Relations Between Circumcision Status,Sexually Transmitted Infection History,and HIV Serostatus Among a National Sample of Men Who Have Sex with Men in the United States

Abstract

Circumcision's potential link to HIV/sexually transmitted infections (STI) has been at the center of recent global public health debates. However, data related to circumcision and sexual health remain limited, with most research focused on heterosexual men. This study sought to assess behavioral differences among a large sample of circumcised and noncircumcised men who have sex with men (MSM) in the United States. Data were collected from 26,257 U.S. MSM through an online survey. Measures included circumcision status, health indicators, HIV/STI screening and diagnosis, sexual behaviors, and condom use. Bivariate and regression analyses were conducted to determine differences between HIV/STI status, sexual behaviors, and condom use among circumcised and noncircumcised men. Circumcision status did not significantly predict HIV testing (p > 0.05), or HIV serostatus (p > 0.05), and there were no significant differences based on circumcision status for most STI diagnosis [syphilis, gonorrhea, chlamydia, human papilloma virus (HPV)]. Being noncircumcised was predictive of herpes-2 diagnosis, however, condom use mediated this relationship. These data provide one of the first large national assessments of circumcision among MSM. While being noncircumcised did not increase the likelihood of HIV and most STI infections, results indicated that circumcision was associated with higher rates of condom use, suggesting that those who promote condoms among MSM may need to better understand condom-related behaviors and attitudes among noncircumcised men to enhance the extent to which they are willing to use condoms consistently.

Introduction

C ircumcision's potential link to HIV and sexually transmitted infection (STI) prevention has been at the center of recent global public health debates. Approximately one quarter of men worldwide have been circumcised for various reasons including religion, culture, medical reasons and/or parental/nonmedical preference,¹ yet researchers are divided in terms of gauging the potential benefits and risks of recommending the procedure as a mechanism to prevent HIV and STIs.^2
–4

Data from several studies have identified a higher prevalence of HIV among men who were not circumcised compared to those who were circumcised.^3

–6 In addition, researchers have found that circumcision acts as a protective factor against some STIs including syphilis and gonorrhea.^1,2 Halperin and Baily⁵ found that men who were circumcised were two to eight times less likely to become infected with HIV, while Weiss et al.⁶ found that among men in sub-Saharan Africa, circumcision was associated with a reduction in risk of risk of HIV infection.

However, although some researchers recommend circumcision as an HIV prevention strategy,⁶ others have been less confident. According to a meta-analysis conducted by Van Howe,⁴ of the 29 articles examined, circumcised men were at increased risk for both HIV acquisition and transmission compared to uncircumcised men, leading the author to suggest that routine circumcision as a means to prevent HIV infection would be unfounded. Similarly, Baily et al.³ reported no difference among circumcised and uncircumcised men with regard to STI histories and Cook and colleagues² reported mixed findings; in their study uncircumcised men had an increased likelihood of syphilis and gonorrhea infection yet had a decreased likelihood of genital warts, and no associations existed between circumcision status and genital herpes, chlamydia, or nongonococcal urethritis.

Among studies that have discussed the efficacy of circumcision in protecting against HIV transmission among men, most tend to focus on transmission via heterosexual sex.⁷ Vaginal intercourse may present differences with regard to disease transmission and circumcision status compared to anal intercourse due to the differences between vaginal and anal mucosa (J. Sanchez, unpublished data). Limited research exists highlighting circumcision as an HIV prevention technique among MSM. According to Millett and colleagues,⁸ there was no relationship among circumcision status and self-reports of STIs among a sample of 1,154 black and 1,091 Latino MSM nor did a meta-analysis of 15 studies, examining circumcision and HIV/STI rates, find any sufficient evidence to support findings that circumcision provides a protective effect against HIV/STI infection.⁹

While there seems to be inconclusive findings regarding the risks and benefits of circumcision, previous research tends to focus on STI status and/or diagnosis. Little research has examined the differences among circumcised and uncircumcised men with respect to sexual behavior. One study indicated that some behavioral characteristics such as engaging in sexual activity the last time they spent a night away from home, ever having used a condom, or having used a condom during last intercourse did not differ among circumcised and uncircumcised men.³ Furthermore, Buchbinder and colleagues¹⁰ found no relationship between unprotected insertive anal intercourse and circumcision, nor was circumcision associated with a reduction in HIV infection in men who engaged in protective receptive anal sex and unprotected insertive anal sex.

While MSM in the United States continue to be disproportionally impacted by HIV and STIs, data from large populations of men related to circumcision status and sexual health behaviors remains limited, inconclusive,^1,2,4 and most research available tends to be heterosexually focused.⁷ The current study sought to assess behavioral differences among a large, fairly representative sample of circumcised and noncircumcised MSM in the United States.

Methods

Participant recruitment and data collection

This study was conducted via a participatory effort with an Internet-based networking site for men who seek sexual or social interactions with other men. An electronic recruitment message was sent via e-mail to all individuals residing in the United States who had an account with the website at the time of the study (July 2009). The recruitment e-mail provided a brief description of the study, its incentives, and included a link to the study website. Previous research has found that Internet-based convenience samples are useful tools to better understand understudied issues and/or for collecting data from hard-to-reach populations, such as gay men who seek sex partners via the Internet.¹¹ Interested individuals proceeded to the study questionnaire; most participants took approximately 20 min to complete the questionnaire. Compensation for participation included the opportunity to receive an electronic coupon valued at $10 that could be redeemed for merchandise from an affiliate of the site from which men were recruited.

The recruitment message remained in each individual's electronic mailbox for a period of 7 days, after which time any unopened e-mails were automatically removed. A total of 127,489 individuals opened the recruitment email and 43,477 (34.1%) of these men clicked on the link to the study website. Of those viewing the study information and consent form, 26,257 (60.4%) consented to and subsequently participated in the study. All study protocols were reviewed and approved by Institutional Review Board at the academic institution of the authors.

Main outcome measures

Participants completed items related to sociodemographics, circumcision status, health status, health-related behaviors, and sexual behaviors.

Sociodemographics

Measures included those related to a participant's age, sexual orientation, race/ethnicity, and level of education completed. Participants also responded to items related to their relationship and current sexual relationship status.

Circumcision status

Measures included a single item, asking men to respond to the question, “Are you circumcised?” with one of the following response options: no, yes, or no response.

Health status

Measures included those related to physical and mental quality of life using the four-item Healthy Days Core Module from the Centers for Disease Control and Prevention (CDC) health-related quality of life measure (HRQOL-4),¹² a widely validated set of survey measures used to assess a person's sense of well-being.^13

–18 In addition, we assessed history of participation in testicular self-examinations, provider-administered (general physical examination) examinations, and HIV and other STI screenings and diagnoses (including HIV, chlamydia, gonorrhea, syphilis, HPV, or herpes within the past 2 years).

Sexual behaviors

Participants were asked to indicate whether or not they had engaged in specific sexual behaviors with men within the past month, past 3 months, past year, more than a year ago, or never. Behaviors included receptive and insertive anal intercourse, and performing and receiving oral sex. Sexual behaviors were summed across participants.

Condom behaviors

Participants were asked to indicate the number of times they used a condom during the past ten times they engaged in insertive anal intercourse, and the number of times their partner used a condom during the past ten times they engaged in receptive anal intercourse.¹⁹ Condom use variables were created for insertive behaviors, (top, when the participant used a condom during insertive sex) and receptive behaviors (bottom, when the participants' partner wore a condom during receptive anal intercourse). These items were treated as a continuous variables from using a condom 0 of 10 times to using a condom 10 of the 10 times during the past 10 sexual acts.

Data analyses

The overall sample was narrowed further to include (1) only men who indicated that they currently have at least one male sexual partner (eliminating 334, 1.3%, participants from the sample) and (2) men who responded to the item related to circumcision status (removing an additional 764, 2.9%, men from the sample). The final sample utilized for these analyses included 25,159 men. All analyses were conducted using PAWS version 17.0 (SPSS Inc., Chicago, IL). Descriptive analyses were conducted to present participant characteristics including demographics, health behaviors, and sexual behaviors. Bivariate analyses including independent sample t-tests, and a series of χ² and Mann-Whitney U tests were calculated to compare demographic differences between men who had been circumcised and men who had not. In addition, logistic regression analyses with adjusted odds ratios (adjusted for age, race, education, and relationship status) were conducted to determine if circumcision was a predictor of HIV/STI testing behaviors and HIV/STI serostatus. Multiple regression analyses were conducted to determine if circumcision status was a predictor of condom use behaviors. Baron and Kenny's²⁰ suggested methods were used for testing mediators between circumcision and herpes diagnosis. Sobel's test was also used to test for partial mediation.

Results

Participant characteristics

Of the 25,159 men who had participated in the current study, 21,312 (82.1%) indicated that they were circumcised whereas 3,847 (14.8%) indicated that they were not circumcised.

Table 1 provides a summary of the main demographics from the entire sample and a comparison of men based on circumcision status. For the entire sample, age was diverse, although most participants tended to be 30 years of age or older (n = 16,905; 69.0%), and circumcision rates were highest among Caucasian participants (n = 18472; 86.9%), followed by African Americans (n = 815; 77.5%) and Latinos (n = 896; 46.2%). Most participants had received some college education or higher (87.4%), identified as white (n = 21,295, 81.9%), and self-identified as homosexual/gay (n = 21,162; 81.4%). In terms of relationship status, a majority of men indicated that they were not currently dating anyone (n = 14,625; 56.2%), although most (n = 20, 900; 83.7%) were currently sexually active. Comparison of circumcised and noncircumcised men based on demographic characteristics, indicated that differences existed with regard to age, race, education, and relationship status.

Table 1.

Participant Characteristics and Circumcision Status ( n = 25,159)

			Circumcision status
	Total sample		Yes		No
Characteristics	n	%	21,097		3803		p
Age (n = 24,900)							<0.001
18–23	3560	14.2	2851	80.1	709	19.9
24–29	4435	17.6	3666	82.7	769	17.3
30–39	5762	22.9	4935	85.6	827	14.4
40–49	6970	27.7	6122	87.8	848	12.2
50–59	3318	13.2	2868	86.4	450	13.6
60 +	855	3.4	655	76.6	200	23.4
Education (n = 25,104)							0.02
High school or less	2817	11.3	2316	82.2	501	17.8
Some college or associates degree	8485	33.7	7157	84.3	1328	15.7
Bachelors degree	8112	32.2	6976	86.0	1136	14.0
Masters degree	3826	15.2	3245	84.8	581	15.2
Professional	1625	6.5	1374	84.6	251	15.4
Other	239	0.9	199	83.3	40	16.7
Race/ethnicity (n = 25,028)							<0.001
Black	1014	4.0	815	80.4	199	19.6
White	20650	82.1	18472	89.5	2178	10.5
Hispanic/Latino	1857	7.4	896	48.2	961	51.8
Asian/pacific islander	729	2.9	426	58.4	303	41.6
Other	777	3.1	599	77.1	178	22.9
Sexual orientation (n = 25,112)							0.087
Homosexual/gay	20542	81.6	17435	84.9	3107	15.1
Bisexual	3927	15.6	3317	84.5	610	15.5
Heterosexual/straight	126	0.5	115	91.3	11	8.7
Relationship status (n = 25,099)							<0.001
In a relationship: 5 years	4767	18.9	4153	87.1	614	12.7
In a relationship: 1–5 years	2658	10.6	2252	84.7	406	15.3
In a relationship: 1 year or less	2044	8.1	1672	81.8	352	18.2
Currently dating more than one person	1439	5.7	1209	84.0	230	16.0
Not currently dating anyone	14105	56.1	11892	84.3	2213	15.7
Current sexual relationship (n = 24,999)							0.08
In a monogamous sexual relationship	4643	18.5	3923	84.5	720	15.5
Sexual relationships with more than one person	6025	23.9	5120	85.0	905	15.0
Sexually active but not in a sexual relationship	10232	40.7	8757	85.6	1475	14.4
Not sexually active with another person	4099	16.3	3381	82.5	718	17.5

Circumcision status and general health

The sample was healthy overall, with the majority of men indicating their general health to be very good to excellent (71.2%). Overall, based upon the CDC (CDC, 2003) guidelines¹⁹ which designate “good” health as under 14 unhealthy days per month, participants reported a relatively low number of physically, mentally or activity limited days (M = 1.73–4.46, SD = 4.55–7.35). Using the HRQOL health outcomes including general health, physically unhealthy days (t(24,260) =0.75), mentally unhealthy days (t(24,374) = −0.098) and activity limited days (t(24,238) = 0.65) there were no significant differences based on quality of life on any of the measures for those individuals who were circumcised compared to those who were not (p > 0.05).

When looking at general assessments of non-disease related health issues, there were no differences based on circumcision status for getting a physical exam (χ²(25,113) = 0.175, p > 0.05), however, groups were significantly different based on circumcision status for conducting a testicular self-examination (χ²(25,103) = 41.345, p < 0.001), with noncircumcised men having been more likely to conduct a testicular self-examination in the last month.

Sexual behaviors and condom use

Circumcision status was predictive of men's engagement in sexual behavior in the past month (p < 0.001) with circumcised men being more likely to have engaged in sexual behavior (oral and anal sex) within the past month, compared to uncircumcised men (odds ratio [OR] = 0.882, p < 0.001).

At the bivariate level, there were no differences based on circumcision status for condom use during insertive anal sex (t(17,554) = −0.897, p > 0.05) and partner's use of a condom during receptive anal sex (t(16,299) = −1.908, p > 0.05). However, when circumcision was entered into a multivariate model, controlling for age, race, relationship status, and education level, circumcision status was predictive of men's condom use behaviors during both insertive and receptive anal sex (p < 0.05).

Upon further exploration of condom use behaviors, results indicate that about one fifth of men used a condom 10 of 10 times they engaged in anal sex (insertive, 23.3%; receptive, 23.9%), %), while one third reported using a condom zero 0 of 10 times or never using a condom (32.2%; receptive, 33.9%) for both insertive and receptive sexual behaviors. Nearly one half of the men in this sample fell into the range of using a condom 1 of 10 times to 9 of 10 times out of the past 10 acts of anal sex (insertive, 44.5%; receptive, 42.2%).

STI testing and diagnosis

Participant's history of getting tested for HIV did not significantly differ based on circumcision status, however, group differences did exist for STI testing (p < 0.01) in the last year, with men who were circumcised reporting higher rates of STI testing. Among those participants who reported having previously been diagnosed with any STI or HIV, no differences existed with the exception of herpes, which was more likely in men who were not circumcised (Table 2).

Table 2.

χ² and Mann-Whitney U

	Circumcision status
	Yes		No
Health variable	n	%	n	%	χ²
	21,312	84.7	3847	15.3
Physical examination in last year	15,531	73.0	2784	72.7	0.175
Testicular examination in last month	9412	44.2	1909	49.9	41.345^a
STD test in last year	9224	43.5	1575	41.2	7.127^b
Positive syphilis diagnosis (2 year)	936	4.4	170	4.5	0.028
Positive gonorrhea diagnosis (2 year)	1009	4.8	193	5.1	0.713
Chlamydia diagnosis (2 year)	886	4.2	165	4.4	0.220
HPV (2 year)	911	4.4	153	4.1	0.555
Herpes (2 year)	878	4.2	187	5.0	4.987^a
HIV test	18697	88.3	3330	87.2	3.348
HIV positive	3012	14.3	543	14.3	0.009
Sexual behavior (any oral, or anal sex) in past month	12300	59.1	2094	56.3	10.983^b

	Circumcision status
	Yes		No
Condom Use & Health Variables	n	%	n	%	Mann-Whitney U
Health status					−1.081 (ns)
Excellent	5990	28.1	1125	29.2
Very good	9653	45.3	1706	44.3
Good	4830	22.7	873	22.7
Fair	745	3.5	127	3.3
Poor	79	0.4	13	0.3

STD, sexually transmitted disease; HPV, human papilloma virus; ns, not significant.

p < 0.01 level; ^b p < 0.05 level.

Logistic regressions were conducted to determine if circumcision status was predictive of having been diagnosed with an STI or HIV in the past 2 years (Table 3), when controlling for age, racial status, education level, and relationship status. Results indicate that circumcision status was not predictive of HIV/STI diagnoses, exclusive of herpes. Because group differences at the bivariate level existed for STI testing and at the multivariate level for condom use, these variables were entered in the logistic regression model as potential mediators using Baron and Kenny's four-step test of mediation.²¹ Both STI testing and condom use met Baron and Kenny's first three criteria. When STI testing was entered into the model, the relationship from circumcision to herpes was significantly reduced, indicating partial mediation (p < 0.01).

Table 3.

Logistic Regression (Control for Race/Ethnicity, Age, and Education)

Outcome variable	B	SE	ExpB	Wald statistic
Positive syphilis diagnosis (2 year)	0.001	0.087	1.001	0.000
Positive gonorrhea diagnosis (2 year)	−0.077	0.082	0.926	0.868
Chlamydia diagnosis (2 year)	−0.05	0.088	0.951	0.325
HPV (2 year)	0.086	0.091	1.090	0.893
Herpes (2 year)	−0.186	0.084	0.831	4.875^b
HIV test	0.062	0.055	1.064	1.292
HIV positive	−0.041	0.052	0.959	0.640
Herpes with condom use mediator	−0.168	0.086	0.845	3.791
Condom use (top)	−0.196	0.053	0.822	13.485^a
Condom use (bottom)	−0.133	0.054	0.876	5.974^b
Sexual behavior (any oral or anal sex) in past month	−0.126	0.037	0.882	12.272^a

p < 0.01 level.

p < 0.05 level.

HPV, human papilloma virus.

Additionally, when condom use was entered into the model, it emerged as a significant predictor of herpes diagnosis (p < 0.05) while circumcision status dropped out of the model (p > 0.05), indicating full mediation.

Discussion

This study provides data of one of the first, large national assessments of associations between circumcision status, sexual behaviors, and STI screening and diagnosis among MSM in the United States. Study participants were fairly representative of national rates of the MSM population which are weighted toward metropolitan areas. Consistent with previous studies,^3,4 results from this study suggest that circumcision status (being circumcised) shows no protective factors for most STIs (chlamydia, HPV, syphilis) or HIV, as diagnoses are not significantly different across groups. Such findings support claims made by the American Pediatric Association and other organizations, that there is not enough substantial evidence to indicate the necessity of recommending neonatal circumcision as a preventive procedure to protect against future STI or HIV infection.¹⁸

Unlike previous research, the current study examined sexual and condom-use behaviors among MSM based on circumcision status. Most of the men in this sample, regardless of circumcision status were currently sexually active with a partner, however, circumcised men were more likely to engage in a higher frequency of sexual behavior (any oral or anal sex) in the past month or three months. Furthermore, 65.1% of circumcised men indicated they were currently having sex with more than one person, potentially increasing their likelihood for STI exposure.

In addition, being circumcised was associated with higher rates of never using a condom whereas not being circumcised was associated with higher rates of always using a condom. These findings suggest that those who promote condoms among MSM may need to better understand condom-related behaviors and attitudes among both circumcised and noncircumcised men to enhance the extent to which all MSM are willing to use condoms consistently. Future research should potentially explore better understanding this discrepancy in condom use behavior based on circumcision status.

Similar to findings of condom use among heterosexual men,¹⁹ for both circumcised and uncircumcised men, condom use behaviors were inconsistent as just over 20% (23.6%) reported always using a condom for insertive and receptive behaviors and approximately one third (33.1%) reported never using a condom. In addition, 43.4% of men reported using a condom between 1 and 9 times out of the past 10 instances of intercourse, indicating a range of inconsistency in condom use behaviors. While this measure of condom use may not be optimal, as it is quantified in such a wide range, it does provide an important insight for researchers studying condom use. Previous research regarding condom use often asks participants about condom use in terms of a dichotomous (yes/no) variable. Given that nearly half the sample reported condom use somewhere within the range of 1 to 9 of the previous 10 sexual encounters, measuring condom use with a dichotomous (yes/no) variable may not be appropriate in all situations especially if the intent is to asses individual HIV/STI risk on a longer term basis (e.g., to assess condom use in the last week). In addition, a “yes” response to a question about whether or not a condom was used in a sexual encounter may include individuals who used a condom for part of the time, but did not use it the entire time penetration was occurring. Researchers and clinicians should make efforts to be more accurate in terms of the realization and conceptualization of the variation in this behavior. More clearly worded questions, taking into account contextual and situational factors, regarding condom use should be considered in order to gain the most accurate information from research participants or patients about their condom use behaviors.

While this study represents an exploration of health behaviors and STI diagnoses based on circumcision status among a large sample of MSM from all states in the United States, it remains limited by the extent to which the sample is one of convenience, as challenges remain with establishing true nationally representative samples of MSM for research. Furthermore, data were collected on the Internet and only from members of a single Internet-based site, yet this method of recruitment was purposive and closely resembles national representations of MSM in the United States.

In addition, because this study was not specifically designed to examine circumcision, future large scale research should take a more in-depth approach to understanding important issues associated with circumcision and condom use, including pleasure and satisfaction.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Moses

, Bailey

, Ronald

. Male circumcision: Assessment of health benefits and risks. Sex Transm Infect, 1998; 74:368–373.

Cook

, Koutsky

, Holmes

. Circumcision and sexually transmitted diseases. Am J Public Health, 1994; 84:197–201.

Bailey

, Neema

, Othieno

. Sexual behaviors and other HIV risk factors in circumcised and uncircumcised men in Uganda. J Acquir Immune Defic Syndrome, 1999; 22:294–301.

Van Howe

. Circumcision and HIV infection: A review of the literature and meta-analysis. Int J Sex Transm Dis AIDS, 1999; 10:8–16.

Halperin

, Bailey

. Male circumcision and HIV infection: 10 years and counting. Lancet, 1999; 354:1813–1815.

Weiss

, Quigley

, Hayes

. Male circumcision and risk of HIV infection in sub-Saharan Africa: A systematic review and meta-analysis. AIDS, 2000; 14:2361–2370.

Sulliva , Kilmarx , Peterman

et al. Male circumcision for prevention of HIV transmission: What the new data mean for HIV prevention in the United States. PLOS Medicine, 2007; 4:1162–1166.

Millett

, Ding

, Lauby

et al. Circumcision status and HIV infection among black and Latino men who have sex with men in 3 US cities. AIDS, 2007; 46:643–650.

Millett

, Flores

, Marks

et al. Circumcision status and risk of HIV and sexually transmitted infections among men who have sex with men: A meta- analysis. JAMA, 2008; 14:1674–1684.

10.

Buchbinder

, Vittinghoff

, Heagert

et al. Sexual Risk, nitrite, inhalant use, and lack of circumcision associated with HIV seoconversions in men who have sex with men. AIDS, 2005; 39:82–89.

11.

Carballo-Dieguez

, Miner

, Dolezal

, Rosser

, Jacoby

. Sexual negotiation, HIV-status disclosure, and sexual risk behavior among Latino men who use the Internet to seek sex with other men. Arch Sex Behav, 2006; 5:473–481.

12.

Centers for Disease Control and Prevention. Measuring healthy days. Atlanta, GA: Centers for Disease Control and Prevention, 2000.

13.

Centers for Disease and Control Prevention. Health-related quality of life among persons with epilepsy—Texas 1998. MMWR, 2001; 50:24–26.

14.

Centers for Disease Control and Prevention. State differences in reported Healthy Days among adults—United States, 1993–1996. MMWR, 1998; 47:239–244.

15.

Centers for Disease Control and Prevention. Community indicators of health- related quality of life—United States, 1993–1997. MMWR, 2000; 49:281–285.

16.

Andresen

, Catlin

, Wyrwich

, Jackson-Thompson

. Retest reliability of surveillance questions on health related quality of life. J Epidemiol Community Health, 2003; 57:339–343.

17.

Moriarty

, Zack

, Kobau

. The Centers for Disease Control and Prevention's Healthy Days Measures—Population tracking of perceived physical and mental health over time. Health Qual Life Outcomes, 2003; 1:37.

18.

American Academy of Pediatrics. Circumcision policy statement. Pediatrics, 1999; 103:686–693.

19.

Noar

, Cole

, Carlyle

. Condom use measurement in 56 studies of sexual risk behavior: Review and recommendations. Arch Sex Behav, 2006; 35:327–345.

20.

Baron

, Kenny

. The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. J Pers Soc Psychol, 1986; 51:1173–1182.