!!!!use of psychotropic drugs during pregnancy and breast feeding

8/17/2019 !!!!Use of Psychotropic Drugs During Pregnancy and Breast Feeding

1/28

Clinical overview

Use of psychotropic drugs during pregnancyand breast-feeding

Larsen ER, Damkier P, Pedersen LH, Fenger-Gron J, Mikkelsen RL,Nielsen RE, Linde VJ, Knudsen HED, Skaarup L, Videbech P. Use of psychotropic drugs during pregnancy and breast-feeding.

Objective: To write clinical guidelines for the use of psychotropic drugsduring pregnancy and breast-feeding for daily practice in psychiatry,obstetrics and paediatrics.Method: As we wanted a guideline with a high degree of consensus

among health professionals treating pregnant women with a psychiatricdisease, we asked the Danish Psychiatric Society, the Danish Society of Obstetrics and Gynecology, the Danish Paediatric Society and theDanish Society of Clinical Pharmacology to appoint members for theworking group. A comprehensive review of the literature was hereafterconducted.Results: Sertraline and citalopram are first-line treatment amongselective serotonin reuptake inhibitor for depression. It is recommendedto use lithium for bipolar disorders if an overall assessment finds anindication for mood-stabilizing treatment during pregnancy.Lamotrigine can be used. Valproate and carbamazepin arecontraindicated. Olanzapine, risperidone, quetiapine and clozapine canbe used for bipolar disorders and schizophrenia.Conclusion: It is important that health professionals treating fertile

women with a psychiatric disease discuss whether psychotropic drugsare needed during pregnancy and how it has to be administered.

E. R. Larsen1

, P. Damkier2

,L. H. Pedersen

3, J. Fenger-Gron

4,

R. L. Mikkelsen5

, R. E. Nielsen6

,V. J. Linde

7, H. E. D. Knudsen

8,

L. Skaarup1

, P. Videbech1

1Department of Affective Disorders, Aarhus University

Hospital, Risskov, 2Department of Clinical Biochemistry

and Pharmacology, Odense University Hospital, Odense,3

Department of Clinical Medicine –

Gynecological/ Obstetric Ward Y, Aarhus University Hospital, Skejby,4Neonatal Ward, Sygehus Lillebælt, Kolding, Denmark,5Psychiatry in the Capital Region of Denmark, Psychiatric

Centre Copenhagen, Section 6211, Rigshospitalet,

Copenhagen, Denmark, 6Psychiatry, Aalborg University

Hospital, Aalborg, 7Psychiatry in the Capital Region of

Denmark, Psychiatric Centre Copenhagen, Affective

Ward 6203, Rigshospitalet, Copenhagen, Denmark and8District Psychiatry Center Psychiatric Center, Hvidovre,

Denmark

Key words: psychotropic drugs; pregnancy; breast-

feeding; clinical guidelines; recommendations

Erik Roj Larsen, Department of Affective Disorders,

Aarhus University Hospital, Skovagervej 2, Dk-8240Risskov, Denmark. E-mail: [email protected]

Accepted for publication July 6, 2015

Clinical recommendations

• Fertile women with a psychiatric disease should be advised before pregnancy whether psychotropicdrugs are needed during pregnancy.

• Valproate and carbamazepine should be avoided if possible for fertile women.

• If antipsychotics are needed, olanzapine is primarily recommended based on the amount of safetydata, but risperidone, quetiapine and clozapine can be used.

Additional comments

• This is a literature review. More clinical studies are needed for risk estimation of the differentpsychotropic drugs with long-term follow-up of the children.

Introduction

Drugs and pregnancy: general observations

Since the thalidomide scandal, the use of drugsduring pregnancy has been a sensitive and often

controversial topic among patients, politicians, journalists and health professionals. This has beenreinforced in recent years along with the growth of the internet and the development of our scientificmethods, resulting in a rapid stream of scientific

1

Acta Psychiatr Scand 2015: 132 (Suppl. 445): 1–28 © 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd All rights reserved DOI: 10.1111/acps.12479

ACTA PSYCHIATRICA SCANDINAVICA


2/28

articles in the area. Particularly, the use of psy-chotropic drugs during pregnancy seems to attractattention. For instance, in 2011, almost 100 newscientific articles on the use of selective serotoninreuptake inhibitors (SSRIs) during pregnancy werepublished. These turned out to have far from

always consistent results. The communication of such results to the individual patient/patient popu-lation will very often rest on personal convictionand philosophy of life with regard to empathy andrisk perception. For the media, it is quite easy tofind unfortunate events and put them in relation toan apparent inconsistency between professionalinformation sources and scientific messages.

Consumption of psychotropic drugs during pregnancy

The best Danish data show that between 4 and 9%

of all pregnant women fill at least one prescriptionfor an antidepressant during pregnancy (Sund-hedsstyrelsen www.dkma.dk 2013). A Danish reg-ister-based study of more than 800 000 pregnantwomen in the period from 1997 to 2009 showedthat 4183 (0.5%) pregnant women were treatedwith a SSRI during pregnancy, while 806 (0.1%)discontinued at least 3 months prior (1). There areinsufficient data for other psychotropic drugs. Thetotal use of prescription-only medicine duringpregnancy is widespread. A Danish populationstudy concerning North Denmark Region showedthat the share of first-time pregnant women filling

at least one prescription during pregnancyincreased from 55% to 61% over a 9-year periodfrom 1999 to 2009 (2).

The perception of teratogenic potential

Congenital abnormalities are common among allnewborns. If everything is included, from smalland harmless abnormalities such as webbing of fin-gers or toes (syndactyly) to severe and life-threat-ening abnormalities with abnormal developmentof the brain (anencephaly), the incident is approx.

3 –

4% Therefore, the possibility of a coincidencebetween malformation and medication is present,thus without the causality being proven.

The risk of abortions and malformations

The background frequency of congenital malfor-mations and other unwanted pregnancy outcomes,such as miscarriage, premature delivery, low birthweight, neonatal failure to thrive and possibleimpact on the child’s later development, is notalways well defined. For practical aspects, it may,however, be assumed that the frequency of con-

genital malformations in the population, as men-tioned, is about 3 – 4% (3, 4), and that theincidence of miscarriages in Denmark in women,who know that they are pregnant, is around 15 – 20% (5). However, more recent numbers with amore stringent definition of severe malformations

reduce the background frequency to around2.5%, according to statements from the EuropeanSurveilliance of Congenital Abnormalities,EUROCAT (http://www.eurocat-network.eu). Butthis estimate represents a certain under-reporting.The actual frequency of drug-induced congenitalmalformations is unknown. Schardein hasattempted to estimate the risk and finds that lessthan 1% of all congenital malformations arecaused by medicine (3).

The risk of not treating a psychiatric disorder during pregnancy

About 10 – 15% of all pregnant women experiencea depression during pregnancy. The risk of depres-sion seems to be greatest in the last two-thirds (sec-ond and third trimester) of the pregnancy. Womenpreviously known with depressive disorder have anincreased risk of depression in connection withpregnancy. Pregnancy appears to increase the riskof anxiety, as 10 – 30% of pregnant women experi-ence such symptoms. In one-third of the personsdiagnosed with obsessive-compulsive disorder(OCD), the condition worsens during pregnancyand immediately thereafter, just like the prevalence

of depressive disorder is higher. Some studies findthat approximately 35% of pregnant women witha bipolar disorder experience relapse despite medi-cal treatment, whereas 85% experience relapsewithout drugs, primarily in the shape of depressiveepisodes or mixed episodes with both manic anddepressive symptoms at the same time.

Aims of the study

As we wanted a guideline with a high degree of consensus among health professionals treating

pregnant women with a psychiatric disease, weasked the Danish Psychiatric Society, the DanishSociety of Obstetrics and Gynaecology, the DanishPaediatric Society and the Danish Society of Clini-cal Pharmacology to appoint members for theworking group. A comprehensive review of the lit-erature was hereafter conducted.

Material and methods

The amount of available knowledge about theadverse effects of the individual drugs during preg-nancy varies greatly depending on the drug’s indi-

2

Larsen et al.

http://www.dkma.dk/http://www.eurocat-etwork.eu/http://www.eurocat-etwork.eu/http://www.dkma.dk/


3/28

cation area and for how long it has been marketed.Casuistic reports only have limited value and can-not demonstrate causal relations, and therefore,useful data about drug exposure and pregnancyoutcome must be provided through epidemiologi-cal studies. Cohort studies, case – control studies

and systematic pharmacovigilance can providevaluable information. However, none of thesemethods are ideal, and the scientific quality of datawill typically vary a great deal. The controlledprospective cohorts, where the data quality is thehighest, are usually small and insufficient to serveas a basis for meaningful risk estimation. There-fore, the assessment will often include data, wheredoubt can be raised as to the drug dose and dura-tion of the exposure (3, 6, 7), for example unpub-lished (but typically available to the public) datafrom, for example the Swedish Birth Register (8),

various international abnormality registers andunpublished large cohorts such as Michigan Medi-caid (9). A scientific rigorous approach to data willmean that the attending doctor in most situationswill be left with a limited basis for decision whenchoosing a drug for a pregnant patient. Thus, thetotal data volume often represents a certain neces-sary compromise between quality and quantity. Inaddition, the predictive value of reproductive ani-mal studies is unknown, and the interpretation of these is quite complex (3, 6). Therefore, these arenot included to a greater extent in the currentguidelines which are based on human studies.

Risk assessment on the basis of available data

How many first trimester exposed without detect-able signal do we require before we consider a drugto be safe? Consensus does not exist in scientificsocieties or from regulatory authorities as to whichlevel of certainty is acceptable. Using historicaldata for the background frequency (applied value:3.5%) of congenital abnormalities, it is possible toroughly estimate the following (7, 10):

i) 200 exposed in the first trimester without signs

of excess incidence of malformations provide agood certainty (80% power and 5% level of significance) that the real risk is not more thanthree times higher than the background risk;

ii) 700 exposed in the first trimester without signsof excess incidence of malformations provide agood certainty (80% power and 5% level of significance) that the real risk is not more thantwo times higher than the background risk;

iii) 2000 exposed in the first trimester without signsof excess incidence of malformations provide agood certainty (80% power and 5% level of sig-

nificance) that the real risk is not more than 1.5times higher than the background risk.

If certainty for an even lower risk is demanded,it will require data for many thousand first trime-ster-exposed pregnant women. Such data are, how-

ever, only available for very few medical products.FDA’s classification system, which is under a fun-damental restructuring (11), has operated with avery conservative algorithm for the classificationof drugs for use during pregnancy. This has led toa very low practical applicability of the formal reg-ulatory classification, as only around 30 medicalproducts have achieved the classification ‘can beused during pregnancy’ (12).

The above estimates apply to the general risk;for specific rare malformations, considerably morecohort data are needed, and these kinds of specificmalformations are therefore typically described in

case –

control studies. Neonatal complications andlong-term effects are not included in the above con-siderations. One example is that in the 1940s,diethylstilbestrol exposure in utero was found tolead to an increased risk of uterine cancer, whenthe child had become an adult (13).

Classification strategy

In this document, the decision as to whether a drugmay be used during pregnancy without signifi-cantly increased risk of malformations is primarily

based on the criteria in EMAS’s guideline. Theserecommendations, which are discussed in detailabove, mean in general the following:

i) As a rule, the guideline demands data on atleast 1000 pregnant women without signs of excess incidence of congenital malformationsto recommend a drug during pregnancy.

ii) In the cases where this data volume is notavailable, the recommendations are basicallyarranged according to the available amount of data.

iii) There may be justified deviation from the

above.The risk of other unwanted foetal impacts is dis-

cussed on an ad hoc basis. Data are primarilyobtained from peer-reviewed publications, but insome cases supplemented with unpublished, butwell-documented available exposure data from theSwedish Birth Register www.janusinfo.se.

Drugs and breast-feeding: general observations

All drugs under approx. 1000 kDa transfer intothe breast milk. In modern mother-and-child

3

Psychotropic drugs during pregnancy and breast-feeding

http://www.janusinfo.se/http://www.janusinfo.se/


4/28

medicine, risk assessment in connection withbreast-feeding is primarily based on a quantitativeestimate: how much medication is transferred tothe child during breast-feeding? In the overallassessment, however, importance is also attachedto available information about possible/probable

side-effects in nursed children.The quantitative estimate can be expressed as a

relative infant dose (RID). This is calculated onthe basis of measurements of drug concentrationsin breast milk, preferably at different maternaldoses of the given drug. Based on the highest con-centration, the child’s exposure is now estimated asa weight-adjusted percentage of the mother’s expo-sure. In these estimates, the child is assumed toconsume 150 ml breast milk per kg per day.Regarding nortriptyline, which has a relativeweight-adjusted dose of 1%, it is possible to calcu-

late the following:i) dose of the mother: 100 mg/70 kg = 1.4 mg/

kg;ii) exposure of the child: 1% of 1.4 mg/

kg = 0.014 mg/kg;iii) thus, a 5-kg child will be transferred a total of

0.07 mg nortriptyline per day through thebreast milk.

Milk – plasma ratio is unfit to contribute withmeaningful risk assessment and is an obsoleteparameter. In contrast to use during pregnancy,there are no regulatory guidelines indicating crite-

ria for an acceptable exposure of nursed children.Internationally, no formal consensus exists either.As a point of departure, a RID below 10% isconsidered compatible with breast-feeding. Mostdrugs have a RID below 1% (14). The decisionalgorithm is modified in practice by observation of probable side-effects; drugs with undesirable prop-erties regardless of RID (antineoplastic drugs;immune-modulating drugs); or drugs with a verylong half-life (risk of accumulation). Finally, itmust be added that regardless of RID, a possiblehypersensitivity in the child towards a given drug

can of course contraindicate the use of it.

Specifically in relation to patients with mental illness

It is a fundamental consideration that exposureduring breast-feeding can be avoided by notbreast-feeding. As a rule, breast-feeding should ingeneral be established, but in relation to thispatient population, circumstances in clinical prac-tice can speak against breast-feeding, even thoughno risk is estimated to be connected with the childbeing exposed to the drug via the breast milk inaccordance with the following section. For severely

depressed or psychotic patients, breast-feeding canthus in practice be impossible or even dangerousfor the child.

Disturbed sleep increases the risk of relapse inmany psychiatric disorders. Therefore, it is impor-tant that mothers with a mental disorder are

ensured a good circadian rhythm and a goodnight’s sleep. This speaks in favour of discontinuingbreast-feeding and of formula feeding the child.

Classification strategy

As a rule, this guideline uses data from Medica-tions and Mothers’ Milk 2012 (14) for the purposeof quantitative estimates. In certain cases, datamay be supplemented with data from peer-re-viewed publications. As a rule, this guideline onlyrecommends breast-feeding when RID is below

5%, and any deviations from this are specificallysubstantiated.

Results

Non-pharmacological intervention

Specific information about the mental conditioncan help break the isolation and anxiety often feltby the affected individuals. If suspecting symptomsin the mother, an assessment by her own medicaldoctor or treatment offer in the municipalityshould take place. If depression, severe anxiety,

OCD or psychotic state is suspected, referral isgiven to psychiatric evaluation and treatment.Depending on the severity, the treatment takesplace either in a private practice or at the hospital.Obstetric wards have established ‘vulnerabilityteams’ to help in this situation across Denmark.

When the treatment is initiated, it should first beassessed if talk therapy can help. Talk therapy maybe the best solution for some conditions to elimi-nate or reduce the symptoms and to improve fam-ily life and the relationship to the unborn child. If there is no improvement or prospects of improve-

ment with talk therapy, medical treatment must beintroduced. By medical treatment, both supportingconversations and information are important partsof the total treatment.

Cognitive and interpersonal therapy has a cer-tain effect on depression. Cognitive therapy canalso be used to treat anxiety and OCD.

Pharmacological intervention

Depression. About 20% of the Danish populationwill develop a depression at some point in theirlives. The depressive phase can be divided into

4

Larsen et al.


5/28

varying degrees of severity: mild, moderate andsevere depression where the development can hap-pen acutely in a couple of days or come slowly dur-ing months.

In a mild depression, the person is marked bysadness and fatigue. Sleep disorders may be pre-

sent in a mild degree, and the interest in, for exam-ple, hobbies, studies or work can be slightlyreduced. At this stage, most people are still able tomaintain some study activity or do a job.

In moderate depressions, the low spirits, the fati-gue, the sleep disorders and the loss of interests aremore pronounced. The ability to concentrate willbe reduced, and the desire to go to work or toschool will disappear. Some people choose to iso-late themselves at home, because they cannot copewith going out and be among other people, butalso because they lack energy. Things that are used

to make them happy no longer have the sameeffect. The depressed person often has thoughtsabout death and accidents, some have thoughtsabout suicide and are tormented by self-recrimina-tion and guilt. Typically, they have a pronouncedfeeling of despair and lack of self-esteem. Likewise,the sexual desire can also be reduced. The individ-ual can have a tendency to cry a lot, but can alsobe tormented by irritability, anger and anxiety.The appetite may be reduced, but the opposite – eating for comfort – is also seen.

In the severe depressions, the depressed personis unable to take care of himself and must often be

assisted in even the most basic tasks. Some areparalysed to the extent that even facial expressionsdisappear and they may stop eating and drinking.Severe memory problems and poor concentrationare observed. In some cases, delusions and halluci-nations emerge, and some persons feel such a pro-nounced despair that they will attempt to commitsuicide.

The risk of not treating depression in pregnant

women. Historically, a pregnancy, which by mostpeople is considered to be a happy circumstance, is

believed to protect against mental illness. How-ever, this cannot be substantiated. On the contrary,it is now generally accepted that pregnant womenhave a greater risk of having a depression thannon-pregnant women. About 10 – 15% of all preg-nant women experience a depression during preg-nancy. The risk of depression seems to be greatestin the last two-thirds (second and third trimester)of the pregnancy. About half of these conditionswill continue after the birth of the child (15).

Women previously known with depressive disor-der have an increased risk of depression in connec-tion with pregnancy and in the period immediately

after the birth of the child. The use of prophylactictreatment with antidepressants during pregnancyhas shown to reduce the risk of relapse from about70% without prophylactic medical treatment toapproximately 25% with prophylactic medicaltreatment (16). Psycho-social and biological fac-

tors in connection with the pregnancy probablyplay a role in the emergence of the depression, butclinically, the depressive symptoms do not differsignificantly from the symptoms of non-pregnantwomen.

Depression in pregnant women does not onlypose a direct risk to the woman, but also to theunborn child. Untreated depression in pregnantwomen has been linked to an increased risk of abnormal bleeding during pregnancy, miscarriage,premature birth, foetal death, pre-eclampsia andother birth complications as well as the child’s fail-

ure to thrive after the birth and decreased breast-feeding initiation (17).For the woman, the illness can result in poor

nutritional condition, increased alcohol andtobacco consumption, other unhealthy habits of life, suicide-related behaviour and regular suicideattempts (18 – 20) as well as a poor utilization of theoffers of the pregnancy care. All these factors cancause harm to her unborn child. Suicide amongpregnant women is rare, but a British studyshowed that of all the women who died duringpregnancy, 28% committed suicide (21). Alsoabnormal biological factors (e.g. increased concen-

tration of adrenocortical hormone), which are seenin at least half of the depressed patients, are sus-pected to be harmful to the foetus, partly by affect-ing the brain development in the foetus in anegative way, partly by the fact that the foetus willlater be hypersensitive to stress factors. Thus, anincreased prevalence of depression is shown in 18-year-old children of mothers who were not treatedduring pregnancy for their depression (22). In chil-dren exposed to antidepressants during pregnancy,no signs of behavioural or emotional problemswere detected when assessed at the age of 4 –

5 years. Untreated depression during pregnancyappears to increase the risk of behavioural prob-lems in the child (23). Finally, a depression whichcontinues after the birth will affect the mother’sability to take care of her child.

Medical treatment of women of child-bearing age with

depression. It is important that medical doctors,midwives and other healthcare professionals areaware of any depressive symptoms in pregnantwomen. This applies in particularly to women witha previous depression. According to the guidelinesin the field of the Danish Health and Medicines

5



6/28

Authority, the possibilities of psychotherapeutictreatment should always be considered in depres-sion in pregnant women (24). Medical antidepres-sive treatment can in some cases be necessary and

justified, for instance in case of a severe depressionor in case of high risk of recurrence of the depres-

sion if an already initiated medical treatment is dis-continued. According to the Danish Health andMedicines Authority’s guidelines, the treatmentshould take place in consultation with a specialistin psychiatry. The potential benefits and the harm-ful effects must be balanced against the risk aninsufficiently treated depression poses to the womanand her unborn child. Moreover, the treatment of pregnant women with antidepressants should inDenmark observe the guidelines in the field of theDanish Health and Medicines Authority.

Anxiety and OCD. About 10 –

15% of the Danishpopulation will have a severe anxiety disorder dur-ing their lives. These disorders are more frequentin women than in men and are divided into a num-ber of different types. Social phobia is characterizedby severe nervousness and anxiety if the personmust be together with others. Others experienceanxiety in specific situations, for example whenthey have to sit for an exam or speak in front of agroup. Some people will experience panic disorder,where the anxiety is completely unexpected and inshort-term attacks. Others suffer from agorapho-bia, which is characterized by fear to leave home,

to take the bus, to take the train, to be in crowdsor to be alone at home. Even others suffer from

generalized anxiety disorder with constant nervous-ness and worry.

Obsessive-compulsive disorder is an anxiety dis-order which may consist of obsessive thoughtsand/or patterns of compulsive behaviour andwhich may vary a lot in terms of both the symp-tomatology and how much it affects everyday life.Doubt and uncertainty are key parts of the disor-der and can be seriously disabling.

Obsessions are often dramatic, unpleasant

thoughts occurring over and over, causing anxiety,stress and perhaps feelings of guilt and self-recrimi-nation. Compulsions may consist of washing ritu-als, compulsion to check, rituals of counting andvisualizing (mental compulsive behaviour), andthat you need to think positive (mental obsessivethoughts), to manage and organize things in cer-tain ways, symmetry and superstitious rituals.

The risk of not treating anxiety and OCD in pregnant

women. Pregnancy appears to increase the risk of anxiety, as 10 – 30% of pregnant women experiencesuch symptoms. Approximately 4% of pregnant

women without previous OCD show signs of OCD6 weeks after the birth (25). In one-third of thepersons diagnosed with OCD, the condition wors-ens during pregnancy and immediately thereafter,

just like the prevalence of depressive disorder ishigher (26). Persons diagnosed with OCD appear

to have a prolonged birth and several obstetriccomplications (27) as the woman can be so tor-mented by anxiety or OCD that she is unable togive birth vaginally. After the birth, OCD can beseen in, for example, the mother refraining frombathing or taking care of the child, because she isabnormally afraid to harm it. Violent fantasiesaimed at the child may appear leading to a tor-mented and concerned mother fearing she will acton these fantasies. This is in contrast to the psy-chotic person who does not worry about this. Bothanxiety and OCD can be treated with good results

with evidence-based psychotherapy. In a few cases,the treatment with an antidepressant may be indi-cated. In very rare cases, combination with anantipsychotic drug may be indicated.

Treatment with SSRI antidepressants

Recommendations. Planned pregnancy and start-up

during pregnancy.

i) Prescription and control of the patient’s treat-ment should ideally be performed in collabo-ration with a specialist in psychiatry.

ii) Sertraline and citalopram are first-line treat-ment among SSRI. The advantage of sertra-line is that the treatment apparently can becontinued during breast-feeding without prob-lems.

iii) Fluoxetine is not first-line treatment despitethe highest amount of data, among otherthings because of casuistic reports on neonataldeath, and because the drug is released slowlyin the body of the newborns.

iv) Paroxetine may be associated with anincreased prevalence of malformations as wellas neonatal complications and is therefore not

recommended.v) Escitalopram is not found to constitute any

problems during pregnancy.vi) Fluvoxamine is not recommended due to lim-

ited or no data.

Pregnancy occurring during an existing treatment with

SSRI.

i) Ideally, the patient should be assessed by aspecialist in psychiatry.

ii) In case of pregnancy during an existingtreatment with fluoxetine, change is only

6

Larsen et al.


7/28

recommended if it is considered to be safe inrelation to the patient’s disease.

iii) Regarding treatment with escitalopram, thetreatment can be continued during pregnancy.

iv) Regarding treatment with fluvoxamine,change to another treatment is recommended,

but only if it is considered to be safe in relationto the patient’s disease.

v) Paroxetine treatment should only continueunder extraordinary strict indication.

Birth. In treatment with SSRI up to birth, there isa risk of neonatal complications. These need nor-mally no treatment. Studies have shown anincreased risk of persistent pulmonary hyperten-sion, which, however, occurs very rarely. Overall,outpatient birth is not recommended, and the par-ents should receive information about typical

symptoms in the child, so that the child can beobserved and the right treatment initiated. Discon-tinuation of SSRI is not recommended, if the treat-ment is still indicated.

Breast-feeding.

i) It is important to be discussed with the preg-nant women whether she wants to breast-feed.In some cases, formula feeding is a goodalternative.

ii) If breast-feeding is wanted, sertraline andparoxetine are recommended, as these two

drugs have the fewest reported side-effects andthe smallest transfer into the breast milk.

iii) There are casuistic reports of accumulation of fluoxetine in nursed children, and most reportson symptoms by use of fluoxetine and citalo-pram, which are therefore not recommended.However, in connection with treatment duringpregnancy, the treatment can continue duringbreast-feeding provided that informationabout possible side-effects in the child is given.In particular, one must pay attention to lackof wellbeing, and whether the child achieves

the expected weight gain. In cases of doubt,the drug can be measured in the child’s blood.

iv) Escitalopram and fluvoxamine are not recom-mended due to limited or no data.

Advice for pregnant woman not already in

treatment. When advising a pregnant womanabout antidepressive treatment of depression, thedecision is simple in case of mild as well as severeillness: mild cases should not be treated with drugs(24), and in severe cases, medical antidepressivetreatment is inevitable. It is the middle group whichcauses problems and where non-pharmacological

treatments of course should be considered first,namely the possibility of evidence-based psy-chotherapeutic treatment, that is with interpersonaltherapy or cognitive therapy. If this is not sufficientor possible, the above points should be discussedthoroughly with the woman, so she can make her

choice on the best basis possible. In particular, it isimportant to inform about the risk of congenitalabnormalities in all circumstances, also without theuse of medical treatment. So if the woman choosesto take drugs and later gives birth to a child with abirth defect, it does not have to be due to herchoice, but may have occurred anyway.

From everything we know, the problem is par-ticularly heart malformations, but number neededto treat, one to harm is high. Paroxetine and per-haps also fluoxetine are considered to be the mostrisky. Newer drugs should be avoided as we have

at least knowledge about those. With regard to flu-oxetine, it should be remembered that it shouldnot be taken by breast-feeding women due to therisk of accumulation in the child.

If the woman has previously had an effect of aparticular drug and nothing in particular speaksagainst this substance, it should be used. The riskof neonatal complications should be mentioned, sothat the woman is prepared for them, and in par-ticular, the increased risk of persistent pulmonaryhypertension (PPHN) should be mentioned,because this complication, although rarely, may befatal.

Advice for woman who is in medical treatment, and

who wants to be pregnant. In good time, it shouldbe considered with the woman if there should bean attempt to phase out the treatment. These con-siderations should, among other things, contain anestimate of the risk of recurrence, how severe thedepression was when it was worst and whether thewoman was suicidal. If the woman has previouslytried to phase out her treatment and it lead torecurrence, it may speak against trying to phaseout. Also, residual symptoms in spite of treatment

predict recurrence of the depression when phasingout.

If you choose to phase out, the woman shouldbe offered careful follow-up to provide quick assis-tance in the event of recurrence. If the woman istreated with paroxetine, and phasing out is con-traindicated, she should be switched to anotherSSRI as an attempt, in accordance with the above.Some women are treated with several different psy-chotropic drugs at the same time, for example anantidepressant and an antipsychotic as a sedative.Such treatment should as far as possible be simpli-fied to include only one single SSRI [or possibly

7



8/28

one single tricyclic antidepressant (TCA)], becauseit is not possible to predict how several differentsubstances will interact in the growing foetus.

A routine phasing out of SSRI 2 weeks prior tothe birth to avoid neonatal complications is notrecommended (28, 29).

Advice to woman already in treatment, who discovers

she is pregnant. The decision about what to dodepends on for how long the woman has beenpregnant when she discovers that she is pregnant.Often, she will be well through the first trimester,and the risk of severe malformations must beassumed to be over. Therefore, there will be noreason to stop the treatment. However, the risk of PPHN still persists. Therefore, some authors rec-ommend that the drug which the woman is alreadytreated with is continued to not risk recurrence if

the treatment is changed.

Background

The effect of selective serotonin reuptake inhibitors(SSRI) for the treatment of depression, OCD andanxiety is scientifically well documented. Cognitivetherapy and interpersonal psychotherapy are alsowell documented in non-pregnant women, and afew studies have shown they are effective in preg-nant women too. Therefore, these types of thera-pies are in principle preferred in mild to moderatedepression and in anxiety disorders in pregnant

women instead of medical treatment. One of theproblems, however, is that they are often not easilyaccessible, among other things, due to a lack of qualified therapists. In addition to this, not allrespond to psychotherapy. In many studies, thesuccess rate in non-pregnant women is around60%. It also often happens that women already inantidepressive treatment want to become pregnant.Or that the woman who is treated with SSRI dis-covers that she is a few weeks to months into herpregnancy. The general practitioner, obstetriciansand psychiatrists will therefore often have to

advise pregnant women about such a treatment.

What risks are present?

It is well known that the incidence of miscarriagesand stillbirths can be influenced by, for example,medication intake during pregnancy, even thoughit normally occurs very rarely. There may beneonatal complications, and a relatively new areais the so-called behavioural teratology, that iswhether the child’s motor, intellectual andemotional development can be influenced by drugsgiven during pregnancy.

On the other hand, a number of studies indicatethat untreated depression in itself may harm theunborn child. However, none of the SSRI drugsare categorically recommended for the treatmentof depression during pregnancy. All of these drugsshould only be used if the benefits for the pregnant

woman (and her unborn child) outweigh thepotential side-effects.

Miscarriages and stillbirth

It is estimated that around 20% of all pregnanciesend up with a miscarriage, but it is difficult toassess for certain, as many miscarriages are soearly that they can be confused with a late comingperiod. Moreover, the risk depends on thewoman’s age and lifestyle. The frequency of mis-carriages, where the woman knows that she is

pregnant, is about 15 –

20%. The results concerningthe impact of SSRI on the frequency of abortionare not clear (30).

An important study of 937 women who tookSSRI during pregnancy, compared with a similarnumber who did not, showed 13% abortions in theSSRI group against 8% in the control group (31).Other studies suggest something similar (32). Thestudies are not without methodological problems.For instance, in the first mentioned study, therewere more smokers and more women who had pre-viously had miscarriages in the group of womenwho took SSRI than in the control group.

Possibly, the increased incidence of miscarriageis specifically associated with paroxetine and ven-lafaxine. To be on the safe side, the woman, whenadvised about choice of treatment, should bewarned about the risk of miscarriage possiblybeing increased. In two large studies from Norwayand Denmark, SSRI during pregnancy was notassociated with significantly increased risk of still-birth or neonatal death (33, 34). A large Danishregister study from 2013 with more than 1 millionpregnancies found a slightly increased risk of mis-carriage by the use of antidepressant drugs during

pregnancy, but this risk could not be replicatedwhen age and mental illness were taken intoaccount. Hence, no increased risk was found inusing SSRI in pregnant women with depressioncompared to women with untreated depression(35).

Congenital malformations

There are many statements, the majority of epi-demiological nature, which shed light on the riskof congenital malformations in children exposed toSSRI during pregnancy. These are rather heteroge-

8

Larsen et al.


9/28

neous both in terms of quality and quantity of dataas well as definitions of exposure and outcome.Therefore, it is difficult to get an overview of methodological details, multiple stratifications,odds ratios and incomparability between thesestudies. Thus, here comes a short summary of the

largest and best statements:A very large meta-analysis was published in

2013 (36). More than 50 000 SSRI-exposed preg-nant women were included, and no overallincreased risk of congenital abnormalities wasfound, RR 0.93 (CI: 0.85 – 1.02) In studies with out-come data for cardiovascular malformationsincluding a total of more than 20 000 SSRI-ex-posed pregnant women, a slightly increased riskwas found, RR 1.36 (CI: 1.08 – 1.71). A significantpart of this signal is carried by septal defects, RR1.40 (CI: 1.10 – 1.77). Some studies – but not all –

have shown a specific increased risk of heart mal-formations by paroxetine and fluoxetine exposure.In the meta-analysis, a signal specifically for parox-etine, but not for fluoxetine, was found. Eventhough there is no consensus as to whether this is areal signal, most recommend avoiding these twodrugs during pregnancy. However, there may beclinical cases where switching treatment from acurrent effective treatment with paroxetine or flu-oxetine is not rational, because the risk for themother with a drug change is greater than thehypothetical risk for the foetus.

Some important specific studies (all of which are

included in the above meta-analysis) must bementioned.

Very large Swedish and Finnish studies showthat if SSRIs are resulting in malformations, theyare located at the heart (37 – 39). However, this dif-ference is not statistically significant compared tothe background population (40). Heart malforma-tions in general were not found more frequentlythan you would expect, but atrium and ventricularseptal defects were more frequent. Among non-medicated women, the risk was 0.5%, whereas itwas 0.8% among those taking SSRI. Number

needed to treat, one to harm, (NNH) was calculatedto 246. This means that more than 246 womenmust be treated with SSRI to cause one additionalcase of septal defect. In addition to this, a numberof these are asymptomatic. It cannot be ruled outthat many septal defects were only found, becausethe foetuses known to have been exposed to SSRIduring pregnancy were examined extra carefully.

In a methodological elegant Danish registerstudy, the frequency of heart malformations of the4183 children exposed to SSRI was compared withthe frequency of children whose mothers had previ-ously been taken SSRI, but who discontinued with

the drug prior to pregnancy and resumed treatmentafter the birth. The risk of heart malformations wasincreased equally for both groups (1). This suggestsconfounding by indication or selection bias, whichmeans that the abnormalities were not caused bythe SSRI exposure, but by other factors, which are

common for the women who took the drug and thewomen who discontinued the treatment. A largepharmacoepidemiological study from the USAwith 65 000 pregnant women who have usedantidepressants in the first trimester has found noincreased risk of cardiac malformations (41).

Reduced foetal development and premature birth

Some studies, but not all, suggest that the treat-ment with SSRI during pregnancy can lead to pre-mature birth and that children born at term are

smaller. However, it is difficult to estimate as somestudies show that untreated depression in itself may result in premature birth and low birthweight. To be on the safe side, however, informa-tion about the risk should be provided (42, 43).

Complications in connection with the birth

Between 15 and 30% of children whose mothershave taken a SSRI in the last time up to the birthwill hours to days after birth show symptoms,which possibly are discontinuation symptoms (44).These comprise, among other things, irritability,

quivering, limpness, and trouble to suck or sleep.Spasms have also been seen (36). The symptomscease spontaneously and require generally no treat-ment, but serious cases requiring treatment, how-ever, occur. The significance of the symptoms isfirst and foremost that they can be confused withsimilar symptoms of serious diseases, such as lowblood glucose and brain damage, and there may beproblems in relation to the establishment of breast-feeding or mother-and-child bonding. Therefore,in some obstetrics wards, the children are observedfor about 24 h, before mother and child are dis-

charged. If the symptoms are especially pro-nounced, you can choose to observe the child in aneonatal ward. A Danish study of children born toSSRI-treated mothers showed lower averageApgar score and a higher degree of hospitalizationon neonatal wards for these children comparedwith control children (45). A Danish register studyfrom 2013 shows that the use of SSRI increases therisk of low Apgar score independent of themother’s depression (46). In principle, the symp-toms can occur in connection with all types of SSRI, but one study has found paroxetine to beparticularly frequently associated with these (47).

9



10/28

The woman should be informed about the risk thatthe child can have these symptoms, and what sheshould do in this case. A register study of morethan 120 000 newborns from the period 1998 to2001 showed no effect on the newborn child’shealth by reducing exposure to SSRI towards the

end of the pregnancy (29).

Persistent pulmonary hypertension (PPHN)

In the minutes after the birth, a number of changesoccur in the child’s circulatory system. This pro-cess is controlled, among other things, by prosta-glandins, and it is expected that serotonin outsidethe central nervous system also participates. Thismay be due to the properties of this substance onblood vessels. A number of aspects and diseasescan interfere with this important process. Malde-

velopment of the lungs, diaphragmatic hernia andmeconium aspiration can thus among many dis-eases and conditions result in PPHN, but studiesindicate that the risk is increased if the mother is intreatment with SSRI after the 20 pregnancy week.The symptoms appear immediately after the birthand include dyspnoea and cyanosis. The conditioncan often be treated, but is serious and potentiallylife-threatening.

A recent population-based cohort study fromthe Nordic countries comprised 1.6 million new-born babies born after pregnancy week 33. The fre-quency of PPHN among non-exposed children was

1.2 per 1000, whereas among children exposed toSSRIs after week 20, it was 3 per 1000. This givesan adjusted odds ratio of 2.1 (95% CI 1.5 – 3.0).The risk was increased roughly in the same degreefor all types of SSRI, which implies that it is a classeffect (48). Other small studies have also pointed toan increased risk of this condition. A new meta-analysis of PPHN showed that early exposure toSSRI does not increase the risk, whereas exposurelate in the pregnancy (after week 20) increases therisk by a factor 2.50 (95% Cl 1.32 – 4.73), the abso-lute frequency was 2.9 to 3.5 per 1000 newborns,

which lead to NNH being between 286 and 351(49). The increased risk can probably be eliminatedif the mother can do without SSRI in, for example,the last month before the birth. However, theproblem is that the risk of relapse is increased justas she is about to give birth.

Brain development and emotional development

In experimental animal models, primarily forrodents, SSRI affects the early brain development.A prospective study of 99 pregnant women who

used SSRI and 570 pregnant women with depres-sive symptoms who took no antidepressive drugsshowed that untreated depression was associatedwith lower rate of growth of the foetus’ body andhead, whereas the treated women’s foetuses hadless growth rate of the head, but not of the rest of

the body. The difference between the head size of the two groups was, however, only 4 mm (50). Asmall MRI study of 33 children who had beenexposed to SSRI suggested that they had anincreased incidence of the so-called Chiari malfor-mation type 1, which is caused by herniation of thecerebellar tonsils into the spinal canal and whichmay be asymptomatic or cause, for example head-ache. The result is interesting, but must be repli-cated before therapeutic consequences can betaken (51).

Unfortunately, only few human follow-up data

for many types of antidepressants exist, but theexisting studies show no correlation between thedrug and the brain development of the foetusmeasured with regard to intelligence developmentand on a number of emotional and motor param-eters. An American study (52) and a Danish (53),however, found associations with slightly delayedmotor development. In a follow-up on the Danishstudy, however, no correlation was found betweenexposure to antidepressants and behaviour at theage of 4 – 5 (23). The studies are difficult to inter-pret due to the many sources of error. For exam-ple, the depression in the mother, according to

several studies, can in itself delay the child’s devel-opment, and this effect cannot be distinguishedfrom the effect of the drug. Such a correlationwas found in the before-mentioned Danish study,where maternal depression, probably after thebirth, was a predictor for the child’s behaviouralproblems.

One study should be mentioned to illustratethe problems of assessing such scientific results:Croen et al. (54) studied 298 cases of Asperger’ssyndrome (AS) and found that 6.7% of thewomen who had a child with this syndrome had

taken SSRI during pregnancy, whereas only3.3% of healthy children’s mothers had takenSSRI. This means that the frequency of ASseems to have doubled if the mother took SSRI.Fortunately, the risk of AS is very low: approxi-mately 0.26 per 1000 children. So, even if therisk is doubled, it is still very low. The authorsdo not rule out that what they show is in fact agenetic predisposition for AS: the women whoreceived treatment had psychological difficultiesdue to Asperger features, which increased therisk of depression or anxiety and subsequently

10

Larsen et al.


11/28

SSRI treatment. At the same time, their childrenhad an increased risk of AS, as there is a consid-erable hereditary element in the disease. Anotherpossibility is that the children’s illness is causedby the mother’s illness and not her treatment,that is that her depression disturbed the develop-

ment of their connection with other people. In aDanish register study of more than 600 000 chil-dren, no link was found between exposure toantidepressant drugs in utero and autism spec-trum disorders (55). The problems relating tochange of the child’s behaviour during childhoodis important to be aware of in future, so thatnew and more thorough studies can be designedto obtain certain knowledge. Furthermore, it is amajor problem in the existing studies that thechild is only followed until it was 5 – 10 years old.However, one study shows that the worser the

depressive symptoms the woman had duringpregnancy, the greater the risk that her child hadhad a depression at the age of 18 (22).

Treatment with serotonin norepinephrine reuptake inhibitors


during pregnancy. Venlafaxine and duloxetine arenot recommended as first-line treatment duringpregnancy. For venlafaxine, however, the data vol-ume is extensive and meets the criteria, but cannotmatch the number of studies of SSRI. The data areinsufficient regarding duloxetine.

Pregnancy occurring during an existing treat-

ment. Venlafaxine is not recommended as first-line treatment during pregnancy. The same appliesto duloxetine. Ongoing clinical satisfactory treat-ment with venlafaxine should be continued. Ongo-ing duloxetine treatment should be changed.However, if the woman has not previouslyresponded to other treatment, it calls for a contin-uation of the ongoing treatment.

Breast-feeding. As a rule, venlafaxine is not rec-

ommended as RID is around 7 –

8%, althoughside-effects have not been described in the nursedchild. Duloxetine can be used as RID is around1%. No side-effects have been described innursed children.

Background. There are only scarce data on the useof these drugs during pregnancy. No studies havebeen published of possible long-term effects on thechild after exposure in foetal life. One prospectivecomparative study of 150 women using venlafaxine(56) and a case series with 11 women have been

published, overall, without signs of teratogenicity(57).

Regarding venlafaxine, Swedish data for morethan 1600 first trimester-exposed pregnant womenshow no increased risk of unwanted foetal impact.Regarding duloxetine, data exist for around 350

first trimester-exposed pregnant women (58 –

60).There are casuistic reports on problems in new-borns, but there are no comparative studies. Swed-ish statements of women who have used serotoninnorepinephrine reuptake inhibitors (SNRI) foundthat the symptoms in newborns reminded verymuch about those seen during treatment with anSSRI. This study found no signs of teratogenicity,but rather an increased risk of premature birth (61).

A large pharmacoepidemiological study fromthe USA with 6900 women exposed to SNRIfound no significant increased risk of heart malfor-

mations in the child after checking for confoundersas physical illness in the mother (41).Venlafaxine has been detected in the blood of

nursed infants without detectable effects on thechildren. However, only data on a total of 12exposed children have been published (62).

Treatment with noradrenergic and specific serotonergic

antidepressant (NaSSA)


during pregnancy. Due to inadequate data, mir-tazapine and mianserin are not recommended dur-

ing pregnancy.


ment. Mirtazapine and mianserin are not recom-mended during pregnancy. However, if the womanhas not previously responded to other treatment, itcalls for a continuation of the ongoing treatment.

Breast-feeding. Due to inadequate data, mirtazap-ine and mianserin are not recommended duringbreast-feeding.

Background. Mirtazapine and mianserin exhibittheir antidepressant effect by impacting themonoaminergic system in the brain. There are onlyscarce data on the use of these drugs during preg-nancy. No studies have been published on possiblelong-term effects on the child after exposure duringpregnancy. In a prospective follow-up study of 104women who took mirtazapine, but where 35% alsotook another type of antidepressant drug, sedativedrugs or mood-stabilizing drug, there were nosigns of teratogenicity (63). Increased risk of premature birth on a par with other antidepressive

11



12/28

drugs and increased risk of miscarriage werefound. But this may be a matter of confounding byindication.

Treatment with tricyclic antidepressants

Recommendations. In general, tricyclic antidepres-sants are not recommended as first-line treatmentof a depression. The data volume for some TCAsmeets the criteria, but significantly more data existon the use of SSRI in pregnant women and womenbreast-feeding.

Planned pregnancy and start-up during pregnancy. Tri-cyclic antidepressants (amitriptyline, clomipramine,imipramine and nortriptyline) do not seem to beassociated with an increased risk of malformations.The data volume of these drugs meets the criteria

stated in the introduction. The other TCAs cannotbe recommended due to scarce data.

Pregnancy occurring during an existing

treatment. Plasma monitoring of TCAs should becarried out continuously during the pregnancy, forexample every 3 months, as there may be a risk thatthe conversion of TCA is changed, so that concen-tration in the blood becomes higher than intended.

Birth. Tricyclic antidepressants are associated withan increased risk of temporary impact on the child

just after the birth. The newborn child may present

with symptoms in the shape of increased crying,constipation, problems with urinating and nausea.

Breast-feeding. RID for amitriptyline, nortripty-line, clomipramine and imipramine is about 1 – 3%.No serious side-effects are reported in nursed chil-dren by maternal treatment with amitriptyline, clo-mipramine, imipramine and nortriptyline.

Background. Tricyclic antidepressants. Some of the substances are chemically related. For example,amitriptyline is metabolized to nortriptyline and

similarly, clomipramine and imipramine share thesame basic structure. Only scarce or no humandata have been published on the use of maprotilineand dosulepin during pregnancy.

Previous studies suggested a possible associationbetween malformations of the extremities and theuse of clomipramine, imipramine, nortriptyline oramitriptyline during pregnancy (64). Subsequentstudies, including three prospective and a largecase series, however, have not been able to confirmthe association. The studies are based on about400 exposed, but with very different study design(65). A Swedish register study of 14 821 women

treated with antidepressants finds that clomipra-mine leads to an increased risk (OR: 1.84) for ven-tricular septal defect and atrial septal defect (39).But at the same time, it is stated that there may bea confounding by indication, which is why it is rec-ommended to avoid start-up in case of planned

pregnancy, whereas there is no reason to discon-tinue during a pregnancy. There are no signs of ter-atogenic effect of doxepin, which, however, isbased on scarce and unpublished data from a mon-itoring programme. Symptoms compatible with adiscontinuation reaction in newborns exposed toimipramine or clomipramine are described. Noreports of symptoms after exposure to amitripty-line or nortriptyline are published. One case of uri-nary retention in a newborn child exposed tonortriptyline in utero as well as one case of neona-tal paralytic ileus after exposure to doxepin and

chlorpromazine in the third trimester is described.There are a few studies concerning the develop-ment of children until preschool age, who duringpregnancy have been exposed to TCA (66). Twostudies have not been able to demonstrate impactof IQ, language or behavioural development afterexposure to TCA. Within the group of TCAs, nor-triptyline has been recommended, partly due toless pronounced cholinergic effect, and partly dueto the fact that it is possible to measure the concen-tration in the blood and thus reliably determinethe right dose.

Amitriptyline, nortriptyline, desipramine and

clomipramine have not been found in the blood of nursed infants, and no side-effects in the child aredescribed. Doxepin has been associated with symp-toms of the child, including drowsiness and vomit-ing. Dosulepin is measured in breast milk as wellas in the blood of nursed children, but no side-ef-fects in the child have been reported. Maprotilinecan be measured in breast milk, but otherwise onlyscarce data are available. It is uncertain whetherTCA affects the brain development. There are nosigns of serious problems in their childhood afterexposure in the pregnancy assessed on the basis of

two studies with a total of 116 exposed (66).

Treatment with monoamine oxidase inhibitors. Notrecommended due to insufficient data.

Treatment with agomelatine. Not recommendeddue to insufficient data.

Treatment with vortioxetin. Not recommended dueto insufficient data.

Bipolar affective disorder. At least 40 000Danes are affected by bipolar affective disorder

12

Larsen et al.


13/28

(previously called manic-depressive disorder). Theterm ‘bipolar’ covers the disorder’s two poles, themanic and the depressive pole. All people experi-ence mood swings, but in the case of bipolar disor-der, the mood swings are extreme and exaggerated.This can result in behaviour, which can be socially

disabling or life-threatening. The disorder mostoften appears at the age of 16 – 18 years or in theearly adulthood, and it affects an equal amount of men and women. People with bipolar disorder areoften diagnosed many years after the onset of thedisease.

Persons with a bipolar disorder have a risk of sui-cide of 10 – 15%, and the symptoms are often accom-panied by other mental problems such as anxietyand abuse. Due to the cyclic phases of the disorderand the risk of relapse, preventive mood-stabilizingtreatment will often be indicated. Untreated, the dis-

order often has serious consequences for the individ-ual and the family in the form of serious financial,work-related and social costs.

What is mania?. In the manic phase, the mood iselevated, and the person is most often described asbeing ‘all keyed up’ with abnormally elevatedenergy level and as being socially flattering andcharming bordering towards a socially inappropri-ate behaviour with a lack of understanding of lim-its. The manic phase is divided into degrees of severity: hypomania, moderate mania and severemania. Generally, mania can occur very suddenly

and develop either in a few hours or in a fewweeks.

In the hypomanic (mild manic) phase, the per-son is more active, committed and energetic thanusual. At the same time, the person is more opti-mistic, outgoing and speaks more. However, theremay also be an increased tendency towards irri-tability. Many people with bipolar disorderdescribe the hypomanic state as being very produc-tive and actually a good state to be in.

In a moderate manic phase, the symptomsincrease in strength. The person is hyperactive, has

several projects running at the same time, and isrestless and uneasy, and the need for sleep is oftendecreased. There may be an increased ideation andproblems with sticking to one subject at a time.The person can easily lose the thread and be dis-tracted. Some people become more intrusive, irri-table or angry if their many ideas are rejected. Theperson is usually excited, has an increased self-es-teem, may experience an increased sexual desireand develop an increased spending of money with-out thinking about the consequences. In the sameway, some people begin to behave irresponsiblyand recklessly.

In the severe manic phase, the symptoms areextremely pronounced: the person sleeps no morethan 3 – 4 h, is in constant activity, is out of controland has a feeling of being invincible. The conditioncan be associated with severe anxiety, but the per-son can also be threatening, aggressive and explo-

sive. In some cases, there may be psychoticsymptoms such as delusions and hallucinations. Inuntreated cases, a mania can result in so-calledacute delirium, which is a life-threatening condi-tion in which the person is extremely troubled andrestless, the body temperature rises and the personmay lose a lot of fluid. At worst, the condition canresult in circulatory collapse and death.

The risk of not treating the bipolar condition in preg-

nant women. The risk of relapse of depression ormania seems to be highest in month 4 – 9 of the

pregnancy. Some studies find that approximately35% experience relapse despite medical treatment,whereas 85% experience relapse without drugs,primarily in the shape of depressive episodes ormixed episodes with both manic and depressivesymptoms at the same time (67). If the mother isnot treated, the foetus can risk being exposed tolarger concentration of stress hormones, especiallyif the patient with bipolar condition is destabilized.Disorder breakout in the mother can have seriousconsequences for her child, as it results in anincreased risk of excessive use of alcohol, tobaccoand illegal drugs, poor nutrition, sexually transmit-

ted diseases, suicide, reduced maternal care for thechild and physical violence against the child (19,68 – 71).


bipolar disorder. The possibility of pregnancyshould always be discussed with this group of women. Furthermore, psychotropic drugs specifi-cally problematic during pregnancy, such as val-proate and carbamazepin, should be avoided.Evidence of the preventive effect is poor for val-proate and carbamazepin, which is why they

should primarily be used in the treatment of acutemania if other treatments have appeared to be inef-fective or not tolerated (72). At the same time, itmust be ensured that the woman is not pregnant.Valproate also increases the risk of polycysticovary syndrome, which can cause irregular periodsand infertility. Oral contraceptives should be takeninto consideration in the event of valproate treat-ment, partly to inhibit the development of polycys-tic ovary syndrome, partly as safe contraception.At start-up and discontinuation of oral contracep-tives containing oestrogen, it has to be taken intoconsideration that the oestrogen increases the

13



14/28

breakdown of lamotrigine by approximately 50%.Antipsychotics may increase prolactin and resultin infertility. This is specifically evident in connec-tion with risperidone and typical antipsychotics,but can occur in all antipsychotics (73).

Treatment with mood stabilizers


during pregnancy.

i) Prescription and control of the patient’s treat-ment should be performed in collaborationwith a specialist in psychiatry, ideally in a spe-cialist clinic in the regional psychiatric.

ii) Lithium is still a cornerstone in the treatmentof bipolar disorder, and as the absolute risk of teratogenic damage is small, it is recom-mended to use lithium if an overall assessment

finds an indication for mood-stabilizing treat-ment during pregnancy.

iii) Valproate and carbamazepin are contraindi-cated during pregnancy due to the high risk of neural tube defects.

iv) In bipolar disorder with primarily or exclu-sively manic episodes, lithium, olanzapine,risperidone, quetiapine and clozapine can beused. Alternatively perphenazine*.

v) In bipolar disorder with primarily depressiveepisodes or without manic episodes duringmany years, lamotrigine may be used. In gen-

eral, use of doses up to 300 mg is recom-mended. Special caution is recommended inthe case of higher doses, where serum mea-surement is recommended each month.

vi) Antidepressants are generally not recom-mended in bipolar disorder, as the effect of thisis much debated.


ment. The patient should be assessed by a special-ist in psychiatry. The indication for medicaltreatment during pregnancy of women with bipo-lar disorder depends, among other things, on the

following factors:i) The general evidence of the effect of mainte-

nance treatment of various drugs.ii) The generally increased risk of relapse/devel-

opment of new episodes during and after

pregnancy with and without mood-regulatingdrugs.

iii) Risk of teratogenic damage for the individualdrugs.

iv) The individual patient’s course of illness (timeof onset, current age, the number of illness epi-

sodes, type and severity of affective episodes,any previous course of pregnancy).

v) Effect and possible side effects of mood-stabi-lizing drug for the individual patient.

On the basis of items 1 – 5, a professional assess-ment of the patient’s individual risk of affectiveepisodes vs. teratogenic damages is made, formingthe basis for the decision of the patient and spouseregarding medical treatment during pregnancy.

Background. Pharmacological treatment of bipo-lar disorder during pregnancy is a highly special-

ized function, which according to the DanishHealth and Medicines Authority should only beundertaken in specialist clinics for affective disor-ders. Guidance in medical treatment of pregnantwomen requires thorough knowledge of the patientand the specific illness course, of special risk fac-tors and of the treatment. This should take placein close cooperation with obstetrics wards.

Pregnancy does not protect against new epi-sodes, and discontinuation of mood-stabilizingdrugs during pregnancy increases the risk of relapse considerably (67). In principle, it is recom-

mended to use the lowest possible effective dose of mood-stabilizing drug as well as to use as fewdrugs at the same time as possible. One must beaware that an increase of dose often is necessary inthe third trimester, where the blood volume isincreased by approximately 30% (74). Plasmamonitoring is recommended, for example withblood test every month, among other thingsbecause several of the liver’s enzyme systems haveincreased activity during pregnancy, which meansthat for instance, lamotrigine breaks down morequickly.

As the plasma volume and the kidneys’ filtration

speed are increased during pregnancy and normal-ized immediately after the birth, it means that itmay be necessary to increase the lithium dose dur-ing pregnancy and similarly reduce immediatelyafter the birth.

General evidence of the effect of maintenance treat-

ment of various drugs. The evidence of mainte-nance treatment can be summed up on the basis of new guidelines from the World Federation of Soci-eties of Biological Psychiatry (75). Lithium is thedrug with the best evidence for a mood-stabilizing

*After Trilafon (perphenazine) has been deregistered in Denmark,

patients can be treated with the similar drug Peratsin instead. The med-

ical doctor may apply for general permission to prescribe Peratsin

‘Orion’ 2, 4 and 8 mg via the Danish Health and Medicines Authority

(http://laegemiddelstyrelsen.dk/da/topics/godkendelse-og-kontrol/

udleveringstilladelser). A copy of the permit is handed out to the

patient, and the pharmacist can then hand out the drug.

14

Larsen et al.

http://laegemiddelstyrelsen.dk/da/topics/godkendelse-og-kontrol/http://laegemiddelstyrelsen.dk/da/topics/godkendelse-og-kontrol/http://laegemiddelstyrelsen.dk/da/topics/godkendelse-og-kontrol/http://laegemiddelstyrelsen.dk/da/topics/godkendelse-og-kontrol/


15/28

effect in bipolar disorder. Quetiapine has, aslithium, shown mood-stabilizing effect in relationto the prevention of depression, mania and mixedepisodes, whereas there is less evidence that olan-zapine prevents depressive episodes. Aripiprazoledoes not prevent depressive episodes, but only

manic and mixed episodes, like all other atypicalantipsychotics. Valproate is effective in the eventof acute mania, whereas the evidence of the pre-ventive effect on mania is poor. Lamotrigine pre-vents in particular depressive episodes.

General risk of relapse/development of new episodes

during and after pregnancy with and without mood-sta-

bilizing drugs. No randomized studies are avail-able, which have compared the effect of mood-stabilizing medical treatment versus no treatmentduring pregnancy. However, there is a consider-

able risk of development of affective episodes dur-ing and after pregnancy.

During pregnancy. The largest and best conductedstudy is from Viguera et al. In a prospective non-randomized trial, they found that 71% of a total of 89 women with a known bipolar disorder devel-oped a new affective episode during pregnancy(67). Most episodes were depressive or mixed epi-sodes, and 47% appeared in the first trimester.Women who discontinued with medical treatmentimmediately before pregnancy had two timesincreased risk of developing new episodes and were

ill five times as many weeks during pregnancy com-pared with women who continued the mood-stabi-lizing medical treatment.

After pregnancy. A Danish register-based studyshowed that 26.9% of women with known bipolardisorder were hospitalized within a period of 1 year after the birth of the child. The greatest riskof hospitalization was 10 – 19 days after the birth(76). Patients with bipolar disorder have anincreased risk of developing a postpartum psy-chosis (77), particularly in the first 4 weeks after

the birth of the child. The risk of relapse is eighttimes higher in the first months after the birth,which increases the need for close follow-up andsupport in this period. In the event of acute maniaduring pregnancy, antipsychotics or ECT is used,which does not affect the foetus inappropriately,and which is described further in a later section.

Risk of teratogenic damage for the individual

drugs. Lithium: The risk of Ebstein’s anomaly(right ventricular hypoplasia and displacement of the tricuspid valve) in children is 1 : 20 000. Inchildren exposed to lithium in the first trimester,

the risk seems to be between 1 : 1000 and 1 : 2000,which is 10 – 20 times increased compared to therisk of the background population (71, 78 – 80), butin absolute figures still a very small risk. A recentmeta-analysis published in Lancet identified 62studies of possible teratogenic effects of lithium

(81). The risk of Ebstein’s anomaly was not statis-tically significantly increased among childrenexposed to lithium compared to the risk amongunexposed children, but the estimate is uncertaindue to the low number of cases with Ebstein’sanomaly. The article also refers to a case – controlstudy of children born with malformations in gen-eral, which did not find statistically significantlymore women treated with lithium in the malforma-tion group (6 of 10 698) compared with the controlgroup of children without malformations (5 of 21 546) (82). Lithium must still be considered a

cornerstone in the treatment of bipolar disorder,and as the absolute risk of teratogenic damage issmall, it is recommended to use lithium if an over-all assessment finds an indication for mood-stabi-lizing treatment during pregnancy (83).

Antiepileptics: Generally, there is a correlationbetween dose and the risk of malformations, whichis why it is important to evaluate whether the preg-nant woman can be treated with a lower dose dur-ing pregnancy (84).

Lamotrigine: Lamotrigine used as monotherapyappears to constitute a low risk (71, 85, 86), eventhough a few reports as the North American

Antiepileptic Drug Pregnancy Register indicate anincreased risk of cleft lip and palate increasingfrom 1 : 500 among the population to 1 : 120 (87,88). The risk seems to be increased in a dose of more than 300 mg/day (89, 90), as the risk of heartmalformations and hypospadias increases to 1%.The plasma level of lamotrigine falls by 60 – 65% insecond to third trimester (90, 91) and should bemonitored closely during the pregnancy, for exam-ple with blood test every month. After the birth,the serum concentration of lamotrigine increasesagain. Prophylactic treatment with folic acid (5 mg

daily) should be initiated in case of pregnancy wishto prevent neural tube defects. Treatment withfolic acid will continue in the first trimester (92). Ingeneral, use of lamotrigine up to 300 mg is recom-mended. Special caution is recommended in thecase of higher doses, where serum measurement isrecommended each month.

Valproate and carbamazepine: The risk of neuraltube defect using carbamazepine over 1000 mg is2% (90). When using valproate below 700 mg, therisk of neural tube defects is 1% and over 700 mg2%. When using over 1500 mg valproate, the riskof multiple malformations is 7%, of heart malfor-

15



16/28

mations is 7% and of hypospadias is 5%. Val-proate may possibly affect the foetal brain, and ahigh association between valproate exposure inutero and later risk of autism is found (93). The useof valproate and carbamazepine is contraindicatedduring pregnancy.

Breast-feeding. Lithium: RID is very high, between12 and 30%. Nursed children achieve plasmawhich is up to 24 – 72% of the mother’s (94). Use of lithium during breast-feeding is usually not recom-mended, but the drug can be used under closeobservation of the child and possibly guided bymeasurements of the concentration of lithium inthe breast milk. Lithium can lead to dehydration,exhaustion, hypotonia and electrocardiographicchanges in the child.

Valproate: Breast-feeding is possible. RID is 1 –

2%. No side-effects have been described in nursedchildren.Carbamazepine: Breast-feeding is possible. RID

is 4 – 6%. No side-effects have been described innursed children.

Lamotrigine: Breast-feeding is possible at dosesof no more than 200 mg daily. RID is relativelyhigh from 9 to 18%. Nursed children achieveplasma concentrations of approximately 30% of the mother’s. Many courses without side-effects innursed children are described. One case of neona-tal apnoea is described where the mother had850 mg daily.

Generally, antidepressants should not be used inbipolar disorder, as it is still being debated whetherthere is evidence for this (95).

Psychoses. This group of disorders is divided intothree subgroups consisting of schizophrenia, acutepsychoses and chronic psychoses. Around 25 000people in Denmark are diagnosed with schizophre-nia. In the course of a lifetime, 1% of the popula-tion will be diagnosed with schizophrenia. For themajority of people with schizophrenia, the disordermanifests itself between the ages of 16 and

25 years, but the disorder is also seen in childrenand elderly. The gender distribution is equal, how-ever, with a tendency that women have a higher ageat onset and a milder course than men. The disor-der affects the ability to think, feel and perceivereality correctly, and there will often be problemswith entering into relations with others. The mostnoticeable symptoms are hallucinations and delu-sions. By hallucinations, the person, for example,hears voices which speak about the person in ques-tion, and hallucinations in other sensory formsmay also occur, for example olfactory, gustatoryand tactile. Delusions are characterized by an

unshakable conviction which is considered unreal-istic or impossible. The delusions are groupedaccording to which theme they concerns, for exam-ple persecution, megalomania, physical illness etc.A psychosis increases the risk of suicide, social iso-lation and exclusion from the labour market.

The risk of not treating the psychotic condition in preg-

nant women. The risk of a non-planned pregnancyis increased in women with psychotic disorders.Approximately 60% of women with such a condi-tion will carry out a pregnancy (96 – 98). Data onthe risk of relapse of psychotic symptoms in preg-nant women with a general psychotic disorder arenot present, but the risk of relapse is 65% in preg-nant women with schizophrenia. The attendingmedical doctor should inform the woman of risksof the current treatment already before any preg-

nancy, explain about a possible need for change of medicine during pregnancy and inform her aboutthe risk of abrupt discontinuation of medical treat-ment, such as insufficient nutrition, pregnancy com-plications and an increased risk of drug abuse (99).

If the woman has a desire for pregnancy, but themedical treatment is considered suboptimal in rela-tion to this, safe contraception must be agreedupon until changes in medical treatment have beenmade, and the initial risk of relapse by change of medicine is over. If the woman is already pregnant,a detailed plan must be made for collaborationbetween relevant parties, such as own medical

doctor, psychiatrist, obstetrician and municipalpartners.

In the treatment with antipsychotics, an increasein prolactin can be seen. The risk is highest in first-generation antipsychotics and risperidone, amisul-pride and paliperidone. The increased prolactinlevels may cause breast tension, milk secretion,irregular periods and reduced fertility.

In addition to focus on the pharmacologicaltreatment, the psychosocial support for the womanmust be optimized to ensure that the pregnancyruns as smoothly as possible and to minimize the

risk of recurrence of the psychotic symptoms.After the birth of the child, psychological condi-tions in connection with expectations and fearfrom the woman and her family as well as thephysical and psychological pressure to care for anewborn can contribute to an increased risk of relapse.


psychotic disorder. An increased incidence of pre-mature birth has been described among peopletreated with antipsychotics, but it is unclearwhether this is due to the pharmacological treat-

16

Larsen et al.


17/28

ment, or whether it is an expression of the underly-ing illness (100, 101). However, the risk is notincreased by the use of atypical antipsychotics asopposed to the use of first-generation antipsy-chotics. This may, however, be due to the fact thatit is the most seriously ill women who use first-gen-

eration antipsychotics (100, 101). Addition of anti-cholinergics against side-effects from antipsychoticsappears to increase the risk of malformations andshould if possible be avoided (102).

There are known metabolic side-effects of antipsychotics, but the risk of metabolic complica-tions in the newborn has not been clarified (100,103, 104). An increased risk of hypoglycaemia innewborns has been found if the mother has beentreated with atypical antipsychotics during thepregnancy (103, 104). The metabolic side-effectsfrom treatment with antipsychotics should be

assessed on an ongoing basis in the mother, as therisk of diabetes is increased.In children who have been exposed to antipsy-

chotics during the foetal stage, increased wakingstate, restlessness, crying and sleep problems (his-taminergic rebound) and increased saliva secretionand diarrhoea (anticholinergenic rebound) imme-diately after the birth can be observed.

Treatment with antipsychotics


during pregnancy. Olanzapine is primarily recom-

mended based on the amount of safety data. Otherconditions may, however, be of importance for thechoice of drug, including metabolic risk profile.Earlier, perfenazine† was recommended, but datafor olanzapine, quetiapine and risperidone arecomparable with this and do not indicate anincreased risk of congenital malformations. ECTmay be considered.


ment. The pregnant woman will most often be intreatment with an atypical antipsychotic. As

described below, data for a number of medicinalproducts are available, and the need for changewill depend on a specific assessment of the patient’srisk of recurrence compared to the product’s riskprofile.

Breast-feeding. RID is typically in the range of 0 – 2% for most of the antipsychotics, and side-effectsin nursed children are only described exceptionally.Factual data for all the drugs, however, are not

available (14). We refer to the specific drug descrip-tions below.

Background. Prescription and control of patientsundergoing treatment with antipsychotics shouldbe performed in collaboration with specialist in

psychiatry. In 2011, FDA issued a general warningin relation to the use of antipsychotics during preg-nancy, primarily due to an increased incidence of extrapyramidal symptoms and discontinuationsymptoms (9). The first symptoms indicate that thebrain of the child as well as of the mother isaffected. The data volume for use during preg-nancy is significantly smaller for antipsychoticscompared with antidepressants, and in fact, onlyolanzapine meets the requirements of this guide-line’s formal criteria of more than 1000 exposedpregnant women. The body’s enzyme systems have

increased activity during pregnancy, which maylead to olanzapine and clozapine being metabo-lized quicker than normal. Certain antipsychotics,such as haloperidol and perfenazine, have a rea-sonable defined therapeutic window, which is whyit is recommended to regularly measure antipsy-chotics in the blood, for example every 3 monthsduring the pregnancy.

An increased incidence of premature birth hasbeen described among women treated withantipsychotics, but it is unclear whether this is dueto the pharmacological treatment, or whether it isan expression of the underlying illness (confound-

ing by indication) (100, 101).

Newer antipsychotics. Pregnancy. Olanzapine:Olanzapine is the newer antipsychotic with themost data for use during pregnancy. Data forabout 1100 first trimester exposed show no signs of excess incidence of undesired foetal impact (105,106). Treatment with olanzapine during pregnancycan be used if there is a clear indication for this.Dose during the last part of the pregnancy shouldbe kept as low as possible. A case report has foundan increased risk of changed glucose tolerance

during pregnancy in the mother (107).Quetiapine: About 450 cases of pregnant women

treated with quetiapine have been reported. Nopattern of malformations which could indicate aspecific teratogenic effects of quetiapine isobserved, and no increased prevalence of malfor-mations compared to baseline risk is seen (8, 105,108, 109). The safety and the effect of quetiapineduring pregnancy are insufficiently described, andthe treatment is not recommended as a rule. How-ever, there may well be clinical courses, where it isnot rational to switch from an efficient ongoingtreatment with quetiapine.

†Regarding prescription of perphenazine – see under treatment with

mood stabilizers.

17



18/28

Aripiprazole: Not recommended during preg-nancy due to insufficient data.

Risperidone: Data for more than 400 first trime-ster pregnant women treated with risperidoneshow no increased risk of malformations (8, 105,108). As a rule, risperidone is not recommended

during pregnancy, as the safety is inadequatelyestablished. However, there may well be clinicalcourses, where it is not rational to switch from anefficient ongoing treatment with risperidone.

Paliperidonie: Not recommended due to insuffi-cient data. There are few specific data, but extrapo-lation from risperidone data is to some extentacceptable, as paliperidone is the active metaboliteof risperidone.

Clozapine: About 200 cases of pregnant womentreated with clozapine have been reported. No pat-tern of malformations is observed. Clozapine is not

recommended due to insufficient data. It can beclinically indicated to use clozapine during preg-nancy for women who have had no effect of anothertreatment, and who therefore cannot be switched toone of the recommended antipsychotics.

Ziprasidone: Not recommended due to insuffi-cient data.

Breast-feeding.

Olanzapine: Can be used. RID is below 2%, andno side-effects have been described in nursed chil-dren.

Quetiapine: Can be used. RID is below 1%, andside-effects in nursed children are not reported.Aripiprazole: Can be used. RID is below 1%, andside-effects in nursed children are not described.Risperidone: As a rule, not recommended. RID isbetween 3 and 9%.Paliperidone: Not recommended due to insufficientdata.Clozapine: As a rule, not recommended. RID isbelow 2%, but because of the side-effect profile,clozapine is not recommended during breast-feed-ing (14, 110).Ziprasidone: Not recommended due to insufficient

data.

Older antipsychotics. Pregnancy.

Perphenazine: Data exists for 500 exposed withoutsigns of increased incidence of undesired foetalimpact. In addition, there are some data for per-phenazine administered in lower doses as aantiemetics.Haloperidol : Not recommended. Data exist forabout 300 first trimester-exposed without signs of significantly increased incidence of malformations.

!!!!use of psychotropic drugs during pregnancy and breast feeding

Documents