AMERICAN JOURNAL OF PUBLIC HEALTH, Volume 77 Number 4: Pages 452-454.
April 1987.

Circumcision as a Risk Factor for Urethritis in Racial Groups


Abstract: A retrospective population-based case-control study of sexually transmitted urethritis was conducted at a large military base over a 21-month period. During the study, 9514 patients were seen for sexually transmitted disease. The analysis was restricted to active duty males and showed that Blacks had 14.8 times the incidence rate for gonococcal urethritis (GCU) and 4.7 times the rate of nongonococcal urethritis (NGU) compared to Whites. There were slightly fewer cases of NGU than GCU. A case-control study of active duty soldiers were 1.65 times as likely to have NGU as uncircumcised subjects (95% CI): 1:37-2.00). However, circumcision was not associated with an increased incidence of GCU. (Am J Public Health 1987; 77:452-454.)


Gonococcal Urethritis (GCU) is the most commonly occurring reportable communicable disease in the US,1 with over a million cases reported annually. Non-gonoccocal urethritis (NGU) secondary to Chlamydia trachomatis, Ureaplasma urealyticum, and other agents is not a reportable disease in many states. However, the incidence is estimated to be twice that of GCU.2-5 Population based incidence studies,6-7 cohort studies,8-10 and clinic based studies11 have all shown that GCU occurs two to 10 times more frequently in the Black population, whereas the NGU rates are about the same or slightly higher among Whites. Risk factors shown to be associated with the acquisition of GCU and NGU include unmarried status, multiple sex partners, low socioeconomic status, and a history of previously sexually transmitted urethritis (STU). However, when all of these factors are controlled for, the rate of GCU remains higher in Blacks.3-5

The data are sparse regarding the importance of the prepuce in the acquisition of STU. Parker et al,12 in their case control study of 1,350 men attending a public health clinic in Perth, Australia showed that the uncircumcised men were twice as likely to have GCU. There was no association with NGU. However, confounding by number of sexual partners, marital status, and use of condoms was not controlled for in that clinic-based study.

The prepuce may provide a physical or immunological barrier against infection or an excellent milieu for infectious agents to grow and multiply.12 This prompted us to investigate the possibility that the elevated risk of STU found in Blacks was due to the influence of the prepuce, since circumcision is more prevalent in the White population.

In order to test the hypotheses that circumcision is associated with the risk of acquiring GCU or NGU, a retrospective, population-based case-control study was conducted at a large US military base.

Study Site
           The study was conducted at a US military post with a history of high rates of sexually transmitted disease (STD) in active duty members, dependents, and the local civilian community.13 Regulations at the post, enforced with good success for a number of years, included a strict policy of referring all STD cases to the Epidemiological Disease Clinic (EDC). In addition to STDs, the EDC is used for all those diseases of epidemiologic importance, and all reportable communicable disease; it serves a population averaging 48,472 on active duty and 101,791 eligible dependents.

Case and Control Definitions
           Persons seen by a military physician on the post with signs or symptoms consistent with any STD and persons eligible for care at the military hospital who presented to the local county health department with signs or symptoms consistent with any STD were referred without treatment to the EDC. Only those active duty members who elected to seek care from and pay a private physician were missed from the STD surveillance.

Criteria for inclusion in the study population as a case included: presentation to the EDC during the period of the study January 1, 1983 to September 30, 1984; granting of informed consent (there were no refusals); completion of a standardized history and physical and laboratory examination; active duty male; receiving the diagnosis of GCU or NGU according to Centers for Disease Control criteria.14

Criteria for inclusion in the study population as a control were similar except that the patients presented with a history of recent sexual contact, and were worried about having a STD, yet after examination were found to be without signs and symptoms consistent with any STD and their urethral smear gram stains and modified Thayer-Martin cultures were negative.

           Every subject entering the EDC during the study period had a standardized history by a STD technician including sociodemographic information, time since last exposure and use of condoms. This was followed by a contact interview and a standardized genitourinary physical examination. For this study, circumcision was defined as the objective surgical absence of a prepuce.

All subjects had a urethral smear gram stain and a modified Thayer-Martin Plate innoculated and incubated in 5 per cent carbon dioxide incubator. Neisseria gonorrhoeae was identified and confirmed by standard technique. Those subjects with a urethral smear positive for gram negative intracellular diplococci (GNID) and/or modified Thayer-Martin culture positive for Neisseria gonorrhoeae were diagnosed as GCU.14,15 Those subjects with a urethral smear negative for GNID and four or more polymorphonuclear leukocytes per oil immersion field (PMNs/HPF), and a modified Thayer-Martin culture negative for Neisseria gonorrhoeae were diagnosed as NGU.14,15

Centers for Disease Control Guidelines14 were used to treat all cases of GCU. Those subjects with GCU who had post-gonococcal urethritis (four or more PMNs/HPF in the absence of GNID, with a negative modified Thayer-Martin urethral culture test-of-cure) after adequate treatment were given the additional diagnosis of NGU.

Over 97 per cent of the infections were treated initially with one antibiotic. In this way, mixed GCU/NGU infections were more likely to be identified and coded. A subject with GCU who was diagnosed after treatment as having post-gonococcal urethritis secondary to NGU was coded as one incident case of GCU and NGU. If a subject returned to the EDC at any time after a successful course of treatment with an interval without symptoms, he was treated as a new incident case.

Table 1--Case and Control Demographic Comparison

                                    Non-Gonococcal    Gonococcal
      Demographics       Control      Urethritis      Urethritis
                        (N = 563)     (N = 2543)      (N = 2982)
                                           %              %

   White                    29.0          33.0           14.0
   Black                    65.0          60.0           81.0
   Other Ethnic Group        6.0           7.0            5.0
   Junior Enlisted          25.5          28.4           31.2
   Junior Non-Commissioned
     Officer                51.3          56.2           58.8
   Senior Non-Commissioned
     Officer                21.6          12.3            9.3
   Commissioned Officers     1.5           3.1            0.7
   Non-High School Graduate  4.4           4.9            6.2
   High School Graduate     73.9          73.2           80.3
   Some College             21.7          21.9           14.5

Data Analysis
           The data was put into a Vax 11/700 mini-computer and checked for accuracy and coding errors. Only active duty males with complete information were included in this analysis. The data set was analyzed using the Statistical Analysis System (SAS) for cross-tabulations. The PLR program on the BMDP statistical package was used to perform non-stepwise multiple logistic regression on the data set. Mantel-Haenszel summary odds ration16-18 were performed on the data set.

Table 2--Multiple Logistic Regression Analysis of Gonococcal
            Urethritis by Circumcision and other Covariates

   Variable                         Odds Ratio          (95% CI)

   Circumcision versus
      Uncircumcised                   0.88             (0.71-1.09)
   Black vs. White                    1.20             (1.11-1.32)
   Other vs. White                    1.00             (0.87-1.16)
Sexual Partners
   2-3 vs.0-1                         1.09             (0.96-1.22)
   4+ vs. 0-1                         1.27             (1.05-1.54)
Marital Status
   Married vs 
      Unmarried                       0.74             (0.50-0.93)
   High-school graduate
      vs Non-graduate                 0.76             (0.59-0.99)
Age (years)                           0.98             (0.96-1.01)

Table 3--Multiple Logistic Regression Analysis of Non-Gonococcal
            Urethritis by Circumcision and Other Covariates

   Variable                          Odds Ratio          (95% CI)

   Circumcision versus
      Uncircumcised                   1.65             (1.37-2.00)
   Black vs. White                    1.01             (0.90-1.14)
   Other vs. White                    0.81             (0.64-1.01)
Sexual Partners
   2+vs.0-1                           0.93             (0.83-1.04)
Marital Status
   Married vs 
      Unmarried                       0.95             (0.76-1.19)
   High-school graduate
      vs Non-graduate                 0.97             (0.88-1.06)
Age (years)                           0.98             (0.92-1.04)


During the 21 months of the study, there were 9,514 STD subjects seen in the EDC. Eighty-four per cent were male, 67 per cent Black, 27 per cent White, and 6 per cent oter ethnic groups. Eighty-six per cent were active duty members and 14 per cent were military dependents. During the study the active duty male population averaged 45,250 resulting in 950,250 persons/months of observation.

In Table 1, the race, rank, and educational parameters of the control population and two individual case-control studies being evaluated are shown. The proportion of Blacks in the GCU population was markedly higher than that of the control group, and the proportion of GCU cases with "some college" was somewhat lower than the control group.

There were 3,057 cases of GCU of which 2,982 were in active duty males with a resultant incidence density of 3.13/1000 person-months for the active duty population. There were 2,653 cases of NGU of which 2,543 were in active duty males with a resultant incidence density of 2.68/1000 persons-months for the active duty population.

Blacks had 4.7 times the rate of the Whites, for NGU and 14.8 times the White rate for GCU. The other ethnic groups were intermediate between Blacks and Whites for both diseases.

Circumcision Effects
           Eighty-five per cent of the Whites, 34 per cent of the Blacks, and 33 per cent of the other ethnic groups were circumcised. Less than 1 per cent of both cases reported using condoms. There was no association between GCU and circumcision in any of the three racial groups (Mantel-Haenzel summary odds ratio 0.90 (95 per cent CI 0.73, 1,11).

In order to control for a variety of factors potentially confounding such an association, a maximum likelihood estimate (MLE) logistic regression analysis was done using GCU as the dependent variable in the model. Table 2 shows the results of the analysis, odds ratio, and 95 per cent CI for each variable, after controlling for the other variables. GCU was not associated with circumcision in this analysis.

In contrast to GCU, circumcision was associated with an increased incidence of NGU in both Blacks and Whites. The Mantel-Haenzel summary odds ratio was 1.70 (95% CI 1.37, 2.10) (data available on request to authors).

In Table 3, the same logistic regression analysis was conducted using NGU as the dependent variable and the same independent variables. Only circumcision was associated with NGU in the model, with an odds ratio of 1.65 (95 per cent CI 1.37-2.00)>


Racial differences in STD incidence were higher than those previously reported.2-5 Some of this difference could reflect racial differences in seeking care from private physicians but this is not felt to be important based on information provided by private physicians, and the representativeness of the ranks in the study population.

The incidence figures also indicate that NGU is occuring less frequently than GCU in this population in contrast to previously reported ratios of 2-2.5 NGU cases per GCU case in civilian populations.2,3 It is possible that the nature of this young, military population may account for the different ratio.

The control group was representative of members of the military post community who are sexually active and worried about having been in contact with a partner with STD but who have not acquired STU as shown by the closeness of the case and control populations in Table 1 with respect to all the variables except the presence or absence of diagnosed disease.

It is clear from Tables 2 and 3 that the racial differences almost disappear in this case-control analysis , after controlling for the other variables. This suggests that for both NGU and GCU there is no biologic mechanism accounting for the difference in rates among the races identified in the population-based incidence data.

The data in Table 2 re-emphasize the association of marital status, race, education, and number of sexual partners in the last month with the risk of acquiring GCU, as well as providing evidence for no association between circumcision and the risk of acquiring GCU.

For NGU, the results in Table 3 show that among this population of sexually active men, having a prepuce (being uncircumcised) was protective for acquiring NGU and there were no associations with the other variables and NGU.

Since the symptoms of NGU are often subtle, it is possible that uncircumcised men were more likely than circumcised men to overlook the presence of urethral discharge. On the other hand, the law proportion of NGU to GCU in this study population may reflect the fact that more symptomatic NGU cases were seen in this study.

The findings of this study are intriguing because circumcision is much less common in Black and other ethnic groups compared with Whites and one a priori hypothesis was that the presence of a prepuce may increase the risk of acquiring NGU and GCU independent of other factors. Wiswell et al19 found a protective effect of circumcision when looking at the rate of urinary tract infections in infant males, and Parker, et al12 found circumcision to be protective for GCU. However, the findings of this study suggest that the prepuce has a protective effect against the acquistion of NGU and that there is no association with GCU.

Having a prepuce may be protective against NGU by effecting the physiologic milieu of the glans penis, by association with post-coital hygiene behavior, or by local immune defense mechanisms acting against the agent. Further basic research is necessary to clarify these relations.


The views of Dr. Smith, Dr. Takafuji, and Mr. Greenup are their own and do not purport to represent those views of the US Department of the Army or the US Department of Defense.


  1. Centers for Disease Control: Sexually Transmitted Diseases. Annual Summary, 1984.
  2. Holmes KK: The Chlamydia Epidemic. JAMA 1981: 245:1718-1723.
  3. Thompson, SE, Washington AE: Epidemiology of venereal urethritis: Comparison of chlamydia trachomatic infections, Epidemiol Rev 1983; 5:96-123.
  4. McCutchan JA: Epidemiology of venereal urethritis: Comparison of gonorrhea and nongonococcal urethritis: Rev Infect Dis 1984; 6:669-688.
  5. Hook EW, Holmes KK: Gonococcal infections. Ann Intern Med 1985; 102:229-243.
  6. Gale JL, Hinds MW: Male urethritis in King County, Washington, 1974-1975. I. Incidence. Am J Pub Health 1978; 68:20-25.
  7. Morton W, Horton HB, Baker HW: Effects of socioeconomic status on incidences of three sexually transmitted diseases. Sex Transm Dis 1979; 6:206-210.
  8. Levine JB, Erickson JM, Dean LM: Social aspects of venereal diseases aboard a US Navy destroyer. J Am Vener Dis Assoc 1976; 3:35-39.
  9. Hooper RR, Reynolds GD, Jones CG, et al: Cohort study of venereal disease. I. The risk of gonorrhea transmission from infected women to men. Am J Epidemiol 1978; 108:136-144.
  10. Jones CG: Cited in: Hobson D. Holmes KK (eds): Non-gonococcal Urethritis and Related Infections. Washington, DC: American Society for Microbiology, 1977; 11-12.
  11. Judson FN: Epidemiology and control of nongonoccocal Urethritis and genital chlamydial infections: a review. Sex Transm Dis 1981; 8:117-126.
  12. Parker SW, Stewart AJ, Wren NM. et al: Circumcision and sexually transmitted disease. Med J Aust 1983; 2:228-290.
  13. North Carolina State, Department of Public Health, Vital Statistics, Communicable Disease Surveillance, 1981.
  14. Centers for Disease Control: Sexually transmitted diseases treatment guidelines 1982; 31:33S-60S.
  15. Washington State, Division of Health and Venereal Disease Program: Sexually Transmitted Diseases: Diagnostic and Treatment Guidelines 1982 Olympia: Washington State Department of Social and Health Services (Pamphlet), 1982.
  16. Mantel N. Tests and limits for common odds ration of several 2X2 contingency tables: Methods in analogy with the Mantel-Haenzel procedure. J Statist Plann Infer 1977; 1:179-189.
  17. Mantel N. Hagenstel W. Statistical aspects of the analysis of data from retrospective studies of disease. NCI 1959; 22:739-748.
  18. Woolf B: On estimating the relation between blood group and disease. Ann Human Genet 1955; 9:251-253.
  19. Wiswell TE, Smith FR, Bass JW: Decreased incidence of urinary tract infections in circumcised male infants. Pediatrics 1985; 75:901-903.

This paper submitted to the Journal May 7, 1986, was revised and accepted for publication August 12, 1986.

(File revised 15 September 2006)