Xu, Fujie MD, PhD; Markowitz, Lauri E. MD; Sternberg, Maya R. PhD; Aral, Sevgi O. PhD
MALE CIRCUMCISION, usually performed during the neonatal period, is a common surgical procedure in the United States, yet there are few data about circumcision prevalence in the general population.1,2 Male circumcision became popular as a routine medical operation in the United States in the late 19th century as a result of advocacy by physicians for preventing a wide variety of conditions and behaviors ranging from paralysis to masturbation.3,4 After systematic medical review, however, American Academy of Pediatrics (AAP) recommended against routine circumcision in 1971.5 In most English-speaking countries except the United States, the prevalence of neonatal circumcision is fairly low.4,6
Using an evidence-based approach, AAP reported in its most recent policy statement that there was “substantial body of evidence” linking male circumcision with lower risk for human immunodeficiency virus (HIV) infection.7 Studies in recent years consistently support this link.8–12 Furthermore, male circumcision reduced the risk of female-to-male transmission of HIV by 61% in a randomized control trial.13 One of the postulated mechanisms by which circumcision can decrease the risk for HIV infection is through decreased propensity to ulcerative sexually transmitted infections (STIs) in circumcised men.10 In the same AAP statement, the association between male circumcision and other STIs, including herpes simplex virus Type 2 (HSV-2), was reported as “complex and conflicting.”7 Several recent studies did not find any association between circumcision and HSV-2 infection,12,14–17 while others found circumcision associated with lower risk for HSV-2 infection.18,19 The continued controversy suggests that the effect of circumcision on HSV-2 infection may vary by population characteristics such as HSV-2 seroprevalence.
In this study, we analyzed the prevalence of circumcision in the United States and investigated the association between circumcision and HSV-2 infection in various subpopulations. The data for our analyses are from cross-sectional national representative surveys, National Health and Nutrition Examination Surveys, from 1999 to 2004 (NHANES 1999–2004).
Study Population and Survey Design
The NHANES is a series of cross-sectional surveys using a complex, stratified, multistage probability sampling design. Details of the survey methodology have been published previously.20 Briefly, during each year of the survey, a nationally representative sample of the US civilian, noninstitutionalized population was selected, interviewed, and examined by medical professionals. Some populations such as adolescents, non-Hispanic blacks, and Mexican Americans were oversampled to increase the sample sizes for these subpopulations. Survey data from 2 or more years are usually combined to achieve adequate sample size to provide national statistics.
Circumcision Status and HSV-2 Serology
During NHANES 1999–2004, health examinations were conducted in specially equipped mobile examination Centers (MECs). The questionnaire module with questions on circumcision status and sexual behaviors was administered to persons aged 14 to 59 years in a private room using audio computer-assisted self-interview (ACASI) in either English or Spanish. The ACASI enables the respondents both to hear questions through earphones and read questions on the computer and to move at their own speed as they touch the screen to indicate their response. Medical artwork of a circumcised and an uncircumcised penis was used as visual aids (Fig. 1), and the respondent indicated his circumcision status by selecting the picture illustrating the appearance of his penis. In addition to circumcision status, questions were asked about sexual behaviors. Of all boys/men aged 14 to 59 years who were interviewed for NHANES 1999–2004, 95% were examined at the MECs.
Blood samples from survey participants were collected during health examinations at the MECs. Only persons aged 14 to 49 years were tested for antibodies to HSV-2, using a type-specific immunodot assay.21 Purified glycoprotein gG-2 of HSV-2 was used as antigen in the assay. Of boys/men aged 14 to 49 years who were selected for NHANES 1999–2004, 96% were examined at the MECs, and 92% of those examined at the MECs were tested for HSV-2.
SUDAAN software (Release 9.0, Research Triangle Institute, Cary, NC) was used for statistical analyses to account for the complex survey design. We estimated the prevalence of circumcision by demographic factors according to NHANES design domain and the year of birth. The year of birth was calculated from survey participants’ age and the year in which the survey was conducted. The calculated year of birth in NHANES 1999–2004 ranged from 1940 to 1990. In this survey, race/ethnicity was defined by self-report as non-Hispanic white, non-Hispanic black, Mexican American, and other. The “other” group includes Asians and Pacific Islanders, Native Americans, non-Mexican American Hispanics, persons reporting more than one race, and those whose race/ethnicity was unknown.22
All prevalence estimates were weighted to represent the civilian, noninstitutionalized US population and to account for oversampling and nonresponse to the interview and the examination.23 Confidence intervals (CI) for the prevalence estimates were calculated based on a log transformation with the standard error calculated using the delta method.24 The Wald Chisq test was used to test the association between two categorical variables, and the Cochran-Mantel-Haenszel test was used to test the association between two categorical variables when stratified by a third variable. An analysis of variance was used to compare continuous variables by categorical variables and the P value from the Wald F test is reported. A log transformation was used when appropriate to improve the assumption of normality of the continuous variable. Logistic regression models were used to examine the association between two variables while adjusting for potential confounders. The adjusted Satterthwaite F test was used to assess the statistical significance of variables in the logistic regression model.
Prevalence and Trends of Circumcision
During NHANES 1999–2004, 6,898 men born from 1940 to 1989 and aged 14 to 59 years at the time of the survey were examined at MECs. A participant who was examined at MECs might skip the ACASI module due to time restraints; 6313 (92%) were interviewed about circumcision status (“Are you circumcised or uncircumcised?”). Of these, 53 refused to answer the question and 86 responded with “Don’t know,” leaving a sample of 6174 persons (90% of those examined at MECs) for further analyses.
The overall prevalence of circumcision was 79% (95% CI 77–80) (Table 1). Circumcision prevalence varied by the year of birth. The overall prevalence was fairly stable in those born in 1950s through the 1980s (78% to 81%) but was significantly lower in those born in the 1940s (71% P <0.001). Circumcision prevalence differed greatly by race/ethnicity, with non-Hispanic whites having the highest prevalence (88%) and Mexican Americans the lowest (42%) (Table 1). The prevalence of circumcision was higher in persons with a higher educational level and those born in the United States (both P <0.001, Table 1).
To further examine the trend in circumcision practice in the United States, we analyzed the prevalence of circumcision by the year of birth among those born in the United States (Fig. 2). Men who reported their birthplace was outside the 50 states or Washington, DC were excluded from this trend analysis because most circumcisions are performed soon after birth. Among non-Hispanic whites, the prevalence of circumcision was 80% in those born during 1940–1949. The prevalence of circumcision increased in those born from 1950 through 1979 and peaked in those born in the 1970s at 94%. In those born in the 1980s, the prevalence of circumcision (88%) decreased significantly from its peak (P <0.001, Fig. 2). Among non-Hispanic blacks, the prevalence of circumcision increased steadily from 50% in those born in the 1940s to 91% in those born in the 1970s; however, in those born in the 1980s, the prevalence of circumcision decreased to 81% (a statistically significant decrease from the peak of 91% in those born in 1970s, P = 0.002). A similar pattern was observed in Mexican Americans: the prevalence of circumcision increased from 34% in those born in 1940s to 57% in those born in 1970s and then decreased to 51% in those born in the 1980s (Fig. 2). When the three race/ethnicity groups above were combined, the overall prevalence of circumcision was 7 percentage points lower in those born in the 1980s compared to those born in the 1970s (84% vs. 91%, P <0.001).
In NHANES 1999–2004, no data were collected on parents’ education level or other socioeconomic variables that may correlate with the ability to pay for circumcision at the time of the participant’s birth. However, total household income in the past year was asked for all survey participants. Comparing persons born in the 1970s with those in the 1980s (limiting to those born in the United States), significant decreases in circumcision prevalence occurred in those below poverty level (from 84% to 75%, P = 0.03) as well as in those at or above poverty level (from 93% to 84%, P <0.001). In addition, circumcision prevalence decreased in those who had high income (annual household income >$55,000, from 96% to 85%, P <0.001) as well as in those who had modest income (annual household income between $35,000 and $54,999, from 92% to 85%, P = 0.03). Circumcision prevalence by participants’ education level were not compared because some participants were of school age.
Circumcision and Sexual Behaviors
Comparisons of sexual behaviors were limited to 4872 boys/men (1397 uncircumcised and 3475 circumcised) who reported having had sex. The mean age of sexual initiation did not differ by circumcision status (16.7 years in uncircumcised men and 16.9 years in circumcised men, P = 0.3), nor the percent of men who had ever had a male partner (3.4% in uncircumcised men and 4.9% in circumcised men, P = 0.07). The median number of lifetime sex partners was 5.8 (95% CI 5.1–7.6) in uncircumcised men and 7.0 (95% CI 6.1–7.8) in circumcised men; the geometric mean was also similar (6.8 in uncircumcised vs. 7.5 in circumcised, P = 0.1). The median number of lifetime sex partners differed by race-ethnicity (6.5 in non-Hispanic whites, 11.0 in non-Hispanic blacks, and 4.6 in Mexican Americans, P <0.001). Stratified analyses by race/ethnicity found no statistically significant differences in the median number of sexual partners by circumcision status.
Circumcision and HSV-2 Infection
This analysis was limited to 3850 boys/men 14 to 49 years of age who reported having had sex and were tested for HSV-2 antibodies. The overall HSV-2 seroprevalence was 13.7% (95% CI 11.0–17.2) in uncircumcised men and 11.6% (95% CI 10.1–13.4) in circumcised men (P = 0.2) (Table 2). Stratified analyses by common demographic and behavior factors are presented in Table 2.
Compared with circumcised men, uncircumcised men appeared to have higher HSV-2 seroprevalence among those born from 1950 through 1959, among Mexican Americans and among those who were born in Mexico (Table 2). A logistic regression model was used to simultaneously control for potential confounders presented in Table 2. HSV-2 infection was associated age, race/ethnicity and sexual behaviors, but was not associated with circumcision status (odds ratio (OR) = 1.1, 95% CI 0.8–1.5) after controlling for potential confounders (Table 3). We also assessed whether the effect of circumcision on HSV-2 infection differed by selected demographic factors, by adding interaction terms between circumcision status and birth cohort, race/ethnicity and birthplace (variables significantly associated with HSV-2 infection in bivariate analyses, Table 2) in the logistic regression model; no interaction was statistically significant.
Of men aged 20 to 49 years who had had sex, 2.7% had the dual biomedical factors that are potentially important for HIV acquisition—infection with HSV-2 and lack of circumcision. The prevalence differed by race-ethnicity: 8.5% in non-Hispanic blacks, 7.5% in Mexican Americans and only 0.8% in non-Hispanic whites (P <0.001).
The overall prevalence of circumcision in men in the United States who were born from the 1940s through the 1980s was high at 79%. Among those born in the United States, the prevalence of circumcision decreased in those born in the 1980s compared to the 1970s. Because circumcision at older ages (i.e., after 14 years of age) is infrequent, our data indicate that the rate of circumcision peaked in the 1970s and then decreased. This decrease may have been in part due to AAP’s recommendations against routine circumcision since 1971. The first AAP policy statement was published in 1971 and reiterated in 1975 and 1983. In the 1989 and 1999 policy statements, AAP concluded that male circumcision has potential health benefits, but routine neonatal circumcision was not recommended.7 The decreases in circumcision in those born in the 1980s apparently were not limited to people who might have difficulty paying for medical cost of circumcision.
In the United States, health benefits of circumcision may be among the reasons parents and healthcare providers choose to circumcise newborns. Our study examined the association between circumcision and HSV-2 infection, and found that HSV-2 seroprevalence did not significantly differ in the overall general population, nor in subpopulations defined by demographic or behavior factors. If circumcision would provide partial protection again HSV-2 infection, persons at relatively lower risk for HSV-2 infection may be more likely to be protected (before the partial protection is “overwhelmed” by risk). In this analysis, circumcision was not associated with lower HSV-2 infection in subpopulations with relative low HSV-2 seroprevalence (such as those with ≤9 lifetime sex partners), nor in populations with relatively high seroprevalence (such as those with ≥10 lifetime sex partners).
The prevalence of self-reported circumcision in our study is consistent with results from a previous study, which described circumcision prevalence for persons born between 1933–1974 (77% in US-born and 42% of non-US-born).1 Data from national hospital discharge surveys and other sources support our findings concerning the differences in circumcision prevalence by demographic factors, such as race/ethnicity, education, and birthplace.1,2,25,26 Among non-Hispanic black and Mexican American men born during the 1940s through 1970s, the prevalence of circumcision increased dramatically. As a result, the differences in the prevalence of circumcision between racial-ethnic groups were smaller in the younger generations. Despite these changes and the recent changes in HSV-2 seroprevalence in the US population,27 the proportion of men who have the two factors (lack of circumcision and HSV-2 infection) that potentially make them more susceptible to sexually acquired HIV infection are still much higher in non-Hispanic black and Mexican American men than in non-Hispanic white men. Even with the decrease from its peak, the prevalence of circumcision in the youngest birth cohort in the United States is still the highest among English-speaking countries.4 A population survey in 2000 found that circumcision prevalence was 16% in British men aged 16 to 44 years, and was 12% in those aged 16 to 19 years.15
Our analysis has some limitations. Circumcision status was based on self-report, and the association between circumcision and HSV-2 infection may be underestimated due to nondifferential misclassifications.28 Variations in the degree of foreskin removal may have caused confusion when study respondents saw the sketches of a circumcised and an uncircumcised penis, resulting in misclassifications. Prior studies found very good agreement (≥95%) between self-reported and clinician-assessed circumcision status, even in the absence of visual aids.14,29 However, the agreement was lower (93%) in a study among mostly minority youth (mean age 15 years) in the United States.30 Second, no data were collected on the timing of circumcision. The majority of circumcision occurs during the neonatal period in the United States, although a considerable number of men are circumcised at older ages. About 1.2 million newborn infants were circumcised before they were discharged from the birth hospital in 2002.2 In contrast, it was estimated that 142,000 male circumcision procedures were performed beyond neonatal period, and of these, 49,000 were in persons older than 15 years of age in 1996.26 From these numbers, we estimate that for every one circumcision in persons older than 15 years, there are more than 26 such procedures performed in younger persons in the United States. Circumcision in late teenage years or older would result in a small increase in circumcision prevalence we observed in the younger birth cohorts, but due to the small number, it is unlikely to eliminate the decrease in circumcision prevalence among those born in the 1980s. Third, although we used 6 years of NHANES data for this analysis, in subpopulation analyses of HSV-2 seroprevalence by circumcision status, the differences were statistically significant only when the prevalence ratio was greater or equal to 1.6 (Table 2). It is possible that small, yet from the public health perspective, important protection by circumcision for HSV-2 infection was missed due to limited sample sizes.
Our findings suggest that the reported protective effect of circumcision against HIV infection found in other studies is unlikely to be mediated by differences in sexual behaviors or the risk for HSV-2 infection. Circumcision status was not associated with any sexual behaviors included in the survey. The primary biologic mechanism for the protective effect of male circumcision may be the removal of Langerhans and other HIV target cells that have been found in high densities in the inner surface of human foreskin.9,11 In contrast, HSV-2 virus infects a wide range of cells and replicates in cells of the epidermis and dermis.31 For HSV-2, there is no biologic mechanism other than that the presence of foreskin may increase occurrence of scratches, tears and abrasions during sexual intercourse. Therefore, it is plausible that the removal of foreskin provides less protection against HSV-2 than HIV infection.
1. Laumann EO, Masi CM, Zuckerman EW. Circumcision in the United States. Prevalence, prophylactic effects and sexual practice. JAMA 1997; 277:1052–1057.
2. Kozak LJ, Hall MJ, Owings MF. National hospital discharge survey: 2002. Annual summary with detailed diagnosis and procedure data. National Center for Health Statistics. Vital Health Stat 13 2002; 153:1–194.
4. Alanis MC, Lucidi RS. Neonatal circumcision: a review of the world’s oldest and most controversial operation. Obstet Gynecol Surv 2004; 59:379–395.
5. American Academy of Pediatrics, Committee on Fetus and Newborn. Standards and Recommendations for Hospital Care of Newborn Infants. 5th ed. Evanston, IL: American Academy of Pediatrics; 1971.
6. Nelson CP, Dunn R, Wan J, et al. The increasing incidence of newborn circumcision: data from the nationwide inpatient sample. J Urol 2005; 173:978–981.
7. Circumcision policy statement. American Academy of Pediatrics. Task force on circumcision. Pediatrics 1999; 103:686–693.
8. Weiss H, Quigley M, Hayes R. Male circumcision and risk of HIV infection in sub-Saharan Africa: a systematic review and meta-analysis. AIDS 2000; 14:2361–2370.
9. Siegfried N, Muller M, Volmink J, et al. Male circumcision for prevention of heterosexual acquisition of HIV in men. Cochrane Database Syst Rev 2003;CD003362.
10. Bailey CR, Plummer FA, Moses S. Male circumcision and HIV prevention: current knowledge and future research directions. Lancet Infect Dis 2001; 1:223–231.
11. Patterson BK, Landay A, Siegel JN, et al. Susceptibility to human immunodeficiency virus-1 infection of human foreskin and cervical tissue grown in explant culture. Am J Pathol 2002; 161:867–1873.
12. Reynolds SJ, Shepherd ME, Risbud AR, et al. Male circumcision and risk of HIV-1 and other sexually transmitted infections in India. Lancet 2004; 363:1039–1040.
13. Auvert B, Taljaard D, Lagarde E, et al. Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: the ANRS 1265 Trial. PLoS Med 2005; 2:e298.
14. Lavreys L, Rakwar JP, Thompson ML, et al. Effect of circumcision on incidence of HIV-1 and other sexually transmitted diseases: a prospective cohort study of trucking company employees in Kenya. J Infect Dis 1999; 180:330–336.
15. Dave SS, Johnson AM, Fenton KA, et al. Male circumcision in Britain: findings from a national probability sample survey. Sex Transm Infect 2003; 79:499–500.
16. Dickson N, van Roode T, Paul C. Herpes simplex virus type 2 status at age 26 is not related to early circumcision in a birth cohort. Sex Transm Dis 2005; 32:517–519.
17. Gray R, Azire J, Serwadda D, et al. Male circumcision and the risk of sexually transmitted infections and HIV in Rakai, Uganda. AIDS 2004; 18:2428–2430.
18. Weiss HA, Buve A, Robinson NJ, et al. The epidemiology of HSV-2 infection and its association with HIV infection in four urban African populations. AIDS 2001; 15 (Suppl 4):S97–S108.
19. Auvert B, Buve A, Legarde E, et al. Male circumcision and HIV infection in four cities in sub-Saharan Africa. AIDS 2001; 15 (Suppl 4): S31–S40.
20. National Center for Health Statistics. In Plan and Operation of the Third National Health and Nutrition Examination Survey, 1988–94. (Vital and Health Statistics Series 1, No. 32). Washington, DC: Government Printing Office; 1994. (DHHS Publication No. (PHS) 94–1308).
21. Lee FK, Coleman RM, Pereira L, et al. Detection of herpes simplex virus type-2-specific antibody with glycoprotein G. J Clin Microbiol 1985; 22:641–644.
23. Modadjer L, Montaquila J, Waksberg J, et al. National Health and Nutrition Examination Survey III: Weighting and Estimation Methodology. Hyattsville, MD: National Center for Health Statistics; 1996.
24. Elandt-Johnson RC, Johnson NL. Survival Models and Data Analyses. New York: Wiley; 1980:69–71.
25. Mansfield CJ, Hueston WJ, Rudy M. Neonatal circumcision: associated factors and length of hospital stay. J Fam Pract 1995; 41:370–376.
26. Owings MF, Kozak LJ. Ambulatory and inpatient procedures in the United States, 1996. National Center for Health Statistics. Vital Health Stat 13 1998; 139:1–119.
27. Xu F, Sternberg M, Kottiri B, et al. National trends in herpes simplex virus type 1 and type 2 in the United States: Data from the National Health and Nutrition Examination Survey (NHANES). JAMA 2006; 296:964–973.
28. Mertens TE. Estimating the effects of misclassification. Lancet 1993; 342:418–421.
29. Castellsague X, Bosch FX, Munoz N, et al. Male circumcision, penile human papillomavirus infection, and cervical cancer in female partners. N Engl J Med 2002; 346:1105–1112.
30. Risser JM, Risser WL, Eissa MA, et al. Self-assessment of circumcision status by adolescents. Am J Epidemiol 2004; 159:1095–1097.
31. Pertel PE, Spear PG. Biology of herpesviruses. In: Holmes KK, Sparling PF, Mardh P, et al, eds. Sexually Transmitted Diseases. 3rd ed. New York: McGraw-Hill; 1999:269–283.