| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Department of Obstetrics and Gynecology, Columbia University, New York, New York; and School of Public Health and Family Medicine, University of Cape Town, Cape Town, South Africa.
Address reprint requests to: Address correspondence to: Carolyn Westhoff, MD, MSc, Department of Obstetrics and Gynecology, Columbia University, PH 16–80, 630 West 168th Street, New York, NY 10032; e-mail: clw3{at}columbia.edu.
 |
ABSTRACT |
|---|
TOP
 ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: This was an open-label controlled trial. We randomly assigned 201 women to immediate start of a contraceptive pill or immediate start of the ring in a 1:1 allocation ratio. Our primary outcome was difference in mean bleeding–spotting days per woman according to treatment assignment. Secondary outcomes were differences in World Health Organization—defined menstrual indices, differences in perceived bleeding changes, and differences in bleeding according to cycle day at the start of method.
Results: The mean bleeding–spotting days in the 84-day reference period for all subjects was 19.2 days (17.0 days for ring users and 21.4 days for pill users, mean difference 4.4 days). Using the World Health Organization menstrual indices, the ring users experienced fewer days or episodes of bleeding–spotting and shorter intervals. Among ring users, no baseline characteristics were associated with bleeding outcomes. Older nulliparous pill users, however, reported more bleeding–spotting days. Significantly more ring users reported a decrease in duration of bleeding compared with pill users (P < .01). We found no significant differences in bleeding patterns based on analysis of cycle day at study enrollment.
Conclusion: Our study shows advantageous bleeding patterns for subjects using the contraceptive vaginal ring. It also confirms our previous findings that immediate start of hormonal contraception is an acceptable alternative to waiting for menses.
Level of Evidence: II-1
|
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We recruited women for participation using medical center flyers and Internet postings. Figure 1 shows the flow of participants through recruitment, enrollment, and follow-up. We screened women by telephone for the following eligibility criteria: English-speaking, aged between 18 and 40 years, regular menstrual cycles of 21–35 days in the past 12 months, no contraindications to use of hormonal contraception, no use of hormonal contraception for more than 2 menses (more than 6 menses for injectables), more than 2 menses since last pregnancy, not currently pregnant, not recent users of emergency contraception, and no unprotected sexual intercourse in the 10 days before the enrollment visit.
|
After consent and before randomization, subjects underwent a baseline interview to collect demographic information and menstrual, contraceptive, and reproductive history. We performed blood pressure readings and sensitive urine pregnancy tests on all participants. We then randomly assigned participants to the immediate start of either a low-dose triphasic oral contraceptive (Ortho Tricyclen Lo, Ortho-McNeil Pharmaceuticals, Raritan, NJ) or immediate start of a monthly vaginal contraceptive ring (NuvaRing, Organon USA, Roseland, NJ). Ortho Tricyclen Lo has a formulation of 0.18 mg norgestimate/25 µg ethinyl E2 for the first 7 tablets, 0.215 mg norgestimate/25 µg ethinyl E2 for the second 7 tablets, and 0.25 mg norgestimate/25 µg ethinyl E2 for the third 7 tablets. NuvaRing releases on average 0.120 mg etonogestrel and 15 µg ethinyl E2 per day over 3 weeks. We instructed participants in the pill group to take 1 pill each day beginning with enrollment day and ending on the last day of the third treatment cycle. We instructed participants in the ring group to use the first ring continuously for 3 weeks, discard it, have 1 ring-free week, and repeat the procedure for the second and third rings. Participants in both groups received either 3 packs of pills or 3 contraceptive rings with instructions for use, and either swallowed the first pill or inserted the first ring in our office during the enrollment visit. We gave all participants condoms and instructed them to use them for the first 7 days of the method. We also gave them home pregnancy tests and emergency contraception, which we instructed them to use in the case of unprotected sex, after 3 or more missed pills, or after an unscheduled removal of the ring.
Each participant received a bleeding diary with standardized definitions of bleeding, spotting, and no-flow days. Bleeding was defined as loss of blood requiring use of a sanitary pad or tampon; spotting was defined as loss of blood requiring use of a panty liner or no protection; and no-flow was defined as neither bleeding nor spotting. Participants recorded the occurrence of bleeding, spotting, and no-flow days in a daily bleeding diary for 84 days from the enrollment day.
A researcher not otherwise involved in the study generated the treatment assignments using a random number table and simple randomization to generate treatment assignments in a 1:1 allocation ratio. Treatment assignments were recorded on cards and kept in sequentially numbered, sealed, opaque envelopes. The allocation was concealed until the interventions were assigned, and the study coordinator and interviewers were blinded to the assignment before the subject’s opening the envelope. In our previous study, participants using a monophasic 35-µg combined oral contraceptive reported a mean number of bleeding–spotting days of 19.1. with a standard deviation of 8.0 days.6 We were interested for this study in detecting a difference between the assignment groups of 4 or more bleeding–spotting days during the 84-day reference period, a difference that seemed clinically significant. A sample size of approximately 85 women in each group gives the study more than 90% power to detect this difference with a 2-tailed 
of 0.05.
After 3 cycles of method use, participants completed an exit interview regarding method satisfaction and continuation, side effects, and perceptions of bleeding changes and returned the bleeding diary. At the exit visit, subjects received an additional ring or oral contraceptive pack and a prescription for continued contraceptive protection, if desired.
A researcher abstracted the diary data and entered it into SPSS 10.0 (SPSS Inc., Chicago, IL). We examined baseline and follow-up variables to describe the randomized groups. The main outcome measure was the total number of bleeding and spotting days, according to the reference-period method outlined by the World Health Organization (WHO).8,9 We chose an 84-day period instead of the 90-day period described by the WHO because it corresponds to 3 complete cycles of either method. The reference period began on the enrollment day (also the first day of method use). We included both bleeding and spotting days in the calculation of all menstrual events; for simplicity, we will use the general term "bleeding–spotting " to indicate bleeding or spotting. We calculated the total number of bleeding days, spotting days, and bleeding–spotting days in the reference period. Following the WHO method, we then repeated our calculations excluding menstrual events in progress at the start and termination of the reference period. For each woman, we then calculated the following menstrual indices as defined by WHO: 1) number of bleeding and spotting days; 2) number of bleeding–spotting episodes (any set of 1 or more bleeding or spotting days bounded at each end by 2 or more no-flow days); 3) length of bleeding–spotting episodes; and 4) length of bleeding–spotting-free intervals (a set of 2 or more consecutive no-flow days bounded at each end by bleeding or spotting days). We then summarized these indices for the participants in each group.
As planned a priori, we analyzed the data both by trial assignment (intent-to-treat analysis: pill compared with ring) and by menstrual cycle day of method initiation (start-day analysis: day 8+ compared with days 1–7). For the start-day analysis, we analyzed the entire group (ring plus pill) according to cycle day at start. We also analyzed each treatment group (ring and pill) separately according to cycle day at start. We calculated means, mean differences, and 95% confidence intervals for the WHO indices; we used Student t tests to compare means and Fisher exact tests or 
2 tests, as appropriate, to compare proportions. We calculated all mean differences as pill minus ring; positive values therefore indicate excess bleeding days in the pill group.
|
RESULTS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
One hundred seventy-four participants (87%) completed an exit interview regarding method satisfaction and continuation, side effects, and perceptions of bleeding changes. The proportion who completed the exit interview was similar for the ring users (88%), and pill users (85%). Cycle day at study enrollment was unrelated to whether the subject completed an exit interview. Of the 27 subjects who did not complete the exit interview, 4 (1 ring, 3 pill) discontinued the method and withdrew from the study before completing 3 cycles. One of the pill users withdrew because of erratic bleeding; the rest withdrew for reasons not related to bleeding. The remaining 23 subjects reported completing 3 cycles of method use during our telephone attempts to schedule exit interviews, but did not return for those exit interviews. There were no changes in blood pressure, no pregnancies, and no other serious adverse events during the 84-day reference period.
Of the 174 women who completed the exit interview, 156 returned a completed bleeding diary. One participant’s diary had missing information for 81 of the 84 days of the reference period and was deemed noninformative for our analyses. (For the remaining 155 women, the mean number of missing days per diary was 2.5.) Therefore, our analyses of bleeding patterns are based on data from 155 women. Of these, 143 (92%) continued using the method until the end of the third treatment cycle. Of the 13 women who discontinued using the method before the end of the reference period (8 pill, 10%; 5 ring, 6%), none discontinued for reasons related to bleeding. We repeated all of our analyses excluding the women who discontinued their assigned method and results were similar (data not shown).
Table 1 compares baseline characteristics of the 2 randomized groups. The population was, on average, 24.9 years old. The majority (77.6%) had used oral contraceptives in the past, and 29.9% had prior pregnancies. They were majority white (52.2%), and approximately 15.9% were Hispanic. Participants were balanced on all baseline characteristics except ethnicity. Twenty-one percent (n = 21) of the pill group was Hispanic, whereas only 10.9% (n = 11) of the ring group was (P = .05). Participants were balanced for race, however. Those randomized to the pill were slightly more likely to be students, slightly more likely to have prior pregnancies, and slightly more likely to have terminated a pregnancy, but these differences did not reach statistical significance. Participants were also balanced for cycle day at study enrollment.
|
Baseline characteristics were similar for those who returned diaries and those who did not. We also compared the ring and pill groups for baseline characteristics after excluding those women who did not return diaries. We found that the 2 groups remained balanced. The ring and pill groups among the 156 who returned diaries were also balanced for the cycle day on which they started the method. Twenty-one women (27%) in the ring group began on cycle days 1–7, and 57 women (73%) on cycle day 8+. In the pill group, 17 women (22%) began on cycle days 1–7 and 61 (78%) on day 8+ (P = .58). We also compared baseline characteristics for the women according to cycle day at start, and found that the 2 groups (cycle day 1–7 compared with 8+) were similar on all baseline characteristics.
The mean number of bleeding–spotting days in the 84-day reference period for all subjects was 19.2 days (17.0 days for ring users and 21.4 days for pill users, P < .01). Ethnicity, weight, body mass index, and smoking were not associated with the number of bleeding–spotting days. Twenty-nine percent of our subjects were smokers, but 81% of these smoked 5 or fewer cigarettes per day. Thus we were unable to detect any effect that might be limited to heavy smokers. In the ring group, no baseline characteristics were associated with the number of bleeding–spotting days. In the pill group, there was a weak association between nulliparity and bleeding–spotting days, particularly for older nulliparous participants. Nulliparous pill users experienced 21.9 bleeding–spotting days, compared with 17.4 for parous pill users (P = .18). This difference is attributable entirely to a difference in the number of spotting days (8.3 compared with 5.3). Nulliparous oral contraceptive users aged older than 30 years reported 24.1 bleeding–spotting days, compared with 18.8 for those aged younger than 30 years (P = .02). Because there were only 9 parous pill users, this difference should be interpreted with caution.
We repeated these analyses according to cycle day at study enrollment. We found no bleeding differences between subjects who started the method during days 1–7 of their cycle (20.2 bleeding–spotting days) and those who started on day 8 or later (18.8 bleeding–spotting days, P = .44). We also compared those who started during days 1–7 to those in each subsequent cycle week (ie, days 8–14, 15–21, and 22+). Again, we found that bleeding patterns for these subjects were similar to those for subjects who started in the first week of their cycle (days 8–14, 16.7 days, P = .13; days 15–21, 17.0 days, P = .14; day 22+, 21.2 days, P = 0.63). We stratified these analyses according to ring and pill use and found similar patterns.
Twenty-nine ring users (37%) reported briefly removing the ring at least once during the 84-day reference period. Sixty pill users (77%) reported missing at least 1 pill. The total number of bleeding–spotting days was similar among inconsistent users and perfect users of both methods (data not shown).
Table 2 shows the WHO menstrual indices for both groups over the 84-day reference period. These indices exclude any menstrual events in progress at the beginning and end of the reference period. Figure 2 shows the percentage of subjects experiencing bleeding–spotting on each day of the reference period by trial assignment. The mean number of bleeding–spotting days for the ring group was 14.5 days, and for the pill group was 19.2 days (mean difference 4.7 days; 95% confidence interval 2.1–7.3, P < .001). The number of bleeding-only days, number of spotting-only days, and number of bleeding–spotting episodes were also all different between treatment groups. The ring group experienced fewer days or episodes and shorter intervals for all of these comparisons. In addition, the mean length of bleeding–spotting-free intervals was greater for ring users than for pill users. These differences are statistically significant and are likely to be clinically significant as well. The only WHO menstrual index that was similar between groups was the mean length of bleeding–spotting episodes (6.0 days for ring group, 6.5 days for pill group; mean difference 0.5 days; 95% confidence interval –0.3 to 1.2). We also calculated these indices comparing the participants according to cycle day at enrollment; we found no differences between those who started the assigned contraceptive method during cycle days 1–7 compared with those who started on cycle day 8 or later.
|
|
The WHO reference period analysis defines clinically important bleeding patterns such as frequent bleeding, irregular bleeding, and prolonged bleeding. Table 3 shows the proportion of women in the pill and ring groups who experienced these patterns during the 84-day reference period. Ring users were somewhat less likely to experience clinically important adverse bleeding patterns. Overall, 36% of those using the ring and 46% of those using the pill experienced some kind of altered bleeding. Fifteen percent of ring users experienced prolonged bleeding, whereas in the pill group 31% did (P = .04). We repeated these analyses for both groups according to cycle day at start. Altered bleeding patterns were similar among those who started during days 1–7 compared with those who started on day 8 or later.
|
All 174 subjects who completed follow-up answered exit questions about their perceptions of bleeding during the 84-day reference period compared with their bleeding at times when they are not using hormonal contraception. Table 4 shows that a significantly greater proportion of women using the ring perceived a decrease in duration of bleeding than did those on the pill (P < .01). Similarly, ring users were less likely to report an increase in flow; this difference, however, did not reach statistical significance. Finally, ring users were more likely to report no change or a good change in their bleeding, whereas pill users were more likely to report a bad change (P < .01).
|
Thirteen women discontinued their assigned method during the 84-day reference period but completed both a bleeding diary and an exit interview. We repeated all of the above analyses excluding these women; the results were similar (data not shown).
|
DISCUSSION |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We found no significant differences in bleeding patterns when we compared subjects according to cycle day at study enrollment (day 1–7 compared with day 8+). This was true for both pill and ring users, and it remained true even when we broke down our results further to look at women in each week of the cycle. This confirms our previous study of the Quick Start method, which showed no bleeding differences according to cycle day at start of method.6 This study has numerous strengths. It was a randomized controlled trial with prospective data collection. The study had low loss to follow-up, and high rates of method continuation. We chose a priori a method of analysis (the WHO reference period method) that was designed specifically for comparing bleeding patterns among contraceptive users. This method has standardized definitions of bleeding and spotting, clear rules for deciding which events to include, and standardized definitions of altered bleeding patterns. Our analytical approach thus offered no opportunity for bias.8,9
Most baseline characteristics—including smoking, weight, and ethnicity—were not associated with the number of bleeding–spotting days. We were surprised to note in pill users a weak association between nulliparity and bleeding–spotting days, particularly for older nulliparous participants. These differences are attributable entirely to a difference in the number of spotting days. Although the available literature does not address this question, we discovered a similar trend in our previous Quick Start study of bleeding patterns. This question therefore warrants further investigation.
In addition to showing advantageous bleeding patterns for our subjects using the ring, our study showed that the Quick Start approach we have reported on in other studies is applicable to the vaginal ring. The benefits to immediate start, including quicker onset of contraceptive protection for some women and much less time needed for counseling on how and when to start the method, can thus be applied to use of the vaginal ring.
|
Footnotes |
|---|
Additional work on this study was performed by Nancy Gallagher, RN, Katherine O’Connell, MD, and Debra Teplin, PA.
Presented at the 2004 annual meeting of the American Society for Reproductive Medicine, October 16–20, 2004, Philadelphia, Pennsylvania.
Reprints are not available.
doi:10.1097/01.AOG.0000164483.13326.59
|
REFERENCES |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Saleh WA, Burkman RT, Zacur HA, Kimball AW, Kwiterovich P, Bell WK. A randomized trial of bleeding patterns by contraceptive types and steroid levels. Am J Obstet Gynecol 1993;168:1740–5.[Medline]
3. Rosenberg MJ, Long SC. Oral contraceptives and cycle control: a critical review of the literature. Adv Contracept 1992;8 suppl:35–45.
4. Rosenberg MJ, Waugh MS. Oral contraceptive discontinuation: a prospective evaluation of frequency and reasons. Am J Obstet Gynecol 1998;179:577–82.[Medline]
5. Westhoff C, Kerns J, Morroni C, Cushman LF, Tiezzi L, Murphy P. Quick start: a novel oral contraceptive initiation method. Contraception 2002;66:141–5.[Medline]
6. Westhoff CL, Morroni C, Kerns J, Murphy PA. Bleeding patterns after immediate vs. conventional oral contraceptive initiation: a randomized, controlled trial. Fertil Steril 2003;79:322–9.[Medline]
7. Akerlund M, Røde A, Westergaard J. Comparative profiles of reliability, cycle control and side effects of two oral contraceptive formulations containing 150 µg desogestrel and either 30 µg or 20 µg ethinyl estradiol. Br J Obstet Gynaecol 1993;100:832–8.[Medline]
8. Belsey EM, Machin D, d’Arcangues C. The analysis of vaginal bleeding patterns induced by fertility regulating methods. Contraception 1986;34:253–60.[Medline]
9. Belsey EM, Carlson N. The description of menstrual bleeding patterns: towards fewer measures. Stat Med 1991;10:267–84.[Medline]
10. Bjarnadottir RI, Tuppurainen M. Killick SR. Comparison of cycle control with a combined contraceptive vaginal ring and oral levonorgestrel/ethinyl estradiol. Am J Obstet Gynecol 2002;186:389–95.[Medline]
11. Roumen FJ, Apter D, Mulders TM, Dieben TO. Efficacy, tolerability and acceptability of a novel contraceptive vaginal ring releasing etonogestrel and ethinyl oestradiol. Hum Reprod 2001;16:469–75.
12. Dieben TO, Roumen FJ, Apter D. Efficacy, cycle control, and user acceptability of a novel combined contraceptive vaginal ring. Obstet Gynecol 2002;100:585–93.
This article has been cited by other articles:
|
|
 |
|
  The ESHRE Capri Workshop Group Endometrial bleeding Hum. Reprod. Update, September 1, 2007; 13(5): 421 - 431. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Westhoff, S. Heartwell, S. Edwards, M. Zieman, L. Cushman, C. Robilotto, G. Stuart, C. Morroni, D. Kalmuss, and for the Quick Start Study Group Initiation of Oral Contraceptives Using a Quick Start Compared With a Conventional Start: A Randomized Controlled Trial Obstet. Gynecol., June 1, 2007; 109(6): 1270 - 1276. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Quick Start of the Contraceptive Ring: Bleeding Patterns Acceptable Journal Watch Women's Health, August 23, 2005; 2005(823): 8 - 8. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
