Contraceptive Efficacy of a Novel Spermicidal Microbicide Used With a Diaphragm A Randomized Controlled Trial Kurt T. Barnhart, MD, MSCE, Michael J. Rosenberg, MD, MPH, H. Trent MacKay, MD, MPH, Diana L. Blithe, PhD, Jim Higgins, PhD, Terri Walsh, MPH, Livia Wan, MD, Michael Thomas, Mitchell D. Creinin, MD, Carolyn Westhoff, MD, William Schlaff, MD, David F. Archer, MD, Charletta Ayers, MD, Andrew Kaunitz, MD, Sutapa Das, MS, and Thomas R. Moench, MD
MD,
OBJECTIVE: Women need products that protect against both pregnancy and sexually transmitted infections, including human immunodeficiency virus (HIV). The acid buffer-
ing gel is a nondetergent spermicide that may provide this dual protection by reinforcing normal vaginal acidity to inactivate both sperm and acid-sensitive sexually transmitted pathogens. The objective of this study was to assess the gel’s contraceptive effects, safety, and acceptability.
From the University of Pennsylvania, Philadelphia, Pennsylvania; Health Decisions, Inc, Chapel Hill, North Carolina; National Institute of Child Health and Human Development, Bethesda, Maryland; California Family Health Council Los Angeles, California; New York University, New York, New York; University of Cincinnati, Cincinnati, Ohio; University of Pittsburgh and Magee-Women’s Research Institute, Pittsburgh, Pennsylvania; Columbia University, New York, New York; University of Colorado Health Sciences Center, Denver, Colorado; CONRAD Clinical Research Center, Eastern Virginia Medical School, Norfolk, Virginia; University of Medicine and Dentistry of New Jersey, New Brunswick, New Jersey; University of Florida Health Science Center, Jacksonville, Florida; and ReProtect, Inc, Baltimore, Maryland.
METHODS: We conducted a multicenter, randomized, double-masked, noninferiority study at 11 centers, comparing 621women who used an acid buffering gel plus diaphragm with 300 women who used a nonoxynol-9 spermicide plus diaphragm for 6 months. A doublemasked study extension followed 234 women for an additional 6 months of use.
Funded by the National Institute of Child Health and Development (NICHD) Contraceptive Clinical Trials Network by contract NIH N-01-HD-6-3261 from the National Institutes of Health (NIH) and, in part, by NIH General Clinic Research Center Grants MO1RR000056 at the University of Pittsburgh and M01RR00096 at New York University. BufferGel was provided by ReProtect, Inc (Baltimore, Maryland). Gynol II and OrthoFlex diaphragms were provided by Johnson & Johnson (New Brunswick, NJ). The authors thank the co-investigators, Ron Frezieres, Anita Nelson (California Family Health Council, UCLA, Harbor-UCLA Medical Center); Bryna Harwood, Michelle Fox, Courtney Schreiber, Amitasrigowri Murthy (University of Pittsburgh); Susan Ballagh (CONRAD Clinical Research Center, Eastern Virginia Medical School, Norfolk ,VA) for reviewing and modifying the study protocol, conduct of the study according to good clinical practice, review and interpretation of the data, and participation in the writing of the manuscript. Presented in part at Microbicides 2006, April 23–26, 2006, Capetown, South Africa. Corresponding author: Dr. Kurt Barnhart, Department of Obstetrics-Gynecology and Center for Clinical Epidemiology and Biostatistics University of Pennsylvania Medical Center, Penn Fertility Care, 3701 Market Street, 8th Floor, Philadelphia, PA 19104; e-mail:
[email protected]. Financial Disclosure Dr. Barnhart received research funding from Personal Care Products (Raritan, NJ) and Johnson & Johnson (Raritan, NJ). Dr. Kaunitz consults with and owns stock in Johnson & Johnson. Dr. Moench is employed by and holds equity in ReProtect, Inc (Baltimore, MD) and is an inventor on BufferGel patents assigned to ReProtect, Inc. The other authors have no potential conflicts to disclose. © 2007 by The American College of Obstetricians and Gynecologists. Published by Lippincott Williams & Wilkins. ISSN: 0029-7844/07
VOL. 110, NO. 3, SEPTEMBER 2007
RESULTS: The 6-month pregnancy rate per hundred women was 10.1% (95% confidence interval [CI] 7.1– 13.1%) for acid buffering gel and 12.3 (95% CI 7.7–16.9) for nonoxynol-9 spermicide users. The difference in rates was –2.2% with a 95% CI –7.7 to 3.3%. Consistent and correct use 6-month pregnancy rates were 4.7% for acid buffering gel and 6.1% for nonoxynol-9 spermicide users, calculated from those cycles where diary entries indicated such use. Adverse events and acceptability were similar between the two groups. Pregnancy probabilities were similar between groups participating in the 12month study extension. CONCLUSION: An acid buffering gel used with a diaphragm is a safe, acceptable contraceptive with efficacy comparable to that of a common commercial spermicide with diaphragm. CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov, www.ClinicalTrials.gov, NCT00065858 (Obstet Gynecol 2007;110:577–86)
LEVEL OF EVIDENCE: I
F
emale barrier contraceptives are important contraceptive options that have been available for more than a century. Methods available today include
OBSTETRICS & GYNECOLOGY
577
spermicides, female condoms, diaphragms, and cervical caps, of which the latter two are typically used in combination with a spermicide. Advantages of these methods include availability over the counter, local rather than systemic action, rapid reversibility, and female control of the method. Moreover, vaginal methods may provide women with dual protection, that is, protection against pregnancy and sexually transmitted infections, by functioning as both spermicides and microbicides.1 Currently available spermicides are all based on detergents, most commonly nonoxynol-9, that potently disrupt the cell membranes of sperm as well as those of human immunodeficiency virus (HIV) and other sexually transmitted pathogens.2,3 However, detergents can also disrupt epithelial cells of the cervix and vagina, compromising protective cell layers and causing inflammation.2,4 Trials of nonoxynol-9 have demonstrated inconsistent protection against HIV and other sexually transmitted infections,5 including possible increased risk of HIV in some studies of female sex workers.6,7 Accordingly, nonoxynol-9 has been abandoned as a microbicide candidate2,8 and has been judged by some to be inappropriate even as a contraceptive for women at high risk for HIV or other sexually transmitted infections and women who have multiple daily acts of intercourse.8 Thus, there is a need for safer spermicides, especially ones with potential utility as vaginal microbicides. BufferGel (ReProtect, Inc, Baltimore, MD) is a nondetergent spermicide that reinforces the mild acidity (pH 4) of the normal human vagina,10,11 a natural vaginal protective mechanism.9 It contains water, 3.8% carbomer 974P as both a buffering and gelling agent, sorbic acid as a preservative, ethylenediaminetetraacetic acid (EDTA), and physiologic salts (K2HPO4, NaH2PO4, MgSO4 ␥ 7H2O) to achieve physiologically appropriate tonicity, viscosity, and pH. The acid buffering gel is formulated at pH 3.85, and a 5-g dose can maintain a pH less than 5.0 after addition of 0.79 mEq of NaOH. This buffering capacity is sufficient to block the alkalinizing action of semen that otherwise abolishes vaginal acidity for hours after intercourse.12 By acidifying semen, the acid buffering gel reinforces normal vaginal acidity to inactivate sperm and acid-sensitive sexually transmitted pathogens including HIV.13–15 Women were instructed to dispense 5 g of gel into the concavity of the diaphragm not more than 2 hours before intercourse and to apply an additional applicator full of gel into the vagina for each additional act of intercourse occurring with the diaphragm still in place. The acid
578
Barnhart et al
Efficacy of a Nondetergent Spermicide
buffering gel used without a diaphragm is currently being studied as a vaginal microbicide in a multicenter HIV prevention trial (www.clinicaltials.gov NCT00074425). We conducted a multicenter, randomized, double-masked noninferiority trial comparing the acid buffering gel and Gynol II (2%), a nonoxynol-9 spermicide (Ortho Pharmaceutical, Raritan NJ), both used with a diaphragm to evaluate the contraceptive efficacy, safety, and user issues. Pregnancy was the primary outcome of the trial, and secondary outcomes included safety, incidence of vaginal infection, and user acceptability.
MATERIALS AND METHODS The study was conducted at 11 centers in the United States between October 2001 and August 2004. The protocol was approved by each organization’s institutional review board, and the study was monitored by the Data and Safety Monitoring Board that met twice during the study. Eligibility requirements included being a sexually active female, 18 – 40 years old, at risk for pregnancy and desiring contraception, and having normal (24- to 35-day) menstrual cycles. Participants needed to be willing to engage in at least four acts of sexual intercourse per month, use the test products as their primary method of contraception, accept a moderate risk of pregnancy, and keep a diary of coital activity, product use, and adverse events. Subjects also needed to be at low risk for HIV or other sexually transmitted infection, which was defined as having a single sexual partner (for at least 4 months before participating in the study), who also would be at low risk of sexually transmitted infection, and expecting to remain with this partner for the entire study. Subjects were ineligible if they had an allergy or sensitivity to spermicides or latex, a history of toxic shock syndrome, two or more urinary tract infections (UTIs) in the past year, a history of infertility, any contraindication to pregnancy, abnormal cervical cytology within the past 12 months, or a sexually transmitted infection diagnosed within the previous 6 months. This study was designed as a noninferiority trial. Sample size was determined by using binomial distribution assumptions to determine power. We defined a pregnancy percentage in the acid buffering gel group no greater than 6 percentage points of that in the nonoxynol-9 spermicide group as noninferior. A difference of 6% was chosen as clinically significant and is equal to differences evaluated in similar trials.16 We chose a 2:1 ratio of recruitment to obtain more information about adverse effects and acceptability in
OBSTETRICS & GYNECOLOGY
the acid buffering gel group. Accordingly, the target enrollment was 650 and 325 in the acid buffering gel and nonoxynol-9 spermicide groups, respectively (nQuery Advisor 3.0; Statistical Solutions, Saugus, MA), which allowed for a pregnancy rate ranging from 8% to 12% and a power ranging from 85% to 94%, depending on pregnancy rate. After providing informed consent, each participant was interviewed and underwent a pelvic examination, including a Pap test, cervical swab for Chlamydia trachomatis and Neisseria gonorrhoeae, vaginal smear for Gram stain, wet mount (for bacterial vaginosis, yeast, and trichomonas), dipstick urinalysis, urine culture, and urine pregnancy test. Colposcopy17 and quantitative vaginal microbiology studies were performed on subsets of women who participated at centers with specialized expertise in performing these evaluations. Each woman returned after the results of all screening evaluations were available and eligibility confirmed. At this visit, she was fitted with a diaphragm and randomly assigned to acid buffering gel or nonoxynol-9 spermicide in blocks of six. The randomization schedule was generated by statisticians at Health Decisions Incorporated, the clinical trial monitor for the study. Treatment assignments were randomly permuted within each block by using a random number generator. The blinded kits of drug within each block were numbered sequentially. The site distributed kits in numerical order to subjects as they were randomized into the study. Each participant was provided a supply of assigned gel and a diaphragm (Ortho All-Flex; Ortho Pharmaceutical, Raritan NJ). Study products were repackaged for this study to ensure masking of the products to the participants and to the study personnel. The study products had similar color, texture, viscosity, and smell, were packaged in indistinguishable unlabeled tubes, and were distributed with identical cartons, study labels, and applicators. Investigators were blinded to the arm of the study to which patients were assigned when they were seen in their follow-up visits Two weeks after admission, participants were contacted by telephone to inquire about any problems with the diaphragm and test product and to assess compliance. Participants were scheduled to return to the study site within 2 weeks after cycles one, three, and six. At each visit, a gynecologic examination was performed, which included a wet mount and vaginal Gram stain. Urine was collected for a pregnancy test and dipstick urinalysis. After cycle one, if the dipstick urinalysis was positive for blood, leuko-
VOL. 110, NO. 3, SEPTEMBER 2007
cyte esterase, protein, or nitrites, the urine was sent to the laboratory for urine culture and microscopic analysis. At the visit after cycle three, testing for chlamydia and gonorrhea and a urine culture were performed. At the visits that followed cycles one and three, compliance with test product was assessed and additional test product and coital diaries were distributed. At all visits, women were reminded to contact the study center if they suspected they might have a UTI (to provide urine for culture) and any time they suspected that they might be pregnant to undergo further evaluation. All women were offered the use of emergency contraception if needed during the study. A final visit was scheduled after 183 days of diaphragm and product use, which was typically after cycle six. At this visit, any remaining study product was collected. Urine was collected for a pregnancy test, dipstick urinalysis, and culture. Participants underwent a gynecologic examination that included a Pap test, chlamydia and gonorrhea testing, vaginal smear for Gram stain, and wet mount. Each woman completed an acceptability questionnaire regarding her experience with the study method. An extension arm was added to the study after it had already begun, and from that point forward, all women completing the 6-month primary follow-up were offered the opportunity to continue for an additional 6 months. Participants were seen after cycles 9 and 12, with procedures for these visits as at cycles 3 and 6, respectively. The primary outcome was contraceptive efficacy. Secondary outcomes included safety; incidence of vaginal or cervical lesions and irritation; incidence of UTIs, bacterial vaginosis, symptomatic yeast vaginitis, asymptomatic vaginal yeast colonization, vaginal Escherichia coli colonization; and product acceptability. Subset analyses were conducted on results of colposcopy evaluations, quantitative vaginal cultures, and fungal cultures. An adverse event was a change in health status from that at baseline, regardless of its possible relationship with the product, consistent with good clinical practice guidelines. Three populations were defined for analysis. The intent-to-treat population included all women who were allocated to intervention. Side effect, adverse event, and acceptability results are based on the all-treated population, which was the group of women who used the product at least once. Primary analysis of contraceptive efficacy is based on a modified intent-to-treat population. This population is defined as randomized participants who reported at least one act of coitus while using the assigned study product and for whom there was at least one report of
Barnhart et al
Efficacy of a Nondetergent Spermicide
579
Fig. 1. Patient flow diagram. The top portion addresses women screened for the trial. The lower portion describes women randomized. *The denominator for these percentages was the total number of subjects screened. †The denominator for these percentages was the total number of subjects allocated to intervention (ITT population) in each treatment group. The only exceptions are the percentages for the allocated to intervention line. Those percentages are calculated by using the total number of subjects randomized as the denominator. ITT, intent-to-treat; MITT, modified intent-to-treat. Barnhart. Efficacy of a Nondetergent Spermicide. Obstet Gynecol 2007.
pregnancy status. Correct and consistent use pregnancy rates were calculated from cycles in which the study product was used correctly for every coital act, compressing these cycles to provide contiguous cycles of correct use. Cycles in which participants used a backup method (other than the study gel plus diaphragm) or emergency contraception were excluded. Any pregnancies that started before randomization (based on serum pregnancy test and estimated date of ovulation) were excluded from the modified intentto-treat and correct and consistent use analyses. Safety was assessed using the all-treated popula-
580
Barnhart et al
Efficacy of a Nondetergent Spermicide
tion. For the 6-month (183 days) analysis, a subject contributed data until the earlier of the following events: conception, 183 days of participation, or the latter of the last product usage date and the last date a diary entry was recorded. For the 12-month (365 days) analysis, a subject contributed data until the earlier of the above listed events used for the 6-month analysis, except that 365 days of participation was used instead of 183 days. Data from partial cycles at the beginning and the end of study participation were included in the analysis. Pregnancies were classified by using the estimated date of ovulation or concep-
OBSTETRICS & GYNECOLOGY
tion recorded. For analyses done by month (days), the estimated ovulation date was compared with the date of enrollment to determine when the pregnancy occurred. The primary hypothesis was that the cumulative pregnancy rate at 6 months (183 days) in subjects using the acid buffering gel and a diaphragm is not inferior to that of subjects using the nonoxynol-9 spermicide and a diaphragm. Using Blackwelder’s approach18 for equivalence testing, the null hypothesis is HO:BⱖG⫹0.06, where B and G represent the 6-month cumulative pregnancy probabilities for two groups, respectively. Kaplan-Meier methods were used to estimate the respective 6-month cumulative pregnancy probabilities, with standard errors to construct confidence intervals calculated by the method of Peto et al.19
RESULTS The number, allocation, and follow-up information on the women enrolled in the trial are summarized in Figure 1. Participants were predominantly young, white, unmarried, nonsmokers who lived with their partners. Approximately three fourths of participants relied on their partners’ use of the male condom as their primary contraceptive in the prior 6 months; most had never used a diaphragm or a spermicide alone. Most had been pregnant at least once, and approximately one third had undergone elective abortion. Slightly more than half of subjects had no children. These and other characteristics were similar between the acid buffering gel and nonoxynol-9 spermicide groups (Table 1). The 6-month pregnancy probability was 10.1% (95% CI 7.1–13.1%) for acid buffering gel and 12.3% (95% CI 7.7–16.9%) for nonoxynol-9 spermicide users (Table 2). The difference in probabilities (acid buffering gel – nonoxynol-9 spermicide) had a one-sided upper 97.5% confidence bound of 3.3, which was less than the noninferiority delta of 6. In participants whose diary entries indicated cycles of correct and consistent use, 6-month pregnancy probabilities were 4.7% (95% CI 0.9 – 8.4%) and 6.1% (95% CI 0.2– 12.0%), respectively. Cumulative pregnancy probability, presented as a life-table analysis, is shown in Figure 2. This analysis of the modified intent-to-treat population is based on 2,836 months of exposure for acid buffering gel and 1,381 months of exposure for the nonoxynol-9 spermicide. Participants reported a mean of 13.0 and 12.8 coital acts monthly in the acid buffering gel and nonoxynol-9 spermicide groups, respectively, and an average rate of product use of 8.9 per month for
VOL. 110, NO. 3, SEPTEMBER 2007
Table 1. Characteristics of Participants
Characteristic Age (y, mean⫾SD) Race Asian African American Native American White Native Hawaiian/ Pacific Islander Other Ethnicity Hispanic or Latino Marital status Never married Married Separated Divorced Living arrangement Living with partner Not living with partner Smoking status Current smoker Former smoker Weight (lb, mean) Contraceptive use (past 6 mo) Male condom Diaphragm Oral contraceptive pill Spermicides Rhythm method Withdrawal Abstinence Never used diaphragm Never used spermicide Number of pregnancies 0 1 2 3 or more Elective abortion (ever) Number of living children 0 1 2 or more
Acid Buffering Gel (nⴝ621)
Nonoxynol-9 Spermicide (nⴝ300)
27.7⫾5.8
28.4⫾5.8
5 23 1 60
5 27 0 58
1 11
1 8
20
17
60 33 2 5
59 32 3 7
68
66
33
34
24 20 162.3
20 25 168.0
77 8
79 8
19 24 13 40 7 80 58
16 22 17 48 9 82 56
33 26 17 25 36
29 24 17 29 39
57 22 21
53 18 29
SD, standard deviation. Data are expressed as % unless otherwise indicated.
both products. Emergency contraception was used by 44 (6%) women in the acid buffering gel group and 26 (7%) women in the nonoxynol-9 spermicide group. Experience was also similar between the two groups with regard to vaginal culture, colposcopy, and user symptoms (Table 3), with the exception of
Barnhart et al
Efficacy of a Nondetergent Spermicide
581
Table 2. Pregnancy Outcomes Acid Buffering Gel
Population 6-Month cumulative probability Intent-to-treat* Modified intent-totreat† Correct and consistent use‡ 12-Month cumulative probability Modified intent-totreat† Correct and consistent use‡
Number of Pregnancies
Pregnancy Probability (%)
66
Nonoxynol-9 Spermicide
95% CI
Number of Pregnancies
Pregnancy Probability (%)
10.3
7.7–12.8
35
51
10.1
7.1–13.1
12
4.7
64 12
95% CI Difference
95% CI
11.0
7.3–14.6
–0.7
–5.1 to 3.8
30
12.3
7.7–16.9
–2.2
–7.7 to 3.3
0.9–8.4
6
6.1
0.2–12.0
–1.4
–8.4 to 5.6
16.7
10.2–23.1
36
17.0
8.2–25.9
–0.3
–11.3 to 10.6
4.7
0.0–14.2
6
6.1
0.0–18.6
–1.4
–17.1 to 14.4
CI, confidence interval. * Intent-to-treat: all women who were allocated to intervention. † Modified intent-to-treat: randomly assigned subjects with at least one act of coitus while using the assigned study product and for whom there was at last one report of pregnancy status. ‡ Correct and consistent use: cycles in which the study product was used correctly with all coital acts.
incidence of symptomatic UTIs (9% compared with 14%, P⫽.03) and the rate of recovery of Ureaplasma urealyticum (20% compared with 43%, P⫽.04), which were significantly lower in acid buffering gel compared with nonoxynol-9 spermicide users. Acceptability was also similar between groups, with approximately two thirds of women in both groups reporting strongly or somewhat liking the method. A much smaller proportion, 17% and 14% in acid buffering gel and nonoxynol-9 spermicide groups, respectively, reported somewhat or strongly disliking the method. Partner-use issues were infrequent and relatively equally distributed between the two comparison groups. Adverse events that were considered possibly or probably related to the product were similar in both groups in terms of overall frequency and the occurrence of most frequent adverse events (Table 4). Of all reported adverse events, 24% were considered by the investigator to be probably or highly probably related to study product in the acid buffering gel group and 25% in the nonoxynol-9 spermicide group (P⫽.54). Common events included vaginal symptoms, as well as general complaints such as upper respiratory infection not believed to be related to product use. A single serious adverse event, a thigh abscess, occurred in a user of acid buffering gel, but this was judged by the investigator to be unlikely to be related to product.
582
Barnhart et al
Efficacy of a Nondetergent Spermicide
DISCUSSION This randomized controlled trial is the first to assess the contraceptive efficacy of a nondetergent spermicide, an agent that may potentially also be microbicidal (PubMed search using key words “microbicide,” “spermicide,” and “contraceptive gel,” restricted to English language and dates 1965 to present). Strengths of this trial include its rigorous design, successful recruitment, and detailed follow-up assessment. The randomized and blinded design allowed an
Fig. 2. Cumulative pregnancy probability using the modified intent-to-treat population. Dotted line, acid buffering gel; solid line, nonoxynol-9 spermicide. Barnhart. Efficacy of a Nondetergent Spermicide. Obstet Gynecol 2007.
OBSTETRICS & GYNECOLOGY
Table 3. Secondary Outcomes Outcome Infections (cumulative from admission through cycle 6) Symptomatic urinary tract infection All urinary tract infections Bacterial vaginosis (based on Nugent’s scoring) Symptomatic yeast vaginitis Irritation (cumulative from admission through cycle 6) Participant (self-reported) Partner irritation Pap test, change since admission Improved Worsened † Vaginal culture at cycle 6 Gardernerella vaginalis Lactobacillus spp (H202⫹) Escherichia coli Ureaplasma urealyticum Candida albicans‡ Colposcopy (percentage of women with lesions)§ Baseline Cycle 1 Cycle 3 Cycle 6 Acceptability (at cycle 6)储 Overall impression of method Liked strongly Liked somewhat Neutral Disliked somewhat Dislike strongly Definitely or probably use product if available Prefer diaphragm and product to condoms
Acid Buffering Gel (nⴝ621)
Nonoxynol-9 Spermicide (nⴝ300)
P*
9 16 27 6
14 18 27 8
.03 .30 .94 .43
35 12
34 10
.82 .42
6 6
4 7
.33
32 77 23 20 20
23 82 14 43 13
.44 .66 .37 .04 .51
42 36 29 29
40 47 37 29
.84 .19 .39 .97
24 44 16 13 4 68 54
24 48 14 12 2 70 57
.66
.56 .50
Data are expressed as %. * All P values were obtained from the Cochran-Mantel-Haenzel statistic for general association to compare the two treatment groups. † n⫽59 for acid buffering gel and n⫽23 for nonoxynol-9 spermicide. ‡ n⫽82 for acid buffering gel and n⫽27 for nonoxynonl-9 spermicide. § n⫽76 for acid buffering gel and n⫽35 for nonoxynol-9 spermicide. 储 n⫽528 for acid buffering gel and n⫽270 for nonoxynol-9 spermicide.
objective assessment of effectiveness, adverse effects, and acceptability. The study population was young (mean age 25 years), had proven fertility (approximately 70% had a prior pregnancy), and reported a high frequency of sexual intercourse. The 6-month pregnancy rate of the acid buffering gel used with a diaphragm was noninferior to that of a commercially available nonoxynol-9 spermicide used in the same fashion. The 6-month cumulative pregnancy probability for the acid buffering gel was 2.2% less than that of the nonoxynol-9 spermicide (95% CI –7.7% to 3.3%). Estimates of differences between the two products were similar in the intent-to-treat, modified intent-to-treat, and consistent and correct use analyses. Our confidence in these relative estimates of effectiveness is supported by compatible observations in the 12-month extension. Finally, the contraceptive effi-
VOL. 110, NO. 3, SEPTEMBER 2007
cacy measured in the nonoxynol-9 spermicide arm is compatible with that reported for diaphragm plus detergent spermicide in a summary of prior studies.20 The availability of emergency contraception was unlikely to have affected these results because it was used by a small and equal number of women in each comparison group. Most studies of coitally dependent family planning methods report correct and consistent-use pregnancy rates that are substantially better than typical use rates.20 –23 The correct and consistent use rates were half as much as typical use (modified intent-totreat analysis) for both the acid buffering gel and the nonoxynol-9 spermicide in this study. The determination of correct and consistent use pregnancy rates relies on self-report of method adherence in participant diaries. We have no means of gauging whether
Barnhart et al
Efficacy of a Nondetergent Spermicide
583
Table 4. Adverse Events Acid Buffering Gel (nⴝ621)
Event* Women who experienced one or more adverse event (probable or highly probable related to product use) Fungal vaginitis Bacterial vaginosis Nasopharyngitis Intermenstrual bleeding Irregular menstruation Urinary tract infection Genital pruritus (female) Vaginal irritation Bacteriuria Vaginal pain Vaginal discharge Influenza Product discontinuation because of adverse event
24 19 17 14 15 14 11 10 9 7 6 6 5 3.2
Nonoxynol-9 Spermicide (nⴝ300)
25 19 15 18 17 13 17 11 11 10 6 4 6 3.7
P†
.54
.73
Data are expressed as %. * Self-reported symptom of moderate or greater severity. † All P values were obtained from the Cochran-Mantel-Haenzel statistic for general association to compare the two treatment groups.
participants accurately reported method adherence in this trial. Although there were no pregnancies in women who consistently and correctly used either product for more than 6 months, the CI for 12-month cumulative probability of pregnancy reflects the relatively small number of cycles contributing to this analysis. Both products used in this trial appeared to be safe and acceptable. There was a low and nondifferential incidence of physical abnormalities assessed via clinical evaluations. Compared with baseline, there was no increase in the number of women with lesions found by colposcopy during use of either product. The incidence of bacterial vaginosis and vulvovaginal fungal infection were low and similar in each group as assessed by incidence of clinical infection or Nugent score and yeast culture, respectively. Changes to vaginal microflora differed minimally between products or from baseline values. There were fewer symptomatic UTIs and less frequent recovery of U urealyticum in the acid buffering gel group compared with the nonoxynol-9 spermicide group, representing a potential benefit of nondetergent spermicide. Nonoxynol-9 has been linked to an increased risk of vaginal colonization of uropathogens.24 It will be important to see if further study confirms this finding. There were no differences in type or number of adverse events between groups. Acceptability was nearly identical between the acid buffering gel and the nonoxynol-9 spermicide groups. The acid buffer-
584
Barnhart et al
Efficacy of a Nondetergent Spermicide
ing gel was acceptable to women, with 68% reporting they liked or strongly liked the product. Additionally, 68% stated they would use it in the future, and 54% prefer it to a condom. Recruitment and retention have been difficult for barrier contraceptive trials.21–25 Loss to follow-up in this trial (12%) was intermediate within the range of recently reported spermicide or barrier contraceptive studies (4 –26%)23–25 and was equally distributed between the two arms. The rate of early discontinuation for reasons other than pregnancy or loss to follow-up was 33% for the acid buffering gel and 32% for nonoxynol-9 spermicide users, a limitation common in other barrier contraceptive studies.21 Analysis of early discontinuers in a recent trial25 suggests that early discontinuation cannot be predicted from participant characteristics during recruitment and screening, thus limiting the ability to reduce loss due to early discontinuation with targeted subject selection. Many women in actual use frequently change their contraceptive practices, and, in particular, barrier methods are often used for relatively short periods.23 Thus, the substantial early discontinuation rates noted in clinical trials evaluating barrier contraceptive methods may reflect the pattern of actual use. The similar rates of early withdrawal in the two randomized arms suggest that early discontinuation did not bias the conclusions of similarity between the two products. Prior studies of intensive nonoxynol-9 use have
OBSTETRICS & GYNECOLOGY
documented epithelial disruption,2,4 but at much more frequent dosing (one dose or more per day) than in the present study, which averaged one exposure every 3.4 days. In contrast to intensive exposure to nonoxynol-9, twice daily exposure to the acid buffering gel for 14 days did not cause epithelial disruption.26,27 A limitation of spermicidal contraceptives is their lower efficacy relative to hormonal methods and intrauterine devices.20 To maximize contraceptive efficacy in this trial, both the acid buffering gel and the nonoxynol-9 spermicide were used with diaphragms because diaphragms improve the contraceptive efficacy of existing spermicides and provide the highest contraceptive efficacy of internal barrier methods.20,28 Similarly, diaphragms provide a mechanical barrier that reduces exposure of the cervix and upper tract to pathogens contained in semen or those shed by penile epithelial lesions and, thus, may enhance the disease-prevention efficacy of microbicides.28 The cervix is susceptible to common sexually transmitted infections due to the presence of columnar epithelium which is relatively fragile, exposes living cells at the luminal surface, and is often inflamed or friable.29 Moreover, potentially infected vaginal fluids can be peristaltically transported by uterine contractions to expose the extensive columnar epithelium of the upper tract,29 –31 a hazard the barrier function of a diaphragm should reduce. This acid buffering gel is being developed as a spermicidal microbicide. Sperm and many sexually transmitted pathogens are acid-sensitive, and the acid buffering gel has been reported to be effective in animal models for contraception, herpes simplex virus, chlamydia, gonorrhea, papillomavirus, and HIV-infected leukocytes.13,14,32 Additional clinical studies will assess whether it can provide disease protection in women. These data represent a paramount first step in the development of female-controlled product providing dual protection from pregnancy and sexually transmitted infection. REFERENCES 1. Dhawan D, Mayer KH. Microbicides to prevent HIV transmission: overcoming obstacles to chemical barrier protection. J Infect Dis 2006;193:36–44. 2. Hillier SL, Moench T, Shattock R, Black R, Reichelderfer P, Veronese F. In vitro and in vivo: the story of nonoxynol 9. J Acquir Immune Defic Syndr 2005;39:1–8. 3. Hicks DR, Martin LS, Getchell JP, Heath JL, Francis DP, McDougal JS, et al. Inactivation of HTLV-III/LAV-infected cultures of normal human lymphocytes by nonoxynol-9 in vitro. Lancet 1985;2:1422–3. 4. Roddy RE, Cordero M, Cordero C, Fortney JA. A dosing study of nonoxynol-9 and genital irritation. Int J STD AIDS 1993;4:165–70.
VOL. 110, NO. 3, SEPTEMBER 2007
5. Wilkinson D, Tholandi M, Ramjee G, Rutherford GW. Nonoxynol-9 spermicide for prevention of vaginally acquired HIV and other sexually transmitted infections: systematic review and meta-analysis of randomised controlled trials including more than 5000 women. Lancet Infect Dis 2002;2:613–7. 6. Kreiss J, Ngugi E, Holmes K, Ndinya-Achola J, Waiyaki P, Roberts PL, et al. Efficacy of nonoxynol 9 contraceptive sponge use in preventing heterosexual acquisition of HIV in Nairobi prostitutes. JAMA 1992;268:477–82. 7. Van Damme L, Ramjee G, Alary M, Vuylsteke B, Chandeying V, Rees H, et al; COL-1492 Study Group. Effectiveness of COL-1492, a nonoxynol-9 vaginal gel, on HIV-1 transmission in female sex workers: a randomised controlled trial [published erratum appears in Lancet 2002;360:1892]. Lancet 2002;360: 971–7. 8. WHO/CONRAD technical consultation on nonoxynol-9, World Health Organization, Geneva, 9 –10 October 2001: summary report. Reprod Health Matters 2002;10:175–81. 9. Olmsted SS, Dubin NH, Cone RA, Moench TR. The rate at which human sperm are immobilized and killed by mild acidity. Fertil Steril 2000;73:687–93. 10. Hillier S, Holmes KK. Normal vaginal flora. In: Holmes KK, Sparling PF, Mardh PA, Lemon SM, Stamm WE, Piot P, Wasserheit JN, editors. Sexually transmitted diseases. 3rd ed. New York (NY): McGraw-Hill; 1999. p 193–4. 11. Boskey ER, Telsch KM, Whaley KJ, Moench TR, Cone RA. Acid production by vaginal flora in vitro is consistent with the rate and extent of vaginal acidification. Infect Immun 1999;67: 5170–5. 12. Masters WH, Johnson VE. Human sexual response. Boston (MA): Little, Brown; 1966. 13. Zeitlin L, Hoen TE, Achilles SL, Hegarty TA, Jerse AE, Kreider JW, et al. Tests of BufferGel for contraception and prevention of sexually transmitted diseases in animal models. Sex Transm Dis 2001;28:417–23. 14. Olmsted SS, Khanna KV, Ng EM, Whitten ST, Johnson ON 3rd, Markham RB, et al. Low pH immobilizes and kills human leukocytes and prevents transmission of cell-associated HIV in a mouse model. BMC Infect Dis 2005;5:79. 15. Ongradi J, Ceccherini-Nelli L, Pistello M, Specter S, Bendinelli M. Acid sensitivity of cell-free and cell-associated HIV-1: clinical implications. AIDS Res Hum Retroviruses 1990;6: 1433–6. 16. Mauck C, Callahan M, Weiner DH, Dominik R. A comparative study of the safety and efficacy of FemCap, a new vaginal barrier contraceptive, and the Ortho All-Flex diaphragm. The FemCap Investigators’ Group. Contraception 1999;60:71–80. 17. CONRAD/World Health Organization. Manual for the standardization of colposcopy for the evaluation of vaginal products. CONRAD, 2004. Available at: http://www.who.int/ reproductive-health/stis/colposcopy.html. Retrieved June 12, 2007. 18. Blackwelder W. “Proving the null hypothesis” in clinical trials. Control Clin Trials 1982;3:345–53. 19. Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 1977;35:1–39. 20. Trussell J. Contraceptive failure in the United States. Contraception 2004;70:89–96. 21. Raymond EG, Chen PL, Luoto J. Contraceptive effectiveness and safety of five nonoxynol-9 spermicides: a randomized trial. Obstet Gynecol 2004;103:430–9.
Barnhart et al
Efficacy of a Nondetergent Spermicide
585
22. Steiner MJ, Dominik R, Rountree RW, Nanda K, Dorflinger LJ. Contraceptive effectiveness of a polyurethane condom and a latex condom: a randomized controlled trial [published erratum appears in Obstet Gynecol 2003;101:1026]. Obstet Gynecol 2003;101:539–47. 23. Trussell J, Vaughan B. Contraceptive failure, method-related discontinuation and resumption of use: results from the 1995 National Survey of Family Growth. Fam Plann Perspect 1999; 31:64–72. 24. Hooton TM, Roberts PL, Stamm WE. Effects of recent sexual activity and use of a diaphragm on the vaginal microflora. Clin Infect Dis 1994;19:274–8 25. Raymond E, Chen PL, Bosny P, Luoto J, Barnhart KT, Bradley L, et al. Participant characteristics associated with withdrawal from a large randomized trial of spermicide effectiveness. BMC Med Res Methodol 2004;4:23. 26. Mayer KH, Peipert J, Fleming T, Fullem A, Moench T, Cu-Uvin S, et al. Safety and tolerability of BufferGel, a novel vaginal microbicide, in women in the United States. Clin Infect Dis 2001;32:476–82. 27. van De Wijgert J, Fullem A, Kelly C, Mehendale S, Rugpao S,
586
Barnhart et al
Efficacy of a Nondetergent Spermicide
28.
29.
30.
31.
32.
Kumwenda N, et al. Phase 1 trial of the topical microbicide BufferGel: safety results from four international sites. J Acquir Immune Defic Syndr 2001;26:21–7. Fu H, Darroch JE, Hass T, Ranjit N. Contraceptive failure rates: new estimates from the 1995 National Survey of Family Growth. Fam Plann Perspect 1999;31:56–63. Moench TR, Chipato T, Padian NS. Preventing disease by protecting the cervix: the unexplored promise of internal vaginal barrier devices. AIDS 2001;15:1595–602. Kunz G, Beil D, Deininger H, Wildt L, Leyendecker G. The dynamics of rapid sperm transport through the female genital tract: evidence from vaginal sonography of uterine peristalsis and hysterosalpingoscintigraphy. Hum Reprod 1996;11:627–32. Barnhart KT, Stolpen A, Pretorius ES, Malamud D. Distribution of a spermicide containing nonoxynol-9 in the vaginal canal and the upper female reproductive tract. Hum Reprod 2001;16:1151–4. Spencer SE, Valentin-Bon IE, Whaley K, Jerse AE. Inhibition of Neisseria gonorrhoeae genital tract infection by leadingcandidate topical microbicides in a mouse model. J Infect Dis 2004;189:410–9.
OBSTETRICS & GYNECOLOGY
