Contraception 87 (2013) 567–575

Review article

Treatment of unscheduled bleeding in women using extended- orcontinuous-use combined hormonal contraception: a systematic review☆

Emily M. Godfreya,b,⁎, Maura K. Whitemana, Kathryn M. CurtisaaDivision of Reproductive Health, Centers for Disease Control and Prevention, Atlanta, GA 30341, USA

bDepartment of Obstetrics and Gynecology, University of North Carolina Chapel Hill, Chapel Hill, NC 27599, USA

Received 3 July 2012; revised 31 July 2012; accepted 7 August 2012

Abstract

Background: Unscheduled uterine bleeding is common among women who choose extended or continuous combined hormonalcontraception and may subsequently lead to method discontinuation. This systematic review evaluates the evidence on treatments forunscheduled bleeding for women using continuous or extended combined hormonal contraception.Study design: We searched the PubMed database for peer-reviewed articles that were published in any language from inception of thedatabase through January 2012 and were relevant to therapeutic or prophylactic treatments for unscheduled uterine bleeding during extendedor continuous combined hormonal contraception use. We used standard abstract forms and grading systems to summarize and assess thequality of the evidence.Results: Four articles met the inclusion criteria. Evidence from two randomized controlled trials, both of fair quality, suggested that a shorthormone-free interval of 3 or 4 days improved a current unscheduled bleeding episode. Evidence from one randomized controlled trial of fairquality suggested that oral doxycycline (100 mg twice daily) did not improve a current unscheduled bleeding episode. One good qualityrandomized controlled trial suggested that prophylactic treatment with a daily low dose of oral doxycycline (40 mg) caused earlier onset ofamenorrhea in new extended combined oral contraceptive users.Conclusions: Limited evidence suggests that introducing a short hormone-free interval when unscheduled bleeding occurs during continuousor extended hormonal contraceptive use may reduce unscheduled bleeding days; one study suggests that twice-daily oral doxycyclineinitiated at the time of unscheduled bleeding is an ineffective treatment. Limited data suggest that prophylactic low-dose daily oraldoxycycline may induce earlier amenorrhea among new extended combined hormonal contraceptive users.© 2013 Elsevier Inc. All rights reserved.

Keywords: Contraception; Combined hormonal contraception; Unscheduled bleeding; Extended contraception; Continuous contraception

1. Introduction

Alternative regimens of the combined oral contraceptivepill, the most commonly used reversible contraceptivemethod in the United States (US), have produced consider-able interest by health care providers and women alike overthe last decade [1]. Extending the duration of the active oralcontraceptive pill from the standard 28-day monthly cyclicregimen has been gaining popularity, particularly followingthe introduction of US Food and Drug Administration-

☆ Disclaimer: The findings and conclusions in this article are those ofthe authors and do not necessarily represent the official position of theCenters for Disease Control and Prevention.

⁎ Corresponding author. Centers for Disease Control and Prevention,Division of Reproductive Health, Atlanta, GA 30341.

0010-7824/$ – see front matter © 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.contraception.2012.08.005

approved and specially packaged combined oral contracep-tion (COC) regimens: the 91-day extended regimen and the365-day continuous regimen [2,3]. COCs, when taken as anextended or continuous regimen, have been shown to be safeand effective [4–7], as well as acceptable to many women[8–10]. Other available combined hormonal contraceptive(CHC) methods, the contraceptive transdermal patch and thevaginal ring, have also been studied using extended orcontinuous regimens [11–13]. When compared to monthlyregimens, extended CHC use has been shown to decrease thenumber of bleeding days and to improve menstrualsymptoms, including endometriosis-associated recurrentdysmenorrhea [14,15]. Extended or continuous regimensmay also alleviate other menstrual-related symptoms[16,17]. Extended or continuous CHC regimens areattractive to women who wish to reduce or eliminate the

568 E.M. Godfrey et al. / Contraception 87 (2013) 567–575

hormone-free interval (HFI) for convenience, to experiencefewer bleeding days or to help alleviate an existingmenstrual-related disorder.

Despite their growing popularity, unscheduled bleedingand spotting are common with extended or continuous CHCregimens, particularly within the first few months ofinitiation [18]. The fundamental mechanisms underlyingunscheduled bleeding and spotting are not fully understood.Women who are using extended or continuous CHC aremore likely to discontinue their method than women usingthe standard 28-day monthly cycle regimen due tounscheduled bleeding [19]. There is good evidence tosuggest, however, that with continued CHC use, unsched-uled bleeding will decrease [7,20]. Although unscheduledbleeding is common, continued use of the combinedhormonal method often alleviates the problem.

Currently, there are no US-based national guidelines thataddress evidence-based treatments, including preventivetreatments or other management strategies, of unscheduledbleeding for women using extended or continuous CHC.Additionally, recommendations regarding extended orcontinuous CHC use do not exist in the World HealthOrganization (WHO) Selected Practice Recommendationsfor Contraceptive Use (SPR) [21]. The WHO SPR iscurrently being adapted for use in the United States andwill include guidance for women using continuous orextended CHC. Therefore, the objective of this review wasto evaluate the evidence for therapeutic or prophylactictreatments for unscheduled uterine bleeding during contin-uous or extended CHC use.

2. Materials and methods

We searched the PubMed database for all articles (in alllanguages) published in peer-reviewed journals frominception of the database through January 2012 using thesearch strategy in Appendix A. Reference lists from articlesidentified by the search, as well as key review articles, werehand-searched to identify additional articles. Articles thatwere in press in peer-reviewed journals and available onlineahead of publication were also considered.

2.1. Selection of studies

We reviewed titles as well as abstracts to identifyrelevant articles. We included articles that were relevant totreatments (therapeutic and prophylactic) or to manage-ment strategies for irregular uterine bleeding duringextended or continuous CHC use (COCs, patch andring). Although no formal definition exists, for thepurposes of this document, extended contraceptionreferred to the use of CHC with a planned HFI after atleast two contiguous 28-day cycles, and continuouscontraception referred to the uninterrupted use of CHC,without an HFI [22]. We excluded studies that solelydescribed bleeding patterns during continuous or extended

use, and had not included a specific treatment intervention.We examined outcomes related to the amount or durationof unscheduled bleeding, as well as satisfaction andacceptability with the treatment (where provided).

2.2. Assessment of study quality

We summarized and systematically evaluated the evi-dence through the use of standard abstract forms [23]. Thequality of each individual piece of evidence was assessedusing the US Preventative Services Task Force gradingsystem [24].

2.3. Data synthesis

Study results were summarized by treatment type(therapeutic or prophylactic) and by CHC method. Thera-peutic treatment was defined as a treatment for a currentunscheduled bleeding episode, and prophylactic treatmentwas defined as treatment to prevent unscheduled bleeding inwomen who otherwise did not have any irregular bleeding.We did not estimate overall summary measures ofassociation due to the heterogeneity across the four studieswith respect to study design, study population, contraceptivemethod examined and treatment type.

3. Results

The search strategy in PubMed identified 1516 articles.After reviewing the titles and abstracts of these articles, andfull articles when necessary, three articles met our inclusioncriteria. These three studies were randomized controlledtrials (RCTs) of therapeutic treatments for current irregularbleeding or spotting. Two of the therapeutic treatment trialsexamined treatment for unscheduled bleeding for womenusing continuous COCs [25,26] and one for women usingcontinuous contraceptive ring (Table 1) [14]. One addi-tional RCT that addressed a prophylactic treatment amongwomen using continuous COCs was in press ahead ofpublication (Table 2) [27]. References of relevant papersidentified no additional trials. A total of four trials wereincluded in this review. No serious adverse events werereported among the participants in any of the studiesincluded in this review.

3.1. Treatment trials

3.1.1. Combined hormonal oral contraceptive pillTwo trials addressed the treatment of unscheduled uterine

bleeding with continuous COC use. The first trial was anopen-label RCT in which 111 current cyclic COC users of atleast 3 months’ duration were switched to a 168-dayextended regimen [30 mcg ethinyl estradiol (EE)/3 mgdrospirenone] (Table 1) [26]. Women experiencing 7consecutive days or greater of breakthrough bleeding(BTB) or breakthrough spotting (BTS) were randomized toeither an immediate 3-day HFI or a continuation of the

569E.M. Godfrey et al. / Contraception 87 (2013) 567–575

method. Among the 102 subjects who completed the study,63 episodes of BTB/BTS occurred; 33 of these events wererandomized to immediate 3-day HFI, and 30 wererandomized to COC continuation. Estimation of bleedingwas evaluated using self-reported “flow-scores,” rangingbetween 0 and 4, in which 0=no spotting or bleeding and 4=bleeding with an estimate of greater than six pads ortampons saturated per day. The authors presented the patternof daily flow scores over 12 days among women randomizedto an HFI versus COC continuation. On average, those in theHFI group had an initial increase in flow, followed by anabrupt decrease by day 7 with eventual cessation of flow byday 11 so that by end of the observational period, flow scoreswere lower in the HFI group than in the COC continuationgroup (pb.0001). The proportion of women randomized tothe COC continuation group with continued unscheduledbleeding was significantly higher than the proportion of thewomen in the initial HFI group (17/30, 57% versus 1/33,3.0%, respectively) (pb.0001) [26].

Kaneshiro et al. [25] conducted a double-blind RCT thatcompared the use of doxycycline with placebo initiated atthe time of unscheduled bleeding or spotting in extendedCOC users (Table 1). Sixty-six current cyclic CHC users ofthe pill, patch or ring were asked to take four cycles ofextended COC, or 112 days of hormonally active pills [20mcg EE/90 mcg levonorgestrel (LNG)]. The researchersrandomized the participants to either 5 consecutive days ofplacebo or doxycycline 100 mg, to be taken orally twice aday starting the first day of unscheduled bleeding. Theparticipants were instructed to repeat their treatment if theyexperienced another episode of unscheduled bleeding morethan 7 days after completion of their prior treatment. After84 days (or three consecutive 28-day cycles) of hormonallyactive pills and treatment for unscheduled bleeding, theresearchers instructed the participants to continue COC usefor the fourth 28-day cycle, but without any treatment forunscheduled bleeding or spotting. All of the women in bothgroups received their allocated intervention, implying thateach participant experienced unscheduled bleeding at somepoint during the study. The investigators found nosignificant differences during the first 84 days of thestudy between the treatment groups in regards to meannumber of bleeding and/or spotting days [doxycycline: 23.1days (standard error (SE) 3.5), placebo: 19.2 days, SE (3.4);p=.32] or mean number of bleeding-only days [doxycy-cline: 8.7 days, SE (2.2), placebo: 6.5 days, SE (1.6); p=.63]. Additionally, they found no significant differencesduring the fourth cycle, or the last 28 days of the study,between the treatment groups in regards to mean number ofbleeding and/or spotting days [doxycycline: 7.8 days, SE(1.7), placebo: 7.9 days, SE (1.7); p=1.0] or mean numberof bleeding-only days [doxycycline: 3.2 days, SE (.9),placebo: 2.9 days, SE (1.0); p=.57]. Duration of longestbleeding/spotting episode during either the first 84 days[doxycycline: 9.4 days, SE (2.4), placebo: 8.9 days, SE(2.0); p=.70] or the last 28 days of study [doxycycline: 5.3

days, SE (1.6), placebo: 5.7 days, SE (1.7); p=.95] did notdiffer [25].

3.1.2. Combined hormonal transvaginal contraceptivering (ring)

One open-label RCT addressed the treatment ofunscheduled uterine bleeding with continuous contracep-tive ring use (Table 1) [14]. Seventy-four current CHC usersof pill, patch or ring of at least 2 months’ duration wereswitched to 6 months of continuous ring use without anyhormone-free days. At initiation of continuous contracep-tive ring use, participants were randomized to either a 4-dayHFI or ring continuation if they experienced 5 days orgreater of BTB/BTS. Among the 74 subjects, 108 episodesof BTB/BTS occurred; 50 of these events occurred inwomen randomized to immediate 4-day HFI, and 58occurred in women randomized to ring continuation.Subjects recorded their daily bleeding using self-reported“flow-scores,” which ranged between 0 and 4, in which 0=no spotting or bleeding and 4=bleeding with an estimate ofgreater than six pads or tampons saturated per day. Theinvestigators found significantly more women in the HFIgroup reporting a greater number of days of amenorrheacompared to those using the ring continuously without anHFI [95 mean days of amenorrhea (range 92–97) versus 89mean days of amenorrhea (range 85–93), respectively; p=.016]. Differences between groups in average daily self-reported flow scores were not significant (0.17 for HFI and0.21 for continuous; p=.38). The authors also presented thepattern of daily flow scores over 14 days among womenrandomized to an HFI versus ring continuation; those in theHFI group had an initial increase in flow on days 1, 2 and 3following ring removal, followed by lower flow scores byday 8, and by the end of the observational period, flowscores were lower in the HFI group than in the ringcontinuation group (pb.001) [14].

3.2. Prevention trials

3.2.1. Combined hormonal oral contraceptive pillKaneshiro et al. [27] performed a double-blind RCT of a

prophylactic treatment that compared the use of low-doseoral doxycycline with placebo given during the first 84 daysof continuous hormonally active COC use (Table 2). Sixty-five current cyclic CHC users of either the pill, patch or ringand without current unscheduled bleeding were switched to112 days (or four consecutive 28-day cycles) of hormonallyactive pills (20 mcg EE/100 mcg LNG). They were thenrandomized to either controlled-release low-dose doxycy-cline 40 mg or placebo, to be taken orally once a day for thefirst 84 days of continuous COC use. After 84 days, or threeconsecutive 28-day cycles, of continuous COC use and thedaily preventive treatment regimen, subjects were instructedto continue COC use for the fourth 28-day cycle, but withoutany treatment. The investigators found no significantdifferences between the treatment groups in regards tonumber of mean spotting or bleeding days in the first 84 days

Table 1Treatment of unscheduled bleeding with extended or continuous combined hormonal contraception (by method)

Author (year)Sources of support

Study design Population Outcome Results Strengths Weaknesses Quality

Combined oral contraceptionSulak et al. (2006) [26]

Financial supportfrom Berlex

RCT168-day hormonally activeCOC (DRSP/EE)Treatment groups:Immediate HFI × 3 days vs.continuing COC × 7 daysand if still bleeding takeHFI × 3 days

111 current COC users of≥3 months’ duration(at least 2 months wereDRSP/EE prior to startingextended phase)Age: 18–48

(1) Unscheduledbleeding/spotting(2) Average dailyflow scores

63 events of ≥7 days ofunscheduled bleeding/spotting,33 to HFI, 30 to no HFILower average daily flowscores in the HFI groupby end of observationalperiod (pb.0001)HFI:97%=↑ flow→ abrupt ↓ →cessation1 required additional HFI ×3 daysNo HFI:57%=continued BTB/BTS.1 did not follow protocolTreatment failure rates:HFI: 1/33No HFI: 17/30(pb.0001)

Intention-to-treatprinciples followed

(1) No power size calculation(2) Subjective measurementof outcomes(3) Nonstandardized bleedingscale used(4) Unblinded

Level I. Fair.Direct

Kaneshiro et al.(2010) [25]Financial supportfrom Wyeth

RCT112-day hormonally activeCOC (LNG/EE)Treatment: Doxycycline ×5 days vs. placebo.Repeat if experienced anotherbleeding episode N7 days aftercompletion of prior treatment.Treatment during first 84 daysof trial. Observation withouttreatment last 28 days of trial

66 current CHC users(patch, pill or ring) w/o hxof unscheduled bleedingAge: 18–45

(1) Unscheduledbleeding/spottingfor first 84 daysof trial(2) Comparison ofbleeding duringtreatment phasev. phase withouttreatment

60 of 66 women completedthe trialMean # of BTB/BTS days (SE)First 84 days:Doxy: 23.1 (3.5)Placebo: 19.2 (3.4); p=.32Last 28 days:Doxy: 7.8 (1.7)Placebo: 7.9 (1.7); p=1.0Mean # of bleeding (only)days (SE)First 84 days:Doxy: 8.7 (2.2)

(1) Intention-to-treatprinciples followed(2) Appropriateblinding procedures(3) Bleeding diarybased on standardcriteria

(1) Subjective measurementof outcomes(2) Discrepancy in numberof episodes of spotting andbleeding and number oftreatment regimens completed

Level I. Fair.Direct

570E.M

.Godfrey

etal.

/Contraception

87(2013)

567–575

Placebo: 6.5 (1.6); p=.63Last 28 days:Doxy: 3.2 (0.9)Placebo: 2.9 (1. 0); p=.57Duration of longest bleeding/spotting episode (SE)First 84 days:Doxy: 9.2 (2.4)Placebo: 8.9 (2.0); p=.70Last 28 days:Doxy: 5.4 (1.6)Placebo: 5.7 (1. 7); p=.95

Contraceptive ringSulak et al. (2008) [14]Financial supportfrom Organon

RCT168-day hormonally activecontraceptive ringTreatment groups:Immediate 4-day HFI v.no HFI

74 current CHC users(pills, patch or ring) of≥2 months’ duration.Prior to enrollment, patchand pill users completed2 additional baselinecycles, ring userscompleted 1 additionalbaseline cycle.Age: 18–45

(1) Completion rate(2) Proportion offlow-free days(3) Average dailyflow scores

108 episodes ≥5 days ofunscheduled bleeding, 50 toHFI, 58 to no HFI.Completion rates:HFI: 36/37 (97%)No HFI: 29/37 (78%)p=.01% Flow-free days:HFI: 95 days,(range: 92–97 days)No HFI: 89 days(range: 85–93 days)p=.02Ave daily flow scoresHFI: 0.17(range: 0.12–0.23)No HFI: 0.21(range: 0.15–0.27)p=.4

(1) Intervention armsclearly explained(2) Intention-to-treatprinciples followed

(1) Possible enrollment bias(2) Subjective measurementof outcomes(3) Nonstandardized bleedingscale used(4) Unblinded(5) High attrition rate innontreatment group

Level I. Fair.Direct

Abbreviations: Ave=average; BTB=breakthrough bleeding; BTS=breakthrough spotting; CHC=combined hormonal contraception; COC=combined oral contraceptives; Doxy=doxycycline; DRSP=drosperinone;EE=ethinyl estradiol; HFI=hormone-free interval; hx=history; LNG=levonorgestrel; RCT=randomized controlled trial; SE=standard error; v.=versus.

571E.M

.Godfrey

etal.

/Contraception

87(2013)

567–575

Table 2Prevention of unscheduled bleeding with extended or continuous combined hormonal contraception

Author (year)Sources of support

Study design Population Outcome Results Strengths Weaknesses Quality

Combined oral contraceptionKaneshiro et al.(2012) [27]Financial supportfrom NIH

RCT112-day hormonallyactive COC (LNG/EE).Treatment: doxycycline40 mg po/day × 84 daysv. placebo.Treatment during first84 days of trial.Observation withouttreatment last 28 daysof trial.

65 current CHC users(patch, pill or ring)w/o hx of unscheduledbleedingAge: 18–45

(1) Number of unscheduledbleeding/spotting for first84 days of trial(2) Comparison of bleedingduring treatment phase vs.phase without treatment

59 of 65 women completedthe trialMean # BTB/BTS days first84 days (SE):Doxy: 14.8 (2.3)Placebo: 17.8 (2.3); p=.4Mean # bleeding (only) daysfirst 84 days (SE):Doxy :5.0 (1.3)Placebo: 7.2(1.3); p=.2Mean # BTB/BTS days last28 days (SE):Doxy :4.1 (1.2)Placebo: 7.9 (1.4); p=.04Mean # bleeding (only) dayslast 28 days (SE):Doxy :1.0 (0.3)Placebo: 2.6(0.9); p=.1Mean last day of BTB/BTS(SE):Doxy: 61.6 (7.8)Placebo: 85.2 (6.7); p=.03

(1) A priori sample size calculation(2) Intention-to-treat principlesfollowed(3) Appropriate blinding procedures(4) Bleeding diary based onstandard criteria

(1) Subjectivemeasurementof outcomes

Level I. GoodDirect

Abbreviations: BTB=breakthrough bleeding; BTS=breakthrough spotting; CHC=combined hormonal contraception; COC=combined oral contraceptives; Doxy=doxycycline; EE=ethinyl estradiol; hx=history;LNG=levonorgestrel; RCT=randomized controlled trial; SE=standard error; v.=versus.

572E.M

.Godfrey

etal.

/Contraception

87(2013)

567–575

573E.M. Godfrey et al. / Contraception 87 (2013) 567–575

of the study [doxycycline 14.8 days, SE (2.3), placebo 17.8days, SE (2.3); p=.36]. However, the authors noted thatwomen who received low-dose oral doxycycline hadsignificantly earlier onset of amenorrhea (mean last day ofspotting in the doxycycline group occurred at 61.6 daysversus the placebo group at 85.2 days; p=.03). Thepreventive treatment effects persisted during the fourth 28-day cycle, after the 84 days of low-dose daily doxycyclinetreatment had been stopped [27].

4. Discussion

Overall, there is limited evidence to suggest thatintroducing a short HFI when unscheduled bleeding occursduring continuous hormonal contraceptive use may reduceunscheduled bleeding days. Only two open-label random-ized trials of women continually using 6 months of CHC(COCs or contraceptive ring) and experiencing at least 5days of unscheduled bleeding compared a short 3- or 4-dayHFI to those without an HFI [14,26]. With the exception ofreporting flow patterns, results of these two trials werereported with different outcomes, making it difficult tocomment on the overall HFI effectiveness.

Two studies examined the effect of doxycycline related tobleeding with continuous CHC use [25,27]. Doxycycline hadno effect on a current bleeding episode compared to placebo,even though doxycycline, a matrix metalloproteinase inhib-itor, is thought to interrupt unscheduled bleeding [28].Doxycycline may play a role in the prevention of bleedingamong new continuous COC users, which demonstratedearlier amenorrhea and extended treatment effects beyond theactive preventive treatment period [27].

The studies included in this review have severallimitations. In each of the four trials, the main outcome ofthe study used self-reported unscheduled bleeding andspotting. This method of data collection is less problematicfor the two double-blind RCT studies [25,27], but likelycontributes to treatment bias for the two open-label studies inwhich neither the subjects nor investigators were blinded[14,26]. In addition, the two open-label studies examiningthe HFI excluded subjects who missed pills or werenoncompliant with the method, and used a nonstandardizedscale to record bleeding and spotting, making it difficult tocompare results with other studies [14,26]. All of thesestudies enrolled current users of cyclic CHC prior toinitiating them on a continuous COCs or ring regimen,which may have underestimated the amount of unscheduledbleeding and spotting that might have occurred if theparticipants had been new to CHC altogether or had switchedfrom a progestin-only method.

Although limited, there are national guidelines addressingunscheduled bleeding among women using CHC forclinicians practicing in the United Kingdom and Canada.The Royal College of Obstetricians and Gynaecologists(RCOG) offers clinical guidance for unscheduled bleeding

among women using CHC, although not specificallyextended or continuous regimens. Providers are encouragedto assess other reasons for unscheduled bleeding, if clinicallyappropriate [29]. If bleeding occurs within the first 3 monthsof starting the method, RCOG suggests reassurance withfollow-up. Medical management can be provided ifrequested and includes advice such as increasing theestrogen component in the COC to 35 mcg. Othersuggestions are made, but with the caveat that no evidenceexists, such as trying a different COC or changing theprogestin dose or type [29]. The Society of Obstetricians andGynaecologists of Canada also has national guidelines tohelp clinicians address unscheduled bleeding during CHCuse and specifically addresses extended or continuous use[22]. Similar to the guidance from RCOG, the Canadianrecommendations reiterate the notion that unscheduledbleeding is extremely common the first 3 months of useand recommend including a clinical history to exclude thepossibility of inconsistent use, malabsorption, STIs andpregnancy. The guidelines state that if bleeding or spottingpersists without an underlying medical disorder, cliniciansmay prescribe a short HFI after a minimum of 21 days ofCHC use [22].

In spite of national guidelines, few treatment optionsexist for women experiencing unscheduled bleeding whileusing continuous or extended CHC. In studies examiningcyclic or extended/continuous CHC use, evidence suggeststhat the number of bleeding or spotting days often decreasesafter several months of use [7,20]. Thus, those experiencingunscheduled bleeding should be encouraged to continuetheir regimen for at least several cycles. Some evidencesuggests that COC formulation may influence unscheduledbleeding, although neither estrogen nor progestin dosecontained within the COC appears to affect unscheduledbleeding [30]; progestin type, however, might play a role[31]. In an RCT of continuous COC users, 139 women wererandomized to one of four types of hormonally active COCpills for 180 days. The formulations in this study consistedof two different doses of estrogen (20 or 30 mcg of EE) andtwo separate types of progestin [LNG or norethindroneacetate (NETA)]. Regardless of the EE dosage, theinvestigators found that fewer days of unscheduled bleedingoccurred in the women using the COC pills that containedNETA [31]. Another study characterizing bleeding patternsfor extended (91 days) COC use showed that a particularpill formulation (30 mcg EE and 150 mcg LNG) thatincluded 10 mg daily EE during the hormone-free week (7days) had less unscheduled bleeding in the second and third91-day cycles compared to women using the same pillformulation, but with a placebo-only hormone-free week.By the fourth 91-day cycle, however, similarities inunscheduled bleeding were reported by women using eitherthe placebo-only hormone-free week or the 10-mg daily EEhormone-free week [32]. Only limited evidence suggeststhat progestin type may contribute to less unscheduledbleeding in continuous CHC users or that adding low-dose

574 E.M. Godfrey et al. / Contraception 87 (2013) 567–575

estrogen during the hormone-free week may be beneficial inextended CHC users. Despite implications that one pillformulation may be superior over another, evidence islacking to suggest that changing COC formulation forextended or continuous use results in an improvement ofunscheduled bleeding.

In conclusion, limited evidence suggests that there aretreatments to help alleviate or prevent unscheduled bleedingand spotting among extended or continuous CHC users.Although there is “Level I, fair” evidence from two RCTs tosuggest that an immediate 3- or 4-day HFI demonstrateslower flow scores after the introduction of the HFI, furtherstudies with larger samples using standardized measuresare needed. In terms of preventing unscheduled bleedingamong new COC users, one study demonstrated thatamenorrhea may occur earlier with low-dose daily doxycy-cline [27]. Although there is “Level I, good” evidence thatsuch prophylactic modalities may be beneficial, the impactof this treatment on continued use of the CHC regimenremains unclear.

Appendix A. Search strategy

(“contraceptives, oral"[MeSH Terms] OR (“contracepti-ves"[All Fields] AND “oral"[All Fields]) OR “oral contra-ceptives"[All Fields] OR (“oral"[All Fields] AND“contraceptives"[All Fields]) OR “contraceptives, oral"[Phar-macological Action]) OR ((“contraceptive devices"[MeSHTerms] OR “contraceptive agents"[MeSH Terms]) AND(“transdermal patch"[MeSH Terms] OR (“transdermal"[AllFields] AND “patch"[All Fields]) OR “transdermal patch"[AllFields] OR “patch"[All Fields] OR “Ortho Evra"[Supplemen-tary Concept] OR “Ortho Evra"[All Fields] OR “orthoevra"[All Fields] OR “Ortho Evra"[Supplementary Concept]OR “Ortho Evra"[All Fields] OR “ortho evra"[All Fields]) OR((“contraceptive devices"[MeSH Terms] OR “contraceptiveagents"[MeSH Terms]) AND ((ring[All Fields] OR“NuvaRing"[Supplementary Concept] OR “NuvaRing"[AllFields] OR “nuvaring"[All Fields]) OR (Nuva[All Fields]AND Ring[All Fields])) AND ((extended-cycle[All Fields]OR (extended[All Fields] AND “regimen"[All Fields]) ORcontinuous-regimen[All Fields]) OR extended[All Fields])AND (“utilization"[Subheading] OR “utilization"[All Fields]OR “use"[All Fields]))) AND ((((“therapy"[Subheading] OR“therapy"[All Fields] OR “treatment"[All Fields] OR“therapeutics"[MeSH Terms] OR “therapeutics"[All Fields])OR manage[All Fields] OR (“prevention and control"[Sub-heading] OR (“prevention"[All Fields] AND “control"[AllFields]) OR “prevention and control"[All Fields])) OR((“Therapeutics"[Mesh] OR “therapy"[Subheading]) OR“Anti-Inflammatory Agents, Non-Steroidal"[Mesh]) AND“humans"[MeSH Terms]) AND ((“unscheduled bleeding"OR “Metrorrhagia"[Mesh]) OR ((“menorrhagia"[MeSHTerms] OR “menorrhagia"[All Fields]) OR (“metrorrha-gia"[MeSH Terms] OR “metrorrhagia"[All Fields]) OR

(“metrorrhagia"[All Fields] OR “spotting"[All Fields]) OR“abnormal menstruation"[All Fields] OR “menstrual abnor-malities"[All Fields]) OR “bleeding"[All Fields] OR(bleeding pattern[All Fields] OR bleeding patterns[AllFields]))) Limits: Humans.

References

[1] Andrist LC, Arias RD, Nucatola D, et al. Women's and providers'attitudes toward menstrual suppression with extended use of oralcontraceptives. Contraception 2004;70:359–63.

[2] US Food and Drug Administration. Label and approval history:seasonale NDA 021544. Available at http://www.accessdata.fda.gov/Scripts/cder/DrugsatFDA/index.cfm?fuseaction=Search.DrugDetails2003 Last accessed March 5, 2012.

[3] US Food and Drug Administration. Lybrel (90 mcg levonorgestrel and20 mcg ethinyl estradiol) tablets, prescribing information and patientpackage insert 2008. Available at http://www.accessdata.fda.gov/drugsatfda_docs/label/2008/021864s002lbl.pdf Last accessed March5, 2012.

[4] Anderson FD, Gibbons W, Portman D. Safety and efficacy of anextended-regimen oral contraceptive utilizing continuous low-doseethinyl estradiol. Contraception 2006;73:229–34.

[5] Archer DF, Jensen JT, Johnson JV, et al. Evaluation of a continuousregimen of levonorgestrel/ethinyl estradiol: phase 3 study results.Contraception 2006;74:439–45.

[6] Teichmann A, Apter D, Emerich J, et al. Continuous, dailylevonorgestrel/ethinyl estradiol vs. 21-day, cyclic levonorgestrel/ethinyl estradiol: efficacy, safety and bleeding in a randomized,open-label trial. Contraception 2009;80:504–11.

[7] Wiegratz I, Stahlberg S, Manthey T, et al. Effect of extended-cycleregimen with an oral contraceptive containing 30 mcg ethinylestradioland 2 mg dienogest on bleeding patterns, safety, acceptance andcontraceptive efficacy. Contraception 2011;84:133–43.

[8] Coffee AL, Sulak PJ, Kuehl TJ. Long-term assessment of symptom-atology and satisfaction of an extended oral contraceptive regimen.Contraception 2007;75:444–9.

[9] Edelman AB, Gallo MF, Jensen JT, et al. Continuous or extended cyclevs. cyclic use of combined oral contraceptives for contraception.Cochrane Database Syst Rev 2005:CD004695.

[10] Foidart JM, Sulak PJ, Schellschmidt I, Zimmermann D. YasminExtended Regimen Study Group. The use of an oral contraceptivecontaining ethinylestradiol and drospirenone in an extended regimenover 126 days. Contraception 2006;73:34–40.

[11] Guazzelli CA, Barreiros FA, Barbosa R, de Araujo FF, Moron AF.Extended regimens of the vaginal contraceptive ring: cycle control.Contraception 2009;80:430–5.

[12] Miller L, Verhoeven CH, Hout J. Extended regimens of thecontraceptive vaginal ring: a randomized trial. Obstet Gynecol 2005;106:473–82.

[13] Stewart FH, Kaunitz AM, Laguardia KD, et al. Extended use oftransdermal norelgestromin/ethinyl estradiol: a randomized trial.Obstet Gynecol 2005;105:1389–96.

[14] Sulak PJ, Smith V, Coffee A, et al. Frequency and management ofbreakthrough bleeding with continuous use of the transvaginalcontraceptive ring: a randomized controlled trial. Obstet Gynecol2008;112:563–71.

[15] Vercellini P, Frontino G, De Giorgi O, et al. Continuous use of an oralcontraceptive for endometriosis-associated recurrent dysmenorrheathat does not respond to a cyclic pill regimen. Fertil Steril 2003;80:560–3.

[16] LaGuardia KD, Fisher AC, Bainbridge JD, LoCoco JM, Friedman AJ.Suppression of estrogen-withdrawal headache with extended transder-mal contraception. Fertil Steril 2005;83:1875–7.

575E.M. Godfrey et al. / Contraception 87 (2013) 567–575

[17] Sulak P, Willis S, Kuehl T, Coffee A, Clark J. Headaches and oralcontraceptives: impact of eliminating the standard 7-day placebointerval. Headache 2007;47:27–37.

[18] Bachmann G, Korner P. Bleeding patterns associated with oralcontraceptive use: a review of the literature. Contraception 2007;76:182–9.

[19] Hitchcock CL, Prior JC. Evidence about extending the duration of oralcontraceptive use to suppress menstruation. Womens Health Issues2004;14:201–11.

[20] Miller L, Hughes JP. Continuous combination oral contraceptive pillsto eliminate withdrawal bleeding: a randomized trial. Obstet Gynecol2003;101:653–61.

[21] World Health Organization. Selected practice recommendations forcontraception use2nd edition. . Geneva: WHO; 2004.

[22] Guilbert E, Boroditsky R, Black A, et al. Canadian consensus guidelineon continuous and extended hormonal contraception, 2007. J ObstetGynaecol Can 2007;29:S1–32.

[23] Mohllajee AP, Curtis KM, Flanagan RG, et al. Keeping up withevidence: a new system for WHO's evidence-based family planningguidance. Am J Prev Med 2005;28:483–90.

[24] Harris RP, Helfand M, Woolf SH, et al. Current methods of the USPreventive Services Task Force: a review of the process. Am J PrevMed 2001;20:21–35.

[25] Kaneshiro B, Edelman A, Carlson N, et al. Treatment of unscheduledbleeding in continuous oral contraceptive users with doxycycline: arandomized controlled trial. Obstet Gynecol 2010;115:1141–9.

[26] Sulak PJ, Kuehl TJ, Coffee A, Willis S. Prospective analysis ofoccurrence and management of breakthrough bleeding during anextended oral contraceptive regimen. Am J Obstet Gynecol 2006;195:935–41.

[27] Kaneshiro B, Edelman A, Carlson NE, et al. A randomized controlledtrial of subantimicrobial-dose doxycycline to prevent unscheduledbleeding with continuous oral contraceptive pill use. Contraception2012;85:351–8.

[28] Li R, Luo X, Pan Q, et al. Doxycycline alters the expression ofinflammatory and immune-related cytokines and chemokines in humanendometrial cells: implication in irregular uterine bleeding. HumReprod 2006;21:2555–63.

[29] Faculty of Sexual & Reproductive Healthcare Clinical Guidance.Management of unscheduled bleeding in women using hormonalcontraception. Available at http://www.fsrh.org/pdfs/UnscheduledBleedingMay09.pdf2009 Last accessed July 30, 2012.

[30] Kaneshiro B, Edelman A, Carlson NE, Nichols M, Jensen J.Unscheduled bleeding with continuous oral contraceptive pills: acomparison of progestin dose. Contraception 2012;86:22–7.

[31] Edelman AB, Koontz SL, Nichols MD, Jensen JT. Continuous oralcontraceptives: are bleeding patterns dependent on the hormonesgiven? Obstet Gynecol 2006;107:657–65.

[32] Kaunitz AM, Portman DJ, Hait H, Reape KZ. Adding low-doseestrogen to the hormone-free interval: impact on bleeding patterns inusers of a 91-day extended regimen oral contraceptive. Contraception2009;79:350–5.