Male circumcision and common sexually transmissible diseases in a developed nation setting

Add Correspondence to: A/Prof Basil Donovan, Sydney Sexual Health Centre, Sydney Hospital, GPO Box 1614, Sydney, NSW, 2001, Australia.

Accepted for publication 4 May 1994

Introduction

It has been suggested that uncircumcised men are at higher risk of certain sexually transmissible diseases (STDs), in particular genital herpes,^1,2 gonorrhoea,¹ syphilis,¹ human immunodeficiency virus type 1 (HIV-1) infection,³ candidiasis^1,2 and chancroid.³ The evidence for male circumcision protecting against most STDs is tentative⁴ but is strong for chancroid and candidal balanoposthitis (though not necessarily subclinical yeast infection⁵). In some populations where chancroid and syphilis may be the common causes of genital ulcers, these ulcers and the uncircumcised state of the men (both independently and synergistically) appear to be major risk factors for the female to male sexual transmission of HIV-1.³ The uncircumcised penis is hypothetically at increased risk of STDs because of larger surface area, thinner epidermal barrier, more opportunity for epithelial microtrauma and the warm, moist niche under the foreskin favouring the persistence of fastidious microorganisms. However, note of these hypotheses has been proven.

We sought to determine any effect the presence of a male foreskin may exert on the acquisition of common STDs by heterosexual men attending an STD clinic and also its effect on the clinical manifestations of herpes simples virus type 2 (HSV-2) infection.

Subjects and methods

A sample site of over 270 was calculated to determine a two-fold risk of serological evidence of HSV-2 for uncircumcised men with 80% power at the 5% level. However, this calculation was based on an expected HSV-2 seroprevalence of about 30% for heterosexual men detected in a 1985 survey at the same clinic (AL Cunningham; unpublished data).

The subjects were 300 consecutive heterosexual male patients who saw a particular clinician (IB) and required venopuncture for any other purpose at the Sydney Sexual Health (previously STD) Centre, Sydney Hospital, between December 1990 and May 1991. Over 95% of the men were Caucasian with the bulk of the remainder of Asian origin. The subjects were representative of the general clinical load of exclusively heterosexual men at the Centre: their reasons for attendance or diagnosis at the time of consultation are outlined in Table 1. HIV-1 infection is rare among heterosexual men attending the Centre⁶ and, as HIV-l infection may alter the natural history of other STDs, it was an exclusion criterion. Homosexually active men were also excluded because anal infections could have confused the objectives of the study. The aims of the study were explained to all subjects and signed informed consent obtained. There were no refusals.

A data collection form with an anonymous identifier code was completed by the clinician and an extra five milliliters of blood was collected for this study. The data collected included age, level of education, circumcision status (confirmed by physical examination), past history of genital herpes or symptoms suggestive of genital herpes (undiagnosed genital ulcer or recurrent genital lesions/itches), nongonococcal urethritis (NGU), genital warts, urethral gonorrhoea, syphilis and acute hepatitis B, as well as reported lifetime number of sexual partners (women with whom they had had vaginal intercourse). For the analysis, current and past diagnoses of STDs were combined for each subject. All STDs had been diagnosed by a physician, mostly at the same clinic. The diagnosis of genital herpes was by cell culture except for the current visit where the diagnosis was on clinical criteria (the delinked anonymous study design precluded incorporating current HSV culture results). Similarly urethral gonorrhoea was diagnosed by culture except for the current visit where the diagnosis was based on clinical findings plus characteristic gram negative intracellular diplococci on a urethral smear. NGU was diagnosed by clinical picture (discharge/dysuria) and microscopy of a distal urethral smear (5 or more polymorphs per oil immersion field). Genital warts were a clinical diagnosis. All syphilis cases required confirmation by specific treponemal serology (Treponema pallidum haemaglutination assay and fluorescent treponemal antibody test). The diagnosis of acute hepatitis B required both an acute hepatitis illness andlaboratory confirmation.

Those subjects who had a history of genital herpes were asked to grade the severity of their initial episode (mild, moderate, or severe), and estimate the number, and similarly grade the severity of any recurrences in the previous 12 months.

The serological analyses for HSV were per-formed at the Virology, Department Centre for Infectious Diseases and Microbiology, Westmead Hospital in Sydney on an anonymous and blinded basis. As described elsewhere,⁷ sera were screened for total HSV antibody by a complement fixing antibody (HSV-CFA) test. Positive specimens on HSV-CFA were then subject to an indirect IgG enzyme immune assay (EIA) specific to the 92 kDa HSV-2 glycoprotein G (gG-2).⁷ All 15 randomly selected sera that were positive for HSV-2 by EIA testing were confirmed by Western blot.

Results

One hundred and eighty five (61.7%) of the men were circumcised and 115 (38.3%) were not: this was not age dependent (table 2). The associations between male circumcision status and demographic, behavioural and STD variables of the study group are summarized in Table 2. None of these variables approached significance at the 5% level. Though gonorrhoea, syphilis and acute hepatitis B were reported too infrequently to determine any association with the subjects circumcision status, histories of genital herpes, serological evidence of HSV-2 infection, genital warts and NGU were common and no distinct trends emerged.

Circumcised and uncircumcised men with previously diagnosed genital herpes reported a similar duration since the initial episode and similar frequency and severity of recurrences in the preceding 12 months (Table 3). While there was a trend for more uncircumcised men to describe their initial genital herpes episode as severe and for circumcised men to describe it as moderate the difference was not significant.

Discussion

In this clinic-based prospectively collected survey we found no association between male circumcision status and STDs that are common in our population. Perhaps importantly, our study group was relatively racially homogeneous, lack of circumcision was not a marker of lower socioeconomic status (using the index of education level; Table 2), and we controlled for a major parameter of sexual behaviour (lifetime number of sexual partners). We did not investigate the relationship between the presence of a prepuce and candidal balanoposthitis as we regard the causal relationship as proven. lack of circumcision is probably also a risk factor for chancroid but this condition was rare in our population. Similarly syphilis, acute hepatitis B and gonorrhoea were so uncommon among heterosexual men attending our clinic that we were unable to exclude any potential association. HIV-1 infection was an exclusion criterion for the study. Symptomatic genital herpes and serological evidence of HSV-2 infection were both common in our population. We previously reported on the risk factors for HSV-2 seropositivity in this study group.⁸ On multi-variate analysis the risk factors for HSV-2 infection were lifetime number of sexual partners, lower level of education and (marginally) a history of sexual contact with a women with genital herpes.⁸ In the present analysis, circumcision status did not correlate with a history of genital herpes, recurrent genital ulcers or itches (which might have represented undiagnosed herpes), or serological evidence of HSV-2 infection. As an unexpectedly high proportion (two thirds) of our study group were HSV-2 seropositive it would not have been possible to demonstrate a two-fold protective effect for circumcision. Nevertheless no trend toward protection was apparent (Table 2).

A previous Australian study which relied on the culture diagnosis of genital herpes did find a correlation between symptomatic genital herpes and lack of circumcision.¹ However this study did not quantify sexual partner numbers or other parameters of risk. Another possible explanation for this finding in the previous study might have been that men with intact foreskins harbouring HSV-2 may be more symptomatic and thus be more likely to present to a clinic. However we found that men who were HSV-2 seropositive were no more symptomatic than circumcised men, nor were their lesion recurrences significantly more frequent or severe. As only a limited number of our study subjects were symptomatic, a larger study would be needed to exclude the possibility that uncircumcised men suffer more severe initial episodes of genital herpes.

Previous studies of the effect of circumcision status on genital warts have produced conflicting results. In more heterogeneous populations lack of circumcision has variously been reported as making symptomatic genital warts both more^1,9,10 and less¹¹ common. We found it exerted no effect. One retrospective study found that circumcision status affected the distribution of warts on the penis. Uncircumcised men were more likely to present with distal lesions and circumcised men with proximal lesions on the penis.¹² We did not investigate this issue. We are unaware of any study to date that has attempted to correlate circumcision status with subclinicalHPV infection.