- Email: [email protected]
Latex condom breakage and slippage in a controlled clinical trial
Latex Condom Breakage and Slippage in a Controlled Clinical Trial Michael J. Rosenberg*t
and Michael S. Waugh*
Although millions of couples rely on male latex condoms to protect against unintended pregnancy and sexually transmitted infections, their use is limited in part by questions about their performance. Rates of condom breakage and slippage, two measures of performance, vary broadly across studies. This variation in part reflects study variability and limitations, including sample size, reliance on subjects’ memory, user populations, and products evaluated. In an effort to define condom performance in a group of monogamous couples typical of those using condoms for contraception, we conducted a clinical trial of a single brand of lubricated condoms (Durex Ramses@). A total of 4637 attempts to use the condom were evaluated. Six breaks occurred before intercourse (nonclinical breaks), and 10 condoms broke during intercourse or were only noted to have broken upon withdrawal (clinical breaks), resulting in a nonclinical breakage rate of 0.13% (95% confidence interval, 0.05-0.2%%), clinical breakage rate of 0.28% (0.15-0.4%%), and a total breakage rate of 0.41% (0.25-0.64%). The rate of complete slippage was 0.63% (0.42-0.90%), and total failure (clinical breaks plus complete slips) was 1.04% (0.76-1.37%). These rates are lower than those in other studies with the exception of one, a prospective investigation in a population of female prostitutes. Results indicate that condoms can, in experienced, motivated populations, provide excellent performance and suggest that their efficacy at preventing pregnancy may equal that of the most reliable forms of contraception. Because this study involved a single condom brand, these results may not be generalizabie to other brands. CONTRACEPTION 1997;56: 17-21 0 1997 Elsevier Science Inc. All rights reserved. KEY
WORDS:
condom,
latex, breakage,
slippage,
failure
*Health Decisions. Inc.. Chaoel HIII. NC and tDeoartments of ObstetricsGynecology and &idemioiogy, University of Norih daroiina, Chapel Hill, NC Name and address for correspondence: Dr. Michael Rosenberg, Health Decisions, Inc., 1516 E. Franklin St., Suite 200, Chapel Hill, NC 27514. Tel: (919) 967-1111; Fax: (919) 967-1145 Submitted for publication March 18, 1997 Revised April 30, 1997 Accepted for publication April 30. 1997
0 1997 Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010
Introduction ondoms are widely used for preventing pregnancy as well as the spread of HIV and other sexually transmitted infections (STIs). In 1988, the most recent year for which national statistics are available, an estimated 5.1 million US women (13.2% of women at risk for unintended pregnancy) used condoms to prevent pregnancy, and 4.2 million additional women used them to prevent ST1s.l In addition to their consistent use, their ability to limit both outcomes depends on how infrequently condoms break or slip off during intercourse.2 In addition, condom breaks and slips may in themselves be barriers to use, as many men consider breaks and slips to be common problems.3 Reported rates of breakage and slippage vary widel~.~ Most studies report total breakage rates between 2.4% and 6.7%, though a rate of 18.6% has been reported with an 81-month-old lot of condoms.2t4-6 Total breakage is divided into clinical breakage, which is breakage during intercourse or breaks that are only noted upon withdrawal, and nonclinical breakage, which occurs before intercourse, typically when the condom is being put on. Clinical breakage rates in prospective studies vary between 0.5% and 3.7%; in retrospective studies, total condom breakage rates range between 0.6% and 7.9%.4 Complete condom slippage, where the condom falls completely off the penis during intercourse or withdrawal, varies from 0.6% to 5.4% in prospective studies and 0.3% to 5.1% in retrospective studies.4 One prospective study, performed in prostitutes, reflects very low breakage (0%) and clinical slippage (0.6%) rates, suggesting that condoms can provide very reliable performance.4 Condom performance may depend on use of oil-based lubricants or additional water-based lubricant; condom brand, age, and storage conditions; and user age, condom experience, and marital status.7-‘0 Subjective considerations associated with increased rates of breakage include vigorous or intense sex, inadequate lubrication, and tearing with fingernails.’ The majority of condom studies are retrospective, meaning that subjects rely on memory of events that may have taken place months earlier. Prospective
C
ISSN OOIO-7824/97/$17.00 PII SOOIO-7824(97)00069-3
18
Rosenberg
Contraception 1997;56:17-21
and Waugh
studies often have relatively small sample sizes, frequently including fewer than 50 subjects4 Because few studies have been conducted with larger numbers of typical users as part of a controlled clinical trial, we conducted a multicenter clinical study using a single brand of male latex condom.
l
l
Methods Study participants were monogamous couples (males 18-50 years old; females 18-40 years old) who had sex at least six times a month and did not have a history of fertility-impairing conditions, including STIs. Couples were at risk of pregnancy, and condoms were their sole method of birth control during the study. Female partners were tested for pregnancy at study entry, then again at study exit. Other clinical evaluations were performed at the request of the participants or as clinically indicated (cervical cytology, wet mount, ST1 tests, and physical examination). Pregnancy tests could be requested by the patient at any time, and postcoital contraception was available on request. The study protocol was approved by our institutional review board and participants provided informed consent before entry. Enrolled couples used lubricated Durex Ramses (London International Group, London, England) brand male latex condoms. Detailed information was recorded on the first five coital acts. At each sexual encounter, a diary was completed that included information on condom use, breaks or slips, as well as type of sexual activity (vaginal vs. anal), and use of additional lubrication. In addition, male partners were provided with a kit by which penis length and circumference were measured. Couples were provided addressed, stamped envelopes to return diaries on a monthly basis, and returned to the clinic after 3 months to review forms and receive additional condoms as needed. Couples were contacted by telephone if forms were not received within 2 weeks of the expected monthly submission date, and also to schedule the 3 month interim visit as well as the exit interview. Couples were followed for 6 months, after which time the final pregnancy test and exit interview were conducted. Statistical analysis included comparison of condom breakage and slippage rates for couples that reported at least one attempt to use the condom, with exact 95 % confidence intervals for binomial data. l1 Terms used for this evaluation are l
l
Clinical breakage rate: breaks that occur during intercourse or that are only noted upon withdrawal divided by the total number of acts of intercourse; Nonclinical breakage rate: breaks that occur before
l
l
intercourse divided by the total attempts to use the condom; Total breakage rate: the sum of clinical and nonclinical breaks divided by the total number of attempts to use the condom; Complete slippage rate: events where the condom comes completely off the penis during intercourse or withdrawal divided by the total acts of intercourse; Total clinical failure rate: the sum of clinical breaks and complete slips divided by the total acts of intercourse; Total failure rate: the sum of total breaks and complete slips divided by the total number of attempts to use the condom.
Results Study participants were relatively young and primarily white (Table I). Nearly all subjects were either married or had never been married, in approximately equal proportions. Slightly less than half the women were parous. Nearly all participants completed high school, and slightly over half had also graduated from college or had completed some college. Three-quarters of subjects lived with their current partner or spouse. The group was sexually experienced, having been involved in a sexual relationship with their study partner for an average of four years. Approximately 70-75% of the study participants were experienced condom users, having used a condom more than 50 times with their current partner, while only approximately 8% had used a condom 10 or fewer times with their current partner. Two men and two women reported no previous condom experience, but did not have any breakage or slippage during the current study. Of the 92 couples that contributed data to this analysis, 66% completed 6 months of follow-up, 18% were lost to follow-up, 2% discontinued for condomrelated reasons (complaints of breakage, discomfort), and the remaining 14% discontinued for reasons unrelated to the study condom (no longer with partner, study requires too much time, moving out of the area, etc.). Those that discontinued early or were lost to follow-up were not different sociodemographically from those that completed 6 months of follow-up. A total of 4637 attempts to use the condom were reported (Table 2). Six of these resulted in nonclinical breaks, leaving 4631 condoms that were used for intercourse. The overwhelming majority of intercourse was vaginal; only four acts of anal intercourse were reported, in one case using the same condom for vaginal intercourse, and none resulted in breakage or
Contraception 1997;56:17-21
Table
Condom
Breakage
and Slippage
in a Controlled
Clinical
Trial
19
1. Characteristics of study participants
Characteristic Mean age (SD), years Ethnicity White Hispanic Black Native American Other Asian/Pacific Islander Marital status Single Married Divorced Separated Widowed Education level Did not complete high school Still in high school High school graduate Some college, no longer enrolled Still in college College graduate Any postgraduate work Living arrangements Live with current partner Live with spouse Live with parents Live alone Live with roommate(s) Employment status Full time Part time Student Unemployed Homemaker Total household income $O-10,000 $lO,OOl-20,000 $20,001-30,000 $30,001-40,000 $40,001-50,000 $50,001-60,000
Greater than $60,001 Don’t know/prefer not to answer Years in sexual relationship with study partner (mean, median) Age at first intercourse (mean, median) Lifetime no. sexual partners (mean, median) Parous Yes No Previous condom use with study partner Never l-2 times 3-10 times 1l-50 times >50 times Previous condom use with any partner Never l-2 times 3-10 times 1l-50 times >50 times
Male partner (n = 92), %
Female partner (n = 92), %
28.8 (6.2)
26.8 (5.5)
75.0 12.0 7.6 4.3
75.0 7.6 8.7 2.2 2.2 4.3
1.1 45.7 46.7 7.6 -
46.7 46.7 5.4
2.2 3.3 19.6 28.3 19.6 22.8 4.3
3.3 10.9 29.3 27.2 22.8 6.5
38.0 39.1 7.6 5.4 9.8
34.8 41.3 6.5 4.3 13.0
63.0 17.4 115.2 4.3 -
44.6 20.7 21.7 4.3 8.7
9.8 17.4 30.4 10.9 8.7 5.4 7.6 9.8 4.0, 3.1 16.8, 16.0 15.2, 7.0
-
1.1
9.8 17.4 27.2 15.2 10.9 4.3 7.6 7.6 4.0,3.1 16.9, 17.0 12.1, 5.5
-
50.0 50.0
2.2
2.2
1.1
1.1
2.2 25.0 69.6
2.2 23.9 70.7
2.2
2.2
1.1 5.4 20.7 70.7
1.1 3.3 18.5 75.0
20
Rosenberg
and Waugh
Contraception 199756: 17-21
Table 2. Condom usage Vaginal intercourse only Anal intercourse only Anal and vaginal intercourse Acts of intercourse First five usesonly All subsequentuses Total acts Number acts per couple (mean, SD) Attempts to use condom First five attempts All subsequentattempts Total attempts to use condom
4626 4 1 456
4175 463 1 50.3, 41.0 457
4180 4637
slippage. Thirteen clinical breaks occurred, resulting in a clinical breakage rate of 0.28%, and a total breakage rate of 0.41% (Table 3). Nine couples broke one condom each, and two couples broke two condoms each. Twenty-nine complete slips were reported, resulting in a slippage rate of 0.63%. Of these complete slips, 21 were reported by a single couple. Total clinical failure, at 42 events, was 0.91%.
Discussion Less than 1 out of every 100 condoms used failed in a way that might place a user at risk of pregnancy or STI. These breakage and slippage rates are similar to those in a recent study of condom performance among female prostitutes and are lower than all others in the published literature.4 These results, recorded among couples that were sexually active, for the most part experienced with condom use, and monogamous, suggest that condoms can provide an extremely reliable method of contraception and ST1 protection in everyday practice. The very small number of events, however, makes it impossible with the current study to explore risk factors, including user characteristics, sexual activities, and penis size. These results may be more broadly generalizable than other condom evaluations because of the population we studied. Many prospective clinical trials include subjects who may not be typical of the end user population. l2 This is often a concern with condom studies, which exclude individuals at risk of becoming pregnant or acquiring a STI.2~12,‘3 The absence of such risk may alter sexual practices or behaviors. Since participants in this study were at risk of unintended pregnancy and possibly of STI, couples would have remained strongly motivated to use the condom correctly. Though self-selection is an issue in any clinical trial, including this one, this form of bias is difficult to quantify. The characteristics of users in this study, including diverse geographical distribution, thus suggest that our results are
generalizable to monogamous couples using condoms as their primary method of birth control. Also impacting generalizability of results is the size of the study. The number of acts of intercourse (4631) is considerably larger than that in most studies, and the number of participants in this study, 92, is in the mid-range of other prospective studies of condom breakage and slippage. Smaller sample size studies of condom breakage and slippage can be adversely affected by the inclusion of one or more couples that are prone to breakage and slippage; that certain couples are prone to such problems has been noted in other studies.9J4,1 5 One such couple in this study reported 21 of the 29 complete slips. Removing this couple results in a clinical failure rate of 0.46%, or one-half of that currently reported. In a study with fewer couples and acts of Table 3. Condom performance
Condom breakage First five usesonly Nonclinical breaks Clinical breaks Total breaks All subsequentuses Nonclinical breaks Clinical breaks Total breaks All usescombined Nonclinical breaks Clinical breaks Total breaks Condom slippage First five usesonly, complete slips All subsequent uses, complete slips All usescombined, complete slips Condom failure First five usesonly Clinical failure Total failure All subsequentuses Clinical failure Total failure All usescombined Clinical failure Total failure
Events
Rate (%)
95% CI
1 5 6
0.22 1.10 1.31
0.01-l .21 0.36-2.54 0.48-2.84
2 13
0.12 0.19 0.31
0.04-0.28 0.08-0.38 0.17-0.53
6 13 19
0.13 0.28 0.41
0.05-0.28 0.15-0.48 0.25-0.64
3
0.66
0.14-1.91
26
0.62
0.41-0.91
29
0.63
0.42-0.90
9 9
1.75 1.97
0.76-3.43 0.90-3.71
34 39
0.81 0.93
0.56-l. 14 0.66-1.27
42 48
0.91 1.04
0.65-1.22 0.76-l .37
Nonclinical breaks are breaks before intercourse divided by attempts to use condom. Clinical breaks are breaks during intercourse or only noted upon withdrawal divided by acts of intercourse. Total breaks are clinical and nonclinical breaks divided by attempts to use condom. Complete slips are slips completely off the penis divided by acts of intercourse. Clinical failure is clinical breaks and complete slips divided by acts of intercourse. Total failure is clinical breaks, nonclinical breaks, and complete slips diwded by attempts to use condom.
Contraception 1997;56:17-21
Condom
intercourse, this one couple would have had an even larger effect on interpretation of condom performance, greatly limiting the generalizability of study findings. Thus, the size of the study reported here increases our confidence in generalizability of its findings. Among study limitations was the small number of breaks or slips, which precluded our exploring other factors for their association with condom performance. These other factors include use of additional lubricants, particularly if oil-based, age and storage conditions of the condom, user age, user experience with condoms, marital status, vigorous sex, inadequate lubrication, tearing with fingernails, and condom brand.“‘O We eliminated the potential variable of differences in condom brands by using a single, commonly available condom. The excellent condom performance reflected in this study may thus reflect the condom, users, or some combination of these factors in addition to others. The use of a single brand of condom means that our results may not be generalizable to others. An additional concern is loss to follow-up and early drop-outs due to problems using the condom. At 19% (18% loss to follow-up and 1% who cited excessive breakage as discontinuation reason), while consistent with other studies of similar duration, it raises the possibility that if couples dropped out early because of problems they experienced with the condom, then the results we report actually underestimate true breakage and slippage rates; that is, study results may be based on acts of intercourse the large majority of which are contributed by users who have little to no problems with breakage or slippage. As a sensitivity analysis, we compared rates of clinical breakage and complete slippage for couples lost to follow-up or who discontinued early citing condom-related problems with those who completed the study. This group reported three clinical breaks and two complete slips in 348 acts of intercourse, resulting in clinical breakage and slippage rates of 0.86% and 0.57%, respectively. For these early dropouts, the total clinical failure rate is 1.43%. Breakage rates are, thus, slightly higher than those reported by couples that completed the investigation (clinical breakage rate for this group was 0.23%); this difference approaches but does not achieve statistical significance (Fisher’s exact test, p = 0.07). Complete slippage rates are slightly higher for couples who finished the study (0.63%) and lower for total clinical failure (0.86%), but neither of these differences are statistically significant. Based on this sensitivity analysis, it appears that if the couples lost to follow-up or who dropped out because of problems with the condom had continued in the study to completion, study findings would not have changed significantly; results would still show condom failure to be an extremely uncommon event among this population of users.
Breakage
and Slippage
in a Controlled
Clinical
Trial
21
Overall, this study indicates that lubricated Durex Ramses latex condoms can provide very reliable performance in a group of motivated, experienced users. Although results presented here of breakage and slippage do not translate directly into risk of pregnancy or STI, the rates-lower than any other reported thus far in a group of typical couples and more reliable because of the number of users-are suggestive that efficacy may compare favorably with those of other methods such as oral contraceptives that have been considered among the most reliable forms of userdependent contraception.16
References 1. London KA, Mosher WD. What is Current Use of a
Contraceptive? Hyattsville, MD: National Center for Health Statistics, 1990. 2. Trussell J, Warner DL, Hatcher RA. Condom slippage and breakagerates. Fam Plann Perspect 1992;24:203. 3. Grady WR, Klepinger DH, Billy JO, Tanfer K. Condom characteristics:the perceptionsand preferencesof men in the United States. Fam Plann Perspect 1993;25:67-73. 4. Albert A, Warner DL, Hatcher RA, Trussell J, Bennett C. Condom use among female commercial sex workers in Nevada’s legal brothels. Am J Public Health 1995j85: 1514-20. 5. Russell-Brown P, Piedrahita C, Foldesy R, Steiner M,
Townsend J. Comparison of condom breakage during human use with performance in laboratory testing. Contraception 1992;45:429-37. 6. Steiner M, Foldesy R, Cole D, Carter E. Study to determine the correlation between condom breakage in human use and laboratory test results. Contraception 1992;46:279-88. 7. Steiner M, et al. The impact of lubricants on latex condoms. Int J STD AIDS 1994;5:30-6. 8. Steiner M, Piedrahita C, Glover L, Joanis C. Can condom users likely to experience condom failure be identified. Fam Plann Perspect 1993;25:220-6. 9. SparrowMJ, Lavill K. Breakageandslippageof condomsin family planning clients. Contraception 1994j50:117-29. 10. Joanis C, Brookshire T, Piedrahita C, Steiner M. Evaluation of Two Condom Designs: a Comparison of Standard and Larger Condoms in Ghana, Kenya, and Mali. Research Triangle Park: Family Health International, 1990. 11. StatXact version 2.0. Cambridge: Cytel Software, 1992. 12. Grady WR, Tanfer K. Condom breakage and slippage among men in the United States. Fam Plann Perspect 1994;26:107-12. 13. Trussell J, Warner DL, Hatcher RA. Condom perfor-
mance during vaginal intercourse: comparison of Trojan-Enz and Tactylon condoms. Contraception 1992j45: 1 l-9. 14. Free MJ, Skiens EW, Morrow MM. Relationship be-
tween condom strength and failure during use. Contraception 1980;22:31-7. 15. Richters J, Donovan B, Gerofi J. How often do condoms break or slip off in use?Int J STD AIDS 1993;4:90-4, 16. Hatcher RA, Trussell J, Stewart F, et al. Contraceptive Technology. New York: Irvington Publishers, 1996: 107-43.
and Michael S. Waugh*
Although millions of couples rely on male latex condoms to protect against unintended pregnancy and sexually transmitted infections, their use is limited in part by questions about their performance. Rates of condom breakage and slippage, two measures of performance, vary broadly across studies. This variation in part reflects study variability and limitations, including sample size, reliance on subjects’ memory, user populations, and products evaluated. In an effort to define condom performance in a group of monogamous couples typical of those using condoms for contraception, we conducted a clinical trial of a single brand of lubricated condoms (Durex Ramses@). A total of 4637 attempts to use the condom were evaluated. Six breaks occurred before intercourse (nonclinical breaks), and 10 condoms broke during intercourse or were only noted to have broken upon withdrawal (clinical breaks), resulting in a nonclinical breakage rate of 0.13% (95% confidence interval, 0.05-0.2%%), clinical breakage rate of 0.28% (0.15-0.4%%), and a total breakage rate of 0.41% (0.25-0.64%). The rate of complete slippage was 0.63% (0.42-0.90%), and total failure (clinical breaks plus complete slips) was 1.04% (0.76-1.37%). These rates are lower than those in other studies with the exception of one, a prospective investigation in a population of female prostitutes. Results indicate that condoms can, in experienced, motivated populations, provide excellent performance and suggest that their efficacy at preventing pregnancy may equal that of the most reliable forms of contraception. Because this study involved a single condom brand, these results may not be generalizabie to other brands. CONTRACEPTION 1997;56: 17-21 0 1997 Elsevier Science Inc. All rights reserved. KEY
WORDS:
condom,
latex, breakage,
slippage,
failure
*Health Decisions. Inc.. Chaoel HIII. NC and tDeoartments of ObstetricsGynecology and &idemioiogy, University of Norih daroiina, Chapel Hill, NC Name and address for correspondence: Dr. Michael Rosenberg, Health Decisions, Inc., 1516 E. Franklin St., Suite 200, Chapel Hill, NC 27514. Tel: (919) 967-1111; Fax: (919) 967-1145 Submitted for publication March 18, 1997 Revised April 30, 1997 Accepted for publication April 30. 1997
0 1997 Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010
Introduction ondoms are widely used for preventing pregnancy as well as the spread of HIV and other sexually transmitted infections (STIs). In 1988, the most recent year for which national statistics are available, an estimated 5.1 million US women (13.2% of women at risk for unintended pregnancy) used condoms to prevent pregnancy, and 4.2 million additional women used them to prevent ST1s.l In addition to their consistent use, their ability to limit both outcomes depends on how infrequently condoms break or slip off during intercourse.2 In addition, condom breaks and slips may in themselves be barriers to use, as many men consider breaks and slips to be common problems.3 Reported rates of breakage and slippage vary widel~.~ Most studies report total breakage rates between 2.4% and 6.7%, though a rate of 18.6% has been reported with an 81-month-old lot of condoms.2t4-6 Total breakage is divided into clinical breakage, which is breakage during intercourse or breaks that are only noted upon withdrawal, and nonclinical breakage, which occurs before intercourse, typically when the condom is being put on. Clinical breakage rates in prospective studies vary between 0.5% and 3.7%; in retrospective studies, total condom breakage rates range between 0.6% and 7.9%.4 Complete condom slippage, where the condom falls completely off the penis during intercourse or withdrawal, varies from 0.6% to 5.4% in prospective studies and 0.3% to 5.1% in retrospective studies.4 One prospective study, performed in prostitutes, reflects very low breakage (0%) and clinical slippage (0.6%) rates, suggesting that condoms can provide very reliable performance.4 Condom performance may depend on use of oil-based lubricants or additional water-based lubricant; condom brand, age, and storage conditions; and user age, condom experience, and marital status.7-‘0 Subjective considerations associated with increased rates of breakage include vigorous or intense sex, inadequate lubrication, and tearing with fingernails.’ The majority of condom studies are retrospective, meaning that subjects rely on memory of events that may have taken place months earlier. Prospective
C
ISSN OOIO-7824/97/$17.00 PII SOOIO-7824(97)00069-3
18
Rosenberg
Contraception 1997;56:17-21
and Waugh
studies often have relatively small sample sizes, frequently including fewer than 50 subjects4 Because few studies have been conducted with larger numbers of typical users as part of a controlled clinical trial, we conducted a multicenter clinical study using a single brand of male latex condom.
l
l
Methods Study participants were monogamous couples (males 18-50 years old; females 18-40 years old) who had sex at least six times a month and did not have a history of fertility-impairing conditions, including STIs. Couples were at risk of pregnancy, and condoms were their sole method of birth control during the study. Female partners were tested for pregnancy at study entry, then again at study exit. Other clinical evaluations were performed at the request of the participants or as clinically indicated (cervical cytology, wet mount, ST1 tests, and physical examination). Pregnancy tests could be requested by the patient at any time, and postcoital contraception was available on request. The study protocol was approved by our institutional review board and participants provided informed consent before entry. Enrolled couples used lubricated Durex Ramses (London International Group, London, England) brand male latex condoms. Detailed information was recorded on the first five coital acts. At each sexual encounter, a diary was completed that included information on condom use, breaks or slips, as well as type of sexual activity (vaginal vs. anal), and use of additional lubrication. In addition, male partners were provided with a kit by which penis length and circumference were measured. Couples were provided addressed, stamped envelopes to return diaries on a monthly basis, and returned to the clinic after 3 months to review forms and receive additional condoms as needed. Couples were contacted by telephone if forms were not received within 2 weeks of the expected monthly submission date, and also to schedule the 3 month interim visit as well as the exit interview. Couples were followed for 6 months, after which time the final pregnancy test and exit interview were conducted. Statistical analysis included comparison of condom breakage and slippage rates for couples that reported at least one attempt to use the condom, with exact 95 % confidence intervals for binomial data. l1 Terms used for this evaluation are l
l
Clinical breakage rate: breaks that occur during intercourse or that are only noted upon withdrawal divided by the total number of acts of intercourse; Nonclinical breakage rate: breaks that occur before
l
l
intercourse divided by the total attempts to use the condom; Total breakage rate: the sum of clinical and nonclinical breaks divided by the total number of attempts to use the condom; Complete slippage rate: events where the condom comes completely off the penis during intercourse or withdrawal divided by the total acts of intercourse; Total clinical failure rate: the sum of clinical breaks and complete slips divided by the total acts of intercourse; Total failure rate: the sum of total breaks and complete slips divided by the total number of attempts to use the condom.
Results Study participants were relatively young and primarily white (Table I). Nearly all subjects were either married or had never been married, in approximately equal proportions. Slightly less than half the women were parous. Nearly all participants completed high school, and slightly over half had also graduated from college or had completed some college. Three-quarters of subjects lived with their current partner or spouse. The group was sexually experienced, having been involved in a sexual relationship with their study partner for an average of four years. Approximately 70-75% of the study participants were experienced condom users, having used a condom more than 50 times with their current partner, while only approximately 8% had used a condom 10 or fewer times with their current partner. Two men and two women reported no previous condom experience, but did not have any breakage or slippage during the current study. Of the 92 couples that contributed data to this analysis, 66% completed 6 months of follow-up, 18% were lost to follow-up, 2% discontinued for condomrelated reasons (complaints of breakage, discomfort), and the remaining 14% discontinued for reasons unrelated to the study condom (no longer with partner, study requires too much time, moving out of the area, etc.). Those that discontinued early or were lost to follow-up were not different sociodemographically from those that completed 6 months of follow-up. A total of 4637 attempts to use the condom were reported (Table 2). Six of these resulted in nonclinical breaks, leaving 4631 condoms that were used for intercourse. The overwhelming majority of intercourse was vaginal; only four acts of anal intercourse were reported, in one case using the same condom for vaginal intercourse, and none resulted in breakage or
Contraception 1997;56:17-21
Table
Condom
Breakage
and Slippage
in a Controlled
Clinical
Trial
19
1. Characteristics of study participants
Characteristic Mean age (SD), years Ethnicity White Hispanic Black Native American Other Asian/Pacific Islander Marital status Single Married Divorced Separated Widowed Education level Did not complete high school Still in high school High school graduate Some college, no longer enrolled Still in college College graduate Any postgraduate work Living arrangements Live with current partner Live with spouse Live with parents Live alone Live with roommate(s) Employment status Full time Part time Student Unemployed Homemaker Total household income $O-10,000 $lO,OOl-20,000 $20,001-30,000 $30,001-40,000 $40,001-50,000 $50,001-60,000
Greater than $60,001 Don’t know/prefer not to answer Years in sexual relationship with study partner (mean, median) Age at first intercourse (mean, median) Lifetime no. sexual partners (mean, median) Parous Yes No Previous condom use with study partner Never l-2 times 3-10 times 1l-50 times >50 times Previous condom use with any partner Never l-2 times 3-10 times 1l-50 times >50 times
Male partner (n = 92), %
Female partner (n = 92), %
28.8 (6.2)
26.8 (5.5)
75.0 12.0 7.6 4.3
75.0 7.6 8.7 2.2 2.2 4.3
1.1 45.7 46.7 7.6 -
46.7 46.7 5.4
2.2 3.3 19.6 28.3 19.6 22.8 4.3
3.3 10.9 29.3 27.2 22.8 6.5
38.0 39.1 7.6 5.4 9.8
34.8 41.3 6.5 4.3 13.0
63.0 17.4 115.2 4.3 -
44.6 20.7 21.7 4.3 8.7
9.8 17.4 30.4 10.9 8.7 5.4 7.6 9.8 4.0, 3.1 16.8, 16.0 15.2, 7.0
-
1.1
9.8 17.4 27.2 15.2 10.9 4.3 7.6 7.6 4.0,3.1 16.9, 17.0 12.1, 5.5
-
50.0 50.0
2.2
2.2
1.1
1.1
2.2 25.0 69.6
2.2 23.9 70.7
2.2
2.2
1.1 5.4 20.7 70.7
1.1 3.3 18.5 75.0
20
Rosenberg
and Waugh
Contraception 199756: 17-21
Table 2. Condom usage Vaginal intercourse only Anal intercourse only Anal and vaginal intercourse Acts of intercourse First five usesonly All subsequentuses Total acts Number acts per couple (mean, SD) Attempts to use condom First five attempts All subsequentattempts Total attempts to use condom
4626 4 1 456
4175 463 1 50.3, 41.0 457
4180 4637
slippage. Thirteen clinical breaks occurred, resulting in a clinical breakage rate of 0.28%, and a total breakage rate of 0.41% (Table 3). Nine couples broke one condom each, and two couples broke two condoms each. Twenty-nine complete slips were reported, resulting in a slippage rate of 0.63%. Of these complete slips, 21 were reported by a single couple. Total clinical failure, at 42 events, was 0.91%.
Discussion Less than 1 out of every 100 condoms used failed in a way that might place a user at risk of pregnancy or STI. These breakage and slippage rates are similar to those in a recent study of condom performance among female prostitutes and are lower than all others in the published literature.4 These results, recorded among couples that were sexually active, for the most part experienced with condom use, and monogamous, suggest that condoms can provide an extremely reliable method of contraception and ST1 protection in everyday practice. The very small number of events, however, makes it impossible with the current study to explore risk factors, including user characteristics, sexual activities, and penis size. These results may be more broadly generalizable than other condom evaluations because of the population we studied. Many prospective clinical trials include subjects who may not be typical of the end user population. l2 This is often a concern with condom studies, which exclude individuals at risk of becoming pregnant or acquiring a STI.2~12,‘3 The absence of such risk may alter sexual practices or behaviors. Since participants in this study were at risk of unintended pregnancy and possibly of STI, couples would have remained strongly motivated to use the condom correctly. Though self-selection is an issue in any clinical trial, including this one, this form of bias is difficult to quantify. The characteristics of users in this study, including diverse geographical distribution, thus suggest that our results are
generalizable to monogamous couples using condoms as their primary method of birth control. Also impacting generalizability of results is the size of the study. The number of acts of intercourse (4631) is considerably larger than that in most studies, and the number of participants in this study, 92, is in the mid-range of other prospective studies of condom breakage and slippage. Smaller sample size studies of condom breakage and slippage can be adversely affected by the inclusion of one or more couples that are prone to breakage and slippage; that certain couples are prone to such problems has been noted in other studies.9J4,1 5 One such couple in this study reported 21 of the 29 complete slips. Removing this couple results in a clinical failure rate of 0.46%, or one-half of that currently reported. In a study with fewer couples and acts of Table 3. Condom performance
Condom breakage First five usesonly Nonclinical breaks Clinical breaks Total breaks All subsequentuses Nonclinical breaks Clinical breaks Total breaks All usescombined Nonclinical breaks Clinical breaks Total breaks Condom slippage First five usesonly, complete slips All subsequent uses, complete slips All usescombined, complete slips Condom failure First five usesonly Clinical failure Total failure All subsequentuses Clinical failure Total failure All usescombined Clinical failure Total failure
Events
Rate (%)
95% CI
1 5 6
0.22 1.10 1.31
0.01-l .21 0.36-2.54 0.48-2.84
2 13
0.12 0.19 0.31
0.04-0.28 0.08-0.38 0.17-0.53
6 13 19
0.13 0.28 0.41
0.05-0.28 0.15-0.48 0.25-0.64
3
0.66
0.14-1.91
26
0.62
0.41-0.91
29
0.63
0.42-0.90
9 9
1.75 1.97
0.76-3.43 0.90-3.71
34 39
0.81 0.93
0.56-l. 14 0.66-1.27
42 48
0.91 1.04
0.65-1.22 0.76-l .37
Nonclinical breaks are breaks before intercourse divided by attempts to use condom. Clinical breaks are breaks during intercourse or only noted upon withdrawal divided by acts of intercourse. Total breaks are clinical and nonclinical breaks divided by attempts to use condom. Complete slips are slips completely off the penis divided by acts of intercourse. Clinical failure is clinical breaks and complete slips divided by acts of intercourse. Total failure is clinical breaks, nonclinical breaks, and complete slips diwded by attempts to use condom.
Contraception 1997;56:17-21
Condom
intercourse, this one couple would have had an even larger effect on interpretation of condom performance, greatly limiting the generalizability of study findings. Thus, the size of the study reported here increases our confidence in generalizability of its findings. Among study limitations was the small number of breaks or slips, which precluded our exploring other factors for their association with condom performance. These other factors include use of additional lubricants, particularly if oil-based, age and storage conditions of the condom, user age, user experience with condoms, marital status, vigorous sex, inadequate lubrication, tearing with fingernails, and condom brand.“‘O We eliminated the potential variable of differences in condom brands by using a single, commonly available condom. The excellent condom performance reflected in this study may thus reflect the condom, users, or some combination of these factors in addition to others. The use of a single brand of condom means that our results may not be generalizable to others. An additional concern is loss to follow-up and early drop-outs due to problems using the condom. At 19% (18% loss to follow-up and 1% who cited excessive breakage as discontinuation reason), while consistent with other studies of similar duration, it raises the possibility that if couples dropped out early because of problems they experienced with the condom, then the results we report actually underestimate true breakage and slippage rates; that is, study results may be based on acts of intercourse the large majority of which are contributed by users who have little to no problems with breakage or slippage. As a sensitivity analysis, we compared rates of clinical breakage and complete slippage for couples lost to follow-up or who discontinued early citing condom-related problems with those who completed the study. This group reported three clinical breaks and two complete slips in 348 acts of intercourse, resulting in clinical breakage and slippage rates of 0.86% and 0.57%, respectively. For these early dropouts, the total clinical failure rate is 1.43%. Breakage rates are, thus, slightly higher than those reported by couples that completed the investigation (clinical breakage rate for this group was 0.23%); this difference approaches but does not achieve statistical significance (Fisher’s exact test, p = 0.07). Complete slippage rates are slightly higher for couples who finished the study (0.63%) and lower for total clinical failure (0.86%), but neither of these differences are statistically significant. Based on this sensitivity analysis, it appears that if the couples lost to follow-up or who dropped out because of problems with the condom had continued in the study to completion, study findings would not have changed significantly; results would still show condom failure to be an extremely uncommon event among this population of users.
Breakage
and Slippage
in a Controlled
Clinical
Trial
21
Overall, this study indicates that lubricated Durex Ramses latex condoms can provide very reliable performance in a group of motivated, experienced users. Although results presented here of breakage and slippage do not translate directly into risk of pregnancy or STI, the rates-lower than any other reported thus far in a group of typical couples and more reliable because of the number of users-are suggestive that efficacy may compare favorably with those of other methods such as oral contraceptives that have been considered among the most reliable forms of userdependent contraception.16
References 1. London KA, Mosher WD. What is Current Use of a
Contraceptive? Hyattsville, MD: National Center for Health Statistics, 1990. 2. Trussell J, Warner DL, Hatcher RA. Condom slippage and breakagerates. Fam Plann Perspect 1992;24:203. 3. Grady WR, Klepinger DH, Billy JO, Tanfer K. Condom characteristics:the perceptionsand preferencesof men in the United States. Fam Plann Perspect 1993;25:67-73. 4. Albert A, Warner DL, Hatcher RA, Trussell J, Bennett C. Condom use among female commercial sex workers in Nevada’s legal brothels. Am J Public Health 1995j85: 1514-20. 5. Russell-Brown P, Piedrahita C, Foldesy R, Steiner M,
Townsend J. Comparison of condom breakage during human use with performance in laboratory testing. Contraception 1992;45:429-37. 6. Steiner M, Foldesy R, Cole D, Carter E. Study to determine the correlation between condom breakage in human use and laboratory test results. Contraception 1992;46:279-88. 7. Steiner M, et al. The impact of lubricants on latex condoms. Int J STD AIDS 1994;5:30-6. 8. Steiner M, Piedrahita C, Glover L, Joanis C. Can condom users likely to experience condom failure be identified. Fam Plann Perspect 1993;25:220-6. 9. SparrowMJ, Lavill K. Breakageandslippageof condomsin family planning clients. Contraception 1994j50:117-29. 10. Joanis C, Brookshire T, Piedrahita C, Steiner M. Evaluation of Two Condom Designs: a Comparison of Standard and Larger Condoms in Ghana, Kenya, and Mali. Research Triangle Park: Family Health International, 1990. 11. StatXact version 2.0. Cambridge: Cytel Software, 1992. 12. Grady WR, Tanfer K. Condom breakage and slippage among men in the United States. Fam Plann Perspect 1994;26:107-12. 13. Trussell J, Warner DL, Hatcher RA. Condom perfor-
mance during vaginal intercourse: comparison of Trojan-Enz and Tactylon condoms. Contraception 1992j45: 1 l-9. 14. Free MJ, Skiens EW, Morrow MM. Relationship be-
tween condom strength and failure during use. Contraception 1980;22:31-7. 15. Richters J, Donovan B, Gerofi J. How often do condoms break or slip off in use?Int J STD AIDS 1993;4:90-4, 16. Hatcher RA, Trussell J, Stewart F, et al. Contraceptive Technology. New York: Irvington Publishers, 1996: 107-43.