Secondary Logo

Share this article on:

A Nested Case–Control Study of Sexual Practices and Risk Factors for Prevalent HIV-1 Infection Among Young Men in Kisumu, Kenya

Mattson, Christine L. MS*; Bailey, Robert C. PhD, MPH*; Agot, Kawango PhD, MPH; Ndinya-Achola, J O. MBchB; Moses, Stephen MD, MPH§

doi: 10.1097/01.olq.0000261335.42480.89

Objectives: To investigate sexual practices and risk factors for prevalent HIV infection among young men in Kisumu, Kenya.

Goal: The goal of this study was to identify behaviors associated with HIV in Kisumu to maximize the effectiveness of future prevention programs.

Study Design: Lifetime sexual histories were collected from a nested sample of 1337 uncircumcised participants within the context of a randomized controlled trial of male circumcision to reduce HIV incidence.

Results: Sixty-five men (5%) tested positive for HIV. Multiple logistic regression revealed the following independent predictors of HIV: older age, less education, being married, being Catholic, >4 lifetime sex partners, prior treatment for an STI, sex during partner's menstruation, ever practicing bloodletting, and receipt of a medical injection in the last 6 months. Prior HIV testing and postcoital cleansing were protective.

Conclusions: This analysis confirms the importance of established risk factors for HIV and identifies practices that warrant further investigation.

This study confirms established risk factors for HIV infection and identifies practices that warrant further investigation: sex during menstruation, blood letting, postcoital cleansing, and receipt of a medical injection.

From the *Division of Epidemiology and Biostatistics, School of Public Health, University of Illinois at Chicago, Chicago, Illinois; †UNIM Project, Kisumu, Kenya; ‡Department of Medical Microbiology, University of Nairobi, Nairobi, Kenya; and §Department of Medical Microbiology, University of Manitoba, Manitoba, Canada

The authors thank all of the participants, without whom this work would not have been possible. They are grateful to Evans Otieno, Nicholas Ouma, and the UNIM Project staff for their assistance in data collection and recruitment efforts and to Drs. Ronald Hershow, Richard Campbell, Donald Hedeker, and Supriya Mehta for their helpful comments on the manuscript.

This work was funded by the Division of AIDS, NIAID, NIH.

S. Moses is the recipient of an Investigator Award from the Canadian Institutes of Health Research.

Correspondence: Christine L. Mattson, MS, Division of Epidemiology and Biostatistics, School of Public Health, University of Illinois at Chicago, 1603 W. Taylor St., Chicago, IL 60612. E-mail:

Received for publication October 23, 2006, and accepted February 8, 2007.

BY THE END OF 2005, The Joint United Program on HIV/AIDS estimated that ∼38.6 million people (33.4–46.0) were living with HIV and over 60% of them resided in sub-Saharan Africa.1 Even though evidence suggests HIV incidence is declining in some parts of sub-Saharan Africa,1,2 it is estimated that 2.7 million people became infected with HIV in 2005,1 reaffirming the need to refine and implement improved prevention strategies. Thus, to evaluate risk factors for prevalent HIV-1 infection, we analyzed cross-sectional data from a nested sample of 1337 uncircumcised men aged 18–24 years participating in the context of a randomized controlled trial (RCT) of male circumcision (MC) to reduce HIV-1 incidence in Kisumu, Kenya.

Back to Top | Article Outline

Materials and Methods

To investigate the safety and effectiveness of MC to reduce HIV-1 incidence, we are conducting a RCT of MC in Kisumu, the Provincial Capital of Nyanza Province in western Kenya. In 2003, the overall prevalence of HIV infection among adults in Kenya was estimated to be 6.7%3,4 and the majority of HIV infections are thought to be acquired through heterosexual intercourse.1 To investigate sexual practices and risk factors for HIV, we recruited men who were screened in the RCT between March 2004 and September 2005. These included consenting seropositive and seronegative men who had experienced sex within the last 12 months, were uncircumcised, aged 18–24 years, and resident in Kisumu District. Interviews were conducted by men fluent in the 3 most common languages spoken in Kisumu (English, DhoLuo, and Kiswahili) and interviewers were blind to the HIV status of participants. This study received ethical approval from the institutional review boards at the University of Illinois at Chicago, the University of Nairobi, and the University of Manitoba.

To obtain comprehensive lifetime sexual histories, information about every sexual relationship since sexual debut was collected for up to 12 partners. Concurrent partnerships were determined by extrapolation from the dates obtained in the interview when the start and end dates of any 2 partners overlapped. All participants provided permission to obtain their HIV test results from the trial using their unique study identifier. Men were tested for HIV-1 using 2 parallel rapid tests: UniGold (Trinity Biotech, Ireland) and Determine (Abbott Laboratories, Inc., Abbott Park, IL). Men with concordant positive results were informed of their HIV status and were referred to a posttest support group, which provides counseling, treatment of opportunistic infections, income-generating activities, and referral to a collaborating clinic that provides comprehensive HIV/AIDS care, including the provision of antiretroviral therapy if indicated. Specimens from men with discordant rapid test results underwent further testing by double ELISA.

Associations between prevalent HIV infection and self-reported demographic characteristics and behaviors were evaluated singly and through multiple logistic regression analysis. Because of the small number of HIV-positive men, Fisher's exact tests were used to test associations between dichotomous variables. Associations between HIV infection and continuous variables were evaluated by independent t-tests or by the Wilcoxon rank sum test (depending on their distributions). A final model was built by adding demographic characteristics and behavioral risk factors that were significant at P <0.10 in bivariate analyses 1 at a time, and calculating log likelihood ratio tests to identify variables that significantly added to the model. Multiplicative effect modifiers and potential confounding factors were also assessed. The Hosmer and Lemeshow test was used to assess the goodness-of-fit of the final model and regression diagnostics were used to identify discriminative capacity and potential covariate patterns that were poorly fitted or highly influential.5 All statistical analyses were conducted in SAS Version 8.2 (SAS Institute Inc., Cary, NC).

Back to Top | Article Outline


Between March 2004 and September 2005, 1348 of the 2059 eligible men enrolled in the study, yielding an overall response rate of 65%. Eleven men were excluded from the analysis: 6 did not complete the interview, 3 were later determined to be ineligible, 1 was missing data on HIV status, and 1 was judged to be giving false answers. Thus, 1337 men were included in the final sample. For a complete description of the sample, see Table 1.





The median number of lifetime sex partners was 5 (IQR 3–9). Only 66 men (5%) reported 1 lifetime sex partner. Eleven hundred sixty-three men (87%) reported 12 partners or less, and thus, provided information on all of their lifetime partners. The median age of reported sexual debut was 15 (IQR 13–17). Receiving oral sex, performing oral sex, and engaging in insertive anal intercourse were reported infrequently (7%, 5%, and 1%, respectively). Only 4 men reported sex with another man and 3 reported receptive anal intercourse. Most men (93%) reported unprotected sex with 1 or more of the partners: 206 (17%) reported unprotected sex with 1 partner, 239 (19%) reported unprotected sex with 2 partners, and 792 (59%) reported unprotected sex with 3 or more partners. Furthermore, 210 men (16%) reported never using condoms with any of the partners discussed in their sexual histories. Eight hundred forty-two men (63%) had at least 1 concurrent sexual partnership (e.g., 2 overlapping partners) in their lifetime sexual history. Of these, 249 (30%) had only 1 instance of overlapping partners, 166 (19%) had 2 instances, and 416 (49%) had 3 or more instances where 2 partnerships overlapped.

Back to Top | Article Outline

Bivariate Analyses

Of the 1337 men who enrolled, 65 (5%) were HIV antibody positive (Table 1). As hypothesized, HIV infection was associated with number of lifetime partners. The median number of partners for HIV-negative men was 5.0, compared with 8.0 for HIV-positive men (Wilcoxon rank sum test = 4.6, P <0.01). Moreover, it is notable that none of the 66 men who reported just 1 lifetime sex partner tested positive for HIV.

Lack of consistent condom use, as defined by not always using a condom during penetrative sex with a partner, was also associated with increased odds of infection. In fact, all of the HIV-positive men reported unprotected sex with more than 1 partner. The median number of unprotected partners for HIV-negative men was 3, compared with 5 for HIV-positive men (Wilcoxon rank sum test = 4.5, P <0.01). HIV infection was also associated with concurrent partnerships. In general, as the number of concurrent partnerships increased, so did HIV seroprevalence. The median number of concurrent partnerships for HIV-negative men was 1 versus 2 for HIV-positive men (Wilcoxon rank sum test = 2.6, P <0.01).

Men who experienced possible exposure to blood through either iatrogenic or traditional practices were also more likely to be HIV-positive. Those who reported an injection in the last 6 months were almost 3 times more likely to be HIV-positive, those who ever practiced bloodletting (saro) were 2 times more likely to be infected, and those who reported tattoos were 2 times as likely to be HIV-positive. There was no association between HIV infection and ever donating blood or ever receiving a blood transfusion.

Finally, we also identified practices that were associated with reduced likelihood of HIV infection. Men who reported previous HIV testing and men who reported cleansing their genitals immediately after their last vaginal intercourse (i.e., postcoital washing) were also less likely to be diagnosed with HIV.

Back to Top | Article Outline

Multivariable Analyses

The final logistic model is presented in Table 2. According to the Hosmer Lemeshow goodness-of-fit test, the model adequately fit the data (P = 0.82) and regression diagnostics indicated that the discriminative capacity of the model was good (area under ROC curve = 0.87). In addition to demographic characteristics previously found to be related to HIV infection in Kenya, men who identified themselves as Catholic as compared with all others were almost twice as likely to be HIV-positive. Although Catholics were no less likely to use condoms (OR 0.9, 95% CI 0.7–1.3) or report greater number of lifetime partners (OR 1.1, 95% CI 0.9–1.4), they were more likely to report 2 or more concurrent partners (OR 1.4, 95% CI 1.1–1.7). In addition to number of sex partners and previous STI, men who ever had sex with a woman while she was menstruating were twice as likely to be HIV-positive. Men who practiced bloodletting or received an injection in the last 6 months were also more likely to be HIV-positive. Behaviors that were less likely to be reported by men who were HIV-positive were prior HIV testing and postcoital cleansing. No multiplicative effect modifiers were identified in this analysis.



Back to Top | Article Outline


In this nested case–control study of young men participating in a RCT of MC in Kisumu, Kenya, we confirmed the importance of previously established risk factors for HIV infection and identified several practices that warrant further study. As consistently noted elsewhere,6–9 we found that the number of lifetime sexual partners was of pivotal importance. Greater number of lifetime partners was also associated with more unprotected partners (OR 5.4, 95% CI 4.0–7.1) and more concurrent partnerships (OR 7.7, 95% CI 6.0–9.9). Similarly, men who used condoms consistently or were faithful to 1 partner were less likely to be infected. However, as has been found in other studies in Africa,10–12 condom use did not remain significant after adjustment for other risk factors in multivariable models. This lack of association could be due to inaccuracies in self-reported data or possibly because we measured frequency, but not correctness of condom use.

Some have suggested that concurrent partnerships may contribute to the spread of HIV.13–16 Although bivariate analyses revealed an association between concurrent partnerships and HIV, this association was not significant after adjustment was made for other variables. Duration of concurrent partnerships may contribute as much or more to the risk of HIV16 as presence or number of such partnerships, which should be explored in future studies.

Consistent with previous studies, men who reported a history of a sexually transmitted infection (STI) had increased odds of being HIV-positive.6,17–22 Sex with a woman while she was menstruating was also more common among the HIV-positive men. Although some studies early in the epidemic found evidence to suggest that sex during menstruation may increase the risk of HIV9,23,24 or other STIs,25 and this is biologically plausible, we are not aware of any recent studies that have reported an association. Also consistent with previous studies in sub-Saharan Africa, the following demographic characteristics were associated with HIV infection: older age,3,26 less education,27 being married,4,17,20 and religious affiliation.28,29

Although there is a general consensus that heterosexual transmission accounts for 90% or more of HIV infections in sub-Saharan Africa,1 some have argued that iatrogenic transmission of HIV, through blood transfusions and medical injections, has been grossly underestimated.30–38 Although transfusions were not associated with HIV infection, only 3% of men reported transfusions which did not provide sufficient power to detect an association. Receipt of injections was associated with HIV, even after controlling for treatment of prior STIs. However, because this is a cross-sectional analysis, it is also possible that HIV-positive men were more likely to report recent injections to treat HIV/AIDS related illnesses. We also found an association between HIV and the practice of bloodletting (saro), which persisted in multivariable analyses. It is possible that bloodletting may increase the risk of HIV through exposure to another person's blood if a knife or other instrument is used in succession without proper sterilization. However, this practice is not typically performed in succession, but rather on an individual basis, making it unlikely that knives serve as fomites for transmission.39 Further, it is less likely that residual blood would stay infective on a knife for as long as it would on a hallow device such as a needle.39 Some prior studies found types of scarification to be a risk for HIV,40,41 while others have not.41,42

We also documented practices associated with reduced odds of infection: prior HIV testing and postcoital cleansing. The fact that men who reported prior testing for HIV have reduced odds of infection could be because men who seek HIV testing practice less risky behaviors,43,44 or it could be attributed to the beneficial effects of risk reduction counseling, which is a standard component of voluntary counseling and testing.45,46 Similar to a recent study in Nairobi, men who reported cleansing their genitals immediately after sex had reduced odds of being infected with HIV.47 In multivariable analyses (not shown) postcoital cleansing was associated with fewer unprotected partners, not drinking alcohol, and not practicing bloodletting, and so it is possible that the association between men testing positive for HIV and postcoital washing is confounded by other factors. Also, hygiene practices reported by HIV-positive men may not reflect their behaviors at the time of HIV acquisition.47

Although the overall response rate was 65%, a differential response rate was noted according to HIV serostatus. Of the 1769 eligible HIV seronegative men, 1272 (72%) enrolled; however, of the 290 eligible HIV-positive men, only 65 men (22%) enrolled. We hypothesize that fewer HIV-positive men chose to participate in this study as they had just been informed of their serostatus and did not want to be interviewed about their sexual behavior; however, it is possible that those who participated engaged in different behavior than those who did not. We compared the demographic characteristics of HIV-positive men who enrolled in the study with those who did not enroll and found no significant differences. We did not have behavioral data available for the HIV-positive men who chose not to enroll in this study, and thus it was not possible to compare sexual practices.

Despite these limitations, this analysis has important strengths. It is noteworthy that we obtained lifetime sexual histories on 87% of participants and described sexual practices not commonly reported elsewhere. Though this study confirms previously documented risk factors for HIV (e.g., greater number of lifetime partners, older age, less education), it also identifies others that have not been consistently documented in the literature, including sex with a woman while she is menstruating, bloodletting, receipt of injections, being Catholic, and the protective effect of postcoital cleansing. Some of these practices have the potential for modification and if the results are replicated in other studies, they could be incorporated into prevention campaigns.

Back to Top | Article Outline


1. Joint United Program on AIDS (UNAIDS) WHO. Report on the Global AIDS Epidemic. Geneva: UNAIDS; 2006.
2. Cheluget B, Baltazar G, Orege P, et al. Evidence for population level declines in adult HIV prevalence in Kenya. Sex Transm Infect 2006; 82 (Suppl 1):i21–i26.
3. Central Bureau of Statistics (CBS) [Kenya] MoHMKaOM. Kenya Demographic and Health Survey 2003. Calverton, MD: CBS, MOH, ORC Macro; 2004.
4. Johnson K, Way A. Risk factors for HIV infection in a national adult population: Evidence form the 2003 Kenya demographic and health survey. J Acquir Immune Defic Syndr 2006; 42:627–636.
5. Hosmer D, Lemeshow, S. Applied Logistic Regression, 2nd ed. New York: Wiley; 2000.
6. Quigley M, Munguti K, Grosskurth H, et al. Sexual behaviour patterns and other risk factors for HIV infection in rural Tanzania: A case-control study. AIDS 1997; 11:237–248.
7. Wawer MJ, Sewankambo NK, Berkley S, et al. Incidence of HIV-1 infection in a rural region of Uganda.[see comment]. BMJ 1994; 308:171–173.
8. Quigley MA, Morgan D, Malamba SS, et al. Case-control study of risk factors for incident HIV infection in rural Uganda. J Acquir Immune Defic Syndr 2000; 23:418–425.
9. Malamba S, Wagner H-U, Maude G. Risk factors for HIV-1 infection in adults in a rural Ugandan community: A case-control study. AIDS 1994; 8:253–257.
10. Lagarde E, Auvert B, Chege J, et al. Condom use and its association with HIV/sexually transmitted diseases in four urban communities of sub-Saharan Africa. AIDS 2001; 15 (Suppl 4):S71–S78.
11. Ferry B, Carael M, Buve A, et al. Comparison of key parameters of sexual behaviour in four African urban populations with different levels of HIV infection. AIDS 2001; 15 (Suppl 4):S41–S50.
12. Kajubi P, Kamya MR, Kamya S, et al. Increasing condom use without reducing HIV risk: Results of a controlled community trial in Uganda. J Acquir Immune Defic Syndr 2005; 40:77–82.
13. Morris M, Kretzschmar M. Concurrent partnerships and transmission dynamics in networks. Soc Netw 1995; 17:299–318.
14. Morris M, Kretzschmar M. Concurrent partnerships and the spread of HIV. AIDS 1997; 11:641–648.
15. Morris M, Kretzschmar M. A microsimulation study of the effect of concurrent partnerships on the spread of HIV in Uganda. Math Popul Stud 2000; 8:109–133.
16. Halperin D, Epstein H. Concurrent sexual partnerships help to explain Africa's high HIV prevalence: Implications for prevention. Lancet 2004; 364:4–6.
17. Auvert B, Buve A, Ferry B, et al. Ecological and individual level analysis of risk factors for HIV infection in four urban populations in sub-Saharan Africa with different levels of HIV infection. AIDS 2001; 15 (Suppl 4):S15–S30.
18. Konde-Lule JK, Wawer MJ, Sewankambo NK, et al. Adolescents, sexual behaviour and HIV-1 in rural Rakai district, Uganda. AIDS 1997; 11:791–799.
19. Lagarde E, Congo Z, Meda N, et al. Epidemiology of HIV infection in urban Burkina Faso. Int J STD AIDS 2004; 15:395–402.
20. Nunn AJ, Kengeya-Kayondo JF, Malamba SS, et al. Risk factors for HIV-1 infection in adults in a rural Ugandan community: A population study. AIDS 1994; 8:81–86.
21. Todd J, Grosskurth H, Changalucha J, et al. Risk factors influencing HIV infection incidence in a rural African population: A nested case-control study. J Infect Dis 2006; 193:458–466.
22. Rakwar J, Lavreys L, Thompson ML, et al. Cofactors for the acquisition of HIV-1 among heterosexual men: Prospective cohort study of trucking company workers in Kenya. AIDS 1999; 13:607–614.
23. Miguez-Burbano MJ, Page JB, Angarita I, et al. High-risk behaviours in men from Bogota, Colombia and the spread of HIV. Int J STD AIDS 2001; 12:739–743.
24. Padian NS, Shiboski SC, Jewell NP. The effect of number of exposures on the risk of heterosexual HIV transmission. J Infect Dis 1990; 161:883–887.
25. Tanfer K, Aral SO. Sexual intercourse during menstruation and self-reported sexually transmitted disease history among women. Sex Transm Dis 1996; 23:395–401.
26. Buve A, Carael M, Hayes RJ, et al. Multicentre study on factors determining differences in rate of spread of HIV in sub-Saharan Africa: Methods and prevalence of HIV infection [see comment]. AIDS 2001; 15 (Suppl 4):S5–S14.
27. Glynn JR, Carael M, Buve A, et al. Does increased general schooling protect against HIV infection? A study in four African cities. Trop Med Int Health 2004; 9:4–14.
28. Sagay AS, Kapiga SH, Imade GE, et al. HIV infection among pregnant women in Nigeria. Int J Gynaecol Obstet 2005; 90:61–67.
29. Hawken MP, Melis RD, Ngombo DT, et al. Opportunity for prevention of HIV and sexually transmitted infections in Kenyan youth: Results of a population-based survey. J Acquir Immune Defic Syndr 2002; 31:529–535.
30. Gisselquist D, Rothenberg R, Potterat J, et al. HIV infections in sub-Saharan Africa not explained by sexual or vertical transmission [see comment]. Int J STD AIDS 2002; 13:657–666.
31. Gisselquist D, Rothenberg R, Potterat J, et al. Non-sexual transmission of HIV has been overlooked in developing countries. BMJ 2002; 324:235.
32. Gisselquist D, Potterat JJ. Heterosexual transmission of HIV in Africa: An empiric estimate [see comment]. Int J STD AIDS 2003; 14:162–173.
33. Gisselquist D, Potterat JJ, Brody S, et al. Let it be sexual: How health care transmission of AIDS in Africa was ignored [see comment]. Int J STD AIDS 2003; 14:148–161.
34. Gisselquist D, Potterat JJ, Brody S. Running on empty: Sexual co-factors are insufficient to fuel Africa's turbocharged HIV epidemic. Int J STD AIDS 2004; 15:442–452.
35. Gisselquist D, Minkin SF, Okwuosah A, et al. Unsafe injections and transmission of HIV-1 in sub-Saharan Africa [comment]. Lancet 2004; 363:1648–1649; author reply 1649–1650.
36. Gisselquist D, Potterat JJ. Review of evidence from risk factor analyses associating HIV infection in African adults with medical injections and multiple sexual partners. Int J STD AIDS 2004; 15:222–233.
37. Gisselquist D, Potterat JJ, Brody S. HIV transmission during paediatric health care in sub-Saharan Africa—Risks and evidence. S Afr Med J 2004; 94:109–116.
38. Gisselquist D, Potterat JJ. Questioning Wawer et al.'s estimated rate of sexual HIV transmission from persons with early HIV infections [comment]. J Infect Dis 2005; 192:1497–1499; author reply 1499–1500.
39. Schmutzhard E. Aids and the African traditional healer (comment). Lancet 1987; 330:459.
40. Gomo E, Chibatamoto PP, Chandiwana SK, et al. Risk factors for HIV infection in a rural cohort in Zimbabwe: A pilot study [see comment]. Cent Afr J Med 1997; 43:350–354.
41. Kanki P, M'Boup S, Marlink R, et al. Prevalence and risk determinants of human immunodeficiency virus type 2 (HIV-2) and human immunodeficiency virus type 1 (HIV-1) in west African female prostitutes. Am J Epidemiol 1992; 136:895–907.
42. N′galy B, Ryder RW, Bila K, et al. Human immunodeficiency virus infection among employees in an African hospital. N Engl J Med 1988; 319:1123–1127.
43. Balmer D, Grinstead OA, Kihuho F, et al. Characteristics of individuals and couples seeking HIV-1 prevention services in Nairobi, Kenya: The voluntary HIV-1 counseling and testing efficacy study. AIDS Behav 2000; 4:15–23.
44. Nyblade LC, Menken J, Wawer MJ, et al. Population-based HIV testing and counseling in rural Uganda: Participation and risk characteristics. J Acquir Immune Defic Syndr 2001; 28:463–470.
45. Group TVH-CaTES. Efficacy of voluntary HIV-1 counseling and testing in individuals and couples in Kenya, Tanzania, and Trinidad: A randomized trial. Lancet 2000; 356:103–112.
46. Weinhardt LS, Carey MP, Johnson BT, et al. Effects of HIV counseling and testing on sexual risk behavior: A meta-analytic review of published research, 1985–1997 [see comment]. Am J Public Health 1999; 89:1397–1405.
47. Meier A, Bukusi E, Cohen C, et al. Independent association of hygiene, socioeconomic status, and circumcision with reduced risk of HIV infection among Kenyan men. J Acquir Immune Defic Syndr 2006; 43:117–118.
© Copyright 2007 American Sexually Transmitted Diseases Association