Article Text

Download PDFPDF

Syphilis among female sex workers in Indonesia: need and opportunity for intervention
  1. Nurholis Majid1,
  2. Liesbeth Bollen1,
  3. Guy Morineau1,
  4. Syaiful Fahmi Daily2,
  5. Dyah E Mustikawati3,
  6. Nurhayati Agus1,
  7. Atiek Sulistyarni Anartati1,
  8. Chawalit Natpratan1,
  9. Robert Magnani1
  1. 1Family Health International, Country Office, Jakarta, Indonesia
  2. 2Department of Dermato-venereology, University of Indonesia and Cipto Mangunkusumo Hospital, Jakarta, Indonesia
  3. 3Ministry of Health, Republic of Indonesia, Jakarta, Indonesia
  1. Correspondence to Liesbeth J M Bollen, Family Health International, Komplek Ditjen PP & PL Depkes RI, Jalan Percetakan Negara 29, Jakarta 10560, Indonesia; lbollen{at}


Objective To study the prevalence and factors associated with syphilis among female sex workers (FSWs) in Indonesia.

Methods Direct and indirect FSWs were sampled in 10 major cities in Indonesia. A behavioural survey was conducted and samples obtained and tested for HIV (Bioline and Determine) and syphilis (RPR and Determine). Syphilis prevalence and potential factors associated with syphilis were assessed in bivariable and multivariable analysis. Syphilis prevalence among brothel-based sex workers from previous surveillance in 2003 and 2005 was compared to 2007.

Results A total of 2436 direct and 1888 indirect FSWs participated in both the behavioural and biomarker surveys. Prevalence of active syphilis (RPR≥1:8) was high among direct and indirect FSWs (7.5% vs 3.1%) and was not lower among those who had visited an STI clinic in the last 3 months. Prevalence of active syphilis was lower among those who had received at least one dose of a prior periodic presumptive treatment programme (PPT) for chlamydia and gonorrhoea with 1 g azithromycin and 400 mg cefixime compared to those who had not received PPT (3.9% vs 6.0%; p=0.008). Older age (AOR=1.4), longer duration of sex work (AOR=1.7) and PPT (AOR=0.6) were associated with active syphilis in multivariable analysis. Syphilis prevalence among brothel-based FSWs increased from 2005 to 2007 (7.8% vs 14.5%; p<0.001).

Conclusions Syphilis prevalence among FSWs in Indonesia was high and increased from 2005 to 2007. Receipt of PPT was associated with lower syphilis prevalence. Current syphilis control programmes need to be evaluated and the possibility of alternative syphilis treatment with azithromycin explored.

  • Syphilis
  • prevalence
  • azythromycin
  • cefixime -PPT
  • control programs
  • STD control
  • surveillance
  • syphilis

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


With the exception of Tanah Papua, where a low-level general population epidemic has emerged,1 HIV in Indonesia remains highly concentrated in most-at-risk populations. Although injecting drug use has been the main driver of expanding HIV epidemics in a number of countries in the region over the past 5–10 y, new projections by both the Indonesian National AIDS Commission and the UNAIDS-sponsored Commission on AIDS in Asia point to a crucial role of commercial sex in determining the future of the epidemic in Indonesia and the region as a whole.2 3 The extent to which commercial sex could spark a rapid spread of HIV depends upon a number of factors, including numbers of sex workers and clients, prevalence of male circumcision, frequency of unprotected sex and the prevalence of HIV and other sexually transmitted infections (STIs) among sex workers.

Although the prevalence of HIV among Indonesian female sex workers (FSWs) has remained modest in comparison with those in other countries in the Asia region, the prevalence of other STIs is relatively high.4 5 Previous reports have documented high prevalence of gonorrhoea and chlamydia among FSWs and interventions, such a periodic presumptive treatment (PPT), are being implemented.6 However, less emphasis has been placed on syphilis control among FSWs in Indonesia. Syphilis infection fuels the HIV epidemic by increasing the risk of HIV acquisition and by increasing HIV infectiousness among those already infected with HIV. Combined estimates from 18 studies showed that positive syphilis serology increased the risk of HIV acquisition by 2.5 times.7 A study among HIV-discordant partnerships in Haiti showed that positive syphilis serology in the HIV-infected person doubled the risk of HIV transmission to the partner.8 In addition, untreated syphilis results in short term (eg, genital lesions, lymphadenopathy, alopecia) and long term (cardiovascular complications such as aortic aneurysm and neurological complications such as meningitis) morbidity and could be transmitted during pregnancy resulting in congenital syphilis.9 10

Sex workers are encouraged to receive regular serological screening for syphilis as part of STI screening services provided at community health centres in Indonesia. The recommended treatment for syphilis according to national and international guidelines is injection with benzathine penicillin. Although anaphylactic shock resulting from treatment with benzathine penicillin is relatively rare,11 anecdotal reports indicate reluctance of Indonesian physicians to inject penicillin (personal communication 2009, Professor Syaiful Fahmi Daily, Department of Dermato-venereology, University of Indonesia and Cipto Mangunkusumo Hospital, Jakarta, Indonesia). In addition, limited availability of penicillin (personal communication 2009, Professor Syaiful Fahmi Daily) also results in many clinicians providing alternative treatment with doxycycline twice daily for 14 days leading to inadequate treatment because of low compliance. Coverage of adequate syphilis treatment among FSWs is not known and may vary from province to province.

We assessed the prevalence of syphilis, gonorrhoea and chlamydia among FSWs in large Indonesian cities. This study focuses on syphilis prevalence, factors associated with syphilis and trends in syphilis prevalence over time. These data will assist in developing targeted interventions in this resource-constrained setting.


Sampling procedures

Two samples of FSWs were identified independently, including direct (ie, brothel-based and street-based FSWs) and indirect FSWs (ie, women working in entertainment places such as massage parlours, karaoke clubs, discotheques and bars). Local health authorities and non-governmental organisations providing services to FSWs obtained city and group-specific sampling frames after mapping sex-work venues. A two-stage cluster sampling approach was used. In the first stage, block censuses were selected proportional to the number of FSWs enumerated in the block. All venues listed in selected blocks were included. A number of FSWs proportional to the venue size was selected through systematic sampling among the women available at the time of the visit.

This study was part of a larger behavioural survey; we limited this analysis to FSWs in cities selected to participate in both the behavioural and biomarker survey. Apart from one city chosen for programme evaluation, the selected cities were either capitals of provinces or hotspots for commercial sex, including one main economic trade and cross-border area with Singapore (Batam), one transit area between Java and Bali (Banyuwangi) and two economically booming areas in Papua (Jayapura and Sorong). The sample size was 250 per group per city. Data were collected from ten cities for direct FSWs and from eight cities for indirect FSWs.

Survey procedures

Trained staff from the National Bureau of Statistics and the Provincial Health Department explained the study objectives, procedures, potential benefits and risks to participants. Only FSWs of at least 15 y of age were asked to participate in the survey. Staff obtained verbal informed consent and collected information on demographics and sexual behaviours during face-to-face interviews using a standardised questionnaire. Data were collected in a manner ensuring sufficient privacy. No personal identifiable information was collected.

Phlebotomy was performed according to routine national surveillance methods. Participants provided self-collected vaginal swabs; swabs were frozen at −20°C until thawed for gonorrhoea and chlamydia testing. Results of biological tests were anonymously linked to behavioural data using a unique ID number assigned to each participant.

Periodic presumptive treatment programme

A pilot PPT programme (supported by the WHO and Family Health International (FHI)) for chlamydia and gonorrhoea with 1 g azithromycin and 400 mg cefixime was provided three times with monthly intervals for FSWs in Denpasar (Bali), Banyuwangi, Surabaya and Semarang (Java) from December 2006 until July 2007. This study assessed the impact of PPT on STI prevalence.

Laboratory analysis

Whole blood was centrifuged and aliquoted. Serum specimens were stored at 4–6°C and tested within 3 days of collection at the nearest local laboratory. Presence of antibodies against Treponema pallidum was tested using the rapid plasma reagin (RPR) test (Shield Diagnostics, Bridpot, Dorset, UK) as a screening test and the T pallidum Rapid Test (Determine TP Syphilis, Inverness Medical, Bedford, UK) as a confirmatory test. Specimens with both positive RPR and T pallidum haemagglutination test (TPHA) results were titred until 1/128 dilution. Those with RPR titre ≥1:8 and TPHA positive were classified as active syphilis.

HIV testing was done with two parallel rapid tests: SD Bioline HIV 1/2 3.0 (Standard Diagnostic Inc, Suwon City, South Korea) and Determine HIV-1 (Inverness Medical, Matsudo, Japan). Discrepant results were retested at the National Research Laboratory using two ELISA assays: Murex (Murex Biotech, Dartford, UK) and Vironostika (Biomérieux, Marcy l'Etoile, France). Detection of Chlamydia trachomatis and Neisseria gonorrhoeae was performed by PCR analysis (Cobas Roche Amplicor Diagnostic System, Basel, Switzerland).

Endocervical swabs and cervico-vaginal smears were examined by light microscopy and treatment was provided free-of-charge on the same day based on patient's evaluation and according to the national STI treatment guidelines (eg, treatment of cervicitis presumptively for gonorrhoea/chlamydia with cefixime 400 mg and 1 g azithromycin).

Women were recommended to receive their syphilis test results after a week at a nearby clinic by showing their participant number; treatment was provided according to national treatment guidelines free-of-charge. HIV test results were not provided. FSWs received vouchers for free voluntary counselling and testing for HIV at a nearby clinic; data from these clinics were not collected as part of this survey.

The study protocol was approved by the Indonesia Ministry of Health Ethics committee as well as the Family Health International Protection of Human Subjects Committee.

Syphilis trends 2003–2007

Data from surveys in 2003 and 2005 among FSWs were used to compare syphilis prevalence. The study in 2003 involved Banyuwangi in East Java, Semarang in Central Java and Jayapura in Papua.12 In 2005, Jakarta, Bandung in West Java and Surabaya in East Java were added as surveillance sites; this study has been described in detail elsewhere.5 Laboratory procedures for the 2003 and 2005 surveys were similar as described for the 2007 project. For the analysis of syphilis prevalence trends from 2003, 2005 to 2007, we limited the analysis to brothel-based sex workers and excluded indirect and street-based sex workers as the proportion of different type of sex workers varied between surveys and difference in prevalence between sex worker types made the trends hardly interpretable.

Statistical analysis

Behavioural data were double entered using Census and Survey Processing System (CSPro) 2.6.007 (US Census Bureau). Laboratory data were entered using Microsoft Excel. Analysis was performed using Stata 9.0 (Stata Corporation, College Station, TX, USA). While sample stratification was taken into account for analysis, insufficient cluster information did not allow weighting for cluster sampling. Variables of interest were described in terms of frequency, median and range. Prevalence of HIV, syphilis (any titre) and active syphilis (titre ≥1:8) were assessed with 95% exact binomial CIs. Associations between categorical variables were assessed using the Wald test and p values <0.05 were considered as significant. Comparisons between medians were tested using Wilcoxon rank sum test. Logistic regression was used to assess associations with syphilis in both bivariable and multivariable models. Continuous variables were dichotomised at median levels except for duration of sex work, which was dichotomised at 12 months to provide a better fit for the model. Multivariable modelling was carried out using factors with p<0.2 in the bivariable analysis and following backward stepwise elimination; significant associations with p<0.05 were retained in the model. ORs and adjusted OR (AOR) were calculated with 95% CIs.


Demographic and behavioural characteristics

A total of 2436 direct and 1888 indirect FSWs participated in the survey from August through October 2007 (participation rates were 94.4% and 97.4%, respectively); their median age was 27 y (range 15–59) (table 1). Most women were divorced, separated or widowed, and most had children. Education levels among FSWs were low.

Table 1

Demographic and behavioural characteristics among 4324 female sex workers in 10 cities in Indonesia, 2007

The median number of clients in the week prior to the survey was five (range 0–76). A higher proportion of indirect FSWs reported consistent condom use with clients in the last week compared to direct FSWs (38.0% vs 31.4%; p<0.001) (table 1). Only 61 (1.4%) FSWs reported to ever have injected drugs.

Genital ulcers in the last year were reported by 15.0% of direct and 7.9% of indirect FSWs (p<0.001). Direct FSWs were significantly more likely than other FSWs to have visited a STI clinic in the prior 3 months and to have sought medical treatment for the last STI episode (both p<0.01). More direct than indirect FSWs accessed the PPT pilot programme (table 1) and most FSWs (80–90%) in the pilot sites accessed PPT (data not shown).

STI prevalence

The prevalence of STIs is shown in table 2. The highest prevalence of active syphilis among direct sex workers was found in Batam, Riau Island Province (16.8%), in Banyuwangi, East Java (9.7%) and among indirect sex workers in Deli Serdang, Sumatra province (12.9%) (figure 1). The prevalence of chlamydia and/or gonorrhoea was lower among those who had participated in the PPT programme at least once compared to never (28.4% vs 38.9%; p<0.001) (data not shown).

Table 2

Prevalence of HIV and other STIs among 4324 female sex workers in 10 cities in Indonesia, 2007

Figure 1

Prevalence of HIV and syphilis among direct* female sex workers in Indonesia, 2007. PPT, periodic presumptive treatment (for chlamydia and gonorrhoea). *Prevalence shown for direct female sex workers at all sites except for Sumatra where indirect sex workers participated (no sex establishments present).

Determinants of syphilis

Direct FSWs had higher prevalence of active syphilis than indirect FSWs (7.5% vs 3.1%; p<0.001) (table 3). HIV-positive FSWs had higher prevalence of active syphilis than HIV-negative FSWs (8.6% vs 2.8%; p=0.01). Prevalence of active syphilis was lower among those who had participated in the PPT programme at least once compared to never (3.9% vs 6.0%; p=0.008); there was no evidence of a difference in prevalence between those receiving one, two or three doses of PPT (data not shown). In multivariable analysis, direct FSWs had a higher risk for active syphilis (AOR 2.9); age older than the median of 27 y (AOR 1.4), longer than 1 y duration of sex work (AOR 1.7) and gonorrhoea (AOR 1.6) were independently associated with syphilis (table 3). Having received PPT at least once was protective against syphilis (AOR 0.6).

Table 3

Characteristics associated with active syphilis (RPR≥1:8) among 4324 female sex workers in 10 cities in Indonesia, 2007

Syphilis trends 2003–2007

Trends in syphilis prevalence among brothel-based FSWs in five cities in Indonesia are shown in figure 2. In all but one city, the prevalence increased from 2005 to 2007. The overall prevalence remained similar from 2003 to 2005 (8.5% vs 7.8%; p=0.6) but increased from 2005 to 2007 (7.8% vs 14.5%; p<0.001). The overall increase was largely due to the prevalence rise observed in Banyuwangi (7.2% vs 14.6%; p=0.01) and Jayapura (3.8% vs 11.6%; p=0.03).

Figure 2

Prevalence of syphilis among brothel-based sex workers in five Indonesian cities, 2003-2007. PPT, periodic presumptive treatment (for chlamydia and gonorrhoea). Prevalence of syphilis from 2005–2007 in Bandung (p=0.3), Semarang (p=0.5), Banyuwangi (p=0.01), Surabaya (p=0.06), Jayapura (p=0.03), overall (p<0.001).


We observed high prevalence of syphilis among FSWs in 10 major cities in Indonesia. This is alarming as untreated syphilis infection can result in serious complications and can fuel the HIV epidemic. Also, the prevalence of syphilis has increased among brothel-based sex workers from 2005 to 2007, indicating the need to review the syphilis control programme in Indonesia.

The prevalence of syphilis was high in our survey compared to FSWs in other countries of South-East Asia. In Northern Thailand, the prevalence of syphilis among FSWs prevalence fell from 21% in the early 1990s13 to around 2% a decade later.14 Successful syphilis control programmes in Bangladesh resulted in declining rates of syphilis from 34% in 1999 to 7% in 2006 among street-based sex workers in Dhaka with <1% HIV prevalence.15 Syphilis control also remains challenging in India where around 10% of sex workers have reactive syphilis serology.16 In addition, syphilis outbreaks have been reported among populations with high HIV prevalence in Asia and elsewhere.17 18

Treponema infections include the non-venereal treponematoses, such as yaws caused by T pallidum subspecies pertenue. The WHO led a campaign to eradicate yaws, which has been successful in most countries.19 Cases of yaws have been detected in Indonesia, especially in rural areas and among children with one study reporting 0.3% prevalence in a community in West-Sumatra.20 As FSWs are at no greater risk for yaws than other population groups, it seems unlikely that yaws contributes to the high sero-prevalence among sex workers from urban areas in our survey.

Only a small proportion of sex workers reported consistent condom use with clients during the last week. This is in agreement with earlier reports from Indonesia5 21 and efforts continue to strengthen the condom promotion programme. Although consistent condom use will contribute to the prevention of syphilis transmission, it will take time to establish efficient and effective programmes.

Prevalence of gonococcal and chlamydial infection was also very high among FSWs in our survey, as previously reported4 5; the prevalence was lower among FSWs who had accessed the PPT programme. We demonstrated recently the importance of providing syndromic treatment with new and adequate drugs (cefixime and azithromycin instead of ciprofloxacin and doxycycline) after completing PPT for a sustainable decline in prevalence.6 This success led the Indonesian National AIDS Commission to scale up the PPT treatment intervention to 12 additional sites in Indonesia in March 2009.

HIV prevalence was higher among FSWs in Papua than elsewhere reflecting the fact that Papua has a generalised epidemic. Although most Indonesian men outside Papua are circumcised, providing them with some protection against HIV, their HIV acquisition risk during unprotected sexual intercourse with a HIV-infected partner will increase after contracting syphilis or another STI. FSWs in Sumatra and Batam had higher syphilis prevalence than HIV prevalence (figure 1), which may contribute to new HIV seroconversions in this population. Alternatively, syphilis control will also be important among populations with high HIV prevalence (such as FSWs in Papua) as syphilis increases HIV infectiousness and the risk of HIV transmission to sexual partners. In addition, syphilis infection also has an effect on the natural history of HIV infection by increasing HIV viral load, decreasing CD4 cell counts22 and increasing the risk of neurosyphilis.23

Prevalence of active syphilis was higher among direct compared to indirect sex workers. However, HIV prevalence was also higher among direct sex workers, which may somewhat explain the higher syphilis prevalence in this population as HIV-infected populations have higher syphilis prevalence.7 In addition, unmeasured factors such as client characteristics might be at play.

The finding that those who visited the STI clinic did not have lower risk for active syphilis could reflect re-infection or failure to access adequate treatment and prevention services. In addition, older FSWs and those involved in sex work for a longer duration had higher active syphilis rates suggesting longer exposure without adequate treatment. However, this study is limited to a cross-sectional assessment and only shows associations without understanding the causality. Although we hypothesise that some factors are associated with syphilis due to poor treatment coverage, we did not assess and evaluate access and provision of adequate treatment.

We found lower active syphilis prevalence among FSWs who had previously accessed the pilot PPT programme, providing 1 g azithromycin and 400 mg cefixime for treatment of chlamydia and gonorrhoea, directly observed by clinic staff. Azithromycin seems to also cure syphilis infection as demonstrated in a randomised trial comparing 2 g azithromycin with benzathine penicillin, which showed similar cure rates.24 However, a non-random treatment design study in Uganda evaluated only 1 g azithromycin and also found similar cure rates, although women had higher cure rates compared to men.25 It might be the case that 1 g azithromycin was sufficient to treat syphilis among (some) FSWs in Indonesia, most of them probably having a body weight <55 kg (the average weight among 29-year-old women in Central Java, Indonesia was 47.7±8.0 kg according to a study about energy intake).26 In addition, cefixime, which was provided as part of PPT, is likely to have had some curative effect on syphilis since other cephalosporins (eg, ceftriaxone) have shown efficacy against syphilis9 27 It is not clear why a similar study among FSWs in India did not show an effect of PPT on syphilis prevalence.16 Although treatment with benzathine penicillin will remain the first treatment of choice according to guidelines,27 28 it would be desirable to determine the optimal azithromycin dose to cover treatment for both chlamydia and early syphilis among persons not able to access or not willing to receive penicillin injection at a clinic. However, there would be limitations to the use of azithromycin for alternative syphilis treatment. First, several azithromycin treatment failures were observed in San Francisco29 and azithromycin-resistant T pallidum strains were documented in the US and Ireland.30 However, no resistance to azithromycin was found in 141 patients with syphilitic lesions in Madagascar suggesting resistance is geographically isolated.31 Active surveillance for resistance would be recommended in sites where azithromycin is used. Second, the placental transfer of azithromycin is very low,32 which renders azithromycin not suitable to prevent congenital syphilis. However, access to oral treatment in community settings could help lower the syphilis prevalence at the population level thereby reducing syphilis prevalence among pregnant women and consequently reducing the risk for congenital syphilis.

Although assessing trends from available data should be undertaken cautiously due to differences in sampling methodology, the assembled data indicate that the overall syphilis prevalence among brothel-based FSWs in Indonesia increased significantly from 2005 to 2007. The reason for this increase is not clear and may result from increased transmission due to inadequate treatment provision in combination with continued low condom use in commercial sex. The significant increase in syphilis prevalence among FSWs in Papua most likely reflects efficient male-to-female transmission as most men in Papua are not circumcised in contrast to the majority of men in Java (eg, figure 2: Bandung, Surabaya, Banyuwangi, Semarang). Based on information from STI clinics, coverage of syphilis screening is high and adequate treatment with benzathine penicillin is being provided in Semarang (personal communication, Dr Robert Magnani, Country Director, Family Health International, Indonesia). In addition, the decrease in syphilis prevalence among FSWs in Semarang might result from previous receipt of PPT. Certainly, the current syphilis control programmes at other sites should be evaluated and strengthened. The need for alternative treatment for early syphilis should be assessed and its effectiveness in treating early syphilis among non-pregnant women who cannot access penicillin treatment will be evaluated as appropriate. A strengthened syphilis control programme would eventually be integrated with the PPT programme for gonorrhoea and chlamydia.

Key messages

  • Syphilis prevalence among female sex workers (FSWs) was high across 10 major cities in Indonesia; active syphilis (rapid plasma reagin ≥1:8) prevalence was 7.5% among direct FSWs.

  • Prevalence of active syphilis was significantly lower among those who had received at least one dose of a prior periodic presumptive treatment (PPT) programme for chlamydia and gonorrhoea with 1 g azithromycin and 400 mg cefixime.

  • Syphilis prevalence among brothel-based FSWs increased significantly from 2005 to 2007.

  • Current syphilis control programmes need to be evaluated and the possibility of alternative treatment for early syphilis with azithromycin explored.

  • Prevalence of chlamydia (33%) and gonorrhoea (25%) was also high among FSWs but significantly lower among those who had participated in the PPT programme.


We thank Dr Graham Neilsen for reviewing this paper and for providing useful comments.



  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was conducted with the approval of the Indonesia Ministry of Health Ethics committee; Family Health International Protection of Human Subjects Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.