Miller, Leslie MD; Hughes, James P. PhD
The combination oral contraceptive (OC) is currently packaged and labeled for cyclic use with 21 hormone or active pills followed by seven spacer or placebo pills, to produce monthly withdrawal bleeding. This bleeding can signal the absence of pregnancy but is not essential to the contraceptive mechanism of the OC. In fact, stopping the progestin for 7 days allows ovarian follicular development and can contribute to method failure. Under the influence of the OC, the endometrium is typically thin, atrophic, protected against future endometrial cancer, and not in need of monthly shedding. 1
In spite of early research 2 with continuous OC administration, the original marketed OC was designed to appear “natural” in the hopes it would be viewed by the Catholic Church as a “morally permissible variant of the rhythm method,” and this required a lunar or monthly cycle. 3 Twenty-six years ago, Loudon et al demonstrated that a withdrawal bleed every 3 months was acceptable and effective, but still no change in the OC schedule was introduced. 4 Recently, the medical requirement to menstruate while on the pill has been questioned, and it has been suggested that women could skip the spacer pills to prevent bleeding. 5–7 A recent review 8 concluded that there is a lack of evidence regarding the best formulation, duration of active pill use, and the long-term effects of continuous OC use, and called for randomized trials to address these issues.
Women appear to be interested in reducing menstruation. When queried, 65% of Dutch women aged 25–34 preferred bleeding every 3 months or less often, and 31.1% desired no bleeding. 9 Use of extended OC cycles to reduce cyclic symptoms with tolerable irregular bleeding has been shown to be effective. 10 Extension of active pill use to produce cycles varying from 42 to 84 days has been studied in small populations, with different OC formulations, and with variable success regarding cycle control. 8,11 Extending cyclical hormone pill use from 21 to 42 days reduced bleeding and need for hygiene products. 11 However, the number of spotting days among extended cycle subjects was similar to the 28-day cycle subjects, and spotting was still present at the year's end. 11 Extended pill cycles imply that at least some scheduled withdrawal bleeding is necessary, rather than the complete omission of scheduled bleeding implied by continuous, uninterrupted, or daily OC use.
However, as with hormone replacement therapy, scheduled hormone withdrawal and bleeding may destabilize the endometrium, induce proliferation, and promote the persistence of irregular bleeding. 12 Histology from continuous hormone replacement users indicates greater endometrial suppression compared with cyclic hormone users. 13,14 Irregular bleeding can trigger pill discontinuation and dissatisfaction, and fears of an increase in unexpected bleeding have been used to justify the promotion of regular monthly withdrawal bleeding as the optimal profile. But for many cyclic pill users, irregular bleeding events are quite common, 15,16 and a continuous schedule with expected initial breakthrough bleeding might be tolerated to attain the goals of improved contraception and no bleeding.
Little work has been published regarding continuous or daily OC use. Coutinho et al compared 3726 continuous cycles with 3364 cyclic cycles in 900 women using a 50-μg ethinyl estradiol/250-μg levonorgestrel OC tablet administered intravaginally. 17 Amenorrhea was rapidly induced, with 67.6% of continuous-use subjects reporting no bleeding by 3 months. This increased to 87.5% by the end of the year without an increase in OC discontinuation. 17 In Italy, after surgical treatment for endometriosis, 90 women were randomized to either daily progestin only (cyproterone acetate 12.5 mg) or a continuous OC (20-μg ethinyl estradiol/150-μg desogestrel) for 6 months to prevent disease reoccurrence. 18 At 6 months, the mean endometrial double layer thickness was 3.4 ± 0.9 mm, and amenorrhea was present in 46% of the continuous OC subjects. 18 Overall satisfaction, pain symptom scores, and continuation were similar between groups, and the authors concluded continuous OC was best for long treatments to avoid the effects of estrogen deprivation. 18 We initiated a year-long randomized trial, among women choosing contraception, to demonstrate that complete elimination of bleeding is possible with a low-dose monophasic OC preparation.
MATERIALS AND METHODS
The study was approved by the local Institutional Review Board, and women were recruited by flyer for contraception without an emphasis on the potential for bleeding reduction. Women were screened by telephone for exclusions before the clinic visit. The study was conducted from January 2000 to September 2001 at a University-affiliated, metropolitan, dedicated research clinic. Subjects were excluded for age less than 18 years or greater than 45 years; standard contraindications to combined OC use; inability to speak or read English; contraceptive injection use in prior 6 months; current diagnosis of uterine infection, fibroid, or cervical dysplasia; hemoglobin less than 9.0 g/dL; or intention to become pregnant within 1 year. Written informed consent was obtained.
At the first in-person visit, subjects completed a self-administered demographic questionnaire, responded to standardized queries regarding medical and gynecologic history items, and underwent gynecologic examination, including a Papanicolaou smear. Subjects were measured for height, weight, blood pressure (BP), and capillary hemoglobin. Prior OC use history was collected, and a subject was classified as a new user if she had no prior OC use, as a restart if it had been greater than 30 days since her last OC use, or as a switcher if OC use was ongoing.
All subjects were given two 28-day packages of study medication (Alesse 20-μg ethinyl estradiol/100-μg levonorgestrel; Wyeth Laboratories, Philadelphia, PA) and told to take the spacer or hormone-free pills on days 22–28 of the cycle for a planned initial withdrawal bleed. The purpose of this first cycle (designated cycle 0) was to document that bleeding would occur with a pill-free week and as a run-in cycle to minimize study dropouts after randomization. 19 Subjects recorded daily pill taking, and bleeding was scored (none, trace = no sanitary protection needed, mild = requiring one to two pads or tampons, or moderate = requiring more than two pads or tampons). At revisits, subjects returned the empty pill packs, and these were used to verify the diary reports. During cycle day 1 to 21 of the second pill package (designated study cycle 1), subjects returned for the enrollment and randomization visit. At this visit, subjects were not randomized for the following reasons: abnormal cervical cytology, absence of withdrawal bleeding, patient choice, and/or noncompliance defined as absent diary information or missing more than two consecutive or three total pills during days 1 to 21 of the first cycle.
An off-site institutional research pharmacy supplied the study medication, randomization schedule, and allocation verification. A computerized random number generator program with balanced blocks of ten (concealed from clinical study staff performing enrollment) was used to randomize subjects to treatment. Randomization was accomplished with the dispensing of identical sequentially numbered sealed brown bags containing study medication. The enrollment and randomization logs were closely monitored, and it was not possible to open a bag and then reassign it to another subject. There was no masking after randomization because the spacer pills are green in color unlike the pink active pills, and pill-taking instructions were different for the two study arms. Subjects randomized to the continuous arm were given 21-day pill packages without spacer pills and instructed to take only pink pills every day without interruption. Cyclic subjects were given 28-day pill packages and instructed to take the pills as packaged with pink hormone pills on days 1 to 21 and green spacer pills on days 22 to 28.
All subjects were asked to volunteer after randomization for a substudy involving a pelvic ultrasound and endometrial biopsy during cycles 1 and 9 to assess the endometrial changes with continuous compared with cyclic use. In addition, all subjects were told that if bleeding lasted longer than 10 days after cycle 3, they should return for pelvic examination, ultrasound, and endometrial biopsy. The pelvic ultrasound was performed by an outside facility. A hospital-based ultrasonographer, blinded to study assignment, used a 5.0-MHz vaginal probe to measure the double-layer endometrial thickness on sagittal view in the midbody of the uterus, assess endometrial stripe consistency, look for evidence of fibroids or submucous fibroids, and measure the size and number of ovarian follicles including the maximal follicle diameter. The endometrial biopsy was performed after the ultrasound on the same day if possible. Using a flexible plastic 4-mm collection device, the tissue was placed into formalin for routine hematoxylin and eosin staining. A single pathologist, blinded to study assignment, scored the specimen for adequacy and histology findings.
Every 3 months and at exit, subjects returned to the clinic for weight and BP measurement and to return their diaries. A urine pregnancy test was also performed if the subject was assigned to the continuous arm, if the subject was assigned to the monthly arm and reported absence of withdrawal bleeding, or if the subject was concerned about pregnancy. A standardized questionnaire was administered at each visit. Subjects were asked if they were worried about being pregnant, if the amount of bleeding had been less, equal to, or more than expected, how they were feeling in the past month (1 = poor to 5 = great), and to rate their level of satisfaction with their pill schedule using a 5-point scale (1 = not satisfied to 5 = better than other methods). Subjects were asked if they would recommend this birth control method to other women (1 = would absolutely not recommend to 5 = absolutely would recommend). Subjects were also asked, “Are there any symptoms you need to discuss?” and then the symptoms supplied by the woman were recorded. Specific symptoms (breast tenderness, nausea, and abdominal pain) were then queried for and quantified for their absence or frequency. All adverse events were recorded. Women quitting the study were queried about their reason for discontinuation and the birth control method they planned to use after the study.
The primary outcome for the study was bleeding profile comparison between the OC schedules. Bleeding outcomes were evaluated using definitions adapted from the World Health Organization. 20 For each subject, bleeding patterns and pill-taking habits were analyzed both by 28-day and 84-day intervals with the first cycle (cycle 0) separate because this cycle was before randomization, and all subjects were scheduled to have withdrawal bleeding. Analyses included total number of bleeding days, number of spotting days, number of episodes of abnormal bleeding patterns, 20 and number of missed pills (a pill was considered “missed” even if it was taken the next day). Secondary outcomes for the study were patient satisfaction, study completion, and pregnancy.
Data were entered into a statistical spreadsheet. The study variables were entered as continuous variables or dichotomous variables as appropriate. Statistics were calculated using Splus (Unix 3.1; Mathsoft, Seattle, WA). For comparisons between the two study arms, significance values for continuous variables (eg, age, weight) were obtained by the Student t test. Significance values for counts (eg, number of bleeding days) were obtained by the Mann-Whitney test. Significance values for discrete variables were obtained by the Fisher exact test. Except where specifically stated otherwise, we used intention-to-treat analyses and included all collected data from all randomized subjects even if the subject exited early, had an endometrial biopsy, violated the protocol, or did not take study drug as directed.
Withdrawal bleeding for a typical combined OC user lasts on average 4 days. 21 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 same 84-day interval, 60% of the women using the 28-day regimen would bleed 12 days compared with only 20% of women using the continuous schedule. The minimum sample size to give the study adequate power (80%) with a two-tailed test to detect this 67% reduction would require 27 women in each arm (Epi Info 6; Centers for Disease Control and Prevention, Atlanta, GA). With an anticipated study dropout rate of 30%, we decided to enroll 80 women to obtain a complete year of data on 54 women.
Seventy-nine women were enrolled and randomized by August 2000. Data collection was completed by September 2001. Study completion rates by treatment arm and reasons for exclusion and dropout are listed in Figure 1. Table 1 compares the enrollment demographics and baseline characteristics of the two study populations. There were no statistical differences in age, body mass index, gravidity, parity, partner status, ethnicity, smoking status, education, or prior OC use history between the two groups. Many subjects were using OC at the time of enrollment and subsequently switched to the study pill after the scheduled withdrawal or spacer week pill use (Table 1).
The median number of bleeding and spotting days per 84-day reference period for all subjects is shown in Table 2. Throughout the year, the continuous-use subjects experienced fewer bleeding days without a statistically significant difference between the two groups in overall spotting days. However, when we restricted the analysis of spotting days to cycle days 1 to 21 (when the cyclic user is not supposed to have bleeding or spotting), we found that the median number of spotting days increased initially but then declined over time in the continuous users, whereas the cyclic users had little change in spotting over time (Table 2).
The percentage of women experiencing absence of bleeding days during a 28-day interval was significantly higher in the continuous arm during the entire study after the run-in cycle (Figure 2). Using diary data and established criteria 20 for various bleeding patterns, the two study arms were compared between the first and last 84-day intervals of randomized study drug use (Table 3). Amenorrhea, absence of both bleeding and spotting, was reported by 16% of continuous-use subjects during cycles 1–3 and increased to 72% by cycles 10–12 (Table 3). All bleeding and spotting events were considered irregular bleeding in continuous OC subjects. These irregular bleeding events decreased over time but continued to occur in some subjects (Table 2). For example, during the last 84-day interval, nine (28%) of the 32 continuous-use subjects had an occasional spotting or bleeding day and could not be considered to have amenorrhea. However, typically, this bleeding was very light. Only three of the 32 women using the continuous OC in the last 84 days had more than a single episode of bleeding, and only one subject reported a single day of bleeding heavy enough to require the use of more than two hygiene products. However, 12 (43%) of the 28 cyclic subjects reported bleeding heavy enough to necessitate the use of more than two hygiene products for at least 1 day in the last 84-day interval (data not shown).
There were seven women who exited early from the continuous arm (Figure 1), and although the reasons reported for exit were not primarily bleeding, analysis of their bleeding diary data revealed these women recorded more bleeding while on study than women completing the study. If all of these women had completed the study, then the total difference of bleeding days between study arms would have been less but would not invalidate the findings. Upon exiting the study, 16 (41%) of the 39 continuous OC users reported that the bleeding had been more than expected compared with seven of 40 (18%) cyclic arm subjects (P = .07).
In the first 84-day interval, nine (30%) cyclic women and five (15%) continuous-use women reported missing three or more pills (P = .26) compared with the last 84-day interval when seven (32%) cyclic and only two (7%) continuous-use subjects reported missing three or more pills (P = .07). Missed pills were analyzed using days 1 to 21 for the cyclic subjects and days 1 to 28 for the continuous-use subjects because many cyclic subjects refused to swallow the spacer pills or consistently missed spacer pills. An electronic diary study of OC compliance also excluded analysis of the spacer-pill week and made no comment regarding spacer-pill consumption data. 22
A total of 16 women agreed to a baseline pelvic ultrasound and endometrial biopsy with the expectation that these procedures would be repeated during cycle 9. This subset included nine women randomized to continuous use, and seven assigned to 28-day cycles. In addition, two continuous-use subjects with prolonged bleeding underwent these assessments during cycle 9. Endometrial thickness measurements were under 5 mm in all subjects during cycle 9 but did not correlate well with histologic diagnosis, and specimens read as proliferative could have been taken from a subject with an endometrial stripe under 3 mm in thickness. No histology result indicated hyperplasia or neoplasia at any time. During cycle 9, three (60%) of the five cyclic OC users undergoing biopsy had proliferative endometrium compared with seven (88%) of the eight continuous OC users with inactive or atrophic specimens. There were two women at baseline with uterine fibroids detected on ultrasound, both under 10 mm, but one was a 5-mm submucosal fibroid, which was later not seen at the cycle-9 ultrasound, and this subject still had some irregular bleeding during cycles 10–12. One of the subjects undergoing evaluation for prolonged bleeding had two (18 mm and 28 mm) fibroids located within the uterine wall.
There were three protocol violators. One continuous arm subject misunderstood the instructions to begin a new pill pack directly and allowed herself a pill-free week each cycle until the first revisit. This meant she did not have her first true continuous week until cycle 3. The second violator (also continuous arm) decided to use a higher-dose triphasic pill for two cycles to see if it would improve her libido. She took this higher-dose pill continuously for cycles 9 to 10 with a reported improvement in her libido. She was the single continuous-use subject with proliferative histology on endometrial biopsy during cycle 9. When this deviation was discovered, she agreed to switch back to the study drug for the last two cycles of the study and again complained of decreased libido. The third violator was a cyclic arm subject who, during cycle 9, missed two consecutive pills and then stopped that pill pack, had a 7-day pill-free interval, and started a new pack after a total of 9 pill-free days. She did not become pregnant and completed the study.
Concern regarding a possible pregnancy was rare with only ten subjects reporting concern at revisit during the year. Five cyclic and four continuous-use subjects responded with “a little” concern, and a single cyclic subject reported “a lot” because of her absence of withdrawal bleeding during the spacer week. At no time did any subject report or have a positive pregnancy test.
Generally, weight, BP, and hemoglobin were not significantly different between study arms at baseline or exit (Table 4). Exceptions were hemoglobin at entry (P = .04) and systolic BP at exit (P = .02). Comparisons of an individual's change in these parameters from baseline to exit (data not shown) were not significantly different between study arms. Standardized questions at revisit and exit regarding estrogen-related symptoms found no difference in reported breast tenderness or nausea; both were infrequent complaints. However, abdominal pain was reported to have occurred at least once a month at study exit by 72% of cyclic subjects and only 26% of continuous-use subjects (P < .001).
There were three serious adverse events, all in cyclic arm subjects who continued study drug use and went on to complete the study. None was considered to be related to study medication. The events were as follows: pyelonephritis, pneumonia, and a 30-year-old subject underwent emergent oophorectomy for pain from a benign dermoid tumor during study drug cycle 6. Adverse events were reported by 85% of cyclic and 87% of continuous-use subjects at least once during the study, and complaints causing study discontinuation are reported in Figure 1. Headaches were specifically recorded on the diary by eight women using cyclic OC, compared with only one continuous OC user. No continuous-use subject complained of mood changes, but three cyclic subjects reported this adverse event.
Method satisfaction scores using a 5-point scale (mean ± standard deviation) were slightly higher at exit for cyclic subjects (4.1 ± 1.3) compared with continuous-use subjects (3.8 ± 1.7), but this difference was not significant (P = .09). On a 5-point scale, continuous OC users (4.4 ± 1.0) were no more likely (P = .83) to recommend their method to other women than cyclic OC users (4.5 ± 1.0). Similarly, no difference between study arms was seen at exit in response to the general question “In the past month, how have you been feeling?” (cyclic 3.7 ± 1.0, continuous 3.7 ± 1.0, P = .64).
At exit, subjects completing the study were asked what type of birth control they planned to use. Twenty-five (78.1%) of the 32 continuous-use subjects planned to continue on daily active OC pill use without a period week, six (18.8%) planned to switch back to cyclic OC use (with three choosing a different OC formulation), and only one subject planned to use no method. Of the 28 cyclic subjects completing the study, 19 (67.9%) planned to stay on a cyclic OC schedule, six (21.4%) planned to try continuous use, and three women chose no method at exit.
Continuous administration of a 20-μg ethinyl estradiol monophasic OC was well tolerated and significantly reduced bleeding compared with cyclic use. Spotting days declined over time with continuous or daily use, and study dropout because of unacceptable bleeding was rare. By study cycle 2 (56 days of daily OC use), 49% of the continuous-use subjects reported no bleeding. This increased to 68% by cycle 6, and by year's end 88% of continuous-use subjects reported no bleeding (Figure 2). Complete amenorrhea (no bleeding or spotting) was reported by 23 (72%) of the 32 continuous-use subjects during the last 84-day time period.
This absence of bleeding was not associated with a fear of pregnancy. Only the cyclic subjects with a “silent” or missed menses expressed this concern. With the typical OC failure rate 50 times higher than the “perfect user” failure rate of 0.1%, eliminating the pill-free or spacer week might improve efficacy. 23 A previous continuous OC study detected a statistically significant reduction in pregnancies from four with cyclic use in 446 women to zero with daily or continuous OC administration in 454 women over 1 year. 17 Our data suggest there was a small improvement in pill-taking compliance with daily consumption compared with cyclic use. However, we relied on a self-reported daily diary, and without electronic recording, it is likely subjects significantly under-reported missed pills. 22 In our study, we extensively counseled pill taking as the key to preventing irregular bleeding, but we did not observe a correlation between missed pills and bleeding in our data.
It is likely a continuous OC schedule would induce greater ovarian suppression, and prevent ovulation, which continues in some women with cyclic OC use. 24,25 Although shortening the pill-free week to 4 days with 24 days of active pill use produced greater ovarian suppression, it did not eliminate irregular bleeding, and often bleeding occurred beyond the pill-free interval. 26 Extreme individual variation is common when studying follicular and hormonal response to the OC, and this idiosyncratic response could explain how some women claim to miss only one pill and still conceive. 27
If continuous OC use is initiated properly and continued without interruption, then the risk of pregnancy should be very small. Even if a pregnancy were present, the modern OC dose is not a teratogen 28 and does not increase the risk of fetal genital malformations. 29 Nonetheless, pregnancy testing for reassurance should be readily available during the first year of withdrawal bleeding suppression. Other contraceptive methods, such as the progestin-only injection or progestin-releasing intrauterine device, also alter menstrual bleeding (including significant rates of amenorrhea), but a monthly pregnancy test is not usually required. Although the OC efficacy depends on user compliance, it is still a decision under the woman's control and “get a pregnancy test if you are missing pills” is prudent advice.
Only a small percentage of our subjects underwent ultrasound and histology assessments, but the findings provide reassurance that there does not appear to be a “build up” within the uterus with continuous OC use. Instead, it appears the cyclic subjects with the scheduled withdrawal bleeding are the ones who experience a proliferation of the endometrium. It was considered novel and experimental to use continuous combined estrogen and progesterone during menopause years ago, but as early as 1981, there was histological documentation of endometrial atrophy and quiescence with daily progestin administration even after only 4 months of treatment. 13 These same authors concluded that continuous replacement is “at least as effective as sequential therapy but without its inconvenience.” 13
Shedding the endometrium was always presumed to be essential to protecting the endometrium from cancer, but another study 30 of long-term estradiol replacement found the induction of withdrawal bleeding was not necessary to prevent hyperplasia. In fact, higher estrogen and progestin doses were needed to produce cyclic bleeding, and continuous combined therapy with low-dose estradiol induced amenorrhea while providing therapeutic estradiol levels. 30 In a recent study, cyclic progestin did not prevent all hyperplasia, whereas no case was found in continuous hormone replacement users. 14 Continuous combined hormone replacement therapy has also been shown to increase compliance compared with cyclic use. 31
Long-term cyclic OC use is very safe, 32 and it is unlikely the loss of the pill-free week would greatly alter the safety, particularly if the total dose consumed is comparable to cyclic use. With 24 days of the 15-μg ethinyl estradiol product, there were no significant procoagulant changes in homeostasis parameters. 26 However, in women using a 42-day cycle with a 30-μg ethinyl estradiol preparation, there was a measurable trend after 1 year towards enhanced estrogenic metabolic effects. 33 To minimize potential additive estrogenic effects with the loss of the pill-free week, it may be reasonable to use only 20-μg ethinyl estradiol OC products for daily use. The total ethinyl estradiol dose administered over 1 year to the cyclic 30-μg ethinyl estradiol OC user would be 8.2 mg, and this is more than the 7.3-mg ethinyl estradiol exposure if a 20-μg ethinyl estradiol dose product were to be consumed daily. In addition, given the evidence regarding hormone replacement doses, it appears a lower estrogen dose would be most conducive to maintaining amenorrhea. 30
In summary, continuous OC use for 1 year was well tolerated without an increase in OC discontinuation and induced amenorrhea or infrequent bleeding in 88% of subjects with an initial increase but then an actual decrease in spotting days as the year progressed. Although our study is limited by numbers and duration of use, perhaps in time it may seem odd to cycle women on the OC, and “The difficulty is not so much in developing new ideas as in escaping from old ones” (John Maynard Keynes).
1. Cancer and Steroid Hormone Study of the Centers for Disease Control and Human Development. Combination oral contraceptive use and the risk of endometrial cancer. JAMA 1987;257:796–800.
2. Kirstner RW. The use of newer progestins in the treatment of endometriosis. Am J Obstet Gynecol 1958;75:264–78.
3. Marks LV. Sexual chemistry. A history of the contraceptive pill. New Haven, Connecticut: Yale University Press, 2001:222.
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. Coutinho EM, Segal SJ. Is menstruation obsolete? New York: Oxford University Press, 1999 [original work published 1996].
6. Thomas SL, Ellertson C. Nuisance or natural and healthy: Should menstruation be optional for women? Lancet 2000;355:922–4.
7. Kaunitz AM. Menstruation: Choosing whether … and when. Contraception 2000;62:277–84.
8. Clarke AK, Miller SJ. The debate regarding continuous use of oral contraceptives. Ann Pharmacother 2001;35:1480–4.
9. 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.
10. 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.
11. Miller L, Notter K. Menstrual reduction with extended use of combination oral contraceptive pills: Randomized controlled trial. Obstet Gynecol 2001;98:771–8.
12. 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.
13. Staland B. Continuous treatment with natural oestrogens and progestogens. A method to avoid endometrial stimulation. Maturitas 1981;3:145–56.
14. Sturdee DW, Ulrich LG, Barlow DH, Wells M, Campbell MJ, Vessey MP, et al. The endometrial response to sequential and continuous combined oestrogen progestogen replacement therapy. Br J Obstet Gynaecol 2000;107:1392–400.
15. 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.
16. Rosenberg MJ, Meyers A, Roy V. Efficacy, cycle control, and side-effects of low and lower dose oral contraceptives: A randomized trial of 20 mcg and 35 mcg estrogen preparations. Contraception 2000;60:321–9.
17. 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.
18. Vercellini P, De Giorgi O, Mosconi P, Stellato G, Vicentini S, Crosignani PG. Cyproterone acetate versus a continuous monophasic oral contraceptive in the treatment of recurrent pelvic pain after conservative surgery for symptomatic endometriosis. Fertil Steril 2002;77:52–61.
19. Lang JM, Buring JE, Rosner B, Cook N, Hennekens CH. Estimating the effect of the run-in on the power of the physicians' health study. Stat Med 1991;10:1585–93.
20. Belsey EM, Machin D, d'Arcangues C. The analysis of vaginal bleeding patterns induced by fertility regulating methods. Contraception 1986;34:253–61.
21. Reisman H, Martin D, Gast MJ. A multicenter randomized comparison of cycle control and laboratory findings with oral contraceptive agents containing 100 mcg levonorgestrel with 20 mcg EE2 or triphasic norethindrone with EE2. Am J Obstet Gynecol 1999;181:S45–52.
22. Potter L, Oakley D, de Leo-Wong E, Canamar R. Measuring compliance among oral contraceptive users. Fam Plann Perspect 1996;28:154–8.
23. Trussell J, Kowal D. The essentials of contraception. In: Hatcher RA, Trussell J, Stewart F, Cates W Jr, Stewart GK, Guest F, et al, eds. Contraceptive technology. 17th rev ed. New York: Ardent Media Inc., 1998:216.
24. Coney P, DelConte A. The effects on ovarian activity of a monophasic oral contraceptive with 100 mcg levonorgestrel and 20 mcg EE2. Am J Obstet Gynecol 1999;181:S53–8.
25. Schwartz JL, Creinin MD, Pymar HC, Reid L. Predicting risk of ovulation in new start oral contraceptive users. Obstet Gynecol 2002;99:177–82.
26. Fruzzetti F, Genazzani AR, Ricci C, De Negri F, Bersi C, Carmassi F. A 12 month clinical investigation with a 24 day regimen containing 15 mcg EE2 plus 60 mcg gestodene with respect to hemostasis and cycle control. Contraception 2001;63:303–7.
27. Killick SR, Bancroft K, Oelbaum S, Morris, Elstein M. Extending the duration of the pill-free interval during combined oral contraception. Adv Contracept 1990;6:33–40.
28. Bracken MB. Oral contraception and congenital malformations in offspring: A review and meta-analysis of the prospective studies. Obstet Gynecol 1990;76:552–7.
29. Raman-Wilms L, Tseng AL, Wighardt S, Einarson TR, Koren G. Fetal genital effects of first trimester sex hormone exposure: A meta-analysis. Obstet Gynecol 1995;85:141–9.
30. Moyer DL, De Legnieres B, Driquez P, Pez JP. Prevention of endometrial hyperplasia by progesterone during long term estradiol replacement: Influence of bleeding pattern and secretory changes. Fertil Steril 1993;59:992–7.
31. Doren M, Reuther G, Minne HW, Schneider HPG. Superior compliance and efficacy of continuous combined oral estrogen progestogen replacement therapy in postmenopausal women. Am J Obstet Gynecol 1995;173:1446–51.
32. Beral V, Hermon C, Kay C, Hannaford P, Darby S, Reeves G. Mortality associated with oral contraceptive use: 25 year follow up of cohort of 46,000 women from Royal College of General Practitioner's oral contraceptive study. BMJ 1999;318:96–100.
33. 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.