Management of Thyroid Dysfunction during Pregnancy and Postpartum:

An Endocrine Society Clinical Practice Guideline

GUIDELINESCLINICAL T h e E n d o c r i n e S o c i e t y s

The Endocrine Society8401 Connecticut Avenue, Suite 900

Chevy Chase, MD 20815

301.941.0200www.endo-society.org

C0-SPONSORING ORGANIZATIONS: American Association of Clinical Endocrinologists (AACE), Asia &Oceania Thyroid Association (AOTA), American Thyroid Association (ATA), European Thyroid Association(ETA), Latin American Thyroid Society (LATS)

Thyroid Dysfunction during Pregnancy and Postpartum Guideline Task Force: Marcos Abalovich, NobuyukiAmino, Linda A. Barbour, Rhoda H. Cobin, Leslie J. De Groot, Daniel Glinoer, Susan J. Mandel, and AlexStagnaro-Green

Endocrinology Division (M.A.), Durand Hospital, Buenos Aires, Argentina; Center for Excellence in Thyroid Care(N.A.), Kuma Hospital, Kobe 650-0011, Japan; Divisions of Endocrinology and Maternal-Fetal Medicine (L.A.B.),University of Colorado at Denver and Health Sciences Center, Aurora, Colorado 80010; The Mount Sinai School ofMedicine (R.H.C.), New York, New York 10016; Department of Medicine (L.J.D.G.), Division of Endocrinology,Brown University, Providence Rhode Island 02903; Endocrine Division (D.G.), University Hospital Saint Pierre, B-1000Brussels, Belgium; Division of Endocrinology, Diabetes, and Metabolism (S.J.M.), University of Pennsylvania School ofMedicine, Philadelphia, Pennsylvania 19104; Departments of Medicine and Obstetrics, Gynecology, and Women's Health(A.S.-G.), University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey07101; and Touro University College of Medicine (A.S.-G.), Hackensack, New Jersey 07601

Disclaimer: Clinical Practice Guidelines are developed to be of assistance to endocrinologists by providingguidance and recommendations for particular areas of practice. The Guidelines should not be considered inclusiveof all proper approaches or methods, or exclusive of others. The Guidelines cannot guarantee any specificoutcome, nor do they establish a standard of care. The Guidelines are not intended to dictate the treatment of aparticular patient. Treatment decisions must be made based on the independent judgment of healthcare providersand each patient's individual circumstances.

The Endocrine Society makes no warranty, express or implied, regarding the Guidelines and specifically excludesany warranties of merchantability and fitness for a particular use or purpose. The Society shall not be liable fordirect, indirect, special, incidental, or consequential damages related to the use of the information contained herein.

First published in the Journal of Clinical Endocrinology & Metabolism, August 2007, 92(8)(Supplement):S1S47

The Endocrine Society, 2007

Commercial Reprint InformationFor information on reprint requests of more than 101 and commercial reprints contact:

Heather EdwardsReprint Sales SpecialistCadmus Professional Communications

Phone: 410.691.6214Fax: 410.684.2789 Email: endoreprints@cadmus.com

Single Reprint InformationFor information on reprints of 100 and fewer, complete the guideline order form and return using one of thefollowing methods:

Mail: The Endocrine Societyc/o Bank of AmericaP.O. Box 630721Baltimore, MD 21263-0736

Fax: 301.941.0257Email: Societyservices@endo-society.org

Questions & CorrespondencesThe Endocrine SocietyAttn: Government & Public Affairs Department8401 Connecticut Avenue, Suite 900Chevy Chase, MD 20815

Phone: 301.941.0200Email: govt-prof@endo-society.orgWeb: www.endo-society.org

For more information on The Endocrine Societys Clinical Practice Guidelines or to download the completeversion of this guideline, visit http://www.endo-society.org/publications/guidelines/index.cfm.

MTSD07

Objective: The objective is to provide clinicalguidelines for the management of thyroid problemspresent during pregnancy and in the postpartum.

Participants: The Chair was selected by the ClinicalGuidelines Subcommittee (CGS) of The EndocrineSociety. The Chair requested participation by the LatinAmerican Thyroid Society, the Asia and OceaniaThyroid Society, the American Thyroid Association,the European Thyroid Association, and the AmericanAssociation of Clinical Endocrinologists, and eachorganization appointed a member to the task force. Twomembers of The Endocrine Society were also asked toparticipate. The group worked on the guidelines for 2 yrs and held two meetings. There was no corporatefunding, and no members received remuneration.

Evidence: Applicable published and peer-reviewedliterature of the last two decades was reviewed, with aconcentration on original investigations. The gradingof evidence was done using the United StatesPreventive Services Task Force system and, wherepossible, the GRADE system.

Consensus Process: Consensus was achieved throughconference calls, two group meetings, and exchangeof many drafts by E-mail. The manuscript wasreviewed concurrently by the Societys CGS, ClinicalAffairs Committee, members of The EndocrineSociety, and members of each of the collaboratingsocieties. Many valuable suggestions were receivedand incorporated into the final document. Each ofthe societies endorsed the guidelines.

Conclusions: Management of thyroid diseases duringpregnancy requires special considerations becausepregnancy induces major changes in thyroid function,and maternal thyroid disease can have adverse effectson the pregnancy and the fetus. Care requirescoordination among several healthcare professionals.Avoiding maternal (and fetal) hypothyroidism is of major importance because of potential damage to fetal neural development, an increased incidenceof miscarriage, and preterm delivery. Maternalhyperthyroidism and its treatment may beaccompanied by coincident problems in fetal thyroidfunction. Autoimmune thyroid disease is associatedwith both increased rates of miscarriage, for whichthe appropriate medical response is uncertain at thistime, and postpartum thyroiditis. Fine-needleaspiration cytology should be performed for dominantthyroid nodules discovered in pregnancy. Radioactiveisotopes must be avoided during pregnancy andlactation. Universal screening of pregnant women for thyroid disease is not yet supported by adequatestudies, but case finding targeted to specific groups of patients who are at increased risk is stronglysupported.JJ CClliinn EEnnddooccrriinnooll MMeettaabb:: 9922 ((88)) ((SSuupppplleemmeenntt))::SS11--SS4477,, 22000077

Abstract

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

1

THE

ENDO

CRIN

E SO

CIET

YS

CLIN

ICAL

GU

IDEL

INES

2

IINNTTRROODDUUCCTTIIOONN

Over the past 15 yrs there has been a rapid expansionof knowledge regarding thyroid disease andpregnancy. These advances relate to the optimalmanagement of pregnant women on levothyroxinetherapy, the impact of iodine deficiency on themother and developing fetus, the adverse effect ofmaternal hypothyroidism on mental development intheir infants, the syndrome of postpartum thyroiditis,and its relation to permanent hypothyroidism.Furthermore, a doubling of the miscarriage rate hasbeen reported in studies in antibody-positiveeuthyroid women, and an increase in preterm deliveryhas been found in women with subclinicalhypothyroidism and/or thyroid autoimmunity.

Given the rapidity of advances in this field, it is notsurprising that controversy surrounds optimaldetection and management of thyroid disease in thepregnant woman. Thyroid disease during pregnancyhas certain characteristics that make writingguidelines more complicated than for some otherfields. This field is concerned with the managementof pregnant women who may have a variety of knownor undisclosed thyroid conditions, such ashypothyroidism and hyperthyroidism, the presence ofthyroid auto-antibodies, the presence of nodules, orunsatisfactory iodine nutrition. Pregnancy may affectthe course of these thyroid disorders and, conversely,thyroid diseases may affect the course of pregnancy.

Moreover, thyroid disorders (and their management)may affect both the pregnant woman and thedeveloping fetus. Finally, pregnant women may beunder the care of multiple health care professionals,including obstetricians, nurse midwives, familypractitioners, endocrinologists, and/or internists,making the development of guidelines all the morecritical.

MMEETTHHOODDSS OOFF DDEEVVEELLOOPPMMEENNTT

An international task force was created, under theauspices of The Endocrine Society, to review the bestevidence in the field and develop evidence-basedguidelines. Members of the task force includedrepresentatives from The Endocrine Society,American Thyroid Association, Association ofAmerican Clinical Endocrinologists, EuropeanThyroid Association, Asia and Oceania ThyroidAssociation, and the Latin American ThyroidSociety. The task force worked during 2 yrs to developthe guidelines, had multiple phone conversations,and two 2-d retreats. Upon completion of theguidelines, they were reviewed and approved by all ofthe participants.

Our committee undertook to review all material onthese topics published in English during the past twodecades, or earlier at the working group's discretion.We concentrated on original reports and largelyexcluded reviews from our references. At present,with the exception of studies on iodidesupplementation, only two prospective, randomizedintervention trials have been published in this area.We are aware of two large-scale prospectiveintervention trials that are presently ongoing.Nevertheless, in the last 15 yrs, many high-qualitystudies have modified older dogmas and profoundlychanged the ways in which these patients aremanaged. These studies are most often prospective orretrospective clinical evaluations of a particularpatient population and matched groups of controlwomen. Such studies, when carefully performed,adequately matched, and appropriately interpreted;provide the bulk of the evidence presented herein.

The committee evaluated recommendations andevidence using the methodology of the United StatesPreventive Service Task Force (USPSTF), in whichtreatments or medical advice are referred to as aservice. The USPSTF grades its recommendations(level A, B, C, D, or I) on the basis of the strength ofevidence and magnitude of net benefit (benefitsminus harms), as follows:

3A: The USPSTF strongly recommends that cliniciansprovide (the service) to eligible patients. TheUSPSTF found good evidence that (the service)improves important health outcomes and concludes thatbenefits substantially outweigh harms.

B: The USPSTF recommends that clinicians provide(the service) to eligible patients. The USPSTF foundat least fair evidence that (the service) improvesimportant health outcomes and concludes that benefitsoutweigh harms.

C: The USPSTF makes no recommendation for oragainst routine provision of (the service). TheUSPSTF found at least fair evidence that (the service)can improve health outcomes but concludes that thebalance of benefits and harms is too close to justify ageneral recommendation.

D: The USPSTF recommends against routinelyproviding (the service) to asymptomatic patients. TheUSPSTF found good evidence that (the service) isineffective or that harms outweigh benefits.

I: The USPSTF concludes that the evidence isinsufficient to recommend for or against routinelyproviding (the service). Evidence that (the service) iseffective is lacking, or poor quality, or conflicting, andthe balance of benefits and harms cannot be determined.

The USPSTF grades the quality of the overallevidence for a service on a three-point scale (good,fair, or poor), defined as follows:

Good: Evidence includes consistent results from welldesigned, well conducted studies in representativepopulations that directly assess effects on healthoutcomes.

Fair: Evidence is sufficient to determine effects onhealth outcomes, but the strength of the evidence islimited by the number, quality, or consistency of the individual studies, generalizability to routinepractice, or indirect nature of the evidence on healthoutcomes.

Poor: Evidence is insufficient to assess the effects onhealth outcomes because of limited number or powerof studies, important flaws in their design or conduct,gaps in the chain of evidence, or lack of informationon important health outcomes.

In addition to the USPSTF grading ofrecommendations, we have also included theappropriate recommendation level as indicated by theGRADE system. The value of an evidence-basedrecommendation, using the GRADE system, is scoredfrom strong to moderate (1-2) and accompanied bysymbols indicating the value of the evidence: high (1,

or ), moderate (2, ), low( ), and very low ( ). (There are noequivalents in the GRADE system for therecommendation levels C, D, and I used in theUSPSTF system.)

The supporting data for the full committee reportfollow this executive summary. The supporting dataconsists of eight subsections dealing in detail withspecific maternal/fetal thyroid problems. Eachsubsection provides the related background andevidence for recommendations. In the subsectionreports, the task force has indicated specificbibliographic citations on which eachrecommendation is based, and for each report cited asevidence for a given recommendation. We believethat this approach provides an important direct linkbetween the supporting evidence and therecommendation.

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

BBAACCKKGGRROOUUNNDD AANNDD EEVVIIDDEENNCCEE

The complete discussion of background data andevidence is offered in the supporting data that followthis executive summary. Some important issues arenoted here.

Pregnant and lactating women require additionaliodine intake, whether in iodine-poor or iodine-sufficient countries. The recommended averageiodine intake is approximately 250 g/d (1). Severeiodine deficiency, if inadequately treated, is a majorcause of neurological damage worldwide (2).

Both overt and subclinical hypothyroidism haveadverse effects on the course of pregnancy anddevelopment of the fetus (3,4,5). Hypothyroidismshould be corrected before initiation of pregnancy,replacement dosage should be augmented early in pregnancy (6), and euthyroidism should be maintained throughout. Overt maternalhypothyroidism has been associated with damage tofetal intellectual development (7), presumablybecause of inadequate transplacental supply ofhormone during early pregnancy (8). Whethersubclinical hypothyroidism carries this risk remainsunproven, but replacement therapy for this conditionis nonetheless advised.

Propylthiouracil is recommended as the first-line drugfor treatment of hyperthyroidism during pregnancy,because of the probable association of methimazolewith fetal developmental abnormalities (9, 10).Maternal Graves disease, past or present, carries arisk for the pregnancy and for the fetus. Antithyroiddrug (ATD) therapy to the mother can induce fetal hypothyroidism, and transplacental passage of TSH-receptor antibodies (TRAb) can cause fetal hyperthyroidism (11,12,13). Targeting ATDtreatment to maintain maternal serum free T4 levelsat the upper limit of the nonpregnant T4 range usuallyprotects the fetus from hypothyroidism (14). Closefollowing of maternal T4 and TSH levels, assay ofTRAb, and fetal ultrasonography including thethyroid are recommended for guiding therapy (15),

THE

ENDO

CRIN

E SO

CIET

YS

CLIN

ICAL

GU

IDEL

INES

4

and fetal blood sampling is rarely needed (15,16).Fetal hyperthyroidism does not occur duringpregnancies in which TRAb levels are normal andATD is not administered. Surgery may be required insome instances. Propylthiouracil, propranolol, andiodides may be used for preoperative preparation.

Hyperemesis is associated with elevation of thyroidhormone levels above average pregnancy values and suppression of TSH (17,18,19). Occasionally,patients are clinically thyrotoxic. The elevation ofthyroid hormone levels and gestational hyper-thyroidism are typically self-remitting and in mostcases do not require antithyroid treatment (17,20).Subclinical hyperthyroidism, commonly found in thissetting, does not require therapy, and therapy isadvised against because it might induce fetalhypothyroidism (21).

Thyroid nodules recognized during pregnancy, orgrowing, are typically biopsied under ultrasoundguidance (22,23), and if appropriate, surgery isperformed in the mid-trimester (24). Delay intreatment of low-grade tumors until after delivery isnot considered a danger (25). Pregnancy is notthought to adversely affect the course of thyroidmalignancy (26,27,28). TSH suppression for knownthyroid malignancy may be maintained duringpregnancy with detectable TSH and with T4 at theupper end of the range for normal pregnancy.Radioactive iodine (RAI) must not be administeredduring pregnancy or lactation.

Autoimmune thyroid disease is common inpregnancy. The presence of antibodies to thyroidperoxidase or thyroglobulin is associated with asignificant increment in miscarriages (29,30). Oneprospective study has reported that treatment with T4during pregnancy may reverse this risk (31).Additional studies on this important issue are needed.

PPT, a form of autoimmune thyroid disease closelyrelated to Hashimotos thyroiditis, is found in about7% of women in the postpartum period (32). It causeshyperthyroidism and/or hypothyroidism that isusually transient (but often seriously symptomatic)

5(33,34) and increases the risk of later permanenthypothyroidism (35,36). Although depression may bea symptom of hypothyroidism in any setting, PPT perse has not been clearly linked to postpartumdepression (37, 38).

A major unsettled question is the advisability ofuniversal screening of pregnant women for thyroiddisease, through TSH testing, and possibly antibodytesting. The prevalence of overt thyroid disease inthis population is 1%, and there is also a 2-3%prevalence of subclinical hypothyroidism and 10-15%antibody positivity (30, 39). As of this date, only onestudy has demonstrated that treatment of antibodypositive euthyroid women with T4 decreases the rateof miscarriage and preterm delivery (31). Thus, fornow, the committee recommends targeted casefinding during early pregnancy but anticipates thatongoing studies may alter this recommendation (40).Vaidya et al. (41) recently reported a study ofscreening by means of TSH, T4, free T4, and thyroidperoxidase antibodies in 1560 consecutive pregnantwomen. An important result was that screening onlywomen considered high risk on the basis of a personalor family history of thyroid disease or a history ofother autoimmune disease would have missed 30% ofwomen with overt or subclinical hypothyroidism.

RREECCOOMMMMEENNDDAATTIIOONNSS

11.. HHyyppootthhyyrrooiiddiissmm aanndd PPrreeggnnaannccyy:: MMaatteerrnnaallaanndd FFeettaall AAssppeeccttss

1.1.1. Both maternal and fetal hypothyroidism areknown to have serious adverse effects on the fetus.Therefore maternal hypothyroidism should beavoided. USPSTF recommendation level is A;evidence is fair (GRADE 1| ). Targetedcase finding is recommended at the first prenatalvisit or at diagnosis of pregnancy (see Section 8,Screening for thyroid dysfunction during pregnancy).USPSTF recommendation level is B; evidence isfair (GRADE 2| ).

1.1.2. If hypothyroidism has been diagnosed beforepregnancy, we recommend adjustment of thepreconception thyroxine dose to reach a TSH levelnot higher than 2.5 U/mL prior to pregnancy.(USPSTF Recommendation level: I, Evidence-poor).(GRADE 1| )

1.1.3. The T4 dose usually needs to be incrementedby 4-6 wk gestation and may require a 30-50%increase in dosage. USPSTF recommendation level isA; evidence is good (GRADE 1| ).

1.1.4. If overt hypothyroidism is diagnosed duringpregnancy, thyroid function tests (TFTs) should benormalized as rapidly as possible. Thyroxine dosageshould be titrated to rapidly reach and thereaftermaintain serum TSH concentrations of less than 2.5 U/mL in the first trimester (or 3 U/mL in thesecond and third trimester) or to trimester-specificnormal TSH ranges. Thyroid function tests should be remeasured within 30-40 days. (USPSTFRecommendation level: A, Evidence-good) (GRADE1| )

1.1.5. Women with thyroid autoimmunity who areeuthyroid in the early stages of pregnancy are at riskof developing hypothyroidism and should bemonitored for elevation of TSH above the normalrange. (USPSTF Recommendation level: A,Evidence-good) (GRADE 1| )

1.1.6. Subclinical hypothyroidism (serum TSHconcentration above the upper limit of the referencerange with a normal free T4) has been shown to beassociated with an adverse outcome for both themother and offspring. T4 treatment has been shownto improve obstetrical outcome but has not beenproved to modify long-term neurological develop-ment in the offspring. However, given that thepotential benefits outweigh the potential risks, thepanel recommends T4 replacement in women withsubclinical hypothyroidism. For obstetrical outcome,USPSTF recommendation level is B; evidence is fair(GRADE 1| ). For neurological outcome,USPSTF recommendation level is I; evidence is poor

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

1.1.7. After delivery, most hypothyroid women needa decrease in the thyroxine dosage they receivedduring pregnancy. (USPSTF Recommendation level:A, Evidence-good) (GRADE 1| )

22.. MMaannaaggeemmeenntt ooff MMaatteerrnnaall HHyyppeerrtthhyyrrooiiddiissmm::MMaatteerrnnaall((aa)) aanndd FFeettaall AAssppeeccttss((bb))

2.1.a.1. If a subnormal serum TSH concentration isdetected during gestation, hyperthyroidism must bedistinguished from both normal physiology duringpregnancy and hyperemesis gravidarum because of theadverse effects of overt hyperthyroidism on themother and fetus. Differentiation of Graves diseasefrom gestational thyrotoxicosis is supported byevidence of autoimmunity, a goiter, and presence ofTSH receptor antibodies (TRAb). (USPSTFRecommendation level: A, Evidence-good) (GRADE1| )

2.1.a.2. For overt hyperthyroidism due to Gravesdisease or hyper-functioning thyroid nodules,antithyroid drug (ATD) therapy should be eitherinitiated (for those with new diagnoses) or adjusted(for those with a prior history) to maintain thematernal thyroid hormone levels for free T4 in theupper nonpregnant reference range. (USPSTFRecommendation level-A, Evidence-good) (GRADE1| )

2.1.a.3. Because available evidence suggestsmethimazole may be associated with congenitalanomalies, propylthiouracil should be used as a first-line drug, if available, especially during first-trimester organogenesis. Methimazole may beprescribed if propylthiouracil is not available or if apatient cannot tolerate or has an adverse response topropylthiouracil. USPSTF recommendation level isB; evidence is fair (GRADE 1| ).

2.1.a.4. Subtotal thyroidectomy may be indicatedduring pregnancy as therapy for maternal Gravesdisease if (1) a patient has a severe adverse reaction toATD therapy, (2) persistently high doses of ATD arerequired, or (3) a patient is not adherent to ATDtherapy and has uncontrolled hyperthyroidism. Theoptimal timing of surgery is in the second trimester.(USPSTF Recommendation level: I, Evidence-poor)( )

2.1.a.5. There is no evidence that treatment ofsubclinical hyperthyroidism improves pregnancyoutcome, and treatment could potentially adverselyaffect fetal outcome. (USPSTF Recommendationlevel: I, Evidence-poor) ( )

2.1.b.1 TRAb (either TSH receptor-stimulating or -binding antibodies) freely cross the placenta and canstimulate the fetal thyroid. These antibodies shouldbe measured before pregnancy or by the end of thesecond trimester in mothers with current Gravesdisease, with a history of Graves disease andtreatment with 131I or thyroidectomy, or with aprevious neonate with Graves disease. Women whohave a negative TRAb and do not require ATD havea very low risk of fetal or neonatal thyroiddysfunction. USPSTF recommendation level is B;evidence is fair (GRADE 1| ).

2.1.b.2. 131I should not be given to a woman who is ormay be pregnant. If inadvertently treated, the patientshould be promptly informed of the radiation dangerto the fetus, including thyroid destruction if treatedafter the 12th week of gestation. USPSTFrecommendation level is A; evidence is good(GRADE 1| ). There are no data for oragainst recommending termination of pregnancy after131I exposure. USPSTF recommendation level is I;evidence is poor ( )

2.1.b.3. In women with elevated TRAb or in womentreated with ATD, fetal ultrasound should beperformed to look for evidence of fetal thyroiddysfunction that could include growth restriction,hydrops, presence of goiter, or cardiac failure.(USPSTF Recommendation level: B, Evidence-fair)(GRADE 1| )TH

E EN

DOCR

INE

SOCI

ETY

S CL

INIC

AL G

UID

ELIN

ES

6

2.1.b.4. Umbilical blood sampling should beconsidered only if the diagnosis of fetal thyroiddisease is not reasonably certain from the clinical dataand if the information gained would change thetreatment.(USPSTF Recommendation level: B,Evidence-fair) (GRADE 1| )

2.1.b.5. All newborns of mothers with Gravesdisease should be evaluated for thyroid dysfunctionand treated if necessary (USPSTF Recommendationlevel: B, Evidence-fair) (GRADE 2| )

33.. GGeessttaattiioonnaall HHyyppeerreemmeessiiss aannddHHyyppeerrtthhyyrrooiiddiissmm

3.1. Thyroid function tests should be measured in allpatients with hyperemesis gravidarum (5% weightloss, dehydration, and ketonuria) (USPSTFRecommendation level: B, Evidence-poor) (GRADE2| )

3.2. Few women with hyperemesis gravidarum willrequire ATD treatment. USPSTF recommendationlevel is A; evidence is good (GRADE 1| ).Overt hyperthyroidism believed due to coincidentGraves disease should be treated with ATD. USPSTF recommendation level is B; evidence is fair(GRADE 1| ). Gestational hyperthyroidismwith clearly elevated thyroid hormone levels (free T4 above the reference range or total T4 > 150%of top normal pregnancy value and TSH < 0.1 U/ml)and evidence of hyperthyroidism may requiretreatment as long as clinically necessary. USPSTFrecommendation level is I; evidence is poor( ).

44.. AAuuttooiimmmmuunnee TThhyyrrooiidd DDiisseeaassee aannddMMiissccaarrrriiaaggee

4.1. Although a positive association exists betweenthe presence of thyroid antibodies and pregnancy loss,universal screening for antithyroid antibodies, andpossible treatment, can not be recommended at thistime. As of this date, only one adequately designed

intervention trial has demonstrated a decrease in themiscarriage rate in thyroid antibody positiveeuthyroid women (USPSTF Recommendation level:C, Evidence-fair) (GRADE 2| )

55.. TThhyyrrooiidd NNoodduulleess aanndd CCaanncceerr

5.1. Fine needle aspiration (FNA) cytology should beperformed for thyroid nodules >1 cm discovered inpregnancy. Ultrasound guided FNA may have anadvantage for minimizing inadequate sampling.(USPSTF Recommendation level: B; Evidence-fair)(GRADE 1| )

5.2. When nodules are discovered in the first or earlysecond trimester to be malignant on cytopathologicanalysis or exhibit rapid growth, pregnancy shouldnot be interrupted but surgery should be offered in thesecond trimester, before fetal viability. Women foundto have cytology indicative of papillary cancer orfollicular neoplasm without evidence of advanceddisease, which prefer to wait until the postpartumperiod for definitive surgery, may be reassured thatmost well-differentiated thyroid cancers are slowgrowing and that surgical treatment soon afterdelivery is unlikely to adversely affect prognosis.(USPSTF Recommendation level: B, Evidence-fair)(GRADE 1| )

5.3. It is appropriate to administer thyroid hormone toachieve a suppressed but detectable TSH in pregnantwomen with a previously treated thyroid cancer, or anFNA positive for or suspicious for cancer, and thosewho elect to delay surgical treatment untilpostpartum. High risk patients may benefit from agreater degree of TSH suppression compared to lowrisk patients. The free T4 or total T4 levels shouldideally not be increased above the normal range forpregnancy. (USPSTF Recommendation level: I,Evidence-poor) ( )

5.4. RAI administration with 131I should not be givento women who are breastfeeding. USPSTFrecommendation level is B; evidence is fair (GRADE1| ). Furthermore, pregnancy should be

7

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

avoided for 6 months to 1 yr in women with thyroidcancer who receive therapeutic RAI doses to ensurestability of thyroid function and confirm remission ofthyroid cancer. USPSTF recommendation level is B;evidence is fair (GRADE 1| ).

66.. IIooddiinnee NNuuttrriittiioonn dduurriinngg PPrreeggnnaannccyy

6.1. Women in the childbearing age should have anaverage iodine intake of 150 g per day. Duringpregnancy and breast-feeding, women should increasetheir daily iodine intake to 250 g on average.(USPSTF Recommendation level: A, Evidence-good) (GRADE 1| )

6.2. Iodine intake during pregnancy andbreastfeeding should not exceed twice the dailyrecommended nutritional intake for iodine, i.e. 500g iodine per day. USPSTF recommendation level isI; evidence is poor ( ).

6.3. To assess the adequacy of the iodine intakeduring pregnancy in a population, urinary iodineconcentration (UIC) should be measured in a cohortof the population. UIC should ideally range between150 and 250 g/L. (USPSTF Recommendation level:A, Evidence-good) (GRADE 1| )

6.4. To reach the daily recommended nutrient intakefor iodine, multiple means must be considered,tailored to the iodine intake level in a givenpopulation. Different situations must therefore bedistinguished: a) countries with iodine sufficiencyand/or with a well-established universal saltiodization (USI) program; b) countries without a USIprogram or an established USI program where thecoverage is known to be only partial; and finally c) remote areas with no accessible USI program anddifficult socioeconomic conditions. (USPSTFRecommendation level: A, Evidence-good) (GRADE1| )

77.. PPoossttppaarrttuumm TThhyyrrooiiddttiiss

7.1. There are insufficient data to recommendscreening of all women for postpartum thyroiditis(PPT) (USPSTF Recommendation level: I,Evidence-poor) ( )

7.2. Women known to be thyroid peroxidaseantibody positive should have a TSH performed at 3and 6 months postpartum (USPSTF Recommendationlevel: A, Evidence-good) (GRADE 1| )

7.3. The prevalence of PPT in women with type 1diabetes is threefold greater than in the generalpopulation. Postpartum screening (TSHdetermination) is recommended for women with type1 diabetes mellitus at 3 and 6 months postpartum(USPSTF Recommendation level: B, Evidence-fair)(GRADE 1| )

7.4. Women with a history of PPT have a markedlyincreased risk of developing permanent primaryhypothyroidism in the 5- to 10-year period followingthe episode of PPT. An annual TSH level should be performed in these women (USPSTFRecommendation level: A, Evidence-good) (GRADE1| )

7.5. Asymptomatic women with PPT who have aTSH above the reference range but below 10 U/mLand who are not planning a subsequent pregnancy do not necessarily require intervention, but should, if untreated, be re-monitored in 48 weeks.Symptomatic women and women with a TSH above normal and who are attempting pregnancyshould be treated with levothyroxine. (USPSTFRecommendation level: B, Evidence-fair) ( )

7.6. There is insufficient evidence to concludewhether an association exists between postpartumdepression (PPD) and either PPT or thyroid antibodypositivity (in women who did not develop PPT).(USPSTF Recommendation level: I, Evidence-poor)

However, as hypothyroidism is a potentially reversiblecause of depression, women with postpartum

THE

ENDO

CRIN

E SO

CIET

YS

CLIN

ICAL

GU

IDEL

INES

8

depression should be screened for hypothyroidism andappropriately treated (USPSTF Recommendationlevel: B, Evidence-fair) (GRADE 2| )

88.. SSccrreeeenniinngg ffoorr TThhyyrrooiidd DDyyssffuunnccttiioonn dduurriinnggPPrreeggnnaannccyy

Although the benefits of universal screening forthyroid dysfunction (primarily hypothyroidism) maynot be justified by the current evidence (presentedabove), we recommend case finding among thefollowing groups of women at high risk for thyroiddisease by measurement of TSH:1. Women with a history of hyperthyroid or

hypothyroid disease, PPT, or thyroid lobectomy.2. Women with a family history of thyroid disease.3. Women with a goiter.4. Women with thyroid antibodies (when known).5. Women with symptoms or clinical signs suggestive

of thyroid underfunction or overfunction,including anemia, elevated cholesterol, andhyponatremia.

6. Women with type I diabetes.7. Women with other autoimmune disorders.8. Women with infertility who should have

screening with TSH as part of their infertilitywork-up.

9. Women with previous therapeutic head or neckirradiation.

10.Women with a history of miscarriage or pretermdelivery. USPSTF recommendation level is B;evidence is fair (GRADE 1| ).

9

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

THE

ENDO

CRIN

E SO

CIET

YS

CLIN

ICAL

GU

IDEL

INES

1. World Health Organization, Technical consultation ofexperts in Geneva in January 2005. The prevention andcontrol of iodine deficiency in pregnant and lactatingwomen and in children under two years:recommendations of a WHO Technical Consultation.Public Health Nutr, in press

2. Bleichrodt N, Born M 1994 A meta-analysis of researchon iodine and its relationship to cognitive development.In: Stanbury JB, ed. The damaged brain of iodinedeficiency: cognitive, behavioral, neuromotor, educativeaspects. New York: Cognizant Communication; 195-200

3. Davis LE, Leveno KJ, Cunningham FG 1988Hypothyroidism complicating pregnancy. Obstet Gynecol72:108-112

4. Leung AS, Millar LK, Koonings PP, Montoro M,Mestman JH 1993 Perinatal outcome in hypothyroidpregnancies. Obstet Gynecol 81:349-353

5. Wasserstrum N, Anania CA 1995 Perinatalconsequences of maternal hypothyroidism in earlypregnancy and inadequate replacement. Clin Endocrinol(Oxf) 42:353-358

6. Alexander EK, Marqusee E, Lawrence J, Jarolim P,Fischer GA, Larsen R 2004 Timing and magnitude ofincreases in levothyroxine requirements during pregnancyin women with hypothyroidism. N Engl J Med 351:241-249

7. Haddow JE, Palomaki GE, Allan WC, Williams JR,Knight GJ, Gagnon J, OHeir CE, Mitchell ML,Hermos RJ, Waisbren SE, Faix JD, Klein RZ 1999Maternal thyroid deficiency during pregnancy andsubsequent neuropsychological development of the child.N Engl J Med 341:549-555

8. Kester MH, Martinez de Mena R, Obregon MJ,Marinkovic D, Howatson A, Visser TJ, Hume R,Morreale de Escobar G 2004 Iodothyronine levels in thehuman developing brain: major regulatory roles ofiodothyronine deiodinases in different areas. J ClinEndocrinol Metab 89:3117-3128

9. Clementi M, Di Gianantonio E, Pelo E, Mammi I,Basile RT, Tenconi R 1999 Methimazole embryopathy:delineation of the phenotype. Am J Med Genet 83:43-46

10. Johnsson E, Larsson G, Ljunggren M 1997 Severemalformations in infant born to hyperthyroid woman onmethimazole. Lancet 350:1520

11. Peleg D, Cada S, Peleg A, Ben-Ami M 2002 Therelationship between maternal serum thyroid-stimulatingimmunoglobulin and fetal and neonatal thyrotoxicosis.Obstet Gynecol 99:1040-1043

12. Weetman AP 2000 Graves disease. N Engl J Med343:1236-1248

13. McKenzie JM, Zakarija M 1992 Fetal and neonatalhyperthyroidism and hypothyroidism due to maternalTSH receptor antibodies. Thyroid 2:155-159

14. Momotani N, Noh JY, Ishikawa N, Ito K 1997 Effects ofpropylthiouracil and methimazole on fetal thyroid statusin mothers with Graves hyperthyroidism. J ClinEndocrinol Metab 82:3633-3636

15. Luton D, Le Gac I, Vuillard E, Castanet M,Guibourdenche J, Noel M, Toubert ME, Leger J,Boissinot C, Schlageter MH, Garel C, Tebeka B, OuryJF, Czernichow P, Polak M 2005 Management of Graves'disease during pregnancy: the key role of fetal thyroidgland monitoring. J Clin Endocrinol Metab 90:6093-6098

16. Laurberg P, Nygaard B, Glinoer D, Grussendorf M,Orgiazzi J 1998 Guidelines for TSH-receptor antibodymeasurements in pregnancy: results of an evidence-basedsymposium organized by the European ThyroidAssociation. Eur J Endocrinol 139:584-586

17. Goodwin TM, Montoro M, Mestman JH 1992 Transienthyperthyroidism and hyperemesis gravidarum: clinicalaspects. Am J Obstet Gynecol 167:648-652

18. Al-Yatama M, Diejomaoh M, Nandakumaran M,Monem RA, Omu AE, Al Kandari F 2002 Hormoneprofile of Kuwaiti women with hyperemesis gravidarum.Arch Gynecol Obstet 266:218-222

19. Leylek OA, Cetin A, Toyaksi M, Erselcan T 1996Hyperthyroidism in hyperemesis gravidarum. Int JGynaecol Obstet 55:33-37

20. Tan JY, Loh KC, Yeo GS, Chee YC 2002 Transienthyperthyroidism of hyperemesis gravidarum. BJOG109:683-688

21. Casey BM, Dashe JS, Wells CE, McIntire DD, LevenoKJ, Cunningham FG 2006 Subclinical hyperthyroidismand pregnancy outcomes. Obstet Gynecol 107:337-341

22. Choe W, McDougall IR 1994 Thyroid cancer in pregnantwomen: diagnostic and therapeutic management. Thyroid4:433-435

10

References

11

23. Rosen IB, Korman M, Walfish PG 1997 Thyroidnodular disease in pregnancy: current diagnosis andmanagement. Clin Obstet Gynecol 40:81-89

24. Sam S, Molitch ME 2003 Timing and special concernsregarding endocrine surgery during pregnancy. EndocrinolMetab Clin North Am 32:337-354

25. Doherty CM, Shindo ML, Rice DH, Montero M,Mestman JH 1995 Management of thyroid nodulesduring pregnancy. Laryngoscope 105:251-255

26. Moosa M, Mazzaferri EL 1997 Outcome of differentiatedthyroid cancer diagnosed in pregnant women. J ClinEndocrinol Metab 82:2862-2866

27. Herzon FS, Morris DM, Segal MN, Rauch G, Parnell T1994 Coexistent thyroid cancer and pregnancy. ArchOtolaryngol Head Neck Surg 120:1191-1193

28. Schlumberger M, De Vathaire F, Ceccarelli C, FranceseC, Pinchera A, Parmentier C 1995 Outcome ofpregnancy in women with thyroid carcinoma. JEndocrinol Invest 18:150-151

29. Glinoer D, Soto MF, Bourdoux P, Lejeune B, Delange F,Lemone M, Kinthaert J, Robijn C, Grun JP, de Nayer P1991 Pregnancy in patients with mild thyroidabnormalities: maternal and neonatal repercussions. JClin Endocrinol Metab 73:421-427

30. Stagnaro-Green A, Roman SH, Cobin RH, el-HarazyE, Alvarez-Marfany M, Davies TF 1990 Detection of at-risk pregnancy by means of highly sensitive assays forthyroid autoantibodies. JAMA 264:1422-1425

31. Negro R, Formoso G, Mangieri T, Pezzarossa A, DazziD, Hassan H 2006 Levothyroxine treatment in euthyroidpregnant women with autoimmune thyroid disease: effectson obstetrical complications. J Clin Endocrinol Metab91:2587-2591

32. Amino N, Tada H, Hidaka Y 1999 Postpartumautoimmune thyroid syndrome: a model of aggravation ofautoimmune disease. Thyroid 9:705-713

33. Stagnaro-Green A 2002 Postpartum thyroiditis. J ClinEndocrinol Metab 87:4042-4047

34. Lucas A, Pizarro E, Granada ML, Salinas I, Foz M,Sanmarti A 2000 Postpartum thyroiditis: epidemiologyand clinical evolution in a nonselected population.Thyroid 10:71-77

35. Othman S, Phillips DI, Parkes AB, Richards CJ, HarrisB, Fung H, Darke C, John R, Hall R, Lazarus JH 1990A long-term follow-up of postpartum thyroiditis. ClinEndocrinol (Oxf) 32:559-564

36. Tachi J, Amino N, Tamaki H, Aozasa M, Iwatani Y,Miyai K 1988 Long term follow-up and HLA associationin patients with postpartum hypothyroidism. J ClinEndocrinol Metab 66:480-484

37. Kent GN, Stuckey BG, Allen JR, Lambert T, Gee V1999 Postpartum thyroid dysfunction: clinical assessmentand relationship to psychiatric affective morbidity. ClinEndocrinol (Oxf) 51:429-438

38. Lucas A, Pizarro E, Granada ML, Salinas I, Sanmarti A2001 Postpartum thyroid dysfunction and postpartumdepression: are they two linked disorders? Clin Endocrinol(Oxf) 55:809-814

39. Allan WC, Haddow JE, Palomaki GE, Williams JR,Mitchell ML, Hermos RJ, Faix JD, Klein RZ 2000Maternal thyroid deficiency and pregnancycomplications: implications for population screening. JMed Screen 7:127-130

40. Lazarus JH, Premawardhana LD 2005 Screening forthyroid disease in pregnancy. J Clin Pathol 58:449-452

41. Vaidya B, Anthony S, Bilous M, Shields B, Drury J,Hutchinson S, and Bilous R 2007 Detection of thyroiddysfunction in early pregnancy: universal screening ortargeted high-risk case finding. J Clin Endocrinol Metab92:203-207

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

SSUUBBSSEECCTTIIOONN RREEPPOORRTTSS

SSEECCTTIIOONN 11.. MMAANNAAGGEEMMEENNTT OOFFHHYYPPOOTTHHYYRROOIIDDIISSMM:: MMAATTEERRNNAALL AANNDD FFEETTAALL AASSPPEECCTTSS

1.2. BACKGROUND

1.2.1. Clinical epidemiology and causal factors.The prevalence of hypothyroidism during pregnancyis estimated to be 0.30.5% for overt hypothyroidism(OH) and 23% for subclinical hypothyroidism(SCH). Thyroid autoantibodies are found in 515%of women in the childbearing age, and chronicautoimmune thyroiditis is the main cause ofhypothyroidism during pregnancy (15). As anexample, in a prospective population study of 9471 pregnant women in the United States in whomserum TSH was measured during the secondtrimester, hypothyroidism was diagnosed in 2.2% ofthe cohort and autoimmune thyroiditis was present in 55% of women with SCH and more than 80% in women with OH (2). Other causes of thyroidinsufficiency include the treatment ofhyperthyroidism (using radioiodine ablation orsurgery) or surgery for thyroid tumors. Ahypothalamic-hypophyseal origin of hypothyroidismis rare and can include lymphocytic hypophysitisoccurring during pregnancy or postpartum (3). It isimportant to remember, however, that on a worldwidebasis the most important cause of thyroidinsufficiency remains iodine deficiency (ID), knownto affect over 1.2 billion individuals.

1.2.2. Clinical and diagnostic features.

Clinical features. Symptoms and signs may raiseclinical suspicion of hypothyroidism duringpregnancy (weight increase, sensitivity to cold, dryskin, etc.), but others may go unnoticed (asthenia,

THE

ENDO

CRIN

E SO

CIET

YS

CLIN

ICAL

GU

IDEL

INES

12

drowsiness, constipation, etc.). Because many womenmay remain asymptomatic, particular attention isrequired from obstetrical care providers for thiscondition to be diagnosed and to evaluate moresystematically thyroid function when women attendthe prenatal clinic for the first time. Only thyroidfunction tests confirm the diagnosis.

Diagnostic features. A serum TSH elevation suggestsprimary hypothyroidism, and serum free T4 levelsfurther distinguish between SCH and OH, dependingon whether free T4 is normal or clearly below normalfor gestational age. Determination of thyroidautoantibodies titersthyroid peroxidase (TPO) andthyroglobulin (TG) antibodies (TPO-Ab and TG-Ab)confirms the autoimmune origin of the disorder (68). The range of normal serum total T4 ismodified during pregnancy under the influence of arapid increase in T4-binding globulin (TBG) levels.Therefore, the nonpregnant total T4 range (512g/dl or 50150 nmol/liter) should be adapted in thesecond and third trimester by multiplying this rangeby 1.5-fold (9, 10). Reference ranges provided by themanufacturers of most free T4 measurement kits havebeen established using pools of nonpregnant normalsera. Such reference ranges are not valid duringpregnancy because free T4 assays are influenced byserum changes (mainly in TBG and serum albumin).Authors have recently proposed to adapt serum freeT4 reference ranges to laboratory-specific ortrimester-specific pregnancy ranges but, so far, noconsensus has been reached on this issue (912).Therefore, we recommend caution in theinterpretation of serum free T4 levels duringpregnancy and also that each laboratory shouldestablish trimester-specific reference ranges forpregnant women.

Serum TSH values are influenced by the thyrotropicactivity of elevated circulating human chorionicgonadotropin concentrations, particularly (but notonly) near the end of the first trimester. Thus, by

Abbreviations: APLA, Antiphospholipid antibodies; ART, assisted reproductive technology; ATD, antithyroid drug; bw, body weight; DM, diabetes mellitus; FNAB, fine needleaspiration; FNAB, FNA biopsy; ID, iodine deficiency; IVF, in vitro fertilization; MMI, methimazole; OH, overt hypothyroidism; PPD, postpartum depression; PPT; postpartumthyroiditis; PTU, propylthiouracil; RAI, radioactive iodine; RNI, recommended nutrient intake; SCH, subclinical hypothyroidism; TAI, thyroid autoimmunity; TBG, T4-binding globulin;TG, thyroglobulin; TG-Ab; TG antibody; TPO, thyroid peroxidase; TPO-Ab, TPO antibody; TRAb, TSH-R antibodies; TSAb, Thyroid-stimulating antibody; TSI, thyroidstimulatingimmunoglobulins; TSH-R, TSH receptor; TV, thyroid volume; UIC, urinary iodine concentration; UIE, urinary iodine excretion; USI, universal salt iodinization.

13

using the classical reference range for serum TSH (0.4 mIU/liter for the lower limit and 4.0 mIU/ literfor the upper limit), one might misdiagnose asnormal women who already have a slight TSHelevation and, conversely, one might suspecthyperthyroidism in normal women who have ablunted serum TSH value (1316). Dashe et al. (14)have recently published a nomogram for serum TSHchanges during pregnancy. They show that 28% ofsingleton pregnancies with a serum TSH greater than2 sd scores above the mean would not have beenidentified by using the nonpregnant serum TSHrange. Other authors have proposed to use trimester-specific reference ranges for serum TSH duringpregnancy (12, 13). An example, illustrated in Fig. 1,shows that the lower normal limit of serum TSH is0.03 mIU/liter in the first and second trimesters andis still reduced to 0.13 mIU/liter in the third trimester.Conversely, serum TSH levels above 2.3 mIU/liter(first trimester) and 3.13.5 mIU/liter (second andthird trimesters) may already be indicative of SCH.

1.3. EVIDENCE

1.3.1. Repercussions of hypothyroidism onpregnancy: maternal aspects. There is a knownassociation between hypothyroidism and decreasedfertility, although hypothyroidism does not precludethe possibility to conceive. In a study by Abalovich etal. (1), 34% of hypothyroid women became pregnantwithout treatment: 11% of them had OH and 89%SCH. When hypothyroid women become pregnantand maintain the pregnancy, they carry an increasedrisk for early and late obstetrical complications, such asincreased prevalence of abortion, anemia, gestationalhypertension, placental abruption, and postpartumhemorrhages. These complications are more frequentwith OH than with SCH and, most importantly,adequate thyroxine treatment greatly decreases therisk of a poorer obstetrical outcome (1, 1517).

1.3.2. Repercussions of hypothyroidism onpregnancy: fetal aspects. Untreated maternal OH isassociated with adverse neonatal outcomes including

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

MFIG. 1. Median and 95% confidence levels for TSH during pregnancy.

premature birth, low birth weight, and neonatalrespiratory distress. Increased prevalence of fetal andperinatal death has also been described, although ithas not been confirmed in all studies. Obstetricaladverse effects such as gestational hypertension mayalso contribute to the overall increase in neonatalrisks (1, 2, 1518). Though less frequent than withOH, complications have also been described innewborns from mothers with SCH. Casey et al. (19)screened pregnant women before 20 wk gestation andreported a doubling of the rate of preterm delivery inthose with SCH. Stagnaro-Green et al. (20)compared the thyroid status of women with pretermdelivery to matched controls who delivered at termand showed that women with very preterm deliveries(before 32 wk) had a 3-fold increase in SCH. In arecent prospective randomized intervention trial byNegro et al. (21), the authors reported a significantdecrease in the rate of preterm delivery amongthyroid-antibody-positive women who had beentreated with thyroxine, compared with thyroid-antibody-positive women who did not receivethyroxine administration and in whom thyroidfunction showed a gradual evolution toward SCHduring gestation.

1.3.3. Maternal thyroid hormones and fetal braindevelopment. A large body of evidence stronglysuggests that thyroid hormone is an important factorcontributing to normal fetal brain development(2224). At early gestational stages, the presence ofthyroid hormones in fetal structures can only beexplained by transfer of maternal thyroid hormones tothe fetal compartment, because fetal production ofthyroid hormones does not become efficient untilmid-gestation. Thyroid hormone and specific nuclearreceptors are found in fetal brain at 8 wk afterconception (25). Physiological amounts of free T4 arefound in the coelomic and amniotic fluids bathing thedeveloping embryo in the first trimester (26). Studiesof different cerebral areas in human fetuses indicatedthe presence of increasing concentrations of T4 andT3 by 1118 wk after conception (27). The ontogenicpatterns of thyroid hormone concentrations and theactivity of iodothyronine deiodinases show a complexinterplay between the changing activities of the

specific D2 and D3 iodothyronine deiodinases duringgestation. This dual enzymatic system is interpretedto represent a regulatory pathway that fine-tunes theavailability of T3 required for normal braindevelopment and avoids, at the same time, thepresence of excessive amounts of T3 (2830).

1.3.4. Clinical studies on the role of maternalhypothyroidism for the psychoneurological outcomein the progeny. Because of the heterogeneity of whatis commonly referred to as gestationalhypothyroidism, different clinical conditions mustbe considered. Thyroid insufficiency varies widelywith regard to time of onset (first trimester vs. later),degree of severity (SCH vs. OH), progressiveaggravation with gestation time (depending on thecause), and adequacy of treatment. To reconcile thesevariable clinical conditions into a global view of therepercussions of maternal hypothyroidism on theprogeny is difficult. However, a common patternclearly emerges. Overall, the results showed that therewas a significantly increased risk of impairment inneuropsychological developmental indices, IQ scores,and school learning abilities in the offspring ofhypothyroid mothers.

Three decades ago Evelyn Man and colleagues(3133) published a series of articles suggesting thatchildren born to mothers with inadequately treatedhypothyroidism had significantly reduced IQs.However, the first large-scale prospective study on theoutcome of children born to mothers withhypothyroidism during pregnancy was reported byHaddow et al. in 1999 (34). In this study, the severityof hypothyroidism varied from OH to probable SCHamong the women whose children were investigatedat school age. The main finding on extensiveneuropsychological testing was that children born tountreated hypothyroid women had, on the average,an IQ score that was fully 7 points below the mean IQof children born to healthy women and thyroxine-treated women. Furthermore, there were three timesas many children with IQs that were 2 sd scores belowthe mean IQ of controls in the children born tountreated hypothyroid women. The study indicatedthat undisclosed and untreated hypothyroidism (and

THE

ENDO

CRIN

E SO

CIET

YS

CLIN

ICAL

GU

IDEL

INES

14

15

probable SCH) during pregnancy was associated witha risk of a poorer outcome in the progeny and a 3-foldincreased predisposition for having learningdisabilities.

Although still unpublished, a large set of data werereported at the 2004 annual meeting of the AmericanThyroid Association by Rovet et al. (35). The interestof this remarkable work is double. First, the authorsinvestigated children born to women who had beentreated for hypothyroidism during pregnancy, but inwhom thyroxine administration was suboptimal(mean TSH between 5 and 7 mIU/liter). Second, thechildren were followed up and tested with extremelyrefined techniques up to 5 yr of age. Results were thatsome components of intelligence were affected,whereas others were not. At preschool age, the study-case children had a mild reduction in globalintelligence that was inversely correlated with thirdtrimesters maternal TSH. On the other hand, therewas no negative impact on language, visual-spatialability, fine motor performance, or preschool ability.The conclusion was that the offspring of women withsuboptimally treated maternal hypothyroidism may

be at risk for subtle and selective clinically relevantcognitive deficits, which depend specifically onseverity and timing of inadequate maternal thyroidhormone availability.

A Dutch study investigated the developmentaloutcome in children born to women with early (firsttrimester) isolated low T4 levels (i.e. a serum free T410.4 pmol/liter, in the lowest 10th decile of normalpregnant T4 values, with normal TSH) (36). Resultssuggested that early maternal low free T4 wasassociated with a lower developmental index in thechildren at approximately 10 months of age. Thesame authors later published a second study based onsimilar selection criteria, but with a larger cohort andmore refined motor and mental evaluations in infantsaged 1 and 2 yr (37). Results were that children bornto mothers with prolonged low T4 (until wk 24 orlater) showed an 8-to 10-point deficit for motor andmental development. Of interest, infants of womenwith early low T4, whose free T4 level recoveredspontaneously to normal later in gestation, had anormal development, suggesting that prolonged lowT4 was needed to impair fetal neuro-development.

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

TABLE 1. Neuropsychiatric and intellectual deficits in infants and schoolchildren born to mothers residing inconditions of mild to moderate ID

Region Tests Main findings First author(s), date (Ref.)

Spain Locally adapted: Bayley, Lower psychomotor and mental Bleichrodt, 1989 (42)McCarthy, Cattell development

Italy (Sicily) Bender-Gestalt Low perceptual integrative motor ability Vermiglio, 1990 (45)and neuromuscular and neurosensorial abnormalities

Italy (Tuscany) Wechsler Raven Low verbal IQ, perception, and motor Fenzi, 1990 (43)and attentive functions

Italy (Tuscany) WISC reaction time Lower velocity of motor response to Vitti, 1992 (46), visual stimuli Aghini-Lombardi, 1995 (40)

India Verbal, pictorial learning Lower learning capacity Tiwari, 1996 (44)tests, tests of motivation

Iran Bender-Gestalt, Raven Retardation in psychomotor development Azizi, 1993 (41)

Table modified from Glinoer and Delange (38).

Neural development in ID. The consequences ofmaternal hypothyroidism on the progeny must beconsidered separately because ID exposes bothmother and fetus to thyroid underfunction (38). In ameta-analysis of 19 studies of infants outcome inrelation to ID, a significant downward shift of thefrequency distribution of IQs was evidenced, with amean 13.5 IQ points reduction in neuro-motor andcognitive functions (39). Because that meta-analysisencompassed conditions with more or less severe ID,the results cannot be fully extrapolated to mild-moderate ID. For this reason, the main results ofseven studies (reported between 1989 and 1996) thathave investigated the late outcome in children bornto mothers with mild-moderate ID are summarized inTable 1 (4046) (see on page 15). Finally, a recentpublication on the outcome of children born tomothers with ID during pregnancy, carried out inSicily in an area with mild-moderate ID, indicatedthat the children had a greater than 10-point averagedeficit in global IQ. Furthermore, the study alsopointed to attention deficit and hyper-activity disorder, which was found in 69% of thechildren born to mothers who had gestationalhypothyroxinemia (47).

11..44.. TTHHEERRAAPPEEUUTTIICC AASSPPEECCTTSS

The administration of levothyroxine is the treatmentof choice for maternal hypothyroidism, if the iodinenutrition status is adequate. Hypothyroid pregnantwomen require larger thyroxine replacement dosesthan do nonpregnant patients, and women whoalready take thyroxine before pregnancy usually needto increase their daily dosage by, on average, 3050%above preconception dosage. Several reasons explainthe incremented thyroid hormone requirements: therapid rise in TBG levels resulting from thephysiological rise in estrogen, the increaseddistribution volume of thyroid hormones (vascular,hepatic, fetal-placental unit), and finally theincreased placental transport and metabolism ofmaternal T4 (7, 4850).

Thyroxine treatment should be initiated with a doseof 100150 g thyroxine/d or titrated according tobody weight (bw). In nonpregnant women, the fullreplacement thyroxine dose is 1.72.0 g/kg bwd.During pregnancy, because of the increasedrequirements, the full replacement thyroxine doseshould be increased to 2.02.4 g/kg bwd. In initiallysevere hypothyroidism, therapy may be initiated bygiving for the first few days a thyroxine dosecorresponding to two times the estimated finalreplacement daily dose, to rapidly normalize theextrathyroidal thyroxine pool (7, 5153). In womenwho already receive thyroxine before conception, theneed to adjust the preconception thyroxine dailydosage becomes manifest as early as by 46 wkgestation, hence justifying the adaptation ofthyroxine replacement to ensure that maternaleuthyroidism is maintained during early pregnancy.An alternative recommended by some thyroidologistsis to anticipate the expected increase in serum TSHby raising the thyroxine dose before conception or assoon as pregnancy has been confirmed. It is importantto note that 25% of hypothyroid women who are ableto maintain a normal serum TSH level in the firsttrimester, and 35% of those who maintain a normalserum TSH level until the second trimester withoutincreasing their daily dosage, will still require anincrement in thyroxine replacement during lategestation to maintain a euthyroid status (7, 54, 55).The magnitude of thyroxine increment duringpregnancy depends primarily on the etiology ofhypothyroidism, namely the presence or absence ofresidual functional thyroid tissue.

Women without residual functional thyroid tissue(after radioiodine ablation, total thyroidectomy, ordue to congenital agenesis of the gland) require agreater increment in thyroxine dosage thanwomen with Hashimotos thyroiditis, who usuallyhave some residual thyroid tissue. As a simple rule of thumb, the increment in thyroxine dosecan be based on the initial degree of TSHelevation. For women with a serum TSH between510 mIU/liter, the average increment inthyroxine dosage is 2550 g/d; for those with aserum TSH between 10 and 20 mIU/liter, 5075

THE

ENDO

CRIN

E SO

CIET

YS

CLIN

ICAL

GU

IDEL

INES

16

17

g/d; and for those with a serum TSH greater than20 mIU/liter, 75100 g/d.

Serum free T4 and TSH levels should be measuredwithin 1 month after the initiation of treatment. Theoverall aim is to achieve and maintain normal free T4and TSH at levels normal for pregnancy throughoutgestation. Ideally, thyroxine treatment should betitrated to reach a serum TSH value less than 2.5mIU/liter. As already alluded to, because it issometimes difficult to correctly interpret the results offree T4 and TSH measurements in the context ofpregnancy, it is useful to optimize the monitoring oftreatment during pregnancy by establishinglaboratory-specific reference ranges for serum free T4and trimester-specific reference ranges for serumTSH. Once the thyroid function tests have beennormalized by treatment, they should be monitoredevery 68 wk. If thyroid function tests remainabnormal, thyroxine dosage should be adjusted andtests repeated after 30 d, and so on untilnormalization of thyroid function tests. Afterparturition, most patients need to decrease thyroxinedosage received during pregnancy, over a period ofapproximately 4 wk postpartum. It should also beremembered that women with evidence of thyroidautoimmunity are at risk of developing postpartumthyroiditis (PPT), a syndrome that may justifydifferences in the pre-and post-pregnancy thyroxinerequirements. It is therefore important to continuemonitoring thyroid function tests for at least 6 months after delivery (7, 56).

In pregnant women in whom hypothyroidism has notbeen diagnosed until after the first trimester, there areindications that the offspring may suffer fromimpairment in final intellectual and cognitiveabilities. The present consensus is to maintain theongoing pregnancy, while rapidly normalizingmaternal thyroid function by the administration ofthyroxine. However, despite thyroxine treatment, it isimpossible to fully reassure the future parents aboutpotential brain damage that may have occurred iflongstanding intrauterine severe hypothyroidism hasbeen present.

11..55.. RREECCOOMMMMEENNDDAATTIIOONNSS

1.5.1. Both maternal and fetal hypothyroidism areknown to have serious adverse effects on the fetus.Therefore maternal hypothyroidism should beavoided. For OH, the USPSTF recommendationlevel is A; evidence is fair. Targeted case finding isrecommended at the first prenatal visit. The USPSTFrecommendation level is B; evidence is fair)(GRADE 2| ) (1, 5, 34, 57).

1.5.2. If hypothyroidism has been diagnosed beforepregnancy, we recommend adjustment of thepreconception thyroxine dose to reach beforepregnancy a TSH level not higher than 2.5 mIU/liter.The USPSTF recommendation level is I; evidence ispoor) (GRADE 2| ) (9, 12, 13).

1.5.3. The thyroxine dose often needs to beincremented by 46 wk gestation and may require a3050% increment in dosage. The USPSTFrecommendation level is A; evidence is good)(GRADE 1| .) (50, 51).

1.5.4. If OH is diagnosed during pregnancy, thyroidfunction tests should be normalized as rapidly aspossible. Thyroxine dosage should be titrated torapidly reach and thereafter maintain serum TSHconcentrations of less than 2.5 mI/liter in the firsttrimester (or 3 mIU/liter in second and thirdtrimesters) or to trimester-specific normal TSH ranges.Thyroid function tests should be remeasured within3040 d. The USPSTF recommendation level is A;evidence is good (GRADE 1| ) (10, 11, 13).

1.5.5. Women with thyroid autoimmunity (TAI)who are euthyroid in the early stages of pregnancy areat risk of developing hypothyroidism and should bemonitored for elevation of TSH above the normalrange. The USPSTF recommendation level is A;evidence is fair) (GRADE 1| ) (54, 57).

1.5.6. SCH (serum TSH concentration above theupper limit of the reference range with a normal freeT4) has been shown to be associated with an adverseoutcome for both the mother and offspring.

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

Thyroxine treatment has been shown to improveobstetrical outcome, but has not been proved tomodify long-term neurological development in theoffspring. However, given that the potential benefitsoutweigh the potential risks, the panel recommendsthyroxine replacement in women with SCH. Forobstetrical outcome, the USPSTF recommendationlevel is B; evidence is fair) (GRADE 1| ); forneurological outcome, the USPSTF recommendationlevel is I; evidence is poor ( ) (1, 16, 19, 49).

1.5.7. After delivery, most hypothyroid women needto decrease the thyroxine dosage they received duringpregnancy. The USPSTF recommendation level is A;evidence is good) (GRADE 1| ) (56).

11..66.. BBIIBBLLIIOOGGRRAAPPHHYYIITTEEMMSS RREELLAATTEEDD TTOORREECCOOMMMMEENNDDAATTIIOONNSS

Abalovich M, Gutierrez S, Alcaraz G, Maccallini G,Garca A, Levalle O 2002 Overt and subclinicalhypothyroidism complicating pregnancy. Thyroid12:6368

The study concerns 150 pregnancies,corresponding to 114 women with primary hypo-thyroidism. Fifty-one pregnancies (34%) wereconceived under hypothyroidism (16 with overt and35 with SCH) and 99 pregnancies undereuthyroidism with thyroxine therapy. Whenthyroxine treatment was not adequately adapted, aspontaneous miscarriage occurred in 60% of womenwith OH and 71% of women with SCH. Furthermore,premature delivery was observed in 20% of womenwith OH and 7% with SCH. Conversely, whenthyroxine treatment was adequate and thyroidfunction remained normal, 100% of women with OHand 91% of women with SCH carried pregnancies toterm and no abortion was observed in any of thegroups. The authors concluded that the outcome ofpregnancy did not depend on whether hypo-thyroidism was initially overt or subclinical, butprimarily on the adequacy of the thyroxine treatment.

Alexander EK, Marqusee E, Lawrence J, Jarolim P,Fischer GA, Larsen PR 2004 Timing and magnitude

of increases in levothyroxine requirements duringpregnancy in women with hypothyroidism. N Engl JMed 351:241249

Women with hypothyroidism who were planningpregnancy were observed prospectively before andthroughout the pregnancies. Twenty pregnanciesoccurred in 19 women that resulted in 17 full-termbirths. An increase in the levo-thyroxine dose wasnecessary during 17 of 20 pregnancies. On theaverage, levothyroxine requirement increased by 47%during the first half of pregnancy. The median timingof this increase corresponded to a gestational age of 8wk and a plateau was observed by wk 16. Theincreased thyroxine dose was required until delivery.

Allan WC, Haddow JE, Palomaki GE, Williams JR,Mitchell ML, Hermos RJ, Faix JD, Klein RZ 2000Maternal thyroid deficiency and pregnancycomplications: implications for population screening.J Med Screen 7:127130

Prospective population study of 9,471 pregnantwomen in whom serum TSH was measured during thesecond trimester: hypothyroidism was diagnosed in2.2% of them. Autoimmunity features correspondingto chronic thyroiditis were associated with thyroiddysfunction in 55% of women with SCH (serumTSH, 610 mU/liter) and more than 80% of womenwith OH (serum TSH, 11200 mU/liter).Furthermore, the rate of fetal death was increased 4-fold (3.8% vs. 0.9%) in mothers with hypo-thyroidism compared with the control populationwith a normal serum TSH. The authors concludedthat from the second trimester onward, the majoradverse obstetrical outcome associated with raisedTSH in the general population was an increased rateof fetal death. The authors also speculated that ifthyroid replacement treatment avoided this problem,this would be another reason to consider populationscreening for thyroid dysfunction and autoimmunity.

Caixas A, Albareda M, Garcia-Patterson A,Rodriguez-Espinoza J, De Leiva A, Corcoy R 1999Postpartum thyroiditis in women with hypo-thyroidism antedating pregnancy. J Clin EndocrinolMetab 84:40004005

THE

ENDO

CRIN

E SO

CIET

YS

CLIN

ICAL

GU

IDEL

INES

18

19

Forty-one women (31 receiving levothyroxinereplacement therapy and 10 receiving suppressivetherapy for thyroid carcinoma) were followed duringthe first year after delivery. A control group of 31nonpregnant women with hypothyroidism (n = 21) orthyroid carcinoma (n = 10) were also followed duringa similar period. A total of 23 of 41 (56%) haddiscordant requirements at follow up after delivery vs.only three of 31 in the control group (P < 0.001). Therate of patients with discordant prepregnancy vs.postpartum levothyroxine doses was higher in thenoncarcinoma subgroup (68% vs. 20%; P < 0.01),whereas in the control group both subgroupsdisplayed a similar rate of discordance. This studydocuments that women with hypothyroidismantedating pregnancy display changes in levo-thyroxine requirements in the first year after delivery,suggesting the role of PPT.

Casey B, Dashe JS, Wells CE, McIntire DD, ByrdW, Leveno KJ, Cunningham FG 2005 Subclinicalhypothyroidism and pregnancy outcomes. ObstetGynecol 105:239245

Serum TSH and free T4 were measured in 25,756women enrolled for prenatal care and who delivered asingleton infant (from November 2000 to April2003). Among the cohort, a total of 17,298 (67%)women were enrolled at 20 wk gestation (or before),and the overall prevalence of SCH was 2.3%. In thewomen with SCH compared with the controls,pregnancy was 3-fold more likely to be complicatedby placental abruption (relative risk, 3.0; 95%confidence interval, 1.18.2). Preterm birth, definedas a delivery at (or before) 34 wk gestation, wasalmost 2-fold more frequent in women with SCH(relative risk, 1.8; 95% confidence interval, 1.12.9).

Dashe JS, Casey BM, Wells CE, McIntire DD,Byrd EW, Leveno KJ, Cunningham FG 2005Thyroid-stimulating hormone in singleton and twinpregnancy: importance of gestational age-specificreference ranges. Obstet Gynecol 106: 753757

Serum TSH reference range evaluated in 13.599singleton and 132 twin pregnancies, with individualvalues converted into a nomogram based onmultiples of the median TSH at each week of

pregnancy. Serum TSH decreased significantly duringfirst trimester, with an even greater decrease (by 0.4 mIU/liter) in twin, compared with singletonpregnancies. Had a classical (nonpregnant, 0.44.0 mIU/liter) range been used rather than thenomogram, 28% of 342 singletons with TSH greater2 sd scores above the mean would not have beenidentified. The upper TSH limit (97.5th percentilebased on the nomogram) was approximately 3.0 toapproximately 3.5 mIU/liter between wk 10 and 30.The TSH nadir was prolonged until late into thesecond trimester.

Demers LM, Spencer CA 2002 Laboratory supportfor the diagnosis and monitoring of thyroid disease.National Academy of Clinical BiochemistryLaboratory Medicine Practice Guidelines. Availableat: http://www.nacb.org/ lmpg/thyroid_lmpg_pub.stm

The authors consider that, until better databecome available, the nonpregnant upper limit forserum TSH set at 2.5 mU/liter is also appropriate forthe first trimester of pregnancy. Inasmuch as between-method biases for the measurement of serum TSH areminimal, there is no need for developing method-specific TSH ranges.

Glinoer D, Rihai M, Gru n JP, Kinthaert J 1994Risk of subclinical hypothyroidism in pregnantwomen with asymptomatic autoimmune thyroiddisorders. J Clin Endocrinol Metab 79:197204

Prospective cohort study in Belgium of 1660unselected consecutive pregnant women whounderwent systematic screening for thyroid functionand autoimmunity (TAI). Eighty-seven women werefound to have positive thyroid antibodies withnormal thyroid function tests (prevalence, 5.3%) inearly pregnancy and were followed sequentiallyduring gestation without treatment. Another group of20 women, identified to have both positive antibodiesand abnormal thyroid function tests (prevalence,1.2%), were followed separately, because theyrequired medical treatment. Thus, the overallprevalence of TAI in the cohort was 6.5%. In thegroup of women with TAI and normal thyroidfunction tests, main results were the following. In thefirst trimester, despite having TSH within the

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

reference range limits, mean serum TSH was alreadysignificantly higher in TAI-positive women comparedwith controls. Furthermore, as gestation advanced toterm, a significant fraction of these womenprogressively developed thyroid deficiency, with 40%of the women having a serum TSH above 3 mU/literand 16% above 4 mU/liter at delivery. Finally, whencomparing serum free T4 levels between TAI-positivewomen and controls at delivery, it was found that,despite being euthyroid during early gestation, theywere unable to maintain euthyroidism at the end ofpregnancy, because at delivery their mean serum freeT4 was not only 30% below that of controls, but inaddition their average serum free T4 was at the lowerlimit of normality (10 pm/liter) and 42% of themwere in the range of hypothyroid values. Theconclusion was that women with asymptomatic TAIwho are euthyroid in early pregnancy carry asignificant risk of developing hypothyroidismprogressively during gestation.

Haddow JE, Palomaki GE, Allan WC, Williams JR,Knight GJ, Gagnon J, OHeir CE, Mitchell ML,Hermos RJ, Waisbren SE, Faix JD, Klein RZ 1999Maternal thyroid deficiency during pregnancy andsubsequent neuropsychological development of thechild. N Engl J Med 341:549555

Sixty-two children (aged ~8 yr) wereinvestigated prospectively and matched with 124control children from the same schools, and allunderwent extensive neuropsychological testing forIQ and school-learning abilities. The study childrenwere born to mothers who had been identifiedretrospectively to have had hypothyroidism aroundmid-gestation, with elevated serum TSH and serumfree T4 levels that were on average 30% below themean free T4 of control mothers. Main results werethat the full-scale IQ scores of children born tohypothyroid untreated mothers averaged 7 pointslower than the mean IQ score of children born tocontrol mothers (P = 0.005). Furthermore, threetimes as many children from mothers with untreatedthyroid deficiency (19% vs. 5%) had IQ scores thatwere 2 sd scores below the mean IQ of the controls(that is

21

Leung AS, Millar LK, Koonings PP, Montoro M,Mestman JH 1993 Perinatal outcome in hypothyroidpatients. Obstet Gynecol 81:349353

The study concerns the perinatal outcome in 68hypothyroid patients, 23 with OH and 45 with SCH.Gestational hypertension (namely eclampsia,preeclampsia, and pregnancy-induced hypertension)was significantly more common in women with OH(22%) and SCH (15%) than in the controlpopulation (8%). In addition, 36% of the overt and25% of the subclinical hypothyroid patients who hadremained hypothyroid until delivery developedgestational hypertension. The low birth weightsobserved in both women with OH and SCH wassecondary to premature delivery due to gestationalhypertension. The authors concluded thatnormalization of thyroid function tests may preventgestational hypertension and its consequentcomplications in hypothyroid pregnant women.

Mandel SJ, Larsen PR, Seely EW, Brent GA 1990Increased need for thyroxine during pregnancy inwomen with primary hypothyroidism. N Engl J Med323:9196

The authors analyzed retrospectively thyroidfunction in 12 pregnant women who receivedthyroxine treatment for primary hypothyroidism.Because of elevated serum TSH levels, the thyroxinedosage had to be increased in nine of 12 patients. Themean thyroxine dosage was 100 g/d beforepregnancy and was increased to 150 g/d duringpregnancy (P < 0.01). Among the three patients whodid not require an increase in thyroxine dosage, twohad a low serum TSH before pregnancy, suggestingexcessive replacement. During postpartum, the meanthyroxine dose was decreased to 117 g/d (P < 0.01,compared with pregnancy dosage). These resultsindicate the need to increase the thyroxine dosage inthe majority of with primary hypothyroidism duringpregnancy.

Panesar NS, Li CY, Rogers MS 2001 Referenceintervals for thyroid hormones in pregnant Chinesewomen. Ann Clin Biochem 38:329332

To establish gestation-related reference intervalsfor thyroid hormones in a Chinese population, the

authors prospectively studied 343 healthy pregnantwomen and 63 nonpregnant controls. TSH, free T4,and free T3 were measured and the median, 2.5th and97.5th percentiles at 4-wk intervals were calculated.Free T3 decreased during pregnancy, whereas free T4initially increased, peaking between 9 and 13 wk andthen decreased. TSH mirrored changes in free T4. Theauthors concluded that the gestation-related referenceintervals for thyroid hormones should alleviate themisinterpretation of thyroid function in pregnancy.

Pop V, Brouwers EP, Vader HL, Vulsma T, VanBaar AL, De Vijlder JJ 2003 Maternal hypo-thyroxineaemia during early pregnancy andsubsequent child development: a 3-year follow-upstudy. Clin Endocrinol (Oxf) 59:282288

Prospective follow-up study of pregnant womenand their infants up to the age of 2 yr. The womenwere selected to have isolated hypothyroxineaemiaduring early pregnancy, with a serum free T4 belowthe lowest tenth percentile of control pregnantwomen but with a normal serum TSH. Main resultswere that the infants born to mothers withhypothyroxineaemia at 12 wk gestation (n = 63) haddelayed mental and motor function both at the agesof 1 and 2 yr compared with control infants (n = 62),assessed by the means of the Bailey Scales of InfantDevelopment. The study also showed that infantsborn to mothers with hypothyroxinemia in the firsttrimester but who spontaneously recovered normalfree T4 levels during later gestation (at 24 and 32 wk)had a normal development, thus indicating that earlyhypothyroxineaemia alone, when it was notprolonged during later gestational stages, did notadversely affect the infants motor and mentaldevelopment.

Soldin OP, Tractenberg RE, Hollowell JG, JonklaasJ, Janicic N, Soldin SJ 2004 Trimester-specificchanges in maternal thyroid hormone, TSH andthyroglobulin concentrations during gestation: trendsand associations across trimesters in iodinesufficiency. Thyroid 14:10841090

The authors describe the interrelationships ofthyroid tests based on trimester-specific con-centrations in healthy, iodine-sufficient pregnant

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

women across trimesters and postpartum. Trimester-specific total T3, free T4, TSH, and TGconcentrations were significantly different betweenfirst and third trimesters (all P < 0.05). Second andthird trimester values were not significantly differentfor free T4, TSH and TG, although total T3 wassignificantly higher in the third, relative to thesecond trimester. Total T4 was not significantlydifferent at any trimester. These results show that iniodine sufficient women serum total T3, free T4, TSH,and TG tend to change throughout the course ofpregnancy, whereas total T4 after the first trimesterdoes not.

SSEECCTTIIOONN 22.. MMAANNAAGGEEMMEENNTT OOFFHHYYPPEERRTTHHYYRROOIIDDIISSMM:: MMAATTEERRNNAALLAANNDD FFEETTAALL AASSPPEECCTTSS

22..22.. BBAACCKKGGRROOUUNNDD

2.2.1 Maternal hyperthyroidism. The prevalence ofhyperthyroidism in the United States is approximately1% (0.4% clinical and 0.6% subclinical) according tothe NHANES survey (58). The most common causeof hyperthyroidism is Graves disease, which is 5- to10-fold more common in women, with a peakincidence during the reproductive age. Thus,hyperthyroidism in pregnancy is not rare and itsreported prevalence ranges from 0.1% to 0.4%, withGraves disease accounting for 85% of cases (5961).Single toxic adenoma, multinodular toxic goiter, andthyroiditis comprise most of the remaining casesduring pregnancy, whereas significant gestationalthyrotoxicosis, factitious thyrotoxicosis, andhydatidiform molar disease are uncommon (62).

As in other autoimmune diseases, the activity level ofGraves disease may fluctuate during gestation, withexacerbation during the first trimester and gradualimprovement during the later half. Patients with

THE

ENDO

CRIN

E SO

CIET

YS

CLIN

ICAL

GU

IDEL

INES

22

Graves disease may also experience an exacerbationshortly after delivery (63). Rarely, labor, cesareansection, and infections may aggravate hyper-thyroidism and even trigger a thyroid storm (6466).

Autoimmune thyroiditis occurs in up to 10% ofwomen in the reproductive age (67). Generally, theresult is hypothyroidism, although a hyperthyroidphase of Hashimotos thyroiditis and silent thyroiditismay both occur. PPT occurs after up to 10% of allpregnancies and may have a hyperthyroid phase,usually within the first month or two (68). BecausePPT may begin between 6 wk to 6 months afterdelivery, and occasionally as long as 1 yr later, it is notuncommon to find women who have started on theirnext pregnancy within this time frame. Furthermore,postpartum thyroiditis may be triggered by amiscarriage occurring as early as 6 wk gestation andsuch women often conceive again within a fewmonths (69). Because the hyperthyroid phase ofthyroiditis is often followed by a hypothyroid phase,and because hypothyroidism is an important risk forfetal development, careful sequential monitoring isnecessary to detect and treat the hypothyroid phase ofthis illness.

Patients with gestational thyrotoxicosis present in themid to late first trimester, often with hyperemesis.Usually classic hyperthyroid symptoms are absent orminimal, except for weight loss, which may be a resultof vomiting and poor nutrition (70, 71). Gravesdisease must be differentiated from gestationthyrotoxicosis so that a decision about ATD therapycan be reached. (See Gestational Hyperemesis andHyperthyroidism, Section 3.)

2.2.2. Fetal thyroid function. The fetal thyroidbegins concentrating iodine at 1012 wk gestationand is under control of fetal pituitary TSH byapproximately 20 wk gestation. Fetal serum levels ofTSH, TBG, free T4, and free T3 increase throughoutgestation, reaching mean adult levels atapproximately 36 wk (72). TSH does not cross theplacenta. However, clinically significant amounts ofmaternal T4 do cross the placenta. In neonates withcongenital hypothyroidism, enough maternal thyroid

23

hormone crosses the placenta to prevent stigmata ofhypothyroidism at birth and to maintain cord bloodthyroid hormone levels at near 50% of normal (73).In addition, TSH-releasing hormone (TRH), iodine,TSH receptor (TSH-R) antibodies, and antithyroiddrugs (ATDs) cross the placenta readily.

Fetal and neonatal risks of maternal hyperthyroiddisease are related to the disease itself and/or to themedical treatment of the disease. Inadequatelytreated maternal thyrotoxicosis is associated with anincreased risk of medically indicated preterm delivery,intrauterine growth restriction and low birth weight,preeclampsia, congestive heart failure, andintrauterine death (74, 75). In addition, over-treatment of the mother with thioamides can result iniatrogenic fetal hypothyroidism (76).

2.3. EVIDENCE

2.3.1. Diagnosis of maternal hyperthyroidism.Because nonspecific symptoms of hyperthyroidismsuch as tachycardia, warm moist skin, tremor, andsystolic murmur may be mimicked by normalpregnancy, the presence of classic thyroidophthalmopathy, a significant goiter, or pretibialmyxedema (while rare) may point to a diagnosis oftrue Graves disease. A careful physical examinationshould be performed in all patients.

Patients suspected of having hyperthyroidism requiremeasurement of serum TSH, T4, T3 levels, andthyroid receptor antibodies. However, interpretationof thyroid function tests must be made in relation tothe hCG-mediated decrease in serum TSH levels andthe increase in TBG concentrations that occur duringpregnancy (7779). In the normal pregnant woman,TSH levels typically fall in the mid to late firsttrimester coincident with rising hCG levels. Themedian serum TSH level in the first half of pregnancyis approximately 0.8 U/ml, with 95% confidenceinterval lower limit of 0.03 U/ml (13). Therefore,subnormal serum TSH levels in the first half ofpregnancy should not be interpreted as diagnostic ofhyperthyroidism (see Fig. 1, Section 1.2.2).

Free hormone measurements by one-or two-stepanalog methods may perform differently duringpregnancy because of the known alterations in twothyroid hormone binding proteins, increased serumTBG levels, and decreased serum albuminconcentrations (1113, 79, 80). Measurement of freeT4 hormone concentration by equilibrium dialysis iscostly and not universally available. However, itappears that serum free T4 levels measured by bothequilibrium dialysis and automated methodologies areincreased over nonpregnant reference ranges in thefirst trimester and then decrease so that by the thirdtrimester, the 95% confidence interval for serum freeT4 concentration may be 30% lower than thenonpregnant reference range (11, 80). For practicalpurposes, keeping the free T4 in the uppernonpregnant normal range is appropriate.Alternatively, because the changes in total T4 levelsduring gestation are predictable, the nonpregnantreference limits just need to be adjusted by a factor of1.5 to determine the normal range for pregnancy(81).

Up to 60% of women with hyperemesis gravidarumhave a subnormal TSH and nearly 50% have anelevated free T4 concentration (70). (See GestationalHyperemesis and Hyperthyroidism, Section 3.) Insituations where doubt exists, measurement of serumtotal T3 concentration and T3 resin uptake may behelpful, as only 12% of women with hyperemesisgravidarum have an elevated free T3 index (70).

Most patients with Graves disease will havedetectable TSH-R antibodies (TRAb) (12).Measurement of TRAb may also help to distinguishGraves disease from gestational thyrotoxicosis in thefirst trimester as TRAb are negative in gestationalhyperthyroidism. Because Graves disease tends toundergo immunological remission after the latesecond trimester (63), detection of TRAb maydepend upon gestational age at measurement.

2.3.2. Diagnosis of fetal hyperthyroidism. Althoughfetal hyperthyroidism requiring treatment is rare(under 0.01% of pregnancies), it should be consideredpossible in any woman with a past or current history

MAN

AGEMEN

T OF TH

YROID DYSFU

NCTIO

N DU

RING PREGN

ANCY AN

D POSTPARTU

M

of Graves disease. The likelihood of developing fetalhyperthyroidism requiring treatment is related to thelevel of maternal stimulating TRAb levels, medicaltreatment of maternal disease, and patient history.Fetal hyperthyroidism can be associated with growthrestriction, fetal tachycardia, fetal goiter, fetalhydrops, preterm delivery, and fetal death (8286).The diagnosis is suggested by fetal tachycardia,intrauterine growth restriction, fetal goiter, fetalcardiac failure, or fetal hydrops. Treatment may begiven on the presumptive diagnosis, but definitivediagnosis (if needed) requires umbilical cord bloodsampling for fetal thyroid function and carries asignificant risk to the fetus (8790). Depending onthe gestational age at presentation, umbilical cordblood sampling in a fetus of a mother with Gravesdisease with the above signs or symptoms may bewarranted if the diagnosis cannot be adequatelyinferred on clinical grounds and if the informationwould change management.

Symptomatic neonatal hyperthyroidism should beconsidered an emergency and treated appropriately.Neonatal hyperthyroidism is typically due totransplacental passage of maternal TRAb, butactivating mutations of TSH-R or of G proteins, andde novo Graves disease in the neonate, should beconsidered in the differential diagnosis (9193).

2.3.4. Adverse effects of maternal hyperthyroidism:pregnancy outcome. Lack of control of hyper-thyroidism is associated with adverse pregnancyoutcomes (62). The risk of complications for bothmother and fetus is related to the duration andcontrol of maternal hyperthyroidism. The highestincidence of complications occurred in women withthe poorest control and the lowest incidence in thosewith adequate treatment (62, 74, 75). Inadequatelytreated maternal thyrotoxicosis is associated with anincreased risk of medically indicated pre-termdelivery. In one study 88% of the untreated,compared with 25% of the partially treated and 8% ofthe adequately treated mothers, had a medicallyindicated preterm delivery (75). In addition,untreated women are twice as likely to developpreeclampsia during pregnancy as are women