The prevalence of depression during pregnancy has been estimated to be approximately 10%, with no established major differences among countries.1 There are no epidemiologic studies specifically evaluating the use of selective serotonin reuptake inhibitors (SSRIs) among pregnant women from different countries, but less than 1% of pregnant women in Europe,2,3 and approximately 3–4% in the United States4 have been reported to use antidepressants in general. Inadequate treatment of maternal depression may lead to worsening of the condition and therefore also indirectly affect fetal well-being.5,6 When safe use during pregnancy has not been established, physicians may hesitate to prescribe antidepressants to pregnant women. The SSRIs have a favorable side effect profile and are, accordingly, being used to an increasing degree. However, possible untoward effects of SSRIs on the unborn child or the neonate could raise serious public health concerns and should be recognized. Most previous studies on the safety of SSRIs in pregnancy have been based on small pregnancy numbers. In addition, study designs and methods for detecting adverse pregnancy outcome have varied, making conclusions difficult. No definitive increase in risk of malformations has emerged so far,7–13 but some studies have found evidence for increased risk of preterm birth.8,14 In addition, several case reports, cohort studies based on small numbers, and 1 database analysis based on spontaneous reporting have described an association between neonatal problems and third trimester exposure.8,14–19
We carried out a population-based study to investigate the safety of SSRIs during pregnancy. Our aim was to study whether exposure to SSRIs during early pregnancy is associated with an increased risk of major malformations. The second objective was to study the effect of continuous exposure (throughout pregnancy) on length of gestation and birth weight. We also wanted to investigate the possible association of third trimester exposure and risk for adverse perinatal outcome.
MATERIALS AND METHODS
The data for our study were derived from a national joint project in Finland, established by 3 governmental organizations: the National Research and Development Centre for Welfare and Health (STAKES), the Social Insurance Institution in Finland, and the Finnish National Agency for Medicines for continuing surveillance of drug-related safety during pregnancy. In this project, the following registers have been linked by a personal identification number unique to all citizens.
The Medical Birth Register collects maternal background data, maternal pregnancy-related medical data, delivery data (live births and stillbirths), and neonatal outcome data including malformations, until the age of 7 days. All infants born in hospitals are examined by a pediatrician before discharge. All births of infants or fetuses with gestational age of at least 22 weeks or birth weight of 500 g or more are included in the register. Data for the register are collected from all maternity hospitals, and in the case of home births, from the assisting midwife or physician. The register data are forwarded to STAKES and missing data are supplemented by data from birth and death certificates. The Medical Birth Register data are considered to be a complete record of all births and newborns in Finland.20–22
The National Register of Congenital Malformations defines a major congenital anomaly as a significant congenital structural anomaly, chromosomal defect, or congenital hypothyroidism. The registry collects information from the whole of Finland on all newborns with a birth defect, using several data sources.23 In cases of a malformation, the delivery units are obliged to complete and forward a special data collection form to the Malformation Register. Data on diagnosis are also collected and confirmed from the Medical Birth Register, the Hospital Discharge Register, the Cause-of-Death Register, and from cytogenetic laboratories until the age of 1 year.23 In addition to maternity hospitals, health care professionals in outpatient clinics, such as child welfare clinics, provide cases that have been detected after discharge from the delivery unit to the register. All pregnancy terminations due to fetal malformation are also recorded in the register.23
The National Register on Induced Abortions is another register maintained in Finland. According to national legislation in Finland, termination of pregnancy is allowed until the end of the 20th gestational week for social reasons, and until the end of the 24th gestational week if a severe major fetal malformation has been detected. A notification of pregnancy termination must be sent to the registry within 1 month after the termination. The registry covers more than 99% of induced abortions registered in the hospital records.24 Due to the liberal legislation and free access to public health care and social welfare units, illegal terminations do not occur. The Cause-of Death Statistics complied by Statistics Finland contains data on postnatal mortality.
The Drug Reimbursement Register covers all permanent residents in Finland. The registry contains data on all reimbursed prescription drug purchases and data on chronic illnesses requiring continuous drug treatment. In Finland, prescription-only medicines deemed necessary for the treatment of an illness are reimbursed under the Health Insurance Scheme.25 Medicines are supplied to the patient only for 3 months at a time from the pharmacy; therefore, a 1-year prescription will contribute to several purchases in the reimbursement database.
The estimated length of gestation at delivery—based on the last menstrual period and ultrasound examination—is recorded in the Medical Birth Register. In the national joint project, beginning of pregnancy has been defined by subtracting the number of days corresponding to gestation length from date of birth. In the combined database, data on drug purchases have been collected during preconception (3 months before pregnancy), each pregnancy trimester, and 3 months postpartum. Pregnancy trimesters have been defined as 0–12 weeks (first trimester), 13–26 weeks (second trimester), and 27 weeks onward (third trimester). A total of 405,892 mother plus child or fetus pairs have been registered in the database during the period of 1996–2001.
The study protocol was approved by the national health authorities, the ethical committee in STAKES, and the Finnish Data Protection Authority in conjunction with establishment of the joint project.
The study material was derived from the joint project database described above and covers the years 1996–2001. Cases were defined as women who had at least 1 purchase of an SSRI drug (defined by the 5th level of the Anatomic Therapeutic Chemical Classification of Drugs code) during the period of 1 month before pregnancy and the day when pregnancy ended (n = 2,077). Only singleton pregnancies were included in the study. Women with chronic illnesses requiring continuous medication (such as hypertension, epilepsy, psychosis) were excluded from the analysis (n = 273). Additionally, another 22 cases were excluded because their matched pair in the control group had a chronic illness. The number of women with SSRI purchases included in the study was 1,782.
Controls were women with no reimbursed drug purchases during the period covering 1 month before and throughout pregnancy. They were first matched with cases by year of pregnancy ending, age, parity (no previous deliveries or 1 or more previous deliveries), geographic area (university hospital district), and social status. Matching was then done on one-to-one basis by randomly selecting 1 control for each case from the case-specific matched control pool. The matching control pairs for the cases that were excluded due to chronic illnesses were excluded (n = 273), as were the 22 women who had records of having a chronic illness. The number of controls included in the study was 1,782.
Low birth weight was defined as birth weight less than 2,500 g and was considered only in the analysis for birth register cases. Small for gestational age (SGA) was defined as a birth weight of more than 2 standard deviations less than the sex- and length-of-gestation–specific national standards26 and was considered only in the analysis for birth register cases. In the Medical Birth Register, only 1-minute Apgar score is recorded. Low Apgar score was defined as Apgar score less than 7. A crude classification of social status was made by grouping upper officials in the higher category and lower officials, workers, students, and housewives in the lower category. Previous deliveries were defined as no previous deliveries (nulliparity) or 1 or more previous deliveries. Smoking status was considered positive if the records showed tobacco smoking anytime during pregnancy. All other reimbursed drug purchases were also recorded in relation to pregnancy trimesters. Only major malformations, defined according to the definition of the Finnish Register of Congenital Malformations, which follows International Classification of Diseases, 9th Revision, and, excluding minor malformations according to the exclusion list of Eurocat,23 were included in the study. Minor malformations are not routinely reported and therefore were not analyzed.
The statistical analyses were performed using the SAS 8.2 (SAS Institute Inc., Cary, NC) statistical software. In comparisons made between one-to-one matched case-control pairs, McNemar test for categorical variables and paired t tests for continuous variables were performed. In comparisons made between different subgroups of cases to all controls or to cases with only first trimester purchase(s), univariate analysis to study differences between proportions were performed using χ2 or Fisher exact test for dichotomous variables and t test for continuous variables (all 2-tailed). Variables found to differ between the study and control groups (P < .1), and those that could be expected to confound the association between other independent variables and dependent variables, were included in the multivariate analysis. Logistic regression analysis was carried out applying major malformations, low Apgar score, treatment in special or intensive care unit, low birth weight, and SGA as dependent outcome variables. Independent variables considered in the model were purchase(s) of SSRIs, low social status, smoking, artificial reproductive techniques (ART), previous deliveries, age, and other purchased medications. All variables were entered simultaneously. Logistic regression analysis was performed between different subgroups of cases or between subgroups of cases and all controls. Crude and adjusted odds ratios (OR) and their 95% confidence intervals (CI) were calculated. Statistical significance was set at a P value of less than .05 (2-tailed).
Purchases of SSRIs were much more common in the first trimester than later during pregnancy (Table 1). There were 229 women with SSRI purchases in each trimester.
The mean age of the women in both cohorts at delivery was 30.0 years (standard deviation ± 0.7) (Table 2). There were more than twice as many tobacco smokers in the cohort of women with SSRI purchases than in the cohort of control women (P < .001) (Table 2). Similarly, pregnancy induced by artificial reproductive techniques was 6 times more common in women with SSRI purchases (P < .001), and termination due to fetal malformation was 5 times more common in the cohort with SSRI purchases than in controls; however, this difference was not statistically significant (P = .7) (Table 2). Mean length of gestation was shorter in SSRI-exposed women (P < .001), and similarly, mean birth weight was slightly lower (P < .001). The rate of perinatal death, caesarean delivery, and total number of malformations did not differ significantly from the control group (Table 2).
A total of 1,398 women purchased at least 1 SSRI drug 1 month before the beginning of pregnancy or during the first trimester (Table 1). Malformations were not more common in this cohort when compared with all control women (P = .4). Also, when cases with first trimester SSRI purchase(s) were compared with their matched controls (n = 1,398), the malformation rate did not differ between the 2 groups (P = 1.0). The occurrence of major malformations was not independently associated with any of the confounding variables considered in the logistic model. There were too few pregnancies induced by ART for meaningful comparison in the logistic regression analysis, but in the unadjusted model ART was not significantly associated with major malformations (P = .5).
In the first trimester, citalopram was purchased most frequently (554 women) followed by fluoxetine (525 women, Table 1). Major malformations were more common in women exposed to fluoxetine in the first trimester when compared with all control women with no drug purchases (P = .03), but after adjusting for confounders, the association did not reach statistical significance (OR 1.7, 95% CI 0.9–3.3) (Table 3). Women with fluoxetine exposure during the first trimester had 29 (5.5%) infants or fetuses with major malformations, compared with 62 of 1,782 (3.5%) in the control group. There were 12 cases of isolated cardiovascular anomalies (2.3%), 2 cardiovascular anomalies in 4 infants with chromosomal abnormalities (all Trisomy 21), and 4 cases with urinary tract malformations. After excluding chromosomal abnormalities, there were 25 (4.8%) malformations in the fluoxetine-exposed pregnancies and 52 (2.9%) in the control cohort.
There were 229 women who purchased 1 or more SSRI drugs during each trimester. To study possible effects of exposure during the period of fetal growth, we added to this cohort those women with SSRI purchases only in the second and third trimesters (total number of women 360). We first compared these cases (continuous exposure, n = 360) to all controls (n = 1,782). Small for gestational age was significantly more common in the continuous exposure group (P = .001), whereas low birth weight was not (P = .3). After adjusting for confounders, the risk of SGA remained statistically significant (OR 2.4, 95% CI 1.1–5.3). We then compared these women (n = 360) to cases in which the woman had purchased SSRIs only during the first trimester or the preconception period (n = 1,010). In this analysis, mean birth weight and length of gestation did not differ significantly between the 2 groups (Table 4). Low birth weight was not more common (P = .8) and neither was preterm birth (< 37 gestational weeks) (P = .2), and there were no cases of births at or earlier than 32 gestational weeks in the continuous-exposure cohort. Small for gestational age was not more common in the continuous-exposure cohort (P = .06), and the risk did not reach statistical significance after adjusting for confounders (OR 1.9, 95% CI 1.0–3.8) (Table 4). Of the confounders, SGA was independently correlated only to smoking (OR 3.2, 95% CI 1.6–6.4) and nulliparity (OR 2.4; 95% CI 1.2–4.8).
A total of 597 women purchased SSRIs during the third trimester. We compared these women with those with exposure only during the first trimester (n = 1,000). If the woman had purchases in both the first trimester and the third trimester, the case was defined as third trimester exposure. There was no difference between the groups in the mode of delivery (vaginal or cesarean delivery) (P = .6) (Table 5). Treatment in a special or intensive care unit was more common in the offspring of women exposed in the third than in the first trimester (15.7% compared with 11.2%) (P = .009), and this difference remained statistically significant after adjusting for confounding variables (OR 1.6, 95% CI 1.1–2.2). Treatment in a special or intensive care unit correlated independently also with smoking (OR 1.7; 95% CI 1.2–2.3). Low Apgar scores (< 7) were more common after third trimester exposure (P = .05), but after adjusting for confounders, the association did not reach statistical significance (OR 1.6, 95% CI 1.0–2.4). There were 64 women who had purchased paroxetine during the third trimester. Treatment in a special or intensive care unit was less common in the neonates exposed to paroxetine (4.7%) than in those exposed to other SSRIs (9.5%) (P = .02).
We found no increased risk of major malformations after exposure to SSRIs in early pregnancy. Treatment in neonatal a special or intensive care unit was more common after third trimester exposure; however, our results do not confirm previous findings of drug-related risk of preterm birth or low birth weight.8,14
The strength of our study is the large number of pregnant women and the access to their pregnancy outcome data enabled by population-based registers with a high quality of data. Validation studies have shown that the coverage of the Medical Birth Register data and the National Register of Induced Abortions data are close to 100%,20,21,24 and the data quality of most variables has been reported to be good in both registers with agreement with medical records of 95% or more.21,22,24 All neonates are examined by a pediatrician before discharge from the hospital, and diagnoses of possible illness or malformations are recorded in the register. By including data from the Register of Induced Abortions, we could add pregnancy terminations due to fetal malformation to the analysis. There are no validation studies concerning the accuracy of the Malformation Register, but the coverage of the register is considered good.23 In our study, data derived from the Malformation Register provided additional cases of malformations that were detected after hospital discharge and were therefore not reported in the Birth Register. It is therefore likely that practically all major malformations were recorded in our study.
The Medical Birth Register data include maternal background data such as tobacco smoking that could be taken into account in the analysis. The Drug Reimbursement Register covers claims data on 97% of all reimbursed drugs.25 The register data also made it possible to control for other purchased prescription medicines. Timing of exposure could be defined by a period corresponding to each trimester because prescribed drugs for continuing treatments are delivered from pharmacies for a maximum period of 3 months at a time.
A study limitation is that pregnant women may not be compliant with the prescribed drug even after purchasing it, and consequently, exposure and timing of exposure cannot be confirmed. Another weakness is that we did not calculate drug doses, and we could not control for alcohol or street drug use. Neither did we have information on nonprescription drug use. The registers we used did not include information on later neurodevelopment of the children, an issue of high importance to be studied in the future.27,28 Furthermore, minor malformations and problems may have remained undetected.
Some of the observed associations may be random when several analyses are run, therefore a certain reservation when considering the statistical significance of P = .05 is justified. However, we decided not to adjust the outcome variables for multiple comparisons in the statistical analyses because we feared that in doing this we might possibly lose some clinically relevant associations.
We found no increased risk of major malformations and exposure to SSRIs in the first trimester. Several controlled studies on first trimester SSRI exposure have been published without suggestion of increased risk of major malformations.8,12,13 However, the numbers in cohorts studied have been relatively small, and the samples may not have been representative of the whole population.8,13 A large study from Sweden, based on data derived from the national birth register and prospectively collected exposure data, did not uncover a risk of major malformations, but pregnancy terminations due to fetal malformations were not included in that study.12 Our data covered 1,398 pregnancies in which the woman had purchased 1 or more SSRI drugs during the preconception period or during the first trimester. Our study had a 85% power to detect a 2-fold increase in major malformations.
In a subgroup analysis of the different SSRIs, major malformations were more common in infants born to women with fluoxetine exposure in the first trimester (P = .03), but no statistically significant association was found after adjusting for confounders (OR 1.7, 95% CI 0.9–3.3). Previous studies have not suggested increased risk of major malformations after first trimester exposure to fluoxetine.7,8,11,12 The fluoxetine pregnancy registry maintained by the manufacturer reported nearly 800 first-trimester exposures with no suggestion of increased risk.29 However, postmarketing surveillance data are based on spontaneous reporting and has serious limitations due to the high percentage of cases lost to follow-up, making definitive conclusions difficult. In our study, there were 12 isolated cardiac malformations in infants of the fluoxetine-exposed women, representing a prevalence of 23 per 1,000 newborns. This is nearly 3-fold more than the prevalence of cardiovascular malformations in Finland (8 per 1,000 newborns). There were 8 isolated cases of ventricular septal defects. Ventricular septal defects are recorded as major malformations, but they often close spontaneously. Quite recently, however, 2 preliminary reports have suggested an association between first-trimester paroxetine exposure and cardiovascular anomalies (Diav-Citrin O, Shechtman S, Weinbaum D, Arnon J, Di Gianantonio E, Clementi M, Ornoy A. Paroxetine and fluoxetine in pregnancy: a multi-center, prospective, controlled study [abstract]. Reprod Toxicol 2005;20:459),30 raising further concerns.
Pregnancies induced by artificial reproduction techniques were more common in women with SSRI purchases. We are not aware of studies that would evaluate the prevalence of antidepressant use in women trying to conceive by ART; however, repeated treatments may render these women more prone to anxiety and depression.31
Our findings support previous reports of increased risk of neonatal problems.8,14 Several case reports and case series have suggested an association between SSRI exposure in late pregnancy and neonatal respiratory difficulties and central nervous system excitation.15–19 Cohort studies based on small samples have also showed increased risk for needing treatment in special care nurseries or low Apgar score.8,18,32,33 A large register-based study reported a 2-fold increase in low Apgar score and respiratory distress, but in that study the timing of SSRI exposure could not be defined in nearly 40% of the cases.14 We found that 15.7% of the infants of mothers exposed in the third trimester were treated in neonatal special or intensive care units compared with 11.2% of infants exposed only in the first trimester, and the association was significant after adjusting for confounders (adjusted OR 1.6). We examined the diagnoses for neonatal special or intensive care unit stay (International Classification of Diseases 10 in the Birth Register), but they were nonspecific, and analyzing them further would not result in definite conclusions about the origin of neonatal problems. Furthermore, we had no data on duration of stay in these units. However, a likely explanation for our results could be SSRI-induced withdrawal symptoms or drug toxicity in the newborn.
We found no increased risk of preterm birth (before 37 weeks) or birth before or at 32 weeks gestation. Two studies have suggested an increased risk for preterm birth in women using SSRIs during pregnancy.8,14 A prospective cohort study based on cases collected by Teratology Information Services found a relative risk of 4.8 for preterm birth after fluoxetine exposure.8 In a Swedish study based on registry data and prospectively collected interviews, a significantly increased risk was found for preterm birth after adjustment for confounders (OR 2.1).14 A controlled study based on medical records reported birth at or less than 36 gestational weeks being significantly more common in the exposed cohort (OR 4.4).32
We found no increased risk of low birth weight, contrary to a previous study which found a 2-fold increase in the occurrence of low birth weight in pregnancies exposed to SSRIs during the second or third trimesters.14 In our study, SGA was significantly more common in infants born to mothers with continuous exposure to SSRIs than in infants of mothers with no SSRI purchases (adjusted OR 2.4). However, no difference was found when comparison was made with women with only first trimester exposure. Our results suggest—but do not confirm—a drug-specific risk of SGA. One previous study based on a small cohort has suggested an increased risk of SGA in full-term infants after exposure to fluoxetine in late pregnancy,8 but in a large register-based study such a risk was not observed.14
SSRIs seem to be rather safe when used during pregnancy. However, because of the common occurrence of neonatal problems requiring observation or treatment and the relative lack of data concerning later neurodevelopment these drugs should be used during pregnancy only when clearly indicated.
1. Bennett HA, Einarson A, Taddio A, Koren G, Einarson TR. Prevalence of depression during pregnancy: systematic review. Obstet Gynecol 2004;103:698–709.
2. Egen-Lappe V, Hasford J. Drug prescription in pregnancy: analysis of a large statutory sickness fund population. Eur J Clin Pharmacol 2004;60:659–66.
3. Malm H, Martikainen J, Klaukka T, Neuvonen PJ; Finnish Register-Based Study. Prescription drugs during pregnancy and lactation—a Finnish register-based study. Eur J Clin Pharmacol 2003;59:127–33.
4. Andrade SE, Gurwitz JH, Davis RL, Chan KA, Finkelstein JA, Fortman K, et al. Prescription drug use in pregnancy. Am J Obstet Gynecol 2004;191:398–407.
5. Andersson L, Sundstrom-Poromaa I, Wulff M, Astrom M, Bixo M. Implications of antenatal depression and anxiety for obstetric outcome. Obstet Gynecol 2004;104:467–76.
6. Bonari L, Pinto N, Ahn E, Einarson A, Steiner M, Koren G. Perinatal risks of untreated depression during pregnancy. Can J Psychiatry 2004;49:726–35.
7. Addis A, Koren G. Safety of fluoxetine during the first trimester of pregnancy: a meta-analytical review of epidemiological studies. Psychol Med 2000;30:89–94.
8. Chambers CD, Johnson KA, Dick LM, Felix RJ, Jones KL. Birth outcomes in pregnant women taking fluoxetine. N Engl J Med 1996;335:1010–5.
9. Hendrick V, Smith LM, Suri R, Hwang S, Haynes D, Altshuler L. Birth outcomes after prenatal exposure to antidepressant medication. Am J Obstet Gynecol 2003;188:812–5.
10. Suri R, Altshuler L, Hendrick V, Rasgon N, Lee E, Mintz J. The impact of depression and fluoxetine treatment on obstetrical outcome. Arch Women Ment Health 2004;7:193–200.
11. Nulman I, Koren G. The safety of fluoxetine during pregnancy and lactation. Teratology 1996;53:304–8.
12. Ericson A, Kallen B, Wiholm B. Delivery outcome after the use of antidepressants in early pregnancy. Eur J Clin Pharmacol 1999;55:503–8.
13. Kulin NA, Pastuszak A, Sage SR, Schick-Boschetto B, Spivey G, Feldkamp M, et al. Pregnancy outcome following maternal use of the new selective serotonin reuptake inhibitors: a prospective controlled multicenter study. JAMA 1998;279:609–10.
14. Kallen B. Neonate characteristics after maternal use of antidepressants in late pregnancy. Arch Pediatr Adolesc Med 2004;158:312–6.
15. Nordeng H, Lindemann R, Perminov KV, Reikvam A. Neonatal withdrawal syndrome after in utero exposure to selective serotonin reuptake inhibitors. Acta Paediatr 2001;90:288–91.
16. Mohan CG, Moore JJ. Fluoxetine toxicity in a preterm infant. J Perinatol 2000;20:445–6.
17. Jaiswal S, Coombs RC, Isbister GK. Paroxetine withdrawal in a neonate with historical and laboratory confirmation. Eur J Pediatr 2003;162:723–4.
18. Costei AM, Kozer E, Ho T, Ito S, Koren G. Perinatal outcome following third trimester exposure to paroxetine. Arch Pediatr Adolesc Med 2002;156:1129–32.
19. Sanz EJ, De-las-Cuevas C, Kiuru A, Bate A, Edwards R. Selective serotonin reuptake inhibitors in pregnant women and neonatal withdrawal syndrome: a database analysis. Lancet 2005;365:482–7.
20. Gissler M, Berg C, Bouvier-Colle MH, Buekens P. Methods for identifying pregnancy-associated deaths: population-based data from Finland 1987-2000. Paediatric and Perinat Epidemiol 2004;18:448–55.
21. Gissler M, Teperi J, Hemminki E, Merilainen J. Data quality after restructuring a nationwide medical registry. Scand J Soc Med 1995;23:75–80.
22. Teperi J. A multi method approach to the assessment of data quality in the Finnish Medical Birth Registry. J Epidemiol Community Health 1993;47:242–7.
23. National Research and Development Centre for Welfare and Health (STAKES). Statistical summary. Available at: http://www.stakes.info/2/1/2,1,5.asp
. Retrieved March, 21, 2005.
24. Gissler M, Ulander V-M, Hemminki E, Rasimus A. Declining induced abortion rate in Finland: data quality of the abortion register. Int J Epidemiol 1996;25:376–80.
25. National Agency for Medicines and Social Insurance Institution. Finnish statistics on medicines. Helsinki (Finland): Edita Prima OY; 2004.
26. Pihkala J, Hakala T, Voutilainen P, Raivio K. Characteristic of recent fetal growth curves in Finland [in Finnish]. Duodecim 1989;105:1540–6.
27. Gentile S. SSRIs in pregnancy and lactation: emphasis on neurodevelopmental outcome. CNS Drugs 2005;19:623–33.
28. Oberlander TF, Grunau RE, Fitzgerald C, Papsdorf M, Rurak D, Riggs W. Pain reactivity in 2-month-old infants after prenatal and postnatal serotonin reuptake inhibitor medication exposure. Pediatrics 2005;115:411–25.
29. Goldstein DJ, Corbin LA, Sundell KL. Effects of first- trimester fluoxetine exposure on the newborn. Obstet Gynecol 1997;89:713–8.
30. GlaxoSmithKline: Clinical trial register. Available at: http://ctr.gsk.co.uk/medicinelist.asp
. Retrieved October 17, 2005.
31. Eugster A, Vingerhoets AJ. Psychological aspects of in vitro fertilization: a review. Soc Sci Med 1999;48:575–89.
32. Simon GE, Cunningham ML, Davis RL. Outcomes of perinatal antidepressant exposure. Am J Psychiatry 2002;159:2055–61.
33. Casper RC, Fleisher BE, Lee-Ancajas JC, Gilles A, Gaylor E, DeBattista A, et al. Follow-up of children of depressed mothers exposed or not exposed to antidepressant drugs during pregnancy. J Pediatr 2003;142:402–8.