Currently 25% of American women using contraception use the birth control pill.1 The combination oral contraceptive (OC) includes both an estrogen and a progestin in the active pills. For four decades, the labeled OC regimen has been to take active pills for 21 days, followed by either a pill-free week or a week of inactive, spacer pill use. This cyclic 28-day regimen induces withdrawal bleeding and mimics monthly menstruation.
The belief that a monthly period is normal and healthy for women using contraception is being challenged.2,3 Historically, women without access to contraception could have had as few as 50 menstrual cycles in a life time, whereas the modern woman with reduced fecundity could have up to 450 cycles.2 Birth control pill use can markedly reduce the morbidity of menstruation yet current OC packaging and labeling does not allow women to choose an extended cycle to reduce hormonal withdrawal symptoms such as bleeding, dysmenorrhea, or headaches.
Over 20 years ago, the “tri-cycle” study, which used a 50-μg ethinyl estradiol (E2) OC product with withdrawal bleeding every 84 days, was successful and acceptable to women.4 Table 1 summarizes the published experience with extended OC regimens.4–12 All but two studies5,12 used monophasic OC preparations; all but three4,11,12 prescribed solely 30-μg ethinyl E2 products; one11 administered the OC intravaginally; and one12 allowed variable products to be used for variable extended cycles. Skipping the pill-free week has been prescribed to treat the menstrual migraine,13 estrogen withdrawal symptoms during the hormone-free interval,12 and to suppress endometriosis.14 The impact of extended OC use on the subsequent bleeding pattern has been studied inconsistently.
Most studies reported only the percentage of women with successful postponement of bleeding4–7,10,12 or did not report specific bleeding patterns.9 With extended use OCs, there appears to be an increase in unscheduled or breakthrough bleeding days which decreases over time.8,11 If postponement of withdrawal bleeding and extension of active pills results in the loss of cycle control or a regular bleeding pattern, this could lead to dissatisfaction and discontinuation.15,16 We conducted a randomized controlled trial comparing a traditional 28-day OC cycle to an extended 49-day OC cycle to measure differences in bleeding pattern, hygiene product use, compliance, symptoms, and method satisfaction.
MATERIALS AND METHODS
The study was submitted to both the local institutional review board and the Food and Drug Administration (FDA) for review. From April 1998 to April 2000, the study was conducted at four clinics in the greater Seattle King County area, including three public health clinics and one private practice clinic. All women offered an OC prescription during their clinic visit were recruited for participation in the study using posted flyers on the clinic and restroom walls. Subjects were excluded from enrollment for age less than 18 years or greater than 45 years, standard contraindications to combined OC use, inability to speak or read English, or intention to become pregnant within 1 year. Subjects were informed of the study procedures and requirements along with risks and benefits, and written informed consent was obtained.
After consent was obtained, subjects were randomly assigned to study groups by opening a sequentially numbered opaque envelope. There was carbon paper inside and subjects were asked to imprint their signature across the flap before opening the envelope. Each envelope contained a card with the designated assignment. The randomization schedule was constructed from permuted blocks of six with a 1:1 assignment ratio, these blocks were assigned a numeric range, and then randomly a number was selected from a random number table, and the corresponding block was then used to construct the card assignment. The randomization schedule was kept locked in the office, and the allocation cards were sealed in the envelopes by the first author, who did not enroll subjects. The assignment card with the randomization allocation, study number, and carbon etching of the subject's signature was then stapled to the front of each study chart as evidence of correct allocation. There was no masking after randomization because the spacer pills are green and the active pills are pink, thus subjects and staff could clearly see which pill regimen the subject was taking.
After randomization, the subject was given an envelope containing the assigned study medication, 30 μg ethinyl estradiol (E2) and 300 μg norgestrel (Lo/Ovral-28; Wyeth-Ayerst Laboratories, Philadelphia, PA). For 28-day cycle assignment, 28-day pill packages were given intact, with the spacer pills in the package. For 49-day cycle assignment, pill packages were altered by removing spacer pills from every other month to create 6 weeks of active pills followed by 1 week of spacer pills. The subjects were instructed to take one pill per day at approximately the same time each day. Subjects were counseled that late or missed pills could increase the likelihood of spotting or breakthrough bleeding.
At enrollment, the subjects were measured for height, weight, and blood pressure (BP). Subjects completed a demographic questionnaire on age, education, and tobacco use. The clinician completed a medical history questionnaire documenting no contraindications to OC use or study enrollment, recent contraceptive method, gravidity, parity, an extensive menstrual history, and history of uterine leiomyomata or infection. History of OC use was collected. A subject was classified as a new user if she had no prior OC use, as a restart if it had been more than 30 days since her last OC use, and a switcher if OC use was ongoing.
Subjects were given a daily diary sheet and instructed to record pill taking (taken on time, taken more than 4 hours early or late, or taken the next day) and bleeding episodes (trace = no sanitary protection needed, mild = requiring one to two pads or tampons, or moderate = requiring more than two pads or tampons). In addition, the presence of various side effects and possible menstrual or cyclic symptoms (cramping, tiredness, headache, breast tenderness, and genital irritation) were scored, if present, on a scale of 1 to 5, with 1 being very mild and 5 being severe.
Every 3 months and at exit, the subjects returned to the clinic for weight and BP measurement, OC dispensing, and to return their diaries. A standardized questionnaire was administered at each visit. Subjects were asked to estimate the number of late or missed pills, using a five-point scale (1 = no late or missed pills to 5 = many late and missed pills with “many” defined as more than two pills), the number of days requiring sanitary protection, and the amount of money spent on sanitary protection during the prior 3 months. In addition, their level of satisfaction with their pill schedule was reported using a five-point scale (1 = not satisfied to 5 = better than other methods and would recommend to other women). Each subject was asked to report her impression of her bleeding experience at interval clinic visits using the scale of “bleeding less than expected,” “as expected,” or “more than expected,” in the context of her study assignment. Women quitting the study were asked about their reason for discontinuation. Women exiting the study were queried about the birth control method that they planned to use after the study.
Data were entered into a statistical spreadsheet. The study variables were entered as continuous variables or dichotomous variables, as appropriate. Diary data were entered as trimester or 84-day reference period totals and not as individual days. For this reason, data were analyzed only from diaries for which there were complete trimesters of data. A trimester was considered complete if a minimum of 11 weeks of data were entered in the diary. For a subject to be described as having completed the study she needed to have documented evidence that she took the entire 12 cycles of study pills, although she might not have attended all four follow-up visits or completed all 12 diaries.
Bleeding outcomes were evaluated using definitions adapted17 from the World Health Organization's Special Programme of Research, Development and Research Training in Human Reproduction (Table 2). For each subject, several items describing bleeding patterns and pill-taking habits were counted per trimester. Included were total number of bleeding episodes, number of bleeding or spotting episodes, number of bleeding days, number of spotting days, number of episodes of prolonged bleeding, number of late or early pills (over 4 hours from regular time), and number of missed pills or pills taken the next day. Subjects who did not return a complete diary for a given trimester were not included in the statistics for that trimester.
Statistics were calculated using SPSS (Statistical Package for the Social Sciences for Windows, release 10.0.7 [1 June 2000] Standard Version; SPSS Inc., Chicago, IL). For comparisons between the two study arms, significance values for continuous variables were obtained by t test for variables with normal distributions or by Mann-Whitney test for those variables without normal distributions. Significance values for discreet variables were obtained by Pearson χ2 or Fisher exact test, as appropriate.
Withdrawal bleeding lasts on average 4 days for OC users.15 The goal of the study was to reduce bleeding days; hence bleeding days were used to calculate the sample size of the study. It was assumed that during the fourth 84-day trimester, 50% of the women using the 28-day regimen would bleed 12 days compared with only 10% of women using the 49-day regimen. The minimum sample size using a two-tailed test to give the study adequate power (80%) to detect this 40% difference would require 24 women in each arm (EpiInfo 6.04a, July 1996; Centers for Disease Control and Prevention, Atlanta, GA). For this reason, and in consideration of an anticipated study dropout rate of 40%, we decided to enroll 90 women to complete 1 year of data on the required 48 women.
Enrollment and randomization of 90 women was complete by April 1999 and data collection complete by April 2000. Late in the enrollment phase, one subject was allowed to enter the 49-day cycle arm despite 28-day cycle assignment because an earlier subject randomized to 49-day cycles was not able to start the study drug (Figure 1). This protocol deviation was approved to ensure that there were adequate numbers of subjects in the experimental arm or 49-day cycle. Figure 1 illustrates the subject dropout and completion rates for the two study assignments.
Table 3 compares the enrollment demographics and baseline characteristics of the two study populations. There were no statistically significant differences in age, body mass index, gravidity, parity, smoking, or education between the two groups. Five women switched from injection contraception to study pill use, five women had never used birth control pills, and 23 women (25% of subjects) had not used pills for at least 30 days before enrollment. Most subjects (67%) were using OCs at the time of enrollment and switched to the study pill after the scheduled withdrawal or spacer week. Of women reporting current or prior OC use, the average years of use were no different between groups (28-day cycle = 6.0 ± 5.2 years, 49-day cycle 6.6 ± 5.4 years, P = .6).
Average number of bleeding episodes, bleeding days, bleeding or spotting episodes, and spotting days were calculated for both groups per 84-day reference period or trimester; each trimester was considered separately and for only those who completed the study. Throughout all four trimesters, the women in the 49-day cycle arm had fewer bleeding days, bleeding episodes, and bleeding or spotting episodes on average than women in the 28-day cycle arm (Table 4). Because 49 does not divide equally into the 84-day reference period, there is only one withdrawal bleeding week in the first and fourth trimesters and two withdrawal weeks in the second and third trimesters for 49-day cycle subjects. All 28-day cycle subjects had three withdrawal weeks scheduled each 84-day trimester. There was no statistically significant difference between the two groups in average number of spotting days (Table 4). When new OC users and restart OC users were compared with OC switchers, there was no difference in spotting days or episodes, and this was true regardless of study assignment. There was also no difference in spotting days or bleeding or spotting episodes compared by morning versus evening pill-taking time as planned by subjects at enrollment.
Interval data collected at clinic revisits were also used to compare bleeding patterns between the two groups. The total number of days for the entire year requiring hygiene product use were significantly different between the study arms (28-day cycle = 53.5 days, 49-day cycle = 27.3 days, P < .001); this finding is supported by a reduction in spending on hygiene products by women who had extended OC use. The sum of the hygiene-product costs for the entire study year, for all subjects completing all revisits, was an average of $41.45 for subjects with a 28-day cycle and $17.54 for subjects with a 49-day cycle (P < .001).
The percentage of women who had amenorrhea, infrequent menses, frequent menses, or prolonged bleeding (as defined in Table 2) were compared between groups. There were no significant differences in amenorrhea, frequent menses, or prolonged menses. However, 58.7% of subjects in the 49-day cycle arm had infrequent menses at least once during the study, whereas only 9.1% of subjects in the 28-day cycle arm did so (P < .001). Five women reported prolonged bleeding during 1 trimester of the study, three of whom had 28-day cycles and two had 49-day cycles. Two prolonged bleeding episodes occurred in the first trimester, one in the second, one in the third, and one in the fourth trimester of study drug use.
There were no statistically significant differences between reported bleeding experience impression, with 41.7% of women with a 49-day cycle reporting less than expected, 50% as expected, and 8.3% with more bleeding than expected. This distribution was similar to women with a 28-day cycle, with 28% having less bleeding than expected, 67% as expected, and 5.6% more bleeding than expected when queried at the fourth trimester visit (P = .56).
Compliance, measured by number of late or missed pills, was no different between the study arms when measured by diary data or by subject report. Interestingly, no subject reported “many late and a few missed pills” or “many late and many missed pills” at the interval clinic visits, despite several diary counts of greater than 15 late or missed pills per trimester. On average, subjects recorded approximately five late pills and at least one missed pill per trimester on their diary sheets. Subjects missed or took pills late in all trimesters regardless of study assignment (28-day = 1.3 and 49-day = 1.2 missed pills, P = .52; 28-day = 2.6 and 49-day = 2.4 late pills reported fourth trimester, P = .81).
The 49-day cycle subjects in the fourth trimester reported significantly lower headache and genital itch symptom scores by diary (Table 5). Weight was measured at all revisits, and although the difference was not statistically significant, the average 28-day cycle subject gained almost 5 pounds during the study, and the average weight gain for the 49-day cycle subject was less than 1 pound (Table 5).
Adverse events resulting in study discontinuation are reported in Figure 1 and common symptoms are given in Table 5. There were two serious adverse events during the course of the study–a pelvic fracture while skiing and a laparoscopic cholecystectomy for gallstones in an obese subject on the 49-day cycle. Women completing the study were an average of 27.7 years old, whereas women not completing the study were younger (P = .009). More than half of those who completed the study (61.5%) had at least some college education, whereas only 38.9% of those who dropped out had completed college courses (P = .06).
The OC method satisfaction scores expressed as a median (25%, 75% with significance calculated by Mann-Whitney test) for those completing the study were no different between study arms (28-day cycle = 4.0 (4.0, 5.0), 49-day cycle 5.0 (4.0, 5.0), P = .13). At exit, subjects were asked what type of birth control they planned to use. Extended OC prescription outside a research protocol was not offered by the clinic sites at the time of study exit. A subset analysis of only women who planned to continue with hormonal contraception found that 52.4% of the extended cycle subjects wished to continue with that regimen and 16.7% of 28-day cycle subjects indicated that they were switching to an extended OC schedule.
A recent prospective, nonrandomized study18 reported significant reductions in headaches, dysmenorrhea, and premenstrual symptoms with extended OC use. In our study women randomly assigned to the 49-day extended OC cycle had significantly fewer bleeding episodes, bleeding days, and less hygiene product use and expenditure. Notably, the reduction in bleeding was not associated with a detectable increase in spotting days or method discontinuation. However, our sample size is small, and spotting or breakthrough bleeding is inherently unpredictable, with much individual variation.
Spotting and breakthrough bleeding events are common, even with cyclic, traditional OC regimens, and are often the reason for discontinuation.15,16 Interestingly, a study of extended OC use,12 specifically to treat menstrual cycle symptoms, found that women tolerated irregular bleeding and spotting if extended OC use produced perceived benefit overall. Because spotting and breakthrough bleeding persisted even in the fourth trimester for some women, future studies could use a cycle control index score19 and evaluate the etiology of the unscheduled bleeding more closely when it occurs.
The formulation and regimen that will be most effective for extended OC cycles is not known. It is likely that monophasic preparations containing gonane-type progestins with strong endometrial activity and long serum half-life will succeed for extended cycles.19–21 With cyclic use, phasic and norethindrone or estrane progestin type OC preparations resulted in significantly more breakthrough bleeding and spotting compared with monophasic levonorgestrel or gonane progestin OC preparations.20,21 Estrogen dose also can affect cycle control. When initiating OC use, decreased breakthrough bleeding was shown with increased estrogen dose, but the incidence of breast tenderness and nausea also increased, and it is not known whether these observations would hold true for extended use.19 The only metabolic study comparing 28-day to 49-day cycle OC use over 1 year detected no overall difference but found a trend toward increased serum hormone binding hormone and high-density cholesterol levels in the extended cycle subjects, suggesting a possible increased estrogen effect with the loss of the pill-free week.9
To date there are no products approved or packaged for extended OC use available in the United States. Yet, as quoted in a recent article, “long cycle regimens with infrequent bleeds are an option and should be offered.”3 We believe that this recommendation might be difficult to follow because women will need to purchase additional OC packages, and insurance coverage might be limited. In our study, extended OC use reduced annual hygiene-product expenditures from roughly $40 for subjects on the 28-day cycle to $20 for subjects on the 49-day cycle. However, the cost of hygiene product is small compared with the OC prescription cost. The cost of an 84-day OC regimen was calculated to not be cost-effective unless the woman requires 48 tampons per month, with an OC package cost of less than $25.20 per cycle.22 But perhaps “Not everything that can be counted counts, and not everything that counts can be counted” (Albert Einstein). Counting hygiene product use might not be as important as an increase in workdays, quality of life, or other productivity measures.
The pill-free week perpetuates conditions, such as anemia and dysmenorrhea, and those associated with feminine hygiene product use, such as vulvar inflammation and toxic shock syndrome. Most hormonal contraceptive methods alter menstrual bleeding, and women might choose a contraceptive method for the possible benefit of reduced bleeding. Depot medroxyprogesterone acetate injections induce hypoestrogenic amenorrhea in 90% of women after 2 years of use.23 A levonorgestrel intrauterine system just approved for use in the United States induces amenorrhea in 20% of first-year users, and after 12 years of continued use in one study, 60% of women reported amenorrhea.24 A recent survey of OC user attitude toward menstruation indicates an increasing trend toward the acceptance of menstrual reduction and even suppression.25 In Australia26 and Europe25 many women already manipulate their active OC use to postpone withdrawal bleeding when desired. Our study confirmed that an extended OC cycle can reduce cyclic bleeding and hormonal withdrawal symptoms. Further research is needed to determine the most effective schedule and formulation.
1. Hatcher R, Trussel J, Stewart F, Cates W, Stewart G, Guest F, et al. The essentials of contraception. In: Hatcher RA, Trussell J, Stewart F, Cates W Jr, Stewart GK, Guest F, et al, eds. Contraceptive technologies. 17th ed. New York: Ardent Media, Inc. 1998:212–3.
2. Thomas SL, Ellertson C. Nuisance or natural and healthy: Should menstruation be optional for women? Lancet 2000;355:922–4.
3. Kubba A, Guillebaud J, Anderson RA, MacGregor EA. Contraception. Lancet 2000;356:1913–9.
4. Loudon NB, Foxwell M, Potts DM, Guild AL, Short RV. Acceptability of an oral contraceptive that reduces the frequency of menstruation: The tri-cycle pill regimen. BMJ 1977;2:487–90.
5. Hamerlynck JV, Vollebregt JA, Doornebos CM, Muntendam P. Postponement of withdrawal bleeding in women using low-dose combined oral contraceptives. Contraception 1987;35:199–205.
6. de Voogd WS. Postponement of withdrawal bleeding with a monophasic oral contraceptive containing desogestrel and ethinyl estradiol. Contraception 1991;44:107–12.
7. Kornaat H, Geerdink MH, Klitsie JW. The acceptance of a 7-week cycle with a modern low-dose oral contraceptive. Contraception 1992;45:119–27.
8. Cachrimanidou AC, Helberg D, Nilsson S, Waldenstrom U, Olsson SE, Sikstrom B. Long interval treatment regimen with a desogestrel containing oral contraceptive. Contraception 1993;48:205–15.
9. Cachrimanidou AC, Hellberg D, Nilsson S, von Schoulz B, Crona N, Siegbahn A. Hemostasis profile and lipid metabolism with long-interval use of a desogestrel-containing oral contraceptive. Contraception 1994;50:153–65.
10. Kovacs G, Rusden J, Evans A. A trimonthly regimen for oral contraceptives. Br J Fam Plann 1994;19:274–5.
11. Coutinho EM, O'Dwyer E, Barbosa IC, Gu ZP, Shaaban MM, Aboul-Oyoon M, et al. Comparative study on intermittent versus continuous use of a contraceptive pill administered by vaginal route. Contraception 1995;51:355–8.
12. Sulak PJ, Cressman BE, Waldrop E, Holleman S, Kuehl TJ. Extending the duration of active oral contraceptive pills to manage hormone withdrawal symptoms. Obstet Gynecol 1997;89:179–83.
13. Boyle CA. Management of menstrual migraine. Neurology. 1999;53(Suppl 1):S14–S18.
14. Kistner RW. Current status of the hormonal treatment of endometriosis. Clin Obstet Gynecol 1966;9:271–92.
15. Stubblefield PG. Menstrual impact of contraception. Am J Obstet Gynecol 1994;170:1513–22.
16. Hillard PJA. The patient's reaction to side effects of oral contraceptives. Am J Obstet Gynecol 1989;161:1412–5.
17. Suvisaari J, Lahteenmaki P. Detailed analysis of menstrual bleeding patterns after postmenstrual and postabortal insertion of a copper IUD or a levonorgestrel-releasing intrauterine system. Contraception 1996;54:201–8.
18. Sulak PJ, Scow RD, Preece C, Riggs MW, Kuehl TJ. Hormone withdrawal symptoms in oral contraceptive users, continuous use. Obstet Gynecol 2000;95:261–6.
19. Rosenberg MJ, Meyers A, Roy V. Efficacy, cycle control, and side effects of low and lower dose oral contraceptives: A randomized trial of 20 micrograms and 35 micrograms estrogen preparations. Contraception 1999;60:321–9.
20. Grace E, Emans SJ, Havens KK, Merola JL, Woods ER. Contraceptive compliance with a triphasic and a monophasic norethindrone containing oral contraceptive pill in a private adolescent practice. Adolesc Pediatr Gynecol 1994;7:29–33.
21. Schilling LH, Bolding OT, Chenault CB, Chong AP, Fleury F, Forest K, et al. Evaluation of the clinical performance of three triphasic oral contraceptives: A multi-center, randomized comparative trial. Am J Obstet Gynecol 1989;160:1264–8.
22. Schwartz JL, Creinin MD, Pymar HC. The trimonthly combination oral contraceptive regimen: Is it cost effective? Contraception 1999;60:263–7.
23. Kaunitz AM. Long-acting injectable contraception with depot medroxyprogesterone acetate. Am J Obstet Gynecol 1994;170(Part 2):1543–9.
24. Ronnerdag M, Odlind V. Health effects of long-term use of the intrauterine levonorgestrel-releasing system. A follow-up study over 12 years of continuous use. Acta Obstet Gynecol Scand 1999;78:716–21.
25. den Tonkelaar I, Oddens BJ. Preferred frequency and characteristics of menstrual bleeding in relation to reproductive status, oral contraceptive use, and hormone replacement therapy use. Contraception 1999;59:357–62.
26. Rutter W, Knight C, Vizzard J, Mira M, Abraham S. Women's attitudes to withdrawal bleeding and their knowledge and beliefs about the oral contraceptive pill. Med J Aust. 1988;149:417–9.