ORIGINAL RESEARCH–ED PHARMACOTHERAPY: Do Phosphodiesterase Type 5 Inhibitors Protect Against Condom-Associated Erection Loss and Condom Slippage?

ORIGINAL RESEARCH–ED PHARMACOTHERAPY: Do Phosphodiesterase Type 5 Inhibitors Protect Against Condom-Associated Erection Loss and Condom Slippage?

1451 ORIGINAL RESEARCH—ED PHARMACOTHERAPY Do Phosphodiesterase Type 5 Inhibitors Protect Against Condom-Associated Erection Loss and Condom Slippage?...

90KB Sizes 0 Downloads 21 Views

1451

ORIGINAL RESEARCH—ED PHARMACOTHERAPY Do Phosphodiesterase Type 5 Inhibitors Protect Against Condom-Associated Erection Loss and Condom Slippage? Stephanie A. Sanders, PhD,*†‡ Robin R. Milhausen, PhD,*‡§ Richard A. Crosby, PhD,*‡¶ Cynthia A. Graham, PhD,*‡** and William L. Yarber, HSD*†‡†† *The Kinsey Institute for Research in Sex, Gender, and Reproduction, Indiana University, Bloomington, IN, USA; † Department of Gender Studies, Indiana University, Bloomington, IN, USA; ‡Rural Center for AIDS/STD Prevention, Indiana University, Bloomington, IN, USA; §Department of Family Relations and Applied Nutrition, University of Guelph, Ontario, Canada; ¶College of Public Health at the University of Kentucky, Lexington, KY, USA; **Oxford Doctoral Course in Clinical Psychology, University of Oxford, Oxford, England; ††Department of Applied Health Science, Indiana University, Bloomington, IN, USA DOI: 10.1111/j.1743-6109.2009.01267.x

ABSTRACT

Introduction. Some physicians prescribe phosphodiesterase type 5 inhibitors (PDE5i) for men who experience condom-associated erection difficulties with a view to increasing condom use and reducing risk of sexually transmitted infections. Aim. To examine whether the prevalence of erection-related condom problems differs between men using and not using PDE5i at the last condom-protected penile–vaginal (PVI) or penile–anal intercourse. Methods. Seven hundred-five men who had used a male condom during the past 3 months for PVI were selected from a sample recruited through advertisement to an electronic mailing list for a large, internet-based, sexual-enhancement product company. An internet-based questionnaire posted in 2006 assessed condom-use errors and problems. Main Outcome Measures. Men who did and did not use PDE5i during the last time a condom was used were compared on: (i) erection loss while applying a condom; (ii) erection loss during sex while using a condom; (iii) condom slipped off during sex; (iv) delayed condom application (penetration of the vagina or anus prior to application of the male condom); (v) early condom removal (condom taken off and intercourse continued without it); (vi) “problem with the way the condom fit”; (vii) “problem with the way the condom felt”; and (viii) condom breakage. Results. Controlling for age, marital status (yes/no), and having children (yes/no), PDE5i users, compared with nonusers, were: (i) three times more likely to report erection loss during sex while using a condom (adjusted odds ratio [AOR] = 3.21, 95% confidence interval [CI] = 1.40–7.39, P = 0.006); (ii) almost five times more likely to report the condom slipped off during sex (AOR = 4.75, 95% CI = 1.68–13.44, P = 0.003); and (iii) more than twice as likely to remove condoms before sex was over (AOR = 2.46, 95% CI = 1.09–5.56, P = 0.03). Conclusions. Physicians prescribing PDE5i may want to evaluate whether men are experiencing condom-associated erection difficulties and, if they are, consider titrating dosages and/or making referrals for psychosexual therapy and/or condom skills education. Sanders SA, Milhausen RR, Crosby RA, Graham CA, and Yarber WL. Do phosphodiesterase type 5 inhibitors protect against condom-associated erection loss and condom slippage? J Sex Med 2009;6:1451–1456. Key Words. Phosphodiesterase 5 Inhibitors; Condoms; Erection; Sexually Transmitted Infections Risk; Erectile Dysfunction

Introduction

P

revious research has suggested that condomassociated erection difficulties (ED) may be

© 2009 International Society for Sexual Medicine

common among men at risk for sexually transmitted infections (STIs) [1–5]. Men who reported ED with condom use were more likely to be inconsistent condom users [2,3,6], have had a greater J Sex Med 2009;6:1451–1456

1452 number of casual partners with whom no condoms were used [1,6], remove condoms before sex was over (incomplete use) [3], and report condom slippage [3,7]. The potential impact of these findings is that men who experience condom-related ED may be at substantially higher risk for STI/HIV. Clinicians may logically believe that PDE5i will alleviate condom-associated erection problems and prescribe them with a view to increasing condom use and reducing risk of STIs [8–10]. For example, PDE5i are often prescribed to HIVpositive men [8–10]. Additionally, young, healthy men may use PDE5i without prescriptions with some believing the drug has the potential to facilitate condom use [11]. However, PDE5i use has been associated with higher rates of risky sexual behavior including unprotected sex in several studies, particularly when combined with other recreational drugs [8–10,12–15]. The relatively high prevalence of erectile dysfunction and of risky sexual practices reported by men is an important context to consider. The estimated prevalence of erectile problems varies across studies. In the U.S. National Health and Social Life survey, Laumann and colleagues [16] reported a prevalence rate of 5% in men under 60 years. Other studies have reported substantially higher rates, e.g., 17% of men at age 40 years and 34% at age 70 years, were found to have moderate erectile problems [17]. In the recent Global Study of Sexual Attitudes and Behaviors, an international survey among adults aged 40–80 years, prevalence rates ranged from 13% to 28% across the 29 countries surveyed [18]. The most important risk factor for erectile dysfunction is age; other correlates are smoking, hormonal factors, and general health [19]. Regarding sexual risk behavior, data from the U.S. National Survey of Family Growth suggests that among men 15 to 44 years of age, classified as being at risk of HIV acquisition, 45.2% had not used condoms during the last episode of penetrative sex. Among same age men who were single, reported ever using condoms, and reported intercourse during the past 4 weeks, 71.4% indicated not using condoms at all in the past 4 weeks. For all men 15 to 44 years of age, 10.4% reported having sex with three or more partners in the past 12 months [20]. These data suggest that erectile dysfunction and risky sexual behavior are fairly common among men, thereby suggesting that research pertaining to PDE5i use and risky sex may be warranted. In particular, it is plausible that PDE5i use may show a relationship with J Sex Med 2009;6:1451–1456

Sanders et al. the quality of condom use in terms of errors and problems. Most studies have focused on condom use/ nonuse and on men who have sex with men. To the best of our knowledge, no previous studies employed detailed assessment of a condom-use event to examine whether PDE5i use reduces condom-associated erection difficulties. Accordingly, the purpose of this study was to examine whether the prevalence of erection-related condom problems differs between men using and not using PDE5i at last condom-protected penile– vaginal or penile–anal intercourse. A predominantly heterosexual sample was assessed.

Methods

Selection and Description of Participants In 2006, participants were recruited from an electronic mailing list for a large, internet-based sexual enhancement product company. The email invitation informed recipients that the study was about sexual arousal and sexual health and included an embedded link to the Study Information page. Only individuals who were at least 18 years old and who were able to read English were eligible to participate. Each link included a random number code that would only be activated one time; this prevented individuals from participating in the study multiple times or forwarding the email invitation to allow others to participate. Consent to participate was given by clicking on a link to the questionnaire. Following completion participants were eligible to enter a drawing for one of 10 $100 gift cards. No identifiers were linked to data. The study procedures were approved by the Human Subjects Committee at the University of Windsor, Ontario, Canada. Nearly 2,000 (N = 1,987; 1,561 men, 426 women) people completed the survey. The following inclusion criteria were applied for this analysis: (i) male (N = 1,561); (ii) used a condom for penile– vaginal or penile–anal intercourse in the past 3 months (N = 805); and (iii) did not use recreational drugs (e.g., crystal meth, cocaine) during last condom use (N = 788). Those younger than 26 years of age (N = 79) and those missing age data (N = 4) were excluded from the analyses because none of the younger men used PDE5i during their last condom use, and thus the samples of PDE5i users and nonusers would be more closely matched on this variable. This resulted in a final sample of 705.

1453

PDE5 Inhibitors and Condoms

Measures Demographic Variables The following sample characteristics are reported: age; marital status (“single—not currently involved with anyone,” “casually dating one or more people,” “seriously dating one person,” “living with partner but not married,” “married,” “widowed,” “separated or divorced”); race/ethnicity (indigenous person, e.g., American Indian or Inuit, Asian, black or African, Native Hawaiian or Pacific Islander, white, Hispanic or Latino/Latina, Other); income (less than $10,000; $10,001 to $25,000; $25,001 to $50,000; $50,001 to 75,000; $75,001 to $100,000; more than $100,000); education (primary school–kindergarten–8th grade; high school or secondary school; technical or trade school; college–2 year; university 4–5-year degree; postgraduate degree); whether or not they had children; sexual orientation (heterosexual/straight, bisexual, lesbian/gay/homosexual, other, uncertain); and country of residence. Event-Specific Measures Assessment focused on the last time participants used a condom for penile–vaginal or penile–anal intercourse within the past 3 months. Men were asked to indicate whether they had used “drugs to enhance sexual performance (e.g., Viagra, Levitra).” Seven erection-related condom problems and breakage were assessed: (i) erection loss while applying a condom; (ii) erection loss during sex while using a condom; (iii) condom slipped off during sex; (iv) delayed condom application (“penetration of the penis in the vagina or anus without the condom, and then it was put on later and penetration continued”); (v) early condom removal (“condom taken off before you finished having sex”); (vi) “problem with the way the condom fit”; (vii) “problem with the way the condom felt”; and (viii) breakage. These items were modified from a previous questionnaire on condom use errors and problems [21]. Vaginal Sex and Condom Use Frequency Measures Participants were asked “In the past 3 months, how many people have you had penile–vaginal sex with?” and “In the past 3 months, how many times have you used a condom for penile–vaginal sex?” Data Analysis Group comparisons using t-tests and chi-square tests were made for demographic variables and consistency of condom use, contrasting those who

used PDE5i (PDE5i) and those who did not use a PDE5i (No PDE5i) the last time they used a condom. Condom use consistency was calculated as the percentage of times condoms were used during intercourse in the past 3 months (number of times condoms were used/number of intercourse events ¥ 100). Each of the eight condom-use problems were treated as an “outcome” variable in separate multiple logistic regression models. In addition to using PDE5i use (no/yes) as a “predictor” variable, demographic variables that differed between those who did and did not use PDE5i during the condom use event were included as covariates (age, being married, having children). These models were used to calculate adjusted odds ratios, their 95% confidence intervals, and respective P values. Multivariate significance was defined by P < 0.05. Results

Characteristics of the Sample PDE5i use during the last time a condom was used for penile–vaginal or penile–anal intercourse was reported by 7.1% of sample. There were no significant group differences (PDE5i vs. No PDE5i) in the consistency of condom use, with just over half (59%) of the sample reporting condom use less than half of the time. Demographic data are presented in Table 1. The mean age of the sample was 38.2 years (standard deviation = 9.1, range 26–70 years). Compared with the No-PDE5i group, men using PDE5i were significantly older (t = 6.34, adjusted degrees of freedom [d.f.] = 52.35, P < 0.001), were less likely to be currently married (c2 = 4.05, d.f. 1, P = 0.044), or have children (c2 = 3.989, d.f. 1, P = 0.046). Given the group differences, these variables were used as covariates in the multivariate regression analyses. There were no group differences for the remaining demographic variables. The large majority of the sample self-identified as white (84.3%), with 4.1% and 3.3% self-identifying as black/African and Hispanic/Latino, respectively. Nearly all (95.6%) self-identified as heterosexual. The sample was highly educated with 56.7% having completed a bachelor’s degree or higher. More than half (57.6%) had a household income of at least $75,000. Most (89.8%) were from the United States, with 6.1% from Canada or the British Isles. The sample reflects the composition of the electronic mailing list for the Liberator company. The majority of persons purchasing products from this American company are from J Sex Med 2009;6:1451–1456

1454 Table 1

Sanders et al. Demographic characteristics of the sample (N = 705)

Demographic characteristics

Total sample, N (%) (N = 705)

PDE5i, N (%) (N = 49)

No PDE5i, N (%) (N = 656)

Significance* P

Age, mean (standard deviation) Currently married Have children Race/ethnicity: white Heterosexual Education: at least bachelor’s degree† Income: over $75,000† U.S. resident

38.2 499 259 594 674 400 406 633

47.8 28 11 43 46 31 33 45

37.5 471 248 551 628 369 373 588

<0.001 <0.05 <0.05 0.62 0.80 0.42 0.20 0.81

(9.1) (70.8) (36.7) (84.3) (95.6) (56.7) (57.6) (89.8)

(11.1) (57.1) (22.4) (87.8) (93.9) (63.3) (67.3) (91.8)

(8.6) (71.8) (37.8) (84.0) (95.7) (56.3) (56.9) (89.6)

*A t-test was used to analyze age. All other P values derived from chi-square tests were corrected for continuity. †Education and income were recorded as six ordinal categories. These variables were dichotomized at the median (education greater than or equal to a bachelor’s degree and income greater than $75,000) for presentation in the table. Mann–Whitney U-tests of the ordinal data found no statistically significant differences between groups. PDEi = phosphodiesterase type 5 inhibitor.

the United States, male, well-educated and with significant incomes as the products are not inexpensive (O. Kaszubski, personal communication, March 7, 2006).

Multivariate Associations Table 2 displays the results from the logistic regression models controlling for age, marital status, and having children. PDE5i users, compared with nonusers, were: (i) three times more likely to report erection loss during sex while using a condom (unadjusted: 20% of PDE5i users; 6% of nonusers); (ii) almost five times more likely to report the condom slipped off during sex (unadjusted: 15% of PDE5i users; 3% of nonusers); and (iii) more than twice as likely to remove condoms before sex was over (unadjusted: 20% of PDE5i users; 8% of nonusers). There was a trend for more PDE5i users to report problems with condom “fit” (unadjusted: 17% of PDE5i users; 9% of nonusers). Significant associations were not found for erection loss during condom application (unadjusted: 15% of PDE5i users; 7% of nonusers), delayed application (unadjusted: 20% of PDE5i users; 22% of nonTable 2

users), problems with the “feel” of the condom (unadjusted: 41% of PDE5i users; 31% of nonusers), and breakage (unadjusted: 0% of PDE5i users; 1.5% of nonusers). Discussion

To our knowledge this is the first published study that has compared the frequency of these condom-use errors (delayed application, early removal) and condom use problems (erection loss during condom application, erection loss during sex while using a condom, condom slippage during sex, breakage, problems with “fit,” problems with “feel”) in a group of men using PDE5i with those not using these medications. We found that men who used PDE5i were substantially more likely to report erection loss during sex while using a condom, condom slippage during sex, and early removal of condoms during the last time they used condoms than men who reported no PDE5i use. Our findings on the prevalence of condomrelated erection problems appear to be consistent with other recent studies of heterosexual men.

Odd ratios associated with PDE5i use for condom use problems and errors

Outcome variable

PDE5i, N (%) (N = 49)

No PDE5i, N (%) (N = 656)

Adjusted OR†, N (%) (95% CI)

Significance P

Erection problem during application Erection problem during sex Slipped off during sex Delayed application Early removal Fit Feel Breakage

7 10 7 10 10 8 20 0

43 40 20 145 56 60 200 10

1.58 3.21 4.75 1.36 2.46 2.34 1.57 0.000

0.32 0.006 0.003 0.43 0.03 0.055 0.16 1.00

(15) (20) (15) (20) (20) (17) (41) (0)

(7) (6) (3) (22) (8) (9) (31) (1.5)

† Odds ratio adjusted for age, marital status (no, yes), and children (no, yes). CI = confidence interval; OR = odds ratio; PDEi = phosphodiesterase type 5 inhibitor.

J Sex Med 2009;6:1451–1456

(0.62–3.99) (1.40–7.39) (1.68–13.44) (0.64–2.93) (1.09–5.56) (0.98–5.57) (0.84–2.93) (0.000–)

1455

PDE5 Inhibitors and Condoms Reporting on the last condom-use event only, 7% of the overall sample reported erection loss during condom application and 7% reported erection loss during sex while using a condom, with a total of 11.6% reporting one or both of these types of condom-associated erection loss. Other studies have used different time frames for reporting and focused on samples of young men. In a U.S. study of 234 sexually active, college-age men, 25% reported ever having had difficulty getting or keeping an erection while putting on a condom; 6% had used PDE5i medication, largely obtained from nonmedical sources, e.g., friends, Internet [6]. Among young men attending an urban STI clinic, 37.1% reported condomassociated erection loss on at least one occasion in the previous 3 months [3]. In a recent Brazilian study of medical students, erectile difficulties occurring with condom use were reported by approximately a quarter of the sample. Nine percent of this sample of young, healthy men reported previous use of PDE5i, and 71.4% thought that the drug had the potential to facilitate condom use [11]. Our findings suggest that although PDE5i may improve erectile functioning, the assumption that the use of these medications will eliminate condom-associated erection problems [11] may be erroneous. In the current study we did not obtain information on the quality of men’s erections during PDE5i use. However, in our view, the most likely reason for the specific condom problems reported by men using PDE5i, e.g., condom slippage during sex, early removal of condoms, was that this group of men was still experiencing some degree of erectile problems. Research employing a within-subjects design examining multiple condom-use events with and without use of PDE5i would be needed to determine whether, within individuals, PDE5i use reduces the likelihood of condom-associated erection difficulties. Much of the previous research on the association between PDE5i use and sexual risk behavior has focused on men who have sex with men, and on HIV-positive men. In these populations a relationship between PDE5i use and risky sexual behavior has been reported [13–15], which appears to be mediated by the use of recreational or “party” drugs [8]. There is also some evidence that young heterosexual men using PDE5i may frequently mix these with other drugs, e.g., cocaine, methamphetamine, or alcohol [6,11]. In the current study, we excluded

men who had used recreational drugs during last condom use. Future studies should assess the effects of use of other drugs and alcohol, in combination with PDE5i, on condom-associated erection loss.

Limitations The small non-probability sample and information from a single event for each participant limit generalization. Additionally, because respondents were recruited from the mailing list of a sexualenhancement production company, it is possible that the sample was biased towards those individuals who were more likely to be experiencing sexual problems; however, this is not necessarily the case. First, potential volunteers were informed that the study was about “sexual arousal and health,” not about sexual problems or erectile dysfunction. Second, we would not assume that individuals who are on the mailing list of a sexual enhancement product company are necessarily experiencing sexual problems; many may be looking to enhance their sexual life but may not necessarily have sexual difficulties. Finally, we did not collect any information about the dosage of PDE5i, or about erection loss in the context of specific sexual behaviors, which would have been informative. Although this was a small, exploratory study, our findings nonetheless have important implications. First, we agree with Korkes and colleagues [11] that condom-related erectile problems seem to be an underestimated problem. We should not assume that PDE5i use, even in young men, is primarily “recreational.” As discussed above, there is evidence that condom-associated erectile problems may be relatively common, even among younger men [3,6,11]. Second, our findings underline the importance of assessing not only whether condoms are used, but also whether they are used correctly and throughout the sexual encounter [5]. Conclusions

We found that the prevalence of erection-related condom problems (erection loss during sex while using a condom, condom slippage, and early removal of condoms) were higher in men using PDE5i compared with men not using PDE5i. Physicians prescribing PDE5i may want to evaluate whether men using these medications are experiencing condom-associated erection difficulties and, if they are, consider titrating dosages and/or making referrals for psychosexual therapy and condom skills education. J Sex Med 2009;6:1451–1456

1456 Corresponding Author: Cynthia A. Graham, PhD, Oxford Doctoral Course in Clinical Psychology, Isis Education Centre, Warneford Hospital, Headington, Oxfordshire UK OX37JX. Tel: +44 (0) 1865 226431; Fax: +44 (0) 1865 226364; E-mail: cygraham@ indiana.edu Conflict of Interest: None declared.

Sanders et al.

11

12

References

1 Bancroft J, Janssen E, Strong D, Carnes L, Vukadinovic Z, Long JS. Sexual risk-taking in gay men: The relevance of sexual arousability. Arch Sex Behav 2003;32:555–72. 2 Cove J, Petrak J. Factors associated with sexual problems in HIV-positive gay men. Int J STD AIDS 2004;15:732–6. 3 Graham CA, Crosby R, Yarber WL, Sanders SA, McBride K, Milhausen R, et al. Erection loss in association with condom use among young men attending a public STI clinic: Potential correlates and implications for risk behaviour. Sex Health 2006;3:1–6. 4 Richters J, Hendry O, Kippax S. When safe sex isn’t safe. Cult Health Sex 2003;5:37–52. 5 Graham CA, Crosby R, Sanders SA, Yarber WL. Assessment of condom use in men and women. Annu Rev Sex Res 2006;16:20–52. 6 Musacchio NS, Hartrich M, Garofalo R. Erectile dysfunction and Viagra use: What’s up with collegeage males? J Adoles Health 2006;39:452–4. 7 Yarber WL, Graham CA, Sanders SA, Crosby RA. Correlates of condom breakage and slippage among undergraduates. Int J STD AIDS 2004;15:467– 72. 8 Rosen RC, Catania JA, Ehrhardt AA, Burnett AL, Lue TF, McKenna K, Heiman JR, Schwarcz S, Ostrow DG, Hirshfield S, Purcell DW, Fisher WA, Stall R, Halkitis PN, PhD, Latini DM, Elford J, Laumann EO, Sonenstein FL, Greenblatt DJ, Kloner RA, Lee J, Malebranche D, Janssen E, Diaz R, Klausner JD, Caplan AL, Jackson G, Shabsigh R, Khalsa JH, Stoff DM. The Bolger Conference on PDE-5 inhibition and HIV risk: Implications for health policy and prevention. J Sex Med 2006;3: 960–75. 9 Goldmeier D, Lamba H, Richardson D. Commentary on Rosen RC, Catania JA, Ehrhardt AA, et al. The Bolger Conference on PDE-5 inhibition and HIV risk: Implications for health policy and prevention. J Sex Med 2006;3:974. 10 Sadeghi-Nejad H. Commentary on Rosen RC, Catania JA, Ehrhardt AA, et al. The Bolger Conference on PDE-5 inhibition and HIV risk:

J Sex Med 2009;6:1451–1456

13

14

15

16 17

18

19

20

21

Implications for health policy and prevention. J Sex Med 2006;3:974–5. Korkes F, Costa-Matos A, Gasperini R, Reginato PV, Perez MDC. Recreational use of PDE5 inhibitors by young healthy men: Recognizing this issue among medical students. J Sex Med 2008;5: 2414–8. Cachy E, Mar-Tang M, Mathews WC. Screening for potentially transmitting sexual risk behaviors, urethral sexually transmitted infection, and Sildenafil use among males entering care for HIV. AIDS Patient Care STDS 2004;18:349–54. Cooperman NA, Arnsten JH, Klein RS. Current sexual activity and risky sexual behavior in older men with or at risk for HIV infection. AIDS Educ Prev 2007;19:321–33. Mansergh G, Shouse RL, Marks G, Guzman R, Rader M, Buchbinder S, Colfax GN. Methamphetamine and sildenafil (Viagra) use are linked to unprotected receptive and insertive anal sex, respectively, in a sample of men who have sex with men. Sex Trans Infect 2006;82:131–4. Paul JP, Pollack L, Osmond D, Catania JA. Viagra (Sildenafil) use in a population-based sample of U.S. men who have sex with men. Sex Transm Dis 2005;32:531–3. Laumann EO, Paik A, Rosen RC. Sexual dysfunction in the US: Prevalence and predictors. JAMA 1999;281:537–44. Feldman HA, Goldstein I, Hatzichristou DG, Krane RJ, McKinlay JB. Impotence and its medical and psychological correlates: Results of the Massachusetts Male Aging Study. J Urol 1994;151:54– 61. Laumann EO, Nicolosi A, Glasser DB, Paik A, Gingell C, Moreira E, Wang T for the GAAAB Investigators’ Group. Sexual problems among women and men aged 40–80 years: Prevalence and correlates identified in the Global study of sexual attitudes and behaviors. Intl J Impot Res 2005; 17:39–57. Lewis RW, Fugl-Meyer KS, Bosch R, Fugl-Meyer AR, Laumann EO, Lizza E, Martin-Morales A. Definitions, classification, and epidemiology of sexual dysfunction. In: Lue TF, Basson R, Rosen R, Guiliano F, Khoury S, Montorsi F, eds. Sexual Medicine. Sexual dysfunctions in men and women. Paris: Health Publications; 2004:39–72. National Center for Health Statistics. Key Statistics from the NSFG. Available at: http://www.cdc. gov/nchs/about/major/nsfg/abclist.htm. (accessed August 7, 2008). Crosby RA, Sanders SA, Yarber WL, Graham CA, Dodge B. Condom use errors and problems among college men. Sex Transm Dis 2002;29:552–7.