In the developing world, female sex workers (FSW) and their clients have been well identified as core groups that play a major role in the propagation of sexually transmitted diseases (STD) and HIV [1–5]. It is therefore crucial that these infections be controlled among these individuals. A high intensity intervention, including condom promotion and monthly laboratory screening of STD, integrated into a study on the interaction between STD and HIV, has been shown to result in a decrease in the incidence of HIV and STD over a 3 year period among FSW in Kinshasa, Zaire . No randomized controlled trial on the impact of such interventions has ever been reported, and very few studies have reported trends in sexual behaviour and HIV and STD prevalences over extended time periods [6,7]. Moreover, because most intervention programmes aimed at FSW cannot achieve the intensity reached within research projects, a better understanding of the trends in HIV and STD prevalence as well as of the changes in sexual behaviour over time may be useful to evaluate such programmes and to refine prevention efforts.
Within an ongoing prevention project aimed at FSW in Cotonou, Benin, we carried out three cross-sectional studies on sexual behaviour and HIV and STD prevalence, which provided us with data on time trends in this population for the period 1993–1999. Furthermore, we describe the significant changes that have occurred during this period in the sociodemographic characteristics of FSW in Cotonou and their potential impact on the observed trends.
Benin is a small West African country with a population of approximately 6 million inhabitants. It is bordered by Nigeria to the east, Togo to the west, Burkina-Faso and Niger to the north, and the Atlantic ocean to the south. Cotonou is the economic capital and the largest city in Benin with approximately 800 000 inhabitants. The majority of self-identified FSW in Cotonou are foreign women who come mainly from Nigeria, Ghana, and Togo. Beninese women, perhaps because of stigmatization, tend to practise a more hidden form of prostitution.
In Benin, the Canadian cooperation, through its West African AIDS Project (Projet Sida-2) is one of the major contributors to the National AIDS and STD Control Programme. A major part of the intervention activities of this project is implemented in Cotonou. Since 1993, the intervention has been aimed at FSW and more recently, since May 2000, at their male sexual partners. Field workers regularly visit FSW at home or at their working sites for HIV/STD prevention activities. A dedicated clinic has also been set up for the care of FSW in Cotonou, including the free offer of monthly screening and treatment (if necessary) of STD, using clinical algorithms that have been adapted for STD screening in this population . Approximately 250 visits per month are recorded at the clinic.
Data and specimen collection
Three serial cross-sectional studies on the prevalence of STD and HIV were conducted among FSW in Cotonou, Benin in 1993 (n = 374), 1995–1996 (n = 365), and 1998–1999 (n = 591), respectively. Participants were consecutive patients consulting for STD screening at the clinic dedicated to FSW. Throughout the intervention, women from all identified prostitution sites in Cotonou were encouraged to consult at the clinic for such screening. Visits to these sites were intensified during each data collection period when participation in the study was promoted by the field workers. At each recruitment round, after verbal informed consent, a similar structured questionnaire was administered to all the participants to record data on sociodemographic characteristics and sexual behaviour. The women underwent a gynaecological examination performed by a physician to diagnose STD clinically according to the current algorithms applied in FSW in Benin . During the examination, cervical swabs were collected and inserted into 2-SP medium. They were frozen at −20°C until they were transported to Canada within a few months to be tested for Neisseria gonorrhoeae and Chlamydia trachomatis. Finally, the women were asked to provide 10 ml of venous blood for HIV and syphilis testing.
N. gonorrhoeae and C. trachomatis were detected using the multiplex Amplicor CT/NG polymerase chain reaction (PCR) test (Roche Diagnostic Systems Inc., Branchburg, NJ, USA) as instructed by the manufacturer. These tests were performed at the Centre Hospitalier Affilié Universitaire de Québec (Quebec City, Canada) in 1993 and 1995–1996 and at the University of Montreal in 1998–1999. In all instances, two of us (F.B. and J.R.J.) were responsible for carrying out these tests. In 1993 and in 1998–1999, culture for N. gonorrhoeae and enzyme immunoassay for C. trachomatis were also performed locally, as well as supplementary confirmatory PCR assays for both bacteria in Canada. As these additional assays were not performed in 1995–1996, we report results based only on the Amplicor PCR test. This test has been shown to provide results almost equivalent to those based on a combination of all the above assays on samples collected in Cotonou [1,8].
HIV and syphilis testing were performed at the laboratory of the National AIDS and STD Control Programme of Benin. In 1993 and 1995–1996, an enzyme immunoassay for the detection of HIV-1 and HIV-2 antibodies (Vironostika HIV mixt; Organon Teknika, Boxtel, the Netherlands), followed by a rapid confirmatory test (Recombigen HIV-1/2; Cambridge Biotech, Galway, Ireland) for positive sera, and a Pepti LAV HIV-1/2 (Diagnostics Pasteur, Marne La Coquette, France) to distinguish between HIV-1and HIV-2, were used. In 1998–1999, Vironostika was also used as the screening test, and all positive samples were confirmed by a second enzyme immunoassay (ICE HIV-1.0.2; Murex Diagnostics SA, France). Confirmed positive samples were assayed with the multispot HIV-1/2 test (Sanofi Diagnostics Pasteur, Marne La Coquette, France) to distinguish between HIV-1 and HIV-2 profiles. Finally, syphilis serology was performed using a rapid plasma reagin test (Becton-Dickinson, Cockeysville, MD, USA), and all positive samples were confirmed using a Treponema pallidum haemagglutination test (Fujirebio, Tokyo, Japan), which was replaced during the course of the last round of testing in 1998–1999 by a T. pallidum particle agglutination test (Fujirebio). Active syphilis was diagnosed when both the rapid plasma reagin and T. pallidum haemagglutination (or T. pallidum particle agglutination) tests were positive.
Data were analysed using the SAS system 6.12 software (SAS Institute Inc., Cary, NC, USA). Chi-square and chi-square for linear trend, when appropriate, were used in the univariate analysis. The Mantel–Haenszel chi-square and its extension for linear trend analysis were used on stratified data. Analysis of variance and linear regression were used for both univariate and multivariate analyses of continuous variables. Multiple logistic regression was used to test the associations between the year of survey and HIV or STD status, while controlling for sociodemographic and behavioural variables. Annual adjusted odds ratios and their 95% confidence intervals, with 1993 as the reference category, were estimated in models in which the year of the survey was treated as categorical, using dummy variables. To test for time trends, we used Wald's chi-square from the logistic regression models with the year of survey treated as a continuous variable. As it was not possible to know if subjects recruited in later surveys had participated in earlier surveys, because data were recorded anonymously, participants in the three different rounds were considered as independent samples for the purpose of this analysis.
Sociodemographic characteristics and sexual behaviour
Table 1 presents the main characteristics of the participants in each survey. The mean age of the subjects decreased progressively over time. Women aged 25–34 years were the largest group in all three surveys, whereas the proportion of younger women, less than 25 years old, increased twofold from 1993 (16.6%) to 1998–1999 (33.7%). The majority of the participants were immigrants from surrounding countries for the three periods (98.1, 90.0 and 77.8%). In 1993, Ghanaian women were the predominant group, accounting for 66.3% of all participants, followed by Togolese (20.1%) and Nigerian (11.0%) women. In 1995–1996, the proportion of Ghanaian women decreased dramatically (38.9%), whereas that of Nigerian women increased to 37.3%. The latter became the largest group in 1998–1999 (38.0% of all the participants) followed by Beninese women (22.2%).
Most women reported having at least one regular sexual partner at each survey, with a slight increase over time. Reported condom use with such partners was very low at each survey, but with the highest rate in 1993. However, between 1995–1996 and 1998–1999 there was a decrease in the proportion of women reporting never using condoms with their regular partners.
The duration of sex work was not recorded in 1993, but was quite stable between the last two surveys. The mean number of clients in the previous week increased from 20 in 1993 to 26 in 1995–1996 and then decreased by half to 13 in 1998–1999. There was a statistically significant difference in the mean number of clients according to the year of the survey (P = 0.0001, analysis of variance). In the same analysis, using the Tukey adjustment for multiple comparisons, the mean number of clients was significantly different in all 2 × 2 comparisons between the years (all P values < 0.05). The mean percentage of condom use with clients in the previous week (calculated for each woman by dividing the number of clients with whom they reported using a condom by the total number of clients) was stable at approximately 60% in the first two surveys, but increased to over 80% in 1998–1999. There was a significant increasing trend in this measure over time (P = 0.0001, linear regression).
Because of the dramatic temporal changes in the distribution of the country of origin of the FSW in Cotonou, we examined the time trends in age and sexual behaviour, stratifying for country of origin (Table 2). The decreasing trend in age was almost entirely explained by the changes in the distribution of the country of origin, Nigerian women being much younger than other women. On the other hand, the multivariate analysis of variance on the number of clients in the previous week, controlling for age and country of origin, yielded results similar to those obtained in the univariate analysis (P = 0.0001). For each country of origin, the same pattern was observed: an increase between the first two surveys and a decrease thereafter. Concerning condom use with clients, the overall results of the multivariate linear regression analysis controlling for age and country of origin were similar to the univariate results (P = 0.0001 for the increase in condom use over time). Finally, the patterns of time trends in condom use with regular partners were similar for each country of origin, except for Beninese FSW, in whom the proportion of women never using condoms with such partners decreased steadily over time.
Prevalence of HIV and sexually transmitted diseases
In the univariate analysis, the prevalences of HIV infection and STD showed a significant decline over time (all P < 0.02; chi-square for trend): HIV infection from 53.3% in 1993 to 40.6% in 1998–1999; syphilis from 8.9 to 1.5%; gonorrhoea from 43.2 to 20.5%; and chlamydia from 9.4 to 5.1% (Table 3 and Table 4). The stratified analysis controlling for age yielded similar results (Table 3). However, in the stratified analysis controlling for the country of origin, only the declines in syphilis and gonorrhoea remained significant (Table 4).
The prevalence of HIV increased with increasing age. Women aged 25 years and over were more likely to be HIV infected than those aged less than 25 years (Table 3). In addition, Nigerian and Beninese women were significantly less infected than Ghanaian and Togolese women (Table 4). HIV-1 was largely predominant over HIV-2, antibodies to the former being detected in more than 95% of the cases in all surveys. Among HIV-infected women, the proportion who were seroreactive to only HIV-2 decreased from 4.6% in 1993 to 0% in 1998–1999, whereas the proportion of dually seroreactive women decreased from 19.5 to 2.2%.
Table 5 shows the results of a multivariate logistic regression model controlling for age and country of origin, with the annual adjusted odds ratios, their 95% confidence intervals, and the P values for trend for all the infections studied. In this analysis, HIV seroprevalence in 1995–1996 (49.4%) and 1998–1999 (40.6%) was not statistically significantly different from that in 1993 (53.3%). The dramatic declines observed in the prevalences of gonorrhoea and syphilis remained highly significant, whereas the downward trend in the prevalence of C. trachomatis was no longer significant.
A multivariate analysis, controlling simultaneously for age, nationality, the mean percentage of condom use, and the number of clients in the previous week yielded results similar to the model including age and country of nationality only (data not shown). Finally, we constructed multivariate models comparing 1998–1999 with 1995–1996, controlling for the same variables as above but also including the duration of prostitution, which was not available for 1993. The results of these models yielded results essentially similar to those that can be drawn form Table 5 when comparing 1998–1999 with 1995–1996 (data not shown).
One of the major findings of this study is the profound change that has occurred over time in the sociodemographic structure of the FSW population in Cotonou. Whereas Ghanaian women were by far the most numerous at the beginning of the 1990s, they have now been largely replaced by a mix of Nigerian and Beninese women, which has also resulted in a decrease in the mean age of this population. The factors determining the entry of women into sex work in Cotonou have not been studied extensively, and probably reflect a complex mix of socioeconomic and political factors. Furthermore, as most FSW in Cotonou are migrants, the epidemiology of HIV and STD in this population is also influenced by the complex relationship between migration and the spread of these diseases . The decrease in the proportion of Ghanaian women could be attributed partly to the high burden of HIV-related disease and death that affected these FSW in the early 1990s, and possibly also to the recent improvement in the economic situation in Ghana, with a resulting decrease in the number of young women migrating abroad to start involvement in sex work. On the other hand, Nigerian women may have massively entered sex work in Cotonou because of the poor economic and political conditions prevailing in Nigeria in the second part of the 1990s and the proximity of this city to the Nigerian border. Finally, decreased purchasing power as a result of the devaluation of the CFA currency in 1994 may have limited employment opportunities for Beninese women, leading some of them to practise overt prostitution, despite the risks of stigmatization.
Another important finding from this study concerns the decline in the number of clients in the second part of the 1990s. This could be partly due to increased awareness with regard to HIV/AIDS in the general population, which could have decreased the demand for FSW services from men. On the other hand, although no exact figures are available on this issue, it appears from our observations in the field that the number of FSW has increased substantially in Cotonou over the past few years. This could also have led to a decrease in the number of clients per woman as a result of an increase in the offer of prostitution services without an increase in the demand.
Despite the fact that the decrease we observed in the prevalence of HIV and STD can be attributed partly to these sociodemographic changes, our results show that there was a substantial decline in the prevalence of syphilis and gonorrhoea over the study period, independent of the changes in sociodemographic characteristics. For syphilis, there is even a trend towards elimination, with a prevalence of only 1.5% in 1998–1999. For gonorrhoea, the decline was drastic and is consistent in the past two surveys. This observation is very important because gonococcal infection is considered to be a good marker for evaluating the recent sexual behavioural risk of HIV transmission .
These declines in the prevalence of STD could be attributed to an improvement in safe sex practices in commercial sexual encounters. Indeed, the mean percentage of condom use with clients in the previous week increased to 80% in 1998–1999 in comparison to 60% in the two previous surveys. Although it is impossible to prove in the context of our observational study, it is very likely that these favourable changes are linked to the Projet Sida-2 intervention, with the development of STD clinical services and educational campaigns on HIV/AIDS/STD targeting FSW in Cotonou since 1993. In the other cities in Benin, where no targeted interventions have been carried out, HIV seroprevalence among FSW increased during the same period, from 34% in 1993 to 51% in 1999 (National AIDS Control Programme of Benin, unpublished data). Furthermore, recent mathematical modelling, using parameters representing the population of Cotonou in terms of demographic characteristics and sexual behaviour and simulating the Projet Sida-2 intervention, suggested that the observed declines in HIV and STD prevalences could be expected from such an intervention, and that it could also have had a favourable impact on the course of the HIV epidemic in the general population .
In the few reports of which we are aware from sub-Saharan Africa, interventions targeting this core group of women have also been found to have a significant impact in reducing HIV/STD rates [6,7,12]. In Dakar, the capital city of Senegal, the prevalence of N. gonorrhoeae decreased from 17.4% in 1991 to 4.7% in 1996 , whereas in Abidjan, it dropped from 33% in 1992 to 11% in 1998 . In the case of Senegal, the seroprevalence of HIV among FSW from Dakar was lower in 1996 (17.3%) than the 49.6% in our data for the same period. The results reported from Senegal may be attributed to the fact that interventions targeting this core group of women started long before the occurrence of the HIV epidemic, with the legalization of prostitution in 1969. Finally, periodic presumptive STD treatment of FSW in a South African mining community resulted in a decline in the prevalence of STD among both FSW and men working in the mines .
Even though the mean percentage of women reporting condom use with clients increased over the study period, condom use with regular partners in FSW from Cotonou remains very low, with more than 80% of the FSW never using condoms with such partners. Given than over two-thirds of the women reported such partners in all three surveys, this issue constitutes a major challenge for prevention programmes targeting FSW. In a study of sexual partners of FSW carried out in Cotonou in 1998 , the prevalence of HIV was twofold higher among the boyfriends of FSW than in their paying clients (16.1 versus 8.4%). Similar results regarding condom use with regular partners have been reported from The Gambia, where only 4% of FSW acknowledged consistently using condoms with their boyfriends . The favourable time trends in Senegal were observed in the context of high rates of consistent condom use with both clients (98%) and regular partners (94%) .
Increasing the usage rates of condoms among FSW in Cotonou thus remains a priority. The achievement of such an objective requires a combination of several different strategies. Therefore, since February 1999, an association of FSW, funded and supervised by Projet Sida-2, has been involved in activities aimed at increasing HIV/STD awareness, to empower FSW in the negotiation of condom use both with their clients and their boyfriends, and to refer newcomers to the STD clinic. Such peer educator-based interventions were shown to impact positively on condom use among FSW in Kenya and Zimbabwe . As the 1998 study among male sexual partners of FSW in Cotonou indicated that many of these men were under-informed or misinformed about HIV/STD transmission, prevention and treatment , Projet Sida-2 trained some men, recruited from within the milieu of prostitution in Cotonou, for HIV/STD education targeting male clients and regular partners of FSW in bars, hotels, and streets, in areas with a high concentration of FSW. Owners of brothels were also solicited and trained because of their influence in the sex work milieu. Finally, an STD clinic dedicated to clients and other sexual partners of FSW in Cotonou was opened in Cotonou in May 2000.
The main limitation of the study was caused by the extreme difficulty in obtaining a random sample of FSW. We used convenience samples at every study round and, in the absence of strictly representative samples, any time trend analysis should be interpreted with caution. However, this potential bias is unlikely to be very important in the context of Cotonou. The coverage of the FSW population by the Sida-2 intervention was very high. For example, a census of sex work venues carried out in 1996 showed the presence of 614 self-identified FSW in Cotonou (Projet Sida-2, unpublished data, 1996). In addition, in a sample of 200 women randomly selected at these prostitution sites in the same year, 96% reported that they knew the clinic dedicated to FSW, and 81% had attended the clinic in the past (Projet Sida-2, unpublished data, 1997). Another limitation was caused by the fact that there was no way to ensure that the study samples were independent. It is very unlikely that the results have been strongly affected by such a problem. Indeed, 58% of the 200 FSW randomly recruited in 1996 reported that they had been living in Cotonou for less than 6 months. In addition, a review of the clinic's records for the women who consulted during the third study round (1998–1999) showed that only 73 had ever consulted at the clinic before 1997. Finally, when comparing proportions, considering samples as fully independent when they are in fact partly dependent will mostly result in conservative P values, whereas it does not have any impact on point estimates.
The data presented here suggest that the intervention has been effective in reducing the prevalence of syphilis and gonorrhoea among FSW in Cotonou. In addition, the levelling off of the prevalence of HIV, independently of sociodemographic characteristics, is encouraging, especially in the context of the increasing prevalence of HIV among FSW in other geographical areas of Benin. Given these results, those obtained by others [3,6,7,12,14], and recent results from mathematical modelling suggesting that interventions focused at high-risk groups have the greatest impact, irrespective of the overall patterns of sexual behaviour and stage of the HIV epidemic , the large-scale development of community-based and clinic-based HIV preventive interventions for FSW should be highly prioritized.
The authors are grateful to the staff of the Cotonou 1 STD clinic and to all the women who agreed to participate in the different surveys.
1. Germain M, Alary M, Guèdèmè A. et al
. Evaluation of a screening algorithm for the diagnosis of genital infections with Neisseria gonorrhoeae
and Chlamydia trachomatis
among female sex workers in Bénin. Sex Transm Dis 1997, 24: 109–115.
2. Lowndes CM, Alary M, Gnintoungbé CAB. et al
. STD management and HIV prevention in men at high risk: targeting clients and non-paying sexual partners of female sex workers in Benin. AIDS 2000, 14: 2525–2534.
3. Laga M, Alary M, Nzila N. et al
. Condom promotion, sexually transmitted diseases treatment, and HIV decline among female Zairian sex workers. Lancet 1994, 344: 246–248.
4. Deceuninck G, Asamoah-Adu C, Khonde N. et al
. Improvement of clinical algorithms for the diagnosis of Neisseria gonorrhoeae
and Chlamydia trachomatis
by the use of a Gram-stained smears among female sex workers in Accra, Ghana. Sex Transm Dis 2000, 27: 401–410.
5. Diallo MO, Ghys PD, Vuylsteke B. et al
. Evaluation of simple diagnosis algorithms for Neisseria gonorrhoeae
and Chlamydia trachomatis
cervical infections in female sex workers in Abidjan, Côte d'Ivoire. Sex Transm Infect 1998, 74 (Suppl. 1) : S106–S111.
6. Meda N, Ndoye I, M'Boup S. et al
. Low and stable HIV infection rates in Senegal: natural course of the epidemic or evidence for success of prevention? AIDS 1999, 13: 1397–1405.
7. Ettiègne-Traoré V, Ghys PD, Diallo MO, et al. Seven years trends in HIV-infection and sexually transmitted diseases (STD) and socio-demographic and behavioural characteristics in female sex workers in Abidjan, Côte d'Ivoire, 1991–1998. XIth International Conference on AIDS and STD in Africa
. Lusaka, September 1999 [Abstract 14PT37-33].
8. Mukenge-Tshibaka L, Alary M, Bernier F. et al
. Diagnostic performance of the Roche Amplicor PCR in detecting Neisseria gonorrhoeae
in genitourinary specimens from female sex workers in Cotonou, Bénin. J Clin Microbiol 2000, 38: 4076–4079.
9. Decosas J, Adrien A. Migration and HIV. AIDS 1997, 11 (Suppl. A) : S77–S84.
10. Renton AM, Whitaker L. Using STD occurrence to monitor AIDS prevention. Soc Sci Med 1994, 38: 1153–1163.
11. Boily MC, Alary M, Lowndes C. The impact of HIV intervention in Cotonou (Benin): past and future. Int J STD AIDS 2001, 12 (Suppl. 2) : 80–81.
12. Steen R, Vuylsteke B, DeCotto T. et al
. Evidence of declining STD prevalence in a South African mining community following a core group intervention. Sex Transm Dis 2000, 27: 1–8.
13. Pickering H, Quigley M, Hayes RJ. et al
. Determinants of condom use in 24 000 prostitute/client contacts in The Gambia. AIDS 1993, 7: 1093–1098.
14. Ngugi EN, Wilson D, Sebstad J, Plummer FA, Moses S. Focused peer mediated educational programs among female sex workers to reduce sexually transmitted disease and human immunodeficiency virus transmission in Kenya and Zimbabwe. J Infect Dis 1996, 174 (Suppl. 2) : S240–S247.
15. van Vliet C, Meester EI, Korenromp EL, Singer B, Bakker R, Habbema JDF. Focusing strategies of condom use against HIV in different behavioural settings: an evaluation based on a simulation model. Bull WHO 2001, 79: 442–454.
© 2002 Lippincott Williams & Wilkins, Inc.