Triphasic versus monophasic oral contraceptives for contraception

Huib AAM Van Vliet; David A Grimes; Laureen M Lopez; Kenneth F Schulz; Frans M Helmerhorst

doi:10.1002/14651858.CD003553.pub3

. 2011 Nov 9;2011(11):CD003553. doi: 10.1002/14651858.CD003553.pub3

Triphasic versus monophasic oral contraceptives for contraception

Huib AAM Van Vliet ^1,^✉, David A Grimes ², Laureen M Lopez ², Kenneth F Schulz ³, Frans M Helmerhorst ⁴

Editor: Cochrane Fertility Regulation Group

PMCID: PMC7154342 PMID: 22071807

Abstract

Background

Side effects of oral contraceptive (OC) pills discourage adherence to and continuation of OC regimens. Strategies to decrease adverse effects led to the introduction of the triphasic OC in the 1980s. Whether triphasic OCs have higher accidental pregnancy rates than monophasic pills is unknown. Nor is it known if triphasic pills give better cycle control and fewer side effects than the monophasic pills.

Objectives

To compare triphasic OCs with monophasic OCs in terms of efficacy, cycle control, and discontinuation due to side effects.

Search methods

We searched the computerized databases of the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library), MEDLINE, POPLINE, EMBASE, and LILACS, as well as clinical trials databases (ClinicalTrials.gov and the World Health Organization Clinical Trials Registry Platform (ICTRP)) in May 2011. Additionally, we searched the reference lists of relevant articles. We also contacted researchers and pharmaceutical companies to identify other trials not found in our search.

Selection criteria

We included randomized controlled trials (RCTs) comparing any triphasic OC with any monophasic pill used to prevent pregnancy. Interventions had to include at least three treatment cycles.

Data collection and analysis

We assessed the studies found in the literature searches for possible inclusion and for their methodological quality. We contacted the authors of all included studies and of possibly randomized trials for supplemental information about the methods used and outcomes studied. We entered the data into RevMan and calculated odds ratios for the outcome measures of efficacy, breakthrough bleeding, spotting, withdrawal bleeding and discontinuation.

Main results

Of 23 trials included, 19 examined contraceptive effectiveness. The triphasic and monophasic preparations did not differ significantly. About half of the included trials reported favorable bleeding patterns, that is less spotting, breakthrough bleeding or amenorrhea, in triphasic versus monophasic OC users. However, meta‐analysis was generally not possible due to differences in measuring and reporting the cycle disturbance data as well as differences in progestogen type and hormone dosages. No significant differences were found in the numbers of women who discontinued due to medical reasons, cycle disturbances, intermenstrual bleeding or adverse events.

Authors' conclusions

The available evidence is insufficient to determine whether triphasic OCs differ from monophasic OCs in effectiveness, bleeding patterns or discontinuation rates. Therefore, we recommend monophasic pills as a first choice for women starting OC use. Large, high‐quality RCTs that compare triphasic and monophasic OCs with identical progestogens are needed to determine whether triphasic pills differ from monophasic OCs. Future studies should follow the recommendations of Belsey or Mishell on recording menstrual bleeding patterns and the CONSORT reporting guidelines.

Plain language summary

Birth control pills with three phases versus one phase

Side effects of birth control pills may keep women from using them as planned. Attempts to decrease side effects led to the three‐phase pill in the 1980s. Pills with three phases provide different amounts of hormones over three weeks. One‐phase pills have the same amount of hormone for three weeks. Whether three‐phase pills lead to more pregnancies is unknown. Nor is it known if the pills give better cycle control or fewer side effects. This review looked at whether three‐phase pills worked as well as one‐phase pills. It also studied whether women had fewer side effects with these pills.

We did a computer search for studies of pills with three phases versus pills with one phase in May 2011. We also wrote to researchers and manufacturers to find other trials. We included randomized trials in any language. The studies had to follow women for at least three treatment cycles.

We found 23 trials that looked at three‐phase versus one‐phase birth control pills. Many studies did not have good methods and the authors did not always report all their methods. The two types of pills did not differ in the numbers of women who got pregnant. About half of the trials found better bleeding patterns with the three‐phase pill. The numbers of women who stopped using the pills were about the same for both types of pills.

The evidence was not strong enough to say whether the three‐phase pill was better than the one‐phase pill for pregnancy prevention, bleeding patterns, or continued use. Therefore, we recommend one‐phase pills for women starting to use birth control pills. Large trials that are of good quality are needed to see if pills with three phases work better than those with one phase.

Background

Side effects of oral contraceptive pills discourage adherence to and continuation of oral contraceptive regimens (Rosenberg 1995; Rosenberg 1998). Three approaches have been used to decrease these adverse effects. These are (a) reduction of the steroid dose; (b) development of new steroids; and (c) new formulas and schedules of administration. These strategies led to the introduction of the triphasic oral contraceptive pill in the 1980s.

Triphasic oral contraceptives allegedly attempt to 'mimic' the rising and falling of estrogen and progesterone levels during the normal menstrual cycle (Upton 1983). This purportedly results in a more 'physiologic' approach and, with some pills, a lower total monthly steroid dosage compared to the older monophasic oral contraceptives. Possible benefits of the triphasic approach are better cycle control and fewer side effects (Guillebaud 1993; Hale 1987). In a cohort study conducted in France, women using 30 to 40 μg pills had similar reports of menstrual symptoms regardless of whether the pills were monophasic, biphasic or triphasic (Moreau 2007). However, potential disadvantages include an increased risk of pill‐taking errors caused by the array of different colored pills, the higher price of the pills and the possible higher incidence of accidental pregnancy (Guillebaud 1993).

Soon after the introduction in Britain of Logynon, a triphasic preparation of levonorgestrel and ethinylestradiol, two case reports described a probable method failure of the pill (Fay 1982; Graham 1982). Studies at abortion clinics in Australia and the Netherlands demonstrated a significant over‐representation of triphasic oral contraceptives used by women with an unplanned pregnancy (Ketting 1988; Kovacs 1989). Whether triphasic oral contraceptives have higher accidental pregnancy rates than monophasic oral contraceptives is unknown. Nor is it known if triphasic contraceptive pills give better cycle control and fewer side effects than the monophasic pills.

Objectives

The aim of this review was to compare triphasic oral contraceptive pills with monophasic oral contraceptive pills. Based on observational studies (Ketting 1988; Kovacs 1989), the a priori hypotheses were: (a) triphasic oral contraceptives are less effective in preventing pregnancy compared to monophasic oral contraceptives; and (b) triphasic oral contraceptives are similar to monophasic pills in terms of cycle control and continuation rates.

Methods

Criteria for considering studies for this review

Types of studies

We included only randomized controlled trials in this review. No language restrictions were placed on the reporting of the trials.

Types of participants

Healthy women of reproductive age were included if they had no contraindications for oral contraceptive use and desired to use oral contraceptives for preventing pregnancy. Women starting oral contraceptives as well as women switching oral contraceptives were included.

Types of interventions

We included any triphasic oral contraceptive pill (OC) compared to any monophasic oral contraceptive pill when used to prevent pregnancy. Both 21‐pill and 28‐pill packages were included. We excluded studies comparing triphasic pills with monophasic pills when the pills were used as a treatment (for example for acne, dysmenorrhea or menorrhagia) and not as a contraceptive. Interventions had to be applied for a minimum of three consecutive cycles to be eligible for inclusion.

Types of outcome measures

We focused on clinically relevant outcome measures. Studies were not included if they primarily looked at metabolic outcome measures or follicular growth. Principal outcomes were contraceptive efficacy, bleeding patterns and trial discontinuation. To be eligible, studies had to report results in a format that could be converted to the outcomes as follow.

Contraceptive efficacy

Proportion of women pregnant.

Cycle control

We used the definitions of spotting and breakthrough bleeding as specified by the authors.

Proportion of cycles with spotting or breakthrough bleeding or intermenstrual bleeding within 3 cycles, 6 cycles and 12 cycles of pill use.
Proportion of cycles with spotting or breakthrough bleeding or intermenstrual bleeding during the third cycle, the sixth cycle and the 12th cycle of pill use.
Proportion of women with spotting or breakthrough bleeding or intermenstrual bleeding within 3 cycles, 6 cycles and 12 cycles of pill use.
Proportion of women with spotting or breakthrough bleeding or intermenstrual bleeding during the third cycle, the sixth cycle and the 12th of pill use.
Proportion of cycles with absence of withdrawal bleeding within 3, 6 and 12 cycles of pill use.
Proportion of women with absence of withdrawal bleeding within 3, 6 and 12 cycles of pill use.
Proportion of women with absence of withdrawal bleeding during the third, the sixth and the 12th cycles of pill use.

Discontinuation

Proportion of women that discontinued within 3, 6 and 12 cycles of pill use.
Proportion of women that discontinued due to bleeding disturbances or adverse events within 3, 6 and 12 cycles of pill use.

Search methods for identification of studies

Electronic searches

We searched the computerized databases of the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library), MEDLINE using PubMed, POPLINE, EMBASE and LILACS for publications comparing monophasic, biphasic or triphasic oral contraceptives. In addition, we searched for recent clinical trials through ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform (ICTRP). Appendix 1 has the 2011 search strategies along with the search dates. Earlier strategies can be found in Appendix 2.

Searching other resources

We reviewed the reference lists of identified studies for additional trials. For the initial review, we examined the references lists from relevant book chapters and review articles identified with the search strategies. We searched the holdings of the Family Health International (FHI) library for relevant trials, book chapters and review articles. In addition, we attempted to contact the authors of all included trials. We also wrote a letter to pharmaceutical companies in the USA and Europe that market oral contraceptives. In the contact letters, we provided a list of studies identified and asked if they knew of unpublished or published trials we had missed.

Data collection and analysis

Selection of studies

Two authors independently evaluated the titles and abstracts identified during the literature searches under unblinded conditions and all potentially relevant articles were photocopied (Berlin 1997). Family Health International employees translated Russian, Chinese, Norwegian and German articles into English (Chen 1987; Dubnitskaia 1988; Engebretsen 1987; Lachnit‐Fixson (1979) of Zador 1979; Lachnit‐Fixson 1984). Then the authors independently examined the retrieved studies for possible inclusion. We excluded studies that were clearly not randomized, were quasi‐randomized controlled trials, or did not focus on interventions included in this review. Discrepancies were resolved through consensus.

Data extraction and management

One author extracted the data from the included studies under unblinded conditions and entered the data into RevMan 4.2 (Berlin 1997). In addition to the outcome measures and methodological quality of the study, we extracted data on participants, inclusion and exclusion criteria, study sites, duration of study, study medication, method of collecting the data and funding. Correct entry of the data was verified by a second author. No disagreements about the extracted and entered data occurred.

We wrote a letter to the authors of the included trials and to the authors of possibly randomized studies. In the letter we asked for additional information about the study methods and the various outcome measures.

Assessment of risk of bias in included studies

The validity of trials was critically appraised by assessing the potential risk for bias (Higgins 2005). We did not calculate summary quality scores but focused on the method of generating the allocation sequence, the use and method of allocation concealment, the use and method of blinding, exclusion of participants after randomization and loss to follow up (Juni 1999).

Assessment of heterogeneity

Most of the meta‐analyses combined a small number of trials. However, when substantial heterogeneity was evident (I² > 50%) we examined the studies separately and discussed the results in Effects of interventions.

Data synthesis

The review was limited to the analytic method used in the paper (for example intent‐to‐treat or per‐protocol).

Contraceptive efficacy

All studies that included contraceptive failure data reported the number of women who became pregnant. We used these proportions of women to calculate the odds ratio (OR) and 95% confidence interval (CI) with the random‐effects model.

Cycle control

Due to the possible relationship between (a) progestogen type and (b) dosage of estrogen and progestogen on bleeding patterns, we only combined trials that compared pills with identical contents (Maitra 2004; Rosenberg 1992), as was done in Gallo 2008. Most trials did not report bleeding pattern data according to the World Health Organization (WHO) recommendations (Belsey 1986) or to the Food and Drug Administration Reproductive Health Drugs Advisory Committee (RHDAC) recommendations (Mishell 2007). They generally reported the proportion of women or cycles with spotting, breakthrough bleeding or amenorrhea within three, six or 12 cycles or at the third, sixth or 12th cycle. We used these proportions to calculate the OR and 95% CI with the random‐effects model. Narrative summaries were provided when the reported bleeding data were not compatible with RevMan. For continuous data, we calculated the mean difference and 95% CI with a random‐effects model.

Discontinuation

We used the proportion of women discontinuing, discontinuing due to adverse events, and discontinuing due to cycle disturbances within three, six and 12 cycles to calculate the OR and 95% CI with the random‐effects model.

Sensitivity analysis

We did not conduct a sensitivity analysis based on the methodological quality of the trials.

Results

Description of studies

Included studies

We included 23 studies comparing triphasic oral contraceptives and monophasic oral contraceptives in this review. Six studies were published in duplicate (Carlborg 1983; Dieben 1984; Lachnit‐Fixson 1984; Saxena 1992; Sulak 1999). The Characteristics of included studies has detailed information regarding participants, inclusion and exclusion criteria, study sites, duration of study, study medication and outcome measures. Six trials included more than two intervention groups.

Triphasic levonorgestrel oral contraceptives versus monophasic oral contraceptives

Eight studies compared a triphasic OC composed of 50‐75‐125 μg levonorgestrel (LNG) and 30‐40‐30 μg ethinylestradiol (EE) with a monophasic pill composed of 150 μg levonorgestrel and 30 μg ethinylestradiol (Carlborg 1983; Chen 1987; Dunson 1993; Engebretsen 1987; Kashanian 2010; Ramos 1989; Saxena 1992; Zador 1979). Pharmaceutical companies market the triphasic preparation under the brand names Logynon, Trionetta 21 and 28, Triquilar, Trinordiol, and Triphasil. The monophasic preparation is marketed under the names Microgynon 30, Stediril, Neovletta, Rigevidon, Lo‐Femenal, Follimin, and Nordette. Chen 1987 evaluated a third preparation, which was composed of 600 μg norethindrone and 35 μg ethinylestradiol. Ramos 1989 also examined a third preparation, that is 400 μg norethindrone and 35 μg ethinylestradiol (Micropil).

The triphasic pill of 50‐75‐125 μg levonorgestrel and 30‐40‐30 μg ethinylestradiol was compared with a monophasic pill containing 150 μg desogestrel and 30 μg ethinylestradiol (Marvelon) in three studies (Dieben 1984; Ismail 1991; Lachnit‐Fixson 1984).

Percival‐Smith 1990 compared the triphasic 50‐75‐125 μg levonorgestrel and 30‐40‐30 μg ethinylestradiol pill, a triphasic pill containing 500‐750‐1000 μg norethindrone and 35 μg ethinylestradiol, and a monophasic pill of 1500 μg norethindrone acetate and 30 μg ethinylestradiol (Percival‐Smith 1990). Brand names of the triphasic norethindrone oral contraceptive are Ortho Novum 7/7/7 and Trinovum. The monophasic norethindrone pill is Loestrin. Percival‐Smith 1990 also examined a biphasic pill of 500‐1000 μg norethindrone and 35 μg ethinylestradiol but those data were not part of this review. The biphasic pill is known as Ortho 10/11.

One study compared: 1) the 50‐75‐125 μg levonorgestrel and 30, 40, 30 μg ethinylestradiol triphasic OC, and 2) the 500‐750‐1000 μg norethindrone and 35 μg ethinylestradiol triphasic OC with a monophasic OC containing 1000 μg norethindrone and 35 μg ethinylestradiol (Reiter 1990). The monophasic pill is marketed as Ortho‐Novum 1/35.

Triphasic norethindrone oral contraceptives versus monophasic oral contraceptives

Two studies compared a triphasic formulation containing 500‐750‐1000 μg norethindrone (NET) and 35 μg ethinylestradiol with a monophasic formulation containing 100 μg levonorgestrel and 20 μg ethinylestradiol (Chavez 1999; Reisman 1999). The monophasic formulation is marketed as Alesse and Loette.

We also included a study which compared an 'estrophasic' 1000 μg norethindrone acetate (NETA) and 20‐30‐35 μg ethinylestradiol combination with a monophasic 1500 μg norethindrone acetate and 30 μg ethinylestradiol combination (Rowan 1999).

Triphasic gestodene oral contraceptives versus monophasic oral contraceptives

Two trials compared a triphasic formulation composed of 50‐70‐100 μg gestodene (GTD) and 30‐40‐30 μg ethinylestradiol with a monophasic pill of 150 μg desogestrel (DSG) and 30 μg ethinylestradiol (Agoestina 1987; Andrade 1993). Brand names of the triphasic preparation are Trimulet and Triodeen.

The triphasic 50‐70‐100 μg gestodene and 30‐40‐30 μg ethinylestradiol oral contraceptive pill was also compared with: 1) a monophasic 75 μg gestodene and 30 μg ethinylestradiol pill, and 2) a monophasic 150 μg desogestrel and 20 μg ethinylestradiol pill (Bruni 2000). Companies market the monophasic gestodene and ethinylestradiol OC under the name Minulet and the monophasic desogestrel and ethinylestradiol pill under the name Mercilon.

Triphasic norgestimate oral contraceptives versus monophasic oral contraceptives

Sulak 1999 compared a triphasic formulation composed of 180‐215‐250 μg norgestimate (NGM) and 35 μg ethinylestradiol with a monophasic formulation of 1000 μg norethindrone acetate and 20 μg ethinylestradiol (Sulak 1999). Companies market the triphasic pill under the name Ortho Tri‐Cyclen and the monophasic pill under the name Loestrin Fe 1/20.

In Rosenberg 1999, the triphasic formulation of 180‐215‐250 μg norgestimate and 35 μg ethinylestradiol was compared with: 1) a formulation of 100 μg levonorgestrel and 20 μg ethinylestradiol (mentioned above for Chavez 1999 and Reisman 1999), and 2) a monophasic formulation of 150 μg desogestrel and 20 μg ethinylestradiol for 21 days and 10 μg ethinylestradiol for five days (Rosenberg 1999). The brand name of the desogestrel containing monophasic preparation is Mircette.

A triphasic preparation of 180‐215‐250 μg norgestimate and 25 μg ethinylestradiol was compared with a monophasic preparation of 1000 μg norethindrone acetate and 20 μg ethinylestradiol (Hampton 2001). The triphasic oral contraceptive is marketed as Ortho Tri‐Cyclen Lo. Hampton 2001 also compared the triphasic to two 'cyclophasic' preparations but those data were not part of this review. Neither 'cyclophasic' regimen has been put on the market yet. The 'cyclophasic' pills had a fixed daily dose of ethinylestradiol and a dose of norgestimate that alternated every other day.

In Kaunitz 2009, the triphasic preparation of 180‐215‐250 μg norgestimate and 25 μg ethinylestradiol was compared with a monophasic pill containing 3 mg drospirenone and 20 μg ethinylestradiol. The drospirenone containing monophasic preparation is marketed as Yaz.

Excluded studies

We excluded one study described as a randomized controlled trial that proved to be a matched cohort study (Dubnitskaia 1988). Seven studies did not report the method used to generate the allocation sequence. After communication with the author, we excluded Grace 1994 because the allocation sequence was not randomized. We excluded the remaining six studies because we were unable to contact the authors (Christie 1989; Dik 1984; Matsumoto 1988; Otolorin 1989; Perrone 1987; Rubio‐Lotvin 1992). We did not include Bancroft 1987 due to the lack of relevant outcomes for this review.

Risk of bias in included studies

Overall, the description of the study methods was poor (CONSORT 2010; DerSimonian 1982).

Allocation

Only three of the 23 included trials reported the method of generating the allocation sequence (Agoestina 1987; Chen 1987; Kaunitz 2009). In addition, only five studies reported the use and method of concealing the treatment allocation sequence (Agoestina 1987; Ismail 1991; Kashanian 2010; Ramos 1989; Reisman 1999). Communication with the authors provided the method of generating the allocation sequence for eight trials (Carlborg 1983; Dunson 1993; Hampton 2001; Ismail 1991; Kashanian 2010; Reiter 1990; Rosenberg 1999; Saxena 1992). Nine authors informed us on whether allocation concealment was done and, if so, what method was used (Carlborg 1983; Chavez 1999; Dieben 1984; Dunson 1993; Hampton 2001; Reiter 1990; Rosenberg 1999; Rowan 1999; Saxena 1992).

Eight studies appeared to have adequate randomization and concealment of treatment allocation (Schulz 2002c). Randomization was done by a computer (Carlborg 1983; Dunson 1993; Hampton 2001; Ismail 1991; Rosenberg 1999; Saxena 1992) or a random number table (Agoestina 1987; Kashanian 2010). The methods used to conceal the allocation sequence included numbered pharmacy packages (Carlborg 1983); numbered containers (Agoestina 1987); sequentially‐numbered, sealed envelopes opened at the time of admission (Dunson 1993; Ismail 1991; Kashanian 2010; Saxena 1992); sequentially‐numbered randomization cards with an opaque scratch‐off dot (Rosenberg 1999); and a centralized voice‐activated randomization system (Hampton 2001). Only Dunson 1993 mentioned that the envelopes were opaque.

Eight studies provided information on either the randomization method or the allocation concealment. Randomization was done by a random number table in Chen 1987 or a computer in Kaunitz 2009 but both studies did not mention the method used to conceal the allocation sequence. Reiter 1990 generated the allocation sequence by a random number table but did not conceal the allocation sequence. We could not find out the method of randomization in four studies, but the studies appeared to use a proper method to conceal the treatment allocation sequence. Acceptable methods include numbered pharmacy packages (Ramos 1989; Rowan 1999) and sequentially‐numbered, opaque, sealed envelopes (Chavez 1999; Reisman 1999). However, Chavez 1999 did not mention whether the envelopes were sealed and Reisman 1999 did not note if the envelopes were opaque. In Dieben 1984, the method of randomization was unclear and the study featured inadequate concealment of allocation.

The remaining seven trials did not mention either the method used to generate the allocation sequence or the method for concealing the allocation sequence (Andrade 1993; Bruni 2000; Engebretsen 1987; Lachnit‐Fixson 1984; Percival‐Smith 1990; Sulak 1999; Zador 1979).

Blinding

Blinding was not mentioned in five trial reports (Agoestina 1987; Carlborg 1983; Kashanian 2010; Lachnit‐Fixson 1984; Saxena 1992). We obtained details on the use of blinding in four trials from the researchers (Carlborg 1983; Kashanian 2010; Rowan 1999; Saxena 1992). None of the studies provided information regarding successful implementation of blinding (Schulz 2002b). Two studies (Hampton 2001; Ramos 1989) reported information to judge the adequacy of the blinding methods (DerSimonian 1982; Schulz 2002b).

Three trials blinded investigators and participants (Carlborg 1983; Kashanian 2010; Ramos 1989). Furthermore, two trials reported that the study was double‐blinded without specifying who was kept unaware of the oral contraceptives assigned (Chen 1987; Rowan 1999) (DerSimonian 1982; Schulz 2002b). One trial was blinded for the outcome assessor (Percival‐Smith 1990). Fifteen trials were open (Andrade 1993; Bruni 2000; Chavez 1999; Dieben 1984; Dunson 1993; Engebretsen 1987; Hampton 2001; Ismail 1991; Kaunitz 2009; Reisman 1999; Reiter 1990; Rosenberg 1999; Saxena 1992; Sulak 1999; Zador 1979). The remaining two trials did not mention the use of blinding; we were unable to contact one team of researchers (Agoestina 1987) and another investigator could not provide additional information (Lachnit‐Fixson 1984).

Incomplete outcome data

Nine trials described detailed information on the number of and reasons for discontinuation (Andrade 1993; Bruni 2000; Chavez 1999; Chen 1987; Ismail 1991; Kaunitz 2009; Ramos 1989; Reisman 1999; Saxena 1992). We acquired information on the number of and reasons for discontinuation from the authors of three studies (Dieben 1984; Kashanian 2010; Rosenberg 1999). The other trials provided insufficient or no information on withdrawals.

Study discontinuation ranged from 4% to 77%. Thirteen trials included data on losses to follow up. Loss to follow up varied from zero to 39%. In Dunson 1993, more than 20% of the participants were lost to follow up. Losses to follow‐up rates greater than 20% may threaten the validity of trials (Strauss 2005).

Of 23 included studies, 17 excluded participants after randomization for reasons like: failure to start oral contraceptives, failure to appear at the first follow‐up visit, incorrect administration of oral contraceptives, protocol violations such as incorrect pill‐taking or skipping the pill‐free interval, inaccurate recording of data, loss to follow up or cycle disturbances (Andrade 1993; Bruni 2000; Carlborg 1983; Chavez 1999; Chen 1987; Dieben 1984; Engebretsen 1987; Hampton 2001; Ismail 1991; Kaunitz 2009; Kashanian 2010; Percival‐Smith 1990; Ramos 1989; Reisman 1999; Reiter 1990; Saxena 1992; Sulak 1999). Exclusion of participants after randomization may lead to bias (Schulz 2002a).

Two trials reported an analysis based on the intention‐to‐treat principle (Dunson 1993; Hampton 2001). Sulak 1999 stated that the analysis was carried out on the intention‐to‐treat population. However, the population included only participants who had started oral contraceptives and who had at least one cycle control measurement after the baseline (Sulak 1999). In addition, cycles were considered invalid and excluded if they had incorrect pill‐taking or were without a pill‐free interval, lasted longer than 31 days or had inaccurate recording of bleeding data. Kaunitz 2009 stated that the evaluation of bleeding patterns was based on the intention‐to‐treat population, defined as all randomly assigned participants who took the study drug and for whom there were post‐baseline bleeding data after day seven. Another 15 trials did not perform an intention‐to‐treat analysis (Agoestina 1987; Andrade 1993; Bruni 2000; Carlborg 1983; Chavez 1999; Chen 1987; Dieben 1984; Engebretsen 1987; Ismail 1991; Kashanian 2010; Percival‐Smith 1990; Ramos 1989; Reisman 1999; Reiter 1990; Saxena 1992). In four studies, it was unclear whether an analysis based on the intention‐to‐treat approach was performed (Lachnit‐Fixson 1984; Rosenberg 1999; Rowan 1999; Zador 1979).

Four studies continued with a proportion of the participants after six cycles of pill use. In Hampton 2001, women were enrolled for six cycles or 12 cycles of pill use at admittance. In Andrade 1993, Carlborg 1983 and Dieben 1984, we could not find out whether the continuation was decided previously or during the study. The latter may result in selection bias.

Other potential sources of bias

A priori hypothesis and sample size calculation

An a priori hypothesis and sample size calculation were provided in four studies (Kashanian 2010; Kaunitz 2009; Percival‐Smith 1990; Reisman 1999) (DerSimonian 1982). However, Kashanian 2010 was inconsistent regarding the power (80% or 85%). Another study stated that the sample size was developed to meet the US regulatory requirements to evaluate the safety and efficacy of oral contraceptives and provided the power (Hampton 2001). One study reported a sample size without explanation (Lachnit‐Fixson 1984).

Funding

Five of the 23 included trials were conducted or supported by independent organizations: World Health Organization (Chen 1987); Family Health International (Dunson 1993; Ismail 1991); Planned Parenthood Federation (Reiter 1990); and Indian Council of Medical Research (Saxena 1992). Fourteen trials were conducted or sponsored by pharmaceutical companies (Agoestina 1987; Bruni 2000; Carlborg 1983; Chavez 1999; Dieben 1984; Hampton 2001; Kaunitz 2009; Lachnit‐Fixson 1984; Percival‐Smith 1990; Reisman 1999; Rosenberg 1999; Rowan 1999; Sulak 1999; Zador 1979). One study was supported by an international organization (United Nations Population Fund) in combination with a pharmaceutical company (Ramos 1989). We could not identify any assistance for three trials (Andrade 1993; Engebretsen 1987; Kashanian 2010). Studies sponsored by pharmaceutical companies are more likely to have outcomes favoring the sponsor than studies funded by other sources (Als‐Nielsen 2003; Lexchin 2003).

Effects of interventions

Contraceptive effectiveness

Nineteen studies comparing a triphasic formulation with a monophasic formulation assessed contraceptive effectiveness (Agoestina 1987; Andrade 1993; Bruni 2000; Carlborg 1983; Chavez 1999; Chen 1987; Dieben 1984; Dunson 1993; Engebretsen 1987; Hampton 2001; Ismail 1991; Kaunitz 2009; Lachnit‐Fixson 1984; Ramos 1989; Reisman 1999; Rosenberg 1999; Saxena 1992; Sulak 1999; Zador 1979). An additional four studies did not report data regarding pregnancy (Kashanian 2010; Percival‐Smith 1990; Reiter 1990; Rowan 1999). Most studies included pregnancies caused by inadequacy of the method as well as imperfect use in the reported number of pregnancies (Trussell 1991). However, in three reports we could not figure out whether the pregnancies were caused by method failures solely or by both method and user failures (Engebretsen 1987; Kaunitz 2009; Saxena 1992). In the analyses, we considered the number of pregnancies reported in these three studies as method and user failures. There was a discrepancy in the described numbers of pregnancies between Cullberg et al (1982) from Dieben 1984 and the later report of Dieben 1984. The Dieben 1984 paper mentioned two pregnancies, and the paper by Cullberg et al (1982) noted three pregnancies. Communication with the author revealed that three pregnancies occurred in the study period.

No significant differences were found between the various pills in contraceptive effectiveness (Analysis 1.1, Analysis 1.2, Analysis 2.1, Analysis 3.1, Analysis 4.1, Analysis 4.2, Analysis 9.1, Analysis 11.1, Analysis 11.2, Analysis 12.1, Analysis 13.1, Analysis 14.1, Analysis 15.1, Analysis 16.1, Analysis 17.1, Analysis 18.1) (Table 19).

1.1 — Comparison 1 Triphasic LNG 50‐75‐125 μg and EE 30‐40‐30 μg versus monophasic LNG 150 μg and EE 30 μg, Outcome 1 Pregnancy per woman within 6 cycles.

1.2 — Comparison 1 Triphasic LNG 50‐75‐125 μg and EE 30‐40‐30 μg versus monophasic LNG 150 μg and EE 30 μg, Outcome 2 Pregnancy per woman within 12 cycles.

2.1 — Comparison 2 Triphasic LNG 50‐75‐125 μg and EE 30‐40‐30 μg versus monophasic NET 600 μg and EE 35 μg, Outcome 1 Pregnancy per woman within 6 cycles.

3.1 — Comparison 3 Triphasic LNG 50‐75‐125 μg and EE 30‐40‐30 μg versus monophasic NET 400 μg and EE 35 μg, Outcome 1 Pregnancy per woman within 12 cycles.

4.1 — Comparison 4 Triphasic LNG 50‐75‐125 μg and EE 30‐40‐30 μg versus monophasic DSG 150 μg and EE 30 μg, Outcome 1 Pregnancy per woman within 6 cycles.

4.2 — Comparison 4 Triphasic LNG 50‐75‐125 μg and EE 30‐40‐30 μg versus monophasic DSG 150 μg and EE 30 μg, Outcome 2 Pregnancy per woman within 12 cycles.

9.1 — Comparison 9 Triphasic NET 500‐750‐1000 μg and EE 35 μg versus monophasic LNG 100 μg and EE 20 μg, Outcome 1 Pregnancy per woman within 4 cycles.

11.1 — Comparison 11 Triphasic GTD 50‐70‐100 μg and EE 30‐40‐30 μg versus monophasic DSG 150 μg and EE 30 μg, Outcome 1 Pregnancy per woman within 6 cycles.

11.2 — Comparison 11 Triphasic GTD 50‐70‐100 μg and EE 30‐40‐30 μg versus monophasic DSG 150 μg and EE 30 μg, Outcome 2 Pregnancy per woman within 12 cycles.

12.1 — Comparison 12 Triphasic GTD 50‐70‐100 μg and EE 30‐40‐30 μg versus monophasic DSG 150 μg and EE 20 μg, Outcome 1 Pregnancy per woman within 13 cycles.

13.1 — Comparison 13 Triphasic GTD 50‐70‐100 μg and EE 30‐40‐30 μg versus monophasic GTD 75 μg and EE 30 μg, Outcome 1 Pregnancy per woman within 13 cycles.

14.1 — Comparison 14 Triphasic NGM 180‐215‐250 μg and EE 35 μg versus monophasic NETA 1000 μg and EE 20 μg, Outcome 1 Pregnancy per woman within 6 cycles.

15.1 — Comparison 15 Triphasic NGM 180‐215‐250 μg and EE 35 μg versus monophasic LNG 100 μg and EE 20 μg, Outcome 1 Pregnancy per woman within 6 cycles.

16.1 — Comparison 16 Triphasic NGM 180‐215‐250 μg and EE 35 μg versus monophasic DSG 150 μg and EE 20 μg + 5 days EE 10 μg, Outcome 1 Pregnancy per woman within 6 cycles.

17.1 — Comparison 17 Triphasic NGM 180‐215‐250 μg and EE 25 μg versus monophasic NETA 1000 μg and EE 20 μg, Outcome 1 Pregnancy per woman within 13 cycles.

18.1 — Comparison 18 Triphasic NGM 180‐215‐250 μg and EE 25 μg versus monophasic DRSP 3 mg and EE 20 μg, Outcome 1 Pregnancy per woman within 3 cycles.

1. Pregnancies and total study cycles for triphasic and monophasic formulations.

	Triphasic		Monophasic
Study	Pregnancies	Total cycles	Pregnancies	Total cycles
Agoestina 1987	1	915	1	903
Andrade 1993	1	1398	0	1245
Carlborg 1983	1	1574	0	3275
	0	1623
Chavez 1999	2	400	1	384
Chen 1987	2	492	3	478
			2	474
Dieben 1984	3	2709	0	2771
Engebretsen 1987	1	1442	1	1416
Hampton 2001	20	11003	19	7497
Ismail 1991	3	741	0	811
Lachnit‐Fixson 1984	1	1536	0	1524
Reisman 1999	2	506	1	453
Rosenberg 1999	3	831	1	819
			0	848
Saxena 1992	0	3319	0	2949
Zador 1979	0	1440	0	1343

Study and COC	Reference period	Number	Acceptable pattern	Infrequent bleeding	Frequent/   prolonged bleeding	No. of bleeding runs	Total bleeding days	Total spotting days
Saxena 1992
Triphasic LNG	1	289	82.7	11.8	5.5	2.9 + 0.6	9.9 + 3.5	2.8 + 3.6
	2	250	84.4	10.4	5.2	3.1 + 0.6	9.8 + 3.5	2.5 + 2.8
	3	195	90.3	8.2	1.5	3.0 + 0.5	9.4 + 3.1	2.7 + 2.9
	4	123	83.7	10.6	5.7	3.1 + 0.6	9.0 + 2.6	3.0 + 3.1
Monophasic LNG	1	248	80.2	15.7	4.0	2.8 + 0.8	8.9 + 2.9	3.3 + 4.1
	2	207	85.0	12.1	2.9	3.0 + 0.6	9.2 + 2.7	2.7 + 3.1
	3	183	85.2	13.7	1.1	3.0 + 0.6	8.8 + 2.8	2.8 + 2.9
	4	129	89.1	10.1	0.8	2.9 + 0.5	8.5 + 2.5	3.5 + 3.5
Chen 1987
Triphasic LNG	1						16.0 + 4.1	5.6 + 4.8
Triphasic LNG	1+2						26.2 + 5.8	8.7 + 7.5
Monophasic LNG	1						15.1 + 4.3	8.0 + 7.1
Monophasic LNG	1+2						25.0 + 7.0	11.2 + 8.4
Monophasic NET	1						14.8 + 5.0	9.4 + 6.3
Monophasic NET	1+2						25.8 + 8.2	14.2 + 8.9

Months	Triphasic LNG	Monophasic LNG	Monophasic NET
0 to 3	39.4	42.6	45.9
4 to 6	88.1	89.7	89.7
7 to 9	95.2	96.0	94.4
10 to 12	93.9	93.7	94.2

Date	Event	Description
11 August 2011	New citation required but conclusions have not changed	Two new trials added (Kashanian 2010; Kaunitz 2009), along with a 2009 secondary paper for Hampton 2001. An ongoing trial was updated (Bayer 2011).
4 May 2011	New search has been performed	Searches updated
20 January 2010	Amended	author order changed
25 November 2008	New search has been performed	Searches were updated; included a secondary article from earlier trial (Hampton 2001). Also added searches of clinical trials databases.
15 April 2008	Amended	Converted to new review format.
8 May 2006	New citation required and conclusions have changed	Substantive amendment

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 6 cycles	2	678	Odds Ratio (M‐H, Random, 95% CI)	0.64 [0.10, 3.91]
2 Pregnancy per woman within 12 cycles	5	4145	Odds Ratio (M‐H, Random, 95% CI)	1.35 [0.25, 7.22]
3 Proportion of cycles with spotting within 3 cycles	1	2367	Odds Ratio (M‐H, Random, 95% CI)	0.52 [0.41, 0.68]
4 Proportion of cycles with breakthrough bleeding within 3 cycles	1	2367	Odds Ratio (M‐H, Random, 95% CI)	0.48 [0.32, 0.72]
5 Proportion of cycles with spotting within 6 cycles	2	7290	Odds Ratio (M‐H, Random, 95% CI)	0.57 [0.48, 0.67]
6 Proportion of cycles with breakthrough bleeding within 6 cycles	2	7290	Odds Ratio (M‐H, Random, 95% CI)	0.63 [0.50, 0.80]
7 Proportion of cycles with spotting within 12 cycles	1	6472	Odds Ratio (M‐H, Random, 95% CI)	0.59 [0.49, 0.72]
8 Proportion of cycles with breakthrough bleeding within 12 cycles	1	6472	Odds Ratio (M‐H, Random, 95% CI)	0.59 [0.45, 0.77]
9 Proportion of women with intermenstrual bleeding within 12 cycles	1	979	Odds Ratio (M‐H, Random, 95% CI)	0.83 [0.53, 1.31]
10 Proportion of women with spotting at cycle 6	1	1032	Odds Ratio (M‐H, Random, 95% CI)	0.24 [0.03, 2.17]
11 Proportion of women with breakthrough bleeding at cycle 6	1	1032	Odds Ratio (M‐H, Random, 95% CI)	2.45 [0.47, 12.67]
12 Proportion of women with spotting at cycle 12	1	896	Odds Ratio (M‐H, Random, 95% CI)	1.04 [0.06, 16.62]
13 Proportion of women with breakthrough bleeding at cycle 12	1	896	Odds Ratio (M‐H, Random, 95% CI)	0.52 [0.05, 5.72]
14 Proportion of women with intermenstrual bleeding within 12 cycles	1	979	Odds Ratio (M‐H, Random, 95% CI)	0.83 [0.53, 1.31]
15 Proportion of women with breakthrough bleeding or spotting at cycle 3	1	314	Odds Ratio (M‐H, Random, 95% CI)	0.10 [0.01, 0.77]
16 Proportion of women with breakthrough bleeding or spotting at cycle 6	1	314	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
17 Proportion of cycles with amenorrhea within 6 cycles	1	2777	Odds Ratio (M‐H, Random, 95% CI)	0.57 [0.28, 1.14]
18 Proportion of cycles with amenorrhea within 12 cycles	1	6472	Odds Ratio (M‐H, Random, 95% CI)	0.27 [0.17, 0.45]
19 Proportion of women with amenorrhea within 12 cycles	1	979	Odds Ratio (M‐H, Random, 95% CI)	1.47 [0.24, 8.83]
20 Total discontinuation within 6 cycles	4	1829	Odds Ratio (M‐H, Random, 95% CI)	1.04 [0.78, 1.37]
21 Discontinuation due to medical reasons within 6 cycles	3	1513	Odds Ratio (M‐H, Random, 95% CI)	1.15 [0.81, 1.61]
22 Discontinuation due to cycle disturbances within 6 cycles	1	189	Odds Ratio (M‐H, Random, 95% CI)	0.97 [0.13, 7.02]
23 Total discontinuation within 12 cycles	4	3310	Odds Ratio (M‐H, Random, 95% CI)	1.13 [0.97, 1.31]
24 Discontinuation due to medical reasons within 12 cycles	3	3010	Odds Ratio (M‐H, Random, 95% CI)	1.12 [0.71, 1.76]
25 Discontinuation due to cycle disturbances within 12 cycles	3	2109	Odds Ratio (M‐H, Random, 95% CI)	1.11 [0.56, 2.21]
26 Discontinuation due to intermenstrual bleeding within 12 cycles	1	1201	Odds Ratio (M‐H, Random, 95% CI)	1.40 [0.44, 4.44]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 6 cycles	1	186	Odds Ratio (M‐H, Random, 95% CI)	0.94 [0.13, 6.79]
2 Total discontinuation within 6 cycles	1	186	Odds Ratio (M‐H, Random, 95% CI)	0.92 [0.44, 1.94]
3 Discontinuation due to medical reasons within 6 cycles	1	186	Odds Ratio (M‐H, Random, 95% CI)	0.93 [0.35, 2.46]
4 Discontinuation due to cycle disturbances within 6 cycles	1	186	Odds Ratio (M‐H, Random, 95% CI)	0.62 [0.10, 3.78]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 12 cycles	1	1200	Odds Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2 Proportion of women with spotting at cycle 6	1	1018	Odds Ratio (M‐H, Random, 95% CI)	0.12 [0.01, 0.94]
3 Proportion of women with breakthrough bleeding at cycle 6	1	1018	Odds Ratio (M‐H, Random, 95% CI)	0.31 [0.11, 0.86]
4 Proportion of women with spotting at cycle 12	1	851	Odds Ratio (M‐H, Random, 95% CI)	0.93 [0.06, 14.98]
5 Proportion of women with breakthrough bleeding at cycle 12	1	851	Odds Ratio (M‐H, Random, 95% CI)	0.47 [0.04, 5.16]
6 Total discontinuation within 12 cycles	1	1200	Odds Ratio (M‐H, Random, 95% CI)	0.80 [0.63, 1.03]
7 Discontinuation due to medical reasons within 12 cycles	1	1200	Odds Ratio (M‐H, Random, 95% CI)	0.78 [0.55, 1.10]
8 Discontinuation due to intermenstrual bleeding within 12 cycles	1	1200	Odds Ratio (M‐H, Random, 95% CI)	0.58 [0.23, 1.47]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 6 cycles	1	555	Odds Ratio (M‐H, Random, 95% CI)	3.0 [0.12, 73.96]
2 Pregnancy per woman within 12 cycles	2	1146	Odds Ratio (M‐H, Random, 95% CI)	7.22 [0.88, 59.00]
3 Proportion of cycles with spotting within 3 cycles	1	2763	Odds Ratio (M‐H, Random, 95% CI)	1.11 [0.88, 1.41]
4 Proportion of cycles with breakthrough bleeding within 3 cycles	1	2763	Odds Ratio (M‐H, Random, 95% CI)	1.16 [0.87, 1.56]
5 Proportion of cycles with spotting and breakthrough bleeding within 3 cycles	1	2763	Odds Ratio (M‐H, Random, 95% CI)	0.71 [0.31, 1.61]
6 Proportion of cycles with spotting within 6 cycles	1	5235	Odds Ratio (M‐H, Random, 95% CI)	1.17 [0.97, 1.41]
7 Proportion of cycles with spotting within 6 cycles	1	3060	Odds Ratio (M‐H, Random, 95% CI)	0.34 [0.27, 0.44]
8 Proportion of cycles with breakthrough bleeding within 6 cycles	1	5235	Odds Ratio (M‐H, Random, 95% CI)	1.13 [0.91, 1.42]
9 Proportion of cycles with breakthrough bleeding within 6 cycles	1	3060	Odds Ratio (M‐H, Random, 95% CI)	0.41 [0.23, 0.71]
10 Proportion of cycles with spotting and breakthrough bleeding within 6 cycles	2	8295	Odds Ratio (M‐H, Random, 95% CI)	0.50 [0.29, 0.86]
11 Proportion of cycles with spotting within 12 cycles	1	5478	Odds Ratio (M‐H, Random, 95% CI)	1.19 [0.99, 1.44]
12 Proportion of cycles with breakthrough bleeding within 12 cycles	1	5478	Odds Ratio (M‐H, Random, 95% CI)	1.09 [0.88, 1.35]
13 Proportion of cycles with spotting and breakthrough bleeding within 12 cycles	1	5478	Odds Ratio (M‐H, Random, 95% CI)	0.64 [0.33, 1.22]
14 Proportion of women with staining/spotting within 12 cycles	1	197	Odds Ratio (M‐H, Random, 95% CI)	1.55 [0.42, 5.67]
15 Proportion of women with moderate flow intermenstrual bleeding within 12 cycles	1	197	Odds Ratio (M‐H, Random, 95% CI)	2.61 [0.49, 13.77]
16 Proportion of women with spotting at cycle 3	1	894	Odds Ratio (M‐H, Random, 95% CI)	1.18 [0.74, 1.90]
17 Proportion of women with breakthrough bleeding at cycle 3	1	894	Odds Ratio (M‐H, Random, 95% CI)	1.10 [0.66, 1.82]
18 Proportion of women with spotting and breakthrough bleeding at cycle 3	1	894	Odds Ratio (M‐H, Random, 95% CI)	0.50 [0.04, 5.50]
19 Proportion of women with spotting at cycle 6	1	797	Odds Ratio (M‐H, Random, 95% CI)	1.33 [0.68, 2.58]
20 Proportion of women with breakthrough bleeding at cycle 6	1	797	Odds Ratio (M‐H, Random, 95% CI)	1.53 [0.80, 2.92]
21 Proportion of women with spotting and breakthrough bleeding at cycle 6	1	797	Odds Ratio (M‐H, Random, 95% CI)	0.50 [0.04, 5.51]
22 Proportion of women with spotting at cycle 12	1	10	Odds Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
23 Proportion of women with breakthrough bleeding at cycle 12	1	10	Odds Ratio (M‐H, Random, 95% CI)	1.0 [0.03, 33.32]
24 Proportion of women with spotting and breakthrough bleeding at cycle 12	1	10	Odds Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
25 Proportion of cycles with amenorrhea within 3 cycles	1	2763	Odds Ratio (M‐H, Random, 95% CI)	1.06 [0.82, 1.39]
26 Proportion of cycles with amenorrhea within 6 cycles	1	5235	Odds Ratio (M‐H, Random, 95% CI)	1.07 [0.87, 1.31]
27 Proportion of cycles with amenorrhea within 6 cycles	1	3060	Odds Ratio (M‐H, Random, 95% CI)	0.21 [0.06, 0.74]
28 Proportion of cycles with amenorrhea within 12 cycles	1	5478	Odds Ratio (M‐H, Random, 95% CI)	1.05 [0.86, 1.28]
29 Proportion of women with amenorrhea within 12 cycles	1	197	Odds Ratio (M‐H, Random, 95% CI)	1.53 [0.25, 9.37]
30 Proportion of women with amenorrhea at cycle 3	1	894	Odds Ratio (M‐H, Random, 95% CI)	1.00 [0.63, 1.57]
31 Proportion of women with amenorrhea at cycle 6	1	797	Odds Ratio (M‐H, Random, 95% CI)	1.35 [0.76, 2.43]
32 Proportion of women with amenorrhea at cycle 12	1	10	Odds Ratio (M‐H, Random, 95% CI)	1.0 [0.03, 33.32]
33 Total discontinuation within 3 cycles	1	948	Odds Ratio (M‐H, Random, 95% CI)	0.80 [0.50, 1.28]
34 Discontinuation due to medical reasons within 3 cycles	1	948	Odds Ratio (M‐H, Random, 95% CI)	0.71 [0.42, 1.19]
35 Discontinuation due to cycle disturbances within 3 cycles	1	948	Odds Ratio (M‐H, Random, 95% CI)	0.73 [0.33, 1.61]
36 Total discontinuation within 6 cycles	2	1503	Odds Ratio (M‐H, Random, 95% CI)	1.00 [0.75, 1.33]
37 Discontinuation due to medical reasons within 6 cycles	2	1503	Odds Ratio (M‐H, Random, 95% CI)	0.71 [0.36, 1.43]
38 Discontinuation due to cycle disturbances within 6 cycles	1	948	Odds Ratio (M‐H, Random, 95% CI)	1.05 [0.58, 1.92]
39 Total discontinuation within 12 cycles	1	197	Odds Ratio (M‐H, Random, 95% CI)	1.44 [0.81, 2.57]
40 Discontinuation due to medical reasons within 12 cycles	1	197	Odds Ratio (M‐H, Random, 95% CI)	1.45 [0.44, 4.72]
41 Discontinuation due to cycle disturbances within 12 cycles	1	197	Odds Ratio (M‐H, Random, 95% CI)	7.29 [0.37, 143.08]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Proportion of cycles with intermenstrual bleeding within 6 cycles	1	987	Odds Ratio (M‐H, Random, 95% CI)	0.76 [0.56, 1.01]
2 Proportion of cycles with amenorrhea within 6 cycles	1	987	Odds Ratio (M‐H, Random, 95% CI)	0.02 [0.00, 0.18]
3 Total discontinuation within 6 cycles	1	236	Odds Ratio (M‐H, Random, 95% CI)	0.87 [0.52, 1.47]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Proportion of cycles with intermenstrual bleeding within 6 cycles	1	1005	Odds Ratio (M‐H, Random, 95% CI)	1.37 [1.05, 1.80]
2 Proportion of cycles with amenorrhea within 6 cycles	1	1005	Odds Ratio (M‐H, Random, 95% CI)	0.59 [0.35, 1.01]
3 Total discontinuation within 6 cycles	1	234	Odds Ratio (M‐H, Random, 95% CI)	0.69 [0.41, 1.18]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Proportion of women with intermenstrual bleeding within 12 cycles	1	260	Odds Ratio (M‐H, Random, 95% CI)	0.59 [0.29, 1.18]
2 Proportion of women with amenorrhea within 12 cycles	1	260	Odds Ratio (M‐H, Random, 95% CI)	0.03 [0.00, 0.43]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Proportion of women with intermenstrual bleeding within 12 cycles	1	245	Odds Ratio (M‐H, Random, 95% CI)	1.06 [0.55, 2.02]
2 Proportion of women with amenorrhea within 12 cycles	1	245	Odds Ratio (M‐H, Random, 95% CI)	0.25 [0.08, 0.76]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 4 cycles	2	729	Odds Ratio (M‐H, Random, 95% CI)	1.97 [0.36, 10.83]
2 Proportion of cycles with spotting within 3 cycles	1	756	Odds Ratio (M‐H, Random, 95% CI)	1.08 [0.70, 1.67]
3 Proportion of cycles with breakthrough bleeding within 3 cycles	1	756	Odds Ratio (M‐H, Random, 95% CI)	0.95 [0.50, 1.78]
4 Proportion of cycles with spotting and breakthrough bleeding within 3 cycles	1	756	Odds Ratio (M‐H, Random, 95% CI)	1.10 [0.77, 1.57]
5 Proportion of cycles with intermenstrual bleeding within 3 cycles	2	1367	Odds Ratio (M‐H, Random, 95% CI)	1.22 [0.96, 1.54]
6 Proportion of women with spotting at cycle 3	1	232	Odds Ratio (M‐H, Random, 95% CI)	2.03 [0.74, 5.54]
7 Proportion of women with breakthrough bleeding at cycle 3	1	232	Odds Ratio (M‐H, Random, 95% CI)	1.52 [0.53, 4.33]
8 Proportion of women with spotting and breakthrough bleeding at cycle 3	1	232	Odds Ratio (M‐H, Random, 95% CI)	0.73 [0.39, 1.38]
9 Proportion of women with intermenstrual bleeding at cycle 3	2	420	Odds Ratio (M‐H, Random, 95% CI)	1.18 [0.79, 1.75]
10 Proportion of cycles with amenorrhea within 3 cycles	2	1367	Odds Ratio (M‐H, Random, 95% CI)	0.87 [0.51, 1.48]
11 Proportion of women with amenorrhea at cycle 3	2	330	Odds Ratio (M‐H, Random, 95% CI)	0.38 [0.07, 1.92]
12 Total discontinuation within 4 cycles	2	729	Odds Ratio (M‐H, Random, 95% CI)	0.83 [0.62, 1.11]
13 Discontinuation due to medical reasons within 4 cycles	2	729	Odds Ratio (M‐H, Random, 95% CI)	0.68 [0.31, 1.47]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 6 cycles	1	480	Odds Ratio (M‐H, Random, 95% CI)	2.77 [0.11, 68.38]
2 Pregnancy per woman within 12 cycles	1	168	Odds Ratio (M‐H, Random, 95% CI)	1.0 [0.06, 16.26]
3 Proportion of cycles with spotting within 6 cycles	1	2515	Odds Ratio (M‐H, Random, 95% CI)	0.96 [0.72, 1.28]
4 Proportion of cycles with breakthrough bleeding within 6 cycles	1	2515	Odds Ratio (M‐H, Random, 95% CI)	0.82 [0.48, 1.43]
5 Proportion of cycles with spotting and breakthrough bleeding within 6 cycles	1	2515	Odds Ratio (M‐H, Random, 95% CI)	0.49 [0.33, 0.73]
6 Proportion of women with spotting at cycle 3	2	579	Odds Ratio (M‐H, Random, 95% CI)	1.13 [0.54, 2.33]
7 Proportion of women with breakthrough bleeding at cycle 3	1	160	Odds Ratio (M‐H, Random, 95% CI)	2.85 [0.73, 11.17]
8 Proportion of women with breakthrough bleeding (with or without spotting) at cycle 3	1	419	Odds Ratio (M‐H, Random, 95% CI)	0.73 [0.33, 1.60]
9 Proportion of women with spotting at cycle 6	2	510	Odds Ratio (M‐H, Random, 95% CI)	1.03 [0.50, 2.12]
10 Proportion of women with breakthrough bleeding at cycle 6	1	158	Odds Ratio (M‐H, Random, 95% CI)	1.0 [0.36, 2.81]
11 Proportion of women with breakthrough bleeding (with or without spotting) at cycle 6	1	352	Odds Ratio (M‐H, Fixed, 95% CI)	0.57 [0.20, 1.65]
12 Proportion of women with spotting at cycle 12	1	144	Odds Ratio (M‐H, Random, 95% CI)	1.5 [0.40, 5.56]
13 Proportion of women with breakthrough bleeding at cycle 12	1	144	Odds Ratio (M‐H, Random, 95% CI)	0.97 [0.27, 3.51]
14 Proportion of cycles with amenorrhea within 6 cycles	1	2403	Odds Ratio (M‐H, Random, 95% CI)	0.60 [0.17, 2.14]
15 Proportion of cycles with amenorrhea within 12 cycles	1	2515	Odds Ratio (M‐H, Random, 95% CI)	0.64 [0.20, 2.01]
16 Proportion of women with amenorrhea at cycle 3	2	579	Odds Ratio (M‐H, Random, 95% CI)	1.30 [0.15, 11.52]
17 Proportion of women with amenorrhea at cycle 6	2	510	Odds Ratio (M‐H, Random, 95% CI)	0.49 [0.04, 5.56]
18 Proportion of women with amenorrhea at cycle 12	2	160	Odds Ratio (M‐H, Random, 95% CI)	0.31 [0.03, 3.10]
19 Total discontinuation within 6 cycles	2	648	Odds Ratio (M‐H, Random, 95% CI)	0.89 [0.59, 1.35]
20 Discontinuation due to medical reasons within 6 cycles	1	480	Odds Ratio (M‐H, Random, 95% CI)	0.87 [0.49, 1.54]
21 Discontinuation due to cycle disturbances within 6 cycles	1	480	Odds Ratio (M‐H, Random, 95% CI)	0.76 [0.23, 2.53]
22 Total discontinuation within 12 cycles	1	168	Odds Ratio (M‐H, Random, 95% CI)	0.82 [0.35, 1.96]
23 Discontinuation due to medical reasons within 12 cycles	1	168	Odds Ratio (M‐H, Random, 95% CI)	2.02 [0.18, 22.76]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 13 cycles	1	1613	Odds Ratio (M‐H, Random, 95% CI)	1.00 [0.14, 7.09]
2 Total discontinuation within 13 cycles	1	1613	Odds Ratio (M‐H, Random, 95% CI)	1.09 [0.88, 1.36]
3 Discontinuation due to medical reasons within 13 cycles	1	1613	Odds Ratio (M‐H, Random, 95% CI)	0.85 [0.60, 1.21]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 13 cycles	1	1614	Odds Ratio (M‐H, Random, 95% CI)	0.66 [0.11, 3.99]
2 Total discontinuation within 13 cycles	1	1614	Odds Ratio (M‐H, Random, 95% CI)	0.93 [0.75, 1.16]
3 Discontinuation due to medical reasons within 13 cycles	1	1614	Odds Ratio (M‐H, Random, 95% CI)	1.11 [0.77, 1.60]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 6 cycles	1	373	Odds Ratio (M‐H, Random, 95% CI)	0.99 [0.14, 7.14]
2 Proportion of women with spotting or breakthrough bleeding at cycle 6	1	231	Odds Ratio (M‐H, Random, 95% CI)	0.26 [0.14, 0.51]
3 Proportion of women with amenorrhea at cycle 6	1	231	Odds Ratio (M‐H, Random, 95% CI)	0.17 [0.06, 0.45]
4 Total discontinuation within 6 cycles	1	373	Odds Ratio (M‐H, Random, 95% CI)	1.04 [0.68, 1.58]
5 Discontinuation due to adverse events within 6 cycles	1	373	Odds Ratio (M‐H, Random, 95% CI)	0.99 [0.50, 1.98]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 6 cycles	1	309	Odds Ratio (M‐H, Random, 95% CI)	3.02 [0.31, 29.35]
2 Proportion of cycles with spotting within 6 cycles	1	1650	Odds Ratio (M‐H, Random, 95% CI)	0.59 [0.42, 0.81]
3 Proportion of cycles with breakthrough bleeding within 6 cycles	1	1650	Odds Ratio (M‐H, Random, 95% CI)	1.16 [0.66, 2.04]
4 Proportion of cycles with amenorrhea within 6 cycles	1	1650	Odds Ratio (M‐H, Random, 95% CI)	0.57 [0.34, 0.96]
5 Total discontinuation within 6 cycles	1	309	Odds Ratio (M‐H, Random, 95% CI)	0.95 [0.52, 1.73]
6 Discontinuations due to adverse events within 6 cycles	1	309	Odds Ratio (M‐H, Random, 95% CI)	0.79 [0.21, 3.00]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 6 cycles	1	309	Odds Ratio (M‐H, Random, 95% CI)	7.09 [0.36, 138.46]
2 Proportion of cycles with spotting within 6 cycles	1	1679	Odds Ratio (M‐H, Random, 95% CI)	0.65 [0.47, 0.91]
3 Proportion of cycles with breakthrough bleeding within 6 cycles	1	1679	Odds Ratio (M‐H, Random, 95% CI)	0.98 [0.57, 1.68]
4 Proportion of cycles with amenorrhea within 6 cycles	1	1679	Odds Ratio (M‐H, Random, 95% CI)	0.37 [0.23, 0.60]
5 Total discontinuation within 6 cycles	1	309	Odds Ratio (M‐H, Random, 95% CI)	1.04 [0.57, 1.92]
6 Discontinuations due to adverse events within 6 cycles	1	309	Odds Ratio (M‐H, Random, 95% CI)	1.33 [0.29, 6.06]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 13 cycles	1	2814	Odds Ratio (M‐H, Random, 95% CI)	0.71 [0.38, 1.34]
2 Proportion of cycles with breakthrough bleeding within 3 cycles	1	7272	Odds Ratio (M‐H, Random, 95% CI)	0.40 [0.35, 0.47]
3 Proportion of cycles with breakthrough bleeding/spotting within 3 cycles	1	7272	Odds Ratio (M‐H, Random, 95% CI)	0.40 [0.35, 0.45]
4 Proportion of cycles with breakthrough bleeding within 6 cycles	1	13692	Odds Ratio (M‐H, Random, 95% CI)	0.45 [0.40, 0.50]
5 Proportion of cycles with breakthrough bleeding/spotting within 6 cycles	1	13692	Odds Ratio (M‐H, Random, 95% CI)	0.44 [0.40, 0.48]
6 Proportion of cycles with breakthrough bleeding within 12 cycles	1	16519	Odds Ratio (M‐H, Random, 95% CI)	0.45 [0.41, 0.50]
7 Proportion of cycles with breakthrough bleeding/spotting within 12 cycles	1	16519	Odds Ratio (M‐H, Random, 95% CI)	0.45 [0.41, 0.49]
8 Proportion of women with breakthrough bleeding at cycle 3	1	2330	Odds Ratio (M‐H, Random, 95% CI)	0.36 [0.27, 0.49]
9 Proportion of women with breakthrough bleeding/spotting at cycle 3	1	2330	Odds Ratio (M‐H, Random, 95% CI)	0.44 [0.35, 0.55]
10 Proportion of women with breakthrough bleeding at cycle 6	1	2118	Odds Ratio (M‐H, Random, 95% CI)	0.42 [0.31, 0.57]
11 Proportion of women with breakthrough bleeding/spotting at cycle 6	1	2118	Odds Ratio (M‐H, Random, 95% CI)	0.40 [0.32, 0.52]
12 Proportion of women with breakthrough bleeding at cycle 12	1	444	Odds Ratio (M‐H, Random, 95% CI)	0.53 [0.24, 1.16]
13 Proportion of women with breakthrough bleeding/spotting at cycle 12	1	444	Odds Ratio (M‐H, Random, 95% CI)	0.48 [0.27, 0.86]
14 Proportion of women with unscheduled bleeding at cycle 3	1	2478	Odds Ratio (M‐H, Random, 95% CI)	0.52 [0.44, 0.62]
15 Proportion of women with unscheduled bleeding at cycle 6	1	2222	Odds Ratio (M‐H, Random, 95% CI)	0.53 [0.44, 0.64]
16 Mean number of days of unscheduled bleeding/spotting at cycle 3	1	719	Mean Difference (IV, Random, 95% CI)	‐0.45 [‐0.85, ‐0.05]
17 Mean number of days of unscheduled bleeding/spotting at cycle 6	1	567	Mean Difference (IV, Random, 95% CI)	0.20 [‐0.22, 0.62]
18 Proportion of cycles with amenorrhea within 13 cycles	1	16957	Odds Ratio (M‐H, Random, 95% CI)	0.05 [0.04, 0.07]
19 Total discontinuation within 6 cycles	1	2089	Odds Ratio (M‐H, Random, 95% CI)	1.01 [0.82, 1.26]
20 Total discontinuation within 13 cycles	1	805	Odds Ratio (M‐H, Random, 95% CI)	1.10 [0.82, 1.46]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pregnancy per woman within 3 cycles	1	334	Odds Ratio (M‐H, Random, 95% CI)	3.02 [0.12, 74.62]
2 Proportion of women with unscheduled bleeding episodes within 3 cycles	1	332	Odds Ratio (M‐H, Random, 95% CI)	0.42 [0.25, 0.70]
3 Unscheduled bleeding days in cycle 3	1	316	Mean Difference (IV, Random, 95% CI)	‐1.0 [‐1.57, ‐0.43]
4 Unscheduled bleeding days within 3 cycles	1	332	Mean Difference (IV, Random, 95% CI)	‐1.5 [‐2.75, ‐0.25]
5 Total discontinuation within 3 cycles	1	355	Odds Ratio (M‐H, Random, 95% CI)	1.16 [0.62, 2.17]
6 Discontinuations due to adverse events within 3 cycles	1	355	Odds Ratio (M‐H, Random, 95% CI)	0.74 [0.16, 3.36]

PERMALINK

Triphasic versus monophasic oral contraceptives for contraception

Huib AAM Van Vliet

David A Grimes

Laureen M Lopez

Kenneth F Schulz

Frans M Helmerhorst

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Background

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Contraceptive efficacy

Cycle control

Discontinuation

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Assessment of heterogeneity

Data synthesis

Contraceptive efficacy

Cycle control

Discontinuation

Sensitivity analysis

Results

Description of studies

Included studies

Triphasic levonorgestrel oral contraceptives versus monophasic oral contraceptives

Triphasic norethindrone oral contraceptives versus monophasic oral contraceptives

Triphasic gestodene oral contraceptives versus monophasic oral contraceptives

Triphasic norgestimate oral contraceptives versus monophasic oral contraceptives

Excluded studies

Risk of bias in included studies

Allocation

Blinding

Incomplete outcome data

Other potential sources of bias

A priori hypothesis and sample size calculation

Funding

Effects of interventions

Contraceptive effectiveness

1.1. Analysis.

1.2. Analysis.

2.1. Analysis.

3.1. Analysis.

4.1. Analysis.

4.2. Analysis.

9.1. Analysis.

11.1. Analysis.

11.2. Analysis.

12.1. Analysis.

13.1. Analysis.

14.1. Analysis.

15.1. Analysis.

16.1. Analysis.

17.1. Analysis.

18.1. Analysis.

1. Pregnancies and total study cycles for triphasic and monophasic formulations.

Cycle control

50‐75‐125 μg levonorgestrel (LNG) and 30‐40‐30 μg ethinylestradiol (EE) versus 150 μg LNG and 30 μg EE (Comparison 1)

1.3. Analysis.

1.8. Analysis.

1.5. Analysis.

1.6. Analysis.

Methods	Randomized controlled trial. Randomization by random number tables. The use of blinding is not described. We were unable to reach the author.
Participants	170 women at 3 sites in Indonesia. Inclusion criteria were healthy women. Exclusion criteria were contraindications to oral contraceptives, use of hormonal contraceptives within the previous 3 cycles before enrollment and current pregnancy. The mean age of the 2 groups of participants differs.
Interventions	Triphasic gestodene/ethinylestradiol (50‐70‐100 μg GTD and 30‐40‐30 μg EE in a 6/5/10 days regimen) [SHD 415 G] versus   monophasic desogestrel/ethinylestradiol (150 μg DSG and 30 μg EE for 21 days) [Marvelon].
Outcomes	Primary outcomes measures are: efficacy; side effects; cycle control; continuation and reasons for discontinuation; pill intake errors. The method to collect data is not described. The report does not describe the definitions of breakthrough bleeding and spotting.
Notes	The report does not provide an a priori hypothesis or a sample size or power calculation. Study duration: 12 cycles. 3 women in the triphasic group and 5 women in the monophasic group discontinued early. The reasons for discontinuation are described. The report does not mention loss to follow up or withdrawals because of protocol violations. Analysis not according to intention‐to‐treat principle. The trial was supported by the manufacturer of the studied triphasic gestodene/ethinylestradiol pill (Schering).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	Numbered containers

Methods	Randomized controlled trial without blinding. The method of randomization and the use of allocation concealment are not described. We were unable to reach the author.
Participants	480 women at 14 study sites in Europe and New Zealand. Inclusion criteria were healthy women under 40 years of age who were at risk of becoming pregnant and had regular 21 to 35 day menstrual cycles. The report does not provide exclusion criteria for the study. Switchers were included in the study.
Interventions	Triphasic gestodene/ethinylestradiol (50‐70‐100 μg GTD and 30‐40‐30 μg EE in a 6/5/10 days regimen, N=250 for 6 cycles of whom N=13 continued for an additional 6 cycles) [no brand name described] versus   monophasic desogestrel/ethinylestradiol (150 μg DSG and 30 μg EE for 21 days, N=230 for 6 cycles of whom N=8 continued for an additional 6 cycles) [no brand name described].
Outcomes	Principal outcome measures are: pregnancy; cycle control; cycle length and bleeding intensity; side effects; laboratory and cytology changes; blood pressure; bodyweight; compliance; discontinuation and reasons for discontinuation. The method to collect data is not described. The report does not describe the definitions of breakthrough bleeding and spotting.
Notes	The report does not describe an a priori hypothesis or sample size or power calculation. Study duration: 6 and 12 cycles. 49 women in the triphasic group and 50 women in the monophasic group discontinued early. The reasons for discontinuation are described. 7 women in the triphasic group and 5 women in the monophasic group were lost to follow up. 6 women in the triphasic group and 8 women in the monophasic group were withdrawn because of protocol violations. Women who missed pills in cycle 6 were excluded from analysis of cycle control. Analysis not according to intention‐to‐treat principle. The paper does not report information on support.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	No information

Methods	Randomized controlled trial. The method of randomization and the use of blinding are not described. Communication with the author indicated a computer‐generated random allocation sequence and blinding of participants and investigators. Method of randomizing is unclear.
Participants	862 women at 12 sites in Sweden. Inclusion criteria were that women had to fulfill the current recommendations for oral contraceptive use. Limited information on baseline characteristics. Switchers were included in the study.
Interventions	Triphasic levonorgestrel/ethinylestradiol (50‐75‐125 μg LNG and 30‐ 40‐30 μg EE in a 6/5/10 days regimen, N=210 for 6 cycles of whom N=89 continued for an additional 6 cycles) [Trionetta 21] versus   triphasic levonorgestrel/ethinylestradiol (50‐75‐125 μg LNG and 30‐40‐30 μg EE in a 6/5/10 days regimen and 7 days of placebo tablets, N=207 for 6 cycles of whom N=93 continued for an additional 6 cycles) [Trionetta 28] versus   monophasic levonorgestrel/ethinylestradiol (150 μg LNG and 30 μg ethinylestradiol, N=418 for 6 cycles of whom N=189 continued for an additional 6 cycles) [Neovletta].
Outcomes	Primary outcomes measures are: pregnancy; side effects; cycle control; continuation rate and reasons for discontinuation. Use of diary cards to collect data on pill‐intake errors and cycle control. Breakthrough bleeding was defined as intermenstrual bleeding which required the use of sanitary protection and spotting as all other cases. Data on side effects were recorded if reported spontaneously.
Notes	The report does not provide an a priori hypothesis or a sample size or power calculation. Study duration: 6 and 12 cycles. 67 women in the triphasic group and 60 women in the monophasic group discontinued early in the 1 to 6 cycles trial period. 26 women in the triphasic group and 24 women in the monophasic group discontinued early in the 7 to 12 cycles trial period. Little information concerning the number and reasons for discontinuation. 27 women entered in the trial are not included in the analysis because they were lost to follow up. The report does not describe withdrawals because of protocol violations. Analysis not according to intention‐to‐treat principle. The trial was supported by the manufacturer of the studied monophasic and triphasic levonorgestrel/ethinylestradiol pills (Schering).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	Report does not mention the use of allocation concealment. Communication with the author indicated allocation concealment by numbered pharmacy packages.

Methods	Double‐blind, randomized controlled trial. Randomization by a WHO random table. The use of allocation concealment and the method of blinding are not described. We were unable to reach the authors.
Participants	279 women aged 23‐34 years in China. Inclusion criteria were healthy women aged 23 to 34 who have the ability to record menstrual cycle on a diary and have normal physical examination and PAP smear. Exclusion criteria were diabetes mellitus, heart, liver, kidney or nervous system disease, cancer, hypertension, use of hormones 2 months prior to the study, use of injectable contraceptives 6 months prior to the study.
Interventions	Triphasic levonorgestrel/ethinylestradiol (50‐75‐125 μg LNG and 30‐40‐30 μg EE in a 6/5/10 days regimen and 7 days of placebo tablets, N=96) [no brand name described] versus   monophasic levonorgestrel/ethinylestradiol (150 μg LNG and 30 μg EE for 21 days and 7 days of placebo tablets, N=93) [Microgynon] versus   monophasic norethindrone/ethinylestradiol (600 μg NET and 35 μg ethinylestradiol for 21 days and 7 days of placebo tablets, N=90) [Pill No 1].
Outcomes	Principal outcomes are: pregnancy; side effects; cycle control; discontinuation and reasons for discontinuation. Use of a diary card to collect data on cycle control. Bleeding pattern was analyzed according to the recommendations of Rodriguez 1976.
Notes	The report does not provide an a priori hypothesis or a sample size or power calculation. Study duration: 6 cycles. 17 women in each treatment group discontinued early. The reasons for discontinuation are described. No woman was lost to follow up. One woman in the triphasic group, 2 women in the monophasic levonorgestrel group and one woman in the monophasic norethindrone group were withdrawn because of protocol violations. Analysis not according to intention‐to‐treat principle. The trial was conducted by the World Health Organization.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	No information

Methods	Randomized controlled trial with blinding of the outcome assessor. The method of randomization, the use of allocation concealment and the method of blinding are not described. Communication with the author revealed no extra information.
Participants	At 4 sites in Canada, 469 women were randomized to one of the pills. However, only 391 women were admitted to the study and used the pills for at least one month. 222 women did not use OC pills at least 90 days before the study, and 247 women did use pills before the study. Inclusion criteria were healthy women aged 15 to 35 years who had a history of regular menses for two months prior to admission. Exclusion criteria were contraindications to oral contraceptives.
Interventions	Triphasic levonorgestrel/ethinylestradiol (50‐75‐125 μg LNG and 30‐40‐30 μg EE in a 6/5/10 days regimen, N=119) [Triphasil] versus   triphasic norethindrone/ethinylestradiol (500‐750‐1000 μg NET and 35 μg EE in a 7/7/7 days regimen, N=117)) [Ortho 7/7/7] versus   biphasic norethindrone/ethinylestradiol (500‐1000 μg NET and 35 μg ethinylestradiol in a 10/11 days regimen, N=116) [Ortho 10/11] versus   monophasic norethindrone acetate/ethinyl estradiol (1500 μg NETA and 30 μg EE, N=117) [Loestrin]. In the pre‐study user group, 16 participants already used Triphasil, 5 Loestrin, 8 Ortho 10/11 and 8 Ortho 7/7/7.
Outcomes	Primary outcomes measures are: side effects; cycle control; continuation, discontinuation rates and reason for discontinuation. Use of daily diary method to collect data on cycle control and side effects. Breakthrough bleeding was defined as free flow, much like menses occurring during the 21 days of active medication and requiring sanitary protection; and spotting as bleeding during the active medication, which is limited to minor staining, whether or not sanitary protection was used.
Notes	The report provides an a priori hypothesis and an adequate sample size calculation. Study duration: 6 cycles. 49 women in the monophasic group, 35 women in the biphasic group, 46 women in the levonorgestrel triphasic group and 39 women in the norethindrone triphasic group discontinued early. The reasons for discontinuation are partially described. The report does not describe the number of women lost to follow up or withdrawn because of protocol violations. Analysis not according to intention‐to‐treat principle. 78 women who were randomized but did not take the oral contraceptives for at least one cycle were excluded from the analysis. Breakthrough bleeding includes all intermenstrual bleeding except continued menstrual flow in this review. The trial was sponsored by the manufacturer of the monophasic norethindrone acetate/ethinylestradiol pill (Parke‐Davis).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	No information

Study	Reason for exclusion
Bancroft 1987	The study examines mood and sexuality.
Christie 1989	The report does not mention how participants were assigned to groups. We attempted without success to reach the author.
Dik 1984	The report does not mention how participants were assigned to groups. We were unable to contact the authors.
Dubnitskaia 1988	Although described as a randomized controlled trial we learned from the author that the study is a matched cohort study.
Grace 1994	Communication with the author indicated no randomization of the allocation sequence.
Kuhl 1985	Insufficient data for analysis of spotting and breakthrough bleeding. Emphasis was on hormonal and metabolic parameters.
Matsumoto 1988	Report does not mention how participants were assigned to groups. We attempted without success to reach the author.
Otolorin 1989	The report describes allocation as systematical. We were unable to contact the author.
Perrone 1987	Report does not mention how participants were assigned to groups. We attempted without success to reach the author.
Rubio‐Lotvin 1992	The report does not mention how participants were assigned to groups. We could not reach the authors.

Trial name or title	Cycle Control and Safety of E2‐DRSP
Methods	Randomized, double‐blind, multicenter
Participants	600 healthy women, 18 to 35 years
Interventions	6 different regimens of drospirenone and ethinyl estradiol (including monophasic and triphasic) over 7 cycles (details not provided)
Outcomes	include bleeding patterns and cycle control
Starting date	March 2008
Contact information	Bayer Study Director (no more information provided)
Notes	Completed June 2009. No report found; searched PubMed and http://www.clinicalstudyresults.org/