Previous studies have been showed an association between previous abortion history and risk of Autism spectrum disorders (ASD). However, there is still controversy about true effect estimate of the association.
This meta-analysis aimed to evaluate the association between previous abortion history and the risk of ASD.
A systematic search was performed using PubMed, Scopus, and Web of Sciences databases to identify potential studies published until December 2021. The presence of statistical heterogeneity was determined using the
Thirteen studies with a total of 331,779 children remained in the present meta-analysis. The estimated odds ratio of the risk of ASD associated with previous abortion history was 1.64 (95% confidence interval [CI], 1.28–2.0;
This meta-analysis suggests that children born from mothers with a history of previous abortion have an increased risk of development of ASD compared to children of mothers without a history of previous abortion.
Statistics have shown that the prevalence of autism spectrum disorder (ASD), a neurological and developmental disorder, among children has increased in recent decades [
A constellation of genetic and environmental factors plays a role in the etiology of ASD; however, the true causal pathways of ASD remain unclear and require further investigation [
The effects of a previous abortion history on ASD are controversial. Some studies reported a significant positive association between previous abortion and ASD [
To resolve such controversy, it is necessary that the observed results of the association between previous abortion history and ASD be systematically gathered, their findings summarized and combined, and the overall effect measure of this association be derived. Therefore, this study aimed to evaluate the association between abortion history and the risk of ASD through a systematic review and meta-analysis of results of previous observational studies.
This systematic review and meta-analysis was performed according to the meta-analyses of observational studies in epidemiology (MOOSE) guideline [
A systematic search was performed of the PubMed, Scopus, and Web of Sciences databases to identify studies published through December 2021. To search these databases, the following MeSH terms and keywords were used: (“spontaneous abortion” or “induced abortion” or “abortion history” or “previous abortion history” or “previous abortion” or “previous miscarriage” or “previous miscarriage history”) and (ASD or “autism spectrum disorder” or “autism spectrum disorders” or autism). We manually searched the reference lists of the included articles to identify additional potential articles. No language restrictions were set.
The main inclusion criteria were as follows: observational study design (cross-sectional, case-control, or cohort) evaluating the association between previous abortion history and ASD and providing the relative risks and corresponding confidence intervals (CIs) for this association. Nonoriginal articles, such as reviews, editorials, commentaries, correspondences, notes, and book reviews, were excluded. According to these criteria, 2 authors (EJ and EA) independently screened the titles and abstracts of the articles identified in the initial searches.
The following data were independently extracted by 2 authors (EJ and EA): first author, year of publication, study design, sample size, diagnosis method for ASD, child age, type of relative risk, and corresponding 95% CI. The 2 authors (EJ and EA) independently assessed the quality of the eligible articles using the Newcastle-Ottawa Scale (NOS) score [
Statistical heterogeneity was determined using
Review of the retrieved studies and their references resulted in the inclusion of 732 studies published through December 7, 2021. Of these, 166 were duplicates and deleted. Thereafter, 547 studies were excluded by the title and abstract review, leaving 19 studies for full-text review. Six studies did not meet the eligibility criteria. Therefore, 13 studies were included in this meta-analysis (
The association between previous abortion history and ASD risk is shown in
Begg and Egger tests were used to assess publication bias; their P values among children with ASD were 0.393 and 0.056, respectively. Publication bias was not identified among the studies that reported an association between previous abortion history and the risk of ASD (
According to the NOS scale, 12 studies were of high quality and 1 was of low quality (
The main objective of this meta-analysis was to estimate the precise effect of the association between a previous abortion history and ASD risk. On stratification of adjusted ORs, our analysis showed that the risk of ASD may increase by 64% among children whose mothers had a history of previous abortions (pooled adjusted OR, 1.64; 95% CI, 1.28–2.00).
Our analysis suggests that the offspring of women with a history of abortion were at higher risk of ASD. Regardless of ASD as a neurodevelopmental disorder, previous studies demonstrated a positive association between abortion and other neurodevelopmental disorders such as attention-deficit/hyperactivity disorder [
Some plausible pathways underlie the association between abortion and ASD risk. For example, regardless of the direct effect of abortion on ASD, the relationship between them may occur through intermediate factors, such as preterm birth, low birth weight, and a lower Apgar score, risk factors for ASD [
There are data on the association between maternal autoimmunity, such as thyroid antibodies and recurrent pregnancy loss (RPL) [
The number of meta-analyses of the determinants of ASD has increased over the last decade. For example, maternal infection during pregnancy [
An important issue worth consideration is that the pooled adjusted OR of 1.64 of the association between previous abortion and ASD may not be generalizable to the association between abortion and autistic disorder or between abortion and Asperger syndrome. However, the pooled OR for abortion and ASD may be modified by abortion number and type. As shown in
This study had several limitations. First, there was a high level of heterogeneity among the pooled studies (
In conclusion, the results of this meta-analysis suggest that the offspring of mothers with a history of previous abortion are at an increased risk of developing ASD compared to those of mothers without a history of previous abortion. Further large-scale prospective cohort studies with better ascertainment of abortion, ASD spectrum components, and confounders are recommended to increase the validity and robustness of our findings.
Supplementary material: Supplementary Table 1 can be found via
MOOSE (meta-analyses of observational studies in epidemiology) checklist for meta-analyses of observational studies
The authors have nothing to disclose.
This study was supported and funded by Hamadan University of Medical Sciences with code 140010218787.
Conceptualization: EJ, EA; Formal Analysis: EJ, EA; Investigation: EJ, SB, MS, MR; Methodology: EJ, EA, SB, MS, MR; Project Administration: EJ, EA; Writing- Original Draft: EJ, EA; Writing-Review & Editing: EJ, EA, SB, MS, MR
Flow of information through the different phases of the systematic review.
Forest plot of the association between previous abortion history and the risk of autism spectrum disorders. OR, odds ratio; CI, confidence interval.
Funnel plot of the association between previous abortion history and the risk of autism spectrum disorders. SE, standard error.
Characteristics of the studies included in the meta-analysis
Design | Study | Sample | Diagnose method | Child age (mean/range) | Estimate | Adjustment | Outcome | Quality |
---|---|---|---|---|---|---|---|---|
Case-control | Stein et al., 2006 [ |
358 | DSM-IV | 4–26 Years | OR | Crude | Previous induced abortion | High |
Cohort | Dodds et al., 2011 [ |
129,733 | ICD-9, 10 | Not reported | OR | Crude | Previous abortion | High |
Cohort | Lyall et al., 2012 [ |
66,445 | PDD-NOS | Not reported | OR | Adjusted | Previous abortion | High |
Case-control | Hamadé et al., 2013 [ |
258 | DSM-IV | 9.94 Years | OR | Crude | Previous abortion | High |
Cohort | Roberts et al., 2013 [ |
116,430 | ADI-R | <12 Years | RR | Adjust | Previous abortion | High |
Cohort | Cheslack-Postava et al., 2014 [ |
7,649 | DSM-IV-TR, ICD-9,10 | Not reported | OR | Crude | Previous abortion | High |
Case-control | Ou et al., 2019 [ |
2,306 | DSM-IV-TR | Not reported | OR | Adjusted | Previous induced abortion | High |
Cross-sectional | Vui et al., 2019 [ |
5,893 | DSM-IV | 18–30 Months | OR | Crude | Previous abortion | High |
Case-control | Bitar et al., 2020 [ |
131 | DSM-IV | 8.61 Years | OR | Crude | Previous abortion | High |
Case-control | Magdalena et al., 2020 [ |
221 | ADOS | 3–12 Years | OR | Crude | Previous abortion | High |
Cross-sectional | Auf et al., 2020 [ |
500 | CARS | 3–12 Years | OR | Crude | Previous abortion | Low |
Case-control | Qutranji et al., 2021 [ |
141 | DSM-IV | <18 Years | OR | Adjusted | Previous abortion | High |
Case-control | Yang et al., 2021 [ |
1,714 | DSM-IV | 4.4 Years | OR | Adjusted | Previous abortion | High |
DSM, Diagnostic and Statistical Manual of Mental Disorders; ICD, International Statistical Classification of Diseases; PDD-NOS, pervasive developmental disorder not otherwise specified; ADI-R, autism diagnostic interview-revised; ADOS, autism diagnostic observation schedule; CARS, childhood autism rating scale; OR, odds ratio; RR, risk ratio.