Superior cycle control with a contraceptive vaginal ring compared with an oral contraceptive containing 30 µg ethinylestradiol and 150 µg levonorgestrel: a randomized trial

K. Oddsson1, B. Leifels-Fischer2, D. Wiel-Masson3, N.R. de Melo4, C. Benedetto5, C.H.J. Verhoeven6 and T.O.M. Dieben6,7

1 Women's Health Associates, Kopavogur, Iceland, 2 Birnauer Srasse, Munich, Germany, 3 Boulevard Chasles, Chartres, France, 4 Department of Gynaecology and Obstetrics, Hospital das Clínicas of São Paulo University, Brazil, 5 Department of Obstetrics and Gynecology, University of Turin, Italy and 6 Clinical Development Department, NV Organon, Oss, The Netherlands

7 To whom correspondence should be addressed at: NV Organon, Clinical Development Department, P.O.Box 20, 5340 BH Oss, The Netherlands. Email: thom.dieben{at}organon.com


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: This trial was conducted to compare cycle control with vaginal ring a combined contraceptive vaginal ring, and a combined oral contraceptive (COC) delivering 30 µg ethinylestradiol (EE) and 150 µg levonorgestrel. METHODS: This open-label, randomized, multi-centre, Phase III study involved adult women from 11 countries. Subjects were treated with either vaginal ring or a COC for 13 cycles (12 months). RESULTS: A total of 1030 subjects (vaginal ring, n=512; COC, n=518) comprised the intention-to-treat (ITT) population. The percentage of women in the ITT population who completed the trial was 70.9% for vaginal ring and 71.2% for the COC group. The incidence of breakthrough bleeding and spotting over cycles 2–13, the primary efficacy parameter, was lower with vaginal ring (range 2.0–6.4%) than the COC (range 3.5–12.6%), and for cycles 2 and 9 the lower incidence with vaginal ring was confirmed as statistically significant (P=0.003 and P=0.002 respectively). The incidence of intended bleeding was significantly higher over all cycles with vaginal ring (58.8–72.8%) than with the COC (43.4–57.9%). CONCLUSIONS: Cycle control with vaginal ring was excellent and superior to that of a COC containing 30 µg EE.

Key words: contraceptive/cycle control/irregular bleeding/NuvaRing/superiority


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Since the introduction of the first oral contraceptive, efforts have been made to improve contraceptive efficacy and tolerability. There has been a progressive reduction in the delivery of daily contraceptive steroid doses; in particular the dose of ethinylestradiol (EE) has been reduced to ≤20 µg daily. However, cycle control, a factor that influences contraceptive acceptability, convenience and compliance, has the potential to be compromised with low doses of estrogen (Åkerlund et al., 1993Go; Rosenberg et al., 1995Go, 1999Go; Kaunitz, 1998Go; Rosenberg and Waugh, 1998Go).

Modification of the oral contraceptive dosing regimen is one potential approach to improving cycle control (Endrikat et al., 2001bGo). For a COC containing 15 µg EE, shortening the pill-free period from 7 to 4 days has been shown to significantly shorten the duration of withdrawal bleeding and reduce bleeding intensity (Gestodene Study Group 324, 1999Go). However, alternative routes of administration may improve cycle control and compliance and offer other benefits, such as avoidance of first pass metabolism and food interactions, and greater convenience. A combined contraceptive vaginal ring has been developed (NuvaRing; NV Organon, The Netherlands) that delivers 15 µg EE and 120 µg etonogestrel (ENG) per day over 3 consecutive weeks. The controlled release design of the vaginal ring produces more uniform circulating concentrations of contraceptive hormones and avoids the daily fluctuations in these hormones associated with the use of COC. NuvaRing's once-monthly dosing means that users do not have to remember to take a pill every day, thus providing greater convenience.

The excellent efficacy, compliance, user acceptability and tolerability profile of NuvaRing have been established in previous large-scale studies. These studies also demonstrated NuvaRing's excellent cycle control (Roumen et al., 2001Go; Dieben et al., 2002Go; Novák et al., 2003Go). In a combined analysis of three small-scale, open-label comparative trials over six treatment cycles, NuvaRing was shown to have superior efficacy and similar tolerability to a combined oral contraceptive (COC) containing 30 µg EE and 150 µg levonorgestrel (LNG) (Bjarnadóttir et al., 2002Go). However, these studies were small and lacked sufficient statistical power to allow definitive conclusions to be drawn.

The current study is the first large-scale, 12 month, randomized controlled trial to compare the vaginal bleeding characteristics of NuvaRing with those of a well-established, commonly used COC (delivering 30 µg EE and 150 µg LNG daily).


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
This Phase III, open-label, randomized, comparative, multi-centre trial was conducted in Belgium, Brazil, Chile, Denmark, Finland, France, Germany, Italy, Norway, Spain and Sweden. The study was approved by the independent ethics committee/institutional review boards of the participating centres and was conducted in accordance with the Declaration of Helsinki and the ICH Guideline for Good Clinical Practice. All subjects provided written informed consent.

The primary objective of this trial was to compare the vaginal bleeding characteristics of NuvaRing (NV Organon) with a standard COC containing 30 µg EE and 150 µg LNG (Microgynon®; Schering AG, Germany). Secondary objectives included comparing efficacy, safety and compliance and these data will be described elsewhere.

Subjects
Healthy women aged ≥18 years who were of childbearing potential and seeking contraception were recruited. Important exclusion criteria included: contraindications for contraceptive steroids; previous use of an injectable hormonal method of contraception within 6 months of the start of trial medication; post-partum or post-abortion within 2 months of the start of trial medication; breastfeeding within 2 months of the start of trial medication; an abnormal cervical smear diagnosed during screening; or use of drugs that interfere with contraceptive hormone metabolism.

Treatment
Subjects were randomized to 13 cycles of treatment with NuvaRing or a COC using an Interactive Voice Response System (IVRS), which provided the treatment group and associated medication number. Each cycle consisted of 3 weeks of ring/COC use followed by a 1 week ring-/pill-free period. For the first cycle, the first day of ring insertion/pill intake was dependent on the prior method of contraception.

Ring insertion
On study entry, subjects received verbal and written instructions on ring use, and received one ring for each cycle. Women taking no hormonal contraception inserted the ring between days 1 and 5 of the menstrual cycle, according to instructions in the package insert. Women using hormonal contraception followed the instructions in the package insert, according to the method they were using.

Pill intake
Women taking no hormonal contraception took one tablet daily for 21 days, starting on the first day of the menstrual cycle. Women using contraception followed the advice on how to start, provided in the package insert for the method they were using.

Tablets were taken in the morning; however, if menstrual bleeding started in the afternoon or evening, the first COC pill was taken on the first morning after the start of bleeding.

Assessments
Assessments occurred at screening (within 1 month before starting treatment), within 1 week after the ring-/pill-free period of cycles 3, 6 and 9 and after cycle 13 or premature discontinuation. At screening and the last study visit, all subjects underwent physical and gynaecological examinations, including a cervical smear test. General medical and gynaecological histories were taken at screening.

Compliance
Subjects used diary cards to record ring/pill use and to document vaginal bleeding patterns. For NuvaRing, a cycle was considered compliant if the period of ring use did not deviate by >48 h from the scheduled 21 days (i.e. from 19 x 24 to 23 x 24 h) and if the ring-free period did not deviate by >24 h. In the COC group, full compliance was defined as a cycle in which all scheduled pills were taken.

Cycle control
Vaginal bleeding was classified as spotting (requiring ≤1 pad/tampon per day) or bleeding (>1 pad/tampon per day). Withdrawal bleeding was classified as any bleeding/spotting that occurred during the ring-/pill-free week; any withdrawal bleeding/spotting starting just before the ring-/pill-free week was termed early withdrawal bleeding; continued withdrawal bleeding was defined as any withdrawal bleeding/spotting that continued into the ring/pill period of the next cycle. All other bleeding/spotting (i.e. breakthrough bleeding/spotting during the 3 week ring/pill period that was not part of early or continued withdrawal bleeding) was classified as irregular bleeding. Intended bleeding was defined as a cycle with withdrawal bleeding but without early or continued withdrawal bleeding and no irregular bleeding. A graphic representation of cycle control parameters is presented in Figure 1.



View larger version (17K):
[in this window]
[in a new window]
 
Figure 1. Illustration of cycle control parameters in relation to ring or pill treatment.

 
Statistical methods
All analyses were performed on both the intention-to-treat (ITT) population (all randomized women who started treatment) and the per protocol (PP) population (all randomized and treated women without any major protocol violations).

Cycle control data were compared between groups using the two-sided Fisher's exact test adjusting for country (if needed, approximated by Cornfield's method) for the ITT and PP populations. The sample size was based on a conservative assumption that 500 randomized subjects per treatment group would result in 350 evaluable ITT cycles in cycles 2–13 for both treatment arms. The trial was designed to detect a statistically significant lower incidence of a breakthrough bleeding/spotting episode for at least one of the cycles 2–13 with >80% probability.

To avoid any possible bias in favour of NuvaRing (due to the different ring and pill starting procedures), the breakthrough bleeding and spotting patterns (primary efficacy variable) in cycle 1 were excluded and assessed for cycles 2–13 only. This parameter was analysed by two-sided testing at the conventional 5% level, corrected for multiplicity by Bonferroni's rule (i.e. at the level {alpha}=1–(0.95)1/12=0.00427). This implied the superiority of NuvaRing if a statistically significant lower incidence was reported in at least one of the cycles 2–13 in the NuvaRing group versus the COC group. Where possible, secondary cycle control parameters were assessed for cycles 1–13. These parameters were analysed using Fisher's exact test adjusted for country at a 5% significance level without adjusting for multiplicity.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Subject disposition
Between September 2001 and August 2003, a total of 1079 subjects was randomized to treatment. Of these, 49 women discontinued prior to treatment; nine (NuvaRing, n=6; COC, n=3) withdrew as they were pregnant at baseline, three (NuvaRing, n=3) because of loss to follow-up, 15 (NuvaRing, n=7; COC, n=8) because of unwillingness to cooperate and 22 (NuvaRing, n=13; COC, n=9) for other reasons.

Of the randomized subjects, 1030 received treatment (NuvaRing, n=512; COC, n=518) and comprised the ITT population. Of these, 298 subjects discontinued treatment prematurely (NuvaRing, n=149; COC, n=149), primarily because of adverse events (NuvaRing, n=58; COC, n=45) and lost to follow-up (NuvaRing, n=33; COC, n=33) (Figure 2). The PP population comprised 899 subjects (NuvaRing, n=440; COC, n=459). The percentage of women in the ITT population who completed the trial was 70.9% for NuvaRing and 71.2% for the COC group.



View larger version (26K):
[in this window]
[in a new window]
 
Figure 2. Subject disposition. ITT = intention-to-treat; COC = combined oral contraceptive.

 
Baseline demographic and clinical characteristics are shown in Table I. There were no notable differences between the two treatment groups. The most commonly used last contraceptive method before the study in both groups was oral contraceptives (65.4% in the NuvaRing group and 66.2% in the COC group). Foam, condoms, suppositories or diaphragms comprised the next most commonly used group of contraceptive methods (19.3% in the NuvaRing group and 20.1% in the COC group).


View this table:
[in this window]
[in a new window]
 
Table I. Baseline demographic and clinical characteristics for the NuvaRing and combined oral contraceptive (COC) treatment groups (intention-to-treat population)

 
Ten pregnancies occurred during treatment in the ITT population (NuvaRing, n=5; COC, n=5). Pregnancy data will be discussed in further detail in a separate publication.

Compliance
Compliance with the prescribed regimen was high in both groups. For NuvaRing, 87.4% of ITT cycles were fully compliant compared with 86.6% of COC ITT cycles.

Cycle control
As results for the ITT and PP cycles were similar, data are only presented for the ITT analysis.

Irregular bleeding: breakthrough bleeding and spotting
For all cycles, the incidence of breakthrough bleeding and spotting was lower with NuvaRing than the COC (Figure 3) and this was statistically significant (P<0.003) for cycles 2 and 9. Because a statistically significant lower incidence was reported in at least one cycle (see Statistical methods), this implied that cycle control with NuvaRing was superior to the COC. In addition, analysis of results without correction for multiplicity revealed a statistically significant lower incidence of breakthrough bleeding with NuvaRing compared with the COC in cycles 2 and 3 and 8–10 (P<0.05 for all).



View larger version (22K):
[in this window]
[in a new window]
 
Figure 3. The incidence of breakthrough bleeding and spotting during the ring/pill period in subjects receiving NuvaRing or a combined oral contraceptive (COC) during cycles 2–13. *P<0.003 versus COC.

 
Breakthrough bleeding or spotting
Breakthrough bleeding in cycles 1–13 occurred less frequently with NuvaRing than the COC (Figure 4), and was significantly lower in cycles 2 (P=0.027) and 9 (P=0.008). The mean number of breakthrough bleeding and spotting days over cycles 1–13 ranged from 1.1 to 4.0 days for NuvaRing and from 2.5 to 4.2 days for the COC group.



View larger version (25K):
[in this window]
[in a new window]
 
Figure 4. The incidence of breakthrough bleeding or spotting during the ring/pill period in subjects receiving NuvaRing or a combined oral contraceptive (COC) during cycles 1–13. *P=0.027 versus COC, **P=0.008 versus COC (both for breakthrough bleeding only).

 
Withdrawal bleeding
Withdrawal bleeding occurred in the majority of ITT cycles in both treatment groups. The incidence of cycles (over 1–13) with no withdrawal bleeding was significantly lower (P=0.01) in the NuvaRing group than in the COC group for cycle 6 (Table II).


View this table:
[in this window]
[in a new window]
 
Table II. Withdrawal bleeding parameters (ranges for cycles 1–13)

 
There was no significant difference in the incidence of early withdrawal bleeding between treatment groups. For cycles 1–13, the incidence of early withdrawal bleeding with only spotting days in the ring/pill period was similar in the NuvaRing and COC groups.

The incidence of continued withdrawal bleeding was lower for the NuvaRing group than the COC group in cycles 1–12 and this was statistically significant for all cycles (P<0.02). Continued withdrawal bleeding with spotting days occurred less frequently in the NuvaRing group in cycles 1–12 and was statistically significant for cycles 1, 3–7, 9, 10 and 12 (P<0.05) compared with the COC group.

Over cycles 1–13, the mean number of withdrawal bleeding days ranged from 4.5 to 5.3 days in the NuvaRing group and from 4.3 to 5.8 days for the COC group.

Intended bleeding
The incidence of intended bleeding was significantly higher in the NuvaRing treatment group than the COC group for cycles 1–12 (P<0.005; Figure 5).



View larger version (44K):
[in this window]
[in a new window]
 
Figure 5. The incidence of intended bleeding in subjects receiving NuvaRing or a combined oral contraceptive (COC) during cycles 1–12. *P≤0.005 versus COC.

 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Good cycle control is a key factor influencing contraceptive acceptability and compliance. The findings from this 1 year study in 1030 women show that NuvaRing, a combined contraceptive vaginal ring, has excellent cycle control that is superior to that of a COC containing 30 µg EE and 150 µg LNG.

Various factors, including unwanted side-effects and bleeding irregularities, are considered as risk factors for poor contraceptive compliance (Rosenberg et al., 1995Go; Rosenberg and Waugh, 1999Go). The incidence of unexpected and irregular bleeding may contribute to a woman's decision to discontinue a particular contraceptive method; 33% of women who stop using oral contraceptives do so because of unexpected vaginal bleeding (Rosenberg et al., 1995Go). Consequently, compliance is likely to be better with a contraceptive that has good cycle control. Because NuvaRing is a vaginal method of contraception, good cycle control is important. Irregular bleeding might be inconvenient for users and could affect NuvaRing's acceptability.

The incidence of irregular bleeding (breakthrough bleeding and spotting) with NuvaRing was very low throughout our study (2.0–4.6%), and NuvaRing was significantly superior to the COC for this parameter. This observed incidence of irregular bleeding for NuvaRing confirms previous findings (Roumen et al., 2001Go; Dieben et al., 2002Go) and is consistent with the results of the smaller study, with the same COC (Bjarnadóttir et al., 2002Go). Furthermore, in our study, the majority of irregular bleeding with NuvaRing was restricted to spotting only, which is also consistent with the two large NuvaRing trials (Roumen et al., 2001Go; Dieben et al., 2002Go).

The low (2.0–4.6%) incidence of irregular bleeding with NuvaRing in the present study is particularly interesting, considering the low dose of 15 µg/day EE released by the ring—half the daily dose of the comparator in this study. This may be related to the achievement of constant serum concentrations of contraceptive hormones with NuvaRing, which is in contrast to the daily variations in concentrations seen with COC.

In the current study, the incidence of irregular bleeding with NuvaRing was low for the second cycle and this was maintained throughout the study. This is in contrast to the COC, where 12.6% of users experienced such bleeding during the second cycle of use. This frequent occurrence of irregular bleeding with the COC is consistent with results presented elsewhere for COC (Archer et al., 1999Go; Gestodene Study Group 322, 1999Go; Reisman et al., 1999Go; Endrikat et al., 2001aGo,bGo).

Regular monthly withdrawal bleeding is considered desirable as it reassures the user of continued absence of pregnancy; however, light bleeding that only lasts for a short period of time is clearly preferable. In both treatment groups, almost all women experienced withdrawal bleeding; the mean duration of withdrawal bleeding was similar and short in both groups. Intended bleeding describes an ‘ideal’ bleeding pattern in which regular withdrawal bleeding occurs during the ring- or pill-free week and no irregular bleeding occurs during ring or pill use. The incidence of intended bleeding occurred to a significantly greater extent in the NuvaRing group than the COC group.

Heavy and/or prolonged withdrawal bleeding is likely to inconvenience the user and, as with irregular bleeding, is a factor associated with poor compliance (Rosenberg et al., 1995Go; Rosenberg and Waugh, 1999Go). Early withdrawal bleeding occurred in a similar proportion of cycles for both groups but the incidence of continued withdrawal bleeding was significantly more frequent with the COC than with NuvaRing. This pattern of early and continued withdrawal bleeding is similar to that previously observed (Bjarnadóttir et al., 2002Go). Also, the majority of both early withdrawal bleeding and continued withdrawal bleeding was restricted to spotting only. These results are again consistent with those reported elsewhere for NuvaRing (Roumen et al., 2001Go; Dieben et al., 2002Go).

In conclusion, cycle control is known to influence contraceptive acceptability, compliance and convenience (Rosenberg et al., 1995Go; Rosenberg and Waugh, 1998Go). Because of the vaginal route of administration, good cycle control with NuvaRing is especially important. Despite the low dose of EE (15 µg/day), cycle control with NuvaRing was superior to that of a COC containing 30 µg EE and 150 µg LNG. In the current study, NuvaRing produced a statistically significant lower incidence of irregular breakthrough bleeding and spotting. NuvaRing provides excellent cycle control in a convenient form that does not require daily dosing.


    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The authors would like to thank and acknowledge the participation of the following investigators during this trial: Belgium: A.Saey, U.Gaspard, M.Renier; Brazil: M.Poli, E.Coutinho, E.A.J.Pellini, R.Pereira de Andrade, L.Bahamondes, M.E.H.D.Yazzle, M.A.Albernaz, J.Sabino Pinho Neto, C.E.Fernandes; Chile: J.C.Montero; Denmark: P.Hein, F.Ahlstrøm, K.Rubeck Petersen, L.Skive; Finland: A.Kivelä, M.Vartiainen, U.Ågren, S.Korhonen, P.Taipale, A.Kiilavuori; France: M.-L.Bollengier, J.Levy-Frebault, A.Mallet, A.Malvy, J.Sentenac, C.Matuchansky; Germany: W.Harlfinger, H.-P.Zahradnik, K.Muschel, F.Starflinger; Italy: G.B.Melis, A.Volpe, A.Genazzani; Norway: B.Eriksen, P.Lossius, K.A.Trosterud, T.Nygaard, M.Blakstad, N.Farnes; Spain: F.Martínez, C.Coll, R.Sánchez, A.Albert; Sweden: B.-M.Landgren, M.van Rooijen, I.Östlund, T.Bäckström, O.Tydén. This study was supported by NV Organon, Oss, The Netherlands. T.Dieben and C.Verhoeven are employees of NV Organon.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Åkerlund M, Røde A and Westergaard J (1993) 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 oestradiol. Br J Obstet Gynaecol 100, 832–838.[ISI][Medline]

Archer DF, Maheux R, DelConte A and O'Brien FB (1999) Efficacy and safety of a low-dose monophasic combination oral contraceptive containing 100 microg levonorgestrel and 20 microg ethinyl estradiol (Alesse). Am J Obstet Gynecol 181, 39–44.[ISI][Medline]

Bjarnadóttir RI, Tuppurainen M and Killick SR (2002) Comparison of cycle control with a combined contraceptive vaginal ring and oral levonorgestrel/ethinylestradiol. Am J Obstet Gynecol 186, 389–395.[CrossRef][ISI][Medline]

Dieben TO, Roumen FJ and Apter D (2002) Efficacy, cycle control, and user acceptability of a novel combined contraceptive vaginal ring. Obstet Gynecol 100, 585–593.[Abstract/Free Full Text]

Endrikat J, Cronin M, Gerlinger C, Ruebig A, Schmidt W and Dusterberg B (2001a) Double-blind, multicenter comparison of efficacy, cycle control, and tolerability of a 23-day versus a 21-day low-dose oral contraceptive regimen containing 20 microg ethinylestradiol and 75 microg gestodene. Contraception 64, 99–105.[CrossRef][ISI][Medline]

Endrikat J, Cronin M, Gerlinger C, Ruebig A, Schmidt W and Dusterberg B (2001b) Open, multicenter comparison of efficacy, cycle control, and tolerabiltiy of a 23-day oral contraceptive regimen with 20 microg ethinylestradiol and 75 microg gestodene and a 21-day regimen with 20 microg ethinylestradiol and 150 microg desogestrel. Contraception 64, 201–207.[CrossRef][ISI][Medline]

Gestodene Study Group 322 (1999) The safety and contraceptive efficacy of a 24-day low-dose oral contraceptive regimen containing gestodene 60 µg and ethinylestradiol 15 µg. Eur J Contracept Reprod Health Care 4, 9–15.

Gestodene Study Group 324 (1999) Cycle control, safety and efficacy of a 24-day regimen of gestodene 60 µg and ethinylestradiol 15 µg and a 21-day regimen of desogestrel 150 µg/ethinylestradiol 20 µg. Eur J Contracept Reprod Health Care 4, 17–24.

Kaunitz AM (1998) Oral contraceptive oestrogen dose considerations. Contraception 58, 15S–21S.[CrossRef][ISI][Medline]

Novák A, de la Loge C, Abetz L and van der Meulen EA (2003) The combined contraceptive vaginal ring, NuvaRing: an international study of user acceptability. Contraception 67, 187–194.[CrossRef][ISI][Medline]

Reisman H, Martin D and Gast MJ (1999) A mutlicenter randomized comparison of cycle control and laboratory findings with oral contraceptive agents containing 100 microg levonorgestrel with 20 microg ethinylestradiol or triphasic norethindrone with ethinylestradiol. Am J Obstet Gynecol 181, 45–52.[Medline]

Rosenberg MJ and Waugh MS (1998) Oral contraceptive discontinuation: a prospective evaluation of frequency and reasons. Am J Obstet Gynecol 179, 577–582.[ISI][Medline]

Rosenberg MJ and Waugh MS (1999) Causes and consequences of oral contraceptive noncompliance. Am J Obstet Gynecol 180, 276–279.[ISI]

Rosenberg MJ, Waugh MS and Meehan TE (1995) Use and misuse of oral contraceptives: risk indicators for poor pill taking and discontinuation. Contraception 51, 283–288.[CrossRef][ISI][Medline]

Rosenberg MJ, Meyers A and Roy V (1999) Efficacy, cycle control, and side effects of low- and lower-dose oral contraceptives: a randomized trial of 20 micrograms and 35 micrograms oestrogen preparations. Contraception 60, 321–329.[CrossRef][ISI][Medline]

Roumen FJ, Apter D, Mulders TM and Dieben TO (2001) Efficacy, tolerability and acceptability of a novel contraceptive vaginal ring releasing etonogestrel and ethinyl oestradiol. Hum Reprod 16, 469–475.[Abstract/Free Full Text]

Submitted on August 13, 2004; accepted on October 15, 2004.