REVIEW

Drug Interactions Between Oral Contraceptives andAntibiotics

Barry D. Dickinson, PhD, Roy D. Altman, MD, Nancy H. Nielsen, MD, PhD, andMelvyn L. Sterling, MD, for the Council on Scientific Affairs, American Medical Association

OBJECTIVE: To evaluate the evidence on possible drug in-teractions between antibiotics and oral contraceptives(OCs) that may lead to OC failure.

DATA SOURCES: MEDLINE and Lexis/Nexis Medical Li-brary searches for 1966–1999 using the key word “oralcontraceptives,” cross-indexed with the terms “antibiotics,”“adverse effects,” and “pregnancy,” and MEDLINE searchusing the additional MeSH term “drug interactions.” Nolanguage restrictions were used.

METHODS OF STUDY SELECTION: A total of 167 articles wereretrieved for analysis. Another 32 articles were identifiedby review of the references cited in these publications.Articles were selected based on their ability to provideinformation on the relationship between antibiotic therapyand OC efficacy in otherwise compliant users (defined aswomen with unplanned pregnancies who reported compli-ance with their OC regimen). Additionally, studies thateither directly measured the effects of antibiotics on thepharmacokinetics of OC components, or that analyzed theeffects of antibiotics on measures of ovulation in OC userswere accepted.

TABULATION, INTEGRATION, AND RESULTS: At least 30 caseshave been reported of pregnancies occurring in womentaking OCs and antibiotics, particularly rifampin. Approx-imately 20% of pregnant women reporting to family plan-ning or abortion clinics reported concomitant OC andantibiotic use. Information from adverse event reportingdatabases generally mirrors the types of informationgleaned from these case reports and clinical surveys andaccounts for approximately one-third of reported cases.Retrospective surveys, primarily from dermatology-based

practices, also have reported 24 pregnancies in OC userswho concomitantly received therapy with antibiotics, mostcommonly tetracyclines and penicillins. Apparent OC fail-ure rates in clinical surveys were within the usual rangeexpected for patterns of typical use. In pooled results ob-tained from relatively small populations, oral antibiotics,with the exception of rifampin, have not significantly af-fected the pharmacokinetics of ethinyl estradiol, levonor-gestrel, and norethindrone or reduced the serum concen-trations of gonadotropins. However, individual patientshave been identified who experienced significant decreasesin the plasma concentration of these components of OCsand who appeared to ovulate.

CONCLUSION: Rifampin impairs the effectiveness of OCs.Pharmacokinetic studies of other antibiotics have notshown any systematic interaction between antibiotics andOC steroids. However, individual patients do show largedecreases in the plasma concentrations of ethinyl estradiolwhen they take certain other antibiotics, notably tetracy-cline and penicillin derivatives. Because it is not possible toidentify these women in advance, a cautious approach isadvised. (Obstet Gynecol 2001;98:853–60. © 2001 bythe American College of Obstetricians and Gynecologists.)

An estimated 10 million women in the United States useoral contraceptives (OCs) annually, including estrogen–progestin combinations or progestin-only products.1

Combination products (monophasic, biphasic, or tripha-sic) pills provide contraception primarily by suppressingfollicle-stimulating hormone (FSH) and luteinizing hor-mone (LH) release, thereby inhibiting follicular growthand development and ovulation. Secondary effects oncervical mucus and endometrial histology that reducethe likelihood of fertilization or implantation also maycontribute. Progestin-only products (“minipills”) inhibitovulation in less than 50% of users, providing contracep-tion mainly by thickening cervical mucus and maintain-ing an endometrial environment that is unsuitable forimplantation and nurturing of a fertilized ovum.

The estrogenic component in virtually all combina-tion products is ethinyl estradiol; a few products containmestranol, which serves as a pro-drug for ethinyl estra-

From the Council on Scientific Affairs, American Medical Association, Chicago,Illinois.

Members and staff of the Council on Scientific Affairs at the time this report wasprepared were: Myron Genel, MD (Chair), New Haven, CT; Michael A.Williams, MD (Chair-Elect), Baltimore, MD; Roy D. Altman, MD, Miami, FL;Scott D. Deitchman, MD, MPH, Duluth, GA; J. Chris Hawk III, MD,Charleston, SC; John P. Howe III, MD, San Antonio, TX; Hillary D. Johnson,St. Louis, MO; Nancy H. Nielsen, MD, PhD, Buffalo, NY; John F. Schneider,MD, PhD, Chicago, IL; Melvyn L. Sterling, MD, FACP, Orange, CA; ZoltanTrizna, MD, PhD, Galveston, TX; and Donald C. Young, MD, Iowa City, IA.Staff: Barry D. Dickinson, PhD (Secretary); James M. Lyznicki, MS, MPH(Assistant Secretary); and Marsha Meyer (Editor), Chicago, IL.

853VOL. 98, NO. 5, PART 1, NOVEMBER 2001 0029-7844/01/$20.00© 2001 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc. PII S0029-7844(01)01532-0

diol. Older progestins used in combination pills arenorethindrone, norethindrone acetate, and ethynodioldiacetate (estrane derivatives), and the second-genera-tion agents norgestrel and levonorgestrel. Newer (third-generation) progestins used in the United States includedesogestrel and norgestimate. These agents may possessless potent androgenic activity.

Standard drug information and drug interaction data-bases continue to issue warnings about the possibility ofOC failure in women who are prescribed rifampin,griseofulvin, penicillins, and tetracyclines.2–4 Addition-ally, a recent review intended for primary care physi-cians identifies amoxicillin, ampicillin, metronidazole,and tetracycline as drugs that may potentially decreasethe effectiveness of OCs, and advises that it may “beprudent for women to use a back-up contraceptivemethod during antibiotic therapy and for 7 days aftercompleting the antibiotic course or having the last epi-sode of vomiting and diarrhea.”5

Given the continued concern about this issue amongphysicians and the problems posed by unintended preg-nancies, the American Medical Association’s Council onScientific Affairs systematically reviewed the literatureconcerning women of childbearing age who becamepregnant after receiving antibiotics while practicing oralcontraception to formulate recommendations for clinicalpractice based on the published scientific evidence.

Because there is no evidence that antibiotics affect thepharmacodynamics of estrogens and progestins, it isgenerally believed that any interaction between antibiot-ics and OCs must be grounded in pharmacokineticeffects on ethinyl estradiol and/or progestins. Therefore,the Council also examined studies that either measureddirectly the effects of antibiotics on the pharmacokineticsof OC components, or analyzed the effects of antibioticson measures of ovulation in OC users.

SOURCES

Published studies from 1966 through December 1999were identified through MEDLINE and Lexis/NexisMedical Library searches (without language restriction)using the term “oral contraceptives,” cross-indexed with“antibiotics,” “adverse effects,” and “pregnancy.” Re-lated articles in the MEDLINE database using the addi-tional MeSH term “drug interactions” also were identi-fied. A total of 167 articles were retrieved for analysis,and another 32 articles were identified by review of thereferences cited in these publications. Articles were se-lected based on provision of information on the relation-ship between antibiotic therapy and OC efficacy in oth-erwise compliant users (defined as women withunplanned pregnancies who reported compliance with

their OC regimen). Case reports, surveys, and informa-tion from adverse reaction databases, as well as studiesthat either directly measured the effects of antibiotics onthe pharmacokinetics of OC components or analyzedthe effects of antibiotics on measures of ovulation in OCusers were reviewed.

RESULTS

Clinical Reports of Antibiotics Apparently Causing OCFailure

The first report of potential interactions between anti-biotics and OCs appeared in 1971 when Reimers andJezek6 reported an increased incidence of intermen-strual breakthrough bleeding in 38 of 51 womentreated concomitantly with OCs and rifampin. Shortlythereafter, rifampin was implicated in five unplannedpregnancies among 88 women with tuberculosis whowere taking both rifampin and OCs. Another 62women suffered from menstrual irregularities.7 Overthe next decade, another 13 reports documentedwomen with tuberculosis who became pregnant onOCs with concomitant use of rifampin, often in asso-ciation with other antitubercular agents, including onewho experienced a second pregnancy under the sameconditions.8 –12 After the reports on rifampin ap-peared, possible linkages between the use of otherantibiotics and failure of OCs were reported in 12additional women; the antibiotics included chloram-phenicol, sulfamethoxypyradizine, ampicillin, oxacil-lin, penicillin G, tetracycline, oxytetracycline, minocy-cline, and other “unspecified.”10,13–16 Thus, the totalnumber of reported pregnancies due to OC failureassociated with antibiotic treatment is 30.

Case reports were also published linking the use ofgriseofulvin to the occurrence of oligomenorrhea andirregular menses in a 32-year-old woman taking a low-dose, triphasic OC and of unplanned pregnancy in twoother women.17,18 Increasing the estrogen dose to 50 mgrestored regular menses. Additionally, 22 case reports ofa possible interaction between griseofulvin and OCswere compiled in spontaneous adverse reaction reports(see below).19

Surveys From Office-Based Practice

Four retrospective surveys have been conducted on pa-tients from office-based practices or outpatient clinicsevaluating possible interactions between antibiotics andOCs (Table 1).20–23 These surveys found OC failurerates of 1.2–1.6% (ie, 1.2–1.6 pregnancies per 100 yearsof women exposure) in women on OCs who also weretreated with antibiotics. These failure rates are largerthan the ideal failure rates predicted with perfect compli-

854 Dickinson et al OC and Antibiotic Interactions OBSTETRICS & GYNECOLOGY

ance, but well within the 1–3% (or larger) range encoun-tered with patterns of typical use. These studies sufferfrom reliance on patient recall and failure to consider thefrequency of sexual intercourse during the period of

possible drug interaction. Also, three studies emanatedfrom dermatology practices, in which the most commonantibiotic used was erythromycin. This drug is not asso-ciated with OC failure, and macrolides may actually

Table 1. Surveys of Apparent Oral Contraceptive Failures in Women Taking Oral Antibiotics

Survey Design Subjects Results Conclusion/Comment

Helms et al20 Retrospective cohortstudy of 3dermatologypractices using chartreview and surveysdependent on patientrecall

356 patients with ahistory ofcombined OC/antibiotic use;

425 controls onOCs alone,including 263from the firstgroups

5 pregnancies in 311woman-years in OC/antibiotics users (1.6%failure rate)

Combined antibiotic/OCis associated with anonsignificant (P 50.4) increase in OCfailure rates. This valueis within the expectedrange associated withtypical patterns of use

12 pregnancies in 1245woman-years for OCsalone (0.96% failurerate)

Antibiotics wereminocycline and a“cephalosporin,” whichhad been taken for atleast 3 months whenpregnancy occurred.No individual analysisby type of OC orantibiotic used

London et al21 Retrospective survey inan outpatientdermatology clinicdependent on patientrecall

34 patients werefound to haveused oralantibiotics andOCs together.

1 pregnancy in 71 woman-years (1.4% failure rate)

Combined oralantibiotics/OCsassociated withnonsignificant increasein OC failure rate.

The most commonantibiotics wereerythromycinand tetracycline

Preponderance oferythromycinrecipients, an antibioticnot noted for thispotential interaction

DeSano et al22 Retrospective survey ofprivate obstetricspractice, dependenton patient recall

16 apparentlycompliantpatients whobecamepregnant whiletaking oralcontraceptives

13 of 16 patients recalledtaking antibioticsduring the cycle inwhich conception tookplace, includingpenicillin (3), ampicillin(5), sulfa antibiotics (3),tetracycline (1), andcephalexin (1) (failurerate N/A)

Antibiotics may causeOC failure in otherwisecompliant patients;however, failure rate iswithin the rangeassociated with typicalpatterns of use

Hughes et al23 Retrospective survey ofprivate dermatologypractice dependenton patient recall

124 patients whohad takenantibiotics foracne and OCsimultaneously

2 pregnancies in 163woman-years (1.2%failure rate); antibioticswere minocycline andoxytetracycline

Antibiotics might causeOC failure in otherwisecompliant patients;however, failure rate iswithin the rangeassociated with typicalpatterns of use

Most common antibioticused waserythromycin, anantibiotic not noted forthis potentialinteraction

OC 5 oral contraceptive.

855VOL. 98, NO. 5, PART 1, NOVEMBER 2001 Dickinson et al OC and Antibiotic Interactions

increase the plasma concentrations of ethinyl estradioland certain progestins.

Interaction Information From Adverse Event ReportingDatabases

The Committee on the Safety of Medicines in the UnitedKingdom and the Netherlands Centre for Monitoring ofAdverse Reactions to Drugs received reports on twowomen who became pregnant on OCs while receivinggriseofulvin and sulfonamide antibiotics.19 Twentyother OC users reported experiencing transient inter-menstrual bleeding or amenorrhea in the first or secondcycle after beginning griseofulvin therapy.19

In an analysis covering the years 1968–1984, a total of 63unplanned pregnancies were reported to the Committee onSafety of Medicines in women taking OCs and concomi-tant antibiotics.24 Penicillins and tetracyclines were in-volved in 70% of the reports. Other antibiotics mentionedwere sulfamethoxazole plus trimethoprim, sulfonamides ortrimethoprim, metronidazole, cephalosporins, erythromy-cin, and fusidic acid. In 51 reports that specified the type ofOC, the largest number of apparent reported failures (n 533) occurred in women using a low-dose monophasic prod-uct containing 30 mg of ethinyl estradiol. The frequencywith which a particular antibiotic was associated with OCfailure roughly mirrored the volume of prescriptions writ-ten for that time period, suggesting that the association mayhave been by chance.

Pregnant Women Presenting for Family PlanningServices

One study over a 4-year period from November 1981 toDecember 1985 documented 163 cases of OC methodfailures in reliable OC users who presented for abortionservices.25 In more than one-third of these cases (36%),there were no predisposing factors. Significant factorsassociated with failure were diarrhea and/or vomiting in35%, and breakthrough bleeding on the combined OC in21%; 23% of the failures were associated with antibioticuse, most commonly amoxicillin. These results are sim-ilar to another study of pregnant women presenting tofamily planning or abortion clinics that also discoveredapproximately 20% of these women (23 of 113) reportedthat conceptions occurred during concurrent use of com-bination OCs and antibiotics.26

Although the strength of association varies in theabove reports of contraceptive failure in patients treatedconcomitantly with various antibiotics, more than 200occurrences have been noted in anecdotal case reports,spontaneous adverse-event monitoring systems, and ret-rospective surveys.

Possible Mechanisms for OC Failure Caused byAntibiotics

Either pharmacodynamic or pharmacokinetic interac-tions could theoretically inhibit the efficacy of OCs.There is no evidence that any antibiotic directly affectssteroid receptor function, or that it serves as a physio-logic antagonist of estrogens or progestins. It is com-monly believed that drug interactions with OCs (notinvolving steroid receptors) have a pharmacokinetic ba-sis. Therefore, the pharmacokinetics of the commonestrogen and progestin components of combined OCsare briefly reviewed. The values reported in the follow-ing sections were obtained from original studies or pre-vious reviews on the subject. If such information waslacking, data was obtained from the manufacturer’s offi-cial product labeling.

Pharmacokinetics of OC Components

Ethinyl estradiol, the major estrogenic component ofcombined OCs, is well absorbed after oral administra-tion but is subject to a first-pass effect in the gut mucosa(sulfate conjugation) and liver. Oral bioavailability aver-ages 38–48%, but great interindividual variability (20–65%) has been reported.27 Ethinyl estradiol is 97%bound to serum albumin. Although it is not bound to sexhormone–binding globulin (SHBG), ethinyl estradiolinduces SHBG synthesis.

The major phase I metabolic reaction undergone byethinyl estradiol is aromatic 2-hydroxylation, which islargely accomplished by the cytochrome P4503A4. Thismetabolite is further methylated and conjugated withglucuronic acid before urinary and fecal excretion. Theparent compound also is subject to direct glucuronida-tion (and sulfation) at position 3, and possibly position17 as well. Conjugates of ethinyl estradiol are secreted inthe bile. These conjugates may be hydrolyzed by gutbacteria to liberate the parent compound, which can thenbe reabsorbed (enterohepatic circulation).28 Treatmentwith either ampicillin 250 mg four times daily or tetracy-cline 250 mg four times daily decreases urinary excre-tion, and increases the fecal excretion of ethinyl estradiol,while reducing fecal bacterial estradiol-b-glucuronidaseactivity.

At steady state, plasma concentrations of estradiolsulfate are higher than that of the parent compound.Other minor hydroxylated metabolites of estradiol arealso formed. The elimination half-life of ethinyl estradiolin most studies varies between 13 and 27 hours at steadystate.29 Based on in vitro analysis, all progestins in com-mon use tend to inhibit the hepatic metabolism of ethinylestradiol to about the same degree.30 Results of somestudies indicate that the hepatic clearance of ethinyl

856 Dickinson et al OC and Antibiotic Interactions OBSTETRICS & GYNECOLOGY

estradiol may be more markedly inhibited by progestinderivatives such as desogestrel.31 Nevertheless, thesteady state concentrations of ethinyl estradiol demon-strate considerable interpatient variability, varying eight-to tenfold.32–34

Levonorgestrel is rapidly and nearly completely ab-sorbed after oral administration with a bioavailability ap-proaching 100%. Levonorgestrel undergoes reduction andhydroxylation in the liver, followed by conjugation withsulfate and glucuronic acid. The parent compound alsomay be subject to sulfation. Levonorgestrel is primarilybound in serum to SHBG. The elimination kinetics oflevonorgestrel are not dose dependent in users of combina-tion OCs because ethinyl estradiol induces the synthesis ofSHBG. The elimination half-life of levonorgestrel is vari-able, averaging 36 6 13 hours at steady state.32–34

Norgestimate is well absorbed after oral administrationand is converted in part to an active, deacetylated derivativethat has a half-life of 12–30 hours. Various other hydroxy-lated metabolites and conjugates also are formed.

Norethindrone (acetate) is well absorbed, with a meanbioavailability of 65% due to presystemic clearance. Itchiefly undergoes reduction of ring A, followed by con-jugation with glucuronide or sulfate. Some hydroxylatedmetabolites also are formed that can undergo conjuga-tion. Like other progestins, the plasma clearance andhalf-life (5–14 hours) of norethindrone exhibit consider-able interindividual variability. Norethindrone acetate iscompletely and rapidly deacetylated to norethindroneafter oral administration.32 Ethynodiol acetate also isquantitatively metabolized to norethindrone.32

Desogestrel is rapidly and well absorbed, with an oralbioavailability of about 84%. The parent compound isconverted to an active metabolite, 3-keto-desogestrel,which is 99% bound to SHBG and has a half-life of28–38 hours. Other minor inactive hydroxylated metab-olites also are formed, which may be converted to sulfateand glucuronide conjugates. As with other progestins,there is a wide variability in individual clearances, andelimination may be nonlinear due to induction of SHBGby ethinyl estradiol.

Enterohepatic circulation is important in the elimina-tion kinetics of ethinyl estradiol. With the possible ex-ception of levonorgestrel, other progestins are not sub-ject to direct conjugation. Rather, they are extensivelyreduced to inactive compounds before conjugation. Ethi-nyl estradiol and all progestins used in OCs are subject tolarge interindividual variations in drug disposition.

Pharmacokinetic Interaction Studies of Antibiotics andOCs

It is well established that drugs that increase the rate ofhepatic drug metabolism (phenobarbital, carbamaz-

epine, phenytoin) can increase the failure rate of OCs.32

Thus, concomitant use of rifampin, which is a potentinducer of cytochrome P4503A4, is associated with anincrease in OC failure rate. Rifampin also inducesSHBG, which may reduce free plasma concentrations ofprogestins. Antibiotics that do not induce cytochromeP450 may reduce the plasma levels of steroids based onindirect interference with the enterohepatic circulation ofethinyl estradiol. With the possible exception oflevonorgestrel, other progestins commonly used in OCsare not subject to direct conjugation with glucuronic acidand enterohepatic circulation.

A number of studies have examined the effect of oralantibiotics on the plasma concentrations of ethinyl estra-diol and the progestin component, as well as FSH andLH.35–49 Some studies also evaluated ovulation, eitherwith ovarian sonography or by measuring progesteroneconcentrations. One study indicated that at least onebroad-spectrum antibiotic (sulfamethoxazole and tri-methoprim) appears to actually increase the plasma con-centration of ethinyl estradiol, while leaving the concen-tration of levonorgestrel unchanged.50

In vitro analysis of the oxidation of ethinyl estradiol inhepatic microsomes prepared from liver biopsies ob-tained from women (n 5 4) treated with rifampin sug-gests that such treatment increases the initial hydroxyla-tion rate of ethinyl estradiol about fourfold.51 The effectof rifampin on the disposition of OCs and the menstrualcycle have been examined both in women with tubercu-losis and in healthy volunteers. The majority of thesewere open studies.35–40,52 Rifampin (450–600 mg dailyor 8–10 mg/kg) significantly accelerates the eliminationof ethinyl estradiol and various progestin componentsthat are contained in combination preparations. Whenmeasured, approximately 30–50% of women appearedto ovulate. Thus, rifampin clearly impairs the efficacy ofmonophasic and triphasic preparations containing 30–50 mg of estradiol.

A number of pharmacokinetic studies have been con-ducted with other antibiotics in small groups of women.Short-term administration of ampicillin (up to 1.5 g dailyin divided doses) did not significantly affect serum con-centrations of gonadotropins or OC steroids. Individualpatients, however, exhibited breakthrough bleeding orelevated FSH concentrations.41–43 Only 3 of 25 womenreceiving low-dose monophasic preparations appearedto ovulate when they were treated briefly with metroni-dazole (400 mg three times daily for 6–8 days).42

In open studies involving volunteers who served astheir own controls, neither tetracycline nor doxycyclinesignificantly affected the disposition of monophasic prep-arations containing 35 mg of ethinyl estradiol and nor-ethindrone.44,45 Large interindividual variation oc-

857VOL. 98, NO. 5, PART 1, NOVEMBER 2001 Dickinson et al OC and Antibiotic Interactions

curred but was obscured in the analysis of pooled results.In studies that examined the effects of fluoroquinoloneson low-dose monophasic preparations containing 30 mgof ethinyl estradiol and levonorgestrel, plasma concen-trations of LH, FSH, and the latter were largely unaf-fected, although one woman experienced a large de-crease in ethinyl estradiol.47–49 Among the macrolides,roxithromycin does not affect ovulation,39 and dirithro-mycin decreases the clearance of ethinyl estradiol ap-proximately 10%, an effect that may be exacerbatedthree- to fourfold in individuals with low baseline val-ues.46 No interaction studies have been conducted usingproducts containing the newer progestins.

Except for rifampin, antibiotics do not significantlyaffect the plasma concentrations of ethinyl estradiol (ornorethindrone and levonorgestrel) based on pooleddata; however, large interindividual variation in plasmaconcentrations of ethinyl estradiol and norethindroneoccurred. Some of the large variance noted historicallycan be attributed to the variability inherent in older assaymethodologies. In some studies, individual patients: 1)exhibited large decreases in the plasma concentrations ofethinyl estradiol; 2) experienced breakthrough bleedingor rises in plasma estradiol indicating incomplete follicu-lar suppression; or 3) appeared to ovulate.41–43,46,49 Onewoman taking a triphasic OC preparation had a historyof breakthrough bleeding whenever she took antibiotics.Measurement of plasma hormone concentrations on day15 showed lower plasma ethinyl estradiol concentrationswith minocycline.52 Moreover, one individual has beenreported who preferentially formed ethinyl estradiolconjugates because of an apparent inability to form the2-hydroxy derivative.53

DISCUSSION

Based on retrospective case series (usually without con-trol groups) and individual case reports, it has long beensuspected that the prescription of oral antibiotics towomen using OCs may occasionally result in OC failurein otherwise compliant patients. Caution must be exer-cised in the interpretation of these data because of recallbias and the possibility that individuals may underreportpoor compliance when they are confronted with anunplanned (or unwanted) pregnancy or are requestingabortion services.

These studies have not demonstrated any systematicinteraction between antibiotics and OC steroids, andclinical pharmacokinetic evidence indicating that antibi-otics alter blood concentrations of OCs is lacking. Nev-ertheless, it is possible that certain individuals may be atrisk of this interaction. Although reported cases of OCfailure in women concurrently taking antibiotics are

relatively few in number compared with the extensiveuse of these products, the true incidence of such failuresis unknown and cannot be predicted in individual cases.

Because the disposition of OC components is so vari-able among individuals, some may be more susceptibleto OC failure. Modern OC preparations contain smalleramounts of estrogen; thus, drug interactions leading todecreased efficacy of OCs may be more likely.54 How-ever, with the exception of rifampin (and griseofulvin),oral antibiotics do not induce hepatic drug metabolism.In fact, the macrolides and azole antifungals tend toinhibit the same cytochrome P450 isoform (3A4) thathydroxylates ethinyl estradiol.

Therefore, clinicians are faced with the task of discuss-ing risks, benefits, and alternatives—with the added bur-den of not knowing the magnitude of the risk of OCfailure in patients who are prescribed oral antibiotics,although the risk appears to be very small. It is so smallthat, given the comparatively large range of OC failure(1% to 3%), under patterns of typical use, it probablycannot be identified. If women who are at risk are thosewho have: 1) very low rates of ethinyl estradiol 2-hy-droxylation; 2) relatively high rates of hepatic conjuga-tion; 3) low plasma ethinyl estradiol concentrations; 4)extensive intestinal hydrolysis of estrogen conjugates;and 5) gut flora particularly susceptible to the antibioticbeing used, then averaging the results in small numbersof women could easily fail to detect any interaction.

Under these circumstances, and given the serious conse-quences of unwanted pregnancy, a cautious approach isadvisable to safeguard the few women using OCs who maybe at risk of OC failure. For those women on short-termbroad-spectrum antibiotic therapy, use of an additionalnonhormonal method of contraception may be justified.Changing contraceptive methods because of short-termantibiotic therapy is not advisable and may have negativeeffects on contraceptive compliance and efficacy. For thosewomen on long-term antibiotic therapy, use of an addi-tional nonhormonal method or alternative method of con-traception may be justified. The period of risk for druginteractions is not known during long-term therapy, butmay exist primarily during the first few weeks of therapy oruntil gut flora become resistant. An alternative form ofcontraception is advised if diarrhea or breakthrough bleed-ing are noted in women taking OCs who receive antibiotictherapy, or if the baseline rate of oral contraception failure isunacceptable.

CONCLUSIONS

The following statements, recommended by the Councilon Scientific Affairs, were adopted as American MedicalAssociation policy in June 2000.

858 Dickinson et al OC and Antibiotic Interactions OBSTETRICS & GYNECOLOGY

1. Women prescribed rifampin concomitantly withOCs faced significant risk of OC failure and shouldbe counseled about the additional use of nonhor-monal contraceptive methods during the course ofrifampin therapy.

2. Women using combined OCs should be informedabout the small risk of interactions with antibioticsand that it is not possible to identify in advance thewomen who may be at risk of OC failure. Womenwho are not comfortable with the small risk of inter-action should be counseled about the additional useof nonhormonal contraceptive methods. Womenwho have had previous OC failures or who developbreakthrough bleeding during concomitant use ofantibiotics and OCs should be counseled about theuse of alternate methods of contraception if theyengage in intercourse during the period of concomi-tant use, as they may be part of the subset of womenat high risk of contraceptive failure.

REFERENCES

1. Abma JC, Chandra A, Musher WD, Leterson LS, Pic-cinino LJ. Fertility, family planning and woman’s health:new data from the 1995 national survey of family growth.Vital Health Stat 23 1997;19:1–114.

2. Gerald K, McEvoy GK, eds. AHFS drug information.Bethesda, MD: American Society of Health System Phar-macists Inc.; 1999.

3. Steven K, Hebel SK, eds. Drug facts and comparisons. St.Louis, MO: Facts and Comparisons, 1999.

4. Hansten PD, Horn JR, eds. Drug interactions analysis andmanagement. St. Louis, Missouri: Facts and Comparisons,1999.

5. Cerel-Suhl SL, Yeager BF. Update on oral contraceptivepills. Am Fam Physician 1999;60:2073–84.

6. Reimers D, Jezek A. The simultaneous use of rifampicinand other antitubercular agents with oral contraceptives.Prax Pneumol 1971;25:255–62.

7. Nocke-Finck L, Breuer H, Reimers D. [Effects of rifampi-cin on the menstrual cycle and on oestrogen excretion inpatients taking oral contraceptives.] Dtsch Med Wochen-schr 1973;98:1521–3.

8. Gallagher JC, Ismail MA, Aladjem S. Reduced urinaryestriol levels with erythromycin therapy. Obstet Gynecol1980;56:381–2.

9. Skolnick JL, Stoler BS, Katz DB, Anderson WH.Rifampin, oral contraceptives, and pregnancy. JAMA1976;236:1382.

10. Baciewicz AM. Oral contraceptive drug interactions. TherDrug Monitor 1985;7:26–35.

11. Stockley IH. Tetracycline and oral contraceptives. J AmAcad Dermatol 1982;7:279–80.

12. Fraser IS, Jansen RPS. Why do inadvertent pregnancies

occur in oral contraceptive users? Contraception 1983;27:531–51.

13. Hempel E, Bohm W, Carol W, Klinger G. Medimkamen-tose enzyminduktion und hormonale kontrazeption. ZblGynak 1973;95:1451–7.

14. Silber TJ. Apparent oral contraceptive failure associatedwith antibiotic administration. J Adolesc Health Care1983;4:287–9.

15. Bainton R. Interaction between antibiotic therapy andcontraceptive medication. Oral Surg Oral Med OralPathol. 1986;61:453–5.

16. DeGroot AC, Eshuis H, Stricker BH. Inefficacy of oralcontraception during use of minocycline. Ned TijdschrGeneeskd 1990;134:1227–9.

17. Cote J. Interaction of griseofulvin and oral contraceptives[see comments]. J Am Acad Dermatol 1990;22:124–5.

18. Bollen M. Use of antibiotics when taking the oral contra-ceptive pill. Aust Fam Physician 1995;24:928–9.

19. Van Dijke CPH, Weber JCP. Interaction between oralcontraceptives and griseofulvin. BMJ 1984;288:1125–6.

20. Helms SE, Bredle DL, Zajic J, Jarjoura D, Brodell RT,Krishnarao I. Oral contraceptive failure rates and oralantibiotics. J Am Acad Dermatol 1997;36:705–10.

21. London BM, Lookingbill DP. Frequency of pregnancy inacne patients taking oral antibiotics and oral contracep-tives. Arch Dermatol 1994;130:392–3.

22. DeSanoEA Jr, Hurley SC. Possible interactions of antihis-tamines and antibiotics with oral contraceptive effective-ness. Fertil Steril 1982;37:853–4.

23. Hughes BR, Cunliffe WJ. Interactions between the oralcontraceptive pill and antibiotics. Br J Dermatol 1990;122:717–18.

24. Back DJ, Grimmer SF, Orme ML, Proudlove C, MannRD, Breckenridge AM. Evaluation of Committee onSafety of Medicines yellow card reports on oral contracep-tive-drug interactions with anticonvulsants and antibiotics.Br J Clin Pharmacol 1988;25:527–32.

25. Sparrow MJ. Pill method failure. NZ Med J 1987;100:102–5.

26. Kakouris H, Kovacs GT. Pill failure and non-use of sec-ondary precautions. Br J Fam Plan 1992;18:41–4.

27. Back DJ, Madden S, Orme MLE. Gastrointestinal metab-olism of contraceptive steroids. Am J Obstet Gynecol.1990;163:2138–45.

28. Jarvenpaa P, Kosunen T, Fotsis T, Adlercreutz H. In vitrometabolism of estrogens by isolated intestinal microorgan-isms and by human faecal microflora. J Steroid Biochem1980;13:345–9.

29. Goldzieher JW, Brody SA. Pharmacokinetics of ethinylestradiol and mestranol. Am J Obstet Gynecol 1990;163:2114–9.

30. Back DJ, Houlgrave R, Tija JF, Ward S, Orme ML. Effectof progestogens, gestodene, 3-keto desogestrel, levonorg-estrel, norethisterone and norgestimate on the oxidation of

859VOL. 98, NO. 5, PART 1, NOVEMBER 2001 Dickinson et al OC and Antibiotic Interactions

ethinyloestradiol and other substrates by human livermicrosomes. J Steroid Biochem Mol Biol 1991;38:219–25.

31. Jung-Hoffman C, Kuhl H. Pharmacokinetics and pharma-codynamic or oral contraceptive steroid: factors influenc-ing steroid metabolism. Am J Obstet Gynecol 1990;163:2183–97.

32. Orme MLE, Back DJ, Breckenridge AM. Clinical pharma-cokinetics of oral contraceptive steroids. Clin Pharmacoki-net 1983;8:95–136.

33. Fotherby K. Variability of pharmacokinetic parameters forcontraceptive steroids. J Steroid Biochem 1983;19:817–20.

34. Brody SA, Turkes A, Goldzieher JW. Pharmacokinerticsof three bioequivalent norethindrome/mestranol and threenorethindrome/ethinyl estradiol OC formulations: Are“low-dose” pills really lower? Contraception 1989;40:269–84.

35. Back DJ, Breckenridge AM, Crawford FE, Hall JM,MacIver M, Orme ML, et al. The effect of rifampicin onthe pharmacokinetics of ethynylestradiol in women. Con-traception 1980;21:135–43.

36. Joshi JV, Joshi UM, Sankolli GM, Gupta K, Rao AP,Hazari K, et al. A study of interaction of a low-dosecombination oral contraceptive with anti-tubercular drugs.Contraception 1980;21:617–29.

37. Bolt HM, Bolt M, Kappus H. Interaction of rifampicintreatment with pharmacokinetics and metabolism of ethi-nyloestradiol in man. Acta Endocrinol 1977;85:189–97.

38. Back DJ, Breckenridge AM, Crawford FE, Hall JM,MacIver M, Orme ML, et al. The effect of rifampicin onnorethisterone pharmacokinetics. Eur J Clin Pharmacol1979;15:193–7.

39. Meyer B, Muller F, Wessels P, Maree J. A model to detectinteractions between roxithromycin and oral contracep-tives. Clin Pharmacol Ther 1990;47:671–4.

40. Gelbke HP, Gethmann U, Knuppen R. Influence of rifam-picin treatment on the metabolic fate of [4-14C] mestranolin women. Horm Metab Res 1977;9:415–9.

41. Friedman CI, Huneke AL, Kim MH, Powell J. The effectof ampicillin on oral contraceptive effectiveness. ObstetGynecol 1980;55:33–7.

42. Joshi JV, Joshi UM, Sankholi GM, Krishna U, MandlekarA, Chowdhury V, et al. A study of interaction of low-dosecombination oral contraceptive with ampicillin and metro-nidazole. Contraception 1980;22:643–52.

43. Back DJ, Breckenridge AM, MacIver M, Orme M, RowePH, Staiger C, et al. The effects of ampicillin oral contra-ceptive steroids in women. Br J Clin Pharmacol 1982;14:43–8.

44. Neely JL, Abate M, Swinker M, D’Angio R. The effect of

doxycycline on serum levels of ethinyl estradiol, norethin-drone, and endogenous progesterone. Obstet Gynecol1991;77:416–20.

45. Murphy AA, Zacur HA, Charache P, Burkman RT. Theeffect of tetracycline on levels of oral contraceptives. Am JObstet Gynecol 1991;164:28–33.

46. Wermeling DP, Chandler MH, Sides GD, Collins D,Muse KN. Dirithromycin increases ethinyl estradiol clear-ance without allowing ovulation. Obstet Gynecol 1995;86:78–84.

47. Maggiolo F, Puricelli G, Dottorini M, Caprioli S, BianchiW, Suter F, et al. The effect of ciprofloxacin on oralcontraceptive steroid treatment. Drugs Exp Clin Res 1991;17:451–4.

48. Csemiczky G, Alvendal C, Landgren BM. Risk for ovula-tion in women taking a low-dose oral contraceptive(Microgynon) when receiving antibacterial treatment witha fluoroquinolone (ofloxacin). Adv Contracept 1996;12:101–9.

49. Back DJ, Tjia J, Martin C, Millar E, Mant T, Morrison P,et al. The lack of interaction between temafloxacin andcombined oral contraceptive steroids. Contraception1991;43:317–23.

50. Grimmer SF, Allen WL, Back DJ, Breckenridge AM,Orme M, Tjia J. The effect of cotrimoxazole on oralcontraceptive steroids in women. Contraception 1983;28:53–9.

51. Bolt HM, Kappus H, Bolt M. Effect of rifampicin treat-ment on the metabolism of oestradiol and 17alpha-ethinyl-oestradiol by human liver microsomes. Eur J Clin Phar-macol 1975;8:301–7.

52. Shenfield GM, Griffine JM. Clinical pharmacokinetics ofcontraceptive steroid: An update. Clin Pharmacokinet1991;20:15–37.

53. Maggs JL, Grimmer SFM, Orme MLE, Breckenridge AM,Park BK, Gilmore IT. The biliary and urinary metabolitesof 3H-17a-ethinyloestradiol in women. Xenobiotica 1983;13:421–31.

54. Szoka PR, Edgren RA. Drug interactions with oral contra-ceptives: Compilation and analysis of an adverse experi-ence report database. Fertil Steril 1988;49(Suppl):31–7.

Address reprint requests to: Barry Dickinson, PhD, Secretaryto the Council on Scientific Affairs, American Medical Associ-ation, 515 North State Street, Chicago, IL 60610; E-mail:barry_dickinson@ama-assn.org.

Received November 30, 2000. Received in revised form April 6, 2001.Accepted June 15, 2001.

860 Dickinson et al OC and Antibiotic Interactions OBSTETRICS & GYNECOLOGY