Bleeding patterns and patient acceptability of standard or continuous dosing regimens of a low-dose oral contraceptive: a randomized trial

1. Introduction 

The administration of combined oral contraceptive (COC) in 21-day cycles of active hormone pills followed by a 7-day placebo or pill-free interval produces predictable withdrawal bleeding in most users. Some women, however, experience nuisance breakthrough bleeding, spotting or amenorrhea. Estrogen dose and progestin formulation influence bleeding patterns. In general, randomized studies have documented that women using 30 μg ethinyl estradiol (EE) pills have less breakthrough bleeding and spotting than 20 μg pills, and that levonorgestrel (LNG)-containing products generally have better bleeding patterns than products containing norethindrone [1], [2], [3]. A survey of Australian women revealed that, when given the choice, they would prefer not to experience any bleeding if there were no adverse health consequences [4].

Although bleeding problems increase, lowering the dose of EE may reduced the risk of serious adverse events such as venous and arterial thrombosis [5], [6] and stroke [7]. In addition, new formulations with 20 μg of EE may decrease the likelihood of less serious affective side effects. A recent trial found no significant difference in weight gain or in the rates of headache, nausea and breast pain between women randomized to receive 20 μg EE/100 μg LNG or placebo over six cycles [8]. Although comparisons using active treatment arms have not demonstrated fewer symptoms or greater compliance in women using the lower estrogen products, these studies failed to keep progestogen product and dose constant [2], [3], [9], [10].

Continuous administration of combined oral contraceptives is often recommended to women to avoid menstrual symptoms including migraine headache [11], endometriosis-associated pelvic pain [12], premenstrual symptoms [13] or inconvenience associated with menses [13]. Avoidance of these unpleasant symptoms may be a more powerful motivator toward compliance than the many known major health benefits, such as reduction in the risk of ovarian cancer. However, few studies have been published comparing the safety, efficacy, side effects and potential advantages and disadvantages of continuous versus traditional administration of oral contraceptives [14].

In order to further the understanding of continuous dosing of oral contraceptives, we designed a study to investigate the hypothesis that women taking combined oral contraceptives without a hormone-free interval have fewer side effects and bleeding days when compared to cyclic administration. Our secondary outcome was to determine whether patients considered this an acceptable method of administration.

2. Materials and methods 

The purpose of this study was to compare the bleeding patterns and acceptance of combined 0.1 mg LNG/20 μg EE taken with and without a hormone-free interval. The study design involved randomization of subjects to six standard 28-day cycles (21 days of active pills and a 7-day hormone-free interval) or 168 continuous days of pill administration without a hormone-free interval. Enrolled subjects received free oral contraceptives. The oral contraceptive pills were obtained through an unrestricted gift from Wyeth Pharmaceuticals. The Institution Review Board of the Oregon Health and Science University (OHSU) approved the study protocol. The study was conducted at the OHSU Women’s Health Clinic between August 2000 and May 2001. Women presenting to the clinic requesting oral contraceptive therapy were given the opportunity to participate in the study provided they were between the ages 18 and 50, willing and able to give informed consent and had no medical contraindication to combined oral contraceptive therapy. Subjects agreed to treatment group assignment by randomization. Subjects and investigators were not blinded to treatment group.

The primary outcome in the study was the total number of bleeding or spotting days occurring over the 6-month duration of treatment. Secondary outcomes included severity of a number of daily symptoms, and overall treatment satisfaction. At enrollment, participants were provided with a calendar and asked to prospectively document any vaginal bleeding events and side effects. Vaginal bleeding was categorized as spotting and bleeding. Spotting was defined as light flow that did not necessitate sanitary protection. Bleeding was defined as heavier flow that required sanitary protection. Subjects reporting bleeding were further asked to subjectively classify each day’s episode as light, moderate or heavy. Participants were also asked to make an additional daily notation regarding symptoms of headache, nausea, bloating, breast tenderness, premenstrual syndrome (PMS) and menstrual pain. PMS was defined as mood-related symptoms not included within the other categories. At the conclusion of the study, subjects completed a satisfaction questionnaire. Overall satisfaction with bleeding patterns and symptom severity was recorded on a 100-mm visual analog scale (VAS).

In order to address potential safety concerns of uninterrupted administration of an oral contraceptive combination on the endometrium, women in the continuous group were asked to volunteer for a transvaginal ultrasound examination to document endometrial stripe thickness. A single measurement of maximum endometrial stripe thickness was obtained in the anterior-posterior orientation using a GE model L418830 scanner equipped with a 6.5-MHz vaginal transducer. The protocol requested women with an endometrial stripe thickness in excess of 5 mm to consent to endometrial biopsy.

We based our sample size on expected differences in the primary outcome of total bleeding days. Premarketing studies of 20 μg EE/100 μg LNG administered cyclically documented a mean of 4.8 days of withdrawal bleeding and an 11% incidence of intramenstrual bleeding [15]. A conservative estimate of total bleeding days within this group would be an average of 5 days/cycle, or 30 days during a 6-month (26-week) trial. Our clinical experience with continuous administration of 20 μg EE/100 μg LNG suggests that many subjects experience spotting during the first two to three cycles of use, and then become amenorrheic. We made the assumption that a 9-day reduction in bleeding, to 21 days over 13-cycle packs (26 weeks) of administration would represent a clinically significant outcome. With an expected standard deviation of 9 days, this provided a standardized effect size of 1.0, and an estimated sample size of 16 in each group [16].

Within each category of bleeding and symptoms, the mean number of days reported by subjects within each group was analyzed for each of the six treatment cycles and for the overall duration of the study. Study discontinuation, pregnancy and adverse events were also tracked. In the event of withdrawal from the study, data were collected and analyzed up to the point of termination. For all variables, mean values were analyzed using the t-test, ordinal variables with the Mann-Whitney U-test and categorical variables with the chi-square test. All statistical tests were performed on a desktop PC running SPSS for Windows (version 10.1; SPSS Corporation, Chicago, IL, USA).

3. Results 

Thirty-two women were enrolled and randomized to six standard 28-day cycles or 168 days without a hormone-free interval. There were two participants who discontinued the study medication. One discontinued in the continuous group due to excessive bleeding, and the other in the cyclic group for intolerable mood changes. Data for each participant were used up to the point of termination of the study, according to intent-to-treat. Two participants were lost to follow-up. One pregnancy occurred in the continuous administration group after completion of the study.

Of the 32 participants randomized, 28 (87%) completed the study and returned completed calendars and surveys for analysis. There were no significant differences between the two groups with regard to age, parity, weight, tobacco use, marital status, prior birth control methods (including prior use of oral contraceptives), prestudy cycle length and bleeding patterns or perceived menstrual side effects (Table 1).

Table 1. Characteristics of the study group

	Characteristic		Cyclic (n = 16)	Continuous (n = 16)a
	Age (years), mean (SD)		26.5 (4.5)	29.3 (6.7)	NS
	Parity mean (SD)		0.19 (0.8)	0.63 (1.2)	NS
	Weight (lbs) mean (SD)		157.8 (58.1)	156.3 (36.7)	NS
	Smoking n (%)		2 (12.5)	1 (6.3)	NS
	Prior bleeding patterns (days), mean (SD)	Cycle length	27.8 (2.5)	29 (5.9)	NS
		Duration	4.4 (1.3)	4.4 (1.4)	NS
	Prior contraceptive method n (%)	None	4 (25)	3 (18.8)	NS
		Cyclic COC	10 (62.5)	9 (56.2)	NS
		Continuous	0	2 (12.5)	NS
		Condoms	2 (12.5)	2 (12.5)	NS
	Marital status n (%)	Single	7 (43.8)	3 (18.8)	NS
		Partner	4 (25)	6 (37.5)	NS
		Married	5 (31.3)	6 (37.5)	NS
		Divorced	0	1 (6.3)	NS

a NS = not significant; COC = combined oral contraceptives.

The data on bleeding days were analyzed with respect to the six 28-day increments as well as the total cumulative days of the study (Fig. 1, Fig. 2, Fig. 3, Fig. 4). The mean cumulative total bleeding and spotting days over the course of the study in the cyclic group was 34.9 (SD = 16.9) compared with 25.9 (SD = 29.2) in the continuous group, a difference of 9 days (p = 0.33). Evaluating the reported types of flow and cycle number revealed no differences in the number of days of spotting (bleeding not requiring sanitary protection). However, subjects randomized to the continuous cycle group required sanitary protection for bleeding significantly less often over the duration of the study (18.4 vs. 33.8 days, p < 0.01), and in every cycle except cycle one (Fig. 3). Overall, women in the continuous group required protection on average less than 2 days compared to almost 4 days among those taking the pill in the standard fashion (p < 0.01).

View Large Image Download to PowerPoint

Fig. 1. Mean number of total bleeding and spotting days in each treatment cycle.

View Large Image Download to PowerPoint

Fig. 2. Mean number of total spotting days in each treatment cycle.

View Large Image Download to PowerPoint

Fig. 3. Mean number of bleeding days in each treatment cycle that required protection. Asterisks indicate significant difference (p < 0.05) within a cycle.

View Large Image Download to PowerPoint

Fig. 4. Percent of subjects reporting no bleeding in each treatment cycle. Asterisks indicate significant difference (p < 0.05) within a cycle.

Women randomized to the continuous administration group were also more likely to be amenorrheic, and this difference reached statistical significance in all cycles except one and three. Only one woman taking pills in the standard fashion reported amenorrhea in cycles 1, 3 and 5. Although 9 of the original 16 subjects assigned to the continuous cycle group reported amenorrhea by cycle 5, by cycle 6 only 5 remained without bleeding (Fig. 4).

Data regarding subjective side effects were collected for daily events of headache, bloating, breast tenderness, nausea, depression, PMS and menstrual pain. Headache was reported infrequently and there was no significant difference between the groups. This was also true of breast tenderness, nausea, depression and PMS. Overall, however, subjects reported significantly less bloating in the continuous group (mean = 0.7 vs. 11.1 days, p = 0.04), and in cycles 3, 5 and 6. Significantly less menstrual pain was also noted in the continuous group overall (mean = 1.9 vs. 13.3 days, p < 0.01), and in each cycle except cycle one.

Satisfaction with bleeding patterns was evaluated at the end of the study with a visual analog scale of 100 mm. Subjects in the cyclic and continuous groups both reported a high level of satisfaction with the bleeding patterns they experienced (mean = 71.2 mm and 70.53 mm, respectively, p = 0.95).

Endometrial stripe measurements were obtained in all 14 women in the continuous administration group. The mean thickness ranged from 2 to 4 mm with a mean of 3.3 mm (SD = 0.73). No endometrial biopsies were necessary.

4. Discussion 

Our study sought to determine whether women taking combined oral contraceptives would prefer to take them in a continuous manner rather than in the standard regimen. Our results demonstrate that both methods of administration have equivalent subjective acceptability. Our primary objective outcome was the total number of bleeding and spotting days over six cycles, and it was hypothesized that there would be less bleeding and spotting in the continuous group. Although the observed difference in bleeding and spotting days (nine fewer in the continuous administration group) matched the deviation on which our sample size was estimated, this failed to reach statistical significance. A relatively high, unanticipated dropout rate explains this outcome; four subjects (two in each group) did not complete the study. Thus, although our effect size was as expected, the final sample size of 28 was insufficient to document a significant difference in overall bleeding and spotting of 9 days.

Although continuous and standard methods of administration of oral contraceptives both resulted in acceptable bleeding patterns and tolerable affective side effects, women randomized to the continuous group reported significantly fewer days of bleeding that required sanitary protection and less severe symptomatology. Although participants reported similar numbers of spotting days, significantly fewer light and moderate bleeding days occurred in the continuous group. Women in the continuous group also reported less bloating and menstrual pain. Our results reinforce the conclusions reached in other studies that document a reduction in bleeding and symptoms with extended use [17] or continuous administration [14] of oral contraceptives.

As we expected women using the pills in the standard fashion to experience cyclic withdrawal bleeding, we felt that it would be impossible to blind subjects to treatment group. Although knowledge of treatment assignment could potentially have influenced the reporting of subjective side effects, the primary outcome, number of bleeding days, is objective and less likely to have been affected by lack of blinding.

Although spotting does not require the use of menstrual pads or tampons, most women still regard it as a nuisance. In our study, women reported no overall difference in the numbers of spotting days regardless of method of administration. Our rate of spotting was similar to that observed in a large efficacy trial of standard dosing of 20 μg EE/100 μg LNG [18]. Although the cause of spotting is unknown, it likely relates to progestin-induced changes of small endometrial blood vessels [19]. Our ultrasound data suggests that continuous use of oral contraceptives results in an inactive endometrium, similar to that obtained with combined continuous hormone replacement therapy.

One pregnancy occurred in the continuous group after completion of the study, and discontinuation of oral contraceptives. This patient had experienced complete amenorrhea throughout the study. A transvaginal ultrasound performed at the time of study discontinuation documented a 4-mm endometrial stripe, consistent with inactive endometrium. After the subject failed to menstruate 4 weeks after stopping the study medication, a pregnancy test was positive. A follow-up ultrasound done at that time suggested an estimated gestational age (EGA) corresponding to approximately 10 to 14 days at the time of study termination. As true embryonic age is estimated as 14 days shorter than EGA, the pregnancy likely was conceived shortly after discontinuation of the method. As women on oral contraceptives occasionally ovulate [20], this pregnancy is not surprising, but inconsistent with data that reducing the pill-free interval reduces the risk of ovulation [21]. The occurrence of this pregnancy underscores the importance of the multiple mechanisms of action of oral contraceptives, as it is likely that discontinuation of the pill at that time reduced the known contraceptive effect of progestins on cervical mucus [22]. The patient continued her pregnancy and had a normal course. Although no conclusions regarding differential efficacy or long-term effects between continuous and cyclic use of oral contraceptives are possible with a study of this size, this outcome provides evidence for rapid reversibility of continuous dosing, a concern of many women who might consider this method.

Noncontraceptive benefits represent an important additional indication for the use of combined oral contraceptives. Providers commonly prescribe the pill to bestow relief from a variety of cycle-related symptoms such as breast tenderness, bloating, premenstrual symptoms and menstrual pain. Our data suggests that continuous administration of oral contraceptives results in less bloating and menstrual pain than cyclic administration. In fact, none of the measured outcomes in our study resulted in a greater rate of affective symptoms for the continuous group. The potential for alleviation of symptoms may be a powerful motivator toward the choice (or continuation) of a specific contraceptive regimen for both provider and patient [23]. Our data suggest that continuous administration will result in additional symptom reduction for women who complain of bloating and menstrual pain.

Inconvenience and/or symptoms associated with menses represent key concerns of women in the reproductive years that providers should address during routine health care visits and contraceptive counseling. Our results with 20 μg EE/100 μg LNG suggest that for women with significant menstrual-related symptoms, or for others who might simply prefer to avoid menstruation, continuous administration will result in fewer bleeding days that require protection and less pain and bloating. Additional studies with a larger sample of women are needed to confirm these results. Future randomized studies should direct attention to estrogen dose, and choice of progestin to determine an optimum regimen for continuous oral contraceptive dosing.