[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure.
Association Between Gestational Age and Obsessive-Compulsive Disorder (OCD) and Between Birth Weight and OCD
Association Between Gestational Age and Obsessive-Compulsive Disorder (OCD) and Between Birth Weight and OCD

Analysis of data as ordinal and continuous variables in fully adjusted, baseline, population-wide estimate and sibling comparison models for gestational age (reference group, 37-41 weeks) (A) and birth weight (reference group, 3501-4500 g) (B) in determination of the risk for OCD in offspring born in Sweden between January 1, 1973, and December 31, 1996. Error bars indicate 95% CI. The y-axis uses a log scale.

Table 1.  
Descriptive Characteristics of Study Population
Descriptive Characteristics of Study Population
Table 2.  
Association Between Perinatal Events and OCD
Association Between Perinatal Events and OCD
Table 3.  
Association Between Perinatal Events and OCD With Common Comorbid Disorders Excludeda
Association Between Perinatal Events and OCD With Common Comorbid Disorders Excludeda
Table 4.  
Post Hoc Analyses
Post Hoc Analyses
1.
Mataix-Cols  D, Boman  M, Monzani  B,  et al.  Population-based, multigenerational family clustering study of obsessive-compulsive disorder.  JAMA Psychiatry. 2013;70(7):709-717.PubMedGoogle ScholarCrossref
2.
Taylor  S.  Molecular genetics of obsessive-compulsive disorder: a comprehensive meta-analysis of genetic association studies.  Mol Psychiatry. 2013;18(7):799-805.PubMedGoogle ScholarCrossref
3.
Brander  G, Pérez-Vigil  A, Larsson  H, Mataix-Cols  D.  Systematic review of environmental risk factors for obsessive-compulsive disorder: a proposed roadmap from association to causation.  Neurosci Biobehav Rev. 2016;65:36-62.PubMedGoogle ScholarCrossref
4.
Uher  R.  Gene-environment interactions in common mental disorders: an update and strategy for a genome-wide search.  Soc Psychiatry Psychiatr Epidemiol. 2014;49(1):3-14.PubMedGoogle ScholarCrossref
5.
Schmitt  A, Malchow  B, Hasan  A, Falkai  P.  The impact of environmental factors in severe psychiatric disorders.  Front Neurosci. 2014;8:19.PubMedGoogle ScholarCrossref
6.
Chudal  R, Sourander  A, Polo-Kantola  P,  et al.  Perinatal factors and the risk of bipolar disorder in Finland.  J Affect Disord. 2014;155:75-80.PubMedGoogle ScholarCrossref
7.
D’Onofrio  BM, Class  QA, Rickert  ME, Larsson  H, Långström  N, Lichtenstein  P.  Preterm birth and mortality and morbidity: a population-based quasi-experimental study.  JAMA Psychiatry. 2013;70(11):1231-1240.PubMedGoogle ScholarCrossref
8.
Perrone-McGovern  K, Simon-Dack  S, Niccolai  L.  Prenatal and perinatal factors related to autism, IQ, and adaptive functioning.  J Genet Psychol. 2015;176(1-2):1-10.PubMedGoogle ScholarCrossref
9.
Class  QA, Rickert  ME, Larsson  H, Lichtenstein  P, D’Onofrio  BM.  Fetal growth and psychiatric and socioeconomic problems: population-based sibling comparison.  Br J Psychiatry. 2014;205(5):355-361.PubMedGoogle ScholarCrossref
10.
Hultman  CM, Torrång  A, Tuvblad  C, Cnattingius  S, Larsson  JO, Lichtenstein  P.  Birth weight and attention-deficit/hyperactivity symptoms in childhood and early adolescence: a prospective Swedish twin study.  J Am Acad Child Adolesc Psychiatry. 2007;46(3):370-377.PubMedGoogle ScholarCrossref
11.
Capstick  N, Seldrup  J.  Obsessional states: a study in the relationship between abnormalities occurring at the time of birth and the subsequent development of obsessional symptoms.  Acta Psychiatr Scand. 1977;56(5):427-431.PubMedGoogle ScholarCrossref
12.
Geller  DA, Wieland  N, Carey  K,  et al.  Perinatal factors affecting expression of obsessive compulsive disorder in children and adolescents.  J Child Adolesc Psychopharmacol. 2008;18(4):373-379.PubMedGoogle Scholar
13.
Lensi  P, Cassano  GB, Correddu  G, Ravagli  S, Kunovac  JL, Akiskal  HS.  Obsessive-compulsive disorder: familial-developmental history, symptomatology, comorbidity and course with special reference to gender-related differences.  Br J Psychiatry. 1996;169(1):101-107.PubMedGoogle ScholarCrossref
14.
Vasconcelos  MS, Sampaio  AS, Hounie  AG,  et al.  Prenatal, perinatal, and postnatal risk factors in obsessive-compulsive disorder.  Biol Psychiatry. 2007;61(3):301-307.PubMedGoogle ScholarCrossref
15.
Sampaio  AS, Miguel  EC, Borcato  S,  et al.  Perinatal risk factors and obsessive-compulsive spectrum disorders in patients with rheumatic fever.  Gen Hosp Psychiatry. 2009;31(3):288-291.PubMedGoogle ScholarCrossref
16.
Cath  DC, van Grootheest  DS, Willemsen  G, van Oppen  P, Boomsma  DI.  Environmental factors in obsessive-compulsive behavior: evidence from discordant and concordant monozygotic twins.  Behav Genet. 2008;38(2):108-120.PubMedGoogle ScholarCrossref
17.
Douglass  HM, Moffitt  TE, Dar  R, McGee  R, Silva  P.  Obsessive-compulsive disorder in a birth cohort of 18-year-olds: prevalence and predictors.  J Am Acad Child Adolesc Psychiatry. 1995;34(11):1424-1431.PubMedGoogle ScholarCrossref
18.
Rutter  M.  Proceeding from observed correlation to causal inference: the use of natural experiments.  Perspect Psychol Sci. 2007;2(4):377-395.PubMedGoogle ScholarCrossref
19.
Thapar  A, Rutter  M.  Do prenatal risk factors cause psychiatric disorder? be wary of causal claims.  Br J Psychiatry. 2009;195(2):100-101.PubMedGoogle ScholarCrossref
20.
D’Onofrio  BM, Lahey  BB, Turkheimer  E, Lichtenstein  P.  Critical need for family-based, quasi-experimental designs in integrating genetic and social science research.  Am J Public Health. 2013;103(suppl 1):S46-S55.PubMedGoogle ScholarCrossref
21.
World Health Organization.  The ICD-10 Classification of Mental and Behavioral Disorders: Diagnostic Criteria for Research. Geneva, Switzerland: World Health Organization; 1993.
22.
Ludvigsson  JF, Otterblad-Olausson  P, Pettersson  BU, Ekbom  A.  The Swedish personal identity number: possibilities and pitfalls in healthcare and medical research.  Eur J Epidemiol. 2009;24(11):659-667.PubMedGoogle ScholarCrossref
23.
Centre for Epidemiology. The Swedish Medical Birth Register: summary of content and quality: 2003 http://www.socialstyrelsen.se/Lists/Artikelkatalog/Attachments/10655/2003-112-3_20031123.pdf. Accessed April 29, 2016.
24.
Ekbom  A.  The Swedish Multi-generation Register.  Methods Mol Biol. 2011;675:215-220.PubMedGoogle Scholar
25.
Ludvigsson  JF, Andersson  E, Ekbom  A,  et al.  External review and validation of the Swedish National Inpatient Register.  BMC Public Health. 2011;11:450.PubMedGoogle ScholarCrossref
26.
Ludvigsson  JF, Almqvist  C, Bonamy  AK,  et al.  Registers of the Swedish total population and their use in medical research.  Eur J Epidemiol. 2016;31(2):125-136.PubMedGoogle ScholarCrossref
27.
Socialstyrelsen. Cause of death: 2013. http://www.socialstyrelsen.se/statistics/statisticaldatabase/help/causeofdeath. Accessed April 29, 2016.
28.
Marsál  K, Persson  PH, Larsen  T, Lilja  H, Selbing  A, Sultan  B.  Intrauterine growth curves based on ultrasonically estimated foetal weights.  Acta Paediatr. 1996;85(7):843-848.PubMedGoogle ScholarCrossref
29.
Apgar  V.  A proposal for a new method of evaluation of the newborn infant: originally published in July 1953, volume 32, pages 250-259.  Anesth Analg. 2015;120(5):1056-1059.PubMedGoogle ScholarCrossref
30.
American Academy of Pediatrics; Committee on Fetus and Newborn; American College of Obstetricians and Gynecologists; Committee on Obstetric Practice.  The Apgar score.  Adv Neonatal Care. 2006;6(4):220-223.PubMedGoogle ScholarCrossref
31.
World Health Organization.  WHO Child Growth Standards: Head Circumference-for-Age, Arm Circumference-for-Age, Triceps Skinfold-for-Age and Subscapular Skinfold-for-Age: Methods and Development. Geneva, Switzerland: World Health Organization; 2007.
32.
Rück  C, Larsson  KJ, Lind  K,  et al.  Validity and reliability of chronic tic disorder and obsessive-compulsive disorder diagnoses in the Swedish National Patient Register.  BMJ Open. 2015;5(6):e007520.PubMedGoogle ScholarCrossref
33.
Allison  PD.  Fixed Effects Regression Models. Thousand Oaks, CA: Sage Publications; 2009.
34.
Oken  E, Kleinman  KP, Rich-Edwards  J, Gillman  MW.  A nearly continuous measure of birth weight for gestational age using a United States national reference.  BMC Pediatr. 2003;3:6.PubMedGoogle ScholarCrossref
35.
Swanson  JD, Wadhwa  PM.  Developmental origins of child mental health disorders.  J Child Psychol Psychiatry. 2008;49(10):1009-1019.PubMedGoogle ScholarCrossref
36.
Barker  DJ.  In utero programming of chronic disease.  Clin Sci (Lond). 1998;95(2):115-128.PubMedGoogle ScholarCrossref
37.
Rees  S, Inder  T.  Fetal and neonatal origins of altered brain development.  Early Hum Dev. 2005;81(9):753-761.PubMedGoogle ScholarCrossref
38.
Huizink  AC, Mulder  EJ.  Maternal smoking, drinking or cannabis use during pregnancy and neurobehavioral and cognitive functioning in human offspring.  Neurosci Biobehav Rev. 2006;30(1):24-41.PubMedGoogle ScholarCrossref
39.
Schlotz  W, Phillips  DI.  Fetal origins of mental health: evidence and mechanisms.  Brain Behav Immun. 2009;23(7):905-916.PubMedGoogle ScholarCrossref
40.
Walhovd  KB, Fjell  AM, Brown  TT,  et al; Pediatric Imaging, Neurocognition, and Genetics Study.  Long-term influence of normal variation in neonatal characteristics on human brain development.  Proc Natl Acad Sci U S A. 2012;109(49):20089-20094.PubMedGoogle ScholarCrossref
41.
Losh  M, Esserman  D, Anckarsäter  H, Sullivan  PF, Lichtenstein  P.  Lower birth weight indicates higher risk of autistic traits in discordant twin pairs.  Psychol Med. 2012;42(5):1091-1102.PubMedGoogle ScholarCrossref
42.
Pinto  R, Monzani  B, Leckman  JF,  et al.  Understanding the covariation of tics, attention-deficit/hyperactivity, and obsessive-compulsive symptoms: a population-based adult twin study [published online February 27, 2016].  Am J Med Genet B Neuropsychiatr Genet. doi:10.1002/ajmg.b.32436PubMedGoogle Scholar
43.
Haworth  CM, Plomin  R.  Quantitative genetics in the era of molecular genetics: learning abilities and disabilities as an example.  J Am Acad Child Adolesc Psychiatry. 2010;49(8):783-793.PubMedGoogle ScholarCrossref
44.
Pettersson  E, Larsson  H, Lichtenstein  P.  Common psychiatric disorders share the same genetic origin: a multivariate sibling study of the Swedish population.  Mol Psychiatry. 2016;21(5):717-721.PubMedGoogle ScholarCrossref
45.
Kuja-Halkola  R, D’Onofrio  BM, Larsson  H, Lichtenstein  P.  Maternal smoking during pregnancy and adverse outcomes in offspring: genetic and environmental sources of covariance.  Behav Genet. 2014;44(5):456-467.PubMedGoogle ScholarCrossref
46.
Frisell  T, Öberg  S, Kuja-Halkola  R, Sjölander  A.  Sibling comparison designs: bias from non-shared confounders and measurement error.  Epidemiology. 2012;23(5):713-720.PubMedGoogle ScholarCrossref
47.
Lahey  BB, D’Onofrio  BM.  All in the family: comparing siblings to test causal hypotheses regarding environmental influences on behavior.  Curr Dir Psychol Sci. 2010;19(5):319-323.PubMedGoogle ScholarCrossref
Original Investigation
November 2016

Association of Perinatal Risk Factors With Obsessive-Compulsive DisorderA Population-Based Birth Cohort, Sibling Control Study

Author Affiliations
  • 1Centre for Psychiatry Research, Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden
  • 2Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden
  • 3Stockholm Health Care Services, Stockholm County Council, Stockholm, Sweden
  • 4Astrid Lindgren Children’s Hospital, Karolinska University Hospital, Stockholm, Sweden
  • 5Department of Psychological and Brain Sciences, Indiana University, Bloomington
  • 6Department of Medical Sciences, Örebro University, Örebro, Sweden
 

Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.

JAMA Psychiatry. 2016;73(11):1135-1144. doi:10.1001/jamapsychiatry.2016.2095
Key Points

Question  Do adverse perinatal events increase the risk for obsessive-compulsive disorder (OCD)?

Findings  In a population-based birth cohort study of 2.4 million children in Sweden, maternal smoking during pregnancy, breech presentation, delivery by cesarean section, preterm birth, low birth weight, being large for gestational age, and Apgar distress scores were associated with a higher risk of developing OCD, independently of shared familial confounders. A dose-response association was identified for a number of perinatal events, with a higher risk for OCD noted in individuals with a greater number of events.

Meaning  The findings of this study are important for the understanding of the cause of OCD and will inform future studies on gene by environment interaction and epigenetics.

Abstract

Importance  Perinatal complications may increase the risk of obsessive-compulsive disorder (OCD). Previous reports were based on small, retrospective, specialist clinic–based studies that were unable to rigorously control for unmeasured environmental and genetic confounding.

Objective  To prospectively investigate a wide range of potential perinatal risk factors for OCD, controlling for unmeasured factors shared between siblings in the analyses.

Design, Setting, and Participants  This population-based birth cohort study included all 2 421 284 children from singleton births in Sweden from January 1, 1973, to December 31, 1996, who were followed up through December 31, 2013. From the 1 403 651 families in the cohort, differentially exposed siblings from the 743 885 families with siblings were evaluated; of these, 11 592 families included clusters of full siblings that were discordant for OCD. Analysis of the data was conducted from January, 26, 2015, to September, 5, 2016.

Exposures  Perinatal data were collected from the Swedish Medical Birth Register and included maternal smoking during pregnancy, labor presentation, obstetric delivery, gestational age (for preterm birth), birth weight, birth weight in relation to gestational age, 5-minute Apgar score, and head circumference.

Main Outcomes and Measures  Previously validated OCD codes (International Statistical Classification of Diseases and Health Related Problems, Tenth Revision, code F42) in the Swedish National Patient Register.

Results  Of 2 421 284 individuals included in the cohort, 17 305 persons were diagnosed with OCD. Of these, 7111 were men (41.1%). The mean (SD) age of individuals at first diagnosis of OCD was 23.4 (6.5) years. An increased risk for OCD remained after controlling for shared familial confounders and measured covariates (including sex, year of birth, maternal and paternal age at birth, and parity), for smoking 10 or more cigarettes per day during pregnancy (hazard ratio [HR], 1.27; 95% CI, 1.02-1.58), breech presentation (HR, 1.35; 95% CI, 1.06-1.71), delivery by cesarean section (HR, 1.17; 95% CI, 1.01-1.34), preterm birth (HR, 1.24; 95% CI, 1.07-1.43), birth weight 1501 to 2500 g (HR, 1.30; 95% CI, 1.05-1.62) and 2501 to 3500 g (HR, 1.08; 95% CI, 1.01-1.16), being large for gestational age (HR, 1.23; 95% CI, 1.05-1.45), and Apgar distress scores at 5 minutes (HR, 1.50; 95% CI, 1.07-2.09). Gestational age and birth weight followed inverse dose-response associations, whereby an increasingly higher risk for OCD was noted in children with a shorter gestational age and lower birth weight. We also observed a dose-response association between the number of perinatal events and increased OCD risk, with HRs ranging from 1.11 (95% CI, 1.07-1.15) for 1 event to 1.51 (95% CI, 1.18-1.94) for 5 or more events.

Conclusions and Relevance  A range of perinatal risk factors is associated with a higher risk for OCD independent of shared familial confounders, suggesting that perinatal risk factors may be in the causal pathway to OCD.

Introduction

With much of the current focus on the discovery of genetic factors conferring risk for development of obsessive-compulsive disorder (OCD), environmental risk factors have received relatively little attention. Because environmental factors are at least as important as genetic factors in the development of OCD,1,2 the identification of such risk factors should be viewed as a research priority.

At the present time, to our knowledge, there are no robust environmental risk factors known to play a causal role in the development of OCD. Previous OCD studies3 have had several important methodologic weaknesses, including the predominant use of retrospective designs and recruitment of patients from specialist clinics, limiting the confidence in the data and generalizability of the findings. The success of gene by environment interaction studies depends on high-quality epidemiologic methods, including a detailed objective assessment of environmental exposures.4

Complications in the perinatal period, including delivery by cesarean section, delivery using vacuum extraction, preterm birth, and low birth weight, have been associated with a range of psychiatric disorders, such as schizophrenia,5 bipolar disorder,6,7 autism spectrum disorder,8,9 and attention-deficit/hyperactivity disorder (ADHD).9,10 Few studies,3,11-17 most of which were retrospective, indicate that perinatal complications may also play a role in OCD, but the methodologic weaknesses of those studies preclude firm conclusions.

Most research into risk factors for mental disorders has primarily relied on adjusting for statistical covariates to account for confounding but has generally not taken familial effects into account.18 Perinatal factors may exhibit a spurious association with a disorder, not representing a causal link but instead being explained by unmeasured confounders, such as parental mental health, social adversity, or maternally transmitted inherited factors.19 Family-based study designs provide a better control for such unmeasured environmental and genetic factors; comparison of full siblings raised in the same family but discordant for the exposure automatically excludes confounding of all shared environmental and a substantial proportion of genetic factors.20

In this longitudinal, population-based cohort study, we aimed to explore the potential causal link between OCD and a range of perinatal factors. Unlike previous studies,3 which tended to explore a limited number of risk factors at a time, we investigated a broad range of exposures and controlled for measured covariates, specifically, year of birth, sex, maternal and paternal age at birth, and parity. A sibling comparison design was used to further control for shared familial confounders.

Methods
Study Population

The study cohort, consisting of all 2 421 284 live singleton births in Sweden from January 1, 1973, through December 31, 1996, was followed up until first diagnosis of OCD (available from 1997, when the International Statistical Classification of Diseases and Health Related Problems, Tenth Revision was introduced21), migration, death, or end of follow-up (December 31, 2013), whichever came first. The data were obtained by linking individuals through their unique personal identification numbers22 from the following population-based registers: (1) the Swedish Medical Birth Register,23 which includes data on more than 99% of all pregnancies and deliveries in Sweden since 1973; (2) the Swedish Multi-generation Register,24 with information about kinship going back to 1932, containing information on 100% of mothers and 98% of fathers of index persons born after 1961; (3) the Swedish National Patient Register,25 which covers all inpatient hospital admissions since 1969 and outpatient care since 2001; (4) the Migration Register,26 providing information about migration in and out of Sweden; and (5) the Cause of Death Register,27 with information on dates and causes of all deaths since 1961. Information from the Cause of Death Register and the Migration Register was used to calculate censoring time. For the sibling comparison analysis, we identified a subsample of 743 885 families with at least 2 full siblings (ie, siblings sharing the same biological mother and father) during the same time period from the Swedish Multi-generation Register.

Ethics approval and waiver of informed consent was obtained from the Regional Ethical Review Board in Stockholm. The requirement for informed consent was waived because the study was register based and data on the included individuals were deidentified.

Exposures

Information about all prenatal exposures was retrieved from the Swedish Medical Birth Register. Unless otherwise specified, information was available from January 1, 1973, to December 31, 2013.

Maternal Smoking During Pregnancy

Information on maternal smoking during pregnancy collected at the first antenatal visit is available in the Swedish Medical Birth Register from 1982, marking the start of cohort inclusion for this exposure (n = 1 547 271). The information is categorized as no daily smoking, 1 to 9 cigarettes per day, and 10 or more cigarettes per day.

Labor Presentation and Obstetric Delivery

Labor presentation was divided into normal presentation, breech presentation, and other malpresentations. Obstetric delivery was divided into 3 hierarchical categories: cesarean section, assisted vaginal delivery (use of forceps or vacuum extraction), and unassisted vaginal delivery.

Gestational Age and Birth Weight

Gestational age and birth weight were analyzed in 2 ways. In the first method, data were evaluated as continuous variables with linear and quadratic terms included. Gestational age was distributed as every week and birth weight was distributed as every 250 g. In the second analysis, gestational age was categorized as very preterm birth (gestational age <32 weeks), preterm birth (32-36 weeks), term birth (37-41 weeks), and postterm birth (≥42 weeks), and birth weight was categorized as 1500 g or less (very low birth weight), 1501 to 2500 g (low birth weight), 2501 to 3500 g, 3501 to 4500 g (normal birth weight [reference category]), and more than 4500 g (high birth weight). Small for gestational age and large for gestational age were defined as a birth weight of more than 2 SDs below and above the mean weight for gestational age, respectively, according to the Scandinavian fetal growth curve adjusted for sex.28

Apgar Score

The Apgar score29 of the index neonate at 5 minutes after delivery was categorized as normal (a score of ≥7), distress (4-6), or near death (≤3) in accordance with neonatal practice.30 The Apgar score is a tool for evaluating heart rate, respiratory effort, reflex irritability, muscle tone, and color after delivery, considered to be 5 useful indicators that could be determined easily without interfering with the care of the infant.

Head Circumference

Small head circumference was defined as head circumference below the 10th percentile for each gestational week, and large head circumference was that above the 90th percentile for each gestational week. These categories were established according to the World Health Organization standards.31

Outcome and Covariates

The first instance of a recorded OCD diagnosis in the National Patient Register constituted the outcome. The OCD diagnosis was defined as code F42 according to ICD-10, which was introduced in Sweden in 1997. The OCD codes in the National Patient Register are reliable and valid.32 Data on all potential measured confounders were collected from the Swedish Medical Birth Register (year of birth, sex, parity, and maternal age at childbirth) and the Swedish Multi-generation Register (paternal age at childbirth).

Statistical Analysis

Differences in sociodemographic and clinical variables between OCD cases and non-OCD cases were determined with χ2 or 2-tailed t tests for independent samples. We performed Cox proportional hazards regression analysis to estimate hazard ratios (HRs) and 95% CIs of the association between perinatal factors and OCD. Three different Cox proportional hazards regression models were fitted for all exposure variables: (1) crude associations with OCD were modeled separately for each exposure variable, (2) analysis was adjusted for sex and year of birth, and (3) all measured confounders, as listed above, were adjusted for in the fully adjusted model.

For continuous variables (ie, gestational age and birth weight), we fitted both a linear and quadratic representation. We used the Akaike information criterion to determine which model (ie, linear or linear + quadratic) best fit the data.

The analyses were replicated in a fixed-effects model of the subsample of clusters of all full siblings by using stratified Cox proportional hazards regression models. By design, these models adjust for shared familial confounders33 and, in particular, for genetic factors and unmeasured shared confounders such as socioeconomic status or stable parental factors. Furthermore, we adjusted for all measured confounders, which typically vary between siblings.

To confirm that the associations were not entirely explained by comorbid conditions, we performed sensitivity analyses in subgroups in which all individuals with comorbid conditions were excluded from analysis. These conditions were organized in 3 clusters: organic disorders (ie, organic brain disorder and epilepsy), psychotic disorders (ie, schizophrenia and bipolar disorder), and neuropsychiatric disorders (ie, ADHD, pervasive developmental disorders, and mental retardation). All disorders were defined as at least 1 registered diagnosis in the National Patient Register according to their ICD-10 code (eTable 1 in the Supplement). These models adjusted for all measured confounders. In addition, we performed a sensitivity analysis on a subsample born in 1987 or later and applied the same models, including the sibling comparison, to examine whether the extended follow-up time for the oldest subsample of the cohort until the introduction of ICD-10 in 1997 was a source of bias.

A post hoc Cox proportional hazards regression analysis was used to determine the association between the number of adverse perinatal events and the risk for OCD. All analyses were conducted using SAS, version 9.4 (SAS Institute Inc).

Results
Descriptive Statistics

Descriptive characteristics of the study population are presented in Table 1. In total, 2 421 284 individuals were included in the cohort; of these, 17 305 were diagnosed with OCD during the study period, resulting in a Kaplan-Meier estimated prevalence of 1.32% at age 40 years (eFigure in the Supplement). The mean (SD) age at first diagnosis of OCD was 23.4 (6.5) years. The individuals with OCD differed significantly from those without OCD in several aspects. For instance, the proportion of women was significantly higher compared with men among persons with OCD (58.9% vs 41.1%; P < .001). Those with OCD also had more comorbid disorders than did those without OCD (37.9% vs 5%; P < .001). Of the 743 885 families with at least 2 children, 11 592 (15.6%) included full siblings discordant for OCD.

Maternal Smoking During Pregnancy

Maternal smoking of 10 cigarettes or more per day during pregnancy was associated with an increased risk of offspring OCD both in the fully adjusted model (HR, 1.20; 95% CI, 1.13-1.28) and in the sibling comparison model (HR, 1.27; 95% CI, 1.02-1.58) (Table 2) compared with offspring of mothers who did not smoke during pregnancy. Maternal smoking of 1 to 9 cigarettes per day during pregnancy exhibited only a small increased risk for OCD in the offspring compared with nonsmoking mothers (HR, 1.06; 95% CI, 1-1.12). The risk remained but with lower precision in the sibling comparison (HR, 1.06; 95% CI, 0.89-1.26).

Labor Presentation

Breech presentation was associated with an increased risk for OCD (fully adjusted model: HR, 1.26; 95% CI, 1.15-1.39; sibling comparison: HR, 1.35; 95% CI, 1.06-1.71) compared with normal presentation (Table 2). A similar association was not found for other malpresentations, neither in the fully adjusted model nor the sibling comparison model (Table 2).

Obstetric Delivery

There was a slightly increased risk of OCD among individuals delivered by cesarean section (HR, 1.09; 95% CI, 1.04-1.15) and assisted vaginal delivery (HR, 1.12; 95% CI, 1.05-1.19) compared with unassisted vaginal delivery in the fully adjusted model (Table 2). In the sibling comparison models, these estimates remained similar in magnitude but had lower precision; the association with cesarean section remained statistically significant (HR, 1.17; 95% CI, 1.01-1.34), whereas the association with assisted vaginal delivery did not (HR, 1.07; 95% CI, 0.94-1.22) (Table 2).

Gestational Age

There was an inverse association between gestational age and OCD that remained when we adjusted for potential confounders (Table 2). Compared with those with term birth, the risk for OCD in individuals with very preterm birth (<32 weeks) was higher (HR, 1.61; 95% CI, 1.35-1.91) than in individuals with preterm birth (32-36 weeks: HR, 1.20; 95% CI, 1.12-1.28), indicating a dose-response association (Figure, A). The continuous representation of gestational age displayed a pattern similar to that of the ordinal representation (Figure, A). The results of the sibling comparison yielded similar estimates but with lower precision (very preterm birth: HR, 1.46; 95% CI, 0.97-2.19; and preterm birth: HR, 1.24; 95% CI, 1.07-1.43) (Table 2 and Figure, A).

Birth Weight

The ordinal representation of birth weight showed an inverse association with OCD that remained when we adjusted for all measured confounders as well as gestational age (known to correlate with birth weight34) (Table 2). Both low and high birth weight were associated with a slightly increased risk for OCD (eg, low birth weight [1501-2500 g]: HR, 1.10; 95% CI, 1-1.21; and high birth weight [>4500 g]: HR, 1.17; 95% CI, 1.07-1.27). The continuous representation mirrored these results (Figure, B).

The results from the sibling comparison models yielded higher estimates for low birth weights but with lower precision and followed a clearer dose-response association, with the highest risk for very low birth weight (≤1500 g) (HR, 1.72; 95% CI, 0.94-3.14) (Table 2 and Figure, B). The estimate for high birth weight (>4500 g) remained similar in magnitude (HR, 1.14; 95% CI, 0.96-1.35). Analyses of both small- and large-for-gestational age individuals followed the same pattern in both the magnitude and precision of the estimates (Table 2).

Apgar Score

At 5 minutes after delivery, infants with Apgar scores at distress (scores 4-6) or near-death (≤3) levels had an increased risk of OCD in both the fully adjusted model (HR, 1.28; 95% CI, 1.09-1.51; and HR, 1.4; 95% CI, 1.03-1.90, respectively) and the sibling comparison model (HR, 1.50; 95% CI, 1.07-2.09; and HR, 1.68; 95% CI, 0.79-3.56, respectively) (Table 2) compared with infants with normal Apgar scores. However, the precision was low in the sibling comparison, especially for near-death scores, which did not remain statistically significant.

Head Circumference

The small association between small head circumference and OCD in the fully adjusted model (HR, 1.07; 95% CI, 1-1.14) did not remain in the sibling comparison (HR, 0.96; 95% CI, 0.84-1.09) (Table 2). No association between large head circumference and OCD was observed.

Sensitivity Analyses

The pattern of results remained largely unchanged when individuals with comorbid disorders were excluded from the analyses (Table 3). Evaluation of the subsample of individuals born in 1987 or later (without an extended follow-up time until the introduction of ICD-10) revealed a similar pattern but with lower precision (eTable 2 in the Supplement); one exception was the association for breech presentation, which did not remain in the sibling comparison.

Post Hoc Analyses: Number of Perinatal Events

A dose-response association between the number of adverse perinatal events and increased risk for OCD was observed. Hazard ratios ranged from 1.11 (95% CI, 1.07-1.15) for 1 event to 1.51 (95% CI, 1.18-1.94) for 5 or more events (Table 4).

Discussion

In this large study of the entire Swedish population, we found that several perinatal factors were associated with a higher risk of developing OCD, confirming and extending the results of the few previous small and often retrospective studies.3,11-17 These associations largely remained when genetic and environmental factors shared by siblings were taken into account, confounders were strictly controlled, and relevant comorbidities were excluded. The associations for preterm birth and low birth weight were further supported by dose-response associations consistent with causal inference. A dose-response association was also identified for the number of perinatal events, whereby the higher the number of events, the greater the risk of OCD.

The specific mechanism linking OCD and maternal smoking, complicated birth, preterm birth, fetal growth, and distress at birth remains to be identified. However, these findings are in line with the fetal programming hypothesis, positing that adaptation to the fetal environment may lead to adverse effects in subsequent life.35,36 An adverse fetal environment or insult (eg, hypoxia-ischemia, white matter injury, reduced blood flow, malnutrition, and differences in development of the serotonergic system) has been observed37-39 to affect brain development. It also has been shown39,40 that variations in fetal growth are associated with brain development well into childhood and adolescence. Research into differences in brain structure and function between people with and without the associations identified in the present study may reveal the mediating biological pathways.

Rather than being unique to OCD, many of these perinatal risk factors may be shared across a range of neuropsychiatric disorders. For instance, using similar family designs, studies have shown associations between low birth weight and ADHD and autism9,10,41 as well as between gestational age and autism, ADHD, and psychotic or bipolar disorders.7 These findings contradict the widely held notion that, although mental disorders share genetic risk factors, the contribution of environmental risk factors is largely disorder specific.42-44 One exception to this lack of specificity may be our finding of an increased risk of OCD in children exposed to maternal smoking during pregnancy. This finding was unexpected since other family-based studies20,45 have suggested that the causal inference between maternal smoking during pregnancy and a range of adverse outcomes in offspring (eg, ADHD, criminality, academic achievement, drug use, adolescent antisocial behavior, adolescent psychological functioning, suicidal behavior, childhood conduct problems, and intellectual abilities) has been overstated. The results of the present study indicate a relatively small, yet robust association between maternal smoking during pregnancy and OCD, even after controlling for familial confounding. If replicated, these findings may have important implications for future research, for example, using relevant animal models.

To our knowledge, this is the first study examining a broad range of perinatal risk factors for OCD using a large, population-based cohort with prospectively collected data during pregnancy and at the time of birth. By comparing clusters of full siblings discordant for OCD, we could control for many (unmeasured) shared familial confounders (genetic and environmental). The validity and reliability of the Swedish ICD-10 codes for OCD have been well established.32

Some limitations of the study need to be considered. The cohort is weighted toward more severe cases and does not represent the totality of all patients with OCD in Sweden. There are missing cases because many individuals do not seek help, coverage of the Swedish National Patient Register between 1997 and 2001 is incomplete, and patients diagnosed with OCD by general practitioners and other nonspecialists are not included. Sibling comparisons can help in inferring, but not proving, causality and have lower statistical power compared with population-based estimates.33 Furthermore, sibling comparison designs are sensitive to random measurement error in the exposure and may be biased owing to variables shared by siblings that are related to the exposure but not the outcome.20,46 The sibling comparison also assumes that there are no carryover effects from one pregnancy to a later pregnancy.46,47 Adjusting for parity addresses this limitation, but just partially. Monozygotic twin comparison studies could cement these findings further, controlling also for potential evocative genetic and more shared environmental factors, but would be limited to exposures that are not shared by monozygotic twins.

Conclusions

In the present study, we found that perinatal factors, especially maternal smoking during pregnancy, breech presentation, cesarean section, preterm birth, low birth weight, being large for gestational age, and Apgar distress scores, were associated with a higher risk of developing OCD independent of shared familial confounders and several other measured covariates. A dose-response association was also identified for the number of perinatal events, with an increasingly higher number of events associated with a greater risk for OCD. The findings are important for the understanding of the cause of OCD and will inform future studies of gene by environment interaction and epigenetics. If the finding is replicated, the association between maternal smoking during pregnancy and OCD may emerge as an interesting disorder-specific risk factor for evaluation in future research.

Back to top
Article Information

Corresponding Author: Gustaf Brander, MSc, Centre for Psychiatry Research, Department of Clinical Neuroscience, Karolinska Institutet, Gävlegatan 22B, 113 30 Stockholm, Sweden (gustaf.brander@ki.se).

Accepted for Publication: July 13, 2016.

Published Online: October 5, 2016. doi:10.1001/jamapsychiatry.2016.2095

Author Contributions: Mr Brander and Dr Rydell had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Brander, Rydell, Lichtenstein, D’Onofrio, Larsson, Mataix-Cols.

Acquisition, analysis, or interpretation of data: Brander, Rydell, Kuja-Halkola, Fernández de la Cruz, Lichtenstein, Serlachius, Almqvist, Mataix-Cols.

Drafting of the manuscript: Brander, Mataix-Cols.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Brander, Rydell, Kuja-Halkola.

Obtaining funding: Lichtenstein, Almqvist, D’Onofrio, Larsson, Mataix-Cols.

Administrative, technical, or material support: Lichtenstein, Serlachius, Almqvist, Mataix-Cols.

Study supervision: Rydell, Serlachius, Rück, Larsson, Mataix-Cols.

Conflict of Interest Disclosures: Dr Larsson has served as a paid speaker for Eli Lilly and Shire and has received a research grant from Shire unrelated to the present study. Dr Lichtenstein has served as a speaker for Medice. No other disclosures were reported.

Funding/Support: Mr Brander is supported by Karolinska Institutet partial funding for new doctoral students for the present study. Dr Rydell has received FORTE grant 2015-00075 from the Swedish Research Council for Health, Working Life, and Welfare. Dr Fernández de la Cruz has received grants from the David and Astrid Hagelén Foundation and FORTE grant 2015-00569 from the Swedish Research Council for Health, Working Life, and Welfare. Dr Almqvist acknowledges financial support through grant 340-2013-5867 from the Swedish Research Council through the Swedish Initiative for Research on Microdata in the Social and Medical Sciences framework. Dr Rück is supported by grant K2013-61P-22168 from the Swedish Research Council.

Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

References
1.
Mataix-Cols  D, Boman  M, Monzani  B,  et al.  Population-based, multigenerational family clustering study of obsessive-compulsive disorder.  JAMA Psychiatry. 2013;70(7):709-717.PubMedGoogle ScholarCrossref
2.
Taylor  S.  Molecular genetics of obsessive-compulsive disorder: a comprehensive meta-analysis of genetic association studies.  Mol Psychiatry. 2013;18(7):799-805.PubMedGoogle ScholarCrossref
3.
Brander  G, Pérez-Vigil  A, Larsson  H, Mataix-Cols  D.  Systematic review of environmental risk factors for obsessive-compulsive disorder: a proposed roadmap from association to causation.  Neurosci Biobehav Rev. 2016;65:36-62.PubMedGoogle ScholarCrossref
4.
Uher  R.  Gene-environment interactions in common mental disorders: an update and strategy for a genome-wide search.  Soc Psychiatry Psychiatr Epidemiol. 2014;49(1):3-14.PubMedGoogle ScholarCrossref
5.
Schmitt  A, Malchow  B, Hasan  A, Falkai  P.  The impact of environmental factors in severe psychiatric disorders.  Front Neurosci. 2014;8:19.PubMedGoogle ScholarCrossref
6.
Chudal  R, Sourander  A, Polo-Kantola  P,  et al.  Perinatal factors and the risk of bipolar disorder in Finland.  J Affect Disord. 2014;155:75-80.PubMedGoogle ScholarCrossref
7.
D’Onofrio  BM, Class  QA, Rickert  ME, Larsson  H, Långström  N, Lichtenstein  P.  Preterm birth and mortality and morbidity: a population-based quasi-experimental study.  JAMA Psychiatry. 2013;70(11):1231-1240.PubMedGoogle ScholarCrossref
8.
Perrone-McGovern  K, Simon-Dack  S, Niccolai  L.  Prenatal and perinatal factors related to autism, IQ, and adaptive functioning.  J Genet Psychol. 2015;176(1-2):1-10.PubMedGoogle ScholarCrossref
9.
Class  QA, Rickert  ME, Larsson  H, Lichtenstein  P, D’Onofrio  BM.  Fetal growth and psychiatric and socioeconomic problems: population-based sibling comparison.  Br J Psychiatry. 2014;205(5):355-361.PubMedGoogle ScholarCrossref
10.
Hultman  CM, Torrång  A, Tuvblad  C, Cnattingius  S, Larsson  JO, Lichtenstein  P.  Birth weight and attention-deficit/hyperactivity symptoms in childhood and early adolescence: a prospective Swedish twin study.  J Am Acad Child Adolesc Psychiatry. 2007;46(3):370-377.PubMedGoogle ScholarCrossref
11.
Capstick  N, Seldrup  J.  Obsessional states: a study in the relationship between abnormalities occurring at the time of birth and the subsequent development of obsessional symptoms.  Acta Psychiatr Scand. 1977;56(5):427-431.PubMedGoogle ScholarCrossref
12.
Geller  DA, Wieland  N, Carey  K,  et al.  Perinatal factors affecting expression of obsessive compulsive disorder in children and adolescents.  J Child Adolesc Psychopharmacol. 2008;18(4):373-379.PubMedGoogle Scholar
13.
Lensi  P, Cassano  GB, Correddu  G, Ravagli  S, Kunovac  JL, Akiskal  HS.  Obsessive-compulsive disorder: familial-developmental history, symptomatology, comorbidity and course with special reference to gender-related differences.  Br J Psychiatry. 1996;169(1):101-107.PubMedGoogle ScholarCrossref
14.
Vasconcelos  MS, Sampaio  AS, Hounie  AG,  et al.  Prenatal, perinatal, and postnatal risk factors in obsessive-compulsive disorder.  Biol Psychiatry. 2007;61(3):301-307.PubMedGoogle ScholarCrossref
15.
Sampaio  AS, Miguel  EC, Borcato  S,  et al.  Perinatal risk factors and obsessive-compulsive spectrum disorders in patients with rheumatic fever.  Gen Hosp Psychiatry. 2009;31(3):288-291.PubMedGoogle ScholarCrossref
16.
Cath  DC, van Grootheest  DS, Willemsen  G, van Oppen  P, Boomsma  DI.  Environmental factors in obsessive-compulsive behavior: evidence from discordant and concordant monozygotic twins.  Behav Genet. 2008;38(2):108-120.PubMedGoogle ScholarCrossref
17.
Douglass  HM, Moffitt  TE, Dar  R, McGee  R, Silva  P.  Obsessive-compulsive disorder in a birth cohort of 18-year-olds: prevalence and predictors.  J Am Acad Child Adolesc Psychiatry. 1995;34(11):1424-1431.PubMedGoogle ScholarCrossref
18.
Rutter  M.  Proceeding from observed correlation to causal inference: the use of natural experiments.  Perspect Psychol Sci. 2007;2(4):377-395.PubMedGoogle ScholarCrossref
19.
Thapar  A, Rutter  M.  Do prenatal risk factors cause psychiatric disorder? be wary of causal claims.  Br J Psychiatry. 2009;195(2):100-101.PubMedGoogle ScholarCrossref
20.
D’Onofrio  BM, Lahey  BB, Turkheimer  E, Lichtenstein  P.  Critical need for family-based, quasi-experimental designs in integrating genetic and social science research.  Am J Public Health. 2013;103(suppl 1):S46-S55.PubMedGoogle ScholarCrossref
21.
World Health Organization.  The ICD-10 Classification of Mental and Behavioral Disorders: Diagnostic Criteria for Research. Geneva, Switzerland: World Health Organization; 1993.
22.
Ludvigsson  JF, Otterblad-Olausson  P, Pettersson  BU, Ekbom  A.  The Swedish personal identity number: possibilities and pitfalls in healthcare and medical research.  Eur J Epidemiol. 2009;24(11):659-667.PubMedGoogle ScholarCrossref
23.
Centre for Epidemiology. The Swedish Medical Birth Register: summary of content and quality: 2003 http://www.socialstyrelsen.se/Lists/Artikelkatalog/Attachments/10655/2003-112-3_20031123.pdf. Accessed April 29, 2016.
24.
Ekbom  A.  The Swedish Multi-generation Register.  Methods Mol Biol. 2011;675:215-220.PubMedGoogle Scholar
25.
Ludvigsson  JF, Andersson  E, Ekbom  A,  et al.  External review and validation of the Swedish National Inpatient Register.  BMC Public Health. 2011;11:450.PubMedGoogle ScholarCrossref
26.
Ludvigsson  JF, Almqvist  C, Bonamy  AK,  et al.  Registers of the Swedish total population and their use in medical research.  Eur J Epidemiol. 2016;31(2):125-136.PubMedGoogle ScholarCrossref
27.
Socialstyrelsen. Cause of death: 2013. http://www.socialstyrelsen.se/statistics/statisticaldatabase/help/causeofdeath. Accessed April 29, 2016.
28.
Marsál  K, Persson  PH, Larsen  T, Lilja  H, Selbing  A, Sultan  B.  Intrauterine growth curves based on ultrasonically estimated foetal weights.  Acta Paediatr. 1996;85(7):843-848.PubMedGoogle ScholarCrossref
29.
Apgar  V.  A proposal for a new method of evaluation of the newborn infant: originally published in July 1953, volume 32, pages 250-259.  Anesth Analg. 2015;120(5):1056-1059.PubMedGoogle ScholarCrossref
30.
American Academy of Pediatrics; Committee on Fetus and Newborn; American College of Obstetricians and Gynecologists; Committee on Obstetric Practice.  The Apgar score.  Adv Neonatal Care. 2006;6(4):220-223.PubMedGoogle ScholarCrossref
31.
World Health Organization.  WHO Child Growth Standards: Head Circumference-for-Age, Arm Circumference-for-Age, Triceps Skinfold-for-Age and Subscapular Skinfold-for-Age: Methods and Development. Geneva, Switzerland: World Health Organization; 2007.
32.
Rück  C, Larsson  KJ, Lind  K,  et al.  Validity and reliability of chronic tic disorder and obsessive-compulsive disorder diagnoses in the Swedish National Patient Register.  BMJ Open. 2015;5(6):e007520.PubMedGoogle ScholarCrossref
33.
Allison  PD.  Fixed Effects Regression Models. Thousand Oaks, CA: Sage Publications; 2009.
34.
Oken  E, Kleinman  KP, Rich-Edwards  J, Gillman  MW.  A nearly continuous measure of birth weight for gestational age using a United States national reference.  BMC Pediatr. 2003;3:6.PubMedGoogle ScholarCrossref
35.
Swanson  JD, Wadhwa  PM.  Developmental origins of child mental health disorders.  J Child Psychol Psychiatry. 2008;49(10):1009-1019.PubMedGoogle ScholarCrossref
36.
Barker  DJ.  In utero programming of chronic disease.  Clin Sci (Lond). 1998;95(2):115-128.PubMedGoogle ScholarCrossref
37.
Rees  S, Inder  T.  Fetal and neonatal origins of altered brain development.  Early Hum Dev. 2005;81(9):753-761.PubMedGoogle ScholarCrossref
38.
Huizink  AC, Mulder  EJ.  Maternal smoking, drinking or cannabis use during pregnancy and neurobehavioral and cognitive functioning in human offspring.  Neurosci Biobehav Rev. 2006;30(1):24-41.PubMedGoogle ScholarCrossref
39.
Schlotz  W, Phillips  DI.  Fetal origins of mental health: evidence and mechanisms.  Brain Behav Immun. 2009;23(7):905-916.PubMedGoogle ScholarCrossref
40.
Walhovd  KB, Fjell  AM, Brown  TT,  et al; Pediatric Imaging, Neurocognition, and Genetics Study.  Long-term influence of normal variation in neonatal characteristics on human brain development.  Proc Natl Acad Sci U S A. 2012;109(49):20089-20094.PubMedGoogle ScholarCrossref
41.
Losh  M, Esserman  D, Anckarsäter  H, Sullivan  PF, Lichtenstein  P.  Lower birth weight indicates higher risk of autistic traits in discordant twin pairs.  Psychol Med. 2012;42(5):1091-1102.PubMedGoogle ScholarCrossref
42.
Pinto  R, Monzani  B, Leckman  JF,  et al.  Understanding the covariation of tics, attention-deficit/hyperactivity, and obsessive-compulsive symptoms: a population-based adult twin study [published online February 27, 2016].  Am J Med Genet B Neuropsychiatr Genet. doi:10.1002/ajmg.b.32436PubMedGoogle Scholar
43.
Haworth  CM, Plomin  R.  Quantitative genetics in the era of molecular genetics: learning abilities and disabilities as an example.  J Am Acad Child Adolesc Psychiatry. 2010;49(8):783-793.PubMedGoogle ScholarCrossref
44.
Pettersson  E, Larsson  H, Lichtenstein  P.  Common psychiatric disorders share the same genetic origin: a multivariate sibling study of the Swedish population.  Mol Psychiatry. 2016;21(5):717-721.PubMedGoogle ScholarCrossref
45.
Kuja-Halkola  R, D’Onofrio  BM, Larsson  H, Lichtenstein  P.  Maternal smoking during pregnancy and adverse outcomes in offspring: genetic and environmental sources of covariance.  Behav Genet. 2014;44(5):456-467.PubMedGoogle ScholarCrossref
46.
Frisell  T, Öberg  S, Kuja-Halkola  R, Sjölander  A.  Sibling comparison designs: bias from non-shared confounders and measurement error.  Epidemiology. 2012;23(5):713-720.PubMedGoogle ScholarCrossref
47.
Lahey  BB, D’Onofrio  BM.  All in the family: comparing siblings to test causal hypotheses regarding environmental influences on behavior.  Curr Dir Psychol Sci. 2010;19(5):319-323.PubMedGoogle ScholarCrossref
×