Aten Primaria. 2019;51(10):626---636 Atención Primaria www.elsevier.es/ap ORIGINAL ARTICLE Association between affective disorders presenting before and during pregnancy and pre-term birth, considering socio-demographic factors, obstetric factors, health conditions, and use of medication Arantxa Aliagaa , Oriol Cunillerab,c , María del Carmen Amadord , Mercedes Aliagae , Martí Arqueronsf , Jesús Almedab,c,∗ a Equip d’Atenció Primària el Castell, SAP Baix Llobregat Litoral, Insitut Català de la Salut (ICS) Castelldefels, Spain b Unitat de Suport a la Recerca Costa de Ponent, Institut Universitari d’Investigació en Atenció Primària Jordi Gol (IDIAP Jordi Gol), Cornellà de Llobregat, Spain c Universitat Autònoma de Barcelona, Bellaterra (Cerdanyola del Vallès), Spain d Atenció a la Salut Sexual I Reproductiva Delta del Llobregat, SAP Baix Llobregat Litoral, ICS, Castelldefels, Spain e ABS la Torrassa, Hospitalet de Llobregat, Barcelona, Spain f Facultad de Medicina, Campus de Bellvitge, Universitat de Barcelona, Hospitalet de Llobregat, Barcelona, Spain Received 15 November 2017; accepted 26 June 2018 Available online 16 November 2018 KEYWORDS Abstract Pregnancy; Objective: To study the relationship between maternal affective disorders (AD) before and Affective disorders; during pregnancy, and pre-term birth. Infant; Design: Retrospective observational study. Premature; Location: Sexual and reproductive health units at the Institut Català de la Salut (ICS) in Cat- Risk factors; alonia, Spain. Antidepressive agents Participants: Pregnant women with a result of live-born child from 1/1/2012 to 30/10/2015. Interventions: Data were obtained from the ICS Primary Care electronic medical record. Main measurements: Diagnosis of AD before and during pregnancy, months of pregnancy, and possible confusion factors were collected. Descriptive statistical analysis (median, interquartile range, and absolute and relative frequency), bivariate analysis (Wilcoxon test and Chi-square test), and multivariate analysis (logistic regression) were performed. Results: 102,086 women presented valid information for the study. Prevalence of AD during pregnancy was 3.5% (4.29% in pre-term and 3.46% in term births; p < 0.004). Pregnant women with pre-term births presented a higher age, smoking habit, lower inter-pregnancy interval, and a lower socio-economic status. Pre-term birth was significantly associated to previous history of stress and dissociative disorder (SDD), anxiety, obsessive-compulsive disorder (OCD) and eating disorders (ED), and use of antidepressants. It was also associated to abuse of alcohol, smoking, ∗ Corresponding author. E-mail address: jalmeda@ambitcp.catsalut.net (J. Almeda). https://doi.org/10.1016/j.aprim.2018.06.009 0212-6567/© 2018 The Authors. Published by Elsevier España, S.L.U. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). Maternal affective disorders and preterm birth 627 and use of psychoactive substances, as well as SDD, ED, use of antipsychotics, and divorce during pregnancy. Multivariate analysis confirmed the relationship between pre-term birth and history of AD, SDD, ED, and smoking, but not with AD during pregnancy. Conclusions: Examining the previous history of SDD and ED in pregnant women, and SDD, and ED during pregnancy is highly relevant to avoid pre-term birth. © 2018 The Authors. Published by Elsevier España, S.L.U. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). PALABRAS CLAVE Asociación entre los trastornos afectivos que se presentan antes y durante Embarazo; el embarazo y el parto prematuro, teniendo en cuenta los factores Trastornos psicóticos sociodemográficos, los factores obstétricos, las condiciones de salud y el uso de afectivos; medicamentos Recién nacido Resumen Prematuro; Objetivo: Estudiar la relación entre diagnósticos de trastornos afectivos (TA) antes y durante Factores de riesgo; el embarazo, y factores de confusión con prematuridad del neonato. Antidepresivos Diseño: Estudio observacional retrospectivo. Emplazamiento: Servicios de atención sexual y reproductiva del Institut Català de la Salut (ICS) en Cataluña, España. Participantes: Embarazadas atendidas con resultado de hijo vivo del 1/1/2012 al 30/10/2015. Intervenciones: Datos recogidos en la base de datos de la historia clínica informatizada. Mediciones: Se recogió los diagnósticos de TA antes y durante el embarazo, meses de gestación y posibles factores de confusión. Se realizó análisis estadístico descriptivo (mediana y rango intercuartílico y frecuencias absoluta y relativa), bivariante (test de Wilcoxon y Chi-cuadrado) y multivariante (regresión logística). Resultados: Ciento dos mil ochenta y seis mujeres presentaban información válida para el estu- dio. La prevalencia de TA durante el embarazo fue del 3,5% (4,29% en prematuros y 3,46% en a término; p < 0,004). Las embarazadas con partos prematuros presentan mayor edad, más tabaquismo, menor tiempo entre embarazos y menor nivel socioeconómico. La prematuridad se asoció a antecedentes previos de trastorno por estrés y disociativo (TED), de ansiedad y obsesivo-compulsivo, de conducta alimentaria (TCA) y uso de antidepresivos. También a abuso de alcohol, tabaco y sustancias psicoactivas; TED, TCA, uso de antipsicóticos y divorcio durante el embarazo. El análisis multivariante confirmó la relación de prematuridad con antecedentes de TA, TED, TCA y tabaquismo, pero no con TA durante el embarazo. Conclusiones: Es importante explorar antecedentes de TED y TA en la embarazada y los TED durante el embarazo, para disminuir la prematuridad. © 2018 Los Autores. Publicado por Elsevier España, S.L.U. Este es un artı́culo Open Access bajo la licencia CC BY-NC-ND (http://creativecommons.org/licenses/by-nc-nd/4.0/). Introduction Other clinical obstetric and therapeutic factors associ- ated to pre-term birth are collected on the Clinical Practice Depression is the most frequent mental disorder in women. Guideline for care in Pregnancy and Puerperium in the It may alter the course of pregnancy, affect the foetus, and Spanish National Health Service (NHS),12 on the ‘Proto- the newborn, having consequences even on adolescence.1,2 col de Seguiment de l’embaràs a Catalunya’ (Protocol on Prevalence of depression during pregnancy ranges from pregnancy monitoring in Catalonia),13 and on numerous 4.8% to 33.2%,3 with under 50% of cases identified in the publications14,15 (Table S1). Moreover, pre-term birth has everyday clinical practice.4 Depression and bipolar disorder been associated with stress,16,17 and low socio-economic have been associated to pre-term birth.5---10 Studies on status.7 the relationship between affective disorders and pre-term The Spanish NHS guideline systematically evaluates post- birth, conducted in the Mediterranean and South European partum depression, but it does not recommend a similar countries, and included in systematic reviews, are scarce, population screening during pregnancy, unlike other coun- therefore, no data on this matter are available in our tries. It neither recommends a systematic mental health country. Pre-term birth has been associated to prolonged examination on women of childbearing age. stays at paediatric intensive care units or neonatal depart- The current study aims to study the relationship between ments, a higher incidence of infection, and of respiratory, maternal affective disorders (AD) and pre-term birth, con- digestive, ophthalmologic, psychomotor conditions, etc.,11 trolling by the confusion factors associated to pre-term resulting in a remarkable impact on the health care system. birth, and differentiating between those factors which 628 A. Aliaga et al. appeared priorly, and those presenting during pregnancy. disorder (OCD), etc. Whether these diagnoses were active Since mental health disorders are not as extensively inves- before or during pregnancy was taken into consideration. tigated as physical conditions, on the patient anamnesis Finally, all treatments related to pre-term birth, based performed on women in family and maternal and child pri- on clinical guides and literature, were registered, and coded mary healthcare services, we will dedicate special attention according to the ‘‘Anatomical Therapeutic Chemical Classi- to them. fication’’ (WHOCC-ATC)21 (Table S2). It was also recorded This study is framed within the primary healthcare area, whether medication was taken before or during pregnancy. being its implementation necessary since to date, no similar clinical study has been performed in our country. Statistical analysis The aim is to study the relationship between maternal affective disorders (AD) before and during pregnancy, and Descriptive statistical analysis, showing median and pre-term birth. interquartile range for numeric variables, and absolute and relative frequency for categorical variables; A bivari- Material and methods ate analysis according to pre-term birth, was performed, through Wilcoxon test for numeric variables and Chi-square test for categorical variables. Design, area of study, and research subjects A multivariate logistic regression model was built (final model obtained by an AIC stepwise backwards variable Retrospective observational study conducted in the primary selection model from all statistically and clinically relevant health care area. variables on the bivariate analysis), without including pos- Data were obtained from the Information System for the sible interactions between them. Enhancement of Research in Primary Care (SIDIAP) database, All analyses were conducted by complete case (with- with an assigned population in 2012 of 5,835,000 patients out missing values being treated). Significance level of 0.05 (80% of the total population in Catalonia). This database is was considered as indicator of statistically significant dif- supplied, among other, by information from the Primary Care ferences. Treatment of data and statistical analyses were electronic medical records (e-CAP), and from the Sexual and performed using software R 3.2.4-revised. Reproductive Health Care Program (ASSIR) of the Catalan Institute of Health (ICS), used by obstetricians, gynaecol- ogists, and midwives in primary care. The ASSIR records Results include data on pregnancy, delivery, and postpartum mon- itoring from women who attended the primary health care Out of the 119,036 pregnancies which fulfilled the inclusion services of the ICS,18 accounting for 65.7% term pregnancies criteria, valid information for the study was obtained from in Catalonia, over the period of our study.19 102,086 women (85%). The prevalence of affective disorders All pregnant women registered in the SIDIAP, aged 15---47, during pregnancy was 3.5%, and the prevalence of pre-term who gave birth to live-born children in Catalonia from 1st birth 4.2%. Patients who presented pre-term birth showed January 2012 to 30 October 2015, with a gestational age a significantly higher age, a lower socio-economic status, a over 25 weeks and under 44 weeks, and with two foetuses higher history of previous pre-term birth, a higher history at most, were included in the study. of abortion, a shorter interval between pregnancies, and a higher rate of twin pregnancies (Table 1). Pre-term birth was also significantly associated with history of AD, OCD, SDD, Variables smoking, and ED, among other (Table 2). With regard to health conditions during pregnancy Pre-term birth was defined as gestational age under (Table 2), pre-term birth was bivariately associated with 37 weeks, based on the number of days from the date of anxiety, OCD, AD, SDD, divorce, smoking, mental disorder last menstruation to the date of birth. due to psychoactive substance use, abuse of alcohol and AD included the following ICD-10 codes: F30 (Manic tobacco, among other. It was also associated with history episode), F31 (Bipolar disorder), F32 (Depressive disorders), of antidepressant intake and use of antipsychotics during F33 (Recurrent depressive disorder), F34 (Persistent mood pregnancy, among other (Table 3). affective disorder), F38 (Other mood affective disorders), Pre-term birth was significantly associated with history F39 (Unspecified mood affective disorders), presenting of AD, SDD, ED, and smoking among other, in the final before pregnancy (at any time of life) or during pregnancy. multivariate model. It was also associated with presence of Socio-economic status was classified according to the smoking and mental disorder due to use of alcohol during Deprivation index MEDEA,20 presented in quintiles (the pregnancy, among other (Table 4). The association between highest quintile, the lower socio-economic status). All the AD before pregnancy and pre-term birth, in terms of variables related to pre-term birth found on the clinical punctual estimation of the adjusted OR (1.32 95%CI: [1.04, guides, and on the conducted bibliographic search, were 1.64]) was inferior than Eclampsia (3.22 [2.63, 3.90]), collected7,12---17 (Table S1). Considering the ICD-10,099 code DM (2.31 [1.56, 3.31]), and HTA (2.15 [1.64, 2.79]) and includes all mental disorders, those most discussed and con- Placenta praevia, detachment, antepartum haemorrhage troverted on the reviewed literature, were chosen. Those (1.80 [1.53, 2.11]), and superior than smoking (non-smokers AD interrelated, especially when scarce in number, were OR: 0.92 [0.86, 0.98]). Presence of SDD during pregnancy gathered, namely, stress and dissociative disorder (SDD), was relevant, providing plausibility to the model, although eating disorder (ED), anxiety disorder, obsessive compulsive its marginal effect did not reach statistical significance. Maternal affective disorders and preterm birth 629 565,544 Pregnancy registers in SIDIAP 16,821 Non-active mother in SIDIAP register: -16,757 reallocated -64 dead 548,723 Pregnancy registers in SIDIAP with active mother 368,778 prior to 2012 179,945 Pregnancy registers in SIDIAP from 01/2012 to 10/2015 with active mother 313 Maternal age out of range: -130 <15 years 179,632 -183 >47 years Pregnancy registers in SIDIAP from 01/2012 to 10/2015 with active mother aged 15-47 20,055 unclassifiable labour 159,577 Pregnancy registers in SIDIAP from 01/2012 to 10/2015,with active mothe raged 15-47 and classifiable labour (normal, caesarean, preterm,other) 40,506 Gestational age out of range: - 16 codified <0 weeks - 14,693 [0-15) weeks - 2,771 [15-25) weeks 119,071 - 23,026 >44 weeks Pregnancy registers in SIDIAP from 01/2012 to 10/2015 with active mother aged 15-47, gestational age between 25 and 44 weeks, and classifiable labour(normal, caesarean, preterm, other) 35 Labour >2 newborns 119,036 Pregnancy registers in SIDIAP from 01/2012 to 10/2015 with active mother aged 15-47, gestational age between 25 and 44 weeks, and classifiable labour (normal,caesarean, preterm, other) under 3 newborns 6,896 unknown labour status 112,140 pregnancies with defined labour status (normal, caesarean, preterm) 10,054 without information on newborn 102,086 pregnancies with 1 or 2 foetus,with information Study schema. Retrospective observational study with the aim to investigate the relationship between maternal affective disorders (AD) before and during pregnancy, and pre-term birth. 630 A. Aliaga et al. Table 1 Descriptive global and bivariate analysis of maternal characteristics, previous pregnancies, and labour regarding preterm birth. N = 102,086 Missing Global (n = 102,086) Normal (n = 97,800) Preterm (n = 4,286) p-Value Socio-demographic characteristics Agea 0 31.00 [27.00, 35.00] 31.00 [27.00, 35.00] 32.00 [28.00, 36.00] <0.001 MEDEA deprivation index 0 0.038 Rural 15,487 (15.17%) 14,804 (15.14%) 683 (15.94%) Quintile 1 (less disadvantaged) 9,553 (9.36%) 9,188 (9.39%) 365 (8.52%) Quintile 2 24,866 (24.36%) 23,864 (24.40%) 1,002 (23.38%) Quintile 3 14,508 (14.21%) 13,913 (14.23%) 595 (13.88%) Quintile 4 17,236 (16.88%) 16,514 (16.89%) 722 (16.85%) Quintile 5 (more disadvantaged) 20,436 (20.02%) 19,517 (19.96%) 919 (21.44%) Partnership status 0 0.468 Single 37,279 (36.52%) 35,691 (36.49%) 1,588 (37.05%) Partner 64,807 (63.48%) 62,109 (63.51%) 2,698 (62.95%) Previous delivery Term deliveryb 1,334 1.07 (1.01) 1.08 (1.01) 0.81 (0.99) <0.001 Preterm deliveryb 1,334 0.03 (0.19) 0.02 (0.16) 0.22 (0.52) <0.001 Abortionb 1,334 0.46 (0.79) 0.46 (0.79) 0.53 (0.88) <0.001 Live bornb 1,334 1.08 (1.01) 1.09 (1.00) 0.98 (1.02) <0.001 Inter-pregnancy interval 0 <0.001 (>18 months) 85,813 (84.06%) 82,292 (84.14%) 3,521 (82.15%) [12---18] months 5,334 (5.23%) 5,102 (5.22%) 232 (5.41%) [0---12] months 10,939 (10.72%) 10,406 (10.64%) 533 (12.44%) Delivery characteristics Type 14,439 <0.001 Eutocic 56,808 (64.81%) 54,974 (65.36%) 1,834 (51.81%) Dystocic 30,839 (35.19%) 29,133 (34.64%) 1,706 (48.19%) Number of foetuses 0 <0.001 1 100,624 (98.57%) 96,872 (99.05%) 3,752 (87.54%) 2 1,462 (1.43%) 928 (0.95%) 534 (12.46%) Qualitative variables are expressed in absolute and relative frequency for each category, and numeric variables are expressed in (a) median and interquartile range, or (b) media and standard deviation. Variable distribution between preterm and term labour groups is compared by Wilcoxon test for numeric variables, and Chi-square test for categorical variables. Association between AD during pregnancy and pre-term the 10 cut-off point for ‘minor’ or ‘major depression’ on the birth was not statistically significant. English version, and 0.81 and 0.92 respectively, on the non- English versions4 ; The sensitivity and specificity values for Discussion the ‘Patient Health Questionnaire’ (PHQ) are 0.79 and 0.75 respectively. Accordingly, the use of this type of screening This study shows an association between previous history tools -with specificity far from ideal- to screen population of maternal AD and pre-term birth. The prevalence of pre- for low prevalence conditions automatically triggers preva- term birth obtained in our study was 4.2%, lower than the 7% lence values. expected.19 This may have been due to the restrictive inclu- Whereas the definition of AD, would lead one to expect sion criteria considered in the study, and to the fact that in a higher prevalence than that found in those studies lim- 2012 not all ASSIR relied on the same electronic register. ited to depression (most of them), it is to be expected a The prevalence of AD during pregnancy obtained in lower prevalence of AD in the population seen in primary our study is considerably lower than that published on care (general population) than in the population seen in most studies which use depression scales without diagnos- specialized centres. tic confirmation (often self-reported),3,4 being comparable Despite identifying a bivariate association between to that reported in studies which use clinical diagnoses,22 AD during pregnancy and pre-term birth initially, we did and even to that found in studies performed in psychiatric not observe this association on the multivariate analysis, in-patient wards.8 after considering history of AD (prior to pregnancy), and Most widely used health evaluation instruments present other factors like the use of relevant medication, whose sensitivity and specificity values which are far from ideal. bivariate association with pre-term birth had also resulted Therefore, the ‘Edinburgh Postnatal Depression Scale’ significant. These results contradict the findings of the two (EPDS) shows a sensitivity of 0.72 and a specificity of 0.85 for meta-analysis published on this matter23,24 ; although the Maternal affective disorders and preterm birth 631 Table 2 [ll1]Health conditions before birth. Descriptive and bivariate analysis according to preterm birth. Previous history During pregnancy Global Normal Preterm p-Value Global Normal Preterm p-Value (n = 97,800) (n = 4,286) (n = 97,800) (n = 4,286) Affective disorder 1,631 1,533 98 <0.001 3,571 3,387 184 0.004 (1.60%) (1.57%) (2.29%) (3.50%) (3.46%) (4.29%) Anxiety disorder and 6,775 6,427 348 <0.001 14,996 14,294 702 0.002 obsessive-compulsive (6.64%) (6.57%) (8.12%) (14.69%) (14.62%) (16.38%) disorder Dissociative and stress 746 695 51 <0.001 1,922 1,818 104 0.009 disorder (0.73%) (0.71%) (1.19%) (1.88%) (1.86%) (2.43%) Smoker 7,962 7,623 339 0.806 22,533 21,491 1,042 <0.001 (7.80%) (7.79%) (7.91%) (22.07%) (21.97%) (24.31%) Non-smoker 2,455 2,372 83 0.046 47,962 46,045 1,917 0.003 (2.40%) (2.43%) (1.94%) (46.98%) (47.08%) (44.73%) Mental and behavioural 2,869 2,756 113 0.511 15,961 15,217 744 0.002 disorders due to use of (2.81%) (2.82%) (2.64%) (15.63%) (15.56%) (17.36%) tobacco Eating disorder 207 192 15 0.044 833 803 30 0.438 (0.20%) (0.20%) (0.35%) (0.82%) (0.82%) (0.70%) Mental and behavioural 125 120 5 1.000 451 422 29 0.024 disorders due to use of (0.12%) (0.12%) (0.12%) (0.44%) (0.43%) (0.68%) psychoactive substances Mental and behavioural 69 63 6 0.118 277 250 27 <0.001 disorders due to use of (0.07%) (0.06%) (0.14%) (0.27%) (0.26%) (0.63%) alcohol Divorce 287 275 12 1.000 1,468 1,387 81 0.013 (0.28%) (0.28%) (0.28%) (1.44%) (1.42%) (1.89%) In Vitro fertilization 147 135 12 0.028 573 503 70 <0.001 (0.14%) (0.14%) (0.28%) (0.56%) (0.51%) (1.63%) Caesarean 1,033 973 60 0.012 334 322 12 0.677 (1.01%) (0.99%) (1.40%) (0.33%) (0.33%) (0.28%) Twin pregnancy 39 36 3 0.491 169 114 55 <0.001 (0.04%) (0.04%) (0.07%) (0.17%) (0.12%) (1.28%) Placental pathology 13 12 1 1.000 23 19 4 0.008 (0.01%) (0.01%) (0.02%) (0.02%) (0.02%) (0.09%) Placenta praevia, 780 752 28 0.446 2,623 2,413 210 <0.001 detachment, (0.76%) (0.77%) (0.65%) (2.57%) (2.47%) (4.90%) antepartum haemorrhage Pre-eclampsia-Eclampsia 417 383 34 <0.001 1,046 892 154 <0.001 (0.41%) (0.39%) (0.79%) (1.02%) (0.91%) (3.59%) Hypertension 110 101 9 0.065 852 751 101 <0.001 (0.11%) (0.10%) (0.21%) (0.83%) (0.77%) (2.36%) 632 A. Aliaga et al. Table 2 (Continued) Previous history During pregnancy Global Normal Preterm p-Value Global Normal Preterm p-Value (n = 97,800) (n = 4,286) (n = 97,800) (n = 4,286) Glomerular disease, 1,432 1,366 66 0.475 1,608 1,524 84 0.045 tubulo-interstitial (1.40%) (1.40%) (1.54%) (1.58%) (1.56%) (1.96%) disease, lithiasis, other Metabolic disorder 310 293 17 0.323 3,829 3,603 226 <0.001 (0.30%) (0.30%) (0.40%) (3.75%) (3.68%) (5.27%) Viral Hepatitis 270 249 21 0.005 769 735 34 0.827 (0.26%) (0.25%) (0.49%) (0.75%) (0.75%) (0.79%) Diabetes Mellitus 49 44 5 0.082 372 333 39 <0.001 (0.05%) (0.04%) (0.12%) (0.36%) (0.34%) (0.91%) Toxic liver disease 5 3 2 0.004 8 7 1 0.772 (0.00%) (0.00%) (0.05%) (0.01%) (0.01%) (0.02%) Sexual transmission 251 233 18 0.028 789 761 28 0.410 disease (0.25%) (0.24%) (0.42%) (0.77%) (0.78%) (0.65%) Viral disease 14,230 13,577 653 0.013 5,219 4,982 237 0.218 (13.94%) (13.88%) (15.24%) (5.11%) (5.09%) (5.53%) Mycosis (Aspergillus, 47 46 1 0.731 96 86 10 0.005 etc.) (0.05%) (0.05%) (0.02%) (0.09%) (0.09%) (0.23%) Blood disorder 58 52 6 0.045 479 459 20 1.000 (0.06%) (0.05%) (0.14%) (0.47%) (0.47%) (0.47%) Congenital malformation 126 119 7 0.591 1,127 1,062 65 0.010 (0.12%) (0.12%) (0.16%) (1.10%) (1.09%) (1.52%) Cerebral palsy and other 4 4 0 1.000 29 25 4 0.035 (0.00%) (0.00%) (0.00%) (0.03%) (0.03%) (0.09%) Vestibular disorder 2325 2,208 117 0.048 1,220 1,165 55 0.638 (2.28%) (2.26%) (2.73%) (1.20%) (1.19%) (1.28%) Benign tumoura 493 452 41 <0.001 2319 2,198 121 0.015 (0.48%) (0.46%) (0.96%) (2.27%) (2.25%) (2.82%) Tumour 37 35 2 1.000 113 102 11 0.007 (0.04%) (0.04%) (0.05%) (0.11%) (0.10%) (0.26%) a Digestive system, respiratory, intrathoracic, bone and cartilage, lipomatous, mesothelial, conjunctive, ovarian, uterus, meningeal, encephalous. Absolute and relative frequency of women presenting the characteristic are shown for each variable. Variable distribution between preterm and term labour groups is compared by Chi-square test. This description and test are conducted on variables prior to pregnancy (previous history) and variables during pregnancy. authors themselves recognize their studies’ limitations: considered significantly fewer confusion factors than our Grigoriadis23 concluded that the association was modest and work, which may explain this assumed difference. the quality of evidence insufficient, and Grote24 reported However, our findings are consistent with more recent lack of homogeneity among the evaluated studies. Both publications: Mei-Danobserved an association between Maternal affective disorders and preterm birth 633 Table 3 Use of medication prior to birth. Descriptive and bivariate analysis according to preterm birth. Previous history During pregnancy Global Normal Preterm p-Value Global Normal Preterm p-value (n = 102,086) (n = 97,800) (n = 4,286) (n = 102,086) (n = 97,800) (n = 4,286) Anticoagulants 3,548 3,385 163 0.249 1,770 1,578 192 <0.001 (3.48%) (3.46%) (3.80%) (1.73%) (1.61%) (4.48%) Antidepressants 11,529 10,962 567 <0.001 4,143 3,936 207 0.010 (11.29%) (11.21%) (13.23%) (4.06%) (4.02%) (4.83%) Androgens 21,109 20,253 856 0.252 9,125 8,589 536 <0.001 (20.68%) (20.71%) (19.97%) (8.94%) (8.78%) (12.51%) Antipsychotics 2,179 2,070 109 0.066 407 374 33 <0.001 (2.13%) (2.12%) (2.54%) (0.40%) (0.38%) (0.77%) ACEI 382 337 45 <0.001 191 160 31 <0.001 (0.37%) (0.34%) (1.05%) (0.19%) (0.16%) (0.72%) Progestagens 18,744 17,993 751 0.153 8,191 7,704 487 <0.001 (18.36%) (18.40%) (17.52%) (8.02%) (7.88%) (11.36%) Absolute and relative frequency of use, are shown for each medication. Variable distribution between preterm and term labour groups is compared by Chi-square test. This description and test are conducted on use of medication prior to pregnancy (previous history) and during pregnancy. ACEI = angiotensin converting enzyme inhibitors. history of maternal AD and pre-term birth, after com- birth.16,17 In our study both LSES and diagnosis of SDD paring pregnant women who had been hospitalized were associated to pre-term birth in the bivariate analysis. due to bipolar disorder or major depression within the Multivariate analysis showed association between previous 5 years prior to pregnancy, with those without docu- history of SDD and pre-term birth; however, the association mented history of mental disorder. We consider that of SDD during pregnancy with pre-term birth was not sta- the higher the diagnostic specificity, the stronger this tistically significant, although by a very small margin. Most association becomes. Furthermore, Mannisto,22 observed authors underline the relevance of identifying and alleviat- an association between pre-term birth and maternal ing exposure to stress during pregnancy, as far as possible, to depression or bipolar disorder, based on electronic medical reduce pre-term birth26 ; our results demonstrate the impor- records and maternal psychiatric diagnoses from discharge tance of evaluating the presentation of these disorders prior summaries. Räisänen7 found a significant association to pregnancy. between depression and pre-term birth, being smoking With regards to use of antidepressants, two meta-analysis a relevant contributing factor, and history of depression recently conducted on the antidepressant-pre-term birth prior to pregnancy the risk factor with the strongest association, Huang and Eke,27,28 considering exclusively association. selective serotonin reuptake inhibitors (SSRIs), found an On the other hand, considering the generalized concept association between the use of SSRIs and pre-term birth. that prevalence of depression may be lower in the Mediter- Viktorin, just observed a small reduction in gestational ranean area, since our study was conducted in a Spanish age associated to SSRIs.29 Sujan published in 2017 a ret- population, we may expect a lower rate of AD than rospective cohort study, in which association between use that found in North European or North American popula- of antidepressants, in general, during the first trimester of tions. However, last report from the WHO contradicts this pregnancy and pre-term birth was observed.30 Our results, assumption,25 having reported a higher relevance of low considering all antidepressants --- except mood stabilizers- socio-economic status on the prevalence of this disorder, are consistent with these findings in the bivariate analysis, regardless of the country in which the study was conducted. although do not coincide in the multivariate analysis. This Our work includes a population from all socio-economic sta- discrepancy suggests the need to conduct further studies on tus, which may explain the lower prevalence of depression, the effect of the various antidepressant drugs on pre-term and is novel since addresses pregnant women resident in birth. Catalonia. With reference to the relationship between use of Some authors have observed association between low antipsychotic medication during pregnancy and pre-term socio-economic status (LSES) and situations of chronic birth, we observed a statistically significant association, stress, and between both with depression and pre-term consistent with the findings of Galbally.14 It seems sensible 634 A. Aliaga et al. Table 4 Multivariate analysis on health conditions prior and during pregnancy and preterm birth. Odds Ratio [CI95%] p-Value (Independent term) 0.00 [0.00, 0.00] <0.001 Affective disorders during pregnancy 1.07 [0.90, 1.26] 0.413 Previous births characteristics Term birth 0.54 [0.45, 0.63] <0.001 Preterm birth 4.68 [4.00, 5.45] <0.001 Abortion 1.07 [1.03, 1.11] 0.001 Live-born 1.44 [1.21, 1.71] <0.001 Socio-economic status Rural -reference- Quintile 1 (less disadvantaged) 0.92 [0.80, 1.05] 0.231 Quintile 2 0.92 [0.82, 1.02] 0.117 Quintile 3 0.91 [0.81, 1.03] 0.144 Quintile 4 0.98 [0.88, 1.10] 0.789 Quintile 5 (more disadvantaged)1.09 [0.97, 1.21] 0.136 Inter-pregnancy interval Without short interval -reference- 12 to 18 months 0.98 [0.84, 1.14] 0.822 Under 12 months 1.19 [1.07, 1.32] <0.001 Number of foetuses 14.32 [12.37, 16.55]<0.001 Previous history Affective disorder 1.32 [1.04, 1.64] 0.017 Dissociative and stress disorder 1.41 [1.02, 1.91] 0.030 Non-smoker 0.72 [0.56, 0.91] 0.007 Eating disorder 1.91 [1.05, 3.20] 0.022 Viral Hepatitis 1.79 [1.04, 2.90] 0.025 Toxic liver disease 12.75 [1.53, 83.93] 0.009 Benign tumoura 1.70 [1.18, 2.38] 0.003 During pregnancy Medication Anticoagulants 1.39 [1.15, 1.67] <0.001 Androgens 1.23 [1.10, 1.36] <0.001 Antipsychotics 1.61 [1.05, 2.36] 0.021 ACEI 1.88 [1.13, 3.03] 0.012 Health conditions Dissociative and stress disorder 1.24 [0.99, 1.53] 0.051 Non-smoker 0.92 [0.86, 0.98] 0.010 Mental and behavioural disorders due to use of alcohol 1.92 [1.20, 2.93] 0.004 In Vitro fertilization 1.46 [1.05, 2.00] 0.021 Placenta praevia, detachment, antepartum haemorrhage 1.80 [1.53, 2.11] <0.001 Eclampsia 3.22 [2.63, 3.90] <0.001 Hypertension 2.15 [1.64, 2.79] <0.001 Glomerular disease, tubulo-interstitial disease, lithiasis, other 1.20 [0.94, 1.52] 0.128 Metabolic disorder 1.30 [1.11, 1.51] <0.001 Diabetes Mellitus 2.31 [1.56, 3.31] <0.001 Mycosis (Aspergillus, etc.) 3.35 [1.60, 6.26] <0.001 Congenital malformation 1.24 [0.94, 1.61] 0.122 Cerebral palsy and other 2.72 [0.74, 7.74] 0.088 Tumour 2.38 [1.12, 4.51] 0.014 a Digestive system, respiratory, intrathoracic, bone and cartilage, lipomatous, mesothelial, conjunctive, ovarian, uterus, meningeal, encephalous. ACEI = angiotensin converting enzyme inhibitors. Results of the multivariate logistic regression model on preterm birth. Final model obtained by ‘‘stepwise backwards’’ variable selection model based on the ‘‘Akaike Information Criteria’’ from all statistically significant and clinically relevant variables on the bivariate analysis, as explicative variables. Results are presented in ‘‘Odds Ratio’’, with a confidence interval of 95%, estimated as the exponential of the obtained coefficient from the multivariate logistic regression model, and p-value of its statistical significance. Maternal affective disorders and preterm birth 635 to follow the general tendency to evaluate the risk-benefit psychosocial factors --- like the already available in many relationship of using this medication in the pregnant woman primary care devices, which have not received the atten- who needs it, thoroughly and individually.31 tion they deserve --- is undoubtedly, a suitable instrument to One of the limitations of this study is the possible exis- perform this task. tence of incorrect or unregistered diagnoses, inherent to the use of electronic medical. In primary care settings, the data are entered by general practitioners, due to their own diagnosis, or because they have seen it in specialist reports Key Points or hospital discharge reports (in these last two cases the data are not downloaded automatically in the computer- ized records of primary care, with the consequent delay in The known on the subject: the coding of hospital and specialist diagnoses). This can also ignore the diagnoses made in private care and not Depression is the most frequent mental disorder in seen in primary care. And finally, it is also of note that women and it may alter the course of pregnancy, affect data obtained from ASSIR of 2012 were incomplete, as the the foetus, and the newborn. ASSIR electronic medical records were not homogeneous Studies on the relationship between affective disorders then. and pre-term birth are scarce, therefore, and few data Although the big sample size and representativeness of on this matter are available in our country. the study is an important strength, it must be advised that it The Spanish NHS guideline systematically evaluates can affect to the statistical signification of tests, easily lead- postpartum depression, but it does not recommend a ing to significance. Nevertheless (and keeping that in mind similar population screening during pregnancy, unlike when valuing p-values), multivariate regression model was other countries. selected according to Akaike Information Criteria, and varia- bles were included or not in the final model with disregard What does this study contribute? of their p-values. On the other hand, our multivariate analysis only took This study shows an association between previous his- into account the effect of the independent variables without tory of maternal affective disorders and pre-term birth. interaction among them; since being such a large number, To investigate the history of affective disorders, and made it practically impossible to examine all variables two the psychosocial environment in the pregnant woman by two. Due to the exploratory nature of the study, it was is of great relevance, considering the potential stress- considered inadequate to force a hypothesis on a limited ful situations which may affect the course of pregnancy number of possible interactions. and avoiding pre-term birth and its severe conse- Nevertheless, the SIDIAP data base has been widely quences. validated. Unlike other data bases, all primary care pro- A thorough anamnesis, which includes psychosocial fac- fessionals in the ICS participate in it, not only voluntaries; tors --- like the already available in many primary care and benefits from a first-rate anonymity system,18 which can devices, which have not received the attention they open the way to conduct further research in our line, and in deserve --- is undoubtedly, a suitable instrument to per- other areas where similar data bases --- which include most form this task. of the population treated in the primary healthcare --- may be established. In conclusion, a previous history of AD in pregnant women is associated to pre-term birth. Classically, general habits Funding and physical conditions have received more attention when evaluating health in pregnancy. This study demonstrates This study received a research grant from the SIDIAP (Infor- that, at least regarding avoiding pre-term birth and its mation System for the Enhancement of Research in Primary severe consequences, to investigate the history of affective Care) of the University Institute for Research in Primary Care disorders, and the psychosocial environment in the preg- (IDIAP) Jordi Gol, in its 4th call for 2014. nant woman is of great relevance, considering the potential stressful situations which may affect the course of preg- Conflict of interest nancy. We coincide with some authors in that the aforemen- The authors declare that there is no conflict of interest tioned should be explored within a general examination, in regarding the publication of this article. a trusted environment, within the framework of the appro- priate physician-patient relationship in primary care; paying attention to avoid that the patient may feel stigmatized or Appendix A. Supplementary data fear social services intervention.32 This becomes easier if we regard that any women of childbearing age can be a mother, Supplementary data associated with this article can be and establish this desirable relationship from adolescence, found, in the online version, at doi:10.1016/j.aprim. whenever possible. A thorough anamnesis, which includes 2018.06.009. 636 A. Aliaga et al. References 17. Loomans EM, van Dijk AE, Vrijkotte TG, van Eijsden M, Stronks K, Gemke RJ, et al. Psychosocial stress during pregnancy is 1. World Health Organization. Depression, Fact sheet. Updated related to adverse birth outcomes: results from a large multi- February 2017. Available from: http://www.who.int/ ethnic community-based birth cohort. Eur J Public Health. mediacentre/factsheets/fs369/en [accessed 26.06.17]. 2013;23:485---91. 2. World Health Organization. Maternal and child mental health. 18. Bolibar B, Fina Avilés F, Morros R, Garcia-Gil M, del M, Her- Available from: http://www.who.int/mental health/maternal- mosilla E, et al., Grupo SIDIAP. Base de datos SIDIAP: la historia child/maternal mental health/en [accessed 26.06.17]. clínica informatizada de Atención Primaria como fuente de 3. Underwood L, Waldie K, D’Souza S, Peterson E, Morton S. información para la investigación epidemiológica. Med Clin. A review of longitudinal studies on antenatal and postnatal 2012;138:617---21. depression. Arch Women’s Ment Health. 2016;19:711---20. 19. Institut d’Estadística oficial de Catalunya --- IDESCAT Avail- 4. O’Connor E, Rossom RC, Henninger M, Groom HC, Burda BU. able from: https://www.idescat.cat/tema/naixe [accessed Primary care screening for and treatment of Depression in preg- 13.04.17]. nant and postpartum women: evidence report and systematic 20. Domínguez-Berjón MF, Borrell C, Cano-Serral G, Esnaola S, review for the US Preventive Services Task Force. J Am Med Nolasco A, Pasarín MI, et al. Construcción de un índice de pri- Assoc. 2016;315:388---406. vación a partir de datos censales en grandes ciudades españolas: 5. Littlewood E, Ali S, Ansell P, Dyson L, Gascoyne S, Hewitt C, (Proyecto MEDEA). Gac Sanit. 2008;22:179---87. Available from: et al. BaBY PaNDA study team Identification of depression in http://www.elsevier.es/es-revista-gaceta-sanitaria-138- women during pregnancy and the early postnatal period using articulo-construccion-un-indice-privacion-partir-S02139111087 the Whooley questions and the Edinburgh Postnatal Depresssion 12329?referer=buscador [accessed 13.04.17]. Scale: protocol for the Born and Bred in Yorkshire: Perinatal 21. WHO Collaborating Centre for Drug Statistics Methodology. Depresssion Diagnostic Accuracy (Baby Panda) study. Br Med J ‘‘ATC/DDD Index’’. Norwegian Institute of Public Health; Open. 2016;6:e011223. 2016. Available from: https://www.whocc.no/atc ddd index/ 6. Ibañez G, Charles MA, Forhan A, Magnin G, Thiebaugeorges O, [accessed 13.04.17]. Kaminski M, et al., EDEN Mother-Child Cohort Study Group. 22. Männistö T, Mendola P, Kiely M, O’Loughlin J, Werder E, Chen Z, Depression and anxiety in women during pregnancy and neona- et al. Maternal psychiatric disorders and risk of pre-term birth. tal outcome: data from the EDEN mother-child cohort. Early Ann Epidemiol. 2016;26:14---20. Hum Dev. 2012;88:643---9. 23. Grigoriadis S, VorderPorten EH, Mamisashvili L, Tomlinson G, 7. Räisänen S, Lehto SM, Nielsen HS, Gissler M, Kramer MR, Dennis CL, Koren G, et al. The impact of maternal depression Heinonen S. Risk factors for and perinatal outcomes of major during pregnancy on perinatal outcomes: a systematic review depression during pregnancy: a population-based analysis dur- and meta-analysis. J Clin Psychiatry. 2013;74:e321---41. ing 2002---2010 in Finland. Br Med J Open. 2014;4:e004883. 24. Grote NK, Bridge JA, Gavin AR, Melville JL, Iyengar S, Katon 8. Mei-Dan E, Ray JG, Vigod SN. Perinatal outcomes among women WJ. A meta-analysis of depression during pregnancy and the with bipolar disorder: a population-based cohort study. Am J risk od pre-term birth, low birth weight and intrauterine growth Obstet Gynecol. 2015;212:367. E1-8. restriction. Arch Gen Psychiatry. 2010;67:1012---24. 9. Liou SR, Wang P, Cheng CY. Effects of prenatal maternal mental 25. World Health Organization. Depression and Other Common distress on birth outcomes. Women Birth. 2016;29:376---80. Mental Disorders Global Health Estimates. Geneva: World 10. Navaratne P, Foo XY, Kumar S. Impact of a high Edimburg Postna- Health Organization; 2017. Licence: CC BY-NC-SA 3.0 IGO. tal Depression Scale score on obstetric and perinatal outcomes. Available from: http://apps.who.int/iris/bitstream/10665/ Sci Rep. 2016;6:33544. 254610/1/WHO-MSD-MER-2017.2-eng.pdf [accessed 13.04.17]. 11. Hansen TN, Hawgood S, Clyman RI, Evans N, Warner B, Bada HS, 26. Lilliecreutz C, Larén J, Sydsjö G, Josefsson A. Effect of maternal et al. In: Rudolph CD, Rudolph AM, Lister GE, First LR, Gershon stress during pregnancy on the risk for pre-term birth. BioMed AA, editors. Disorders specifically related to pre-term birth. Central Preg Child. 2016;16:5. New York: Rudolph’s Pediatrics ed. McGraw Hill Medical; 2011. 27. Huang H, Coleman S, Bridge JA, Yonkers K, Katon W. A meta- p. 233---63. analysis of the relationship between antidepressant use in 12. Guía de Práctica Clínica de atención en el embarazo y el puerpe- pregnancy and the risk of pre-term birth and low birth weight. rio. In: Guías de Práctica Clínica en el SNS. Madrid: Ed. Ministerio Gen Hosp Psychiatry. 2014;36:13---8. de Sanidad, Servicios Sociales e Igualdad; 2014. 28. Eke AC, Saccone G, Berghella V. Selective serotonin reuptake 13. Generalitat de Catalunya. Departament de Salut. Protocol inhibitor (SSRI) use during pregnancy and risk of pre-term birth: de seguiment de l’embaraç a Catalunya, 2a ed. Revisada. a systematic review and meta-analysis. Br J Obst Gynaecol. Barcelona: Direcció General de Salut Pública; 2005. Avail- 2016;123:1900---7. able from: http://salutweb.gencat.cat/web/.content/home/ 29. Viktorin A, Lichtenstein P, Lundholm C, Almqvist C, D’Onofrio ambits tematics/linies dactuacio/model assistencial/ordenacio BM, Larsson H. Selective serotonin re-uptake inhibitor use cartera i serveis sanitaris/pla estrategic dordenacio materno during pregnancy: association with offspring birth size and ges- infantil i atencio salut sexual i reproductiva/material de tational age. Int J Epidemiol. 2016;45:170---7. suport/documents/protsegui2006.pdf [accessed 13.04.17]. 30. Sujan AC, Rickert ME, Öberg AS, Quinn PD, Hernández-Diaz 14. Galbally M, Snellen M, Power J. Antipsychotic drugs in preg- S, Almqvist C, et al. Associations of maternal antidepres- nancy: a review of their maternal and fetal effects. Ther Adv sant use during the first trimester of pregnancy with pre-term Drug Saf. 2014;5:100---9. birth, small for gestational age, autism spectrum disorder and 15. Goel N, Beasley D, Rajkumar V, Banerjee S. Perinatal out- attention-deficit/hyperactivity disorder in offspring. J Am Med come of illicit substance use in pregnancy-comparative and Assoc. 2017;317:1553---62. contemporary socio-clinical profile in the UK. Eur J Pediatr. 31. Howard LM, Molyneaux E, Dennis CL, Rochat T, Stein A, Mil- 2011;170:199---205. grom J. Non-psychotic mental disorders in the perinatal period. 16. Wadhwa PD, Entringer S, Buss C, Lu MC. The contribution of Lancet. 2014;384:1775---88. maternal stress to pre-term birth: issues and considerations. 32. Editorial. Screening for perinatal depression: a missed opportu- Clin Perinatol. 2011;38:351---84. nity. Lancet. 2016;387:505.