Association between antidepressant drug use during pregnancy and child healthcare utilisation

Abstract

Objective  To evaluate healthcare utilisation by children who were exposed to antidepressant drug use during pregnancy and those whose mothers stopped using antidepressants before pregnancy compared with a control group.

Design  Cohort study.

Setting  Health insurance records in the Netherlands.

Population  A total of 38 602 children born between 2000 and 2005.

Methods  Survey of child healthcare utilisation in relation to gestational antidepressant use.

Main outcome measure  Healthcare utilisation rates during the first year of life, with special emphasis to medical care related to cardiac disease.

Results  Children of mothers who used antidepressants during pregnancy showed increased healthcare use during the first year of life, independent of the mother’s healthcare use. The relative risk of more than two visits to general practitioners was 1.5 (95% confidence interval, CI: 1.3–1.8) in the continuous antidepressant users group and 1.3 (95% CI: 1.2–1.5) in the group of children whose mothers stopped taking medication. In both study groups there was a trend towards more drug use for infections and inflammation compared with the control group. Children continuously exposed to antidepressants had an increased risk of cardiac interventions such as cardiovascular surgery or heart catheterisation, relative risk of 5.6 (95% CI: 1.8–17.4). The risk of physiotherapy was twice as high in the antidepressant group compared with the control group (relative risk 2.0; 95% CI: 1.5–2.6).

Conclusion  Antidepressant use during pregnancy is associated with increased child healthcare utilisation and increased risk of major cardiac interventions in early childhood.

Introduction

Antidepressants are frequently prescribed to women of childbearing potential.^1–4 The reasons for use include depression and a variety of other psychiatric and physical illnesses such as anxiety disorders, sleeping disorders and neuropathic pain.⁵

For long there has been a reluctance to prescribe antidepressants during pregnancy because of restrictions mentioned in official labelling, handbooks, the prevailing precautionary principle and lack of human data. The initial results of studies on the safety of antidepressant use during pregnancy were reassuring.⁶ However, recent large database studies found an association between antidepressant use during pregnancy and problems in newborn children, including major congenital heart defects and primary pulmonary hypertension ^7–15, although other studies could not confirm these findings.^6,9,16 Another potential complication is the poor neonatal adaptation, including irritability, tremor, feeding problems and respiratory distress associated with antidepressant use during the last trimester, and thought to result from withdrawal or caused by referring to withdrawal, which is usually mild, but occasionally requires treatment.¹⁷ On the other hand, descriptions of the harmful effects of untreated stress and depression on birth outcome and child development contribute to the dilemma regarding whether or not to continue medication.^18,19 Studies on healthcare use have shown that depression and anxiety are associated with higher healthcare use, even after correcting for physical illnesses.^2,20 Increased primary care visits, emergency department visits and hospitalisations have been reported among children of mothers with mental distress.^21–23 Possible causes that have been suggested are increased health-seeking behaviour among mothers with psychiatric illness, an impaired ability to cope with their children’s diseases, non-compliance with therapy, and over-diagnosis fed by anxiety. The influence of antidepressants on healthcare utilisation by the offspring has not been studied. Researching the effects of medication is difficult, not only because of the multiple factors that influence pregnancy but also the birth outcome. Currently, the use of antidepressants in pregnant women is no longer considered absolutely contraindicated. In the UK, the NICE clinical guideline 45 on antenatal and postnatal mental health has been issued, but at the time of our study this guideline had not been implemented in the Netherlands nor had the warnings been issued against the use of paroxetine, which is the most used antidepressant in the Netherlands.²⁴ According to this guideline an individual risk-benefit balance should be made that takes into account the potential risks of untreated disease and the adverse effects of medication use.²⁵ So far, safety studies on antidepressants have focused primarily on structural birth defects that can be detected soon after birth, but little is known about functional teratology that leads to impaired organ function or to changes in physiological systems such as the immune system. Serotonin reuptake inhibitors like modern antidepressants may have lasting effects on the fetus, considering the important coordinating role of serotonin in the early development of many tissues and processes in the body, especially the maturation of the heart.^26,27

The objective of our study was to evaluate whether, after being exposed to antidepressants in utero, children suffer from more general health problems (reflected in healthcare utilisation rates) than children of mothers who did not use antidepressants during pregnancy. In an attempt to separate the influence of the disease from the effects of the drugs, we also studied cases where mothers stopped using antidepressants during pregnancy.²⁸

Methods

Setting

The study was carried out using insurance information from VGZ Health Insurance Company in the Netherlands on children born during the 5-year period between January 2000 and December 2004 and the mothers of these children. VGZ is one of the largest insurance companies in the Netherlands, with more than 2 million policyholders living predominantly in the centre and south of the country. Their client population is representative for the whole Dutch population with a slight overrepresentation of people who depend on social health care.¹ VGZ extracted anonymised data on healthcare services provided (including prenatal care, hospital admission, other medical facilities and drugs dispensed) from files in their database. In the Netherlands, the majority of healthcare use (including prescription drugs) is covered by compulsory healthcare insurance.

Study population

We selected all of the children from the database who were born during the study period; 964 twins, 19 triplets, 51.1% boys. Only single births were included. The corresponding mothers were identified using the child’s date of birth, the recorded date of obstetric help at delivery, and their home address. The mean maternal age at delivery was 30.3 (±4.7) years which is consistent with the mean age at delivery of 30.9 years in the overall Dutch birth registry data during the observed period.

If there was more than one birth (n = 8459), only the last child was included, leaving 38 989 children. Finally, we excluded those children whose own insurance records or those of the mother did not cover the entire observation window of 6 months before pregnancy, the pregnancy period and 1 year after birth (n = 387). This resulted in a study population of 38 602 children.

We divided the study population into five mutually exclusive cohorts according to patterns of antidepressant use based upon pharmacy dispensing data. Conception was estimated at 266 days before delivery. We examined antidepressant use during four different time windows: 6 months before conception and during each of the 90-day trimesters of pregnancy. Antidepressant use during one or more of these time windows was defined as the pharmacy dispensing at least one prescription for an antidepressant drug (ATC code N06A).^1–3

Women who used antidepressants before pregnancy as well as during their entire pregnancy were defined as CONTINUOUS USERS. Women for whom the pharmacy did not dispense an antidepressant during the 6 months before pregnancy but who did use antidepressants during pregnancy were defined as STARTERS. Women who used antidepressants before pregnancy but did not do so during pregnancy were defined as STOPPERS. The reference group (NON-USERS) consisted of women who did not use antidepressants before or during pregnancy. Finally, women who showed all other possible patterns of antidepressant use during pregnancy were defined as IRREGULAR USERS.

Healthcare utilisation

The extent of child healthcare utilisation was determined for the first 2 weeks of life as well as for 12 months after that. The first weeks after birth may differ from the first year, considering medical observations or interventions applied that are related to the possible occurrence of withdrawal effects after antidepressant use during the last trimester of pregnancy. The use of several types of healthcare services was evaluated. Visits to a general practitioner (GP) or specialist, having to have a medical intervention performed by a specialist, diagnostic procedures (e.g. laboratory tests), physiotherapist treatmen, or a hospital admission was determined by counting the number of subjects whose insurance records mention these items. To investigate special cardiovascular health care, the number of children who received cardiovascular diagnostic tests or cardiac interventions during the first year of life—such as cardiovascular surgery or interventional heart catheterisation, indicating clinically relevant congenital heart defects—were determined. Child medication use was determined to be the number of children in each cohort who received one or more drugs from specified categories during the first year.

In addition, healthcare utilisation by the mothers was determined during the 6 months prior to estimated conception and during pregnancy. Maternal medication use was determined by counting the number of women who received no, one to two and more than two prescriptions for different drugs other than antidepressants.

Data analysis

Database management was performed using Microsoft Access 2000. Descriptive statistics were used to evaluate characteristics of the different cohorts. Relative risks and the 95% confidence intervals (CIs) of the prevalences of specific healthcare utilisation items within the cohorts compared to the cohort with no antidepressant use were calculated. No correction for multiple testing was applied. An uncorrected chi-square test was used for comparing distributions of the categorical variables. For age means were calculated. Two-sided P-values <0.05 were considered to indicate statistical significance of differences between cohorts. The strength of the associations between child healthcare utilisation and maternal characteristics were expressed as relative risks with 95% CI. Statistical analyses were performed using SPSS software for Windows (version 15, SPSS Inc., Chicago, IL, USA).

Results

Characteristics of the study population

Antidepressant use during pregnancy was recorded for 784 mothers (2.2% of the population). Paroxetine was used most frequently (n = 305, 39%), followed by fluoxetine (n = 110, 14%). In total the use of selective serotonin reuptake inhibitors (SSRIs) was 71%, (n = 557). In 14% (n = 109), a tricyclic antidepressant (TCA) (mainly amitriptyline) was used and in 15% (n = 118) another antidepressant. Table 1 shows the characteristics of the mothers according to their antidepressant use before conception and during pregnancy. Mean age at delivery was 30.3 years (range: 15–50). Of the 784 mothers who used an antidepressant drug sometime during pregnancy, 197 women (0.5%) were identified as CONTINUOUS USERS. A small group of antidepressant-naïve women started to use these drugs during pregnancy (STARTERS, n = 44). Further analysis was not performed on this group because of its small size and the possibility that antidepressants were prescribed for use only after delivery. A large group of women showed irregular patterns of antidepressant use, defined as occasionally receiving a prescription (for instance, prescriptions before pregnancy, none in the first trimester, one in the second trimester followed by no antidepressant prescriptions in the third trimester) (IRREGULAR USERS, n = 543). Antidepressant use was stopped before pregnancy in 820 women (2.1%, STOPPERS). Overall, women in the CONTINUOUS USERS, IRREGULAR USERS as well as the STOPPERS group showed higher health care use before conception than those in the NON-USER group [relative risks varying between 1.2 and 2.3 (Table 1)]. Women in the CONTINUOUS USERS cohort were older than those in the other groups (P = 0.002) and showed higher rates of visits to medical specialists, prenatal tests and other diagnostic tests during pregnancy. The STOPPERS also used less medication of other kinds during pregnancy and showed less pre-delivery hospital admissions compared to the CONTINUOUS USERS, but more than the NON-USERS (P = 0.04).

Table 1.   Maternal characteristics categorised according to pre-conceptual and gestational antidepressant use (n = 38 602)

	NON USERS (n = 36 998) n, %	CONTINUOUS USERS (n = 197) n, RR (95% CI)	STARTERS (n = 44) n, RR (95% CI)	IRREGULAR USERS (n = 543) n, RR (95% CI)	STOPPERS (n = 820) n, RR (95% CI)	P-value**
Presented is the number of women who met the criteria, for example who had more than two GP visits during the first 6 months before gestation. RR, relative risk compared to NON USERS; NE, not estimable. *Other than precriptions for antidepressants. **P values calculated for the differences between study cohorts, NON USERS excluded. ***Till 3 days before labour.
Healthcare utilisation during the 6 months before the first day of gestation
GP visits ≥2	1310, 4	16, 2.3 (1.4–3.7)	1, 0.6 (0.1–4.5)	37, 2.0 (1.4–2.6)	61, 2.1 (1.6–2.7)	0.38
Specialist visits ≥2	3034, 8	32, 2.0 (1.4–2.7)	2, 0.6 (0.1–2.2)	70, 1.6 (1.3–2.0)	105, 1.6 (1.3–1.9)	0.44
Diagnostic tests ≥2	11 643, 31	87, 1.4 (1.2–1.6)	17, 1.2 (0.8–1.8)	252, 1.5 (1.4–1.6)	374, 1.5 (1.3–1.6)	0.09
Specialist procedures ≥2	8232, 22	64, 1.5 (1.2–1.8)	7, 0.7 (0.4–1.4)	167, 1.4 (1.2–1.6)	242, 1.3 (1.2–1.5)	0.41
Hospital admissions ≥1	3645, 10	20, 1.0 (0.7–1.6)	0, NE	48, 1.0 (0.7–1.2)	73, 1.0 (0.7–1.1)	0.19
Drug prescriptions ≥3*	11 163, 30	120, 2.0 (1.8–2.3)	16, 1.2 (0.8–1.8)	334, 2.0 (1.9–2.2)	494, 2.0 (1.9–2.1)	< 0.001
Physiotherapy	4290, 12	47, 2.1 (1.6–2.6)	13, 2.6 (1.6–4.0)	120, 2.0 (1.6–2.2)	185, 2.0 (1.7–2.2)	0.70
Healthcare utilisation during pregnancy
GP visits ≥2	3672, 10	43, 2.2 (1.7–2.9)	10, 2.3 (1.3–4.0)	113, 2.1 (1.8–2.5)	163, 2.0 (1.7–2.3)	0.51
Specialist visits ≥2	17 381, 47	142, 1.5 (1.4–1.7)	25, 1.2 (0.9–1.6)	307, 1.2 (1.1–1.3)	468, 1.2 (1.1–1.3)	0.006
Prenatal diagnostic tests ≥1	3588, 10	44, 2.3 (1.8–3.0)	4, 1.0 (0.4–2.4)	75, 1.4 (1.2–1.8)	104, 1.3 (1.1–1.6)	0.004
Other diagnostic tests ≥12	11 725, 32	110, 1.8 (1.6–2.0)	24, 1.7 (1.3–2.3)	263, 1.5 (1.4–1.7)	366, 1.4 (1.3–1.5)	0.002
Specialist procedures ≥2	11 524, 31	117, 1.9 (1.7–2.1)	18, 1.3 (0.9–1.9)	235, 1.4 (1.3–1.5)	343, 1.3 (1.2–1.5)	0.23
Hospital admissions ≥1***	15 841, 6	112, 2.4(1.6–3.4)	26, 2.4 (1.2–5.1)	273, 1.6 (1.2–2.1)	381, 1.7 (1.4–2.1)	0.04
Drug prescriptions ≥3*	14 474, 39	135, 1.8 (1.6–1.9)	25, 1.5 (1.1–1.9)	369, 1.7 (1.6–1.8)	468, 1.5 (1.4–1.6)	<0.001
Physiotherapy	7690, 21	62, 1.5 (1.2–1.9)	19, 2.1 (1.5–2.9)	177, 1.6 (1.4–1.8)	283, 1.7 (1.5–1.8)	0.43
Mean age at delivery (SD)	30.3 (4.7)	32.3 (4.7)	29.6 (5.4)	30.7 (5.1)	30.1 (5.0)	0.002

Child healthcare utilisation during the first 2 weeks of life

Figure 1 shows the risk estimates we found for each cohort in relation to healthcare utilisation of the child during the first 2 weeks of life compared to the NON-USERS group. The majority of children exposed to antidepressants during pregnancy stayed in the hospital after birth, because maternal psychotropic drug use is considered an indication to monitor the newborn for 1 or 2 days after birth. The CONTINUOUS USERS group showed a relative risk of having more than two visits to a medical specialist of 2.4 (95% CI: 1.7–3.3). Also, the risk of having one or more than one second-line medical intervention was increased significantly, with a relative risk of 1.5 (95% CI: 1.2–1.8) and 1.7 (95% CI: 1.1–2.6) respectively. The neonates from the STOPPERS group did not differ from the NON-USERS neonates during the first 2 weeks of life.

Figure 1.  Child healthcare utilisation during the first 2 weeks after birth. Presented here are the relative risks of health care service rates of children whose mothers used antidepressants continuously (A: CONTINUOUS USERS) before as well as during pregnancy (n = 197), used antidepressants irregularly during pregnancy (B: IRREGULAR USERS, n = 543), or stopped using antidepressants during pregnancy (C: STOPPERS, n = 820), compared to the non-user control group (n = 36 998). The symbol’s size represents the absolute risk in the study group: ◆ <10%, 10–25%, >25%.

Child healthcare utilisation during the first year of life

Healthcare utilisation after the neonatal period by children of CONTINUOUS USERS was higher compared to the NON-USERS group (Figure 2). The relative risk of more than two GP visits in that year is 1.5 (95% CI: 1.3–1.8), 1.2 (95% CI: 91.1–1.4) and 1.3 (95% CI: 1.2–1.5) respectively. The relative risk of specialist visits is the highest in the CONTINUOUS USERS group (1.5, 95% CI: 1.2–1.9). Also, the number of children admitted to hospital tends to be higher compared with the control group. Another feature associated with antidepressant use is physiotherapy. The observed risk of 18% is twice (CI 95%: 1.5–2.6) as high among the CONTINUOUS USERS compared with the control population. This risk is also increased among the IRREGULAR USERS but to a lesser extent, with a relative risk of 1.3 (95% CI: 1.1–1.7). It is interesting to note that we observed an increased likelihood of receiving more than three drug prescriptions in the IRREGULAR USERS or STOPPERS group.

Figure 2.  Child healthcare utilisation during the first year of life. Presented here are the relative risks of health care service rates during the first year of life starting at two weeks after birth for children whose mothers used antidepressants continuously (A: CONTINUOUS USERS) before as well as during pregnancy (n = 197), used antidepressants irregularly during pregnancy (B: IRREGULAR USERS, n = 543), or stopped using antidepressants during pregnancy (C: STOPPERS, n = 820) compared to the non-user control group (n = 36 998). The symbol’s size represents the absolute risk in the study group: ◆ <10%, 10–25%, >25%.

Exploring drug prescriptions in more detail (as shown in Figure 3), we found that the risk of prescriptions for laxatives increased in all three study groups (relative risk 1.6–2.3). Also, drugs for infections and inflammation increased in the IRREGULAR USERS group as well as in the STOPPERS and CONTINUOUS USERS groups, with the relative risk of receiving systemic antibiotics at 1.3 (95% CI: 1.2–1.5), 1.2 (95% CI: 1.1–1.3) and 1.2 (95% CI: 1.1–1.4) respectively. We found approximately the same results for drugs that act on the respiratory tract. The increased number of children receiving anticonvulsants in the CONTINUOUS USERS group were mainly given diazepam rectal solution, which is indicated for acute management of febrile convulsions.

Figure 3.  Prescription drugs, cardiovascular diagnostic tests, or cardiac interventions during the first year of life. Presented here are the relative risks of receiving a prescription drug for specific health issues during the first year of life for children whose mothers used antidepressants continuously (A:CONTINUOUS USERS) before as well as during pregnancy (n = 197), used antidepressants irregularly during pregnancy (B:IRREGULAR USERS, n = 543), or stopped using antidepressants during pregnancy (C:STOPPERS, n = 820), compared to the non-user control group (n = 36 998). The symbol’s size represents the absolute risk in the study group: ◆ <10%, 10–25%, >25%. *Specific antidepressant: CONTINUOUS USERS (n = 3): amitriptyline, fluvoxamine, and clomipramine; IRREGULAR USERS (n = 2): paroxetine and paroxetine in combination with temazepam.

When focusing on cardiac health services, we found that the risk of having diagnostic tests was no different than that of the control group. On the other hand, three of the 179 children in the CONTINUOUS USERS group had to undergo major cardiovascular interventions, resulting in a five times higher risk compared with the control group (relative risk 5.6, 95% CI: 1.9–16.3). The individual cases were related to exposure to different antidepressants. The relative risk of having to undergo a cardiac intervention was not significantly increased for children of women who used antidepressants irregularly (Figure 3).

Although we also investigated the relationship between the individual health care use of the mothers and the rates of healthcare use of their children, we found correlation coefficients to be <0.2 for GP and specialist visits as well as hospital admissions (data not shown). Adjustment for differences in maternal characteristics or healthcare utilisation of the mothers during pregnancy did not influence the relative risk estimates for the association between the utilisation patterns of antidepressants during pregnancy and the degree of healthcare utilisation of the child. For example: for physiotherapy of children of mothers in the continous users group the relative risk unadjusted was 1.99 (95% CI: 1.49–2.67) and adjusted to age and maternal healthcare use: 1.81 (1.26–2.60).

Discussion

Irrespective of the extent of the mother’s healthcare utilisation, infants of mothers who used antidepressants during pregnancy as well as infants of mothers who stopped using antidepressants before pregnancy showed more healthcare utilisation during the first year of life.

Our results confirm that women with depressive symptoms tend to show more healthcare-seeking behaviour.²⁹ Depressive patients have more co-morbidity, not in the least because physical diseases increase the risk of developing depression. Mothers whose medical records show antidepressant use were more likely to consult a doctor for their newborn child. An inability to cope with their children’s complaints and diseases may be one reason for this, but it may also be that their infants really do have an increased incidence of medical problems.

Intestinal and respiratory tract problems in babies are difficult to cope with and that might be the reason why we and others found more healthcare use in this area in the index groups.^21,22 On the other hand, it has been shown that children whose parents are depressed tend to have respiratory illnesses more often, independent of behaviour and smoking habits.³⁰ Moreover, children of women with higher levels of psychiatric symptoms show more febrile illnesses.³¹ This may explain the higher rate of antibiotic use in the groups exposed to antidepressants as well as in the STOPPERS group. Increased use of diazepam also points in the direction of a higher risk of febrile convulsions.

Although feeding problems during the first days after birth have been described in several studies on fetal exposure to antidepressants, persistent bowel complaints that could explain the elevated use of laxatives have not been reported before.^32,33 Serotonin plays a coordinating role in the developing intestine and stimulates gastrointestinal motility.³⁴ Withdrawal of serotonin reuptake inhibitors after birth may lead to a relative shortage of serotonin at receptors in the enteric nervous system, resulting in constipation. This theory does not explain the increased use of laxatives in the STOPPERS group and therefore constipation may also be associated with antidepressant exposure during early pregnancy or with the disease state.

By studying different patterns of maternal antidepressant use, we were able to extract some specific phenomena that were only seen in children who had been exposed to antidepressants: higher rates of physiotherapy and more cases with cardiac interventions. Higher rates of physiotherapy during the first year may be related to the observation that antidepressants during pregnancy have negative effects on a child’s motor movement.³⁵ An elevated risk of cardiac interventions like cardiac surgery or interventional heart catheterisation was found only in the CONTINUOUS USERS group, being the group that presumably had the highest level of exposure. Although the association between antidepressants and cardiac anomalies has been described before, this has not been found consistently.^7,9,14–16 Berard et al. showed a correlation between clinically relevant congenital heart defects and the prescribed dose.⁷ Bar-Oz et al. contributed these findings to the demanding behaviour of women on antidepressants leading to more diagnostic testing, therefore suggesting that some serious heart defects would remain undetected in the control group.⁸ Contrary to their hypothesis, we did not observe a higher rate of cardiovascular diagnostic testing. In fact, most of the congenital heart defects requiring intervention during the first year of life will be detected during standard child health screening programmes in the Netherlands.³⁶ The higher incidence of major congenital heart defects after continuous gestational antidepressant use justifies detailed fetal echocardiography screening, as is established for other pregnancies with risk factors for mayor congenital heart defects.^37–39 Although the absolute additive risk of birth defects may be acceptable for the individual antidepressant user, the question is whether the increased risk for the whole population is negligible.

Limitations

The disadvantage of studying healthcare use from the side of the provider is that there are no compliance data for the prescribed medication. Nor was it possible to obtain an overview of the extent of over-consumption of health care or under-treatment. No data was available on factors that may have influenced healthcare utilisation (such as co-morbidity and socio-economic data). Also, the insurance data did not include specific information on the disease state or indication. The contribution of women diagnosed with minor depression or mild symptoms may be larger among STOPPERS than in the CONTINUOUS USERS group. Moreover, gestational age at delivery was not known. We made the assumption that pregnancy started at approximately 38 weeks before birth, thereby ignoring preterm delivery, which may be associated with maternal depression or antidepressant use and which may have led to some misclassification of antidepressant use. Insurance data do not include information on maternal smoking habits, substance use or lactation. These factors may also influence the child’s health.

We did not differentiate between the classes of antidepressants although their differences in pharmacological characteristics may result in differences in effect on the child. Another limitation was the fact that records of visits to paediatricians were not detailed. Because of this, the only way to distinguish between real illnesses and mere complaints was by exploring drug prescriptions and performed procedures. Unfortunately, records of cardiovascular surgery and heart catheterisation included no details on indication.

Conclusions

Early in life, children of mothers with a history of antidepressant use during pregnancy show more healthcare utilisation than other children, also in cases where the mother tapered off or stopped using these drugs during pregnancy. Increased prescription rates of antibiotics and drugs for pulmonary and intestinal problems in these children indicate that symptoms could be related to the mother’s psychiatric illness, since it was also found in case women stopped using antidepressants during pregnancy. Another explanation might be that these symptoms are an effect of exposure tot antidepressants during early fetal development.

Children whose mothers used antidepressants throughout pregnancy have more frequently a heart disease or motor function problems requiring therapeutic interventions compared to non-exposed children. These effects may be drug-induced and therefore prenatal screening on congenital heart defects is recommended. Further research is needed to define the relationship between mental wellbeing, medication status and specific paediatric issues such as functional or structural cardiovascular defects, neuromuscular development, infections and pulmonary diseases. Medical demands of children prenatally exposed to antidepressants are not restricted to the first 2 weeks after birth, but extend into the first year of life.

Disclosure of interests

The authors’ work was independent of VGZ Health Insurance Company.

Contribution to authorship

TV, AE, AS and GV were the main contributors to the conception, design, analysis, interpretation of data. TV and KW were responsible for data management. TV and TE were resposible for statistical analyses and wrote the first draft of the manuscript. MF, MH, LJ and GV contributed to interpreting the data. All authors provided suggestions during the different phases of the preparation of the manuscript and approved the final version.

Details of ethics approval

Not needed. According to the regulations of the ethics committee of the University Medical Center Utrecht our study protocol did not require submission for approval because the study concerned a selected database with anonymised data.

Funding

None declared.

Acknowledgements

We are grateful to VGZ Health Insurance Company for providing the anonymised health care and medication records we used in our study.