Raquel Nogueira Avelar E Silva1,2,3, Yongfu Yu2,4, Zeyan Liew5,6, Anne Vested1, Henrik Toft Sørensen1,2, Jiong Li1,2. 1. Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark. 2. Department of Clinical Medicine, Aarhus University, Aarhus, Denmark. 3. Institute of Biological Psychiatry, Mental Health Center Sct. Hans, Mental Health Services Copenhagen, Roskilde, Denmark. 4. Department of Biostatistics, School of Public Health, and The Key Laboratory of Public Health Safety of Ministry of Education, Fudan University, Shanghai, China. 5. Department of Environmental Health Sciences, Yale School of Public Health, New Haven, Connecticut. 6. Yale Center for Perinatal, Pediatric and Environmental Epidemiology, Yale School of Public Health, New Haven, Connecticut.
Abstract
Importance: Maternal diabetes has been suggested as a risk factor for attention-deficit/hyperactivity disorder and autism in offspring, but evidence on its association with the full spectrum of psychiatric disorders remains lacking. Objective: To investigate the associations between maternal diabetes diagnosed before or during pregnancy and 10 types of psychiatric disorders in offspring during the first 4 decades of life. Design, Setting, and Participants: This population-based cohort study used data from several Danish nationwide medical and administrative registries in Denmark on all 2 413 335 live births from 1978 to 2016. Data were analyzed between October 1, 2019, and July 15, 2021. Exposures: Any maternal diabetes diagnosis during pregnancy (56 206 offspring [2.3%]) and 3 diabetes subtypes (pregestational type 1 diabetes, 22 614 offspring [1.0%]; pregestational type 2 diabetes, 6713 offspring [0.3%]; and gestational diabetes, 26 879 offspring [1.1%]). Main Outcomes and Measures: Outcomes included 10 types of psychiatric disorders: any psychiatric disorder, substance use disorders, schizophrenia, mood disorders, anxiety disorders, eating disorders, personality disorders, intellectual disorders, developmental disorders, and behavioral disorders. Hazard ratios (HRs) and 95% CIs were computed using Cox proportional hazards regression models. Covariates included maternal and paternal history of any psychiatric disorder, offspring sex, calendar period of birth, singleton status, and several maternal characteristics during pregnancy (ie, age, parity, educational level, smoking, cohabitation, residence, and body mass index). Sibship design and competing risk analyses were also conducted. Results: A total of 2 413 335 individuals (1 239 148 male participants [51%]; age range, 1-39 years; median age, 19.0 years [IQR, 5.8-20.8 years]) were included in this study. During the 39-year follow-up time, 151 208 offspring (6.4%) received a diagnosis of a psychiatric disorder. Offspring born to mothers with any diabetes diagnosis during pregnancy were at increased risk of developing any psychiatric disorder (HR, 1.15; 95% CI, 1.10-1.20), schizophrenia (HR, 1.55; 95% CI, 1.15-2.08), anxiety disorders (HR, 1.22; 95% CI, 1.09-1.36), intellectual disabilities (HR, 1.29; 95% CI, 1.11-1.50), developmental disorders (HR, 1.16; 95% CI, 1.03-1.30), and behavioral disorders (HR, 1.17; 95% CI, 1.08-1.27) compared with offspring born to mothers without a diabetes diagnosis during pregnancy. No association was observed for substance use disorders, mood disorders, eating disorders, and personality disorders. Conclusions and Relevance: This study shows a pattern that suggests that prenatal exposure to maternal diabetes during pregnancy was associated with increased risk of psychiatric disorders overall and most specific psychiatric disorders in offspring in their first 4 decades of life.
Importance: Maternal diabetes has been suggested as a risk factor for attention-deficit/hyperactivity disorder and autism in offspring, but evidence on its association with the full spectrum of psychiatric disorders remains lacking. Objective: To investigate the associations between maternal diabetes diagnosed before or during pregnancy and 10 types of psychiatric disorders in offspring during the first 4 decades of life. Design, Setting, and Participants: This population-based cohort study used data from several Danish nationwide medical and administrative registries in Denmark on all 2 413 335 live births from 1978 to 2016. Data were analyzed between October 1, 2019, and July 15, 2021. Exposures: Any maternal diabetes diagnosis during pregnancy (56 206 offspring [2.3%]) and 3 diabetes subtypes (pregestational type 1 diabetes, 22 614 offspring [1.0%]; pregestational type 2 diabetes, 6713 offspring [0.3%]; and gestational diabetes, 26 879 offspring [1.1%]). Main Outcomes and Measures: Outcomes included 10 types of psychiatric disorders: any psychiatric disorder, substance use disorders, schizophrenia, mood disorders, anxiety disorders, eating disorders, personality disorders, intellectual disorders, developmental disorders, and behavioral disorders. Hazard ratios (HRs) and 95% CIs were computed using Cox proportional hazards regression models. Covariates included maternal and paternal history of any psychiatric disorder, offspring sex, calendar period of birth, singleton status, and several maternal characteristics during pregnancy (ie, age, parity, educational level, smoking, cohabitation, residence, and body mass index). Sibship design and competing risk analyses were also conducted. Results: A total of 2 413 335 individuals (1 239 148 male participants [51%]; age range, 1-39 years; median age, 19.0 years [IQR, 5.8-20.8 years]) were included in this study. During the 39-year follow-up time, 151 208 offspring (6.4%) received a diagnosis of a psychiatric disorder. Offspring born to mothers with any diabetes diagnosis during pregnancy were at increased risk of developing any psychiatric disorder (HR, 1.15; 95% CI, 1.10-1.20), schizophrenia (HR, 1.55; 95% CI, 1.15-2.08), anxiety disorders (HR, 1.22; 95% CI, 1.09-1.36), intellectual disabilities (HR, 1.29; 95% CI, 1.11-1.50), developmental disorders (HR, 1.16; 95% CI, 1.03-1.30), and behavioral disorders (HR, 1.17; 95% CI, 1.08-1.27) compared with offspring born to mothers without a diabetes diagnosis during pregnancy. No association was observed for substance use disorders, mood disorders, eating disorders, and personality disorders. Conclusions and Relevance: This study shows a pattern that suggests that prenatal exposure to maternal diabetes during pregnancy was associated with increased risk of psychiatric disorders overall and most specific psychiatric disorders in offspring in their first 4 decades of life.
Diabetes is one of the most common chronic diseases among fertile women worldwide, and its rates are increasing.[1] Pregnancies complicated by maternal diabetes are hypothesized to lead to psychiatric disorders in offspring through several mechanisms, such as oxidative stress and hypoxia of the fetus.[2,3,4,5] Epidemiological studies have indicated that offspring exposed to maternal diabetes during pregnancy are more likely to develop autism spectrum disorders (ASDs)[6,7,8,9] and attention-deficit/hyperactivity disorder (ADHD).[10,11,12]So far, empirical evidence on the associations between maternal diabetes during pregnancy and a full spectrum of psychiatric disorders is limited, to our knowledge.[3] Previous studies followed up with participants only during childhood or adolescence (offspring age range, 4 months to 17 years),[2,3,4] but many psychiatric disorders do not develop until the age of 10 years.[13] Moreover, some findings are inconsistent.[6,7,8,9] For example, while a multiethnic clinical cohort in California found an association between maternal diabetes during pregnancy and ASD in offspring,[6] 3 other studies using the Boston Birth Cohort and Finish registries have not found such an association,[7,8,9] which underlies the continuing need for further investigation.We hypothesized that prenatal exposure to a maternal diabetes diagnosis during pregnancy would increase the risk of offspring developing psychiatric disorders in childhood and adulthood (up to the age of 40 years). Therefore, in this Danish population-based cohort study, we investigated the associations between maternal diabetes diagnosed before or during pregnancy and the onset of 10 broad types of psychiatric disorders in offspring during the first 4 decades of life. Specifically, we examined the risks of maternal diabetes during pregnancy associated with any psychiatric disorder, substance use disorders, schizophrenia, mood disorders, anxiety disorders, eating disorders, personality disorders, intellectual disorders, developmental disorders, and behavioral disorders.
Methods
Study Design and Participants
This cohort study used data from several Danish nationwide registries (eAppendix 1 in the Supplement).[14] We included all live births in Denmark from 1978 to 2016 (N = 2 413 335). Follow-up started at the lowest possible age of onset for each group of psychiatric disorders (minimum, 1 year) (eTable 1 in the Supplement)[13] and ended on the date of a first psychiatric diagnosis, death, emigration, or December 31, 2016, whichever came first. Statistics Denmark and the Danish Health Data Authorities granted access to Danish registries in an anonymous and secure form. The Danish Data Protection Agency approved the study. By Danish law, informed consent is not required for a register-based study based on anonymized data. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Exposure
Exposure was defined as a maternal diabetes diagnosis registered during pregnancy.[1] Thus, offspring born to mothers with a diagnosis of diabetes before childbirth were considered to have been exposed to maternal diabetes during pregnancy. Maternal diabetes was first classified into pregestational diabetes (PGDM) and gestational diabetes (GDM). Pregestational diabetes was further categorized as PGDM type 1 and type 2. To identify PGDM, we applied criteria by extracting the following International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) and International Classification of Diseases, Eighth Revision (ICD-8) equivalent codes of diabetes diagnosis and treatment: (1) a diagnosis of diabetes with ICD-10 codes of E10, E11, H36.0, and O24, excluding codes O24.4 and O24.9, or an ICD-8 code of 249 or 250; (2) receipt of podiatric treatment for diabetic conditions; (3) 2 redeemed prescriptions for insulin (Anatomical Therapeutic Chemical code A10A); or (4) 2 redeemed prescriptions for oral antidiabetics (Anatomical Therapeutic Chemical code A10B). Blood glucose measurements were not used owing to their poor validity.[15] For GDM, we used ICD-10 codes O24.4 and O24.9 and ICD-8 code 634.74. Because diabetes was recorded using a single ICD-8 code (250) during the period from 1977 to 1986, we used 2 approaches to distinguish between type 1 and type 2 diabetes during that period: (1) a specific code for type 1 or type 2 diabetes conferred later or (2) age at diabetes onset (cutoff age for type 1, <30 years; cutoff age for type 2, ≥30 years).[11,12]In the present study, 56 206 offspring (2.3%) were born to mothers with diabetes (PGDM type 1, 22 614 [1.0%]; PGDM type 2, 6713 [0.3%]; and GDM, 26 879 [1.1%]). As most women in our sample were younger than 30 years, the prevalence of PGDM type 1 was higher than the prevalence of PGDM type 2 (Table). A detailed description of the methods used to identify diabetes is in eAppendix 2 and eTable 2 in the Supplement. If a mother received diagnoses of multiple types of diabetes during the same pregnancy owing to misclassification, she was classified in the main analysis according to the first recorded diagnosis (eAppendix 2 and eTable 2 in the Supplement).[16] Information on diabetes diagnosis was obtained from the Danish National Diabetes Register, the Danish National Patient Registry, and the Danish National Prescription Registry (eAppendix 2 in the Supplement).[14] In Denmark, diseases, including diabetes, have been coded according to the Danish modification of the ICD-8 until 1993 and the ICD-10 from 1994 until the present.[14,16]
Table.
Characteristics of the Cohort by Exposure to Maternal Diabetes During Pregnancy, Denmark, 1978-2016
Maternal smoking during pregnancy was available starting in 1991.
Abbreviations: GDM, gestational diabetes; PGDM, pregestational diabetes.Maternal smoking during pregnancy was available starting in 1991.
Outcomes
Outcomes included 10 broad types of psychiatric disorders diagnosed in offspring during their first 4 decades of life as described in previous publications using Danish registries.[13,17,18] The following categories were examined: any psychiatric disorder, substance use disorders, schizophrenia, mood disorders, anxiety disorders, eating disorders, personality disorders, intellectual disabilities, developmental disorders, and behavioral disorders (eTable 1 in the Supplement).[13,17] For every group of psychiatric disorders, we started follow-up according to the lowest possible age at onset as reported in the literature.[13,18] A detailed description of the diagnosis added to each category is presented in eTable 1 in the Supplement. Information on psychiatric disorder diagnoses was obtained from the Danish National Patient Registry and the Danish Psychiatric Central Research Register (eAppendix 1 in the Supplement).[14]
Covariates
We retrieved data on the following potential confounders: maternal and paternal history of any psychiatric disorder (yes or no); maternal age (<20, 20-24, 25-29, 30-34, or ≥35 years); parity (1, 2, or ≥3 children); maternal educational level (0-9, 10-14, or ≥15 years of schooling); maternal smoking during pregnancy (yes or no); maternal cohabitation (single or cohabiting); and maternal residence (Copenhagen, cities with ≥100 000 inhabitants, or other). We further included singleton status (yes or no), offspring sex (male or female), and calendar period (birth year before 1980 or 5-year intervals during 1981-2010, and 2011-2016) to control for possible confounding due to birth cohort effects and changes in disease diagnoses during the study period (eAppendix 3 in the Supplement).[16] We also included data on maternal body mass index (BMI; calculated as weight in kilograms divided by height in meters squared), categorized as less than 18.5, 18.5 to less than 25.0, 25.0 to less than 30.0, 30.0 to less than 35.0, and 35.0 or more. Except for data on maternal smoking (accessible from 1991) and maternal BMI (accessible from 2004), all other variables were available for the entire cohort (eAppendix 3 in the Supplement).[16]
Statistical Analysis
Statistical analyses were performed between October 1, 2019, and July 15, 2021. Descriptive statistics were used to summarize the characteristics of offspring and their mothers according to exposure to maternal diabetes during pregnancy (Table). Incidence rates for the unexposed and exposed groups were calculated as the total number of new cases (first psychiatric disorder diagnosis) divided by the total follow-up time in person-years (eTable 3 in the Supplement). Considering all causes of death as competing risks, competing risk analyses, using the cumulative incidence function in SAS, version 9.4 (SAS Institute Inc) (Fine and Gray model),[19] were performed to estimate the cumulative incidence of a diagnosis of psychiatric disorders (Figure 1 and Figure 2). Individuals were censored at emigration or end of follow-up (December 31, 2016). We used Cox proportional hazards regression models to compute hazard ratios (HRs) and 95% CIs to evaluate the association between maternal diabetes during pregnancy and development of psychiatric disorders in offspring (Figure 3). Model 1 was unadjusted and included only maternal diabetes. Model 2 was adjusted for maternal and paternal history of any psychiatric disorder as well as calendar period of birth.[20] Model 3 was adjusted additionally for sex, singleton status, and maternal characteristics (parity, educational level, smoking, cohabitation, residence, and age).[6] For all 3 models, we used the year of follow-up since birth for each offspring as the underlying time scale.
Figure 1.
Risks of Psychiatric Disorders in Offspring Exposed to Any Maternal Diabetes Diagnosis During Pregnancy Compared With the Unexposed
A, Any psychiatric disorder. B, Substance use disorders. C, Schizophrenia. D, Mood disorders. E, Anxiety disorders.
Figure 2.
Risks of Psychiatric Disorders in Offspring Exposed to Any Maternal Diabetes Diagnosis During Pregnancy Compared With the Unexposed
A, Eating disorders. B, Personality disorders. C, Intellectual disabilities. D, Developmental disorders. E, Behavioral disorders.
Figure 3.
Associations Between Maternal Diabetes During Pregnancy and Psychiatric Disorders in Offspring During the First 4 Decades of Life
Model 1 indicates unadjusted; model 2, adjusted for maternal and paternal history of psychiatric disorder, and calendar year of birth; and model 3, adjusted for sex, singleton status, maternal characteristics (parity, educational level, smoking, cohabitation, residence, age). GDM indicates gestational diabetes; HR, hazard ratio; and PGDM, pregestational diabetes.
Risks of Psychiatric Disorders in Offspring Exposed to Any Maternal Diabetes Diagnosis During Pregnancy Compared With the Unexposed
A, Any psychiatric disorder. B, Substance use disorders. C, Schizophrenia. D, Mood disorders. E, Anxiety disorders.A, Eating disorders. B, Personality disorders. C, Intellectual disabilities. D, Developmental disorders. E, Behavioral disorders.
Associations Between Maternal Diabetes During Pregnancy and Psychiatric Disorders in Offspring During the First 4 Decades of Life
Model 1 indicates unadjusted; model 2, adjusted for maternal and paternal history of psychiatric disorder, and calendar year of birth; and model 3, adjusted for sex, singleton status, maternal characteristics (parity, educational level, smoking, cohabitation, residence, age). GDM indicates gestational diabetes; HR, hazard ratio; and PGDM, pregestational diabetes.Our missing data analysis indicated that, in the final study population, there were no missing values for maternal age, maternal and paternal history of psychiatric disorders, singleton status, parity, maternal age at childbirth, maternal residency, or offspring age (Table). Also, the proportion of missing values was very low for offspring sex (0.0%), maternal cohabitation (0.0%), and maternal educational level (5%). We handled missing data on covariates by multiple imputation techniques using SAS MI procedure (eAppendix 3 in the Supplement). We used a fully conditional specification method to impute 10 replications as advised by White et al.[21] This method made the computation feasible for our large data set with millions of observations.[22] The imputation model included the following variables: maternal diabetes during pregnancy (yes or no), offspring psychiatric disorder (yes or no), maternal age (continuous), parity (1, 2, or ≥3 children), maternal cohabitation (single or cohabitating), maternal educational level (0-9, 10-14, or ≥15 years), maternal residence (Copenhagen, cities with ≥100 000 inhabitants, or other), maternal smoking during pregnancy (yes or no), maternal and paternal history of psychiatric disorder (yes or no), maternal country of origin (Nordic countries or others) (yes or no), sex of offspring (male or female), and calendar period of birth (before 1980 or 5-year intervals during 1981-2010 or 2011-2016).[16] The distributions of observed and imputed variables were similar.[16] Because maternal prepregnancy BMI is a risk factor for both maternal diabetes and psychiatric disorders in offspring[4] but the information on maternal pregnancy BMI was available only beginning in 2004, we performed a sensitivity analysis with additional adjustments for prepregnancy BMI (eTable 5 in the Supplement). Furthermore, to strengthen causal inference, we carried out a sibship analysis to evaluate the association between maternal diabetes and any psychiatric disorder. In association studies, the sibship design can provide a more robust basis for causal inference because it allows us to control for shared genetic and environmental factors between siblings.[23] Thus, in adjusted model 3, we performed stratified Cox proportional hazards regression analysis by including a stratum for each family identified by the mother’s unique civil registration number. Only sibling pairs discordant for both maternal diabetes and offspring psychiatric disorders were informative and contributed to the effect estimate, which provides information on the association of diabetes per se with the psychiatric disorder (eTable 6 in the Supplement). The analyses were performed using SAS, version 9.4 and Stata, version 14 (StataCorp).
Results
This study included 2 413 335 individuals born in Denmark from 1978 to 2016 (1 239 148 male participants [51.3%]; age range, 1-39 years; median age, 19.0 years [IQR, 5.8-20.8 years]). Additional maternal and offspring characteristics at baseline are presented in the Table. For example, offspring whose mothers had a history of psychiatric illness most often had mothers exposed to some form of maternal diabetes during pregnancy (7428 [9.4%]). Participants were followed up for up to 39 years (median follow-up, 19.0 years [range, 1-39 years]), resulting in 412 662 622 million person-years of observations. During the 39-year follow-up, 148 017 (6.4%) offspring (146 053 unexposed and 1964 exposed [PGDM type 1 = 850, PGDM type 2 = 377, and GDM = 737]) received a diagnosis of a psychiatric disorder (eTable 1 in the Supplement). The incidence rates of any psychiatric disorder per 10 000 person-years were higher among the exposed offspring (38.0 [PGDM type 1 = 38.0; PGDM type 2 = 42.1] and GDM = 36.5) compared with the unexposed group (35.8) (eTable 3 in the Supplement). The competing risk analysis (Figure 1) showed that the overall risk of developing any psychiatric disorder during the first 4 decades of life was higher among offspring exposed to maternal diabetes than among unexposed offspring. The risk increased steadily starting at 1 year of age. For schizophrenia and personality disorders, the exposed group’s increased risk was observed only starting at the age of 20 years (Figure 1 and Figure 2).Figure 3 shows the results of models 1, 2, and 3 for 40 possible associations between maternal diabetes during pregnancy and psychiatric disorders in offspring, of which we chose to report the results adjusted for potential confounders (model 3). Offspring born to mothers with any diabetes diagnosis during pregnancy were at increased risk of developing any psychiatric disorder (HR, 1.15; 95% CI, 1.10-1.20), schizophrenia (HR, 1.55; 95% CI, 1.15-2.08), anxiety disorders (HR, 1.22; 95% CI, 1.09-1.36), intellectual disabilities (HR, 1.29; 95% CI, 1.11-1.50), developmental disorders (HR, 1.16; 95% CI, 1.03-1.30), and behavioral disorders (HR, 1.17; 95% CI, 1.08-1.27), compared with offspring born to mothers without a diabetes diagnosis during pregnancy. We did not find associations of any maternal diabetes diagnosis with substance use disorders (HR, 0.99; 95% CI, 0.91-1.08), mood disorders (HR, 1.12; 95% CI, 0.90-1.39), eating disorders (HR, 1.21; 95% CI, 0.98-1.50), and personality disorders (HR, 1.09; 95% CI, 0.71-1.65).Sensitivity analyses, in which we performed multiple imputations of missing values of the covariates (eTable 4 in the Supplement) and the analyses additionally adjusted for maternal BMI (eTable 5 in the Supplement), yielded findings similar to those of the primary analyses. Finally, the sibship analysis results showed that the association between maternal diabetes during pregnancy and any psychiatric disorder was attenuated but remained positive for PGDM types 1 and 2, a more robust association of type 2 than the primary analysis (eTable 6 in the Supplement).
Discussion
Our findings expand the current literature by showing that maternal diabetes during pregnancy was associated with increased risks for psychiatric disorders overall and several specific psychiatric disorders (schizophrenia, anxiety disorders, intellectual disabilities, developmental disorders, and behavioral disorders) in a population-based Danish birth cohort. We did not find associations of maternal diabetes during pregnancy with substance use disorders, mood disorders, eating disorders, and personality disorders in offspring.Most studies that have explored the association between maternal diabetes during pregnancy and psychiatric disorders in offspring have focused on ASD, a developmental disorder,[6,7,8,9] and ADHD, a behavioral disorder.[10,11,12] Xiang et al[6] reported that PGDM type 2 and GDM were associated with the occurrence of ASD during 17 years of follow-up, independently of maternal BMI. After adjusting for maternal BMI, 2 studies in Finland with 10 years of follow-up[7,8] and 1 study in the US with 16 years of follow-up[9] did not find an association between maternal diabetes and ASD. In our study, ASD was associated with prenatal exposure to PGDM type 1 and GDM. Moreover, we found that GDM was associated with the incidence of ADHD during 4 decades of follow-up. This finding accords with previous studies demonstrating that offspring exposed to GDM[14,15] or PGDM types 1 or 2[16] were more likely to receive a diagnosis of ADHD during 6 years of follow-up.Empirical evidence on the long-term associations of maternal diabetes with mental health outcomes other than ASD and ADHD is limited, to our knowledge. So far, only 1 study has addressed schizophrenia in offspring of mothers with diabetes. This 2019 Japanese study of 4478 adolescents investigated the association between GDM and adolescents’ self-reported experiences with psychotic symptoms that characterize schizophrenia, such as hallucinations.[24] Consistent with our study, it reported that GDM was associated with an increased number of psychotic experiences. Nonetheless, schizophrenia starts developing at a minimum age of 10 years, which was the mean age of the participants of the Japanese study. Two studies further showed that PGDM types 1 and 2 and GDM were associated with anxiety and intellectual disabilities only in combination with elevated maternal BMI.[7,8] Increased maternal BMI during pregnancy may also be associated with psychiatric disorders via the inflammatory processes that occur in the uterus owing to the lipotoxic effects associated with overweight.[8] Thus, although exposure to maternal diabetes during pregnancy is hypothesized to be associated with the neurodevelopment of offspring, the joint associations of maternal diabetes and an elevated BMI might engender more pronounced damages to neurodevelopment than the individual associations.Moreover, we found no associations between maternal diabetes during pregnancy and substance use disorders, mood disorders, eating disorders, and personality disorders. We know that genetic factors play an important role in mood disorders, such as major depression throughout life.[20,23] Another possibility is that the environment may play a greater role than fetal programming for some psychiatric disorders.[25] Perhaps substance use, mood disorders, and eating disorders are more primarily associated with sociocultural factors (eg, peer pressure and social norms).[26] It is possible, for example, that someone would start substance use because of perceived benefits from their friends, such as social acceptance.[27] Also, an individual could develop an eating disorder owing to the social norm to be fit, which is present in many societies.[28] However, these examinations are outside of our study’s scope, which leaves something to consider in future research.The potential mechanisms underlying the observed associations between maternal diabetes during pregnancy and schizophrenia, anxiety disorders, intellectual disabilities, developmental and behavioral disorders may include oxidative stress[5,29,30,31] and fetual hypoxia.[1] Experimental animal studies have shown that induced hyperglycemia can affect organ development during the embryonic period. The key mechanism may be associated with a chain of chemical reactions triggered by hyperglycemia that leads to increased production of reactive oxygen species (ROS),[1] which are molecules containing oxygen derived from normal cell metabolism. As hyperglycemia increases and the ROS level increases, antioxidants in the body may be insufficient to counteract all ROS.[1] In excessive amounts, ROS react with tissue molecules, leading to oxidative stress. In embryos, oxidative stress can damage the DNA of all types of cells, including the central nervous system.[1] This damage may lead to long-lasting neurodevelopmental impairments, including psychiatric disorders. In their 2019 study, Wang et al[5] demonstrated in vitro hyperglycemia leading to ROS production, oxidative stress, and damage to the neural tube and brain development. Moreover, the oxidative stress activates the signaling pathways that lead to mitochondria death, which may cause fetal hypoxia, damage brain development, and lead to psychiatric disorders in offspring later.[1]
Strengths and Limitations
Our study has several strengths. First, we used high-quality data from Danish registries with relatively high validity and completeness for most variables, such as maternal diabetes.[32] Second, the nearly complete follow-up minimizes selection bias.[14] Third, we followed up with our participants for 4 decades—the longest follow-up, to our knowledge, examining the long-term associations of maternal diabetes with a full spectrum of nonorganic psychiatric disorders as the outcome of interest. Fourth, our analyses included a wide range of potential confounders, especially parental history of any psychiatric disorder and maternal educational level, which is considered a good proxy for socioeconomic status. As suspected,[6] we observed that the high HR observed in the unadjusted models were primarily confounded by parental history of psychiatric disorders. However, our findings from sibship analysis showed a 9% and 28% increase in risks of any psychiatric disorder associated with PGDM types 1 and 2, respectively. The positive findings suggest that exposure to maternal diabetes during pregnancy might play a role in later susceptibility for psychiatric disorders.This study also has several limitations. First, owing to the observational study design, we could not rule out the possibility of unmeasured confounders, such as lifestyle factors (eg, maternal alcohol consumption). Second, uncertainty regarding complete ascertainment of psychiatric disorders is possible. However, although validity varies across diagnoses, overall, the registry-based psychiatric diagnoses in Denmark are given by psychiatrists and are found to be valid.[13,18,33,34,35,36] Third, diagnoses of psychiatric disorders reflected more severe cases because they were recorded in hospital settings. Thus, our study did not include patients with mild psychiatric disorders diagnosed in general practice alone. The 2019 study by Musliner et al[37] projected that approximately 25% of Danish individuals who start treatment with antidepressants would receive a hospital-based psychiatric diagnosis within 5 years. Not covering all cases of psychiatric disorders could have underestimated our findings, but we cannot be sure. Fourth, although our results might apply to countries similar to Denmark, which provides an established and secure welfare state, including high-quality health care for pregnant women,[14] they might not be generalizable to all countries. Among low- and middle-income countries, prenatal care may not be equally available to all women, hampering the detection of diabetes during pregnancy; health care quality disparities also exist between urban and rural areas.[38] Thus, we recommend that our research be replicated in low- and middle-income countries. Furthermore, given the positive associations between maternal diabetes during pregnancy and psychiatric disorders among Danish offspring, we suspect that these associations would be stronger in populations from low- and middle-income countries.
Conclusions
Our study shows a pattern in which prenatal exposure to maternal diabetes diagnosed during pregnancy was associated with increased risks for several psychiatric disorders in the first 4 decades of life. Children born to mothers with any diabetes diagnosis during pregnancy had increased risks of any psychiatric disorder, schizophrenia, anxiety disorders, intellectual disabilities, developmental disorders, and behavioral disorders. These findings suggest that careful clinical management of maternal diabetes during pregnancy is needed. Overall, our findings signal the importance of effective strategies for preventing, screening, and treating diabetes among women of fertile age for women’s health and children’s mental health.
Authors: Natalie C Momen; Oleguer Plana-Ripoll; Esben Agerbo; Michael E Benros; Anders D Børglum; Maria K Christensen; Søren Dalsgaard; Louisa Degenhardt; Peter de Jonge; Jean-Christophe P G Debost; Morten Fenger-Grøn; Jane M Gunn; Kim M Iburg; Lars V Kessing; Ronald C Kessler; Thomas M Laursen; Carmen C W Lim; Ole Mors; Preben B Mortensen; Katherine L Musliner; Merete Nordentoft; Carsten B Pedersen; Liselotte V Petersen; Anette R Ribe; Annelieke M Roest; Sukanta Saha; Andrew J Schork; Kate M Scott; Carson Sievert; Holger J Sørensen; Terry J Stedman; Mogens Vestergaard; Bjarni Vilhjalmsson; Thomas Werge; Nanna Weye; Harvey A Whiteford; Anders Prior; John J McGrath Journal: N Engl J Med Date: 2020-04-30 Impact factor: 91.245
Authors: Jonathan A C Sterne; Ian R White; John B Carlin; Michael Spratt; Patrick Royston; Michael G Kenward; Angela M Wood; James R Carpenter Journal: BMJ Date: 2009-06-29
Figure 1. 
Figure 2. 
Figure 3.