Volume 87, Issue 3 p. 965-987
ORIGINAL ARTICLE
Free Access

Trajectories of antidepressant drugs during pregnancy: A cohort study from a community-based sample

Aurélie Cabaillot

Aurélie Cabaillot

Département de Médecine Générale, UFR de Médicine, Université Clermont Auvergne, Clermont-Ferrand, France

CHU Clermont-Ferrand, Inserm, Neuro-Dol, Service de Pharmacologie médicale, Centres Addictovigilance et Pharmacovigilance, Observatoire Français des Médicaments Antalgiques (OFMA), Institut Analgesia, Université Clermont Auvergne, Clermont-Ferrand, France

Search for more papers by this author
Alexandra Bourset

Alexandra Bourset

Département de Médecine Générale, UFR de Médicine, Université Clermont Auvergne, Clermont-Ferrand, France

Search for more papers by this author
Aurélien Mulliez

Aurélien Mulliez

Délégation à la recherche clinique et à l'innovation, CHU Clermont-Ferrand, Clermont-Ferrand, France

Search for more papers by this author
Jessica Delorme

Jessica Delorme

CHU Clermont-Ferrand, Inserm, Neuro-Dol, Service de Pharmacologie médicale, Centres Addictovigilance et Pharmacovigilance, Observatoire Français des Médicaments Antalgiques (OFMA), Institut Analgesia, Université Clermont Auvergne, Clermont-Ferrand, France

Search for more papers by this author
Massimiliano Orri

Massimiliano Orri

McGill Group for Suicide Studies, Douglas Mental Health University Institute, Department of Psychiatry, McGill University, Montréal, QC, Canada

Bordeaux Population Health Research Centre, Inserm U1219, Université de Bordeaux, Bordeaux, France

Search for more papers by this author
Mathilde Vicard-Olagne

Mathilde Vicard-Olagne

Département de Médecine Générale, UFR de Médicine, Université Clermont Auvergne, Clermont-Ferrand, France

Npsysydo, Université Clermont Auvergne, Clermont-Ferrand, France

Search for more papers by this author
Marie Christine Zenut

Marie Christine Zenut

CHU Clermont-Ferrand, Inserm, Neuro-Dol, Service de Pharmacologie médicale, Centres Addictovigilance et Pharmacovigilance, Observatoire Français des Médicaments Antalgiques (OFMA), Institut Analgesia, Université Clermont Auvergne, Clermont-Ferrand, France

Search for more papers by this author
Marie Tournier

Marie Tournier

Inserm, Bordeaux Population Health Research Center, Team Pharmacoepidemiology, Université de Bordeaux, Bordeaux, France

Hospital Charles Perrens, Bordeaux, France

Search for more papers by this author
Denis Gallot

Denis Gallot

CNRS 6293, INSERM 1103, GReD, QC G1V 0A6 Clermont-Ferrand; Department of Obstetrics and Gynecology, CHU Clermont-Ferrand, Faculty of Medicine, Université Clermont-Auvergne, Clermont-Ferrand, France

Search for more papers by this author
Nicolas Authier

Nicolas Authier

CHU Clermont-Ferrand, Inserm, Neuro-Dol, Service de Pharmacologie médicale, Centres Addictovigilance et Pharmacovigilance, Observatoire Français des Médicaments Antalgiques (OFMA), Institut Analgesia, Université Clermont Auvergne, Clermont-Ferrand, France

Search for more papers by this author
Chouki Chenaf

Chouki Chenaf

CHU Clermont-Ferrand, Inserm, Neuro-Dol, Service de Pharmacologie médicale, Centres Addictovigilance et Pharmacovigilance, Observatoire Français des Médicaments Antalgiques (OFMA), Institut Analgesia, Université Clermont Auvergne, Clermont-Ferrand, France

Search for more papers by this author
Catherine Laporte

Corresponding Author

Catherine Laporte

Département de Médecine Générale, UFR de Médicine, Université Clermont Auvergne, Clermont-Ferrand, France

Npsysydo, Université Clermont Auvergne, Clermont-Ferrand, France

Correspondence

Catherine Laporte, Département de Médecine Générale, UFR de Médicine, EA7280, 28 Place Henri Dunant, BP 38, 63001 Clermont-Ferrand Cedex 1, France.

Email: [email protected]

Search for more papers by this author
First published: 04 August 2020
Citations: 9

The authors confirm that the Principal Investigator for this paper is Catherine Laporte and that she had direct clinical responsibility for data.

Abstract

Aims

The aim of this study was to monitor the trajectories of antidepressant use during pregnancy and the postpartum period among women chronically treated with antidepressants before their pregnancy, and to assess characteristics associated with each trajectory.

Methods

This cohort study included all pregnant women whose data were included in the General Sample of Beneficiaries (EGB) database affiliated with the French Health Insurance System, from 2009 to 2014. Women were followed up until 6 months after childbirth. Chronic treatment was defined as exposure over the 6-month period preceding pregnancy. A group-based trajectory model (GBMT) was estimated to identify distinctive longitudinal profiles of antidepressant use.

Results

Among 760 women chronically treated with antidepressants before their pregnancy, 55.8% stopped their treatment permanently in the first trimester, 20.4% discontinued it for a minimum of 3 months and resumed it postpartum, and 23.8% maintained it throughout pregnancy and postpartum. No sociodemographic or medical characteristics were associated with any trajectory group. Women who maintained treatment presented more frequent obstetric complications and postpartum psychiatric disorders. Among women who interrupted treatment, prescription of benzodiazepines and anxiolytics decreased initially but rose postpartum to a higher level than before pregnancy.

Conclusions

Pregnant women treated with antidepressant require a re-evaluation of psychiatric treatment. It is necessary to pay attention to obstetric complications for severely depressed women. Additionally, as relapse was associated with increased benzodiazepine use, it is important to carefully monitor all women who stop antidepressant treatment during pregnancy.

What is already known about this subject

  • There is no absolute contraindication to taking antidepressants during pregnancy.
  • The majority of pregnant women treated with antidepressants stop treatment abruptly.
  • Seventy per cent of pregnant women treated with antidepressants who stop them abruptly may experience a relapse.

What this study adds

  • Half of the participating women stopped their antidepressant treatment without complications.
  • Women undergoing antidepressant treatment who stopped during pregnancy and resumed in the following weeks had a rebound effect of increased prescription of benzodiazepines and anxiolytics.

1 INTRODUCTION

In France, the use of psychotropic drugs was of particular concern in women of childbearing age, with a prevalence of 23% versus 13% in men of the same age, and 3.2% of women started pregnancy when exposed to antidepressants1 in 2010. The main indication for antidepressants is depression,2 and women of childbearing age are indeed twice as likely as men to develop depression.3 Pregnancy is also a risk period, with a prevalence of depression during pregnancy in France between 5 and 15%.4

Untreated maternal depression has deleterious consequences for the expectant mother and the fetus.5 In the mother, psychic relapse can lead to suicidal risk, puerperal psychoses and an unstable emotional connection between the mother and the child.4, 6 Although there are fewer deaths and suicide attempts during pregnancy and the postnatal period than in the general female population, deaths by suicide account for 20% of all deaths in the postnatal period.7-9 Additionally, severe depressive state may increase the risk of intrauterine growth retardation and prematurity in the fetus and delayed psychomotor development in children may be due to increased amounts of maternal cortisol released into the bloodstream.10, 11

As all antidepressants pass the placental barrier,12 about 30% of newborns exposed in utero suffer from neonatal withdrawal syndrome, which is usually benign, treatable and resolves under surveillance.13

Selective serotonin reuptake inhibitors (SSRIs) are the antidepressants most commonly prescribed during pregnancy. Despite contrasting evidence, they appear to be safe in pregnant women,14-18 so they are recommended by the National Agency for Health and Medicines (ANSM)19 and the 2018 International Cochrane Review International Publications.20 As there is no absolute contraindication to taking antidepressant medications during pregnancy, they should not be stopped abruptly, which would increase the risk of depression relapse with a rebound effect.21, 22

In France in 2016, 5% of women were exposed to at least one anxiolytic–hypnotic drug during pregnancy.23 Benzodiazepines are often prescribed for anxiety and depression,24 with substantial heterogeneity across countries, and across the different periods of pregnancy (e.g., increase from preconception to pregnancy, with a subsequent decrease to postpartum).25 When used during pregnancy, benzodiazepines and benzodiazepine-related drugs pass readily through the placenta, and are associated with adverse birth outcomes (spontaneous abortion, preterm birth, and when used in the third semester loppy infant syndrome or respiratory problems).25

Many studies have shown that the use of antidepressants decreases during the first 3 months of pregnancy.26-30 A recent population-based study from Denmark, observed that the prescription of antidepressants was less common during pregnancy than before and after it in the period 1997 to 2016.31 To our knowledge, only few previous studies have investigated trajectories of medication use during pregnancy. Hurault et al. were the first researchers to study psychotropic drugs during pregnancy with longitudinal methods23 and to describe use of anxiolytics.24 They encouraged longitudinal methods over this period which evolves over 9 months, and which is studied over 3-months periods. Hurault et al. compared the use of cross-sectional and longitudinal methods during pregnancy: transverse methods will be better suited to looking for a specific effect of a drug during a given time. Longitudinal studies are needed to compare women who discontinue antidepressant treatment with those who did not, with respect to relapse of depression during pregnancy and the postnatal period and to consequences for newborn infants.24 Bandoli et al. studied the prescription trajectories of antidepressants during pregnancy and during the postpartum period but without linking the two periods, and on a small sample (166 women).32

The prevalence of depressive disorders in women and their consequences for the mother and child during and after pregnancy, as well as the high prevalence of use of antidepressants in France, necessitates monitoring the trajectories of antidepressant treatment during and after pregnancy in women who were treated before their pregnancy. Our study is the first to study antidepressant use over the entire pre-intra-post pregnancy period, with a large sample of women chronically treated with antidepressants before their pregnancy. This allows us to (1) assess the prevalence of antidepressant discontinuation, (2) compare women's characteristics and consequences associated with antidepressant treatment maintenance vs discontinuation, and (3) to inform preventive interventions in this vulnerable population.

2 OBJECTIVES

The primary aim of the study was to monitor the time trends of antidepressant treatment during pregnancy and the postpartum 6-month period among women chronically treated with antidepressants before their pregnancy, using trajectory models that help in identifying treatment changes. The secondary objectives were: (i) to identify factors associated with stopping antidepressant treatment during pregnancy; (ii) to investigate complications associated with each trajectory of treatment use; and (iii) to assess the parallel evolution of co-treatment with benzodiazepines.

3 METHODS

3.1 Design and data sources

This cohort study was conducted using data from the General Sample of Beneficiaries database (EGB) during the period from January 1, 2009 to December 31, 2014.33 We used a group-based trajectory model (GBMT) to identify distinctive longitudinal profiles of reimbursement for antidepressant use.34

The EGB database contains a representative 1/97th random sample of the population (600,000 beneficiaries) and is affiliated with the French national health insurance system (approximately 98% of the French population).33 The EGB is updated quarterly and counts entries (births, new affiliations) and exits (deaths, regime changes). Since 2005, the EGB database has included anonymous information on the sociodemographic characteristics of the beneficiaries (age, sex, date of death, affiliation fund, and welfare benefit for low-income people) and some medical information. The sample contains detailed codes for biologicals, medical devices and medical technical procedures according to the Common Classification of Medical Acts (CCAM). It includes information about health professionals who met beneficiaries (specialty, nature of practice) and care facilities. The database is linked to the national hospitalization database of the Medical Information Systems Program (PMSI) with information on the diagnoses (main, connected, associated) according to the 10th version of the International Classification of Diseases (ICD-10), for each hospital stay of the beneficiary (except psychiatric hospitalization). Diagnoses are also available through the long-term condition status (ALD), which gives access to care free of charge for the given disease (collected according to their ICD-10). Finally, the EGB database includes all drug reimbursements, classified according to their Anatomical Therapeutic Chemical (ATC) code, with the exception of drugs retroceded by the hospital. The collected data are kept for 20 years beyond the current year. The EGB database has been widely used for public health and pharmaco-epidemiological purposes for over 5 years.35, 36

Time trends in the reimbursement for antidepressants were monitored for 6 months before conception date, during pregnancy and up to 6 months after childbirth, divided into 3-month periods. The pregnancy trimesters were calculated in 90-day increments, thus following the definition of the World Health Organization (WHO) by indicating gestational age in weeks of gestation (WG) (1–15 WG inclusive for 1st trimester – T1; 16–28 WG inclusive for 2nd trimester – T2; and 29–41 WG inclusive for 3rd trimester –_T3). We chose to analyse postpartum treatment for 6 months.37

3.2 Study sample

The study was conducted on all women chronically treated with antidepressants prior to pregnancy and who gave birth between 2009 and 2014 (see Appendix 1). We selected women who had: (i) a pregnancy ending with birth from January 1, 2009 to December 31, 2014; (ii) a refund for three boxes or more of antidepressant over the 6 months preceding the beginning of pregnancy; and (iii) available data from 6 months before pregnancy to 6 months after childbirth. The conception date was based on the gestational age at delivery (based on early ultrasound) subtracted from the date of birth. We excluded women for whom the term at delivery was unknown, women who had a pregnancy ending in an abortion or miscarriage before 22 weeks, and women without antidepressant treatment before pregnancy.

3.3 Measures and outcomes

3.3.1 Drug exposure

Drug exposure was identified through the reimbursement for antidepressants. Non-selective monoamine reuptake inhibitors (INSRMs), selective serotonin reuptake inhibitors (SSRI), non-selective monoamine inhibitors (IMONSs), monoamine oxidase inhibitors (MAOIs), and serotonin–norepinephrine reuptake inhibitors (SNRIs) were identified in the EGB database via their ATC code (Appendix 2). In France, antidepressants are only available on doctor's prescription and are therefore all reimbursed (the database therefore gives access to the entirety of antidepressants delivered to patients), and the majority of antidepressants are sold in boxes of 7, 14, 28 or 30 tablets (the prescription can be from 1/2 to several tablets per day). Considering at least one reimbursement over 3 months allows us to take into account the participants who had a prescription during this period and not to miss a patient under treatment. For each 3-month time unit, there was at least one antidepressant reimbursement for the patient in the EGB database (1) or there was none (0).

3.3.2 Collected data

Data were extracted from the EGB database using the CCAM, ICD-10 or ALD codes (Appendix 3). The collected data concerned sociodemographic information, pregnancy outcomes and obstetric or non-obstetric history of the patient: age at delivery (a three-category variable: <20 years, 20–35 years, or >35 years), welfare beneficiary, nulliparous or parous, history of complicated pregnancy, sterility or pathway of assisted reproduction, anorexia, diabetes, substance use, ALD status or psychiatric comorbidities, psychiatric disorders or suicide attempt.

Context of ongoing pregnancy: single or multiple pregnancy, substance use, or obesity.

Pregnancy tracking: number of obstetric ultrasounds (usually three per pregnancy) (Appendix 3), number of general practice or psychiatric consultations and number of hospitalizations during the pregnancy, length of maternity stay.

Complications of pregnancy and childbirth: excessive weight gain, preeclampsia/eclampsia, threatened preterm labour (PAD), gestational diabetes, haemorrhage during pregnancy and delivery, Non-reassuring fetal heart rate, Caesarean delivery but also ALD status or psychiatric comorbidities during a hospitalization, psychiatric disorders or suicide attempt during pregnancy (Appendix 3).

Psychiatrics events or complications before, during and after pregnancy: ALD status or psychiatric comorbidities during a hospitalization, psychiatric disorders or suicide attempt (Appendix 3).

Reimbursement for benzodiazepines before, during and after pregnancy, identified in the EGB database via their ATC code (Appendix 2). For each 3-month time unit, there was at least one benzodiazepine reimbursement for the patient in the EGB database (1) or there was none (0).

3.4 Statistical analysis

3.4.1 Trajectories of antidepressant reimbursement

To identify distinctive longitudinal patterns of reimbursement for antidepressant use, we used a semi-parametric mixture model group-based trajectory model (GBTM).34 GBTM is an empirical procedure that uses multinomial modelling with maximum likelihood to identify clusters of individuals who follow similar treatment trajectories. Models with one to five trajectories and different shapes (i.e., constant, linear, quadratic or cubic evolution over time) were estimated and compared using the Bayesian information criterion (BIC). We used Mplus to model the trajectories, so the best-fitting model is the one minimizing the BIC. Participants were assigned to the trajectory with the highest posterior probability. The average posterior probability was used as a criterion of the quality of the classification operated by the model (good if >0.70); because it is a binary variable (0 = no treatment, 1 = treatment), we used a logit model to answer the research questions.38

3.4.2 Analysis

The population was described using numbers and percentages for categorical data and means ± standard deviations (or median and interquartile range when data are not normal) for continuous data. Comparisons among the three groups created by the GBTM were carried out using the chi-squared test (or Fisher's exact test when appropriate), followed by the Marascuilo procedure39 when significance was reached (omnibus p < 0.05) for categorical data. For categorical outcomes such as low birth weight, preeclampsia and premature delivery, results are shown as odds ratios (ORs) with their 95% confidence intervals for each two-group comparisons. Following this, a multivariable logistic regression model was fitted using obstetric and psychiatric complications as dependent variable and GBTM groups and two demographic variables (age and CMU) as independent variables. Results are shown as adjusted ORs and their 95% confidence intervals (95% CIs). Comparisons of means between groups were carried out using analysis of variance (or by the Kruskal–Wallis test when data were not normal), followed by a post hoc Tukey–Kramer test (or Dunn's test when data were not normal). Normality was assessed graphically and using Shapiro–Wilk's test. All tests were two-sided and a p-value of <5% was considered significant. Statistics were computed using STATA v15 (Stata Corp., College Station, TX, USA) and Mplus.

3.5 Ethics approval and consent to participate

The EGB database guarantees the confidentiality of all data and anonymity (agreement of the National Commission for Computing and Freedoms (CNIL) on June 14, 2005).

4 RESULTS

4.1 Trajectories of antidepressant treatment

Among women who gave birth from 2009 to 2014 (n = 41,722), we identified 760 treated with antidepressants for at least 6 months prior to pregnancy (Appendix 4). SSRIs were the most widely reported class of antidepressants (62%, n = 471), followed by SNRIs (19.3%, n = 147) and INSRMs (6.3%, n = 48).

The GBTM showed that three models had a very close BIC: 3-, 4- and 5-trajectory model (see Appendix 5). Even if the BIC was slightly lower in the 4- and 5-trajectories models, we selected the 3-trajectory model as the most clinically pertinent (BIC = +4666.42). Specifically, the fourth and fifth trajectories were obtained by splitting one of the trajectories obtained in the 3-trajectory model, but clinically all these groups would not be treated differently. The three selected trajectories of antidepressant reimbursement (Figure 1) were: the “Stopped” trajectory, which included 424 women (55.8%) who stopped their treatment during the first trimester of their pregnancy without resuming over the study period; the “Interrupted” trajectory, which comprised 155 (20.4%) who discontinued treatment for at least 3 months during pregnancy but resumed it postpartum; and the “Maintained” trajectory, which included 181 (23.8%) who maintained their treatment throughout pregnancy. The average posterior probability was 0.87.

Details are in the caption following the image
Trajectories of evolution of antidepressant use during pregnancy

4.2 Factors associated with treatment change during pregnancy

Table 1 describes the characteristics of the study population and the comparison of the three groups. The maternal age at delivery was slightly higher in the Maintained group (33.3 ± 5.2 years) than in the Stopped group (31.8 ± 5.5 years; p = 0.02. The proportion of patients receiving welfare benefit was the same in the three trajectories, as was a history of pregnancy or complications during pregnancy. Fewer than 2% of the participants used medical assisted procreation (PMA) and it was not associated with the time trends of the treatment. Difference in history of anorexia or obesity, or of substance use (alcohol or tobacco), was not related to any trajectory. Drug use was more common in the Interrupted group than in the Stopped group (p < 0.001).

TABLE 1. Characteristics of the overall sample and of women following distinct trajectories of antidepressant use
Variable All sample (n = 760) Stopped (n = 424) Interrupted (n = 155) Maintained (n = 181) p-value
Sociodemographic characteristics
Age at delivery 32.3 (5.5) 31.8 (5.5) 32.5 (5.3) 33.3 (5.2) 0.02b
<20 6 (0.8) 5 (1.2) 1 (0.7) 0 (0.0) 0.2
20–35 522 (68.7) 302 (71.3) 103 (66.5) 117 (64.6) ()
>35 232 (30.5) 117 (27.6) 51 (32.9) 64 (35.4)
CMU 156 (20.5) 89 (21.0) 38 (24.5) 29 (16.0) 0.15
Multiparous 385 (50.7) 226 (53.3) 79 (51.0) 0.12
History of complicated pregnancy 13 (1.7) 6 (1.4) 2 (1.3) 5 (2.8) 0.48
Medical assisted procreation 14 (1.8) 8 (1.9) 3 (1.9) 3 (1.7) 1
Multiple pregnancy 12 (1.6) 5 (1.2) 5 (3.2) 2 (1.10) 0.23
History before pregnancy
Anorexia 14 (1.8) 6 (1.4) 4 (2.6) 4 (2.2) 0.54
Diabetes 11 (1.5) 8 (1.9) 1 (0.7) 2 (1.1) 0.65
Alcohol 25 (3.3) 9 (2.1) 9 (5.8) 7 (3.9) 0.08
Tobacco 61 (8.0) 28 (6.6) 14 (9.0) 19 (10.5) 0.24
Drugs 24 (3.2) 6 (1.4) 13 (8.4) 5 (2.8) <0.001a
Obesity 41 (5.4) 16 (3.8) 11 (7.1) 14 (7.7) 0.08
Current pregnancy
Pregnancy tracking
No. of obstetric consultations 3 [1;6] 3 [0.5;6] 4 [1;7] 3 [0;6] 0.64
No. of ultrasounds 2 [0;4] 2 [0;4] 2 [0;4] 2 [0;4] 0.51
No. of general practice cs (med) 5 [2;9] 5 [2;8] 6 [3;1] 5 [2;8] 0.05a,c
No. of psychiatric cs (med) 0 [0;0] 0 [0;0] 0 [0;0] 0 [0;2] 0.005a,b,c
Length of hospital stay for childbirth 5.5 (3.5) 5.7 (4.0) 5.0 (2.2) 5.7 (3.2) 0.03a,b,c
No. of hospitalizations during pregnancy 0 [0;1] 0 [0;0] 0 [0;1] 0 [0;1] 0.52
Complications of pregnancy
Excessive weight gain 30 (4.0) 17 (4.0) 6 (3.9) 7 (3.9) 1
Pregnancy-induced hypertension 23 (3.0) 12(2.8) 3 (1.9) 8 (4.4) 0.41
Preeclampsia 27 (3.6) 13 (3.1) 2 (1.3) 12 (6.6) 0.02c
Eclampsia 0 0 0 0
Gestational diabetes 83 (10.9) 44 (10.4) 14 (9.0) 25 (13.8) 0.32
Early haemorrhage during pregnancy 6 (0.8) 3 (0.7) 0 (0.0) 3 (1.7) 0.23
Haemorrhage before childbirth 5 (0.7) 1 (0.2) 2 (1.3) 2 (1.1) 0.15
Complications of childbirth
Gestational age childbirth 38.7 (2.3) 38.9 (2.0) 38.5 (3.0) 38.5 (2.1) 0.03a,c
Threat of premature delivery 40 (5.3) 20 (4.7) 12 (7.7) 8 (4.4) 0.3
Premature delivery 64 (8.4) 26 (6.1) 18 (11.6) 20 (11.1) 0.04
Premature rupture of membranes 66 (8.7) 32 (7.6) 16 (10.3) 18 (9.9) 0.45
Fetal distress 177 (23.3) 85 (20.1) 36 (23.2) 56 (30.9) 0.02a
Non-reassuring fetal heart rate 126 (16.6) 59 (13.9) 27 (17.4) 40 (22.1) 0.04
Fetal dystocia 58 (7.6) 33 (7.8) 14 (9.0) 11 (6.1) 0.59
Haemorrhage during childbirth 4 (0.5) 3 (0.7) 0 (0.0) 1 (0.6) 0.82
Caesarean delivery 196 (25.8) 103 (24.3) 42 (27.1) 51 (28.2) 0.56
Low birth weight 32 (4.2) 15 (3.5) 6 (3.9) 11 (6.1) 0.35
Post partum
Postpartum haemorrhage 47 (6.2) 22 (5.2) 6 (3.9) 19 (10.5) 0.02
  • Data are presented as frequencies (associated percentages).
  • Abbreviations: CMU, Universal health insurance; cs, Consultation.
  • a p < 0.05 between Traj1 and Traj2;
  • b p < 0.05 between Traj1 and Traj3;
  • c p < 0.05 between Traj2 and Traj3.

4.3 Obstetric factors associated with treatment trajectory

During pregnancy, univariate analysis showed fetal distress more frequently in the Maintained group (30.9%, n = 56) than in the Stopped group (20.1%; n = 85; p = 0.02) (Table 1). This result was confirmed by the multivariable analysis (OR, 95% CI: 1.77 [1.19–2.63]), which also showed a higher risk for non-reassuring fetal heart rate (OR, 95% CI: 1.74 [1.11–2.73]) without difference between the Maintained and Interrupted groups (Table 2). Premature delivery was less frequently observed in the Stopped than in the Maintained (OR, 95% CI: 1.84 [0.99–3.41]) and Interrupted (OR, 95% CI: 1.93 [1.02–3.64]) groups. Postpartum haemorrhage were more frequent in the Maintained than in the Interrupted (OR, 95% CI: 2.54 [1.31–4.91]) and Stopped (OR, 95% CI: 3.33 [1.28;8.66]) groups, as it was for preeclampsia (respectively OR, 95% CI: 5.51 [1.21–25.2] and to a lesser extent OR, 95% CI: 2.15 [0.95–4.85]).

TABLE 2. OR for obstetrical and psychiatrical complications
Low birth weight Preeclampsia Premature delivery Fetal distress Postpartum haemorrhage Non-reassuring fetal heart rate
Obstetric complications OR [CI 95%] p OR [CI 95%] p OR [CI 95%] p OR [CI 95%] p OR [CI 95%] p OR [CI 95%] p
Interrupted/Stopped 0.99 [0.37–2.62] 0.981 0.39 [0.09–1.75] 0.220 1.93 [1.02–3.64] 0.042 1.20 [0.77–1.87] 0.418 0.76 [0.30–1.94] 0.571 1.30 [0.79–2.14] 0.305
Maintained/Stopped 1.65 [0.73–3.72] 0.225 2.15 [0.95–4.85] 0.065 1.84 [0.99–3.41] 0.052 1.77 [1.19–2.63] 0.005 2.54 [1.31–4.91] 0.006 1.74 [1.11–2.73] 0.016
Maintained/Interrupted 1.67 [0.59–4.70] 0.330 5.51 [1.21–25.2] 0.028 0.95 [0.48–1.89] 0.893 1.47 [0.90–2.40] 0.123 3.33 [1.28–8.66] 0.014 1.34 [0.78–2.31] 0.293
Postpartum depression Puerperal psychiatric disorder
Post-partum psychiatric complications OR [CI 95%] p OR [CI 95%] p
Interrupted/Stopped 6.46 [1.64–25.4] 0.008 2.13 [1.07–4.25] 0.032
Maintained/Stopped 4.89 [1.20–20.0] 0.027 6.88 [3.87–12.2] <0.001
Maintained/Interrupted 0.76 [0.25–2.32] 0.627 3.23 [1.70–6.13] <0.001
  • OR [CI 95%] odds ratio, [95% confidence interval]. Adjusted on age and CMU.

4.4 Monitoring of pregnancy according to treatment trajectory

Women in the Interrupted group consulted their GP more frequently during pregnancy (8 ± 9.4 consultations vs 6.1 ± 5.1 for the Maintained group and 5.8 ± 4.4 for the Stopped group), and consulted a private psychiatrist more often than women in the Stopped group did (1.7 ± 5.0 vs 0.9 ± 3.4, p < 0.05). Women in the Maintained group consulted a private psychiatrist more often than others did (2.3 ± 6 consultations vs 1.7 ± 5.0 for the Interrupted group and 0.9 ± 3.4 for the Stopped group, p < 0.05) (Table 1).

4.5 Psychiatric factors associated with treatment trajectory

Compared to the Stopped group, women in the Maintained group were more likely to have psychiatric comorbidities during hospitalization or had long-term condition status (ALD) for depression-anxiety (30.9%, n = 56 vs 18.2%, n = 77, p = 0.002), and other psychiatric disorders (17.1%, n = 37 vs 6.6%, n = 28, p = 0.002) in the lifespan (Table 3). However, there was no difference in rates of psychiatric comorbidities during hospitalization or ALD according to psychiatric history (depression, anxiety, psychotic disorders or suicide attempt) in the 6 months preceding pregnancy. During pregnancy, women in the Maintained group, had more frequent psychiatric comorbidities during hospitalization or ALD for depression-anxiety than those in the Stopped group (15.5%, n = 28 vs 2.4%, n = 10, p < 0.001), and those in the Interrupted group (6.5%, n = 10, p < 0.001), and more often suffered from other psychiatric disorders than those in Stopped group (5.5%, n = 10 vs 0.9%, n = 4, p = 0.003). During the postpartum period, the rate of psychiatric comorbidities during hospitalization or ALD for psychiatric disorders related to the puerperium was higher in the Maintained group (24.3%, n = 44) than in the others (Interrupted: 10.3%, n = 16 and Stopped 5.0%, n = 21, p < 0.001 for all). This result was confirmed by the multivariable analysis (Maintained/Interrupted OR, 95% CI: 3.23 [1.70–6.13] and Maintained/Stopped OR, 95% CI: 6.88 [3.87–12.2]).

TABLE 3. Psychiatric events of the overall sample and of women following distinct trajectories of antidepressant use
Variable All sample Stopped (Traj 1) Interrupted (Traj 2) Maintained (Traj 3) p-value
Psychiatric comorbidities during hospitalization or long-term affection before pregnancy
Depression/anxiety 165 (21.7) 77 (18.2) 32 (20.7) 56 (30.9) 0.002b
Psychiatric disorder 80 (10.5) 28 (6.6) 21 (13.6) 37 (17.1) <0.001b
Suicide attempt 62 (8.2) 28 (6.6) 17 (11.0) 17 (9.4) 0.19
Psychiatric comorbidities during hospitalization or long-term affection 6 months before pregnancy
Depression/anxiety 39 (5.1) 23 (5.4) 7 (4.5) 9 (5.0) 0.9
Psychiatric disorder 7 (0.9) 3 (0.7) 1 (0.7) 3 (1.7) 0.58
Suicide attempt 21 (2.8) 12 (2.8) 5 (3.2) 4 (2.2) 0.87
Psychiatric comorbidities during hospitalization or long-term affection during pregnancy and childbirth
Depression/anxiety 48 (6.3) 10 (2.4) 10 (6.5) 28 (15.5) <0.001b,c
Psychiatric disorder 18 (2.4) 4 (0.9) 4 (2.6) 10 (5.5) 0.003b
Suicide attempt 2 (0.3) 1 (0.2) 0 (0.0) 1 (0.6) 0.69
Psychiatric comorbidities during hospitalization or long-term-affection post partum (6 month after childbirth)
Depression/anxiety 15 (2.0) 2 (0.5) 7 (4.5) 6 (3.3) 0.001
Psychiatric disorder 0 0 0 0 n.a.
Suicide attempt 5 (0.7) 1 (0.2) 3 (1.9) 1 (0.6) 0.05
Psychiatric comorbidities during hospitalization or long-term affection during childbirth and post partum
Psychiatric disorder related to the puerperium 81 (10.7) 21 (5.0) 16 (10.3) 44 (24.3) <0.001b,c
  • Data are presented as frequencies (associated percentages).
  • a p < 0.05 between Traj1 and Traj2;
  • b p < 0.05 between Traj1 and Traj3;
  • c p < 0.05 between Traj2 and Traj3.

4.6 Evolution of benzodiazepine treatments

Benzodiazepine reimbursement was not different in the three groups over the 6 months preceding pregnancy: Stopped = 42.0% (181), Interrupted = 40.7% (63) and Maintained = 46.4% (84), p = 0.9 (Figure 2). Over the 6 months following delivery, women had lower benzodiazepine prescriptions than before pregnancy in the Stopped group (8.5%, n = 36 vs 42%, n = 181, p < 0.001) and in the Maintained group (30.4%, n = 55 vs 46.6%, n = 84, p < 0.001). In the Interrupted group, the benzodiazepines dispensed decreased to T3 (13.6%) but then rose to a higher level than before pregnancy (54.2%, n = 84 vs 40.7%, n = 63, p = 0.005). The same time trend was observed for other anxiolytics.

Details are in the caption following the image
Evolution of benzodiazepine treatment during pregnancy by trajectory of antidepressant treatment

5 DISCUSSION

Among women who started pregnancy when exposed to antidepressant treatment, we identified three trajectories of antidepressant dispensing during pregnancy: more than half stopped their treatment; a quarter maintained their treatment throughout pregnancy; and one out of five discontinued it for a minimum of 3 months and then resumed it during the postpartum period. Before pregnancy, only lifetime history of psychiatric disorders was associated with maintenance versus stopping of treatment. There was no sociodemographic or medical characteristic associated with the trajectories of antidepressant treatment during pregnancy. There were more obstetric complications in the Maintained group than in the Stopped (premature delivery, non-reassuring fetal heart rate, postpartum haemorrhage and rhythm disorder) and Interrupted group (preeclampsia and postpartum haemorrhage) groups. The frequency of psychiatric comorbidities during hospitalization over the postpartum period was higher for the Maintained and Interrupted groups than for the Stopped group. The dispensing of benzodiazepines and other anxiolytics differed in the Interrupted group: it initially decreased with the decrease in antidepressants, then rose to a higher level in the postpartum period than before pregnancy.

The women who maintained their antidepressant treatment presented with the most severe depression profile: they more often had ALD status for depression, anxiety or other psychiatric disorders and more frequently consulted a psychiatrist and/or a GP than women in the Stopped group did. In terms of obstetric complications, there were more obstetric complications in the Maintained group than in the Stopped (premature delivery, non-reassuring fetal heart rate, postpartum haemorrhage and rhythm disorder) and Interrupted (preeclampsia and postpartum haemorrhage) groups. Consistent with this finding, a prior study showed that women exposed to antidepressant and/or anxiolytic medication before the 16th week of pregnancy have a 3-fold increased risk for preeclampsia.40 A meta-analysis showed that exposure to antidepressants in late gestation was associated with a significantly increased risk of postpartum haemorrhage.41 Therefore, when a woman is treated with antidepressants in late pregnancy, health professionals should be especially vigilant, and ask the woman about a possible interruption of treatment during pregnancy: if the antidepressant was maintained throughout the pregnancy, the woman is even more at risk of complications than if there was an interruption. Moreover, women who maintained their antidepressant treatment probably had the most severe depression, which in turn is associated with obstetric complications due to increased amounts of maternal cortisol released into the bloodstream during severe depressive states.10, 11

Half of the women stopped their treatment and did not return to their treatment; they did not relapse. It is therefore essential to re-evaluate antidepressant treatment in early pregnancy and to discontinue unnecessary treatments. In France, the prevalence of individuals taking antidepressant drugs is 6%. Their use is higher among women.42

The women in the Interrupted group showed intermediate levels of pre-pregnancy psychiatric events and had fewer depression or anxiety events during pregnancy and the postpartum period than women in the Maintained group. It is possible that women in the Interrupted group had less severe depression than those in the Maintained group, which would explain the interruption of their treatment, and their less frequent follow-up with a psychiatrist. However, they consulted a GP more often; this is a sign of mood disorders, which explains the resumption of their treatment. Finally, their prescription of benzodiazepines and anxiolytics, after a decreasing that paralleled that of the antidepressants, rose to a higher level postpartum than before pregnancy. A first hypothesis is that stopping the antidepressant may have resulted in a relapse of depression21 with resumption of treatment and the prescription of benzodiazepine. A second hypothesis is that the profile of these pregnant women was perhaps different before their pregnancy. Specifically, those women may have been more anxious than depressed, so that their management would have been different. An implication of this finding is that if antidepressants are stopped at the beginning of pregnancy, patients must be reassessed to identify those who will either relapse or will present with an increase in anxiety. These women would require a different management compared to those who stop without psychiatric complications. The drop in benzodiazepines associated with stopping antidepressants is also observed in a study on health insurance data in the Netherlands (from 33% in the 1st trimester to 2.8% in the 3rd trimester),43 but the originality of our study is the observation of this rebound during the postpartum period. Prescribed during pregnancy, treatment with benzodiazepine may be harmful if combined with treatment with antidepressants,44 and prescribed after pregnancy, it carries the known risks of dependence on treatment.

Our study did not identify any somatic, psychological or obstetric risk factors that predict appropriate cessation of the treatment. The evaluation of psychiatric events related to depression over the 6 months preceding pregnancy is therefore not indicative of the evolution of antidepressant use, as no differences among the three groups were identified. To reassess disease severity and therefore obtain an indication of whether to maintain or stop treatment, we should question the whole history of antidepressant treatment in the patient and not just that in the last 6 months. These findings also should encourage an enhancement of maternal follow-up during the postpartum period.

The strength of this study was the use of the French EGB database, which allows an effective supply of patients. The data were reimbursement data; therefore, there was no declarative bias. Trajectory modelling made it possible to follow the purchase of the treatment as closely as possible to the evolution of the pregnancy. Some limitations of this study should be considered in interpreting the results.. The EGB database provides information on the drugs prescribed and purchased by the patient. Therefore, there is the possibility that the were not actually taken. The lack of compliance, which can be 50% in high-income countries45 and especially relevant for psychotropic drugs, was not taken into account in this study. Therefore, there is a risk of having overestimated the actual treatment intake.

However, as the women went to the pharmacy to buy their treatment regularly, this bias seems very limited, and the Maintained group does not seem to be concerned. Additionally, to avoid an overestimation of treatment intake, we observed the 6 months preceding pregnancy as an inclusion criterion. Another limitation is that previous events of depression were identified in our study by measuring psychiatric comorbidities during hospitalization or ALD declarations for depression; however, a large number of cases of depression do not lead to hospitalization or ALD, and this proportion could therefore have been underestimated. Finally, despite the longitudinal design, caution is needed in interpreting our results as causal association; it is possible that several confounders were unmeasured.

In conclusion, an announcement of pregnancy in an antidepressant-treated woman requires re-evaluation of psychiatric treatment, not only assessing the events of the 6 months preceding the pregnancy, but more generally, her history of depression. Women requiring continued treatment, and therefore those with more severe depression, demonstrated a higher rate of perinatal complications. Half of the women treated before pregnancy stop their antidepressant use in early pregnancy and do not resume the treatment postpartum, without this having any consequences on the events of depression during pregnancy or the postpartum period. However, when antidepressants are stopped at the beginning of pregnancy, patients must be reassessed to identify those who will relapse their psychiatric disorder, and manage them differently to prevent an over-prescription of benzodiazepines.

COMPETING INTERESTS

There are no competing interests to declare. The study did not receive any funding.

CONTRIBUTORS

C.L., A.B., J.D., M.C.Z., C.C. and N.A. conceived the presented idea. C.C., J.D. and A.M. verified the analytical methods. A.M. and M.O. performed the analysis of the data. A.C., A.M., M.V.O., M.T., D.G., C.C., N.A. and C.L. contributed to the interpretation of the results. A.C. took the lead in writing the manuscript. A.C., M.V.O., M.T., D.G., N.A. and C.L. discussed the results and contributed to the final manuscript. C.L. supervised the findings of this work.

APPENDIX 1

Flow chart of participants

APPENDIX 2

Anatomical Therapeutic Chemical codes for study antidepressants and anxiolytics

ATC codes of antidepressants

Non-selective monoamine reuptake inhibitor NO6AA
Selective serotonin reuptake inhibitor NO6AB
Non-selective monoamine inhibitors NO6AF
Monoamine oxidase inhibitors N06AG
Others N06AX, NO6CA

ATC codes of benzodiazepines, anxiolytics and hynotics

Benzodiazepines and derivatives N05BA
Diphenylmethane N05BB
Azaspirodecandedione N05BE
Other anxiolytics N05BX
Benzodiazepines derivatives N05CD
Benzodiazepine-related drugs N05CF
Other hypnotics and sedatives N05CM
Associated hypnotics and sedatives, other than barbiturates N05CX

APPENDIX 3

CCAM, ICD-10 or ALD codes of Measures and outcomes

Eligibility criteria

Medical termination of pregnancy and late miscarriages

CODES INTITULE
O049 Avortement médical complet ou sans précision, sans complication = IMG
O039 Avortement spontané complet ou sans précision, sans complication (fausse couche tardive)
O036 Avortement spontané complet ou sans précision, compliqué d'une hémorragie retardée ou sévère (fausse couche hémorragique)

Obstetric or non-obstetrical history before pregnancy

Nulliparous

CODE INTITULE
JQGD013 Accouchement unique par le siège par voie naturelle avec grande extraction, chez Une primipare
JQGD010 Accouchement céphalique unique par voie naturelle, chez Une primipare
JQGD004 Accouchement unique par le siège par voie naturelle, chez Une primipare
JQGD003 Accouchement unique par le siège par voie naturelle avec petite extraction, chez Une primipare
JQGD002 Accouchement multiple par voie naturelle, chez Une primipare
Z355 Surveillance d'une primipare âgée
Z356 Surveillance d'une primipare très jeune

Parous

CODE INTITULE
JQGD012 Accouchement céphalique unique par voie naturelle, chez Une multipare
JQGD008 Accouchement unique par le siège par voie naturelle avec petite extraction, chez Une multipare
JQGD007 Accouchement multiple par voie naturelle, chez Une multipare
JQGD005 Accouchement unique par le siège par voie naturelle avec grande extraction, chez Une multipare
JQGD001 Accouchement unique par le siège par voie naturelle, chez Une multipare
Z354 Surveillance d'une grossesse avec multiparité élevée
Z641 Difficultés liées à Une multiparité

History of complicated pregnancies

CODE INTITULE
Z875 Antécédents personnels de complications de la grossesse, de l'accouchement et de la puerpéralité

Sterility or pathway of assisted reproduction

CODE INTITULE
N97 Stérilité de la femme
N974 Stérilité de la femme associée à des facteurs relevant de l'homme
N989 Complication de la fécondation artificielle, sans précision
Z31.1 Insémination artificielle
Z31.2 Fécondation in vitro
Z31.3 Autres méthodes de fécondation assistée
JHFB001 Prélèvement de spermatozoïdes au niveau du testicule, de l'épididyme ou du conduit déférent, par voie transcutanée
JJFC011 Prélèvement d'ovocytes Sur un ou deux ovaires, par coelioscopie
JJFJ001 Prélèvement d'ovocytes Sur un ou deux ovaires, par voie transvaginale avec guidage échographique
JKHD002 Prélèvement et examen de la glaire cervicale, sans examen de la mobilité des spermatozoïdes
JKHD003 Prélèvement et examen de la glaire cervicale, avec examen de la mobilité des spermatozoïdes [test postcoïtal de Huhner]
JSEC001 Transfert intratubaire d'embryon, par coelioscopie
JSED001 Transfert intra-utérin d'embryon, par voie vaginale
JSLD002 Insémination artificielle intracervicale

Long-term condition status for depression or anxiety

CODE INTITULE
F40 Troubles anxieux phobiques
F41 Autres troubles anxieux
F410 Trouble panique [anxiété épisodique paroxystique]
F411 Anxiété généralisée
F412 Trouble anxieux et dépressif mixte
F419 Trouble anxieux, sans précision
F32 Épisodes dépressifs
F320 Épisode dépressif léger
F321 Épisode dépressif moyen
F322 Épisode dépressif sévère sans symptômes psychotiques
F323 Épisode dépressif sévère avec symptômes psychotiques
F328 Autres épisodes dépressifs
F329 Épisode dépressif, sans précision
F33 Trouble dépressif récurrent
F920 Troubles des conduites avec dépression

Long-term condition status for psychiatric issues other than depression and anxiety

CODE INTITULE
F20 Schizophrénie
F21 Trouble schizotypique
F22 Troubles délirants persistants
F220 Trouble délirant
F23 Troubles psychotiques aigus et transitoires
F25 Troubles schizo-affectifs
F28 Autres troubles psychotiques non organiques
F29 Psychose non organique, sans précision
F31 Trouble affectif bipolaire
F310 Trouble affectif bipolaire, épisode actuel hypomaniaque
F312 Trouble affectif bipolaire, épisode actuel maniaque avec symptômes psychotiques
F319 Trouble affectif bipolaire, sans précision
F34 Troubles de l'humeur [affectifs] persistants
F348 Autres troubles de l'humeur [affectifs] persistants
F39 Trouble de l'humeur [affectif], sans précision
F42 Trouble obsessionnel-compulsif
F44 Troubles dissociatifs [de conversion)
F60 Troubles spécifiques de la personnalité
F608 Autres troubles spécifiques de la personnalité
F68 Autres troubles de la personnalité et du comportement chez l'adulte
F69 Trouble de la personnalité et du comportement chez l'adulte, sans précision
F98 Autres troubles du comportement et troubles émotionnels apparaissant habituellement Durant l'enfance et l'adolescence
F989 Trouble du comportement et trouble émotionnel apparaissant habituellement Durant l'enfance et l'adolescence, sans précision

Substance use

CODE INTITULE
Z720 Usage du tabac
F17 Troubles mentaux et du comportement liés à l'utilisation du tabac
Z71.6 Conseil pour tabagisme
Z721 Consommation d'alcool
F10 Troubles mentaux et du comportement liés à l'utilisation de l'alcool
Z714 Conseil et surveillance pour alcoolisme
Z50.2 Sevrage d'alcool
Y90 Preuves du rôle de l'alcool confirmé par le taux d'alcoolémie
Y91 Preuves du rôle de l'alcool confirmé par le degré d'intoxication
X65 Auto-intoxication par l'alcool et exposition à l'alcool
Y15 Intoxication par l'alcool et exposition à l'alcool, intention non déterminée
O354 Soins maternels pour lésions fœtales (présumées) dues à l'alcoolisme maternel
Z722 Utilisation de drogues
F12 Troubles mentaux et du comportement liés à l'utilisation de dérivés du cannabis
F13 Troubles mentaux et du comportement liés à l'utilisation de sédatifs ou d'hypnotiques
F14 Troubles mentaux et du comportement liés à l'utilisation de cocaïne
F15 Troubles mentaux et du comportement liés à l'utilisation d'autres stimulants, y compris la caféine
F16 Troubles mentaux et du comportement liés à l'utilisation d'hallucinogènes
F18 Troubles mentaux et du comportement liés à l'utilisation de solvants volatils
F19 Troubles mentaux et du comportement liés à l'utilisation de drogues multiples et troubles liés à l'utilisation d'autres substances psychoactives.
Z71.5 Conseil et surveillance pour toxicomanie et pharmaco dépendance

Pregnancy tracking

Obstetric ultrasound

CODE INTITULE
JQQM019 Échographie biométrique et morphologique d'une grossesse multifoetale au 2ème trimestre
JNQM001 Échographie non morphologique de la grossesse avant 11 semaines d'aménorrhée
JQQJ037 Mesure de la longueur du canal cervical du col de l'utérus, par échographie par voie vaginale
JQQM001 Échographie de surveillance de la croissance foetale
JQQM002 Échographie d'une grossesse unifoetale à partir du 2ème trimestre avec échographie-doppler des artères utérines de la mère et des vaisseaux du foetus, pour souffrance foetale
JQQM003 Échographie de surveillance de la croissance foetale avec échographie-doppler des artères utérines de la mère et des vaisseaux du foetus
JQQM007 Échographie d'une grossesse multifoetale à partir du 2ème trimestre avec échographie-doppler des artères utérines de la mère et des vaisseaux des foetus, pour souffrance foetale
JQQM008 Échographie et hémodynamique doppler du coeur et des vaisseaux intrathoraciques du foetus
JQQM010 Échographie biométrique et morphologique d'une grossesse uniembryonnaire au 1er trimestre
JQQM015 Échographie biométrique et morphologique d'une grossesse multiembryonnaire au 1er trimestre
JQQM016 Échographie biométrique et morphologique d'une grossesse unifoetale au 3ème trimestre
JQQM017 Échographie biométrique et morphologique d'une grossesse multifoetale au 3ème trimestre
JQQM018 Échographie biométrique et morphologique d'une grossesse unifoetale au 2ème trimestre

Complications of pregnancy

Monofetal pregnancy

CODE INTITULE
O800 Accouchement spontané par présentation du sommet
O801 Accouchement spontané par présentation du siège
O808 Autres accouchements uniques et spontanés
O809 Accouchement spontané, sans précision
Z370 Naissance unique, enfant vivant
Z371 Naissance unique, enfant mort- né
Z3710 Naissance unique, enfant mort-né, hors interruption de la grossesse pour motif médical
Z3711 Naissance unique, enfant mort-né, à la suite d'une interruption de la grossesse pour motif médical

Multiple pregnancy

CODE INTITULE
Z374 Naissance gémellaire, jumeaux morts-nés
Z3740 Naissance gémellaire, jumeaux morts-nés, hors interruption de la grossesse pour motif médical
Z3741 Naissance gémellaire, jumeaux morts-nés, à la suite d'une interruption de la grossesse pour motif médical
Z373 Naissance gémellaire, l'un des jumeaux né vivant, l'autre mort-né
Z3730 Naissance gémellaire, l'un des jumeaux né vivant, l'autre mort-né, hors interruption de la grossesse pour motif médical
Z3731 Naissance gémellaire, l'un des jumeaux né vivant, l'autre mort-né, à la suite d'une interruption de la grossesse pour motif médical
Z372 Naissance gémellaire, jumeaux nés vivants
O840 Accouchements multiples, Tous spontanés
O841 Accouchements multiples, Tous avec forceps et ventouse
O842 Accouchements multiples, Tous par césarienne
Z3771 Autres naissances multiples, Tous morts-nés, à la suite d'une interruption de la grossesse pour motif médical
Z3770 Autres naissances multiples, Tous morts-nés, hors interruption de la grossesse pour motif médical
Z377 Autres naissances multiples, Tous morts-nés
Z3761 Autres naissances multiples, certains enfants nés vivants, à la suite d'une interruption de la grossesse pour motif médical
Z3760 Autres naissances multiples, certains enfants nés vivants, hors interruption de la grossesse pour motif médical
Z376 Autres naissances multiples, certains enfants nés vivants
Z375 Autres naissances multiples, Tous nés vivants
Z372 Naissance gémellaire, jumeaux nés vivants
O849 Accouchements multiples, sans précision
O848 Autres accouchements multiples
O300 Grossesse multiple: Jumeaux
O301 Grossesse multiple: triplés
O302 Grossesse multiple: quadruplés
O308 Autres grossesses multiples
O309 Grossesse multiple, sans précision
O310 Foetus papyracé
O311 Poursuite de la grossesse après avortement d'un ou plusieurs foetus
O312 Poursuite de la grossesse après mort intra-utérine d'un ou plusieurs foetus
O318 Autres complications spécifiques à Une grossesse multiple

Pregnancy-induced hypertension/preeclampsia/eclampsia

CODE INTITULE
O14 Prééclampsie
O140 Prééclampsie modérée
O141 Prééclampsie sévère
O142 Syndrome HELLP
O149 Prééclampsie, sans précision
ECLAMPSIE
O150 Éclampsie au cours de la grossesse
O151 Éclampsie au cours du travail
O152 Éclampsie au cours de la puerpéralité
O159 Éclampsie, sans précision quant à la période
O13 Hypertension gestationnelle
O100 Hypertension essentielle préexistante compliquant la grossesse, l'accouchement et la puerpéralité
O101 Cardiopathie hypertensive préexistante compliquant la grossesse, l'accouchement et la puerpéralité
O102 Néphropathie hypertensive préexistante compliquant la grossesse, l'accouchement et la puerpéralité
O103 Cardionéphropathie hypertensive préexistante compliquant la grossesse, l'accouchement et la puerpéralité
O104 Hypertension secondaire préexistante compliquant la grossesse, l'accouchement et la puerpéralité
O109 Hypertension préexistante compliquant la grossesse, l'accouchement et la puerpéralité, sans précision
O16 Hypertension de la mère, sans précision

Threat of premature delivery

CODE INTITULE
O470 Faux travail avant 37 semaines entières de gestation [menace d'accouchement prématuré]

Premature membrane rupture

CODE INTITULE
O420 Rupture prématurée des membranes, avec début du travail dans les 24 heures
O421 Rupture prématurée des membranes, avec début du travail au-delà des 24 heures
O422 Rupture prématurée des membranes, travail retardé par traitement
O429 Rupture prématurée des membranes, sans précision

Obesity/excessive weight gain

CODE INTITULE
E66 Obésité
E660 Obésité due à un excès calorique
E6600 Obésité due à un excès calorique de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 30 kg/m2 et inférieur à 40 kg/m2, ou obésité due à un excès calorique de l'enfant
E6601 Obésité due à un excès calorique de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 40 kg/m2 et inférieur à 50 kg/m2
E6602 Obésité due à un excès calorique de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 50 kg/m2
E6609 Obésité due à un excès calorique de l'adulte, indice de masse corporelle [IMC] non précisé
E661 Obésité médicamenteuse
E6610 Obésité médicamenteuse de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 30 kg/m2 et inférieur à 40 kg/m2, ou obésité médicamenteuse de l'enfant
E6611 Obésité médicamenteuse de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 40 kg/m2 et inférieur à 50 kg/m2
E6612 Obésité médicamenteuse de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 50 kg/m2
E6619 Obésité médicamenteuse de l'adulte, indice de masse corporelle [IMC] non précisé
E662 Obésité extrême avec hypoventilation alvéolaire
E6620 Obésité extrême avec hypoventilation alvéolaire, avec indice de masse corporelle égal ou supérieur à 30 kg/m2 et inférieur à 40 kg/m2
E6621 Obésité extrême avec hypoventilation alvéolaire, avec indice de masse corporelle égal ou supérieur à 40 kg/m2 et inférieur à 50 kg/m2
E6622 Obésité extrême avec hypoventilation alvéolaire, avec indice de masse corporelle égal ou supérieur à 50 kg/m2
E6629 Obésité extrême avec hypoventilation alvéolaire, avec indice de masse corporelle inconnu
E668 Autres obésités
E6680 Autres obésités de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 30 kg/m2 et inférieur à 40 kg/m2, ou autres obésités de l'enfant
E6681 Autres obésités de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 40 kg/m2 et inférieur à 50 kg/m2
E6682 Autres obésités de l'adulte avec indice de masse corporelle égal [IMC] ou supérieur à 50 kg/m2
E6689 Autres obésités de l'adulte, indice de masse corporelle [IMC] non précisé
E669 Obésité, sans précision
E6690 Obésité sans précision de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 30 kg/m2 et inférieur à 40 kg/m2, ou obésité sans précision de l'enfant
E6691 Obésité sans précision de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 40 kg/m2 et inférieur à 50 kg/m2
E6692 Obésité sans précision de l'adulte avec indice de masse corporelle [IMC] égal ou supérieur à 50 kg/m2
E6699 Obésité sans précision de l'adulte, indice de masse corporelle [IMC] non précisé
0260 Prise de poids excessive pendant la grossesse

Gestational diabetes

CODE INTITULE
O244 Diabète sucré survenant au cours de la grossesse
O249 Diabète sucré au cours de la grossesse, sans précision

Haemorrhage during pregnancy and delivery

CODE Hémorragie du début de la grossesse
O200 Menace d'avortement
O208 Autres hémorragies du début de la grossesse
O209 Hémorragie du début de la grossesse, sans précision

Long-term condition status for depression or anxiety during pregnancy

CODE INTITULE
F0632 Troubles dépressif organique
F320 Episode dépressif léger
F3200 Episode dépressif léger, sans syndrome somatique
F3201 Episode dépressif léger, avec syndrome somatique
F321 Episode dépressif moyen
F3210 Episode dépressif moyen, sans syndrome somatique
F3211 Episode dépressif moyen, avec syndrome somatique
F322 Episode dépressif sévère sans symptômes psychotiques
F323 Episode dépressif sévère avec symptômes psychotiques
F3230 Episode dépressif sévère avec symptômes psychotiques congruents à l'humeur
F328 Autres épisodes dépressifs
F329 Episode dépressif, sans précision
F330 Troubles dépressif récurrent, épisode actuel léger
F3300 Troubles dépressif récurrent, épisode actuel léger, sans syndrome somatique
F3301 Troubles dépressif récurrent, épisode actuel léger, avec syndrome somatique
F331 Troubles dépressif récurrent, épisode actuel moyen
F3310 Troubles dépressif récurrent, épisode actuel moyen, sans syndrome somatique
F3311 Troubles dépressif récurrent, épisode actuel moyen, avec syndrome somatique
F332 Troubles dépressif récurrent, épisode actuel sévère sans symptômes psychotiques
F333 Troubles dépressif récurrent, épisode actuel sévère avec symptômes psychotiques
F3330 Troubles dépressif récurrent, épisode actuel sévère avec symptômes psychotiques congruents à l'humeur
F412 Troubles anxieux et dépressif mixte
F4322 Réaction mixte, anxieuse et dépressive
F920 Troubles des conduites avec dépression
F53 Troubles mentaux et du comportement associés à la puerpéralité, non classés ailleurs
F530 Troubles mentaux et du comportement légers associés à la puerpéralité, non classés ailleurs
F531 Troubles mentaux et du comportement sévères associés à la puerpéralité, non classés ailleurs
F538 Autres troubles mentaux et du comportement associés à la puerpéralité, non classés ailleurs
F539 Trouble mental de la puerpéralité, sans précision
0993 Troubles mentaux et maladies du système nerveux compliquant la grossesse, l'accouchement et la puerpéralité

Prematurity (childbirth before 37 weeks of gestation)

CODE INTITULE
O60 Travail et accouchement prématurés [avant Terme]

Low birth weight

CODE INTITULE
O365 Soins maternels pour croissance insuffisante du foetus

Suicide attempt/suicide

CODE INTITULE
X60 JUSQU A X84 Lésions auto-infligées
Z03.6 Mise en observation pour suspicion d'effet toxique de substances ingérées
Z91.5X-003 Tentative de suicide

Complications of childbirth

Caesarean delivery

CODE INTITULE
JQGA005 Accouchement par césarienne, par abord vaginal
JQGA002 Accouchement par césarienne programmée, par laparotomie
JQGA003 Accouchement par césarienne au cours du travail, par laparotomie
JQGA004 Accouchement par césarienne en urgence en dehors du travail, par laparotomie
YYYY069 Supplément pour accouchement multiple par césarienne
O820 Accouchement par césarienne programmée
O821 Accouchement par césarienne d'urgence
O822 Accouchement par césarienne avec hystérectomie
O828 Autres accouchements uniques par césarienne
O829 Accouchement par césarienne, sans précision
O842 Accouchements multiples, Tous par césarienne

Delivery with instruments/assistance

CODE INTITULE
O830 Accouchement avec extraction par le siège
O831 Autres accouchements par le siège
O832 Autres accouchements avec l'aide de manipulations
O833 Accouchement d'un foetus viable après grossesse abdominale
O834 Intervention destructrice lors d'un accouchement
O838 Autres accouchements précisés, avec assistance
O839 Accouchement avec assistance, sans précision
O810 Accouchement par forceps bas
O811 Accouchement par forceps à la partie moyenne de l'excavation
O812 Accouchement par forceps à la partie moyenne de l'excavation, avec rotation
O813 Accouchements par forceps, autres et sans précision
O814 Accouchement par extraction pneumatique
O815 Accouchement par association d'un forceps et d'une ventouse
JQGD011 Extraction instrumentale Sur tête dernière, au cours d'un accouchement par le siège

Haemorrhage during delivery

CODE Hémorragie précédent l'accouchement
O460 Hémorragie précédant l'accouchement avec anomalie de la coagulation
O468 Autres hémorragies précédant l'accouchement
O469 Hémorragie précédant l'accouchement, sans précision
CODE Hémorragie pendant l'accouchement
O670 Hémorragie pendant l'accouchement avec anomalie de la coagulation
O678 Autres hémorragies pendant l'accouchement
O679 Hémorragie pendant l'accouchement, sans précision
CODE Hémorragie du post-partum
O720 Hémorragie de la délivrance
O721 Autres hémorragies immédiates du post-partum
O722 Hémorragie du post-partum, tardive et secondaire
O723 Anomalie de la coagulation au cours du post-partum

Fetal distress

CODE INTITULE
O680 Travail et accouchement compliqués d'une anomalie du rythme cardiaque du foetus
O681 Travail et accouchement compliqués de la présence de méconium dans le liquide amniotique
O682 Travail et accouchement compliqués d'une anomalie du rythme cardiaque du foetus avec présence de méconium dans le liquide amniotique
O683 Travail et accouchement compliqués de signes biochimiques de détresse foetale
O688

Travail et accouchement compliqués d'autres signes de détresse foetale

O689 Travail et accouchement compliqués d'une détresse foetale, sans précision

Post partum

Psychiatric disorder related to the puerperium

CODES INTITULES
O993 Troubles mentaux et maladies du système nerveux compliquant la grossesse, l'accouchement et la puerpéralité
F53 Troubles mentaux et du comportement associés à la puerpéralité, non classés ailleurs

APPENDIX 4

Comparison of patients included in the study versus those not included

Non included patients n = 41,722 Included patients n = 760 p-value
Age at delivery 29.6 ± 5.4 32.3 ± 5.5 <0.001
<20 years 1,071 (2.6) 6 (0.8) <0.001
20–35 years 3,595 (82.9) 522 (68.7)
>35 years 6,056 (14.5) 232 (30.5)
Gestational age 39.0+/−2.1 38.7+/−2.3 0.001
CMU 7,423 (17.8) 153 (20.1) 0.10
Multiparous 21,889 (52.5) 385 (50.7) 0.32
Medical assisted procreation 955 (2.3) 14 (1.8) 0.41
Psychiatric comorbidities during hospitalization or long-term affection
Depression/anxiety
  • Before pregnancy
854 (2.1) 165 (21.7) <0.001
  • 6 months before pregnancy
69 (0.2) 39 (5.1) <0.001
Psychiatric disorder
  • Before pregnancy
311 (0.8) 80 (10.5) <0.001
  • 6 months before pregnancy
18 (0) 7 (0.9) <0.001
Suicide attempt
  • Before pregnancy
373 (0.9) 62 (8.2) <0.001
  • 6 months before pregnancy
46 (0.1) 21 (2.8) <0.001
History of complicated pregnancy 188 (0.5) 13 (1.7) <0.001
History before pregnancy
Alcohol 134 (0.3) 25 (3.3) <0.001
Tobacco 1,291 (3.1) 61 (8.0) <0.001
Drugs 115 (0.3) 24 (3.2) <0.001
Obesity 939 (2.3) 41 (5.4) <0.001
  • CMU, Universal health insurance.

APPENDIX 5

Model selection criteria

Model LogLikelihood AIC BIC Entropy % smaller class
1-class −2,941,261 5,892,522 5,915,689
2-class −2,341,353 4,704,707 4,755,673 0.886 30.9
3-class −2,219,131 4,472,261 4,551,028 0.934 9.9
4-class −2,167,099 4,380,198 4,486,765 0.947 10.1
5-class −2,134,851 4,327,703 4,462,069 0.945 5.5
  • Abbreviations: AIC, Akaike information criterion; BIC, Bayesian information criterion.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.