April 2000, Volume 50, Issue 4

Original Article

Multivariate Analysis of Risk Factors Associated with Genital Ulcer Disease among Incarcerated Males in Sindh

S. Akhtar  ( Division of E pidemiology and Biostatistics, Department of Community Health Sciences Medical College, The Aga Khan university, Karachi. )
S.P. Luby  ( Division of E pidemiology and Biostatistics, Department of Community Health Sciences Medical College, The Aga Khan university, Karachi. )
M.H. Rahbar  ( Division of E pidemiology and Biostatistics, Department of Community Health Sciences Medical College, The Aga Khan university, Karachi. )


Objective: To evaluate the potential risk behaviors associated with the lifetime risk of self reported genital ulcer disease (GUD) among prison inmates.
Setting: Prison inmates from 14 prisons of Sindh Province.
Methods: A cross-sectional study was conducted on 3395 prison inmates during July to December, ‘1994. A questionnaire was used to assess the lifetime risk of self-reported GUD (whether or not the subject was ever affected with CUD up to present age) and to investigate demographic markers and risk behaviors for their possible association with lifetime risk of CUD using logistic regression analysis.
Results: The reported lifetime risk of CUD in the study sample was 11.4% (386/3395). In final multivariate logistic regression model the sexual behaviors which were independently associated with CUD were having sexual intercourse with female (adjusted OR = 1.7;95% CI: 1.3-2.3, P = 0.0002), sexual intercourse with a prostitute (adjusted OR = 1.5;95% Cl: 1.2-2.0, P = 0.0008), sexual intercourse with man (adjusted OR = 2.2;95% CI: 1.7-2.7, P = <0.001) and sexual intercourse with man during current incarceration (adjusted OR = 1.9;95% CI: 1.2-2.9, P = 0.0071).
Conclusion: Health education needs to re-enforce monogamous relationship for high risk groups such as in our study. Although infrequent condom use was not a risk factor for CUD in this study, yet based on the results of previous studies, promotion of condom use should be the component of health education program GPMA 50:115, 2000).


Sexually transmitted diseases (STDs) have been identified as a major public health problem world over1, particularly in developing countries, where resources for their management are limited2. The efforts to control STDs in developing countries are hampered by insufficient number of specialist STD clinics to cater for the needs of population, the non-availability of suitable diagnostic facilities and appropriate drugs1,3. Failure to provide effective treatment for patients with STDs will ultimately lead to continued spread of disease, high rates of complications and in additions enhanced rates of sexual transmission of human immunodeficiency virus (HIV)4-10.
Among the STDs genital ulcer disçase (GUD) is a frequent problem in men attending the STDs clinics” and remained consistently high over the past ten years, contributing approximately 25% to the total case load seen in developing countries12,13. Sixty nine percent of all cases of GUD are in 20 to 34 year old age group and 3% in those older then 45 years2. The major causes of GUD include Treponema pallidum (syphilis), donovanosis (granuloma inguinale), genital herpes and Haemophilus ducreyi infection1,11.
It has been shown that genital ulcers were significantly associated with HIV seropositivity in Nairobi14,15. Further studies in homosexual populations have confirmed the association of genital ulcers with HIV transmission8. GUD has been shown to be independently associated with immunosuppression among HIV positive women, but this relationship is complex and may be in part cause and part effect16. Nevertheless, there is convincing evidence that genital ulceration increases the susceptibility to HIV infection and compared with men with other sexually transmitted diseases those with genital ulceration had increased risk of seroconversion against HIV9.
The prevention programs for STDs transmission is becoming a priority of public health departments in developing coutnries17. Such prevention programs would have to be based on the knowledge of modifiable risk behaviors associated with acquisition of STDs among members of high risk segments of population. Previous studies have shown that risk of acquiring a STD increases with the number and type of sexual contacts and lack of condom use18,19. The differences have been shown in the risk behaviors associated with gonorrheal and chalmydial infection20. Contrary to these findings, risk factors for developing either GUD or any other STD were shown to be similar2. In another study, men with GUD compared to those with other STDs have had increased frequency of sexual intercourse with commercial sex workers, history of STD symptoms in the past 12 months and a recent sexual contact. Infrequent condom use was not a risk factor for this population21,22. However, other studies have shown the effectiveness of condoms in reducing the risk of STD transmission23. These inconsistent results of different studies regarding the risk factors for GUD in different populations prompted us to conduct the present study.
Some published data on different high risk groups concerning epidemiology of other STDs are available from Pakistan24. However, to our knowledge no study so far has determined the risk factors for GUD among incarcerated males, a known high risk group that may pose a risk of STDs transmission in general population on completing their prison term. The objective of this study therefore, was to identify the risk factors associated with GUD in a systematic sample of incarcerated male inmates from prison System of Sindh, Pakistan, using rnultivariate analysis.


Study Subjects and Data Collection
The study setting sampling technique used to select the study population has been described elsewhere, Briefly, 3395 male prison inmates were included in the present analysis. They were selected using one-in-three systematic sampling technique from among 10600, male prisoners incarcerated in judicial custody as indicted criminals in 14 prisons of Sindh during July 1994. The subjects interviewed comprised mainly of two self-identified ethnic groups. The subgroups were identified based on their mother-tongue i.e., Sindhi, Urdu. However, a small proportion also comprised other ethnic groups. Inmates were eligible to participate in the study, if they spoke Urdu, being a national language. Because of the varying literacy levels of the prison inmates, a structured risk behavior interview was administered to each study subject in confidence by trained research interviewer in a private area within the prison. The interview focused on seeking information on demographic, sexual and drug use behaviors during the subject’s lifetime up to his present age. The questions on sexual behavior solicited information on number and type of sex partners, homosexuality both before and after incarceration, condom use and illicit drug use. Inmates were asked, if they had painful genital ulcer disease in the past (lifetime risk of genital ulcer disease occurrence i.e., if the respondent ever had this condition up to his present age). The question concerning the past history of genital ulcer disease was phrased to deliver a concise description of the common signs and symptoms associated with genital ulcers of different origins11,25. Specifically, the question asked was: have you ever had a painful genital ulcer disease?
Ethics and Confidentiality
Informed verbal consent of each study was sought and to ensure frank and complete answers, they were assured about complete confidentiality of all interview questionnaire responses. This study was approved by the Aga Khan University’s Committee for Human Subjects Protection.
Data Analysis
For all analyses, the dependent variable, lifetime risk of genital ulcer disease occurrence had two categories: ever affected and never affected. We categorized the continuous variables such as age and duration of imprisonment into quartiles to reduce the influence of outliers. Frequencies (%) of demographic variables and sexual behaviors were computed26. The relationship between the dependent variable and the independent variables was examined by using two-way and multi-way contingency comparisons; the c2 test was used to compare proportions27. The crude measure of association between a single putative risk factor and inmates genital ulcer disease status was expressed as the odds ratio (OR) and the corresponding 95% confidence intervals (CE) was derived by means of first-order Taylor series approximations method26.
A multivariable logistic regression model was used to estimate the effect of each variable on the lifetime risk of genital ulcer disease adjusting for the effects of other variables in the model. For multivariate analysis, a full model was specified with all independent variables significantly (p<0.1) related with outcome variable in un ivariate analyses. Backward stepwi se multiple logistic regression analysis was carried out to arrive at the final multivariable model relating the variables simultaneously to the lifetime risk of genital ulcer disease28. in addition to significant (p<0.1) main effects, identified through univariable analyses, some interaction terms were considered for possible inclusion in the final model. Selection of the final model was based on parsimony, biological interpretability and statistical significance. The parameters of the logistic regression model were estimated by the maximum-likelihood method. The adjusted odds ratios (ORs) and their 95% confidence interval (Cis) were computed using the estimates of parameters of final logistic regression model and were the main focus for substantive interpretation of the model. In all the analyses 5% significance level (a=0.5) was used unless stated otherwise. All the analyses were carried out by using SPSS/PC windows version 7.5 (SPSS Inc. Chicago, IL. USA).


The distribution of study subjects with respect to demographic variables including age, ethnicity, education, marital status, duration of imprisonment and reported risk behaviors is given in Table 1. The reported lifetime risk of GUD in the study sample was 11 .4% (386/3395). On unadjusted analysis being Sindhi and duration of imprisonment were significantly (p<0.001) associated with lifetime risk of GUD (Table 1). All risk behaviors considered except ever injected drugs, ever shared needles and ever having use condom during sexual intercourse were significantly (p<0.001) associated with lifetime risk of GUD for this study population in univariate analysis (Table 1).

Multivariable ‘ogistic regression model
The sexual behaviors which were independently associated with GUD in final multivariable logistic regression model were having had sexual intercourse with female (adjusted OR = 1.7;95% CI: 1.3-2.3, P = 0.0002), had a sexual intercourse with a prostitute (adjusted OR = 1.5;95% CI 1.2-2.0, P = 0.0008), had a sexual intercourse with man (adjusted OR = 2.2;95% CI: 1.7-2.7, p = <0.001) and sexual intercourse with man during current incarceration (adjusted OR 1 .9;95% Cl: 1.2-2.9, p = 0.0071). Finally ethnicity and duration of imprisonment were not significantly associated with GUD.


Recent evidence suggests that STDs facilitate the transmission of HIV29. Among the STDs GUD has been recognized as a major factor in HIV transmission8,30 and a proportion of HIV infections in men attributable to GUD as high as 75-98% has been reported in Africa31. Recently, effective treatment of GUD and other STDs has been shown to reduce the incidence of HIV infection in Africa29. The etiology of GUD varies both geographically and temporally33-34. The primary agents causing GUD in STD clinic patients are Treponema pallidum, Haemophilus ducreyi and herpes simplex virus35. Laboratory tests for the detection of these organisms are relatively insensitive, costly, technically sophisticated, time consuming and are often not available in clinics where GUD patients are seen in developing countries36,37. These difficulties in assessing the etiologic causes of GUD in the developing countries are further compounded by the high incidence of all the sexually transmitted causes of GUD, mixed infections and atypical presentation of long-standing diseases38. Aforementioned diagnostic limitations have been partly overcome by the introduction of syndromic approach to GUD diagnosis and management worldwide. WHO has recommended this syndromic approach for GUD diagnosis in areas with limited resources39 and therefore, was employed in the present study.
It is known that the STDs are directly related to the patterns of sexual behavior and these patterns differ significantly within continents and even within countries40, we therefore, investigated the sexual behaviors associated with GUD among incarcerated male inmates in criminal justice system in Sindh, Pakistan using a cross-sectional study design.
In our final multivariate logistic regression model, the risk factors associated with GUD included independent effects of sexual intercourse with a female, sexual intercourse with a commercial sex worker, sexual intercourse with a man and sexual intercourse with a man during current incarceration. These findings consistent with those of other studies conducted in developing and developed countries21,22, revealed that men continue to engage in risky sexual practices by having unprotected sex with commercial sex workers and ultimately with casual sex partners and their wives in case of married inmates.
Furthermore, the subjects in our study are quite heterogeneous with respect to their sexual behaviors. It has shown that the heterogeneity in sexual behavior as measured by the rate of sex partner change adds substantially to the reproductive rate of the causative agents and thus the likely future rate of growth of an epidemic40. Thus characterization of the heterogeneity of sexual behavior plays an extremely influential role in determining both the course of STDs epidemic and choice of control strategy41. Infrequent use of condom was not a risk factor for GUD in our study population. However, other studies have shown the effectiveness of condoms in reducing the risk of STDs transmission23.
The results of this and previous studies42, showed that there is high prevalence of STDs among the incarcerated. Thus, health care providers at correction system need to increase the efforts at STDs prevention and treatment in the criminal justice system, which may help in suppressing the HIV epidemic. The health education needs to re-enforce monogamous relationships and the use of condoms to reduce the risk of STDs transmission.
The treatment seeking behaviors in these study subjects with GUD is of great concern and could not be addressed in this study. The factors influencing these behaviors may be socio-economic in nature including inability to afford treatment, failure to perceive that they have illness that requires treatment. Further research is required on the factors, which could possibly influence such decisions and to devise possible strategies to change the risky behaviors.


1.Htun Y, Morse SA, Dangor Y, et al. Comparison of clinically detected, disease specific and syndromic protocols for the management of genital ulcer disease in Lesotho. Sex Transm. lnf.. 1998:74(Suppl 1):S23-S28,
2.Wellington M, Ndowa F, Mbengeranwa L. Risk factors for sexually transmitted disease in Harare: A case-control study. Sex Transm. Dis.,1997;34:528-32.
3.Adler MW. Sexually transmitted disease control in developing countries. Genitourin. Med., 1996:72:83-88.
4.Greenblatt RM, Lukehart SA, Plummer FA, et al. Genital ulceration as a risk factor for human immunodeficiency virus infection. AIDS, 1 988:2:47-50.
5.Holmberg SD, Stewart JA, Gerbev AR, et al. Prior herpes simplex virus type
2 infection as a risk factor for HIV infection, JAMA, 1 988:259:1048-50.
6.Simonson JN, Cameron DW, Gakinya MN, et al. Human inimnodeficiency virus infection among men with sexually transmitted disease. N. EngI. J. Med., 1988:319:274-78.
7.Stamm WE, Handsfield HH, Rompallo AM, et al. The association between genital ulcer disease and acquisition of HIV infection in homosexual men. JAMA, 1988;260: 429-33.
8.Piot P, Laga M. Genital Ulcers, other sexually transmitted diseases and the sexual transmission of HIV. Br. Med. J., 1989:298:623-29.
9.Wasserheit JN. Epidem ological synergy: interrelationships between human immunodefieiency virus infection and other sexual ly transmitted diseases. Sex Transmdis., 1992; 19:61-77.
10.Ghys PD, Diallo MO. Etticgnc - Traore V, ci al. Genital ulcers associated with human immunodeficiency virtts—rclatcd immunosupprcssion in feitiale sex workers iii Abidjan Ivory coast. 3 Infect. Dis., 1995:172:1321-29.
11.O’Farraell N, Hoosen AA, Coetzee KD, et al. (lettital ulcer disease in men in Durban, South Africa. Genitouritt. Med., 1991 ;67:327-30.
12.City of Harare. Annual Report of site City Health Department 991. Harsre, Zimbabwe, City Health Department 1991.
13.City of Harare. Annual Report of she City Health Department 1992. Harare, Zimbabwe, City Iealth t)cpartmcnt, 1992.
14.Piol P, Plummcv PA, Mhaltt PS, et at. All): An international perspective. Science, 1988:239:573-79.
15.Simonsen JN, Cameron Dlv, Gakinva MN, et al. Hutnan iintnuncdcficicncy view type 2 infection as a risk factor for lily infection. JAMA, 1988:259 1048-50.
16.Camerott DW, Ngugi EN, Ronald AR, et al. Condont use prevents genital ulcers in women working as prostitotes. Intlttence of humatt utununodeficicncv virus itttèction. Sex Traitsttt. Dia., 1991:18: 1 88-91.
17.Lags M, Mattoka A, Kivuvtt M, et al. Non-ulcerative STDs as a risk factor for lily—I transmission in women. Restilts from a cohort study. AIDS, 1993:7:95­-102.
18.Ds—Costs Li, Plummer PA, l3oumer I, ci al. Prostitutes are a major sotirce of sexually transmitted diseases in Nairobi, Kenya. Sex Transm. Dis., 1985:12:64-67.
19.Hooper R, Reynolds Gi-l, Jones OG, et at. Cohort study of venereal disease: TIte risk of gonorrhoea trattsmission front women so matt. Am. J. Epidemiol., 1978; 108: 136-44.
20.Hook EW, Reichart CA, Upehurelt DM, ci al, Comparative behavioral epidemiology of Gonococal and Chlamidial infections among patients attending a Balsitttore, Marylattd, Sexually Transmitted Disease clinic. Ant. J. Epidcntinl.. 1992:136:662-72.
21.Upehurch DM, Brady WE, Reichart CA, ci at, Behavioral contributions to acquisition of gonorrhoea in patients attending iiiner city sexually transmitted clinic. J. litfeet. Dis., 1990; 161:938-11.
22.Malone JD, Hyams KC, Hawkins RE, et al. Risk factors for sexually transmitted diseases ataong deployed U.S. military personnel. Sex Transin. Dis., 1993:20:294-98.
23.Ngugi EN, Plummcr FA, Simansen JN, ct al, Prevention and transmission of huntan im in unodefi ciency virus in Africa: EtTcctivcness of condom promotion and health education atnong prostitutes. Lancet. 1988:2:887-90.
24.Baqi 5, Nabi N, Hasan SN, ci al. BIV antibody seroprevalence and associated risk factors in sex workers, drug users and prisoners in Sindh. Pakistan. J. Acquir. Immun Defic. Syndr. Hum. Retrovir., 1998:18:73-79.
25.Wiestier P, Browtt S. Kraus S. et al. Genital ulcers, In: Holmes KK, MardI PA, ed International perspectives on neglected sexually transtnitted diseases’ Impact (in venercologv, in ten I its’ and maternal al and infant Iteal th. Washington DC, Heistispherc Publishing Corporation. 1983, pp. 2 19-3-I.
26.Rothman Ki, Moderti Epidemiology Boston, Little Brown and Cotstpanv. 1996, p. 358.
27.Roster B. Fundamentals of Biostatistics t3oston, Duxburv Press, I 986, p. 682.
28.Hosmer DW, Leineshow S Applied Regression Analysis New York, John Wiley. 1989. p. 307.
29.Grosskurth H. Mosla F, Todd i, ci al. littpaet of improved sexually transmitted diseases on I llV infeetiott in rural Tanzania Raitdotttized eotttrolled trial. Lancet, 1995:346.530-36.
30.Jessamine PG. Pluntitter PA, Ndiny a—Aehola JO. et al, I lumatt tttttttuttodeiieuenev virus, genital (leers and the tale foreskin: Synergism itt DIV—I transmission. Seartd i. Infect Os. (, Suppl). 1990.69:18 1-86.
31.I ayes R. Schulz KF. What proportion of I ltV infection are attributable to genital ulcers in sub-Saharan Africa? VIII Ittlernattonal Conference ott AIDS, Abstract No. MoC 0tt29. Amsterdam, 19-24 July, 1992.
32.Morse SA, Chattcroid and Haemophilus ducrvi nit update. Clin. Microbiol. Rev., 1989:2:137-57.
33.Trees Dl, Morse SA. Chancruid aid I laeittophilus dnerevi. Au iipdtite. Cliii. Mierobiol. Rev., 1995:8:357-75.
34.orle KA, Gates CA. Martin DH, ci al Sinsultaneous PCR detection of Haetnopltilus ducrevi, Treponema pallidum aitd herpes simplex viruses type I attd type 2 front genital ulcers J. Cliii. Mierobiul., 1996:34:49-54.
35.Bogaerts 3, Rieart CA, Van Dvck P. ci al. TIte etiology if genital ulceration in Ravanda. Sex. Tm’anstti. Dis., 1989:16 23-26.
36.Sclttilte 3M, March PA, Settniid GR Chancroid in the Limited States 1981—90. Evidence for underreporttitg of eases MMWR CDC Surveillance Sitittmn 1992;41~SS 31:57-61.
37.Larsen SA, I-Inter EF. MeGress BE, Syphilis. In. Wentworth BB, Jitdsoit EN, Gilchrist MJ, eds. Laboratorv methods for the diagnosis of sexually trattsmitted diseases. Washiitgtoit. DC, Aniericaii Public Health Association. l99l,pp. 1-52
38.O’Farrell N, Dorset AA, Coezee KD, et al. fiettital ulcer disease: Accutracy of clinical diagnosis attil strategies to ittprove control in Dsirbatt. South Africa, Genilourin. Med., 994, pp. 711:7—11.
39.World Health Organizattott Progratttne for sexually sransnsitted diseases. Global Progratttme on Al DS. Recent in etidatioti for the nat agetnettt of sexually trananitted diseases. WHO/GPA/TEM/94. Geneva, WHO. 1994.
40.Over M, Piot P. Human Immunodeficiency virus infections attd oiler sexually transmitted diseases in developitig cetititries: Public health itttpurtattee attd priorities for resource allocation. J. Infect. Ois . 1996: 74iSuppl 2):S162-75.
41.Anderson EM. Mathematical and structural studies of die epidemiology of AIDS. AIDS, 1989:33 333-46.
42.Glaser JB, Grei fi tiger R B. Currectiuttal heal lIt care: a public health opportunity. Ann ltttern Med , 1993: 11 8: 139-45.

Journal of the Pakistan Medical Association has agreed to receive and publish manuscripts in accordance with the principles of the following committees: