O'Farrell, Nigel MD, FRCP*†; Morison, Linda MSc†; Moodley, Prashini MB, ChB, PhD‡; Pillay, Keshree MB, ChB‡; Vanmali, Trusha MB, ChB‡; Quigley, Maria MSc†; Hayes, Richard MSc, DSc†; Sturm, A. Willem MD, PhD‡
HIV prevalence varies considerably throughout Africa. In some countries, HIV prevalences in antenatal clinic attenders appear to have peaked at around 5%, whereas levels of more than 30% are reported in others.1 In general, countries in east and southern Africa are worse affected than those in west Africa. A number of reasons have been put forward for this, including time of introduction of HIV to different populations, differences in sexual behavior, male circumcision, and prevalence of genital ulcer disease.2-5 More than 25 years into the epidemic, it seems likely that the incidence of HIV has now peaked in most African countries making the first of these reasons unlikely. Also, it appears that sexual behavior, in the context of numbers of sexual partners, cannot alone explain this variation in HIV prevalence.6
The role of male circumcision and the related issues of genital ulcer disease and genital hygiene still require further clarification. In uncircumcised men, Langerhans cells that are concentrated in the prepuce could be a target for HIV, whereas circumcised men have increased keratinization of the exposed glans that may provide a barrier to HIV entry.7 Uncircumcised men are more likely than those circumcised to develop syphilis and chancroid, both risk factors for HIV infection in diverse populations that may be related to poor penile hygiene.8,9 Delayed healing and secondary infection of genital herpes, now the commonest cause of genital ulceration in Africa, may also be related to poor genital hygiene.10 Aside from any direct effect of male circumcision on HIV acquisition, biological plausibility suggests that HIV transmission is enhanced in uncircumcised men, not only through sexually acquired genital ulcers acting as a portal of entry and exit for the virus, but also through microulcerations after trauma during sex and secondary to poor penile hygiene.11
Despite convincing epidemiological and biological evidence of an association between male circumcision and HIV infection, many researchers and policy makers have been reluctant to consider male circumcision as a measure to control HIV infection. This may be because both the risk of HIV attributable to lack of circumcision is heterogeneous and appears higher in studies of sexually transmitted infection (STI) clinic attenders compared with whole communities12,13 and a fear of increased risk behavior among newly circumcised men. However, a recent intervention trial in South Africa has shown a significant reduction in the incidence of HIV in young men in the first 18 months after circumcision performed by trained skilled operators under sterile conditions.14 If circumcision is to be promoted as a public health intervention, safeguards must be in place to ensure the procedure is undertaken with minimal risk by trained operators under sterile conditions as deaths are not an uncommon complication of the procedure when undertaken by unskilled individuals.15,16 In addition, cultural sensitivities should be addressed so that circumcision is accepted in traditionally noncircumcising communities. By contrast, male genital hygiene might be amenable to simple interventions targeted at the whole male population without the surgical trauma of circumcision. However, although circumcision would be a one-off intervention, long-term efforts may be needed to bring about and sustain any improvements in genital hygiene behavior.
Despite evidence and beliefs that male hygiene is related to risk of STIs and penile cancer,9,17-19 little work has been done in this field since interventions to promote genital hygiene as a prophylactic measure against STIs were undertaken routinely during the First World War.9 Recent studies have assessed genital hygiene through behavioral measures rather than clinically.20,21 However, Prakash et al22 found evidence of a subpreputial collection in 40% of 1000 uncircumcised outpatient attenders in India with evidence of a thin layer on careful inspection in 85% of these. He concluded that if this material was allowed to collect, inflammation and chronic irritation would supervene, and these changes might themselves result in secondary phimosis and further difficulties in achieving good genital hygiene. Moreover, he also believed the subpreputial space to be wet in most individuals.23 By contrast, any degree of penile wetness was observed in only 9.5% of uncircumcised men attending a London STI clinic at their initial presentation.24 The main aim of this study was therefore to determine whether subpreputial wetness assessed clinically was associated with HIV in a population of male STI clinic attenders in Durban, South Africa.
The primary objectives of the study were the following:
I. To conduct a cross-sectional study to describe the prevalence and characteristics of poor penile hygiene as assessed by the presence of subpreputial penile wetness among men attending a STI clinic. Penile wetness was defined as the observation of a diffuse homogenous film of moisture on the surface of the glans and coronal sulcus.
II. To investigate the association between current levels of penile wetness and HIV infection in this population.
This was a cross-sectional observational study among men free of STI-related genital ulcers and urethral discharge. Participants enrolled at the STI clinic were asked to return 2 weeks later. At this time, assessment was made of their penile wetness status after initial treatment of their STI-related symptoms at enrollment.
Consecutive men attending the Prince Cyril Zulu STI clinic in Durban for a new STI-related problem were enrolled. Inclusion criteria for subjects were being sexually active, heterosexual, aged 16 to 75 years, and with the potential ability to return in 2 weeks. Signed consent was obtained from all subjects, and all were informed that the study was to investigate risk factors for HIV and would involve a sociodemographic and sexual behavior questionnaire. Inclusion in the study made no difference to STI management for the current problem. Cases were dealt with according to National STI syndromic guidelines. Ethical approval for the study was obtained from the ethics committees of the Nelson Mandela Medical School, University of KwaZulu- Natal, Ethekwini Municipality Health Department Research Committee, Durban and the London School of Hygiene & Tropical Medicine.
Because both penile wetness and degrees of penile mucosal discontinuity could be influenced by a current STI complaint, consenting men were asked to return to the clinic 2 weeks later after completing STI treatment. Previous studies involving follow-up at the clinic achieved a 70% to 80% return rate.
At enrollment, sociodemographic data were obtained including age, marital status, ethnic group, religion, and income, and a sexual behavioral questionnaire was administered by a trained interviewer in a face-to-face interview. Clinical examination of the genital area was performed by a physician, and information recorded on circumcision status and ease of foreskin retraction subdivided into the following categories, easy full retraction, slight difficulty with retraction, difficult but full retraction, retraction so that half the glans was visible, and nonretractile. The degree of penile mucosal discontinuity, determined by any breaks of the penile skin on clinical examination, and presence of bleeding from the penis during examination and foreskin retraction were recorded. Participants were offered voluntary counseling and testing for HIV at the clinic site.
On returning to the clinic on day 14, participants were interviewed using a structured questionnaire by the same interviewer as that at enrollment about, mobility in the last 30 days, access to water in their current main abode, degree of privacy for washing the genitals, and male genital cleansing practices and beliefs. Genital examination and penile wetness assessment were undertaken. The degree of wetness of the glans observed after foreskin retraction was classified into 4 grades: dry (1), slight wetness (2), wet (3), or very wet with or without smegma (4). Observers were unaware of the HIV status of participants.
Interobserver Variability of Assessment of Penile Wetness
Clinical examination and data collection for the main cross-sectional study were conducted by 3 physicians. In a random sample of 62 (sample size chosen for logistical convenience) participants, a further clinical assessment on day 14 was made by another observer (NO'F) to determine the degree of interobserver variation in the clinical observations made.
HIV was diagnosed using Determine HIV 1/2 (Abbott Laboratories, Chicago, IL). Positive results were confirmed with a second serological test (Capillus HIV1/2, Trinity Biotech, Berkeley Heights, NJ). Serological testing for HSV-2 antibodies was by Focus HSV-2 (Focus Diagnostics, Cypress, CA) and for syphilis with the rapid plasma reagin test (Becton Dickinson, Cockeysville, MD), and Treponema pallidum hemagglutination assay (Omega Diagnostics, Alva, UK). Nucleic acid amplification tests (NAATs) tests for gonorrhea, chlamydia, and trichomoniasis were undertaken on urethral swabs from subjects with urethral discharge/dysuria as described previously.25 Data from the questionnaires were anonymized and linked to the results of serological testing.
Questionnaire data were double-entered and validated using EpiData v3.0 (CDC).26 Laboratory data were entered into an Excel spreadsheet (Microsoft, Reading, UK). Statistical analysis was conducted using Stata v8.0 (Statacorp, College Station, TX). The primary analysis was restricted to uncircumcised men who had no symptoms or laboratory evidence of either genital ulceration or urethral discharge on their return visit at day 14. We compared sociodemographic factors, sexual behavior, HIV prevalence, HSV-2 antibodies, syphilis serology, other STIs by syndrome, and STI infections diagnosed by NAATs in those men included in the primary analysis with those enrolled but not included.
The main exposure variable of interest was subpreputial wetness assessed by clinical examination. This was categorized into a binary variable of dry versus any wetness (slightly wet, wet, or very wet). First, we examined associations between sociodemographic, sexual behavior and hygiene behavior characteristics, and penile wetness. Cross-tabulations and logistic regression models were used to estimate crude prevalence odds ratios (OR). Then those variables associated with penile wetness in the crude analysis were included in a multiple logistic regression model. Backward elimination was used to exclude any variables no longer significantly associated after adjustment for the other variables. Each of all the other variables was then added back into the model in turn to describe its association with penile wetness adjusted for significant predictors.
A similar procedure was followed to obtain a multiple logistic regression model for the association between sociodemographic characteristics, sexual behavior, hygiene behavior, penile wetness, and HIV infection. However, because the primary interest was the association between penile wetness and HIV infection, each variable that was not already included in the model as a significant predictor was added in turn to examine whether it confounded the primary association (defined as altering the OR for the primary association by 10% or more). Those variables found to be confounders were retained in the model. We also decided a priori to conduct the adjusted analysis with and without HSV-2 antibody status in the model. HSV-2 antibodies could be considered as an indicator of past risky sexual behavior, that is, a potential confounder, but might also be on the causal path between the exposure and outcome, hence the need for caution in including it. We also examined gonorrhea, chlamydia, or trichomoniasis infections diagnosed by NAATs testing at recruitment as possible confounders.
Logistic regression models were used (giving OR rather than prevalence risk ratios [RR]), although the outcome was not rare because of convergence problems with the log-link model. The prevalence of the outcome (HIV infection) was expected to be high which would result in an OR that overestimated the RR. Crude and adjusted OR for the association between penile wetness and HIV were therefore converted by hand to RR.27
The primary analysis was to compare HIV prevalence between men with and without penile wetness. The overall HIV prevalence was thought likely to be in the range 40% to 60%, and we assumed that around half the men would have penile wetness. We aimed to obtain data on 370 men for the primary analysis. This would give the study 80% power to detect a RR of 1.5 if the HIV prevalence in the group without penile wetness was 30% or a RR of 1.3 if the HIV prevalence in the group without penile wetness was 50%.
Six hundred and fifty men (589 uncircumcised, 55 circumcised, 6 circumcision status unknown) were enrolled into the study (Fig. 1) of whom 488 (75%) returned after 2 weeks. Of these, 386 satisfied the conditions for inclusion in the primary analysis, that is, uncircumcised and without STI symptoms or signs (genital ulceration or urethral discharge) that might directly cause penile wetness (Fig. 1). There was little difference between those included in the sample for the primary analysis and uncircumcised men enrolled but not included, although gonorrhea was more prevalent (P = 0.016) in the former group and genital warts (P = 0.039) in the latter (Table 1). The remainder of the results refers to analysis of the 386 men eligible for the primary analysis.
The age range of the men included in the sample for the primary analysis was 18 to 59 years, and most of the men (89.5%) were Black African Zulus (Table 2). Most of them (58.5%) received a low income of less than 100 rand per week (US$16). Only 7% of men were married, but 27% reported living with a partner. More than 80% of men reported more than 6 partners during their lifetime and around 50% reported 2 or more partners in the last 3 months. Only 5.2% reported sex with a sex worker ever. Around 80% of men had access to water inside their residence and privacy to wash. More than half reported genital washing more frequently than once a day and around 30% of men reported washing directly after sexual intercourse.
Assessment of Penile Wetness and Genital Hygiene Behavior
Half the men (190/386, 49.2%) were assessed by a clinician as having some wetness around the glans and coronal sulcus. Of those assessed as having penile wetness, 79.5% (151) were assessed as slightly wet, 19.0% (36) as wet, and 1.6% (3) as very wet. The crude analysis showed that penile wetness was significantly more common among those with a low level of attained education, low income, no access to water inside the residence, washing once a day or less, not washing after sexual intercourse, and with a clinical assessment of problems retracting the foreskin (Table 2). The adjusted analysis showed significant associations between penile wetness and younger age, low level of attained education, low income, higher lifetime numbers of sexual partners, and not washing after sex.
The prevalence of penile wetness in circumcised men was low. Of the 36 men circumcised men that returned at day 14 and were clear of urethral discharge and genital ulcers, only 1 (2.8%) was assessed as having penile wetness.
Interobserver Variation in Assessment of Penile Wetness
In 62 cases examined by 2 doctors, both agreed in 54 cases (87%) on whether wetness was present or absent.
HIV and Penile Wetness Assessment
The overall prevalence of HIV among the men was 56.0% (216/386). Table 3 shows that HIV prevalence was significantly higher among those of older age, low level of attained education, not being married, not living with a partner, and reporting ever having sex with a sex worker.
Table 3 shows that in men assessed as having any level of penile wetness, the HIV prevalence was 66.3% (90/196) compared with 45.9% (90/196) in those with no wetness resulting in an OR of 2.32. Table 3 also shows that adjusting for other predictors of HIV and confounders reduced the OR slightly to 2.27, with adjustment for HSV-2 antibodies making little difference (OR, 2.38). Converting the OR to RR gave a crude RR of 1.44 (95% confidence interval [CI], 1.23-1.63), and after adjustment for all predictors, confounders and HSV-2 gave a RR of 1.46 (95% CI, 1.19-1.68). None of the variables indicating STI diagnosis at enrollment confounded the relationship between penile wetness and HIV. The final OR for HIV and penile wetness (OR, 2.38; Table 3) was altered to 2.30 (95% CI, 1.19-4.44) when adjusted for gonorrhea, chlamydia, or trichomoniasis, 2.36 (95% CI, 1.41-3.95) when adjusted for urethral discharge/dysuria, 2.30 (95% CI, 1.36-3.87) when adjusted for genital ulcer syndrome, and 2.32 (95% CI, 1.38-3.88) when adjusted for positive Treponema pallidum hemagglutination assay and rapid plasma reagin.
Table 4 shows the HIV prevalence in each category of penile wetness. HIV prevalence was 45.9% (90/196) in men who were assessed as dry and between 66.2% and 66.7% in the 3 categories of wetness showing no dose-response pattern.
Among circumcised men for whom an HIV result was available, the HIV prevalence was 42.9% (15/35), similar to that in uncircumcised men without penile wetness (45.9%).
This study is the first to examine the association between subpreputial penile wetness assessed by direct clinical examination and HIV prevalence. We found a 44% increase in the HIV prevalence of uncircumcised men assessed as having subpreputial wetness compared with those found to be dry. Adjusting for confounding made little difference to this result. Our study included a small number of circumcised men, and the prevalence of HIV among these was similar to that among uncircumcised men with a dry penis.
We found a good level of agreement between physicians in assessing the degree of wetness supporting our classification of the clinical findings of degrees of wetness as identified in a study of men in India of low socioeconomic status.22 Interestingly, although half the men were clinically assessed as having penile wetness in our study, only 12 (3.1%) perceived their genital hygiene as not good.
In defining poor genital hygiene clinically, no current accepted gold standard exists. However, penile wetness, assessed in the same way as our study, was observed in only 9.5% of 372 uncircumcised men attending a London STI clinic at their initial presentation, a population that included patients with urethral discharge and diabetes-associated balanitis. We believe the wetness of the subpreputial glans reflects poor penile hygiene, a clinical observation reported previously in India.22 This wetness is not the same as overt smegma, which is a collection of necrotic debris under the foreskin also associated with poor hygiene, and we identified only a few cases with smegma using this definition. Although urine spreading out in the subpreputial space in uncircumcised men that do not retract the foreskin when urinating is not thought to be related to this wetness, prostatic, vesicular, and urethral secretions may play a role.23 However, if this were the case, it would be difficult to explain the marked difference in penile wetness assessed clinically between the multiethnic London STI clinic population24 and those in India22,23 and our study. One study has shown that lysozyme, a substance secreted by the apocrine glands of the inner prepuce may kill HIV in vitro, but immunological function of the prepuce is not well developed, and there are no studies investigating in vivo evidence of this phenomenon.28
There are some potential effects of selection and participation bias in our study of STI clinic attenders. However, targeting this group is critical in limiting the spread of HIV in worst affected countries.29 Our follow-up rate of 75% was what we expected based on previous studies at this urban clinic, and although our study sample is not completely representative of all STI patients, we are making internal comparisons with the subgroup that returned, and it is unlikely these associations will have been biased. Also, in determining the sample for the primary analysis, we were careful to exclude those with urethral discharge and genital ulceration that could affect the assessment of penile wetness.
Cross-sectional studies have well-known limitations in their ability to make inferences about causality. Although we adjusted for the potential confounders we collected data on, there is always the possibility of residual confounding. The sexual and hygiene behavior questions we asked are particularly susceptible to social desirability bias. Based on the degree of penile wetness observed on clinical examination, we believe that men in Durban overestimated the frequency with which they washed the penis, giving the answer that they thought acceptable to the interviewer. Establishing the validity of reported male genital hygiene behavior, while recognizing that poor hygiene is a socially undesirable trait, is one reason why few studies have been done in this field. Our study may have underestimated the effect of penile wetness as a risk factor for HIV in several ways. Patients coming to a STI clinic would expect to be examined and may have cleaned themselves before their attendance. Furthermore, although we did not mention hygiene on day 1 at enrollment, patients coming back on day 14 may have discussed genital hygiene with other subjects about to be enrolled that led them to alter their genital hygiene behavior. The differences in rates of gonorrhea and warts at enrollment between those included in the primary analysis and those enrolled but not included are unlikely to have biased our results. In fact, the difference for gonorrhea is likely to be due in part to excluding men with gonorrhea at the follow-up visit from the sample for primary analysis. Penile warts would not have been expected to be associated with penile wetness and were therefore not excluded. STI diagnoses at enrollment were examined as potential confounders of the association between penile wetness (after treatment) and HIV infection but none was found to confound it.
The mechanism of the action whereby subpreputial wetness might predispose to HIV is not clear and must therefore be subject to speculation. We propose 4 possible mechanisms. Firstly, penile wetness might create an environment whereby any trauma to the penile mucosa during sexual intercourse either directly or through a sexually acquired genital ulcer would be less likely to heal quickly either naturally or because of secondary infection. Secondly, penile wetness may lead to subpreputial balanitis that might result in microulcerations that could provide a portal of entry for HIV. Thirdly, a wet penis might enhance adherence of infective HIV virions to the surface of target cells for longer than a dry penis. Fourthly, penile wetness could elicit an immune response recruiting more HIV target cells to the surface of the inner foreskin, predisposing the individual to infection. Our study found no trend as the penis gets wetter toward greater HIV prevalence. It might be that just a small degree of wetness was all that was required to enhance HIV susceptibility by the mechanisms discussed above.
Other studies in Africa have shown that among Luo men in Kenya, HIV prevalence was higher in those with less than adequate hygiene, defined as not reporting retracting the foreskin while washing and the presence of smegma on clinical examination. In that study, 19% were defined as having poor hygiene compared with half with penile wetness in our study.20 A community study in South Africa found that the few men (13%) that washed their penis with soap less than once a day had a higher prevalence of HIV than those that washed more often, although reported washing beneath the foreskin was not recorded.21
If penile wetness were confirmed to be an important risk factor for acquiring HIV, methods of facilitating a dry penile environment would have to be identified. Penile wetness was significantly associated with a low level of attained education, low income, no access to water inside the residence, washing once a day or less, not washing after sexual intercourse, and a clinical assessment of difficulty in retracting the foreskin. The first 3 of these findings are poverty-related and support the need for increased access to water and education as well as more conventional HIV prevention efforts. However, although access to water inside a residence when none exists will be important, water supply alone is not sufficient.30 Information, education, and communication programs at a number of levels would be needed, for instance, encouraging washing related to sexual activity-precoital or postcoital or as an everyday life skill. In Kenya, postcoital cleansing may be a more socially acceptable practice than washing before sex but may be difficult without access to water.31 In another Kenyan study involving focus group discussions centered on the acceptability of male circumcision, improved male hygiene was one of the main benefits perceived by Luo women.17 If good penile hygiene were accepted among communities as a social norm, men's behavior might be encouraged to change if good penile hygiene made them more acceptable as sexual partners. Further observational studies and randomized controlled trials are needed to examine both the possible influence of penile wetness on HIV transmission and methods to improve it.
The authors acknowledge the help of Mark Colvin and Salim Abdool Karim in logistics, Simi Chanderman and Adrian Smith for data entry, Prathna Bhola for collecting data and examining patients, Gladys Nzimande for enrolling patients, and all other staff from the Department of Medical Microbiology and the Prince Cyril Zulu STI clinic involved in facilitating the study. The study received funding from the UK Department for International Development.
2. Bongaarts J, Reining P, Way P, et al. The relationship between male circumcision and HIV infection in African populations. AIDS. 1989;3:373-377.
3. Moses S, Bradley JE, Nagelkerke NJ, et al. Geographical patterns of male circumcision practices in Africa: association with HIV prevalence. Int J Epidemiol. 1990;19:693-697.
4. Buvé A, Carael M, Hayes R, et al. Variations in HIV prevalence between urban areas in sub-Saharan Africa: do we understand them? AIDS. 1995;9(suppl A):S103-S109.
5. O'Farrell N. Targeted interventions required against genital ulcers in African countries worst affected by HIV infection. Bull World Health Organ. 2001;79:569-577.
6. Buvé A, Lagarde E, Carael M, et al. Interpreting sexual behaviour data: validity issues in the multicentre study on factors determining the differential spread of HIV in four African cities. AIDS. 2001;15(suppl 4):S117-S126.
7. Szabo R, Short RV. How does male circumcision protect against HIV infection. BMJ. 2000;320:1592-1594.
8. Weiss HA, Thomas SL, Munabi SK, et al. Male circumcision and risk of syphilis, chancroid and genital herpes: a systematic review and meta-analysis. Sex Transm Infect. 2006;82:101-109.
9. Moore JE. The diagnosis of chancroid and the effect of prophylaxis upon its incidence in the American Expeditionary Forces. J Urol. 1920;4:169-176.
10. O'Farrell N, Hoosen AA, Coetzee KD, et al. Genital ulcer disease: accuracy of clinical diagnosis and strategies to improve control in Durban, South Africa. Genitourin Med. 1994;70:7-11.
11. O'Farrell N, Quigley M, Fox P. Association between the intact foreskin and inferior standards of male genital hygiene behaviour: a cross sectional study. Int J STD AIDS. 2005;16:556-559.
12. O'Farrell N, Egger M. Circumcision in men and the prevention of HIV infection: a meta-analysis revisited. Int J STD AIDS. 2000;11:137-142.
13. Weiss HA, Quigley MA, Hayes RA. Male circumcision and risk of HIV infection in sub-Saharan Africa: a systematic review and meta-analysis. AIDS. 2000;14:2361-2370.
14. Auvert BA, Puren A, Taljaard D, et al. Randomised controlled intervention trial of male circumcision for reduction of HIV infection risk: the ANRS 1265 trial. PLoS Med. 2005;2:1112-1122.
15. Crowley IP, Kesner KM. Ritual circumcision (Umkhwetha) amongst the Xhosa of the Ciskei. Br J Urol. 1990;66:318-321.
16. Magoha GA. Circumcision in various Nigerian and Kenyan hospitals. East Afr Med J. 1999;76:583-586.
17. Bailey RC, Muga R, Poulussen R, et al. The acceptability of male circumcision to reduce HIV infection in Nyanza Province, Kenya. AIDS Care. 2002;14:27-40.
18. O'Farrell N. Soap and water prophylaxis for limiting genital ulcer disease and HIV-1 infection in men in sub-Saharan Africa. Genitourin Med. 1993;69:297-300.
19. Misra S, Chaturvedi A, Misra NC. Penile carcinoma: a challenge for the developing world. Lancet Oncol. 2004;5:240-247.
20. Agot KE, Ndinya-Achola JO, Kreiss JK, et al. Risk of HIV-1 in rural Kenya. Epidemiology. 2004;15:157-163.
21. Auvert B, Ballard R, Campbell C, et al. HIV infection among youth in a South African mining town is associated with herpes-simplex virus-2 seropositivity and sexual behaviour. AIDS. 2001;15:885-898.
22. Prakash S, Jeyakumar S, Subramanyam K, et al. Human subpreputial collection: its nature and formation. J Urol. 1973;110:211-212.
23. Prakash S, Rao R, Venkatesan K, et al. Subpreputial wetness-its nature. Ann Natl Med Sci (India). 1982;18:109-112.
24. O'Farrell N, Chung C-K, Morison L. High standard of penile hygiene in male STI clinic attenders in London. Paper presented at: International Union against Sexually Transmitted Infections Meeting; November 15-18, 2005; Bangkok, Thailand.
25. Sturm PD, Moodley P, Khan N, et al. Aetiology of male urethritis in patients recruited from a population with a high HIV prevalence. Int J Antimicrob Agents. 2004;24(suppl 1):S8-S14.
26. Lauritsen JM, Bruus M. EpiData (version 3). A comprehensive tool for validated entry and documentation of data. Odense, Denmark: The EpiData Association; 2003.
27. Zhang J, Yu KF. What's the relative risk? A method for correcting the odds ratio in cohort studies of common outcomes. JAMA. 1998;280:1690-1691.
28. Lee-Huang S, Huang PL, Sun Y, et al. Lysozyme and RNases as anti-HIV components in B-core preparations of human chorionic gonadotopin. Proc Natl Acad Sci USA. 1999;96:2678-2881.
29. O'Farrell N. Enhanced efficiency of female to male HIV transmission in core groups in developing countries: the need to target men. Sex Transm Dis. 2001;28:84-91.
30. Fewtrell L, Kaufmann RB, Kay D, et al. Water, sanitation, and hygiene interventions to reduce diarrhea in less developed countries: a systematic review and meta-analysis. Lancet Infect Dis. 2005;5:42-52.
31. Steele MS, Bukusi E, Cohen CR, et al. Male genital hygiene beliefs and practices in Nairobi, Kenya. Sex Transm Infect. 2004;80:471-476.
© 2006 Lippincott Williams & Wilkins, Inc.