Alwan, 2007

Retrospective case control study.

N= 13714

Inclusion:

Case infants were born alive or died at 20 weeks or more of gestation.

Pregnancies with reliably ascertained defects that were electively terminated. Only one case or control infant from each multifetal pregnancy.

The controls were live-born infants with no major birth defects who were randomly selected

from hospital or state birth-certificate records from the same geographic areas.

Exclusion:

Infants with recognized or strongly suspected chromosomal abnormalities or single-gene conditions were excluded from the study.

To reduce pathogenetic heterogeneity,

cases with complex sequences (e.g., omphalocele-exstrophy-imperforate anus-spinal defects phenotype) 11 were excluded.

Infants with major birth defects.

N=9622

Exposed to SSRI N=230

N=4092

Exposed to SSRI N=86

Mothers completed a standardized telephone interview regarding exposure to potential risk factors, including medications, before and during pregnancy.

SSRI in general:

There were no significant associations between maternal use of SSRIs overall during early pregnancy and congenital heart defects or most other categories or subcategories of birth defects. Maternal SSRI use was associated with anencephaly (214 infants, 9 exposed; AOR 2.4; 95% CI 1.1-5.1), craniosynostosis (432 infants, 24 exposed; AOR 2.5; 95% CI 1.5-4.0), and omphalocele (181 infants, 11 exposed; AOR 2.8, 95% CI 1.3-5.7).

SSRI seperately:

- fluoxetine use was significant associated

with craniosynostosis (10 exposed infants;

AOR 2.8; 95% CI 1.3-6.1)

- sertraline use was significant associated with anencephaly (4 exposed infants; AOR 3.2; 95% CI 1.1-9.3).

- Four significant associations for paroxetine: with anencephaly (five exposed infants; AOR 5.1; 95% CI 1.7-15.3), right ventricular outflow tract obstruction defects (7 exposed infants; AOR 2.5; 95% CI, 1.0-6.0), omphalocele (6 exposed infants; AOR 8.1; 95% CI

3.1-20.8), and gastroschisis (5 exposed infants; AOR, 2.9 95% CI 1.0-8.4).

Clearly defined groups: Yes.

Selection bias: They did a power measurement. Participation < 100%. -> no relevant effect expected on outcome.

Adequate exposure assessment method: yes. Groups clearly described. Medication not clearly described -> women were prompted by brand names -> leading to possible underreporting of other SSRI use, malformations clearly described. Methods clearly described.

Method of assessing the outcome appropriately: Recall bias (telephone interview). Analysis limited cause of small sample size for each SSRI separately.

Impactassessor blinded: no.

Follow-up: retrospective study infants born from 1997 through 2002. 2 interview moments.

Selective loss to follow up: Infants of mothers with prepregnancy type 1 or 2 diabetes were excluded from adjusted analyses because of the strong association of diabetes with birth defects. Infants for whom complete maternal interviews were unavailable were excluded. Interviews were incomplete for 163

case or control mothers, and 5 mothers reported depression but did not report use of antidepressants. These subjects were excluded from the analysis.

Identification confounders and correction in analysis:

Potential confounders: maternal race or ethnic group (non-Hispanic white vs. other), maternal age,  maternal education, prepregnancy obesity, BMI, maternal smoking or alcohol use, maternal folic acid use, family income, singleton

versus multiple pregnancy, parity and presence or absence of prepregnancy hypertension.

-> they did different analysis for correction for confounders.

Funding: Dr. Friedman reports receiving honoraria for consulting from I3 Research. No other potential conflict of interest relevant to this article was reported.           

Bakker, 2010

Retrospective case-control study (Netherlands)

Exclusion:

Children for whom there was no information on maternal medication use and deceased children and fetuses who were not examined postmortem.

N= 678

Fetuses or children

with isolated congenital heart defects.

N=615

Fetuses and children with a genetic disorder with no heart defect

Exposure paroxetine in first trimester pregnancy.

After excluding mothers who used paroxetine outside the first trimester, or who had used another SSRI, we found no significantly increased risk for heart

defects overall (10 exposed cases; AOR, 1.5; 95% CI 0.5–4.0), but we did find a significantly increased risk for atrium septum defects (3 exposed cases; AOR, 5.7; 95% CI 1.4–23.7)

Clearly defined groups: yes.

Selection bias: Yes.

But significant heterogeneity between groups and no correction. (year of birth, maternal age, mothers educational level and pregnancy outcome). The difference in pregnancy outcome between included and excluded controls remained after stratifying for year of birth.

Adequate exposure assessment method: Yes.

Method of assessing the outcome appropriately: Yes.

Follow-up: children born between 1997-2006

Selective loss to follow up: Reasons for exclusion was described in detail and also checked if there were significant differences between the lost to follow up group and the study group.

Identification confounders and correction in analysis: Maternal age and pregnancy outcome was not

associated with exposure status, so they calculated ORs adjusted for year of birth alone. They did not calculate ORs adjusted for other potential important confounders as smoking, maternal disease, or use of other (teratogenic) drugs.

Funding: Not described.

Bérard, 2007

Case-control study, retrospective

N=1403

- Paroxetine N=542

- Other SSRI, not paroxetine N=443

- Other antidepressants, not SSRI N=418

Three administrative databases (RAMQ, Med-E´ cho, and ISQ), and includes all pregnancies in

Quebec between 01/01/1997 and 06/30/2003.

Inclusion:

1) 15–45 years of age at entry

2) be covered by the

RAMQ drug plan ≥ 12 months before and during pregnancy

3) be using one type of antidepressant during the 1st trimester

4) live birth.

Exclusion:

Women also exposed to known teratogens during pregnancy were excluded.

N=101

1) any major malformations

2) any cardiac malformations diagnosed in the first

year of life

N=1302

no major malformations

1) Association between 1st trimester exposure to paroxetine and congenital (cardiac)

malformations

2) Dose–response relationship between paroxetine

use and cardiac defects.

1) The use of paroxetine (adjusted OR 1.38, 95% CI 0.49–3.92), and the use of other SSRIs (adjusted OR 0.89, 95% CI 0.28–2.84) during the 1st trimester of pregnancy did not increase the risk of congenital cardiac malformations compared with the use of non-SSRI antidepressants.

2) Dose–response relationship women exposed to >25 mg/day of paroxetine during the 1st trimester of pregnancy were at increased risk of having an infant with major congenital malformations (adjusted OR 2.23, 95% CI 1.19-4.17), or major cardiac malformations (adjusted OR 3.07, 95% CI 1.00-9.42).

Clearly defined groups: Yes.

Selection bias: It is a retrospective design, clear described inclusion and exclusion criteria.

Adequate exposure assessment method: Yes. Very detailed description of the methods and groups.

Method of assessing the outcome appropriately: Yes. Also detailed description of outcomes

Impactassessor blinded: No.

Follow-up: Delivery between 01/01/1997 and 06/30/2003

Selective loss to follow up: Not described. But retrospective design.

Identification confounders and correction in analysis: Yes. They were not able to measure smoking status, folic acid intake, alcohol use, pregnancy weight gain, and over-the counter (OTC) medication use. However, they indirectly adjusted for lifestyles such as smoking and alcohol intake by design given that they restricted the study population to antidepressant users and they adjusted for other confounders.

Funding: Not described.

Cole, 2007

Case control studie, retrospective (US)

N=5956

Inclustion:

All women with at least one dispensing of an

antidepressant (including selective serotonin reuptake

inhibitors, selective norephinephrine reuptake inhibitors,

serotonin-2 antagonist reuptake inhibitors, tricyclics, and monoamine oxidase inhibitors) and

who had a live-born delivery between the ages 12 and

49 during January 1995 and September 2004.

Inclusion into

the final case group were: (1) a structural abnormality with surgical, medical, or cosmetic significance,14

(2) available medical record documentation within the year following birth, and (3) documentation of the

defect in question through the abstracted medical record.

Exclusion:

minor anomalies, birthmarks,

physiologic conditions, known chromosomal

or genetic conditions, and cases where evidence was insufficient to confirm the diagnosis.

N=815 (Monotherapy) -> paroxetine

N=1020 (Polytherapy) -> paroxetine

N=4198 (Monotherapy) -> other antidepressants than SSRI’s.

N=4936  (polytherapy)-> other antidepressants than SSRI’s.

The primary outcome under study was all congenital malformations identified up to 9 months of age among infants. They grouped congenital malformations into

categories according to organ system.

Comparison Paroxetine versus antidepressive use in general.

Adjusted ORs for all congenital

malformations associated with paroxetine were 1.89 (95%CI 1.20–2.98) for monotherapy, and 1.76 (95%CI 1.18–2.64)

for mono- or polytherapy.

Adjusted ORs for cardiovascular malformations associated with paroxetine were 1.46 (95%CI 0.74–2.88) for monotherapy, and 1.68 (95%CI 0.95–2.97) for mono- or polytherapy.

Clearly defined groups: Yes.

Selection bias: Yes. Not all of the charts of suspected case infants could be abstracted -> If these infants had been confirmed as malformation cases with the same frequency as the other infants with suspected malformations, we would expect the odds ratios to be closer toward the null value of 1.0 than the observed odds ratios.

Method of assessing the exposure and outcome appropriately: Not entirely clear. No information about other depressants used. Nothing known about the dosage of medication used.

Impactassessor blinded: No.

Follow-up: January 1995 and September 2004.

Selective loss to follow up: Not described.

Identification confounders and correction in analysis: calculated the adjusted odds ratios (AORs) through logistic regression. In all analyses, the crude and AORs were

similar, indicating that there was little or no confounding by variables identified either a priori or empirically in the medical and pharmacy claims data

Funding: Yes. This research was supported by a research contract between i3 Drug Safety and GlaxoSmithKline.

Einarson, 2008

Prospective patient controlled study

N=1174

Population from two sources: teratology information services and database studies.

Database study results not included for this evidence table cause of retrospective onset and no clear information about sort SSRI used.

Inclusion:

women that call to inquire about the use of a drug they are taking and are currently pregnant. Eligible women are prospectively enrolled in the study after providing informed consent over the telephone.

N= 1174

Exposed to paroxetine in 1st trimester of the pregnancy. Maximum daily dose 20mg.

Not exposed to paroxetine in 1st trimester of the pregnancy

Telephone interview twice, in pregnancy and after delivery, questionnaire.

Primary outcome:

Cardiovascular birth defects

The rates of cardiovascular defects in the teratology information service cohort were 0.7% in the exposed group and 0.7% in the unexposed group (OR=1.1, 95% CI=0.36–2.78).

Demographics were matched. No difference between the groups. 

Clearly defined groups: Yes.

Selection bias: Yes. No clearly defined inclusion and exclusion criteria.

Adequate exposure assessment method: Yes.

Method of assessing the outcome appropriately: Yes.

Impactassessor blinded: No

Follow-up: Telephone interview twice, in pregnancy and after delivery, questionnaire. No clear interval and no starting and ending data.

Selective loss to follow up: Not described.

Identification confounders and correction in analysis: Not described. They asked the women about smoking habits, alcohol use and ses.

Funding: No support received has been related to paroxetine or any other SSRI. The remaining authors report no competing interests.

Kornum, 2010

Retrospective cohort study

Danish population of 3 clinical centers 1991-2007.

Inclusion criteria:

All women in the counties who had a live birth after the 20th week of gestation during the periods covered by the prescription registries.

Exclusion criteria:

women who redeemed prescriptions for both a SSRI and a non-SSRI antidepressants in the period from 30 days

before conception until the end of each pregnancy.

Use of antiepileptics or antidiabetic drugs was an exclusion criteria.

Malformations with SSRI use (overall)

Malformations with non SSRI use.

Four categories of malformations:

-malformations

overall

-noncardiac malformations

-cardiac malformations

-septal defects diagnosed within

the infant’s first year of life

SSRI in general:

The 2062 women with SSRI prescriptions during early pregnancy gave birth to

105 (5.1%) infants with malformations, while the 213712 women with no SSRI prescriptionsgave birth to 7449 (3.5%) infants with malformations. SSRI use was associated with an increased risk of malformations overall (OR 1.3; 95% CI 1.1–1.6) and cardiac malformations (OR 1.7; 95% CI: 1.1–2.5).

SSRI seperately:

For specific SSRIs, we found an increased risk for septal defects associated with sertraline. Cardiovasculair defects

(OR 3.0; 95% CI 1.4-6.4), Septal defects (OR 3.3; 95% CI 1.5-7.5).

Clearly defined groups: Yes.

Selection bias: Possible. Retrospective design, missing data. 3 clinical centers with different follow up periods. No information about over the counter medication.

Adequate exposure assessment method: Yes.

Method of assessing the outcome appropriately: Yes. They did not consider malformations leading to a miscarriage or to an elective abortion after prenatal diagnosis in this study, that’s a limitation. No exposure information. Exposure window large.

Impactassessor blinded: No.

Follow-up: 1991-2007

Selective loss to follow up: Not described.

Identification confounders and correction in analysis: Potential confounders, including maternal age, birth order, and maternal smoking status. Adjusted analysis for maternal smoking during pregnancy, birth order, maternal age, and birth year

Funding: authors report no conflicts of interest.

Kulin, 1998

Prospective controlled cohort study

N= 534

Inclusion criteria:

All women who contacted 1 of 9 participating Teratology Information Service centers regarding exposure to fluvoxamine, paroxetine, and sertraline during the first trimester of pregnancy for depression.

Exclusion criteria: women who, in addition to being exposed to a new SSRI, were also exposed to a known human teratogen or drugs of uncertain teratogenicity.

N= 267

Use of:

Sertraline N=147

Paroxetine N=97

Fluvoxamine N=26

(dosage also described)

N= 267

Controls were randomly selected

from the total group of women counseled

and followed by the Mother risk Program

after exposure to agents proven to be nonteratogenic.

Primary outcome: the rates of major malformations, (defined as structural or functional anomalies that have significant medical or social consequences)

The SSRI and control groups were compared on a variety of

baseline characteristics and pregnancy

outcome values.

SSRI use in general:

Pregnancy outcome did not differ between the groups. with similar rates of major malformations, spontaneous and elective abortions and stillbirth, and similar mean birth weight and gestational age. The RR for major malformations among SSRI-exposed neonates was 1.06 (95% CI, 0.43-2.62)

Demographics:

Women exposed to an SSRI were significantly less likely to be primigravid and significantly more likely to smoke cigarettes and to have had a previous therapeutic abortion than the 267control women These trends were homogeneous among the SSRIs. Correction for smoking did not change the outcome.

Clearly defined groups: Yes.

Selection bias: No. But no (blind) randomisation. Clearly described inclusion and exclusions.

Adequate exposure assessment method: No blindness to exposure.

Method of assessing the outcome appropriately: Yes.

Impactassessor blinded: No.

Follow-up: Fixed times for interviewing no data pronounced when the study started and ended.

Selective loss to follow up: No loss to follow up.

Identification confounders and correction in analysis: No description if they corrected for (all) confounders. But they did a correction for smoking.

Funding: Not described.

Louik, 2007

Retrospective case control study

Inclusion:

Infants with any of a wide range of malformations are identified.

Exclusion:

Children with isolated minor defects.

N=9849 with malformations

N=5860 without malformations

Telephone interview

Per malformation they checked the use of SSRI in pregnancy.

SSRI use in general:

Overall use of SSRIs was not associated with significantly increased risks of craniosynostosis (N=115, 2 exposed to SSRIs; AOR 0.8; 95% CI, 0.2-3.5), omphalocele (N=127, 3 exposed; AOR 1.4; 95% CI, 0.4-4.5), or heart defects overall (N=3724, 100 exposed; AOR 1.2; 95% CI, 0.9-1.6).

SSRI use seperately:

Significant associations between the use of sertraline and omphalocele (AOR 5.7; 95% CI 1.6-20.7 N=3) and septal defects (AOR 2.0; 95% CI 1.2-4.0; N=13) and between the use of paroxetine and right ventricular outflow tract obstruction defects (AOR, 3.3; 95% CI, 1.3-8.8; N=6)

Clearly defined groups: Yes

Selection bias: Possible, retrospective study, not clearly described inclusion and exclusion criteria. 

Adequate exposure assessment method: Yes. There is a possible recall bias.

Method of assessing the outcome appropriately: Yes. Telephone interview. Nothing known about the dosage of medication used.

Impactassessor blinded: Interview done bij trained study nurses who are unaware of the study hypotheses.

Follow-up: retrospective.

Selective loss to follow up: Malformation with

< 5 SSRI users in the group were not analysed.

Identification confounders and correction in analysis: They did a correction for important confounders (Maternal age, maternal race, ethnic group, maternal education, year of LMP, study center, smoking and alcohol 1st trimester, history of a birth defect in first degree relative, BMI, parity, presence/absence of seizures, DM, hypertension, infertility, folic acid use 1st trimester).

Funding: Yes. Publication bias.

Malm, 2005

Prospective population based cohort study (Finland)

N=1782

Inclusion:

Singleton pregnancies

Exclusion:

Not described

N=1398

SSRI use 1st, second or/and third trimester:

Citalopram, fluoxetine, paroxetine, sertraline, fluvoxamine, combination.

N=1782

No SSRI use 1st, second or/and third trimester

Primary outcomes:

- Major malformations

- Preterm birth

- SGA

- LBW

- NICU stay

Also division in SSRI intake per trimester.

Major congenital anomaly defined as a significant

congenital structural anomaly, chromosomal

defect, or congenital hypothyroidism.

Malformations 1st trimester exposure (adjusted):

SSRI use in general:

- All SSRI’s (N=1,398) P =0 .40, OR=1.0, 95% CI 0.6-1.7

SSRI use seperately:

- Citalopram (N=554) OR=0.4, 95% CI 0.1-1.4

- Fluoxetine (N=525) OR=0.7, 95% CI 0.9-3.3

- Paroxetine (N=152) OR=0.4, 95% CI 0.1-3.3

- Sertraline (N=118) OR=0.7, 95% CI 0.1-5.3

- Fluvoxamine (N=65) OR=1.0, 95% CI 0.1-7.8

SGA (adjusted):

OR=2.4, 95%

CI 1.1–5.3 -> continues exposure vs only exposure in first trimester.

NICU (adjusted):

OR=1.6, 95% CI 1.1-2.2 -> third trimester SSRI intake.

Apgar score (adjusted):

OR=1.6, 95% CI 1.0-2.4 -> third trimester SSRI intake.

Rest of the outcomes show no significant effect.

Clearly defined groups: Yes.

Selection bias: Not certain. But only one inclusion criteria and no exclusion criteria described.

Adequate exposure assessment method: Yes. But no information about dosage of the medication. No information about co-medication.

Method of assessing the outcome appropriately: Yes.

Impactassessor blinded: No.

Follow-up: 1996-2001

Selective loss to follow up: Not described.

Identification confounders and correction in analysis: They did a correction for confounders.

Funding: Not described.

Merlob, 2009

Prospective cohort study, hospital based

Inclusion:

All women on admittance to the maternity ward (Departments of Neonatology in Rabin Medical Center and Schneider Children’s Medical Center of Israel)

Only inborn infants.

Exclusion:

Not described.

N=235

Use of SSRI in pregnancy.

(Nonsyndromic congenital heart defects 8/235)

N= 67636

No use of SSRI in pregnancy

(Nonsyndromic congenital heart defects 1083/67636 – clearly defined loss to follow up describtion -> exclusions)

A standardized pregnancy questionnaire is administered to all women and  reviewed by the attending neonatologist.

Investigate the rate of nonsyndromic congenital heart defects in newborns exposed in utero to SSRIs compared with unexposed controls

SSRI use in general:

Newborns exposed in utero to SSRIs (without differentiation between sort SSRI), have a 2 fold higher risk of mild nonsyndromic heart defects than unexposed infants (RR, 2.17; 95% CI, 1.07–4.39)

(The types of SSRIs paroxetine (n=92), fluoxetine (n=66), citalopram (n=43), escitalopram (n=13), sertraline (n=8), fluvoxamine (n=4), and venlafaxine (n=9) -> no analysis of RR/OR of each type of SSRI separately.)

Clearly defined groups: Yes.

Selection bias: No. But not clear described inclusion criteria.

Adequate exposure assessment method: Yes. As certainment was based on maternal reporting, and some mothers might have chosen not to report use of SSRIs as well as substance abuse. Nothing known about the dosage of medication used.

Method of assessing the outcome appropriately: Yes. Every infant with a persistent murmur on the second or third day of life were examined by a pediatric cardiologist and referred for electrocardiography and echocardiography.

Impactassessor blinded: During the last 2 years of the study, the cardiologists were not blinded to SSRI exposure.

Follow-up: 2000-2007

Selective loss to follow up: Not real lost to follow up, but they did not include all infants in the control group -> exclusions and reasons were described.

Identification confounders and correction in analysis: lack of information on some potential confounders such as race and maternal BMI. No correction for  confounders described, no adjusted RR analysis.

Funding: No.

Oberlander, 2008

Retrospective cohort study

Inclusion: Not described.

Exclusion:

Women with anticonvulsant exposure in first trimester.

Exposure groups were defined as SRI monotherapy

And SSRI + benzodiazepines (BZ) used in combination.

Monotherapie SSRI

N = 2626

Citalopram (n=101),

Fluoxetine (n=638), Fluvoxamine (n=119),

Paroxetine (n=993),

Sertraline (n=608),

Venlafaxine (n=250)

Non exposure

N=107320

Major congenital anomalies (CA), congenital heart disease (CHD incl. VSD+ASD).

Also risk analysis for each SSRI separately.

Also risk analysis for dosage of medication.

SSRI use in general:

- SSRI monotherapy did not

increase the risk for major CAs

- SSRI monotherapy did not increase the risk for CHD overall compared with no exposure.

- The risk for an atrial septal defect (ASD) 

increased with SSRI monotherapy compared with no exposure RD 50.16% (0.03 to 0.43%)

SSRI use seperately:

- An increased risk for cardiovascular defects was observed following exposure to citalopram (n=53; 2.97%; 95% CI (1.02–8.37); RD 2.49% (0.54–7.89) when compared to no exposure. While this risk difference remained statistically

significant after controlling for potential confounders

Doses relation:

- No association between increased 1st trimester dose and risk for CA or CHD in the simple comparisons or in the regression models, for SRI monotherapy.

Clearly defined groups: Yes

Selection bias: Yes. Retrospective design. Not clearly described inclusion or exclusion criteria.

Adequate exposure assessment method: Yes.

Method of assessing the outcome appropriately: Yes.

Impactassessor blinded: No.

Follow-up: 39 months 1998-2001.

Selective loss to follow up: Not described-> retrospective study design so possible no loss to follow up!

Identification confounders and correction in analysis: Yes. With every analysis they corrected for confounders. But not for all important confounders like use of tobacco and alcohol during pregnancy or maternal weight gain, illicit drug use, or maternal parity.

Funding: Yes. But no pharmaceutical companies.

Pastuszak, 1993

Prospective patient controlled study

Inclustion:

All women that contacted one of the TIS during pregnancy.

There were no differences in any of the characteristics of the women using fluoxetine across the four participating centers.

Fluoxetine use in 1st trimester of pregnancy (mean daily dose of 25.8mg range 10-80mg)

N= 128

1. nonteratogen medication users (NTC) N=128

2. Tricyclic antidepressants users(TC) N= 74

Telephone interview 8-12 months after the expected date of delivery.

Primary outcomes:

- Major congenital anomalies

- Weight gain (mother)

- Gestational age with delivery

- Birth weight

- Type of delivery

No difference in rates of major birth defects

Fluoxetine vs NTC  p=0.38

Fluoxetine vs TC p=0.80

Other primary outcomes showed also no statistic significance between the groups.

(there were significantly more miscarriages in the fluoxetine and the TC group in compare with the NTC group.)

Clearly defined groups: Yes.

Selection bias: No. But each fluoxetine user was only age-matched to two controls. No exclusion criteria.

Adequate exposure assessment method: Yes. Also dosage info.

Method of assessing the outcome appropriately:  Yes. Telephone interview. Clearly described outcomes.

Follow-up: Telephone interview 8-12 months after the expected date of delivery.

Selective loss to follow up: No loss to follow up.

Identification confounders and correction in analysis: Maternal characteristics (obstetric histerory, ethanol and sigarette use) were also compared.  But they did no correction for confounders.

Funding: Not described.

Sivojelezova, 2005

Prospective patient controlled study

N=396

Inclusion:

The women were recruited from the cohort of pregnant women or women planning pregnancy who contacted the Motherisk Program from 1999 to 2002 inquiring about the safety of citalopram and other medications in pregnancy

Exclusion criteria:

Exposure to a known teratogen or a xenobiotic with undetermined safety in pregnancy. Premature infants and infants

exhibiting intrauterine growth retardation.

Citalopram use N=132

Only those women

who had an exposure to the medication during the period of organogenesis were included.

1. Disease-matched Group (other SSRI use) N=132

2. Non teratogen group N=132

Telephone follow-up interview

Risk of adverse pregnancy outcome, birth defects and neonatal complications.

(Major malformations

were defined as structural and/or functional anomalies that have to be corrected surgically or that

may alter the social acceptability of the individual.)

SSRI use seperately:

Of the 108 women who were exposed to citalopram in the first trimester, there was only 1 (0.9%) baby born with a major malformation (umbilical and scrotal

Herniae). Not significant different from the other groups.

Third-trimester exposure to citalopram did not differ statistically from that of the unexposed group of women.

There was a 4-fold increased risk for citalopram use of need for the NICU (relative risk, 4.2 [95% CI 1.71-10.26)

Maternal characteristics not statistically different from the disease match comparison -> but smoking was significant different in comparison with the non exposed group.

Clearly defined groups: Yes.

Selection bias: Not clearly described selection criteria. Selection and measurement method was clearly described. Groups were matched for maternal age and gestational age. But other demographic matches/differences were not described.

Adequate exposure assessment method: Yes.

Method of assessing the outcome appropriately: Yes.

Impactassessor blinded: No.

Follow-up: 1999-2002, fixed intervals for follow up

Selective loss to follow up:

Not described. But most loss to follow up can be explained with the exclusion criteria.

Identification confounders and correction in analysis: They did no correction for smoking. The exposure group had significant more smokers.

Funding: Not described.