected with HIV, over 40 % are infected with HCV and almost all of those infected with HIV are also infected with HCV. High percentages of FSW reported injecting drugs (30.5 % of all participants injected in the past 3 months) and regression analyses found that injection drug use was positively associated with HIV and HCV sero status, as well as Syphilis infection.

The high percentage of FSW infected with both HIV and HCV demands a serious and targeted response. In a BBSS conducted in 2009 in Mauritius, almost 100 % of PWID who had positive HIV sero status also had anti-bodies to HCV [5]. Over the next 10–15 years Mauritius will likely see an epidemic of HCV-fueled liver disease and an unprecedented increase in hepatic linked morbidity and mortality. Although the acquisition of HCV among FSW in Mauritius is most likely due to sharing injection needles/syringes and other equipment, only 17.5 % of injecting FSW reported doing so at their last injection. FSW who injected drugs in the past 3 months were 18 times more likely to have antibodies to HCV in the bivariate regression analysis. Given that over 80 % of FSW who ever injected drugs had done so in the past 3 months, current needle sharing (compared to past or ever needle sharing) may be low due to the intensive harm reduction programs currently available in Mauritius. As the first African country with a government-regulated needle/syringe exchange program, Mauritius currently has 46 locations for syringe exchange (personal communication, Ministry of Health and Quality of Life). In 2009, females made up approximately 10 % or 350 of the 3,500 clients accessing syringe exchange programs. Among the FSW who reported ever injecting drugs, 83 % reported that they were aware of services for PWID and among those who reported injecting in the previous 3 months, 27 % were in the government sponsored needle/syringe exchange program and 35 % were in the methadone treatment program. However, 39 % were not involved in any program. As of May 2011, 232 female injectors were receiving methadone treatment in Mauritius (Personal communication, Ministry of Health and Quality of Life). Possible strategies to increase injecting FSWs’ participation in programs available to PWID include reducing stigma and discrimination by service providers towards FSW, educating service providers about the special needs of FSW, providing more “female friendly” services, increasing the usage of FSW as peer-educators and involving FSW in the planning and provision of services.

The finding that FSW have fairly high rates (87 %) of condom use at last sexual contact with a paying partner is encouraging and in line with median condom use among FSW with the most recent paying partner (84 %) in 56 low and middle income countries in 2008 and 2010 [19] but higher than that found in specific studies conducted in other African countries [9, 14, 15]. Less encouraging is that only 59 % of all FSW always used condoms in the past month with paying partners. However, FSW who always used a condom in the past 1 month with paying partners were 70 % less likely to be HIV seropositive. Also found in the multivariate regression analysis, was that FSW who used a condom at last sex with a client were eight times more likely be HIV seropositive which could indicate that these women know that they are at risk for HIV and either protecting their clients or that they had clients that they knew to be at high risk or that were HIV positive; and that they were trying to protect themselves. However, using condoms at last sex does not necessarily indicate consistent condom use. Requests by clients to not use condoms remains a predominant reason for unprotected sex and many paying partners offer higher payment for sex without condoms [2] highlighting the necessity of designing interventions for FSW which target clients as well.

Similar to findings in other studies of FSW, we found lower percentages of FSW using condoms with their non-paying partners compared to paying partners [2]. Under 40 % of FSW in Mauritius reported using condoms at last sex and under 60 % reported always using condoms in the past month with non-paying partners. These findings are higher than that found in several other African countries including in Kenya in which only 20 % of FSW reported using condoms with their non-paying partners [2].

The prevalence of injection drug use among FSW in Mauritius found in this study was two thirds lower than that found in a rapid assessment conducted among FSW in 2004 [6]. However our quantitative study, compared to a rapid assessment, is more likely to provide reliable results due to the larger sample size and the RDS sampling methodology and analysis. In our study injection drug use prevalence among FSW was much higher than that found in Zanzibar in 2007 (2.8 %) [9] and as high as that found in other East African countries where injection drug use and sex work are reported to overlap significantly [20, 21]. FSW who inject drugs were found to be more likely to have non-paying partners (bi-variate) and less likely to use condoms with non-paying partners (multivariate). More information about whether the non-paying partners of injecting FSW also inject drugs would be useful and warrants further investigation. Many studies have found that women who inject drugs are more likely than men who inject drugs to engage in risky sexual behaviors, sell sex to support their drug use [22] and to inject drugs and use injecting equipment with regular sex partners [23, 24].

Although the majority of FSW have ever been tested for HIV, 33 % did not receive their test result the last time they were tested. Among those who had ever tested for HIV and were seropositive for HIV in this study, 65 % of them received their test results at their last test prior to this study and, among those, 21.9 % reported receiving positive test results. Knowledge about being HIV positive can motivate persons to access education, care and treatment needed to avoid transmission to sexual partners or offspring. This study did not provide rapid testing and participants had to wait 2 weeks for their results, which may account for why only 156 (53 %) of FSW returned to the interview location to receive their test results and referrals. This study used enzyme immunoassays (EIA) and Western Blot to test HIV infection. However, rapid tests, which provide test results-albeit not confirmatory for positive results-within 20 min of testing, coupled with EIA, are as accurate and more cost effective than EIA with Western Blot [25] and should be considered for future BBSS.

The questionnaire was designed to measure indicators of sexual, rather than parenteral HIV transmission and omitted numerous important questions that would have been useful for better understanding injection drug use among this population. Given that injection drug use is so prevalent among FSW, it will be important to add more measures of injection practices, including needle sharing history and drug use behaviors with sexual partners, in future BBSS. Although the estimates presented provide some representativeness of the population from which respondents were recruited, the small number of values for certain variables limit our ability to detect statistically significant differences between groups. In some cases, confidence intervals are too wide for meaningful interpretation.

These data provide a baseline from which to evaluate the overlap of injection drug use and unsafe sexual practices that put FSW and their sexual partners at heightened risk of HIV and HCV infection. These findings indicate that FSW who inject drugs are at high risk for HIV and HCV infection and illustrate the need for gender responsive HIV and injection prevention and treatment models that respond to the unique situations that increase FSW’s risk for injection drug use. Ideally, programs would ensure access to female service providers and include STI testing and treatment, gynecological and prenatal care, and mental health and social services. Future BBSS of the injection and sexual risk behaviors among FSW are being planned as part of an on-going surveillance strategy to monitor trends, identify and respond to failures and measure successes.