Edelman, Alison B. MD, MPH1; Koontz, Stephanie L. MD1; Nichols, Mark D. MD1; Jensen, Jeffrey T. MD, MPH1
A small number of randomized trials have studied extended or continuous dosing of combined oral contraceptives (OCs).1–6 These dosing regimens vary in cycle length from 2 months to more than 1 year. The possible advantages for this approach compared with standard regimens (21 days of active pills, 7 days of placebo) include a decrease in menstrual-associated symptoms and total number of bleeding days.
Although there may be a decrease in the total number of bleeding days women experience with extended or continuous cycle OCs, these regimens frequently result in an increased number of “breakthrough” or nonscheduled bleeding days.1–6 Breakthrough bleeding is among the main reasons cited for discontinuation of OCs dosed cyclically or continuously and may offset the perceived benefit of fewer withdrawal bleeding events for many women taking continuously dosed OCs.7,8 The exact mechanisms responsible for breakthrough bleeding during hormonal contraception are unknown, although research has demonstrated that bleeding patterns improve with increasing duration of use.1,3,5,6 Miller et al6 revealed that 49% of study subjects experienced amenorrhea by cycle 2 of continuous OC use, whereas 88 % experienced amenorrhea by cycle 12. Studies have not addressed strategies to reduce or eliminate abnormal bleeding while using continuous OCs.
Several studies have shown levonorgestrel, a potent gonane, to have better cycle control than norethindrone acetate, a weak estrane progestin, in cyclic-dosed OCs.9–12 Due to characteristics of their classes, levonorgestrel and norethindrone acetate have distinct biochemical differences: levonorgestrel has higher bioavailability, a longer serum half-life, and a higher relative binding affinity in humans.9 Longer serum half-life is expected to give more endometrial stability and therefore improve cycle control. The ethinyl estradiol (E2) dose may also influence cycle control by counterbalancing the progestin component and stabilizing the endometrium.13–17
The primary hypothesis of this study was that characteristics of individual progestins, or estrogen dose, may influence the breakthrough bleeding rates with continuously dosed OCs. This study compares the bleeding patterns and acceptance of those patterns among women initiating continuous dosing of an OC containing either 1,000 μg norethindrone acetate or 100 μg levonorgestrel combined with either 20 μg or 30 μg ethinyl E2.
MATERIALS AND METHODS
A randomized, double-blind trial was conducted at Oregon Health & Science University between October 2002 and August 2004. Women were recruited through radio and print advertisements. In addition, flyers were used to target Oregon Health & Science University and Planned Parenthood of the Columbia Willamette clinics. The Institutional Review Board at Oregon Health & Science University and the Planned Parenthood of the Columbia Willamette Research Oversight Committee approved the study protocol.
Women who met the following criteria were invited to participate: aged 18 to 49 years, good general health, no medical contraindications to OC therapy (including no abnormal bleeding). All participants were required to have taken cyclic OCs for at least 3 months before enrollment to avoid the common transition bleeding seen with OC initiation. All subjects underwent informed written consent.
Subjects were instructed to take the study medication for 180 days without any pill-free intervals. Two of the treatment arms used common, commercially available OCs: 100 μg levonorgestrel/20 μg ethinyl E2 (20LNG group) (Alesse, Wyeth Pharmaceuticals, Madison, NJ) and 1,000 μg norethindrone acetate/20 μg ethinyl E2 (20NETA group) (LoEstrin 1/20, Warner-Chilcott, Rockaway, NJ). To vary the dose of estrogen while keeping the progestin type and dose constant, the other 2 study arms used novel pills: 100 μg levonorgestrel/20 μg ethinyl E2 plus 10 μg ethinyl E2 (30LNG group) and 1,000 μg norethindrone acetate/20 μg ethinyl E2 plus 10 μg ethinyl E2 (30NETA group) that were formulated by adding 10 μg ethinyl E2 to each of the commercially available OCs. We intentionally did not use any commercially available 30 μg ethinyl E2 preparations because this would also have varied the dose of progestin.
Subjects were randomly assigned to 1 of the 4 active treatment arms using a simple computer-generated randomization scheme, with block sizes of 40. To conceal treatment groups to investigators and subjects, individual pills were packaged into identical gel capsules. The 20LNG and 20NETA pills were packed with and without a 10 μg ethinyl E2 tablet in gel caps and packed with an inert substance (Starch, CAC-09005–25–8, Spectrum Chemical Manufacturing Corporation, Gardena, CA) to appear identical. The 10 μg ethinyl E2 tablets were prepared by a compounding pharmacy (Flanders Pharmacy, Portland, OR) using bulk ethinyl E2 (Spectrum Labs, Gardenia, CA) packaged with rice flour (Puratos Baking Company, Seattle, WA). The Oregon Health & Science University Research Pharmacy dispensed the pills into identical bottles containing a 30-day supply each and labeled these bottles according to the study number based on the randomization sequence. Only the research pharmacist had access to the randomization code and block size. No identifiers of treatment group were placed on subject data sheets or pill bottles, only the number in order of enrollment.
After telephone screening, a subject’s first visit consisted of an eligibility evaluation, demographic and medical history taking, weight and blood pressure check, and informed consent. Subjects were then randomly assigned to their treatment arms by the research pharmacist and provided with a 4-month supply of pills and a study calendar to document daily bleeding events, pill compliance, and adverse effects. Bleeding was rated daily by the subject on a scale of 0 to 2: 0 = no bleeding; 1 = spotting; 2 = bleeding greater than spotting. Spotting was defined as light flow that does not necessitate sanitary protection, or the equivalent of 1 panty liner not saturated. In addition, subjects were asked to make a daily notation in their calendars (yes/no) regarding the presence or absence of headache, nausea, bloating, acne, mood/depression, breast tenderness, menstrual pain, and any other events experienced. Subjects were also asked to make a daily notation on their calendars (yes/no) of whether they took their pills. If a pill was missed, they were instructed to take this pill as soon as they remembered and to note the date and time on their calendars.
Approximately 90 days into study participation and at study exit (180 days), subjects attended a visit to review study calendars and adverse effects, recheck blood pressure, and to dispense the remaining study medication. If study calendars were not complete at the time of either visit, the subject was asked to complete the calendar retrospectively. At study exit (180 days), subjects also completed a satisfaction questionnaire. Satisfaction questions included “were you satisfied with the bleeding pattern you experienced with this continuous pill regimen” (100 mm visual analog scale, anchors: 0 mm = very unsatisfied, 100 mm = very satisfied) and “how does this continuous pill regimen compare to the traditional regimen you used previously” (answers: no preference, prefer continuous regimen, or prefer traditional regimen). Subjects requesting early withdrawal from the study were asked the reason for withdrawal, completed the exit satisfaction survey, and had their remaining study medication collected.
The primary outcomes measured in this study were total number of bleeding, spotting, or amenorrhea days. Bleeding patterns were analyzed for each 30-day treatment episode, the first and second 90-day intervals, and the total study period (180 days). Secondary outcomes for the study were subject satisfaction, adverse events (pregnancy, other major medical events), menstrual-associated adverse effects, compliance/missed pills, and the subset of endometrial thickness measurements.
The target sample size of 160 subjects (40 in each group) was based on a power analysis for a 4-arm analysis of variance design, with the number of bleeding, spotting, or amenorrhea days as the primary outcome (percent standard deviation unit was set at 0.75, a β of 0.8, and an alpha of 0.05) (www.math.yorku.ca/SCS/Online/power).
Between-group demographic and menstrual-associated symptoms were analyzed using an analysis of variance, pair-wise comparisons with a post-hoc test (Bonferroni), and χ· tests (categorical variables). Bleeding, spotting, and amenorrhea days were analyzed using nonparametric testing (Kruskall Wallace). In addition, pair-wise comparisons were performed using a Mann-Whitney U test (6 pair-wise comparisons, significant difference set at 0.008). To account for subjects lost to follow-up and subject dropout, the proportion of days with each bleeding pattern (amenorrhea, spotting, bleeding) was calculated by dividing the total number of bleeding, spotting, or amenorrhea days for subjects within a treatment group by the number of women-days of participation. We then multiplied the proportions by 30 to obtain the expected number of days of each bleeding pattern during a 30-day interval. The mixed effects analysis was performed to assess whether a change across time in each bleeding pattern was significantly different among the 4 treatment arms. We used a contrast statement within the framework of the mixed effects analysis to test the overall group difference at each time point and adjusted the P values using the Bonferroni method (6 comparisons, significant difference set at 0.0083). Statistical analyses were performed using the Statistical Program for Social Sciences (SPSS 13.0 for Windows; SPSS inc, Chicago, Ill) and the Statistical Analysis System (SAS 9.0 for Windows; SAS Institute Inc, Cary, NC). All analysis was done based on intent to treat.
One hundred thirty-nine women were randomly assigned to 1 of 4 groups: 20LNG (n = 36), 30LNG (n = 34), 20NETA (n = 33), and 30NETA (n = 36). Study recruitment was terminated after 139 subjects due to a longer than expected recruitment period.
Demographic characteristics showed no differences among groups (Table 1). The average study subject was a 25-year-old, nulligravid, white woman who weighed 157 lb. Although women in the 20NETA group were somewhat heavier (170 lb) than the other treatments (150–153 lb), this did not represent a significant difference. Study retention (from entry to 180 days) was highest in the 30NETA arm (26/36, 72%) as compared with 20NETA (18/33, 55%), 20LNG (20/36, 56%), and 30LNG (13/34, 38%) arms. Reasons for study dropout and when it occurred are illustrated in Figure 1. The “physician recommendation” reason listed for discontinuation was made by the subject’s primary physician and was secondary to an unspecified minor complaint.
Table 2 demonstrates the median number of amenorrhea, spotting, and bleeding days for the first and second 90 study day intervals and for the entire study period (180 days). In the second 90-day study interval, 20NETA and 30NETA each had significantly more days without any bleeding (amenorrhea) than 30LNG (P < .008). The 30LNG group reported more spotting days than 20NETA over the entire study period (P < .008) and more spotting days than 30NETA for the second 90 days (P < .008). Only a small number of bleeding days occurred, with no significant differences noted between groups. Although not statistically significant, both LNG groups had fewer days of amenorrhea than either NETA group for the entire study period (approximately 10 days fewer for 20LNG and 30 days fewer for 30LNG), and 20LNG had more days of spotting for the entire study period (approximately 10 days more).
To better account for dropout, Table 3 and Table 4 and Figure 2, Figure 3, and Figure 4 illustrate the mean expected days of amenorrhea, spotting, and bleeding in the study for each 30-day interval. There was a significant group by time interaction for amenorrhea days [F(15,104) = 2.12; P < .014] and spotting days [F(15,104) = 1.81; P < .043], indicating that the pattern of changes across time for these variables was significantly different among the 4 treatment arms. Results from this analysis are consistent with the overall observations. The 30LNG group reported significantly fewer amenorrhea days and more bleeding and spotting days in comparison with both NETA groups. The 20LNG results were not significantly different from the other 3 groups but trended between the 30LNG and both NETA groups.
Figure 5 demonstrates the percentage of patients who reported 30 consecutive days of amenorrhea within each 30-day increment of the study (groups are significantly different in the second, fourth, fifth, and sixth 30-day intervals). Although the initial percentage of women with 30 days of amenorrhea was significantly greater for 30NETA than either of the 2 LNG groups group in the second 30-day interval (P < .05), the gap between the 30NETA and 20LNG groups diminished across the subsequent intervals of observations whereas the differences between the 30NETA and 30LNG groups persisted (fourth interval P = .03, fifth interval P = .04, sixth interval P = .08). Overall, 20NETA, 30NETA, and 20LNG show an increasing percentage of patients achieving an entire 30 days of amenorrhea during the course of the study, whereas 30LNG did not.
The number of days of menstrual-associated symptoms (nausea, headache, bloating, breast tenderness, cramping, acne and moodiness) experienced by subjects using the 4 treatments did not differ among the groups across the first and second 90-day intervals or for the total study period of 180 days.
Pill-taking compliance was analyzed from subjects’ daily menstrual calendars over a 180 day period. Self-reported compliance for all groups was 96% or higher, with no significant differences seen among or between groups (P = .25)
In direct comparisons between groups at study exit, subjects in 20NETA (75 mm, P = .01) and 30NETA (83 mm, P = .001) reported significantly higher satisfaction with the bleeding pattern as compared with 30LNG (43 mm) but not the 20LNG (72 mm) groups. When compared with traditionally dosed pills that the subjects took before study entry, there was no statistically significance difference in regimen preference among the 4 groups. However, within progestin groups, significantly more NETA subjects reported a preference for continuous dosing over traditional (20NETA 71%, 30NETA 83%) than did LNG users (20LNG 64%, 30LNG 38%). No major adverse outcomes were reported in any of the groups including no pregnancies occurred.
The 4 hormone combinations in this study were carefully chosen to investigate bleeding profiles with continuously dosed OCs. We started with 2 popular formulations of very-low-dose monophasic pills (20 μg ethinyl E2), with progestins from different families (estranes and gonanes), and then created 2 novel formulations by adding 10 μg ethinyl E2 to each of these pills. This approach allowed for direct comparisons between progestins, as well as between estrogen doses, while controlling for the progestin dose and type. At this time, continuous dosing is considered an off-labeled use with these formulations.
We recognize that 2 of the study arms (30LNG and 30NETA) represent products not available by prescription. No commercially available products exist that vary the ethinyl E2 dose while keeping the progestin dose and type constant. Thus, to address our hypothesis, new combinations were created. Although pharmacokinetics studies of these 2 novel pills were not performed, the effect of the gelatin capsules and filler (starch) on the absorption of steroid hormones should be insignificant and similar to a meal. Furthermore, detailed metabolic studies do not suggest that absorption or steady-state kinetics of contraceptive steroids are significantly influenced by dietary factors.18,19 Oral contraceptives do not need to be dosed on an empty stomach. Importantly, even in the worst-case scenario that the compounded ethinyl E2 was not well absorbed or unstable, the results comparing norethindrone acetate and levonorgestrel for the 20 μg ethinyl E2 arms remain unchanged.
Although we experienced a higher than expected discontinuation rate, study retention for subjects who completed the first month was comparable to other studies (59–80%).6 In our study, subjects received a 4-month supply of oral contraceptives upon enrollment. Since no other benefits to participation or payments were provided, perhaps subjects felt no incentive to continue with the study after receiving the study medication (this may have also contributed to our prolonged enrollment period).
The length of the study (6 months), the study population, or both may have affected our results in regard to bleeding. As mentioned previously, breakthrough bleeding with continuously dosed OCs seems to improve with greater duration of use. Breakthrough bleeding rates between our study groups may equalize given more time. However, our study illustrates what occurs in the first 180 days of use, which may be more important for improving OC discontinuation.
In addition, we chose to study a population of OC switchers (women already taking 3 months of cyclic OCs) rather than a study population of new starts or both. This eligibility criterion was a deliberate choice, because we wanted to avoid the breakthrough bleeding commonly attributed to new exposure to hormonal contraception. Although we did not analyze the OC preparation that subjects were switched from, it is not likely that the prior formulation used influenced our outcomes, because subjects were randomly allocated to treatment groups. It is not known whether switchers or new starts are more susceptible to breakthrough bleeding when initiating OCs in continuous-dosing cycles.
We hypothesized that either enhanced endometrial support due to a higher estrogen dose (eg, 30 μg ethinyl E2 pills) or higher bioavailability and longer half-life of progestin (the gonane levonorgestrel) would result in the best bleeding profiles.9 In contrast, we found superior bleeding patterns with norethindrone acetate, and no benefit of additional ethinyl E2. This study had an internal consistency in its results: the trends in our data support 20NETA with the best bleeding profile, followed by 30NETA, then 20LNG, and last, 30LNG. Significant differences were most prominent between the 2 ends of the spectrum (20NETA compared with 30LNG). This would suggest better bleeding patterns with norethindrone acetate and no benefit from the addition of ethinyl E2 (20 μg ethinyl E2 as compared with 30 μg ethinyl E2 [nonsignificant trends]).
Most clinicians view breakthrough bleeding while using OCs as a result of an unstable or atrophic endometrium caused by a progestin-dominant OCs. Standard treatment is to increase the estrogen dose to “stabilize” the lining. Prior research has examined the relationship between estrogen dose and breakthrough bleeding in cyclic-dosed OCs, but the results are conflicting.7,8,13–17,20–28 In addition, the type and dose of progestin often differ between comparison groups in these studies. Our findings suggest that changing the type of progestin may be more beneficial than estrogen dose for decreasing breakthrough bleeding.
In summary, our data do not support the use of higher-dose estrogen preparations to prevent breakthrough bleeding with continuous-dosing OCs. The progestin, norethindrone acetate at a 1,000-μg dose seems to have an improved bleeding profile over levonorgestrel at a 100-μg dose for continuous dosing. Future investigations to confirm these findings and investigate strategies to treat breakthrough bleeding are warranted.
1. Cachrimanidou AC, Hellberg D, Nilsson S, Waldenstrom U, Olsson SE, Sikstrom B. Long-interval treatment regimen with a desogestrel-containing oral contraceptive. Contraception 1993;48:205–16.
2. Anderson FD, Hait H. A multicenter, randomized study of an extended cycle oral contraceptive. Contraception 2003;68:89–96.
3. 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.
4. Kwiecien M, Edelman A, Nichols MD, Jensen JT. Bleeding patterns and patient acceptability of standard or continuous dosing regimens of a low-dose oral contraceptive: a randomized trial. Contraception 2003;67:9–13.
5. Miller L, Notter KM. Menstrual reduction with extended use of combination oral contraceptive pills: randomized controlled trial. Obstet Gynecol 2001;98:771–8.
6. Miller L, Hughes JP. Continuous combination oral contraceptive pills to eliminate withdrawal bleeding: a randomized trial. Obstet Gynecol 2003;101:653–61.
7. Archer DF, Maheux R, DelConte A, O’Brien FB. Efficacy and safety of a low-dose monophasic combination oral contraceptive containing 100 microg levonorgestrel and 20 microg ethinyl estradiol (Alesse). North American Levonorgestrel Study Group (NALSG). Am J Obstet Gynecol 1999;181:39–44.
8. An open-label, multicenter, noncomparative safety and efficacy study of Mircette, a low-dose estrogen-progestin oral contraceptive. The Mircette Study Group. Am J Obstet Gynecol 1998;179:S2–8.
9. Wallach M, Grimes D, editors. Modern oral contraception: updates from the contraceptive report. Totowa (NJ): Emron; 2000. p15–22, 60–6, 70–6.
10. Edgren RA, Nelson JH, Gordon RT, Keifer WS Jr. Bleeding patterns with low-dose, monophasic oral contraceptives. Contraception 1989;40:285–97.
11. Ramos R, Apelo R, Osteria T, Vilar E. A comparative analysis of three different dose combinations of oral contraceptives. Contraception 1989;39:165–77.
12. Rosenberg MJ, Long SC. Oral contraceptives and cycle control: a critical review of the literature. Adv Contraception 1992;8 suppl: 35–45.
13. Endrikat J, Muller U, Dusterberg B. A twelve-month comparative clinical investigation of two low-dose oral contraceptives containing 20 micrograms ethinylestradiol/75 micrograms gestodene and 30 micrograms ethinylestradiol/75 micrograms gestodene, with respect to efficacy, cycle control, and tolerance. Contraception 1997;55:131–7.
14. Akerlund M, Rode A, Westergaard J. Comparative profiles of reliability, cycle control and side effects of two oral contraceptive formulations containing 150 micrograms desogestrel and either 30 micrograms or 20 micrograms ethinyl estradiol. Br J Obstet Gynecol 1993;100:832–8.
15. Akerlund M. Clinical experience of a combined oral contraceptive with a very low dose ethinyl estradiol. Acta Obstet Gynecol Scand Suppl 1997;164:63–5.
16. Appel TB, Kambi AA, Birdsall C, Christenson DA, Clark DO, DeKoning JR, et al. A comparison of a new graduated estrogen formulation with three constant-dosed oral contraceptives. Contraception 1987;35:523–32.
17. Sulak P, Lippman J, Siu C, Massaro J, Godwin A. Clinical comparison of triphasic norgestimate/35 micrograms ethinyl estradiol and monophasic norethindrone acetate/20 micrograms ethinyl estradiol. Cycle control, lipid effects, and user satisfaction. Contraception 1999;59:161–6.
18. Fotherby K. Bioavailability of orally administered sex steroids used in oral contraception and hormone replacement therapy. Contraception 1996;54:59–69.
19. Roberts RJ, Leff RD. Influence of absorbable and nonabsorbable lipids and lipidlike substances on drug bioavailability. Clin Pharmacol Ther 1989;45:299–304.
20. Chavez A, DelConte A. A comparison of cycle control with monophasic levonorgestrel/ethinylestradiol 100 micrograms/20micrograms versus triphasic norethindrone/ethinylestradiol 500-750-1000 micrograms/35micrograms: a multicenter, randomized, open-label study. Eur J Contracept Reprod Health Care 1999;4:75–83.
21. Bannemerschult R, Hanker JP, Wunsch C, Fox P, Albring M, Brill K. A multicenter, uncontrolled clinical investigation of the contraceptive efficacy, cycle control and safety of a new low dose contraceptive containing 20micrograms ethinyl estradiol and 100micrograms levonorgestrel over six treatment cycles. Contraception 1997;56:285–90.
22. DelConte A, Loffer F, Grubb GS. Cycle control with oral contraceptives containing 20 micrograms of ethinyl estradiol. A multicenter, randomized comparison of levonorgestrel/ethinyl estradiol (100 micrograms/20micrograms) and norethindrone/ethinyl estradiol (1000micrograms/20mircrograms). Contraception 1999;59:187–93.
23. Endrikat J, Jaques MA, Mayerhofer, M, Pelissier C, Muller U, Dusterberg B. A twelve-month comparative clinical investigation of two low-dose oral contraceptives containing 20 micrograms ethinyl estradiol/75 micrograms gestodene and 20 micrograms ethinyl estradiol/150 micrograms desogestrel, with respect to efficacy, cycle control and tolerance. Contraception 1995;52:229–35.
24. Fitzgerald C, Feichtinger W, Spona J, Elstein M, Ludicke F, Muller U, et al. A comparison of the effects of two monophasic, low dose oral contraceptives on the inhibition of ovulation. Adv Contracept 1994;10:5–18.
25. Serfaty D, Vree ML. A comparison of the cycle control and tolerability of two ultra low-dose oral contraceptives containing 20 micrograms ethinyl estradiol and either 150 micrograms desogestrel or 75 micrograms gestodene. Eur J Contracept Reprod Health Care 1998;3:179–89.
26. 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.
27. Jung-Hoffman C, Kuhl H. Intra-and interindividual variations in contraceptive steroid levels during 12 treatment cycles: no relation to irregular bleedings. Contraception 1990;42:423–38.
28. Saleh W, Burkman R, Zacur H, Kimball AW, Kwiterovich P, Bell WK. A randomized trial of three oral contraceptives: comparison of bleeding patterns by contraceptive types and steroid levels. Am J Obstet Gynecol 1993;168:1740–5; discussion 1745—7.