Article Text
PDF
Abstract
Objective To investigate whether in utero exposure to antidepressants is associated with increased risk of attention deficit hyperactivity disorder (ADHD).
Design Cohort study.
Setting Denmark.
Participants All Danish singletons born alive from 1996 to 2009 were included. Using national medical registries, we defined in utero exposure to antidepressants as redemption of an antidepressant prescription by the mother 30 days prior to or during pregnancy. We defined maternal former users of antidepressants as women, who had redeemed a prescription up to 30 days prior to pregnancy, and never users as women who had never redeemed a prescription.
Main outcome measures ADHD was defined as redemption of a prescription for ADHD medication or an ADHD hospital diagnosis. Children were followed through 2010, and we used proportional-hazards regression to compute adjusted HRs comparing children exposed in utero and children born to former antidepressant users with children born to never users. To adjust for confounding from family-related factors, we conducted a within-mother between-pregnancy analysis comparing exposed children with unexposed siblings using conditional logistic regression.
Results We identified a cohort of 877 778 children, of whom 1.7% were exposed in utero. The overall median follow-up time was 8 years; selective serotonin reuptake inhibitors were the most commonly used class of antidepressant during pregnancy (78% of users). The adjusted HR comparing children exposed to any antidepressant in utero with children born to never users was 1.2 (95% CI 1.1 to 1.4), and 1.6 (95% CI 1.5 to 1.8) comparing children born to former users to children born to never users of antidepressants. In the within-mother between-pregnancy analysis (n=867), the adjusted OR was 0.7 (95% CI 0.4 to 1.4).
Conclusions This study provides no evidence to support a causal association between in utero exposure to antidepressants and risk of ADHD.
This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 3.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/3.0/
Statistics from Altmetric.com
Article summary
Strengths and limitations of this study
-
The study was based on a large study population with long and virtually complete follow-up. Our use of data from population-based databases in a setting of universal healthcare practically eliminates the risk of recall and selection biases.
-
We were limited by our lack of data on actual antidepressant intake by the mother and on the actual timing of intake; we used prescription redemption as a proxy for this information, which may have biased our results towards the null hypothesis.
-
Attention deficit hyperactivity disorder is a clinical diagnosis based on subjective criteria, and diagnoses therefore may vary among practitioners. Misclassification of the outcome could have led us to overestimate the association if children of mothers with a psychiatric diagnosis were more likely to obtain psychiatric treatment.
Introduction
Up to 13% of pregnant women experience depression,1 but only approximately 2% are treated with antidepressants.2 Untreated depression is associated with maternal tobacco and alcohol use, poor dietary intake and risk of poor birth outcomes, including preterm birth.3–6 However, use of selective serotonin reuptake inhibitors (SSRIs) during pregnancy has also been associated with adverse birth outcomes7 ,8 as well as teratogenic effects.9 ,10
Early-life exposure to SSRIs or tricyclic antidepressive agents may cause behavioural changes, as observed in rodents.11 ,12 The serotonin transporter, which is blocked by SSRIs, is expressed transiently in many brain areas during fetal life; serotonin plays a key role in neural development and maturation.13 Attention deficit hyperactivity disorder (ADHD) is a common neurodevelopmental disorder characterised by impulsiveness, inattention and hyperactivity.14 The incidence of ADHD is increasing, with a current prevalence of approximately 5% in children.15 Proposed risk factors include genetics, preterm birth and prenatal exposure to smoking or maternal stress.16–20 However, evidence about risk of ADHD following in utero exposure to antidepressants is limited to one study, which found no evidence for an association between in utero exposure to SSRIs and ADHD (OR: 0.91, 95% CI 0.51 to 1.60).21
Evidence of an association between antidepressant use during pregnancy and later ADHD in the offspring would have major public health implications. In this study, we investigated whether in utero exposure to antidepressants was associated with an increased risk of ADHD using nationwide Danish medical registries with virtually complete long-term follow-up.
Methods
Setting
The Danish healthcare system (5.5 million inhabitants and a yearly birth rate of approximately 65 000) provides tax-supported health services to all residents guaranteeing access to primary and secondary care free of charge. Except for emergencies, initial contact with the healthcare system is through general practitioners, who either treat patients themselves or refer them to hospitals or to private practice specialists.
Study population and design
Using the Danish Medical Birth Registry, we identified a cohort of all singletons born alive from 1996 until the end of 2009. The Danish Medical Birth Registry contains computerised records of all deliveries in Denmark since 1973. Each record includes the civil registration number of the mother, father and newborn, as well as multiple variables regarding the delivery, the newborn and the mother. Data are collected by midwives or physicians overseeing delivery.22 The civil registration number is a 10-digit number assigned to each Danish citizen at birth and to residents on immigration, enabling accurate and unambiguous linkage of relevant registries at the individual level.23 Siblings born to the same mother were identified through the civil registration system.23 Thus, in addition to a general population comparison cohort, it was possible to identify a sibling comparison cohort that was highly suitable for optimising adjustment for important potential family-related and genetic confounders.
Maternal antidepressant use
In utero exposure to antidepressants was defined as maternal redemption of a prescription for an antidepressant 30 days prior to or during pregnancy, as identified through the Danish National Prescription Registry.24 Since January 1994, the registry has recorded the following information whenever a prescription is redeemed in Denmark: the civil registration number of the patient, the medication classification code (the anatomical therapeutic chemical classification system of the WHO) and the date of dispensing. All antidepressants, as well as drugs for ADHD, are available by prescription only in Denmark. Pregnancy was defined as starting from the first day in the last menstrual period, according to the Danish Medical Birth Registry. First-trimester exposure was defined as redemption of a prescription by the mother 30 days prior to the beginning of pregnancy and up to 12 weeks after the beginning of pregnancy; second-trimester exposure was defined as prescription redemption between 12 weeks and 28 weeks of pregnancy and third-trimester exposure was defined as prescription redemption during the remainder of the pregnancy.25 If the mother redeemed more than one prescription during pregnancy, the first redemption was used to determine the trimester of exposure. We defined maternal former users of antidepressants as women, who had redeemed a prescription up to 30 days prior to pregnancy, and never users as women who had never redeemed a prescription.
Attention deficit hyperactivity disorder
ADHD was detected either as a diagnosis of ADHD or redemption of a prescription for ADHD medication. Using the Danish Psychiatric Registry26 and the Danish National Registry of Patients,27 we identified children in the study population with inpatient and outpatient hospital diagnoses of ADHD. The Danish Psychiatric Registry contains computerised data on all admissions to psychiatric hospitals and psychiatric wards in Denmark. The Danish National Registry of Patients has tracked all inpatient stays in Danish hospitals since 1977, and outpatient clinic and emergency room visits at all public hospitals since 1995. Data recorded in the Danish National Registry of Patients and the Danish Psychiatric Registry include civil registration number of the patient, dates of admission and discharge and up to 20 discharge diagnoses from each admission, classified according to the 8th revision of the International Classification of Diseases until 1993, and the 10th revision thereafter.
Diagnosing and treating ADHD is also handled by private practice psychiatrists and general practitioners in cooperation, without hospital contact. These patients are not recorded in the Danish Psychiatric Registry or the Danish National Registry of Patients. Therefore, to ensure completeness, we defined ADHD as either a diagnosis of ADHD or a redemption of a prescription for ADHD medication. Information on prescription redemption was obtained from the Danish National Prescription Registry.24
Covariates
Risk factors for ADHD that were potentially associated with maternal use of antidepressants were identified. We obtained information on maternal and paternal psychiatric diagnoses anytime before baseline from the Danish Psychiatric Registry. Information was also gathered on maternal epilepsy, infections and use of anxiolytics/hypnotics/sedatives during pregnancy from the Danish National Registry of Patients and the Danish National Prescription Registry. From the Danish Medical Birth Registry we obtained information on maternal age at birth, gender of the child, the child's birth order, maternal smoking during pregnancy, maternal body mass index and marital status. As body mass index was only available starting in 2004 and marital status was incompletely registered, body mass index and marital status were only included as potential confounders in separate subanalyses. We also obtained information on the child's gestational age, birth weight and 5-min Apgar score from the Danish Medical Birth Registry. These variables were investigated because they may play a role in a causal pathway linking in utero antidepressant exposure and ADHD. Since antidepressant use and ADHD prevalence increased between 1996 and 2010, we also adjusted for calendar time.
Statistical analyses
The children were followed from date of birth until the date of redemption of an ADHD prescription, receipt of an ADHD diagnosis, emigration, death or the end of follow-up on 31 December 2010, whichever came first.
General population comparison cohort
We compared children exposed to antidepressants in utero and also children of maternal former users with the unexposed children of never users (women who never redeemed a prescription for antidepressants). We performed separate analyses according to type of antidepressant (SSRIs, serotonin-norepinephrine reuptake inhibitors, tricyclic antidepressive agents, other antidepressants and combinations of antidepressants). For these subgroups, we conducted analyses comparing exposed children to unexposed children of never users. With Cox proportional-hazards regression, we computed crude and adjusted HRs (aHR) with 95% CIs as measures of relative risk. Patients with missing data were excluded from the analyses. The assumption of proportional hazards was graphically verified.
Sibling comparison cohort
To optimise control for family-related factors, such as genetics and socioeconomic or medical status, we conducted a within-mother between-pregnancy cohort analysis by restricting the study population to children of mothers who had more than one child and at least one exposed and one unexposed pregnancy. We then compared exposed children to unexposed children using a conditional logistic regression model. We computed adjusted ORs (aOR) with 95% CIs for receiving an ADHD diagnosis or redeeming a prescription for ADHD medication. In the first model, we adjusted for calendar time at birth to account for differences in length of follow-up. In this way we were able to compare the sibling analysis with the time-to-event analysis used in the general population comparison cohort.28 A stratified Cox proportional-hazards regression showed the same results as the conditional logistic regression model. The second model was fully adjusted. All statistical analyses were performed using SAS (V.9.2; SAS Institute Inc, Cary, North Carolina, USA). This study was approved by the Danish Data Protection Agency (Record no. 2013-41-1790). Codes used to define study variables are provided in the online supplementary appendices.
Results
We identified 877 778 singletons, of whom 15 008 (1.7%) were exposed to antidepressants in utero (Otables 1 and 2). The most commonly used class of antidepressant by the mothers was SSRIs (78%, table 3). The overall median follow-up time was 8 years. A total of 12 841 (1.5%) participants who developed ADHD (8100 (0.9%) were detected as a diagnosis of ADHD, and 4741 (0.5%) were identified through redemption of a prescription for ADHD medication as a first-outcome measure). In the within-mother between-pregnancy analysis we identified 867 children (330 groups of siblings) of whom 348 (40%) were exposed to antidepressants in utero. Of the 348 exposed children, 79 (23%) developed ADHD. Of the 519 unexposed children, 270 (52%) developed ADHD. Of 330 groups of siblings, 311 groups contained one child with ADHD and 19 pairs contained two children with ADHD.
Maternal and paternal characteristics of 877 778 singleton births in Denmark in 1996–2009, according to maternal use of antidepressants during pregnancy
Characteristics of 877 778 singleton births in Denmark in 1996–2009, according to maternal use of antidepressants during pregnancy
Distribution of 15 008 maternal antidepressants users according to classes of antidepressants
Maternal and paternal characteristics
Age at delivery was higher among antidepressant users than never users. Maternal antidepressant users more frequently smoked during pregnancy, were unmarried, had psychiatric diagnoses other than depression, epilepsy or infections during pregnancy. They also used anxiolytics/hypnotics/sedatives during pregnancy more often than never users. Fathers of children born to maternal users of antidepressants were also more likely to have a psychiatric diagnosis than fathers of children born to never users (table 1).
Birth outcomes
Gender distribution was the same for exposed and unexposed children. Children exposed in utero to antidepressants had a higher prevalence of low birthweight (3.7% vs 2.1%), low Apgar score (seven or under) at 5 min (2.6% vs 1.2%) and were born prematurely more often (15.4% vs 8.6%) than unexposed children born by never users (table 2).
Risk estimates
The aHR comparing children exposed to any antidepressant in utero with children born to never users was 1.2 (95% CI 1.1 to 1.4; table 4). The aHR was 1.6 (95% CI 1.5 to 1.8) when comparing children born to former users of antidepressants with children born to never users. Adjusting for the mother's body mass index and marital status in subanalyses did not change the estimates (results not shown). When stratifying according to trimester and class of antidepressants, we detected a minor association between ADHD and exposure in utero to any kind of antidepressant in the first and second trimesters, as well as exposure to SSRIs or the group ‘other antidepressants’ (table 4). In the within-mother between-pregnancy analysis, the full aOR was 0.7 (95% CI 0.4 to 1.4; table 5).
Crude and adjusted HRs and 95% CIs for time to redeem a prescription for attention deficit hyperactivity disorder (ADHD) medication or receiving the diagnosis of ADHD, comparing exposed children to unexposed children born to never users (women who had never redeemed a prescription on antidepressants)
Adjusted ORs and 95% CIs for a within-mother between-pregnancy analysis on a subpopulation of 867 children, restricted to mothers who had more than one child, with at least one exposed and at least one unexposed pregnancy
Discussion
While our analyses based on a general population comparison cohort showed a weak association between in utero exposure to antidepressants and ADHD, the within-mother between-pregnancy analyses were not consistent with a causal association. This observation strongly indicates the presence of unmeasured confounding by family-related factors in our comparisons with the general population. The former user analysis in the general comparison cohort also confirmed the presence of unmeasured confounding factors.
Our results extend the findings of an American study of claims-based data from 38 074 families that concluded, in line with our findings, that children exposed to SSRIs were not at increased risk of ADHD (OR: 0.91, 95% CI 0.51 to 1.60) while children exposed to bupropion were at increased risk (OR: 3.63, 95% CI 1.20 to 11.04), especially after second-trimester exposure.21 The study was limited by maternal smoking during pregnancy as a potential confounder, especially in the results for exposure to bupropion as this antidepressant is used for smoking cessation. Bupropion (which in Denmark is used for smoking cessation only) was contained in the group ‘other’ antidepressants in our study. The number of children exposed to this drug in utero was only 135 and due to sparse data we did not perform separate regression analyses.
The strengths of our study include its large study population with long and virtually complete follow-up. Our use of data from population-based databases in a setting of universal healthcare practically eliminates the risk of recall and selection biases. Importantly, the within-mother between-pregnancy analysis allowed us to effectively adjust for family-related factors, such as genetics and socioeconomic and medical status.
We were limited by our lack of data on actual antidepressant intake by the mother and on the actual timing of intake. We used prescription redemption as a proxy for this information which could lead to misclassification of exposure status. Regarding children classified as exposed in utero, we cannot be sure that women who redeemed a prescription 30 days before or during pregnancy actually used the drugs during pregnancy. Also, drugs redeemed earlier in life could have been stored for later use. Therefore, women categorised as former users might have been using antidepressants during pregnancy. Furthermore, the Danish National Prescription Registry24 has only recorded information since 1994. Among women who were categorised as never users some may have used the drugs before 1994. Such misclassification of exposure may have biased our result towards the null. Also, ADHD is a clinical diagnosis based on subjective criteria, and diagnoses may vary among practitioners. Misclassification could have led us to overestimate the association if children of mothers with a psychiatric diagnosis were more likely to obtain psychiatric treatment. At the same time, children with ADHD were identified based on hospital diagnoses and drug prescriptions. Thus, patients with ADHD diagnosed by private psychiatrists or general practitioners and not prescribed drug treatment would be misclassified as not having ADHD. This could lead to an underestimation of the association. The within-mother between-pregnancy analysis also may be limited by misclassification of exposure status and outcome. Such misclassification may have led us to underestimate an association in this analysis.29 However, the strong association between former maternal antidepressant use and ADHD in offspring, compared with a lack of exposure to antidepressants in utero, further supports the importance of causative factors other than antidepressant use. Comorbid ADHD in mothers using antidepressants would potentially conceal a causal association between in utero exposure to antidepressants and ADHD, given the strong ADHD risk conveyed by parental ADHD. However, given the reported prevalence of ADHD among adults with major depression, ranging from 6% to 9%30 we do not believe parental ADHD can explain the null association.
In conclusion, this large-scale study with complete long-term follow-up for ADHD provides no evidence to support an association between in utero exposure to antidepressants and risk of ADHD.
Additional material is published online only. To view please visit the journal online (http://dx.doi.org/10.1136/bmjopen-2013-003507)
Acknowledgments
We would like to thank all participants in this study.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Footnotes
-
Contributors KL, MSO, ABTA and HTS made primary contributions to writing the article. All authors contributed to the study conception and study design. TF performed data collection and statistical analyses and commented on the article. All authors contributed to the interpretation of results and revising the article critically and all approved the final article. HTS is the guarantor for this study.
-
Funding This study was supported by grants from ‘Max og Anna Friedmanns Legat til Sygdomsbekæmpelse’, ‘Familien Hede Nielsens Fond’ and from the Department of Clinical Epidemiology's Research Foundation. Research was conducted independently of the funders. Department of Clinical Epidemiology, Aarhus University Hospital, receives funding for other studies from companies in the form of research grants to (and administered by) Aarhus University. None of these studies have any relation to the current study.
-
Competing interests None.
-
Ethics approval Register studies do not require ethical approval in Denmark. This study was approved by the Danish Data Protection Agency (Record no. 2013-41-1790).
-
Provenance and peer review Not commissioned; externally peer reviewed.
-
Data sharing statement No additional data are available.