This study explored the role of Hijras (male transgenders) as a core group in Pakistan\'s HIV epidemic. Four hundred and nine Hijras underwent detailed behavioural and biological assessment. Our subjects were young (median age: 24 years), debuted sex early and used alcohol and drugs. Sex with men, women and other Hijras along with co-habitation/marriagewere reported. Most (84%) had sold sex. These reported a median of four times and at least one regular client weekly. Few used condoms. Most (94%) could identify a condom, but 42% reported never needing one. Over two-thirds had correct HIV and protective knowledge. Many reported experiencing physical abuse or forced sex (40%) and sexual orientation-related discrimination (45%). Most (58%) had sexually transmitted infections (STIs) and 38% had multiple infections. The commonest infections were syphilis (50%) and gonorrhoea (18%). Care-seeking was mainly (87%) from the private sector. High STI prevalence, commercial sex, bisexuality and infrequent protective behaviours are described. Their communal living and wider sexual networks suggest a more central role for Hijras in Pakistan\'s HIV epidemic. Effective HIV/STI control programmes must take the above factors into account to enrich their programme content.
Keywords: Hijra, Transgender, Pakistan, HIV, Sexually transmitted, STI.
Hijras, males with female gender identity, have long been a part of the South-Asian culture.1 In today\'s modernizing societies, they live in marginalized communities and participate extensively in the sex trade, which makes them an essential link in Pakistan’s HIV transmission chain.2,3
Transgender males who identify themselves as female are zenanas. However, in practice both society and health/prevention workers conflate the terms, calling all zenanas Hijras. True eunuchs are uncommon in Pakistan (,1-15%).1,4,5 Surgical castration is a major rite of passage, that is performed ceremoniously on the devoted few and enhances the castratee\'s status within the Hijras community.1,4,6
Hijras live in tightly knit, seniority-based hierarchical communities led by gurus (teacher) who are more senior members of the community.1 While gurus do not force disciples to sell sex, disciples seek the guru\'s sanction (often post-facto) for such acts and turn their income over to gurus.6 Income is generated by dancing, blessing births, begging and sex trade; although modernity has diminished the role of the first two.1 Finally, Hijras may have sustained non-commercial sexual relations or commercial sex with a chawa (male non-Hijra).6 Hijra communities abhor inter-Hijra sexual relations and sex with women is considered rare (Abbas, personal communication), implying limited concordant relationships and sexual networks. A previous study had found some syphilis and risk behaviours but no HIV among Hijras.4 No study has looked at other STIs among this group in Pakistan.
Pakistan\'s HIV epidemic is fully established among injecting drug users and is spreading into male sex workers (MSWs) and Hijras in many cities before involving female sex workers.2,7 HIV prevention programmes directed at Hijras with MSWs provide counselling, peer outreach, condoms and STI care and nearly all are funded by the Pakistani Government. According to the national HIV surveillance programme, there are an estimated 36,000 Hijras in cities with populations of 200,000 or more, comprising one-quarter of all sex workers nationwide.3 Since STIs may serve as surrogates of potential HIV transmission,8 linking STI prevalence with behaviours, sexual networks, prevention knowledge and practices of Hijras can provide insights for the design and placement of effective HIV control measures. Female gender identity, stigma, marginalization and communal organization distinguish Hijras from other men who have sex with men (MSM). These factors may influence personal protective decisions of Hijras and have implications for HIV/STI prevention programmes that are premised on advancing the individual\'s decision-making ability about sex behaviours to enhance personal safety.
We present a subanalysis of the Hijras data from a crosssectional study,9 which was conducted to inform about the risk of HIV and STIs among sex workers, Hijras and injecting drug users in order to design effective prevention programmes. We explore sexual behaviours, relationships and prevalence of HIV and STIs among Hijras to better define their potential role in Pakistan\'s HIV epidemic.
We administered behavioural questionnaires and biological tests to 206 Hijras from Karachi and 203 from Lahore. All selfidentified Hijras were eligible. Based on interviews with local informants, we assumed that at least 500 Hijras lived in each city. Using the formula: n ¼ z2 _ (prevalence) _ (1 2 prevalence)/ CI:2 a sample size of 170 in each of the cities was needed to detect 20% prevalence of any STI (a: 0.05, b: 0.8 and confidence interval [CI]: 6%). Interviews and examinations were conducted privately in a local clinic, rented for the purpose. Refusal rate was 5%. The methodology for recruitment, consent, ethical oversight, questionnaire design and biological testing has been described previously.9
The study was explained in local languages by the interviewer to individual potential subjects. Once the subject agreed to participate, separate verbal consents were obtained and recorded for the interview and biological samples (blood, urine and anal/genital swabs). Subjects who refused biological sampling were still included in interviews, if they so consented. Due to widespread illiteracy, a formal written consent was not possible for majority of subjects and therefore not pursued.
Subjects were compensated for their time at Rupees 400 (US$ 7) as determined from local non-governmental organization (NGO) consultations and targeted interviews with subjects.
Data were analysed with SPSSw (SPPS, Chicago, IL, USA). Odds ratios were calculated from univariate analysis and all factors significant at P, 0.2 were included in multiple regression models.
All respondents identified themselves as Hijras and zenanas. None were eunuchs and 98% were circumcised. The main demographic features are described in Table-1.
Median age at first coitus was 14 years. Of the 105 who currently lived with a partner, 79 (78%) did so with a man, 14 (13%) with a woman and seven (7%) with other Hijras; 32 (8%) were currently married. About 20% reported ever having sex with awoman and 7% within the past one year; with 15% condom use at last sex encounter. In the past month, 15% had paid a man to have sex with them. Nearly all (95%) reported having had sex with other Hijras. Most (84%) had sold sex within the past month and nearly all (99%) within the past year, at Rupees 100-200 per act (US$ 1.6-.3.3).
Fifty-one percent reported new clients (median: 1) and 68% regular clients (median: 3) within the past week. Eighteen percent reported using condoms during last sex with a one-time client (12% with receptive anal sex) and 15% with a regular client (13% with receptive anal sex); 83% had never asked a client to use condoms. Oral sex was performed with 39% one-time and 42% regular clients and never with condoms. Clients were businessmen (42%), drivers (35%), students (28%), police/military (25%), rickshaw-drivers (15%) and unemployed (10%). They were contacted at the roadside (53%), parks (29%), in the houses (29%), hotels (13%) or cinemas (11%).
About 40% reported having sex with non-paying partners (median: 2) in the past month. Condoms were never (89%) or infrequently (8%) used during the last such acts. Some (14%) had participated in group sex within the past year.
Although 79% used a lubricant some of the time, only 4% did so simultaneously with a condom. Oil (54%), saliva (30%), antiseptic cream (12%) and water-based lubricants (1%) were the most commonly reported lubricants during the last anal sex act.
STI/HIV Prevention Knowledge:
Most (94%) could identify a condom, but 42% felt they never needed one. Only 1% carried a condom but 56% felt they could obtain one if needed. Most (68%) had heard of HIV and associated risk reduction with condom use (69%), avoidance of anal sex (73%) or needle sharing (87%).
Within the past year 40% reported being physically abused or forced to have sex by either the police (49%: 71% in Karachi and 32% in Lahore) or clients (46%). In the same period 45% felt discriminated against and 19% were blackmailed due to their sexual orientation.
Variations between Lahore and Karachi:
The behaviour patterns varied between the cities. More from Lahore (66% vs. 46%, P, 0.023) were lifetime residents of the city, reported alcohol use within past week (28% vs. 11%, P, 0.001), had paying female clients within past month (28% vs. 1% P, 0.001), cohabited (31% vs. 18%, P, 0.003), used condom with last sex act (16% vs. 9%, P, 0.009), had regular clients (80% vs. 53%, P, 0.001), participated in group sex (19% vs. 7%, P, 0.001) and faced discrimination (55% vs. 34%, P, 0.001). More from Karachi reported sex work as their main source of income (97% vs. 84%, P, 0.001). HIV knowledge was slightly better in Karachi but did not translate into better protective actions.
Clinical Assessment of STIs:
Current STI-related symptoms or examination findings were uncommon. Only 7% had inguinal lymphadenopathy, 1% had urethral discharge and 3% had anal or genital sores or ulcers on examination. Subjects were asked if they knew of STI symptoms as surrogate of previous STIs, but such knowledge correlated with a current diagnosis of STIs only for burning urination (P< 0.016) and non-anal genital ulcers or sores (P< 0.028).
One (59%) or more STIs (38%) were common. STIs prevalence varied considerably between the cities (Table-2).
The probability of having an STI or HIV was predicted using sex behaviours or protective knowledge in separate multiple regression models. The sex behaviours model (R2: 0.074) showed that only having regular clients within the past week decreased probability of an STI or HIV (adjusted odds ratio: 0.85, 95% CI: 0.73-0.98). In this model having women or one-time male clients, condom use (with any category of clients or with any type of sex act), performing oral sex and paying for sex were not associated with STI acquisition.
A separate model (R2: 0.074) predicted the probability of STIs based on knowledge of HIV or STIs and their protective actions. Only awareness of the danger from shared needles was associated with the reduction in the probability of having an STI or HIV (adjusted odds ratio: 0.52, 95% CI: 0.27-0.99).
STI Treatment and Care-seeking:
Main STI symptoms that led to care-seeking in the past year were genital sores (38%), anal (46%) or penile discharge (57%). Care was sought from private clinics (51%), traditional healers (10%), pharmacies (4%) or government clinics (4%). One-third (37%) never sought care. Median time to seeking care was two days and the treatment costs a median of Rupees 50 (US$ 0.80).
We found infrequent condom use, bisexual lifestyles, universal commercial sex, high STI prevalence and dissociation between protective knowledge and actions and STIs among Hijras. These factors, their high numbers and frequency of unprotected sexual encounters and increasing HIV prevalence in some cities3 confirm Hijras as a core group in Pakistan.10 Female gender identity and their unique social organizational structure1,5,11 differentiate them from other MSWs or MSM with whom they are usually grouped by prevention programmes.
A significant minority of our subjects indulged in sex with women. Until now Hijras have been considered to have sex with men only and are, therefore, considered to have a limited role in HIV/STI transmission. Our findings broaden the scope of Hijras’ sex networks by linking women into these networks. These findings may reflect transvestites living among Hijras, individuals still transitioning in their gender identity or lack of rigid boundaries for gender identity. However, harm-prevention programmes for Hijras must account for these individuals.
Nearly all Hijras engaged in commercial sex with per-act fees that place them at the bottom of the commercial scale for sex trade.2 Most reported physical abuse, sexual violence or coercion, and discrimination due to their sexual orientation. These factors limit Hijras\' ability to control the terms of protection during sex by either reducing the number of clients they must engage or by negotiating condom use.
Using standardized questions,12 we found that appropriate prevention knowledge neither improved condom use nor reduced STIs. This may reflect Hijras\' inability to control the terms of protection during sexual encounters or that knowledge alone may not lead to effective actions. Alternatively, these questions may be inadequate for the purpose. Our findings emphasize the need for better measures of protective knowledge, accurate predictors of actions and for programmes or enabling environments that empower Hijras to prevent HIV by effective preventive actions.
High STI prevalence (particularly syphilis) among Hijras in Pakistan is consistent with international experience.13 Nearly all syphilis was acute, likely reflecting our subject\'s young age and possibly their recent admission into this group. Frequent acute and multiple infections reflect very high infection incidence and persistence; suggesting high partner rate, participation in concordant sex networks (with other Hijras or a circumscribed group of clients who seek sex with Hijras) and/or lack of timely treatment of infections, against a backdrop of high infection prevalence. Further research should explore the contributions of sexual network dynamics, re-infections and lack of treatment among Hijras.
Syphilis, anal gonorrhoea and anal chlamydia, which are usually asymptomatic,14-16 accounted for most infections and may explain the infrequent current symptoms or positive examination findings despite very high STI prevalence. Symptoms drive care-seeking, and timely care-seeking for STIs may reduce STI transmission within a community by bringing clients into contact with those working on prevention. By the same logic, asymptomatic infections impede prevention programmes. Future interventions should address such asymptomatic infections through regular check-ups among clients of HIV/STI prevention programmes.
The study examined the sexual risk of Hijras to design effective HIV prevention programmes for them. One limitation of the study was that all self-identified Hijras were included. This may include zenanas with yet-to-mature gender identity, transvestites living among and acting like Hijras or an identity overlap as suggested by anecdotes of gurus fathering children. Inclusion of all of the above is functionally acceptable since Hijra harm-reduction interventions must reach all these individuals. Another limitation was that anal samples from Lahore were all negative. This was due to a single accident with storage of anal samples at the laboratory in Lahore. While extrapolated data from Karachi suggest similarly high levels of anal STIs in Lahore, this cannot be stated definitively.
Our findings suggest high sexual risk among Hijraswhose networks seemto involve men and somewomen. This enhances their significance in the overall HIV transmission scenario in Pakistan. Their communal living practices distinguish them from other MSWs and hence the approaches and messages that appeal to them may be different. Future research must explore their sexual networks and social structure as a means to deliver prevention interventions and improved ways to translate protective knowledge into effective actions and better outcomes.
Funding was received from the Department for International Development (DFID) for HIV control work in Pakistan.
The National AIDS Control Programme provided the study objectives, technical supervision, logistical and financial monitoring, data management, analysis and writing of this manuscript. The initial study design was prepared by the London School of Hygiene and Tropical Medicine, the Population Council of Pakistan and the Aga Khan University, Karachi, Pakistan. Data were collected by the Pakistan Medical Research Council under a subcontract from the Family Health International (Pakistan).
We are particularly indebted to Irfan Masood for help with data management and to Dr J Dennis Fortenberry (Indiana University) for reviewing the manuscript and providing valuable suggestions and comments.
1. Nanda S. Neither Man Nor Woman: The Hijras of India. Belmont, CA, USA: Wadsworth Publishing, 1990.
2. The National AIDS Control Programme MoH, The Family Health International, The Pakistan Medical and Research Council. The National Study of Sexual and Reproductive Tract Infections. See http://www.fhi.org/NR/rdonlyres/eawwvbpc6n4y ccbal2b73hndlgafk54ctb2q3qh5yrtqvao 5y67ybc2mjqevq6rquevj2lctpredre/RTIStudyPakistan3enhv.pdf 2004 (last accessed 23 September 2008).
3. HASP Team, The National AIDS Control Programme MoH. HIV/AIDS Surveillance Project of Pakistan, Round 1. See http://www.nacp.gov.pk/ wp-content/uploads/2007/04/HASP-Round1report-final.pdf 2005 (last accessed May 2008).
4. Baqi S, Shah SA, Baig MA, Mujeeb SA, Memon A. Seroprevalence of HIV, HBV and syphilis and associated risk behaviours in male transvestites (Hijras) in Karachi, Pakistan. Int J STD AIDS 1999;10:300-4.
5. Jami H. Conditions and Status of Hijras (Transgenders, Transvestites etc.) in Pakistan. Islamabad, Pakistan: National Institute of Psychology, Quaid-i-Azam University, 2005.
6. Mithani A, Burfat FM. Hijra - the sex in between. J Independ Stud Res 2003;1. See http://jisr.szabist.edu.pk/servlet/ JISRMSSE.Download?Location¼030303 HijraTheSexinBetween.pdf (last accessed 23 September 2008).
7. UNAIDS. AIDS Epidemic Update; See http://www.unaids.org/epi/2005/doc/ report_pdf.asp 2005 (last accessed May 2008).
8. Pinkerton SD, Layde PM. Using sexually transmitted disease incidence as a surrogate marker for HIV incidence in prevention trials. Sex Transm Dis 2002;29:298-307.
9. Bokhari A, Nizamani NM, Jackson DJ, et al. HIV risk in Karachi and Lahore, Pakistan: an emerging epidemic in injecting and commercial sex networks. Int J STD AIDS 2007;18:486-92.
10. Thomas JC, Tucker MJ. The development and use of the concept of a sexually transmitted disease core. J Infect Dis 1996;174:S134-43.
11. Pappas G, Khan O, Wright JT, Khan S, Kumaramangalam L, O\'Neill J. Males who have sex with males and HIV/AIDS in India: the hidden epidemic. AIDS Public Policy 2001;16:4-17.
12. UNAIDS. Guidelines for the Development of Core Indicators; See http://data.unaids.org/pub/Manual/2007/20070411_ ungass_core_indicators_manual_ en.pdf 2005 (last accessed 23 September 2008).
13. Pisani E, Girault P, Gultom M, et al. HIV, syphilis infection, and sexual practices among transgenders, male sex workers, and other men who have sex with men in Jakarta, Indonesia. Sex Transm Infect 2004;80:536-40.
14. Quinn TC, Goodell SE, Mkrtichian E, et al. Chlamydia trachomatis proctitis. N Engl J Med 1981;305:195-200.
15. Quinn TC, Stamm WE, Goodell SE. The polymicrobial origin of intestinal infections in homosexual men. N Engl J Med 1983;309:576-82.
16. Jones RB, Rabinovitch RA, Katz BP. Recovery of Chlamydia trachomatis from the pharynx and rectum of heterosexual patients at risk for genital infection. Ann Intern Med 1985;6:757-62.
This journal is a member of and subscribes to the principles of the Committee on Publication Ethics.