Dieben, Thom O. M. MD; Roumen, Frans J. M. E. MD, PhD; Apter, Dan MD
Combined oral contraceptives have been available for four decades. Their development has focused on lowering the dosage of ethinylestradiol and progestogens and using newer, more selective progestogens.1 However, reducing the dose of estrogen below 20 μg/day of ethinylestradiol adversely affects cycle control; cycle control is a key factor affecting contraceptive compliance.2 Oral administration of contraceptives has some disadvantages, including hepatic first‐pass metabolism of the contraceptive steroids and susceptibility to reduced uptake because of vomiting or food interactions. In addition, all combined oral contraceptives rely on daily pill intake; this results in daily fluctuations in hormone levels, and may also account for poor compliance in some users.3–5 Furthermore, a recent study of women's wishes in relation to contraceptive methods showed that most American women would prefer a contraceptive that requires monthly, not daily, administration (details of the methodology of this survey have yet to be published, however).6
These observations illustrate the need for new, more convenient methods of hormonal contraception. In response, several alternatives have been developed using different routes of administration and controlled‐release formulations, but cycle control was not always optimal. At present, injections or implantable rods and hormone‐releasing intrauterine devices are the only widely available nonoral hormonal contraceptives, most of which contain only a progestogen. These products have several disadvantages that make them less acceptable to many women, including the need for administration by trained personnel and the unpredictable bleeding patterns associated with progestogen‐only formulations.
Because of such disadvantages, other approaches have been investigated, which led to the development of vaginal rings to administer contraceptive steroids.7,8 Vaginal administration has several advantages including the use of controlled‐release formulations, easy insertion and removal by the user herself, no need for daily intervention, a rapid return to normal cycling after removal, and the possibility of administering both estrogens and progestogens. Vaginal administration also allows a lower dose of estrogen to be administered than with oral forms, which may reduce the occurrence of dose‐related adverse events.
Results have been published from preliminary trials of a vaginal ring designed to release norethindrone acetate and ethinylestradiol over 1 year. Although good efficacy was achieved, the ring did not maintain steroid release in the latter months of use.7 More recently, a combined contraceptive vaginal ring that releases 15 μg of ethinylestradiol and 120 μg of the progestogen etonogestrel per day, at constant rates, has been developed.9 This contraceptive is a flexible, transparent ring made of evatane (ethinyl vinyl acetate), with an outer diameter of 54 mm and a cross‐section of 4 mm. Each ring is intended for one cycle of use only, comprising 3 weeks of ring use and a 1‐week ring‐free period. Recently published data show that the ring completely inhibits ovulation.10 A separate study of 1145 women showed that the ring has good contraceptive efficacy and cycle control, is well tolerated, and highly acceptable to users.8
This international study was designed to estimate the contraceptive efficacy, cycle control, tolerability, and user acceptability of the combined contraceptive ring in a large group of women.
MATERIALS AND METHODS
Two open‐label, noncomparative studies of the contraceptive efficacy, cycle control, tolerability, and user acceptability of a combined contraceptive vaginal ring (NuvaRing, NV Organon, Oss, the Netherlands) were performed using similar protocols. The data represent the combined findings of these two studies. One study was performed in 52 European centers; the other study was carried out in 48 centers in Canada and the United States. The findings from the European part of this study have been published previously.8 The study was performed in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines, and with local ethical committee approval from each center. Participants did not receive any financial compensation except for reimbursement of travel expenses.
The study planned to recruit 2400 women aged between 18–40 years at risk of pregnancy and asking for contraception. The exclusion criteria for the study included: contraindications to contraceptive steroids; previous use of an implant or hormone‐releasing intrauterine device (within 2 months), or an injectable hormonal method of contraception (within 6 months); use of drugs that interfere with the metabolism of contraceptive steroids, or any sex steroid other than the study medication; a cervical smear result of Papanicolaou grade III, IV, or V (or in the Bethesda System, low‐grade squamous intraepithelial lesions or higher); the presence of certain conditions relevant to ring use such as cervicitis, vaginitis, or a bleeding cervical erosion; cystocele, rectocele, or prolapse of the uterine cervix; severe or chronic constipation; dyspareunia, or other coital problems.
Women received a different ring for each cycle of use. For the first cycle, the first day of ring insertion depended on the prior method of contraception. Women who were not using oral contraceptives inserted the ring between days 1 and 5 of the menstrual cycle; these women were instructed to use an additional form of contraception (such as a condom) during the first 7 days of ring use. Women who were using the combined oral contraceptive pill started using the ring 7 days after the start of the usual pill‐free or placebo period. Women who were taking a progestogen‐only pill continued pill intake until the day before ring insertion; if such a user had regular vaginal bleeding, she could also insert the ring on the fifth day after the start of vaginal bleeding. All subjects received written and verbal instructions on ring use, including the methods of insertion and removal. Each ring was inserted and removed by the woman herself, and she was allowed to remove the ring temporarily for intercourse providing that it was reinserted within 3 hours of its removal.
The intended treatment period of this study was 13 cycles of ring use. Each 28‐day cycle consisted of 3 weeks of ring use followed by a 1‐week ring‐free period.
Study assessments were scheduled at the time of initial screening, in the first week after cycles 3, 6, and 9, and after cycle 13, or premature discontinuation. At screening, medical and gynecologic histories of all subjects were obtained, and they underwent a general physical examination, including a pelvic examination with cervical cytology sampling. At all study visits, blood pressure, heart rate, and body weight were measured. In addition, subjects were questioned on the use of any concomitant medication. The pelvic examination and cervical cytologic test were repeated at cycle 6 and at the end of the study treatment. The physical examination was repeated at the last visit.
Cervical cytology was assessed at a central laboratory and classified using Papanicolaou class and/or the Bethesda system.
All subjects were asked to record the duration of ring use on each treatment day and the dates of new ring insertions and ring removals. These data were documented using either an interactive voice recording system in the United States and Canada or diary cards in Europe, and used to calculate the proportion of compliant cycles. A cycle was considered compliant if the ring use period did not deviate by more than 48 hours from the scheduled 3 weeks and the ring‐free period length did not deviate by more than 24 hours from the scheduled 1 week. Shortened and prolonged ring‐free periods imply durations of less than 6 and more than 8 days, respectively. The number and duration of temporary ring removals per subject were also assessed.
Contraceptive efficacy was assessed by determining the occurrence of pregnancy during the study. A home pregnancy test was performed on the day before inserting the first ring, and if pregnancy was suspected during the study. At the end of the last treatment cycle, serum β‐human chorionic gonadotropin was measured at a central laboratory to assess pregnancy status. Any pregnancies that occurred during the study were fully documented.
During the study period, cycle control was assessed daily by the subjects using either the interactive voice recording system in the United States and Canada, or diary cards in Europe. Vaginal bleeding was classified by the subject as either spotting (requiring at least one pad/tampon per day) or bleeding (two or more pads/tampons per day). The analysis of vaginal bleeding during the study focused on the occurrence of withdrawal bleeding (ie, any bleeding occurring during the ring‐free week), the occurrence of early or late withdrawal bleeding (ie, any withdrawal bleeding starting before, or continuing after the ring‐free period), and the occurrence of irregular bleeding (ie, any other bleeding occurring during the 3‐week ring period). Irregular bleeding was further classified as breakthrough bleeding or spotting.
The occurrence of adverse events was assessed and recorded at every study visit. Adverse events were classified into World Health Organization system‐organ classes. In addition, any events considered directly related to ring use, such as vaginal discomfort and device‐related events (ie, coital problems, foreign body sensation, and expulsion) were recorded in specific categories.
Subjects who withdrew from the study classified their reasons as related to either an adverse event (including vaginal discomfort and device‐related events), bleeding irregularity, pregnancy or “other” (nonmedical and non‐device‐related reasons, including no further requirement for contraception, or loss to follow‐up).
The user's acceptability of the ring was assessed by questionnaire at cycles 3, 6, and 13, or on early discontinuation. Acceptability was further evaluated by analyzing the discontinuation rates and the reasons for discontinuation.
All efficacy analyses were performed on both the intent‐to‐treat and the per‐protocol populations. The intent‐to‐treat population was defined as all the enrolled women who started treatment; the per‐protocol population was all the treated women without major protocol violations. These included the use of prohibited concomitant medication, absent or incorrect recording of data on the interactive voice recording system or diary cards, serious violation of the inclusion or exclusion criteria, or any violation that resulted in the exclusion of all that individual's data.
Contraceptive efficacy was estimated by calculating the Pearl Index and its 95% confidence interval (CI), and also the cumulative pregnancy rate derived from life table analysis. The Pearl Index is defined as the number of pregnancies per 100 woman‐years of exposure. The validity of combining the European and North American datasets was confirmed by the Uniform Most Powerful Unbiased test performed on the respective Pearl Indices. No significant difference was observed, showing that it was acceptable to pool results. The probability of discontinuation was estimated by plotting the cumulative discontinuation rates derived from life table analysis. Cycle control was analyzed using descriptive statistics. Tolerability analysis of the intent‐to‐treat group was performed using summary statistics and frequency tables.
A total of 2392 subjects were recruited for the study, of whom 2322 started treatment and so formed the intent‐to‐treat population. Of the 70 women who were enrolled in the trial but did not start treatment, 20 withdrew because they were found to be pregnant at baseline, four because of nonacceptance of the vaginal ring before starting treatment, and 46 because of nonmedical or nondevice‐related reasons. The per‐protocol population comprised 2015 subjects (86.8% of the intent‐to‐treat population) who were exposed to 16,912 treatment cycles. The intent‐to‐treat population was exposed for a total of 23,298 cycles, equivalent to 1786 woman‐years. The subject demographics and baseline characteristics are shown in Table 1.
During the study, 821 subjects (35.4%) discontinued the treatment prematurely; 18.5% discontinued for non‐medical or nondevice‐related reasons, including no further need for contraception and loss to follow‐up, and 15.1% withdrew because of adverse events. The cumulative discontinuation rates showed that the probability of discontinuation was higher during the initial months of treatment (Figure 1).
In the intent‐to‐treat population, 85.6% of cycles (79.9% of cycles in North America and 90.8% in Europe) were considered compliant with the dosing regimen. The incidence of cycles with a prolonged ring‐free period was 4.8%; a shortened ring‐free period occurred in 4.9% of cycles. Over 97% of subjects never temporarily removed the ring during any of the ring periods. There were slightly more temporary ring removals in North America than in Europe: 1.2% versus 0.28% of all ring‐use days, respectively.
The validity of combining the European and North American datasets was confirmed by the Uniform Most Powerful Unbiased test performed on the respective Pearl Indices. No significant difference was observed, showing that it was acceptable to pool results. Throughout the study period, there was a total of 21 pregnancies in the intent‐to‐treat group during 23,298 cycles, giving a Pearl Index of 1.18 (95% CI 0.73, 1.80). The overall cumulative pregnancy rate of in‐treatment pregnancy (derived from life table analysis), 1.18% (95% CI 0.68, 1.69), was comparable with the Pearl Index; the pregnancies were evenly distributed throughout the study. Eleven of these pregnancies occurred in women who did not comply with the study regimen. Noncompliance included an extended ring‐free period (two subjects reported ring‐free periods of 11 and 14 days), two or more consecutive days without ring use during a ring‐use period, or a misunderstanding of how to use the ring. In the per‐protocol group, there were therefore ten in‐treatment pregnancies, giving a Pearl Index of 0.77 (95% CI 0.37, 1.40). There were more pregnancies in the North American study (namely, 15) than in the European study (six); the overall Pearl Index (for both user and method failure) was therefore 1.75 (95% CI 0.98, 2.89) for North America and 0.65 (95% CI 0.24, 1.41) for Europe.
In the per‐protocol population, withdrawal bleeding occurred in 98.5% of all 16,912 cycles. The median day of onset of withdrawal bleeding was 3 (interquartile range 2–4) days after ring removal; mean duration of withdrawal bleeding ranged from 4.5 to 5.2 days. Early withdrawal bleeding (ie, bleeding starting just before the ring‐free period) occurred in 6.1% of cycles. In 23.9% of cycles, the withdrawal bleeding continued into the next cycle of treatment (late withdrawal bleeding). In most women, early or late withdrawal bleeding was restricted to spotting only (Table 2).
During the period of ring use, the incidence of irregular bleeding (classified as breakthrough bleeding and/or spotting) was very low, with an average of 5.5% per cycle over cycles 1–13 in the per‐protocol population (Figure 2). In the majority of cycles, the irregular bleeding was restricted to spotting only. Although the incidence of irregular bleeding was low in all cycles, levels declined slightly during the study. In cycle 1, 5.9% of women experienced breakthrough bleeding and spotting compared with 3.8% in cycle 13. In accordance with the slightly higher compliance seen in the European study, irregular bleeding was less frequent in the European study (4.4% of women per cycle) than in the North American study (7.0% of women per cycle). The intent‐to‐treat and per‐protocol data were comparable for all the cycle control parameters examined.
Over the 1‐year treatment period, 1522 subjects (65.5%) reported at least one adverse event, of which 37.5% were considered possibly, probably, or definitely related to study treatment. The most common treatment‐related adverse events were headache, vaginitis, and leukorrhea (Table 3). The incidence of acne and typically estrogen‐related events such as breast tenderness and nausea was very low. There were no major differences in the incidence of these events between the European and North American studies.
A total of 15.1% of subjects discontinued because of adverse events. The most frequently reported adverse events that resulted in withdrawal were device‐related events (2.5%), headache (1.3%), emotional lability (1.2%), and weight increase (1%). Only 0.8% of subjects withdrew as a result of bleeding irregularities, 0.7% because of vaginitis, and 0.6% because of leukorrhea.
Over the 1 year of treatment, the mean increase in body weight of all subjects from baseline was 0.84 kg (Table 4). Similarly, there was no clinically relevant change from baseline in either diastolic or systolic blood pressure (Table 4).
The majority of subjects (97.2%) had normal cervical cytology results at baseline and their last assessment. A total of 33 subjects (1.3%) changed from Papanicolaou grade I, IIa, or IIb (normal) at baseline to Papanicolaou grade IIIa (low‐grade squamous intraepithelial lesion). In seven subjects (0.4%), a clinically relevant shift in cervical cytology from Papanicolaou grade I, IIa, or IIb to Papanicolaou grade IIIb–IV (abnormal; high‐grade squamous intraepithelial lesion‐carcinoma in situ) was observed during the study. In two subjects, a shift to Papanicolaou grade IIIb–IV occurred at the last assessment, in three subjects, a shift to Papanicolaou grade IIb occurred at cycle 6 followed by a change to Papanicolaou grade IIIb–IV at the last assessment, and in the other two subjects, Papanicolaou grade IIIb–IV was initially reported at cycle 6 (both these subjects discontinued, the other five remained in the study). In the 33 subjects who shifted from Papanicolaou grade I, IIa, or IIb to grade IIIa, the shift was initially reported at cycle 6 in 18 subjects and at the 1‐year assessment in 15 subjects. This shift did not result in any withdrawals from the study. Eleven women entered the study with Papanicolaou grade IIIa cytology and were not excluded; in eight of these women, the cytology was normal at last assessment.
Medical examinations performed before, during, and after the study did not show any clinically relevant changes from baseline. There were no clinically relevant changes from baseline in blood biochemistry, hematology, or heart rate measurement.
The responses to the questions asked at the last assessment are presented in Table 5. Almost all women, including those who discontinued the study, found the ring easy to insert and to remove. The proportion of women who reported at least occasionally feeling the ring during intercourse was 18% and slightly higher in the discontinuers than the completers (23% versus 15%). The percentage of partners feeling the ring during intercourse was 32% and also slightly higher in the discontinuers than the completers (37% versus 29%). However, most partners in both the completers (94%) and the discontinuers (83%) groups did not object to women using the ring.
Throughout the study, most women were satisfied or very satisfied with the ring (Figure 3). The proportion of women reporting that they were satisfied or very satisfied with the ring was 96% at cycle 13 for completers. When all participants were considered, 85% were satisfied with the ring at last assessment. At the end of cycle 13, or the last visit, 90% of respondents indicated that they would recommend the ring to others. The proportion of women who would recommend the ring was high in both groups that completed (97%) and discontinued (75%) the study.
This questionnaire was also completed at cycles 3 and 6. The responses obtained at these visits were similar to those at cycle 13. There were no major differences between Europe and North America.
The findings from this 1‐year study of over 2300 women in the United States, Canada, and Europe show that this combined contraceptive vaginal ring is an effective contraceptive that has exceptional cycle control, is well tolerated, and acceptable to most users.
Good compliance with a contraceptive method is essential to maintain contraceptive reliability.2 We observed that most women used the ring correctly, and temporary ring removal rarely occurred. In 4.8% of cycles, the scheduled ring‐free period was extended by more than 1 day, which may increase the risk of pregnancy.
There was a slight difference in compliance between the two continents, which could have contributed towards the greater number of pregnancies observed in the North American study and the slight difference in irregular bleeding rates. Previous studies of oral contraceptives have also reported higher pregnancy rates in North America than in Europe,11,12 as well as differences in compliance between the continents.13,14 Several risk factors for lack of compliance to combined oral contraceptive use have been identified, including attitude towards contraception, age, parity, ethnicity, socioeconomic factors, and whether users were new contraceptive users or switching from other contraceptive methods.
During the study, there were 21 pregnancies in the total population, of which 11 were attributable to major protocol violations resulting in noncompliance with the regimen. This is illustrated by the markedly lower Pearl Index for the per‐protocol population. This good contraceptive efficacy is supported by pharmacodynamic studies showing that use of this ring causes effective inhibition of ovulation, comparable to that with a 30‐μg ethinylestradiol and 150‐μg desogestrel oral contraceptive.10 In one study designed to examine the effect of noncompliance, ring use effectively prevented ovulation in women with follicles up to 13‐mm diameter.15 These studies show that the ring is a contraceptive method that is robust to deviations from the prescribed usage schedule.
Contraceptive acceptability and compliance are influenced by many factors, key among which is cycle control; our findings demonstrate that the ring has very good cycle control. Almost all women experienced withdrawal bleeding, which did not generally extend outside the ring‐free week; when it did so, it was mainly restricted to spotting. Regular, monthly withdrawal bleeding is considered a positive attribute as it reassures the user of the continued absence of pregnancy. In contrast, irregular bleeding during the treatment cycle is recognized as a risk factor for poor compliance, as well as being inconvenient for the user.2,4 In this study, ring use was associated with a very low incidence of irregular bleeding, which was maintained from the first cycle onwards. This finding is in marked contrast to low‐dose combined oral contraceptives containing 30 μg of ethinylestradiol (double the daily 15‐μg dose delivered by the ring), with up to 20% of users experiencing some irregular bleeding during the first cycles of use.16–19 In a comparative study, the incidence of irregular bleeding was 5% or less in all cycles with the ring, compared with 5.4–38.8% with a 30‐μg ethinylestradiol oral contraceptive.20 With currently available combined oral contraceptives containing 15 μg of ethinylestradiol, irregular bleeding occurred in 29% of women during the first six cycles and in 21% over cycles 1–19,21,22 rates that are higher than those observed during ring use. We found that ring use results in sustained good cycle control despite a release of only 15 μg of ethinylestradiol per day; one possible explanation for this is the continuous steroid release. Only 19 of the 2322 women (0.8%) withdrew from the study because of bleeding irregularities.
The incidence of side effects is also a major factor in determining the overall acceptability of a contraceptive method.23 Throughout this study, the ring was well tolerated and showed a good safety profile. Vaginitis and leukorrhea were reported by 5.6% and 4.8% of subjects, respectively, whereas only 4.4% of subjects reported other device‐related events and 2.4% reported vaginal discomfort. The other treatment‐related side effects that were reported, such as headache (5.8%), nausea (3.2%), and breast tenderness (2.6%), are common with combined hormonal contraceptives. However, the proportions of women reporting nausea and breast tenderness in this study are lower than usually associated with low‐dose combined oral contraceptives.24–26 This may be attributable to the low and constant ethinylestradiol dose released by the ring. In contrast to the present study, trials using a norethindrone acetate and ethinylestradiol‐releasing vaginal ring have reported high levels of nausea, particularly during the first cycle of ring use.7,27 This was attributed to a large release of ethinylestradiol that occurred just after the ring was inserted.
The cervical cytology screening showed that 40 women had a shift in cytology from normal to abnormal (33 to Papanicolaou grade IIIa, seven to Papanicolaou grade IIIb–IV). An association between these shifts and ring use is unlikely. This statement is supported by the absence of adverse changes in cervical cytology with a prototype 120‐μg etonogestrel‐ and 15‐μg ethinylestradiol‐releasing vaginal ring.28 It is possible that the shifts were detected because of the frequent screening; previous data indicate that, in the normal population, many women have frequent shifts in cervical cytology grades.29,30 This is supported by findings from the current study: of the 11 women with Papanicolaou grade IIIa at the start of the study who were accidentally included in the study before the screening results were known, eight had normal smear results by the end of the study. These factors suggest that ring use is not associated with an increased risk of cervical abnormalities.
Throughout the study, physical, pelvic, and breast examinations reported no clinically relevant changes. It was also shown that the ring has little effect on body weight and blood pressure. The slight weight increase noted at the end of the study was not clinically relevant and was comparable with the change in women using a nonhormonal method of contraception.31 Moreover, no clinically relevant findings were reported for any hematologic or biochemical assessments.
These findings, which indicate that the ring is a well‐tolerated and safe contraceptive, suggest that it would be well accepted by users. This was confirmed by the results from the acceptability questionnaire. From the start of the study onwards, most women found the ring easy to insert and to remove, which is essential to ensure the usefulness of the method. Some women reported feeling the ring during intercourse, as did a slightly greater proportion of their partners. This did not appear to be a problem, as most partners did not object to the women using the ring as a contraceptive method, and temporary ring removal was infrequent. Overall, most women, including those who discontinued use for whatever reason, were highly satisfied with the ring and would recommend it to others. The high acceptability of the ring was reflected by the low number of withdrawals from the study because of device‐related events (2.5%).
In summary, this large multicenter study shows that the ring is an effective and reliable contraceptive with excellent cycle control; it is well tolerated, convenient, and highly acceptable to most women and their partners. The findings indicate that the ring is an alternative to combined oral contraceptives, avoiding the necessity of taking medication on a daily basis.
1. Fotherby K, Caldwell AD. New progestogens in oral contraception. Contraception 1994;49:1–32.
2. Rosenberg MJ, Waugh MS, Meehan TE. Use and misuse of oral contraceptives: Risk indicators for poor pill taking and discontinuation. Contraception 1995;51:283–8.
3. Zibners A, Cromer BA, Hayes J. Comparison of continuation rates for hormonal contraception among adolescents. J Pediatr Adolesc Gynecol 1999;12:90–4.
4. Rosenberg M, Waugh MS. Causes and consequences of oral contraceptive noncompliance. Am J Obstet Gynecol 1999;180:276–9.
5. Rosenberg MJ, Waugh MS, Burnhill MS. Compliance, counseling and satisfaction with oral contraceptives: A prospective evaluation. Fam Plann Perspect 1998;30:89–92.
7. Weisberg E, Fraser IS, Lacarra M, Mishell DR Jr, Alvarez F, Brache V, et al. Efficacy, bleeding patterns, and side effects of a 1-year contraceptive vaginal ring. Contraception 1999;59:311–8.
8. Roumen FJME, Apter D, Mulders TMT, Dieben TOM. Efficacy, tolerability and acceptability of a novel contraceptive vaginal ring releasing etonogestrel and ethinyl oestradiol. Hum Reprod 2001;16:469–75.
9. Timmer CJ, Mulders TMT. Pharmacokinetics of etonogestrel and ethinylestradiol released from a combined contraceptive vaginal ring. Clin Pharmacokinet 2000;39:233–42.
10. Mulders TMT, Dieben TOM. Use of the novel combined contraceptive vaginal ring NuvaRing for ovulation inhibition. Fertil Steril 2001;75:865–70.
11. Darney P. Safety and efficacy of a triphasic oral contraceptive containing desogestrel: Results of three multicenter trials. Contraception 1993;48:323–37.
12. Ferguson H, Vree ML, Wilpshaar J, Eskes TK. Multi-center study of the efficacy, cycle control and tolerability of a phasic desogestrel-containing oral contraceptive. Eur J Contracept Reprod Health Care 2000;5:35–45.
13. Potter L. Oral contraceptive compliance and its role in the effectiveness of the method. In: Cramer JA, Spiker B, eds. Patient compliance in medical practice and clinical trials. New York: Raven Press, 1991:195–207.
14. Rosenberg MJ, Burnhill MS, Waugh MS, Grimes DA, Hillard PJ. Compliance and oral contraceptives: A review. Contraception 1995;52:137–41.
15. Mulders TMT, Dieben TOM, Coelingh Bennink HJT. Ovarian function with a novel combined contraceptive vaginal ring. Hum Reprod. In press.
16. Rabe T, Runnebaum B, Kohlmeier M, Harenberg J, Weicker H. Clinical and metabolic effects of gestodene and levonorgestrel. Int J Fertil 1987;32(Suppl.):29–44.
17. Corson SL. Efficacy and clinical profile of a new oral contraceptive containing norgestimate. U.S. clinical trials. Acta Obstet Gynecol Scand 1990;152(Suppl.):25–31.
18. Brill K, Muller C, Schnitker J, Albring M. The influence of different modern low-dose oral contraceptives on inter-menstrual bleeding. Adv Contracept 1991;7(Suppl. 2):51–61.
19. Fotherby K. Twelve years of clinical experience with an oral contraceptive containing 30 micrograms ethinyloestradiol and 150 micrograms desogestrel. Contraception 1995;51:3–12.
20. Bjarnadóttir RI, Tuppurainen M, Killick SR. Comparison of cycle control with a combined contraceptive vaginal ring, and oral levonorgestrel/ethinylestradiol. Am J Obstet Gynecol 2002;186:389–95.
21. Gestodene Study Group 322. 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 1999;4 Suppl 2:9–15.
22. Gestodene Study Group 324. Cycle control, safety and efficacy of a 24-day regimen of gestodene 60 μg/ethinylestradiol 15 μg and a 21-day regimen of desogestrel 150 μg/and ethinylestradiol 20 μg. Eur J Contracept Reprod Health Care 1999;4(Suppl. 2):17–25.
23. Newton JR. Classification and comparison of oral contraceptives containing new generation progestogens. Hum Reprod Update 1995;1:231–63.
24. Mircette Study Group. An open-label, multicenter, non-comparative safety and efficacy study of Mircette, a low-dose estrogen-progestin oral contraceptive. Am J Obstet Gynecol 1998;179 Suppl:S2–8.
25. Archer DF, Maheux R, DelConte A, O'Brien FB. A new low-dose monophasic combination oral contraceptive (Alesse) with levonorgestrel 100 micrograms and ethinyl estradiol 20 micrograms. North American Levonorgestrel Study Group (NALSG). Contraception 1997;55:139–44.
26. Endrikat J, Düsterberg B, Ruebig A, Gerlinger C, Strowitzki T. Comparison of efficacy, cycle control, and tolerability of two low-dose oral contraceptives in a multicenter clinical study. Contraception 1999;60:269–74.
27. Weisberg E, Fraser IS, Mishell DR, Lacarra M, Darney P, Jackanicz TM. A comparative study of two contraceptive vaginal rings releasing norethindrone acetate and differing doses of ethinyl estradiol. Contraception 1999;59:305–10.
28. Roumen FJ, Boon ME, van Velzen D, Dieben TOM, Coelingh Bennink HJ. The cervico-vaginal epithelium during 20 cycles' use of a combined contraceptive vaginal ring. Hum Reprod 1996;11:2443–8.
29. Barron BA, Richart RM. Screening protocols for cervical neoplastic disease. Gynecol Oncol 1981;12 Suppl:S156–67.
30. Syrjanen KJ. Epidemiology of human papillomavirus (HPV) infections and their associations with genital squamous cell cancer. APMIS 1989;97:957–70.
31. Urbancsek J. An integrated analysis of nonmenstrual adverse events with Implanon. Contraception 1998;58 Suppl:109S–15S.