MAJOR ARTICLE
Male Circumcision Decreases Acquisition and Increases Clearance of High-Risk Human Papillomavirus in HIV-Negative Men: A Randomized Trial in Rakai, Uganda Ronald H. Gray,1,7 David Serwadda,7,8 Xiangrong Kong,1 Frederick Makumbi,7,8 Godfrey Kigozi,7 Patti E. Gravitt,2 Stephen Watya,9 Fred Nalugoda,7 Victor Ssempijja,7 Aaron A. R. Tobian,4 Noah Kiwanuka,7,8 Lawrence H. Moulton,3 Nelson K. Sewankambo,7,10 Steven J. Reynolds,5,6 Thomas C. Quinn,5,6 Boaz Iga4,7, Oliver Laeyendecker,5,6 Amy E. Oliver,5 and Maria J. Wawer1,7 1
Department of Population, Family, and Reproductive Health, 2Department of Epidemiology, 3Department of International Health, Johns Hopkins University Bloomberg School of Public Health, 4Department of Pathology, 5Department of Medicine, Johns Hopkins Medical Institutions, Baltimore, and 6National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland; 7Rakai Health Sciences Program, Kalisizo, Rakai District, and 8School of Public Health, 9Urology Unit, Department of Surgery, Mulago Hospital, and 10Faculty of Medicine, College of Health Sciences, Makerere University, Kampala, Uganda
(See the article by Serwadda et al, on pages 1463–1469, and the editorial commentary by Viscidi and Shah, on pages 1447–1449.)
Methods. Uncircumcised human immunodeficiency virus (HIV)–negative men aged 15–49 years were randomized to immediate circumcision (intervention arm, 441 subjects) or delayed circumcision (control arm, 399 subjects). Human papillomavirus (HPV) was detected by Roche HPV Linear Array at enrollment, and at 6, 12, and 24 months. Incident high-risk HPV (HR-HPV) was estimated in men who acquired a new HR-HPV genotype. HR-HPV clearance was determined in men with prior genotype-specific HR-HPV infections. Rate ratios (RRs) and 95% confidence intervals (CIs) of HR-HPV acquisition were estimated by Poisson multiple regression. Results. Enrollment characteristics were comparable between study groups. HR-HPV incidence was 19.7 cases per 100 person-years (PYs) in the intervention arm (70 cases per 355.8 PYs) and 29.4 cases per 100 PYs (125 cases per 424.8 PYs) in the control arm (RR, 0.67; 95% CI, 0.51–0.89; P p .006 ). The incidence of multiple HRHPV infections was 6.7 cases per 100 PYs in the intervention arm and 14.8 cases per 100 PYs in the control arm (RR, 0.45; 95% CI, 0.28–0.73), but there was no significant effect on single infections (RR, 0.89; 95% CI, 0.60– 1.30). HR-HPV incidence was lower in the intervention arm for all genotypes and demographic/behavioral subgroups. The clearance of preexisting HR-HPV infections was 215.8 cases per 100 PYs (205 cases per 95 PYs) in the intervention arm and 159.1 cases per 100 PYs (255 cases per 160.25 PYs) in the control arm (adjusted RR, 1.39; 95% CI, 1.17–1.64). Conclusions. Male circumcision reduces the incidence of multiple HR-HPV infections and increases clearance of HR-HPV infections in HIV-uninfected men. Trial Registration. ClinicalTrials.gov identifier: NCT00425984. Two trials have shown that male circumcision reduces the prevalence of high-risk human papillomavirus (HR-
Received 5 October 2009; accepted 4 December 2009; electronically published 6 April 2010. Address for correspondence: Dr Ronald H. Gray, Johns Hopkins University Bloomberg School of Public Health, 615 N Wolfe St, Ste E4132, Baltimore, MD 21205 (
[email protected]). The Journal of Infectious Diseases 2010; 201(10):1455–1462 2010 by the Infectious Diseases Society of America. All rights reserved. 0022-1899/2010/20110-0004$15.00 DOI: 10.1086/652184
218 • Gray et al
HPV) infection by ∼35% in human immunodeficiency virus (HIV)–negative men [1, 2], and several crosssectional studies have found a lower rate of prevalent penile HPV infection in circumcised HIV-uninfected men [3–7]. However, it is unclear whether circumcision reduces the incidence of HPV acquisition or Potential conflicts of interest: none reported. Financial support: The study was primarily supported by the National Institutes of Allergy and Infectious Disease (NIAID), Division of AIDS, National Institutes of Health (NIH) (grant UO1 AI11171-01-02) and in part by the Division of Intramural Research, NIAID, NIH. This publication was supported in part by a fellowship from the Fogarty International Center/USNIH (grant 2D43 TW000010-19-AITRP).
increases the clearance of preexisting HPV infection. One longitudinal observational study of 285 US men, 25 of whom were circumcised, found no statistically significant difference in the hazards ratio (HR) of acquisition of either HR-HPV infections (adjusted HR, 1.7; 95% CI, 0.6–4.9) or any HPV infection (adjusted HR, 0.8; 95% CI, 0.4–1.9) in circumcised versus uncircumcised men [7]. However, the authors reported a significantly faster rate of HR-HPV clearance (adjusted HR, 6.5; 95% CI, 2.1–19.7) [7]. We used data from a randomized trial of male circumcision conducted in Uganda to assess whether circumcision affected the acquisition of new HR-HPV infections and clearance of preexisting infections. METHODS We conducted a trial of male circumcision for prevention of HIV acquisition in initially HIV-uninfected men in Rakai District, Uganda, during the period 2003–2006 [2, 8]. The effect of circumcision on HPV was a secondary trial end point. The trial design has been previously described [2, 8]. In brief, we enrolled uncircumcised HIV-negative men 15–49 years of age who were randomized to receive either immediate circumcision (intervention arm) or circumcision delayed for 24 months (control arm). At enrollment, and at 6, 12, and 24 months follow-up, trial participants provided information on socio-
demographic characteristics, sexual risk behaviors, and symptoms suggestive of sexually transmitted infections. Penile swab samples were collected at each study visit for HPV detection. Moistened Dacron swab samples were obtained from the coronal sulcus and glans using a standard protocol, placed in Digene specimen transport medium, and stored at ⫺80C. HPV genotyping was performed using the Roche HPV Linear Array (Roche Diagnostics), as described elsewhere [9– 11]. HPV genotypes 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68 were considered the primary HR-HPV, or carcinogenic, viral genotypes. Genotypes 6, 11, 26, 40, 42, 43, 53, 54, 55, 61, 67, 70, 71, 72, 73, 81, 82, 83, 84, and 108 were considered low-risk HPV (LR-HPV) genotypes. Penile swab samples without amplifiable cellular or viral DNA were deemed insufficient for HR-HPV detection [12, 13]. Informative samples with amplifiable DNA were defined as swab samples with detectable b-globin, indicating the presence of cellular DNA, or detectable HPV, indicating sufficient viral DNA levels were present, even in the absence of detectable b-globin. The study was approved by 4 institutional review boards: the Science and Ethics Committee of the Uganda Virus Research Institute, the Uganda National Council for Science and Technology, the Committee for Human Research at Johns Hopkins University, Bloomberg School of Public Health, and the West-
Figure 1. Trial profile. High-Risk HPV in HIV-Negative Men • 219
Table 1. Participant Characteristics at Enrollment No. (%) of participants, by study arm Intervention (n p 441)
Control (n p 399)
Age, years 15–24
93 (21.1)
92 (23.1)
25–29 30–35
135 (30.6) 100 (22.7)
108 (27.1) 89 (22.3)
135 Education
113 (25.6)
110 (27.6)
None Primary
38 (8.6) 311 (70.5)
37 (9.3) 280 (70.2)
92 (20.9)
82 (20.6)
No use Use Use of alcohol with sex No
284 (64.4) 157 (35.6)
259 (64.9) 140 (35.1)
200 (45.4)
164 (41.1)
Yes No. of sex partners
241 (54.7)
235 (58.9)
1 partner Multiple partners
253 (57.4) 188 (42.6)
233 (58.4) 166 (41.6)
151 (39.1)
133 (38.6)
Characteristics
.66
.94
Secondary or postCondom use
.93
.24
.82
a
Prevalent HR-HPV at enrollment a
P
.17
Among participants assayed at enrollment.
ern Institutional Review Board. The trial was registered with ClinicalTrials.gov number NCT00425984 The trial profile is given in Figure 1. The population assessed in this analysis of HR-HPV incidence/clearance was restricted to HIV-negative married men because we planned to assess HPV infections in their female partners; given assay costs and limited resources, this was the most efficient sampling design to address effects in men and women. We also excluded men with HIV-infected wives because of high rates of HPV infection in HIV-positive women [14, 15]. In addition, we excluded men whose wives were not concurrently enrolled with their husbands, because we would not be able to rigorously study the efficacy of circumcision for prevention of HPV infection in women, unless these female partners’ enrollment status was known at the time of their husbands’ enrollment. We then randomly selected married men with concurrently enrolled HIV-negative wives who provided samples obtained at enrollment and at the 24-month follow-up to determine the effects of circumcision on prevalent HPV infection, as reported elsewhere [2]. There were 835 HIV-negative men in the intervention arm with concurrently enrolled, HIV-negative wives. Because of assay cost and resource limitation, we randomly selected 441 (52.8%) of these male participants who provided a penile swab sample at enrollment. Of these men, 7 did not accept circumcision by the time of the 6-month visit and, per protocol, were classified as cross-overs at that time. In the con220 • Gray et al
trol arm, there were 803 HIV-negative men with concurrently enrolled, HIV-negative wives, of whom we randomly selected 399 (49.7%) who provided a penile swab sample at enrollment. Three control participants were circumcised between the 1- and 2-year follow-up visits and were classified as cross-overs at that interval. We conducted HPV assays on randomly selected penile samples at the 6- and 12-month follow-up visits to further reduce assay costs. There were 330 randomly selected, penile samples assayed in the intervention arm at 6 and 12 months. In the control arm, we assayed 321 randomly selected samples at 6 months and 314 samples at 12 months. Because our original objective was to determine HPV prevalence at 24 months [2], we assayed all available samples at that time point (440 in the intervention arm and 399 in the control arm). Statistical analyses. We assessed participant characteristics at enrollment by study arm. We then determined the proportion of penile swab samples with amplifiable cellular or viral DNA by study visit to assess whether detection varied over time between study arms. An intention-to-treat analysis was then used to assess the efficacy of male circumcision for prevention of new HR-HPV infections among participants with amplifiable cellular or viral DNA detected at sequential study visits. An astreated analysis classified the 7 intervention arm cross-overs as uncircumcised if they did not receive surgery by the 6-month visit, and 3 control subjects who received circumcision from nonproject sources between the 12- and 24-month visits were classified as circumcised for the last follow-up interval. Acquisition of HR-HPV was defined as a new infection identified in men who were initially negative for any HR-HPV and who acquired 1 or ⭓2 new HR-HPV infection(s) before the next follow-up, or men whose tests were initially positive for a specific HR-HPV genotype but acquired ⭓1 new HR-HPV genotype(s) before the next follow-up. HR-HPV incidence rates per 100 person-years (PYs) were estimated assuming that the new HR-HPV infection was acquired at the midpoint of the sequential follow-up interval at which the new infection was detected. Similar analyses were conducted for the LR-HPV genotypes. The unit of observation was an individual participant, and each man with an incident HR-HPV infection was counted only once per follow-up interval, irrespective of whether he acquired a single HR-HPV genotype or multiple HR-HPV genotypes. We separately assessed the incidence of single and multiple (⭓2) new HR-HPV genotype-specific infections, again using the individual participant as the unit of observation. We also assessed the incidence of each HR-HPV genotype; for each genotype, the population at risk was men without that genotype at a prior study visit, irrespective of the coinfection with other HR-HPV genotypes. Acquisition of HR-HPV infection was examined by study arm and stratified by sociodemographic and behavioral covariates at enrollment. We estimated the incidence rate ratio (IRR) and 95% confidence intervals (CIs) of HR-
Table 2. High-Risk Human Papillomavirus (HR-HPV) Incidence by Study Arm and Follow-up Interval for Participants with Amplifiable Cellular or Viral DNA at Sequential Study Visits Control arm
Intervention arm Infections Any HR-HPV (at least 1 new infection) 0–12 12–24 a 0–24 Single new HR-HPV Infection 0–12 12–24 0–24a Multiple (11) new HR-HPV Infections 0–12 12–24 0–24a
HRHPV/PYs
HRHPV/100 PYs
HRHPV/PYs
HRHPV/100 PYs
Intervention/control incidence rate ratio (95% CI)
57/210.25 21/145.5
27.1 14.4
100/226.25 45/198.5
44.2 22.7
0.61 (0.44–0.85) 0.64 (0.38–1.07)
70/355.75
19.7
125/424.75
29.4
0.67 (0.50–0.90)
42/210.25
20.0
61/226.25
27.0
0.74 (0.50–1.10)
15/145.5
10.3
21/198.5
10.6
0.97 (0.50–1.89)
46/355.75
12.9
62/424.75
15.6
0.89 (0.60–1.30)
15/210.25 6/145.5
7.1 4.1
39/226.25 24/198.5
17.2 12.1
0.41 (0.23–0.75) 0.34 (0.14–0.83)
24/355.75
6.7
63/424.75
14.8
0.45 (0.28–0.73)
NOTE. CI, confidence interval; PY, person-year. a
The unit of observation was each participant, so the sum of the no. of men with incident infections over the intervals exceeds the no. of men with incident infection during the 0–24 month observation time, because individuals who acquired new genotypes in both the 0–12and 12–24-month intervals are counted only once in the 0–24-month interval.
HPV acquisition in intervention arm versus control arm cumulatively over 12 and 24 months using a Poisson log linear regression model with offset. Multivariate Poisson models were used to adjust for covariates found to be associated with incident HR-HPV at P ! .20 in univariate analyses. Clearance of HR-HPV was estimated among men with preexisting HR-HPV infections who had sequential samples with amplifiable viral or cellular DNA. Clearance was expressed as the proportion of men with preexisting HR-HPV whose tests were negative for that genotype at a subsequent sequential study Table 3.
visit, and as the rate of clearance per 100 PYs observation subsequent to detection of a HR-HPV infection. Clearance was assessed for each HR-HPV genotype, irrespective of the number of HR-HPV infections per individual, and genotype-specific clearance rates were summed to provide global estimates. The rate ratio of clearance was estimated using a log-linear model based on an underdispersed Poisson distribution to account for multiple clearance events within individuals. Potential confounders were examined in univariate analyses and covariates found to be associated at a ! 0.20, or suspected confounders
Type-Specific High-Risk Human Papillomavirus (HR-HPV) Incidence over 24 Months, by Study Arm
Incidence/100 PYs
No./PYs
Incidence/100 PYs
Intervention/control incidence rate ratio (95% CI)
3.6 1.6
21/438.5 24/451.3
4.8 5.3
0.75 (0.38–1.51) 0.30 (0.12–0.75)
Control arm
Intervention arm HR-HPV type 16 18
No./PYs 13/360 6/370
31 33
6/372.5 2/377.5
1.6 0.5
10/462.5 14/456.5
2.2 3.1
0.74 (0.27–2.05) 0.17 (0.04–0.76)
35 39
7/371.75 3/377.25
1.9 0.8
17/452 7/460.5
3.7 1.5
0.50 (0.21–1.21) 0.52 (0.14–2.02)
45 51
6/368.25 14/350.5
1.6 4.0
11/458 23/435.8
2.4 5.3
0.68 (0.25–1.83) 0.76 (0.39–1.47)
52
6/369.5
1.6
16/450.5
3.6
0.46 (0.18–1.17)
56 58 59
4/376.5 10/364.8 10/367.3
1.1 2.7 2.7
8/465.3 22/443 22/447.5
1.7 5.0 4.9
0.62 (0.19–2.05) 0.55 (0.26–1.17) 0.55 (0.26–1.9)
66 68
12/362.8 8/369.3
3.3 2.2
25/445.3 19/452
5.6 4.2
0.59 (0.30–1.17) 0.52 (0.23–1.18)
NOTE. CI, confidence interval; PY, person-year.
High-Risk HPV in HIV-Negative Men • 221
Table 4. The Incidence of High-Risk Human Papillomavirus (HR-HPV) over 24 Months Stratified by Sociodemographic and Behavioral Covariates Control arm
Intervention arm Participant characteristics and behaviors
Incidence of HR-HPV/PYs
Incidence per 100 PYs
Incidence of HR-HPV/PYs
Incidence per 100 PYs
Intervention/control incidence rate ratio of HR-HPV (95% CI)
Age, years 15–24
15/72.8
20.6
34/98.8
34.4
0.60 (0.33–1.10)
25–29 30–35
22/110 16/80
20.0 20.0
40/121 24/92.3
33.1 26.0
0.60 (0.36–1.02) 0.77 (0.41–1.45)
35+ Education
17/93
18.3
27/112.8
23.9
0.76 (0.42–1.40)
None Primary Secondary and higher Condom use
4/27.5 50/264.8 16/63.5
14.5 18.9 25.2
7/31.8 93/303 25/90
22.0 30.7 27.8
0.66 (0.19–2.25) 0.62 (0.44–0.87) 0.91 (0.48–1.70)
No use Use Alcohol use with sex No
44/237.8 26/118
18.5 22.0
71/258 54/166.8
27.5 32.4
0.67 (0.46–0.98) 0.68 (0.43–1.09)
34/163.8
20.8
56/174.5
32.1
0.65 (0.42–0.99)
Yes No. of sex partners during the last 12 months 1 partner Multiple partners
36/192
18.8
69/250.3
27.6
0.68 (0.45–1.02)
42/207.5 28/148.3
20.2 18.9
64/245 61/179.8
26.1 33.9
0.77 (0.53–1.14) 0.56 (0.36–0.87)
NOTE. CI, confidence interval; PY, person-year.
(on the basis of biological reasoning or prior studies) were included in multivariate analyses of men with consecutive follow-up samples. Analyses were performed using Stata software (version 8.0; Stata) and SAS software (version 9.2; SAS Institute). RESULTS The intervention and control arms were comparable with respect to sociodemographic and behavioral characteristics at enrollment (Table 1). The prevalence of HR-HPV at enrollment was 39.1% in the intervention arm and 38.6% in the control arm (P p .17). At enrollment, and prior to circumcision, the proportions of samples with amplifiable cellular DNA were 344 (78.0%) of 441 men in the intervention arm and 306 (76.7%) of 399 men in the control arm. Among the intervention arm, the proportion of samples with amplifiable cellular DNA decreased over time to 208 (63.0%) of 330 men at 12 months and 240 (60.8%) of 395 men at 24 months (P ! .001). However, there was no decrease in the proportion of samples with amplifiable cellular DNA among uncircumcised control participants at 12 months (241 [76.8%] of 314 men) or 24 months (272 [75.3%] of 361 men), and this difference in amplifiable cellular data between study arms at 24 months was statistically significant (P ! .001). The incidence of new HR-HPV infections or genotypes was determined for samples with amplifiable cellular or viral DNA 222 • Gray et al
at sequential study visits (Table 2). In an intention-to-treat analysis at the 24-month follow-up, 70 men in the intervention arm acquired ⭓1 new HR-HPV infection (incidence, 19.7 cases per 100 PYs), whereas 125 men in the control arm acquired ⭓1 HR-HPV infections, with an incidence rate of 29.4 cases per 100 PYs (IRR, 0.67; 95% CI, 0.50–0.90). During the first year of follow-up, the incidence rate was significantly lower in the intervention group (IRR, 0.61; 95% CI, 0.44–0.85), whereas during the second year of follow-up, the effect was of borderline statistical significance (IRR, 0.64; 95% CI, 0.38–1.07). HR-HPV incidence decreased over time in both study arms, partly because participants with higher risk sexual behaviors acquired infections in the first follow-up year and thus could not contribute to incident infections with the same genotype during the second follow-up year. There were no significant differences in HR-HPV incidence between study arms among men who only acquired a single HR-HPV infection (IRR, 0.89; 95% CI, 0.60–1.30). However, the acquisition of multiple HR-HPV infections over 24 months was markedly lower among men in the intervention arm (6.7 cases per 100 PYs) than for men in the control arm (14.8 cases per 100 PYs) with an IRR of 0.45 (95% CI, 0.28–0.73). In an as-treated analysis, the incidence of any HR-HPV infection was 20.6 cases per 100 PYs (72 cases per 350 PYs) in the intervention arm and 28.6 cases per 100 PYs (123 cases per 430.5 PYs) in the control arm, with an IRR of 0.72 (95% CI, 0.54–0.96). The as-treated incidence of mul-
Table 5. Clearance of Genotype-Specific Preexisting High-Risk Human Papillomavirus (HR-HPV) Infection, by Study Arm Clearance rate over 100 PYs, by study arm HR-HPV type 16 18 31 33
Intervention 21/11 15/5.3 16/4.016 7/2
Control
(190.9) (285.7) (400.0) (350.0)
31/18.5 15/9.5 12/5.0 16/7.8
(167.6) (157.9) (240.0) (206.5)
Intervention/control rate ratio of clearance (95% CI) 1.14 1.81 1.67 1.70
(0.65–1.98) (0.88–3.70) (0.79–3.52) (0.70–4.12)
35 39
9/5.75 (156.5) 6/1.75 (342.9)
18/11.5 (156.5) 9/8.8 (102.9)
1.00 (0.45–2.23) 3.33 (1.19–9.36)
45 51
17/7.0 (242.9) 35/17 (205.9)
20/7 (285.7) 27/22 (122.7)
0.85 (0.45–1.62) 1.68 (1.02–2.77)
52 56 58 59
13/7.0 4/3.0 19/7.8 15/7.0
66 68 Total
(185.7) (133.3) (245.2) (214.3)
19/13.8 9/4.5 21/16.8 19/12.5
(138.2) (200.0) (125.4) (152.0)
1.34 0.67 1.96 1.41
(0.66–2.72) (0.21–2.16) (1.05–3.64) (0.72–2.77)
17/9.8 (174.4) 11/6.816 (163.0)
23/12.5 (184.0) 16/10.3 (156.1)
0.95 (0.51–1.77) 1.04 (0.48–2.25)
205/95.0 (215.8)
255/160.3 (159.1)
1.36 (1.13–1.63)
NOTE. CI, confidence interval; PY, person-year.
tiple (⭓2) HR-HPV infections was 7.1 cases per 100 PYs (25 cases per 350 PYs) in the circumcised men and 14.4 cases per 100 PYs (62 cases per 430.5 PYs) in the control participants, with an IRR of 0.50 (95% CI, 0.31–0.79). The incidence of type-specific HR-HPV infection over 24 months was lower in the intervention arm than in the control arm for all high-risk genotypes examined (Table 3), and these differences were statistically significant for HR-HPV genotypes 18 (IRR, 0.30; 95% CI, 0.12–0.75) and 33 (IRR, 0.17; 95% CI, 0.04–0.76). The rates of HR-HPV acquisition were lower in the intervention arm than in the control arm in all sociodemographic and behavioral subgroups (Table 4). Among control subjects, HR-HPV incidence was higher in younger men and decreased with age, but no age-related trend was observed among men in the intervention arm. The effect of circumcision on HRHPV acquisition was statistically significant among men reporting high-risk sexual behaviors such as nonuse of condoms (IRR, 0.67; 95% CI, 0.46–0.98) and multiple sex partners (IRR, 0.56; 95% 0.36–0.87). After adjustment for covariates associated with HR-HPV at enrollment (age, education, condom use, alcohol consumption with sex, and number of sex partners), the adjusted incidence rate ratio of HR-HPV acquisition in circumcised men compared with uncircumcised men was 0.67 (95% CI, 0.50–0.91; P p .008). We also assessed acquisition of LR-HPV genotypes. The 24month incidence of any LR-HPV genotype was 29.9 cases per 100 PYs (102 cases per 341.5 PYs) in the intervention arm and 35.0 cases per 100 PYs (144 cases per 411 PYs) in the control
arm (IRR, 0.84; 95% CI, 0.66–1.10). For single LR-HPV infections, the incidence was 19.3 cases per 100 PYs (36 cases per 41.5 PYs) in the intervention arm and 17.8 cases per 100 PYs (73 cases per 411 PYs) in the control arm (IRR, 1.09; 95% CI, 0.78–1.52). However, the incidence of multiple LR-HPV infections was 10.5 cases per 100 PYs (36 cases per 341.5 PYs) in the intervention arm and 17.3 cases per 100 PYs (71 cases per 411 PYs) in the control arm (IRR, 0.61; 95% CI, 0.41– 0.91). Clearance of HR-HPV infections is shown in Table 5. Clearance rates per 100 PYs observation were higher in the intervention arm (215.8 cases per 100 PYs) than in the control arm (159.1 cases per 100 PYs), with a RR of 1.36 (95% CI, 1.13– 1.63). The rates of clearance per 100 PYs were increased for most genotypes, and this was statistically significant for types 39, 51, and 58 (Table 5). After adjustment for age, education, number of sex partners, and condom use, the ratio of clearance rates in the intervention relative to the control arm was 1.39 (95% CI, 1.17–1.64). Among men in the intervention arm, 77.7% (205 of 264) of preexisting HR-HPV infections were cleared over the 24-month observation period, whereas clearance was 66.9% (255 of 381) among the control participants (RR, 1.16; 95% CI, 1.05–1.28). In multivariate analyses, the adjusted RR of percent clearance was 1.17 (95% CI, 1.05–1.31). DISCUSSION Circumcision was associated with a significant reduction in the acquisition of HR-HPV infection, with an estimated effiHigh-Risk HPV in HIV-Negative Men • 223
cacy of ∼33% over 2 years. Moreover, the efficacy was 55% for prevention of multiple HR-HPV infections (Table 2). The protective effect of circumcision was observed for all HR-HPV genotypes (Table 3) and was consistently observed in all sociodemographic and behavioral subgroups (Table 4). These findings are consistent with the observed reduction in the prevalence of HR-HPV in 2 randomized trials [1, 2] and in several observational studies [3–7]. Therefore, it would appear that circumcision protects men from acquiring new HR-HPV infections. Similar reductions were observed in the acquisition of LR-HPV genotypes in circumcised men. The rate of clearance of HR-HPV per 100 PYs was greater among circumcised men than among uncircumcised men (Table 5), which is consistent with 1 observational study, but the magnitude of the effect of circumcision in this trial was less than that reported in the observational study (adjusted HR, 6.5; 95% CI, 2.1–19.7) [7]. The implications of these findings for male-to-female transmission of HR-HPV are unknown, but observational studies suggest lower HR-HPV carriage in female partners of circumcised men [3, 16]. Recent studies have also shown a high degree of type-specific concordance between sex partners within couples [15], so it is likely that reduced male HR-HPV carriage following circumcision, resulting from lower acquisition and faster clearance, will reduce female HR-HPV infections. We are currently evaluating the efficacy of circumcision for reducing female infections among sex partners of men enrolled in this trial. It is plausible that circumcision might reduce HR-HPV acquisition. In uncircumcised men, the foreskin is retracted over the shaft during intercourse, and the inner preputial mucosa is exposed to vaginal and cervical fluids. In addition, the intact foreskin is vulnerable to microtears during intercourse, which could facilitate viral entry, and the moist subpreputial cavity may provide a favorable environment for HR-HPV survival [17]. Thus, by removing the foreskin, circumcision may both reduce exposure to female genital HR-HPV and reduce access of HPV to epidermal basal cells. Additionally, keratinization of the circumcision scar may reduce the risk of HR-HPV infection. We observed a significant reduction in detection of b-globin in penile samples from circumcised men, but there was no change in detection of cellular DNA among uncircumcised control participants. This suggests that progressive keratinization of the surgical scar may reduce the number of basal cells vulnerable to HPV infections over time. This study has limitations. The analysis was confined to a random sample of married men, so we cannot assess efficacy among single men. The 6-month intervals between visits in the first follow-up year and the 12-month interval between visits in the second year were long, and we likely missed incident infections that cleared before the subsequent follow-up visit. 224 • Gray et al
In addition, the long follow-up intervals did not allow a precise estimate of time to clearance of preexisting infections. We were conservative in our estimates of incidence and clearance, which were restricted to sequential samples with amplifiable cellular or viral DNA to ensure the adequacy of sample collection. Because swab samples with amplifiable DNA decreased in the intervention arm relative to control arm, this led to omission of more follow-up intervals for circumcised men (11.0%) than for control participants (8.0%), which differentially reduced person time for the intervention arm and thus possibly inflated rates of acquisition and clearance for this group. We collected swab samples from the coronal sulcus and glans and the shaft but only had resources to assay the corona sulcus– glans samples. Therefore, we do not know whether HR-HPV carriage at other penile sites is affected by circumcision. However, the South African trial sampled the urethral meatus and found that circumcision reduced HR-HPV prevalence, suggesting that the effect may not be site-specific [1]. Some studies to determine the optimal site for penile HPV detection report a lower prevalence of HPV from swab samples of the coronal sulcus or glans compared with the shaft [12, 13], whereas others found no difference in detection between these 2 sites [18]. Collection of samples from 1 site may be a limitation to this study, but because swab procedures were identical in both study arms, this should not bias the estimates of efficacy. In summary, circumcision significantly reduced the incidence of multiple penile HR-HPV infections and increased clearance of preexisting HR-HPV in married HIV-negative men. Thus, male circumcision may potentially reduce exposure of female partners to HR-HPV infection. Acknowledgments We are grateful for the advice provided by the Rakai Community Advisory Board. We also wish to express our gratitude to the study participants whose commitment and cooperation made the study possible.
References 1. Auvert B, Sobngwi-Tambekou J, Cutler E, et al. Effect of male circumcision on the prevalence of high-risk human papillomavirus in young men: results of a randomized controlled trial conducted in orange farm, South Africa. J Infect Dis 2009; 199:14–19. 2. Tobian AAR, Serwadda D, Quinn TC, et al. Male circumcision for the prevention of HSV-2 and HPV infections and syphilis. N Engl J Med 2009; 360:1298–1309. 3. 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. 4. Baldwin SB, Wallace DR, Papenfuss MR, et al. Human papillomavirus infection in men attending a sexually transmitted disease clinic. J Infect Dis 2003; 187:1064–1070. 5. Bosch FX, Manos MM, Munoz N, et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International Biological Study on Cervical Cancer (IBSCC) Study Group. J Natl Cancer Inst 1995; 87:796–802. 6. Hernandez BY, Wilkens LR, Zhu X, et al. Circumcision and human
7.
8.
9.
10. 11.
12.
papillomavirus infection in men: a site-specific comparison. J Infect Dis 2008; 197:787–794. Lu B, Wu Y, Nielson CM, et al. Factors associated with acquisition and clearance of human papillomavirus infection in a cohort of US men: a prospective study. J Infect Dis 2009; 199:362–371. Gray RH, Kigozi G, Serwadda D, et al. Male circumcision for HIV prevention in men in Rakai, Uganda: a randomised trial. Lancet 2007; 369: 657–666. Gravitt PE, Lacey JV Jr, Brinton LA, et al. Evaluation of self-collected cervicovaginal cell samples for human papillomavirus testing by polymerase chain reaction. Cancer Epidemiol Biomarkers Prev 2001; 10: 95–100. Gravitt PE, Peyton CL, Alessi TQ, et al. Improved amplification of genital human papillomaviruses. J Clin Microbiol 2000; 38:357–361. Gravitt PE, Peyton CL, Apple RJ, Wheeler CM. Genotyping of 27 human papillomavirus types by using L1 consensus PCR products by a single-hybridization, reverse line blot detection method. J Clin Microbiol 1998; 36:3020–3027. Giuliano AR, Nielson CM, Flores R, et al. The optimal anatomic sites for sampling heterosexual men for human papillomavirus (HPV) detection: the HPV detection in men study. J Infect Dis 2007; 196:1146– 1152.
13. Weaver BA, Feng Q, Holmes KK, et al. Evaluation of genital sites and sampling techniques for detection of human papillomavirus DNA in men. J Infect Dis 2004; 189:677–685. 14. Safaeian M, Kiddugavu M, Gravitt PE, et al. Determinants of incidence and clearance of high-risk human papillomavirus infections in rural Rakai, Uganda. Cancer Epidemiol Biomarkers Prev 2008; 17:1300– 1307. 15. Mbulawa ZZ, Coetzee D, Marais DJ, et al. Genital human papillomavirus prevalence and human papillomavirus concordance in heterosexual couples are positively associated with human immunodeficiency virus coinfection. J Infect Dis 2009; 199:1514–1524. 16. Gray RH, Wawer M, Thoma M, et al. Male circumcision and the risks of female HIV and sexually transmitted infections acquisition in Rakai, Uganda. In: Programs and Abstracts of the 13th Conference on Retroviruses and Opportunistic Infections (Boston, MA). Alexandria, VA: Foundation for Retrovirology and Human Health, 2006. Abstract 128. 17. Szabo R, Short RV. How does male circumcision protect against HIV infection? BMJ 2000; 320:1592–1594. 18. Smith JS, Moses S, Hudgens MG, et al. Human papillomavirus detection by penile site in young men from Kenya. Sex Transm Dis 2007; 34:928–934.
High-Risk HPV in HIV-Negative Men • 225
