Review article

Biphasic versus triphasic oral contraceptives for contraception�

Hubertus A.A.M. van Vliet, David A. Grimes*, Frans M. Helmerhorst, Kenneth F. SchulzFamily Health International, Research Triangle Park, NC, USA 27709

Abstract

Side effects caused by oral contraceptives discourage compliance with and continuation of oral contraceptives. A suggested disadvantageof biphasic oral contraceptive pills compared to triphasic oral contraceptive pills is an increase in breakthrough bleeding. We examined thispotential disadvantage by conducting a systematic review comparing biphasic oral contraceptives with triphasic oral contraceptives in termsof efficacy, cycle control, and discontinuation because of side effects.

We included randomized, controlled trials comparing any biphasic oral contraceptive with any triphasic oral contraceptive when usedto prevent pregnancy. Only two trials of limited quality met our inclusion criteria. Larranaga compared two biphasic and one triphasic pills,each containing levonorgestrel and ethinyl estradiol. No important differences emerged, and the frequency of discontinuation because ofmedical problems was similar with all three pills. Percival-Smith compared a biphasic pill containing norethindrone (Ortho 10/11) with atriphasic pill containing levonorgestrel (Triphasil) and another triphasic pill containing norethindrone (Ortho 7/7/7). The biphasic pill hadinferior cycle control compared with the levonorgestrel triphasic pill.

The available evidence is limited and of poor quality; the internal validity of these trials is questionable. Given that caveat, the biphasicpill containing norethindrone was associated with inferior cycle control compared with the triphasic pill containing levonorgestrel. Thissuggests that the choice of progestin may be more important that the phasic regimen in determining bleeding patterns. © 2002 ElsevierScience Inc. All rights reserved.

Keywords: Biphasic; Triphasic; Oral contraceptives; Systematic review

1. Introduction

Side effects caused by oral contraceptives discouragecompliance with and continuation of oral contraceptives [1].Three approaches have been used to decrease these adverseeffects: reduction of the steroid dose, development of newsteroids, and new formulas and schedules of administration.The third strategy led to the development of both bi- andtriphasic pills.

Biphasic and triphasic oral contraceptives purportedlyattempt to “mimic” the pattern of rising and falling ofestrogen and progesterone as seen during the normal men-strual cycle [2]. The biphasic pill soon yielded to the tripha-sic pill because of a suggested increase in breakthroughbleeding associated with biphasic oral contraceptive pill use[3]. We conducted this systematic review to examine this

potential disadvantage of biphasic oral contraceptives com-pared with triphasic pills.

The aim of this review was to compare biphasic oralcontraceptive pills with triphasic oral contraceptive pills.Our a priori hypotheses were that biphasic and triphasicpills have similar contraceptive efficacy but that triphasicoral contraceptives cause fewer side effects, give a bettercycle control, and have higher continuation rates.

2. Materials and methods

We included only randomized, controlled trials ofhealthy women of reproductive age without contraindica-tions for oral contraceptive use who desired to use oralcontraceptives for preventing pregnancy. Trials comparingany biphasic oral contraceptive pill with any triphasic oralcontraceptive pill when used to prevent pregnancy wereeligible for inclusion. Both 21- and 28-pill packages wereincluded. We excluded studies examining sequential pills(those containing estrogen alone early in the cycle, followedby estrogen plus progestin later in the cycle). We alsoexcluded studies comparing biphasic pills with triphasic

� This systematic review was conducted with support from the USNational Institute of Child Health and Human Development and the USAgency for International Development.

* Corresponding author. Tel.: �1-919-544-7040; fax: �1-919-544-7261.

E-mail address: dgrimes@fhi.org (D. Grimes).

Contraception 65 (2002) 321–324

0010-7824/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved.PII: S0010-7824(01)00314-6

pills when used as a treatment and not as a contraceptive.Principal outcome measures included the incidence of ac-cidental pregnancy, spotting, breakthrough bleeding, amen-orrhea, intermenstrual bleeding, and discontinuation be-cause of side effects. We excluded studies that lookedprimarily at metabolic outcome measures and folliculargrowth.

We searched computerized databases MEDLINE, EM-BASE, Popline and Cochrane Controlled Trials Register(CCTR) for publications comparing monophasic, biphasic,or triphasic oral contraceptives.

We searched MEDLINE using the search strategy:contraceptives, oral[MeSH Terms] AND (monophasi-c[ALL] OR biphasic[ALL] OR triphasic[ALL] OR mul-tiphasic[ALL]) AND (clinical trials[MeSH Terms] OR clin-ical trial*[ALL] OR controlled clinical trial*[ALL] ORcomparative stud*[ALL] OR compar* OR randomized con-trolled trial[ALL] OR random allocation[MeSH Terms] ORrandom allocation[Text Word] OR random[ALL] OR dou-ble-blind method[MeSH Terms] OR double blind method-[Text Word] OR single-blind method[MeSH Terms] ORsingle blind method[Text Word] OR multicenterstud*[ALL])

We searched POPLINE using the search strategy:(kw) oral contraceptives AND (tw) (monophasic OR bipha-sic OR triphasic OR multiphasic) AND (tw) (compar* ORclinical trials OR comparative studies OR random OR dou-ble blind studies)

We searched EMBASE using the search strategy:

1. oral contraceptive agent2. biphasic3. triphasic4. multiphasic5. 2 OR 3 OR 46. 1 AND 57. monophasic8. 6 AND 7

We searched the CCTR using the search strategy:

1. (contraceptives and oral)2. monophasic3. biphasic4. triphasic5. multiphasic6. (((#2 or #3) or #4) or #5)7. (#1 and #6)

We used a computer to search the holdings of the FamilyHealth International library for all relevant trials, bookchapters, and review articles. We searched all relevant bookchapters and review articles identified with the above MED-LINE, EMBASE, Popline, CCTR, reference lists, and li-brary searches for all relevant trials. We reviewed the ref-erence lists of all identified studies for additional previouslyunidentified trials.

We attempted to contact the authors of all included trials.

We also wrote a letter to pharmaceutical companies in theUS and in Europe who market oral contraceptives. In thecontact letters we provided a list of studies identified andasked if correspondents knew of unpublished or publishedtrials we had missed.

Two reviewers evaluated the titles and abstracts foundduring the literature searches, and we photocopied all po-tentially relevant articles. We had one French article [4]translated into English. Then each reviewer independentlyexamined the retrieved studies for possible inclusion, andthe methodological quality of the trials was assessed byusing the Cochrane guidelines. We focused especially onthe method of randomization, the use of allocation conceal-ment, the use of blinding, and the exclusion of participantsafter randomization.

After inclusion of a study, each reviewer abstracted thedata. No disagreements about the inclusion of studies or theabstracted data occurred. We wrote a letter to the authors ofthe two included studies and asked for additional informa-tion about the methods of the study and the various outcomemeasures. One reviewer then entered the abstracted datainto RevMan 3.1, later imported into RevMan 4.1. Anotherreviewer verified the correct entry of the data. We calculatedPeto odds ratios (ORs) with a 95% CI for the outcomemeasures breakthrough bleeding, spotting, withdrawalbleeding, intermenstrual bleeding, and discontinuation be-cause of medical reasons.

3. Results

Two studies met the criteria for this review. One study[5] compared three pills (two biphasic and one triphasic):

Y a biphasic pill (termed “Alpha”) containing levonorg-estrel 50 mcg and ethinyl estradiol 50 mcg for 11days, followed by levonorgestrel 125 mcg and ethinylestradiol 50 mcg for 10 days,

Y a biphasic pill (termed “Beta”) composed oflevonorgestrel 150 mcg and ethinyl estradiol 30 mcgfor 7 days, followed by levonorgestrel 200 mcg andethinyl estradiol 40 mcg for 14 days,

Y a triphasic pill (termed “Gamma”) containinglevonorgestrel 50 mcg and ethinyl estradiol 20 mcgfor 7 days, levonorgestrel 50 mcg and ethinyl estra-diol 50 mcg for 7 days, then levonorgestrel 125 mcgand ethinyl estradiol 30 mcg for 7 days.

The other study [6] compared three marketed products:

Y a biphasic pill containing norethindrone 500-1000mcg and ethinyl estradiol 35 mcg (Ortho 10/11),

Y a triphasic pill containing levonorgestrel 50–75-125mcg and ethinyl estradiol 30–40-30 mcg (Triphasil),

Y a triphasic pill containing norethindrone 500-750-1000 mcg and ethinyl estradiol 35 mcg (Ortho 7/7/7).

322 H.A.A.M. van Vliet et al. / Contraception 65 (2002) 321–324

We excluded four studies. One [4] did not report that theallocation method was randomized, and we were unable toreach the author. Another study [7] examined a sequentialpill. Because of probable fraud, [8] we excluded two studiesby Briggs [9,10].

Neither study [5,6] adequately described the method ofrandomization and allocation concealment. One study, [6]sponsored by Parke-Davis, reportedly kept investigatorsblinded as to treatment. This report [6] provided an a priorihypothesis and a sample size calculation. The other study[5] did not report an a priori hypothesis or a sample size orpower calculation. We received supplemental informationfrom Percival-Smith [6] but not from Larranaga [5].

Both trials had high losses after randomization. Of the458 women who participated in one trial [5], 252 womendiscontinued. Losses to follow-up ranged from 40% to 54%in the three treatment arms. In the other trial [6], 169 womenout of 469 participating women withdrew at some time up to6 months. Seventy-eight dropped out before or during thefirst cycle of pill use. This raises the strong possibility ofselection bias in both studies.

We could not aggregate the clinical outcomes of theincluded studies because they examined different kinds ofpills [5,6]. Moreover, even if they had examined the samepills, we would not choose to aggregate two studies of poormethodological quality.

Larranaga et al. [5] examined 1269 user cycles of bipha-sic preparation Alpha, 1163 user cycles of biphasic prepa-ration Beta, and 1154 user cycles of triphasic preparationGamma. No pregnancies occurred during the study period.Life table continuation rates after 12 cycles were lower withthe triphasic pill (Gamma; 40%) than with Alpha (50%) orBeta (44%). The discontinuation rates for medical reasonsdid not differ significantly among the three preparations.Patterns of breakthrough bleeding, spotting, and absence ofwithdrawal bleeding were impossible to interpret because ofinsufficient detail.

Percival-Smith [6] examined 533 user cycles of a bipha-sic Ortho 10/11, 506 cycles of Triphasil, and 524 cycles ofthe triphasic Ortho 7/7/7. In the comparison between Ortho10/11 and Triphasil, the biphasic pill (Ortho 10/11) wasassociated with significantly more cycles with intermen-strual bleeding (OR 1.7; 95% CI 1.3–2.2). Similarly, cycleswithout a withdrawal bleed were significantly more com-mon with the biphasic pill (OR 6.5; 95% CI 3.1–13). How-ever, rates of study discontinuation because of intermen-strual bleeding were similar for both groups.

In the comparison between Ortho 10/11 and Ortho 7/7/7,both preparations were associated with a similar frequencyof cycles with intermenstrual bleeding. The proportions ofcycles without withdrawal bleeding were also comparable.Finally, the proportion of women who discontinued thestudy because of intermenstrual bleeding was the same inboth groups.

4. Discussion

We identified only two randomized controlled trialscomparing a biphasic regimen with a triphasic regimen[5,6]. The reports do not describe the method of random-ization and the use (if any) of allocation concealment. Bothstudies examined a modest number of user cycles. Moreimportant, a large proportion of women in both trials werelost to follow-up after randomization. Because of the largelosses after randomization, these data may be more appro-priately considered observational, and the internal validityof these trials is questionable. Because of small samplesizes, we were unable to address our a priori hypothesisconcerning contraceptive efficacy. However, we foundsome support [6] for our hypothesis of better cycle controlwith at least one triphasic regimen.

The earlier trial [5] contributed little information. How-ever, it showed no important differences among the threepills studied. The later trial [6] found that the biphasic pill(Ortho 10/11) caused significantly more cycles with inter-menstrual bleeding or no withdrawal bleeding than did onetriphasic pill (Triphasil). This difference was not seen whenthe biphasic pill was compared with a triphasic pill (Ortho7/7/7) containing the same progestin (norethindrone).

Other randomized, controlled trials have found thattriphasic pills containing levonorgestrel provide better cyclecontrol than do those containing norethindrone. One trialcomparing two norethindrone-containing triphasic pills ver-sus a levonorgestrel-containing triphasic pill [11] found thatlevonorgestrel was associated with the lowest frequency ofintermenstrual bleeding. Another trial comparing thesethree triphasic formulations [12] corroborated this observa-tion. Thus, the poorer cycle control with the biphasic pill(Ortho 10/11) versus Triphasil in this review [6] may reflectthe progestin used (norethindrone) rather than the phasicformulation itself. Levonorgestrel has a greater bioavailabil-ity, longer serum half-life, and greater relative binding af-finity in humans than does norethindrone [13]. These fea-tures may translate into better control of menstrual bleeding.

In conclusion, the systematic review found one study thatshowed a biphasic pill containing norethindrone (Ortho 10/11) provided inferior cycle control to a triphasic pill con-taining levonorgestrel (Triphasil). Aside from this, clini-cians have limited information to distinguish betweenbiphasic and triphasic pills. However, given the infrequentuse of biphasic pills in most of the world, further compar-ative trials do not appear justified.

Acknowledgment

A version of this review has been published in TheCochrane Library. Cochrane systematic reviews are regu-larly updated to include new research and in response tocomments and criticisms from readers. If you wish to com-ment on this or other Cochrane reviews of biphasic or

323H.A.A.M. van Vliet et al. / Contraception 65 (2002) 321–324

triphasic oral contraceptives, please contact David A.Grimes, M.D., Family Health International, P.O. Box13950, Research Triangle Park, NC 27709, U.S. Email:dgrimes@fhi.org.

The results of a Cochrane Review can be interpreteddifferently, depending on people’s perspectives and circum-stances. Please consider the conclusions presented carefully.They are the opinions of review authors and are not neces-sarily shared by the Cochrane Collaboration.

Dr. Grimes has consulted with or served on a speakersbureau for Berlex Laboratories, GynoPharma, Mead John-son, Organon, Ortho-McNeil, Parke-Davis, Schering, Searleand Wyeth-Ayerst. Dr. Helmerhorst had contacts with AstaMedica, Ferring, Hoechst Marion Roussel, Johnson & John-son, Merck, Novartis, Novo Nordisk, Organon, Pharmacia-Upjohn, Schering, SmithKline & Beecham, Serono, WyethAyerst and Zeneca. He supervised studies sponsored orassigned by Hoechst Marion Roussel, Johnson & Johnson,Merck, Novartis, Organon, Schering, Serono and WyethAyerst. Johnson & Johnson, Organon, Schering, and WyethAyerst have marketed oral contraceptive pills.

References

[1] Hillard PJ. Oral contraception noncompliance: the extent of the prob-lem. Adv Contracept 1992;8(Suppl 1):13–20.

[2] Upton GV. The phasic approach to oral contraception: the triphasicconcept and its clinical application. Int J Fertil 1983;28:121–40.

[3] Mishell DR. Oral contraception: past, present, and future perspec-tives. Int J Fertil 1991;36(Suppl 1):7–18.

[4] Castaigne JP. Triella-Etude clinique multicentrique francaise meneeen double aveugle. Contracept Fertil Sex 1985;13(Suppl):445–53.

[5] Larranaga A, Sartoretto JN, Winterhalter M, Navas Filho F. Clinicalevaluation of two biphasic and one triphasic norgestrel/ethinyl estra-diol regimens. Int J Fertil 1978;23:193–9.

[6] Percival-Smith RK, Yuzpe AA, Desrosiers JA, Rioux JE, Guilbert E.Cycle control on low-dose oral contraceptives: a comparative trial.Contraception 1990;42:253–62.

[7] Gaspard UJ, Romus MA, Gillain D, Duvivier J, Demey-Ponsart E,Franchimont P. Plasma hormone levels in women receiving new oralcontraceptives containing ethinyl estradiol plus levonorgestrel ordesogestrel. Contraception 1983;27:577–90.

[8] Rossiter EJ. Reflections of a whistle-blower. Nature1992;357:434–6.

[9] Briggs M, Briggs M. A randomized study of metabolic effects of fouroral contraceptive preparations containing levonorgestrel plus ethiny-loestradiol in different regimens. In: Greenblatt RB, editor. The de-velopment of a new triphasic oral contraceptive: proceedings of aspecial symposium held at the 10th World Congress on Fertility andSterility. Lancaster, UK: MTP Press, 1980. p. 79–88.

[10] Briggs MH, Briggs M. Randomized prospective studies on metaboliceffects of oral contraceptives. Acta Obstet Gynecol Scand 1982;105(Suppl):25–32.

[11] Droegemueller W, Katta LR, Bright TG, Bowes WA. Triphasicrandomized clinical trial: comparative frequency of intermenstrualbleeding. Am J Obstet Gynecol 1989;161:1407–11.

[12] Schilling LH, Bolding OT, Chenault CB, et al. Evaluation of theclinical performance of three triphasic oral contraceptives: a multi-center, randomized comparative trial. Am J Obstet Gynecol 1989;160:1264–8.

[13] Wallach M, Grimes DA, Chaney EJ, et al. Modern oral contraception.Totowa, New Jersey: Emron, Inc., 2000.

324 H.A.A.M. van Vliet et al. / Contraception 65 (2002) 321–324