Risk factors for HIV infection in people attending clinics for sexually transmitted diseases in IndiaBMJ 1995; 311 doi: https://doi.org/10.1136/bmj.311.7000.283 (Published 29 July 1995) Cite this as: BMJ 1995;311:283
- Jeanette J Rodrigues, coprincipal investigator,
- Sanjay M Mehendale, study coordinatora,
- Mary E Shepherd, data analysta,
- Anand D Divekar, investigatorb,
- Raman R Gangakhedkar, investigatora,
- Thomas C Quinn, investigator,
- Ramesh S Paranjape, investigatorc,
- Arun R Risbud, investigatora,
- Ronald S Brookmeyer, biostatisticiana,
- Deepak A Gadkari, investigatorb,
- Manjusha R Gokhale,, investigatora,
- Anne M Rompaloa,
- Shridhar G Deshpande, consultantb,
- M M Khalandkar, consultant,d,
- Nita Mawar, consultante,
- Robert C Bollinger, principal investigatorb
- aNational AIDS Research Institute, Pune, India
- bJohns Hopkins University, Baltimore, MD, United States
- cNational Institute of Allergy and Infectious Disease, National Institutes of Health, Bethesda, MD, United States
- dBJ Medical College and Sassoon General Hospital, Pune, India
- aDr Kotnis Health Centre, Pune Municipal Corporation, Pune, India
- Correspondence to: Dr R C Bollinger, Division of Infectious Diseases, Johns Hopkins Medical School, Ross Building 1159, Baltimore, MD 21205, United States.
- Accepted 24 May 1995
Objective: To investigate the risk factors for HIV infection in patients attending clinics for sexually transmitted diseases in India.
Design: Descriptive study of HIV serology, risk behaviour, and findings on physical examination.
Subjects: 2800 patients presenting to outpatient clinics between 13 May 1993 and 15 July 1994.
Setting: Two clinics and the National AIDS Research Institute, in Pune, Maharashtra State, India.
Main outcome measure: HIV status, presence of sexually transmitted diseases, and sexual behaviour.
Results: The overall proportion of patients infected with HIV was 23.4% (655/2800); 34% (184) of the women and 21% (459) of the men were positive for HIV infection. Of the 560 women screened, 338 (60%) had a reported history of sex working, of whom 153 (45%) were infected with HIV-1. The prevalence of HIV-1 infection in the 222 women who were not sex workers was 14%. The significant independent characteristics associated with HIV infection based on a logistic regression analysis included being a female sex worker, sexual contact with a sex worker, lack of formal education, receptive anal sex in the previous three months, lack of condom use in the previous three months, current or previous genital ulcer or genital discharge, and a positive result of a Venereal Disease Research Laboratory test.
Conclusions: In India the prevalence of HIV infection is alarmingly high among female sex workers and men attending clinics for sexually transmitted diseases, particularly in those who had recently had contact with sex workers. A high prevalence of HIV infection was also found in monogamous, married women presenting to the clinics who denied any history of sex working. The HIV epidemic in India is primarily due to heterosexual transmission of HIV-1 and, as in other countries, HIV infection is associated with ulcerative and non-ulcerative sexually transmitted diseases
Genital ulcer diseases are common in such clinics and wereassociated with a higher risk of HIV infection
High risk sexual behaviour is also common among clinic attenders
The risk factors for HIV infection and the increasing prevalence ofHIV-1 suggest a similar pattern in India to that seen in epidemics inother countries with high rates of transmission of HIV
There is an immediate need for comprehensive and national efforts tocontrol sexually transmitted diseases and to provide intensive educationon HIV and AIDS targeted at changing high risk behaviour in India
Since the first reported cases of HIV infection diagnosed in India a considerable increase in the prevalence of HIV infection has been detected in high risk populations of sex workers and attenders of clinics for sexually transmitted diseases, particularly in the western state of Maharashtra.1 2 3 4 5 We present the first description of the data from an ongoing investigation to identify the risk factors for HIV-1 infection in these high risk populations.
Subjects and methods
Between 13 May 1993 and 15 July 1994 patients attending two clinics for sexually transmitted diseases in Pune, India, were offered serological screening for the presence of HIV infection. After informed consent, participants answered a structured questionnaire which included data on demographics, medical history, sexual behaviour, reproductive history for female patients, and knowledge of HIV and AIDS. Each participant was then examined by a physician for the presence of sexually transmitted diseases and given appropriate antibiotics.
Serum samples were initially screened with a commercially available enzyme linked immunosorbent assay (ELISA) for detection of antibodies to HIV-1 and HIV-2 (Recombigen HIV1/HIV2, Cambridge Biotech, Galway, Republic of Ireland). Specimens that yielded positive results by this method were confirmed by using a rapid test for HIV-1 and HIV-2 (Recombigen HIV1/HIV2 Rapid Test Device, Cambridge Biotech) and by HIV-1 or HIV-2 western blot (Cambridge Biotech). Syphilis serology was determined by standard Venereal Disease Research Laboratory non-treponemal test (VDRL antigen, Span Diagnostic, Surat, India) on all samples. The presumptive clinical diagnosis of sexually transmitted disease was based on a detailed physical examination and was made by the examining physican without knowledge of whether the participants were infected with HIV.
The proportion of patients who were positive for antibodies to HIV-1 at screening was calculated by age, sex, parity, occupation, and other biological, demographic, and clinical variables and were compared by using statistics and Fisher's exact test. Odds ratios for seropositivity and confidence intervals were computed. Multivariate analyses were performed with logistic regression to determine characteristics independently associated with prevalent seropositivity. The variables included in the logistic regression were those generally significant in the univariate analysis. In addition, several variables that were highly correlated were either combined (for example, previous or current genital ulcer) or excluded (for example, number of lifetime sexual partners, which was correlated with sex working).
Of the 2800 patients screened for HIV-1 and HIV-2, 655 (23.4%) were initially positive for antibodies to HIV, of whom 609 (93%) were positive for HIV-1. Only 12 (0.4%) of 2800 were positive for HIV-2, and 34 (1.2%) were dually reactive. Four hundred and fifty nine (20.5%) of 2240 men and 184 (32.9%) of 560 women were positive for HIV-1 or dually reactive. Table I shows the univariate analysis of risk factors for HIV-1 infection. Sixty per cent (338) of the women were sex workers, and 90% (2010) of the men reported a history of contact with a sex worker within the previous three months. Our analysis identified a high prevalence of HIV-1 of 14% (31/222) in women who were not sex workers. Of these 222 women, 183 (82%) were monogamous and married, of whom 24 (13%) were positive for HIV-1. Twenty of these women presented with genital symptoms, and two presented as contacts of men with a sexually transmitted disease. Other potential risks for transmission of HIV-1 were rare, with use of injected drugs reported in only one participant and a total of 55 (2%) of 2978 participants reporting a history of blood transfusion.
Only 24% (682) of the men and women reported any use of condoms during the previous three months, and 84% (2343) of the patients presented to the clinics for relief of active genital symptoms. Fifty three per cent (1471) of the participants reported a history of a genital ulcer, and 27% (761) a history of genital discharge. Chancroid was the most common clinical diagnosis among both men and women (33% (739) and 20% (112), respectively).
Other factors associated with a higher prevalence of HIV-1 infection in the univariate analysis included older age, not living with family, receptive anal sex within three months, lack of formal education, lack of knowledge of AIDS, use of condoms within three months, lifetime number of sexual partners, having received a tattoo since 1985, a self reported history of genital ulcer or discharge at the time of screening, and a positive result of a Venereal Disease Research Laboratory test. After sex working and other risk behaviours were controlled for, users of condoms had a significantly lower risk of HIV-1 infection.
Table II shows the adjusted odds ratios for characteristics independently associated with prevalent HIV-1 infection based on logistic regression. These include being a female sex worker, men who have had contact with a sex worker, lack of formal education, receptive anal sex, having received a tattoo since 1986, current or previous genital ulcer or genital discharge, and a positive result of a Venereal Disease Research Laboratory test.
Since 1986 the prevalence of HIV infection in female sex workers and patients attending clinics for sexually transmitted disease in India has continued to rise.6 7 Prevalence of HIV infection among sex workers in Pune has increased from 6% in 19892 to over 45% in our current 1993-4 study. By 1992 the prevalence of HIV among patients attending clinics for sexually transmitted disease in Bombay had increased to 11% and was 10% in Pune,2 4 and overall was 23% in our study. Our data identified two key groups at risk: female sex workers and men who had contact with sex workers. Also of concern is the high prevalence of HIV-1 (13%) observed in monogamous, married women, suggesting that the HIV epidemic in India has begun to affect other risk groups.
The association of other sexually transmitted diseases and HIV infection has been previously reported by numerous studies.8 This may be particularly important for India in light of some studies reporting that 50-60% of patients attending urban clinics for sexually transmitted disease in India have genital ulcer disease9 10 and serological studies showing that up to 9% of some high risk groups yielded positive results on a Venereal Disease Research Laboratory test.11 12 13 In our study, a positive result was present in 21% of the men and 22% of the women. A significant proportion of the participants in our study had a current or previous history of genital ulcer disease or genital discharge, which were found to be independently associated with prevalent HIV-1 infection. In addition, high risk sexual behaviour, including lack of condom use and high number of sexual partners, is common and associated with prevalent HIV-1 infection in patients attending clinics for sexually transmitted disease in Pune.
Although previously described risk factors for HIV infection, such as homosexuality, use of injected drugs, or transfusion, were rare in our study, 13% of the men described themselves as bisexual, and receptive anal intercourse was independently associated with HIV infection. In addition, 23% of the men and 29% of the women had received a tattoo since 1986, which was also found to be an independent risk factor for prevalent HIV-1 infection in our study. Sex specific analysis suggests that tattooing is associated with sex working in women, but the association of tattooing and HIV infection in men seems to be independent of other risk factors. Our data alone are insufficient to explain this finding, but it suggests that further study of the potential for HIV-1 transmission in these settings by contaminated needles or equipment may be warranted in India.
Since the first descriptions of the HIV epidemic in India, there have been reports of HIV-2 infection in western India.14 Several studies have suggested that HIV-2 infection may represent 3% to 9% of all HIV infection in Bombay and other parts of Maharashtra.14 15 16 These studies have also reported dually reactive serology in 6% to 29% of all HIV infection. One report has suggested a higher rate of HIV-2 infection, approaching 16% in sex workers in Bombay.17 Our study has confirmed that the prevalence of HIV-2 (1.2%) remains much lower than that of HIV-1 (23%) in clinics for sexually transmitted disease in Pune.
These data provide a comprehensive analysis of prevalent HIV-1 infection in one of the largest studies of people at high risk in India. We have provided current evidence of the association between sexually transmitted diseases, high risk sexual behaviour, and HIV-1 infection in India. These findings strongly suggest that the pattern of the HIV epidemic is similar to that seen in other parts of the world. The dramatic increase in prevalence of HIV-1 in these high risk groups suggests there is an urgent need for comprehensive and national efforts in India to control sexually transmitted diseases and to provide intensive education on HIV and AIDS targeted at changing high risk behaviour.
We thank Dr V D Jadhav, Dr Sule, Dr Tolat, Mahendra Gaikwad, John Kurian, Ram Kate, Sangeeta Patil, Sharmila Kate, Hemalata Mahajani, Varsha Navalgundkar, Jessy Thomas, Manisha Iyer, Suvarna Gurav, and M Attar for the excellent counselling, data collection, laboratory work, and clinical care of study participants. We also thank Dr Smita Kulkarni, Madhuri Thakar, B Surve, and Komal Sukamaran of the National AIDS Research Institute for laboratory assistance; and Mark Moss and Richard Kline of the Johns Hopkins University for laboratory assistance. We offer special thanks to Fabian Swamy and Radhika Pathak for their excellent data management.
Funding This work was supported by the National Institute of Allergy and Infectious Diseases, National Institutes of Health NIAID (AI 33879-02), NIH-NCRR OPD-GCRC (5M01RR00722), and the NIH-Fogarty International Center Program of International Training Grants in Epidemiology Related to AIDS (D43 TW0000).
Conflict of interest None.