AMERICAN JOURNAL OF PUBLIC HEALTH, Volume 84 Number 2: Pages 197-201,
February 1994.

Circumcision and Sexually Transmitted Diseases

Linda S. Cook, M.S.; Laura A. Koutsky, Ph.D.; and King K. Holmes, MD, Ph.D.

The authors are with the Department of Epidemiology School of Public Health, University of Washington, Seattle. Laura A. Koutsky and King K. Holmes are also with the center for AIDS and STDs, School of Medicine, University of Washington.

Requests for reprints should be sent to:
Laura A. Koutsky. Ph.D.,
Center for AIDS and STDs,
1001 Broadway,
Seattle, WA 98122.

Objectives. New evidence linking lack of circumcision with sexually transmitted human immunodeficiency virus revives concerns about circumcision and other sexually transmitted diseases. This study was undertaken to assess the relationship between circumcision and syphilis, gonorrhea, chlamydial infection, genital herpes, nongonococcal urethritis, and exophytic genital warts.

Methods. A cross-sectional study of 2776 heterosexual men attending a sexually transmitted disease clinic in 1988 was used to investigate the relationship between circumcision and sexually transmitted diseases. Subjects with sexually transmitted diseases and those without such diseases were compared after adjustment for age, race, zip code of residence, other sexually transmitted diseases, and number of sexual partners.

Results. A positive relationship was observed between uncircumcised status and both syphilis and gonorrhea. A negative relationship was found between warts and lack of circumcision. No apparent relationship was noted between uncircumcised status and genital herpes, chlamydial infection, or nongonococcal urethritis.

Conclusions. Uncircumcised men were more likely than circumcised men to have syphilis and gonorrhea and were less likely to have visible warts.

(Am J Public Health. 1994;84: 197-201)


The possible role of male circumcision in the acquisition of sexually transmitted diseases has taken on new significance with recent evidence that human immunodeficiency virus (HIV) infection in men is associated with uncircumcised status.1-3 Sexually transmitted diseases previously linked with uncircumcised status include chancroid, syphilis, gonorrhea, and genital herpes infection1,4-7; however, most of the reported associations between these diseases and uncircumcised status have not been adequately adjusted for potentially confounding factors. The present cross-sectional study was undertaken to examine the relationship between circumcision status and genital infection by Treponema pallidum, Neisseria gonorhoeae, and Chlamydia trachomatis, in addition to clinical evidence of nongonococcal urethritis, genital herpes, and exophytic genital warts, with adjustment for potentially confounding factors.


Study Population

The study population included men attending the Seattle-King County Department of Public Health Sexually Transmitted Disease Clinic at Harborview Medical Center between January and December 1988. The clinic, which provides comprehensive services related to sexually transmitted diseases such as diagnosis, treatment, education, and risk reduction counseling, served 80 to 120 patients per day in 1988. The exact number of unique men seen in the clinic in 1988 is unknown; the Department of Public Health tallies new problem visits, to which an individual can contribute multiple times. Among all male visits to the clinic in 1988, 8034 new problem visits were recorded. Study subjects were identified in l988 through patient medical charts that had been archived according to the year of visit. Charts are placed into the archives after approximately 1 year if the patient has not returned to the clinic in the interim. Charts of men who returned to the clinic in 1989 were not included in the study. According to statistics compiled by the clinic, the study population closely matched that of all men seen for a new problem in 1988 in regard to cases of syphilis (0.7% vs 0.8%), gonorrhea (8.1% vs 10.0%), warts (14.5% vs 14.2%), chlamydial infection (6.8% vs 6.9%), nongonococcal urethritis (29.0% vs 25.4%), and genital herpes (4.7% vs 5.0%). Thus, the study population was considered a good representative sample. Of the 4137 subjects identified, 897 did not meet the study inclusion criteria (380 who presented for HIV screening only and did not have a genital examination, 306 who were homosexual or bisexual, and 211 who did not reside in King County). Of the 3240 eligible study subjects, 464 did not have circumcision status recorded and were excluded from the analysis. The resulting study population consisted of 2776 heterosexual men. All data used in this study were abstracted in 1990 from standardized clinic records of men who attended the clinic in 1988. These records included information concerning demographic characteristics of the patient, reason(s) for visit, symptoms, sexual history (new sexual partner in the last month, number of sexual partners in the last month, and days since last sexual exposure), prior sexually transmitted disease history, and results of all physical and laboratory examinations. All consenting patients seen for a new problem underwent genital exam- ination and had urethral swabs taken for a Gram-stained smear and for cultures for N. gonorrhoeae and C. trachomatis. Blood for serologic testing of syphilis (reagent antibody) was obtained from all consenting patients, and darkfield examination was performed on suspected syphilitic lesions. Diagnoses of warts and genital herpes were presumptively based on the clinical appearance of lesions; cultures for herpes simplex virus were obtained when considered necessary by the clinician. Current syphilis was determined by a positive darkfield test from a suspicious lesion or by a positive rapid plasma reagent (RPR) or Venereal Disease Research Laboratory (VDRL) titer (titer above 1:2) with confirmatory T. pallidum.

Hemaglutination assay (TPHA) serology.

This definition included primary, secondary, and some early latent syphilis cases; however, all cases were newly diagnosed. The diagnosis of nongonococcal urethritis was based on the presence of at least two of the following criteria: subject-reported urethral discharge or dysuria, presence of a purulent urethral discharge on exam, or a urethral Gram-stained smear that showed at least five polymorphonuclear leukocytes per 1000X oil immersion field for at least three fields. The combination of the first two criteria was used only if the N. gonorrhoeae culture was negative. A subset of subjects (n = 278) had elevated polymorphonuclear leukocytes on urethral smear; these subjects were given a clinical diagnosis of nongonococcal urethritis and treated for this condition but were not recorded as having an abnormal urethral discharge. Addition of these subjects to those in the more stringently defined category of nongonococcal urethritis did not alter the study results, so they were combined in the final analysis.

For this study, the principal sexually transmitted diseases evaluated were laboratory-confirmed N. gonorrhoeae, C. trachomatis, and T. pallidum infection and clinical evidence of nongonococcal urethritis, genital warts, or genital herpes. The "other sexually transmitted disease" category comprised subjects with a variety of diagnoses, including pediculosis pubis (n = 81), scabies (n = 33), and molluscum contagiosum (n = 69). The reference group was composed of subjects who were not diagnosed with any of the above sexually transmitted diseases or conditions.

Circumcision status was recorded by the clinician as part of the routine examination. Place of residence in King County was categorized by zip code into six Seattle areas and one non-Seattle area, as explained in detail elsewhere.8

Statistical Analysis

The initial 1988 clinic visit of each subject was used for analysis. The prevalence of each of the principal study diseases was determined on the basis of circumcision status. In order to allow direct sexually transmitted disease prevalence comparisons between circumcised and uncircumcised men, prevalences were standardized to the combined age and race distribution of the entire study population.

Characteristics of those with sexually transmitted diseases and those in the reference group without such diseases were compared. Odds ratios (ORs) were used as the measure of association to assess the relationship between circumcision status and each disease category. Logistic regression9,10 was used to calculate adjusted odds ratios and the associated 95% confidence intervals (CIs). Since subjects could have multiple sexually transmitted disease diagnoses, all variables were placed in a single model to allow for the control of multiple diseases. Therefore, adjusted odds ratios reflect an association between a particular sexually transmitted disease and circumcision status that is not influenced by the presence of other sexually transmitted diseases. All analyses were conducted with SAS.


Prevalences of Sexually Transmitted Diseases Under Study Age- and race-standardized prevalence figures revealed that, in this clinic population, the prevalence of current syphilis was higher in uncircumcised men than in circumcised men (2.5% vs 0.6%; P < .001), as was the prevalence of gonorrhea (12.7% vs 8.1%; P < .01). Conversely, the prevalence of exophytic warts was higher in circumcised men (17.6% vs 11.0%; P < .001). No significant differences between uncircumcised and circumcised men were observed for prevalences of nongonococcal urethritis, (30.2% vs 32.7%), genital herpes (8.8% vs 8.7%), or urethral infection with C. trachomatis (7.2% vs 6.4%).

Subject Characteristics Related to Current Diagnosis

The study population, composed of 540 uncircumcised men (19.5%) and 2236 circumcised men (80.5%), is described in Table 1 according to sexually transmitted disease status. In comparison with the reference group, those with gonorrhea, chlamydial infections, and genital warts were younger (P < .05); subjects with a diagnosis of current syphilis, gonorrhea, chlamydial infection, or nongonococcal urethritis were more likely to be African American, whereas those with exophytic warts were more likely to be white. Subjects with current diagnosis of gonorrhea, chlamydial infection, and nongonococcal urethritis more often reported a new sexual partner in the previous month and two or more partners in the previous month than those in the reference group. Subjects diagnosed with any of the principal study sexually transmitted diseases were more likely than those with no disease diagnosis to present at the clinic with symptoms and less likely to present for screening.

TABLE 1-Subject Characteristics (%) According to Sexually
        Transmitted Disease Status

        Other Sexually
        Reference               Chlamydial      Genital
  Genital Transmitted Group*   Syphilis
sonotflea Infection       Herpes             Urethritis        Warts Diseases

             (n=985)  (n=33)  (n=262) (n=181) (n=251)(n=882)(n=463)(n=18O)
Age, y
11-19           8       6       13*     19***   6       10     7     17***
20-24           20      27      28*     30**    21      23    27*    39***
25-29           24      27      23      22      26      26    29*    22
30-34           17      12      17      13      21      15    18      7***
35+             27      27      17**    12***   21      21    14***  11***

Unknown          6       -       3       4       6       5     5      4


White           61      9***    21***   56      65      57*** 73***  62

African American

                23      61**    60***   31***   21      33*** 19     25

Other           13      27      15       9      10       9**   5***  10

Unknown          3       a       4       2       4       1     3      3

Marital status

Single          84      82      91**    93**    85    68***   86     89*

Married         11      12       6**     4**    10     8***   12      4**

Unknown          4       6       3       2       4     4       3      6

No. of sexual partners in previous month
0               14      18       1***    4***   10     6***   13     15
1               60      56      55*     61      65    63      65     57
2+              18      21      41***   30***   21    27***   17     23

Unknown          7       3       3*      4       3*    3*      5      4

New sexual partner in previous month

                31      33      63***   50***   26      43*** 21**   34*

No, of days since last sexual exposure

        07      60      33      69***   50      59      54    50     48
        8-14    13      24"     19**    27""    14      23*** 16*    19*
        15+     30      36       9***   22      20**    20*** 26     28

Reason for visit

Symptoms        51      55      89***   74***   86**    71*** 70***  8O***

Contact with:

Gonorrhea        5       3       8       7       1**     3        2**    3

Syphilis                 0.8    33**     .       .       0.2      0.1    .

Genital warts    6       3       1***    17**    2***   13        4**    4*

Other sexually transmitted disease
                14      3        2***     8**   14      12**     12      9*

Screen          28      3        5*       8***   9***   13***    16***  11***

Consultation     5     12        1**      3      3       2*       5      6

*Group with no sexually transmitted diseases.
*.01    <P< .05.
**.001  <P< .01.
***P < .001 (in comparison with the reference group).

Association of Uncircumcised Status with Current Disease


Crude and adjusted odds ratios for the association of uncircumcised status with the study diseases are presented in Table 2. Adjusted odds ratios were estimated by fitting a multiple logistic regression model adjusted for age, race, number of sexual partners in the previous month, place of residence, and the other sexually transmitted diseases. No adjustment was made for a new partner in the last month, days since last sexual exposure, marital status, and reason for visit because of their negligible effects on the estimated odds ratios. No adjustment was made for a prior history of sexually transmitted diseases because any influence of circumcision status was assumed to be the same for a prior diagnosis as for the current diagnosis, and adjustment would have controlled for the associations the present study sought to quantify. In the adjusted analysis, an elevated odds ratio was noted for the association of uncircumcised status with current syphilis (OR = 4.0, 95% CI = 1.9, 8.4) and gonorrhea (OR = 1.6, 95% CI = 1.2, 2.2). A negative association was noted for warts and uncircumcised status (OR = 0.7, 95% CI = 0.5, 0.9). Uncircumcised status was not associated with nongonococcal urethritis, genital herpes, or urethral infection with C. trachomatis.

Table 2 - Odds Ratios Associating Sexually Transmitted
          Disease with Circumcision Status

Crude Distribution,            No.     Crude     95%     Odds Odds Confidence

Reference group                198      787     1.0     1.0

Syphilis                        20      13      6.1     4.0    1.9,8.4
Gonorrhea                       87      175     2.0     1.6    1.2,2.2
Chlamydial infection            34      147     0.9     1.0    0.7,1.6
Nongonococcal urethritis       161      721     0.9     1      0.8,1.3
Genital herpes                  49      205     1.0     1.2    0.8,1.7
Genital warts                   51      412     0.5     0.7    0.5,0.9
Other sexually transmitted      26      154     0.7     0.8    0.5,1.2

*Adjusted for age group (13-19, 20-24, 25-29, 30-34, 35+)
race/ethnicity (White, African American, other), number of
sexual partners in the last month (0,1,2+), place of
residence (six Seattle areas and one non-Seattle area
defined by zip codes), and the sexually  transmitted diseases.


The results of this study support the hypothesis that circumcision status is related to some, but not all, sexually transmitted diseases. The positive findings associating syphilis and gonococcal infections with uncircumcised status in the present study were similar to those previously reported by the Public Health Department Special Treatment Clinic in Perth, Australia.6

Age-adjusted results from the Perth study indicated a positive relationship with both syphilis (OR = 5.4, 95% CI = 1.4, 21.1) and gonorrhea (OR = 2.3 95% CI = 1.5, 3.7). Contrary to results in the present study, the Perth study also found a positive relationship between uncircumcised status and herpes and warts. Discrepancies between the two studies could be due to the present study's finer adjustment for confounding variables and restriction to heterosexual men. Other studies have also investigated specific relationships between certain sexually transmitted diseases and uncircumcised status.

In a study conducted at the Whitechapel Clinic in London, a comparison of men with culture-confirmed herpes and controls revealed a higher proportion of men with herpes being uncircumcised.5 Interpretation of this finding is limited since the control group consisted of men with a number of other sexually transmitted disease diagnoses that each could have had a specific relationship with circumcision status.

A study of active duty army men found that uncircumcised status was protective for nongonococcal urethritis but not related to gonococcal infection.11 Although that study adjusted for important confounders, direct comparisons with the present study are difficult since the former included subjects who apparently contributed to the case groups several times. In the final analysis, there were approximately five times more men in both the gonorrhea and nongonococcal urethritis groups than in the control (no sexually transmitted disease) group. In addition, the ratio of nongonococcal urethritis cases to gonorrhea cases was much lower than that typically observed in sexually transmitted disease clinic populations. Finally, a study of Canadian army men reported a higher proportion of uncircumcised active duty men with warts and syphilis compared with the proportion of uncircumcised men among new recruits.12 Although suggestive, the results could easily be biased as a result of the comparison of two distinctly different groups of men.

It was not possible to evaluate chancroid because it was not found in this study population. It was also not possible to evaluate HIV infection as a result of its very low prevalence (estimated to be 0.6% among heterosexual men seen in the clinic from June 1988 through December 1990).

Therefore, no comparisons can be made with the many studies linking uncircumcised status with chancroid, including recent studies from East Africa that have linked uncircumcised status with both chancroid and HIV seroconversion.1,3 Mechanisms for differential risks of sexually transmitted diseases based on circumcision status remain undefined. Possible theories include the following:

(1) trauma of the intact foreskin during sexual intercourse produces microscopic lesions that increase susceptibility;

(2) the environment under the foreskin could extend the survival of infectious agents and thus prolong exposure to viable organisms;

(3) the thinner, less cornified epithelium on the glans and preputial sac of an uncircumcised male may provide less of a physical barrier to microbial invasion; and

(4) nonspecific balanitis, not uncommon in uncircumcised men, may predispose to certain sexually transmitted diseases.

The finding that circumcision status is associated with some but not all sexually transmitted diseases may be related to the differing pathogenesis of these diseases. For syphilis, entry by T. pallidum is thought to occur through microscopic abrasions and not through undamaged, keratinated epithelium.4 In uncircumcised men, trauma of the foreskin or the thin epidermis of the glans during sexual activity can open a passageway of infection for treponemes. The thicker, cornified epithelium on the glans of a circumcised male may be less susceptible to abrasions and, therefore, to entry by T. pallidum. For gonorrhea, the inner lining of the foreskin13 might create an environment conducive to the survival or growth of N. gonorrhoeae and, perhaps, to entry into the urethra and enhanced infectivity.

A limitation of the present study arises from the available information on genital herpes and genital warts. For genital herpes, diagnosis was based on the presence of clinical lesions, with or without confirmatory cultures, and subclinical infections could have been undetected. If uncircumcised subjects were more likely to have inapparent infections, this study could have underestimated the true association of uncircumcised status with genital herpes. The finding that genital warts were less likely to be diagnosed in uncircumcised men was of interest. Further information on the anatomic distribution of penile warts is actually the relevant information for assessing the influence of circumcision status on wart lesions and will be addressed in a subsequent paper.

A possible concern is that the study subjects presumably came to the clinic as a result of a problem related to a sexually transmitted disease and that, therefore, an elevated prevalence of one disease would predict a lower prevalence of some other disease. In this cross-sectional population, subjects were not selected on the basis of the presence of symptoms related to, or a diagnosis of, a sexually transmitted disease. Thirty-five percent of the study population did not have a sexually transmitted disease diagnosis. Furthermore, a subject could be diagnosed with multiple diseases, and, because testing for infection with N. gonorrhoeae and C. trachomatis and for syphilis was routine in all patients, diagnoses were not mutually exclusive. Since neither a sexually transmitted disease diagnosis nor a single diagnosis were obligatory for inclusion in the study, the prevalence of one disease should not necessarily have influenced the prevalence of another disease. Because the study population was composed of men presenting at a public clinic, we were unable to control for certain self-selection factors in the analysis. The frequency of exposure to sexually transmitted diseases, health-seeking behavior, and perception of disease could all influence whether or not a person presents at a clinic. Therefore, the results can be generalized only to men who, given the possibility of a sexually transmitted disease, would present at a clinic. Men who do not seek care or who seek care from clinicians in private practice are likely to be different from the men who present at clinics such as the one described here.

In addition, the study population represented a sample of all heterosexual male visits to the clinic in 1988. It is possible that the proportion of uncircumcised and circumcised men in the study population differed from the total 1988 clinic population. Of particular concern are the men who returned to the clinic in 1989 and thus were not included in the study population as a result of their charts not being archived. This potential selection bias cannot be evaluated directly because the circumcision status of all men seen at the clinic in 1988 or 1989 was not tallied in health department statistics. However, the health department does report the age and race/ethnicity distributions for all male heterosexual clinic visits.

Since these two characteristics were the strongest predictors of circumcision status in the study population, we compared the age and race/ethnicity distribution of the study population, all clinic visits of male heterosexuals in 1988, and all clinic visits of male heterosexuals in 1989. No significant differences were observed between the three groups, indicating a similar corn-position in terms of age, race/ethnicity, and, presumably, circumcision status. This implied that the source of the study population, the archived clinic charts, was a reasonable reflection of all heterosexual male clinic visits in 1988. Furthermore, the age and race/ethnicity distributions did not fluctuate from year to year, and, presumably, neither did the circumcision status of heterosexual men attending the clinic. Therefore, although this potential bias cannot be dismissed, we believe it is unlikely that it had a major impact on the study results.

As with all cross-sectional studies, it is difficult to distinguish newly acquired cases from more established cases, especially for sexually transmitted diseases such as warts and herpes. However, all of the study cases were classified as new problem visits at the clinic. The reason that some individuals delayed in seeking medical care is unknown but could be due to variations in disease perception, as well as in behavior with regard to seeking health care. As a result of this limitation, we addressed only the issue of who was or was not diagnosed with a sexually transmitted disease at their initial 1988 clinic visit, not acquisition per se.

A unique feature of circumcision is that it occurs almost exclusively in infancy or early childhood in the United States. A very small percentage of men require or request the procedure as adults. Therefore, we can safely assume that in almost all of the men attending the clinic, circumcision status was determined prior to the occurrence of the disease. Thus, although a cross-sectional design was used to investigate the association between circumcision and sexually transmitted diseases, the temporal sequence of events was not in question.

The determination of circumcision status is not always straightforward be cause of incomplete circumcisions and foreskins that naturally retract to some degree. Information from a larger database of patients seen in the clinic from 1980 to 1989 was used to compare the recorded circumcision statuses for any patient with multiple visits. Of the 5666 patients in the database, 2266(40%) were seen more than once. Only 13 subjects had conflicting circumcision statuses recorded. Therefore, we are confident that the clinicians consistently identified and recorded circumcision status in the vast majority of subjects.

Benefits and disadvantages of male circumcision continue to be controversial, although recent research has attempted to address pertinent health issues associated with circumcision status.14,15 The results of this study do not show a definitive benefit of circumcision; although uncircumcised men were more likely to have syphilis and gonorrhea, circumcised men were more likely to have genital warts and equally likely to have nongonococcal urethritis and genital herpetic lesions.


This work was supported by National Institutes of Health grant AI27757. The research described in this paper was presented, in poster form, at the annual meeting of the Society for Epidemiologic Research, June 1991, Buffalo, NY.


  1. Simonsen JN, Cameron DW, Gakinya MN, et al. Human immunodeficiency virus infection among men with sexually transmitted diseases. N Engl J Med 1988;319:274-278.
  2. Cameron DW, D'Costa U, Maitha GM, et al. Female to male transmission of human immunodeficiency virus type I: Risk factors for seroconversion in men. Lancet. 1989;2:4O3-407.
  3. Hira SK, Kamanga J. Macuacua R, Mwansa N. Cruess DF, Perine PL. Genital ulcers and male circumcision as risk factors for acquiring HIV-1 in Zambia. J Infect Dis. 1990;161:584-585.
  4. Holmes KK, Mardh P, Sparling PF, Wiesner PJ, eds. Sexually Transmitted Diseases. New York, NY: McGraw-Hill; 1990.
  5. Taylor PK, Rodin P. Herpes genitalia and circumcision. Br J Venereal Dir. 1975;51:274-277.
  6. Parker SW, Stewart AJ, Wren MN, Gollow MN, Straton AY. Circumcision and sexually transmissible disease. Med J Aust. 1983;2:288-290.
  7. Thirumoorthy T, Eng EH, Doraisingham S, Ling AE, Lim KB, Ieee CT. Purulent-penile ulcers of patients in Singapore. Genitourin Med. 1986;62:252-255.
  8. Washington State/Seattle-King County HIV-AlDS Quarterly Epidemiology Report. Seattle, Wash: Washington State Health Department; second quarter, 1991.
  9. Breslow N, Powers W. Are there two logistic regressions for retrospective studies? Biometrics. 1978;34: 100-105.
  10. Breslow NE, Day NE. Statistical Methods in Cancer Research. The Analysis of Case-Control Studies. Lyon, France: International Agency for Research on Cancer; 1980.
  11. Smith GL, Greenup R, Takafuji ET. Circumcision as a risk factor for urethritis in racial groups. Am J Public Health 1987; 77:452-454.
  12. Wilson RA. Circumcision and venereal disease. Can Med Assoc J. 1947;56:54-56.
  13. Ross MH, Reith EJ. Histology, a Text and Atlas. New York, NY: Harper & Row; 1985.
  14. Wiswell TE. Routine neonatal circumcision: a reappraisal. Am Fam Physician. 1990;41:859-863.
  15. American Academy of Pediatrics. Report of the Task Force on Circumcision. Pediatrics. 1989;84:388-391.

This paper was accepted May 11, 1993.

(File revised 14 September 2006)

Parent directory

Return to CIRP library