This is the first and largest study to use a population-based dataset to examine the impact of MLBW on subsequent psychiatric disorders. The study findings suggest that term-born children with MLBW were more likely than term-born children without MLBW to be diagnosed as having developmental delays or intellectual disability, ADHD, ASD, conduct disorder, and emotional disturbances.
Our finding that term-born infants with MLBW were more likely to have developmental delays supports the findings of previous studies that have reported an increased association of lower intelligence, cognitive difficulties, or lower school achievements in term-born SGA infants compared with those with normal birth weights (2, 9, 14). The mechanisms underlying developmental delays or intellectual impairment are still unclear but may be a consequence of perinatal brain injuries disrupting cortical growth (33). Placental insufficiency is another possible factor affecting fetal cerebral hemodynamics leading to deficits in brain development (2, 10). However, placental insufficiency may be a confounding factor for poor neurodevelopmental outcomes, such as language, intelligence, or cognitive developmental disorders. Other possible causes, including maternal factors such as mothers’ genetic loading, incident prenatal infections, or intrauterine exposure to cigarettes, alcohol, or narcotic drugs and their associations with intellectual disability or developmental disabilities may also be considered (2). Moreover, placental insufficiency may not be associated with neurodevelopmental problems. Savchev et al. reported that although SGA infants had poorer neurological developments, parameters from their prenatal umbilical artery Doppler assessment showed no signs of placental insufficiency (13). Future investigations are required to examine possible causal factors for developmental issues in term-born infants with MLBW and genetic or intrauterine exposures.
Our finding that term-born infants with MLBW had an increased risk of ADHD is consistent with those of previous studies that have reported an increased association of LBW or SGA with attentional problems or aggressive behaviors (16, 17, 34). Multifactorial interplays of genes and environmental factors are implicated in the causal mechanisms underlying ADHD (35). Cerebral magnetic resonance imaging (MRI) studies have revealed reductions in the brain surface area and cortical gray matter volume; such reductions were associated with placental insufficiency and may be related to poor neurodevelopment and later attentional problems in SGA babies (34, 36, 37). Although white matter reduction and corpus callosum thinning were related to ADHD in adolescents born with prematurity and VLBW, this was not demonstrated in term-born adolescents with SGA (16, 38). A recent study demonstrated reductions in brain connectivity, which represent possible disruptions in the maturation of functional networks among children with ADHD (39, 40). To explore possible causal mechanisms for these findings, additional long-term follow-up studies that involve well-defined phenotypes of ADHD and combine structural or functional neuroimaging techniques for term-born children with MLBW are recommended.
The increased risk of ASD observed in our MLBW cohort is consistent with the findings of previous research that has described associations of prematurity (41) and VLBW (42) with the development of ASD. Individuals with ASD are characterized by limited social interactions, deficits in communication or language skills, restricted interests, or stereotypy (43). Although ASD is considered a neurodevelopmental disease strongly influenced by genetic factors, the exact cause is unclear and may include genetic, environmental, perinatal, maternal cervical–vaginal infections, or immunological effects (43). Placental insufficiency that leads to LBW may also cause a brain maturation issue, deficiency of neuroprotective factors, or increased inflammation, thus increasing the risk of brain injury or neurodevelopmental disorders (41). Most studies to date have focused on the associations between ASD and prematurity or VLBW. Investigations into relevant causal mechanisms underlying the association between term- born MLBW and ASD are still warranted.
Emotional disturbance is characterized by shyness, introversion, elective mutism, relationship problems, academic underachievement, or oppositional defiant disorders. We found that the MLBW cohort had an increased risk of subsequent diagnoses of emotional disturbance and conduct disorder. This finding is supported by reports that emotional and behavioral problems are more common in children born preterm with VLBW than in term-born children (44). However, studies that included children born at 32 ~ 36 weeks or full term and conducted follow-up evaluations for preschool behavioral and emotional problems have not revealed SGA as a risk for such difficulties when compared with a term-born pediatric population (45, 46). The discrepancy between our findings and those of the aforementioned studies may be that the inclusion criteria for LBW in our study is different from past research. Past research focused on incidences of emotional or behavioral problems among term babies with SGA or those born at 32 ~ 36 weeks, while our study focused on that among term-born marginally LBW infants (2000 to 2499g). Furthermore, our outcome of emotional disturbance was determined based on clinical diagnoses from insurance claims, while previous literature determined their outcomes through responses from questionnaires (45, 46). Studies have demonstrated that the prevalence of poor social skills, inadequate emotional regulation, and associated poor clinical outcomes in patients with ADHD or ASD may be as high as 70% (44, 47). Research on the probable mechanisms of these conditions should consider the cerebral pathology associated with LBW that might alter ventral striatal activation related to impulsive symptoms (48). Impairments in functional connectivity between the amygdala and the prefrontal or anterior cingulate cortex may also affect emotional processing (7, 47, 49, 50). Similarly, white matter pathology and hemodynamic abnormalities influencing brain activity associated with emotions and visual cortical regions have been revealed by MRI studies on the possible etiology of conduct disorders (7, 51, 52). Although parental demographic and child-rearing factors such as mother’s education level, family’s problem-solving abilities, child-rearing style, or parental physiological or psychological distress (7, 53, 54) may be associated with incident emotional disturbances or conduct disorders in MLBW term-born children, it may be of interest why the incidence of personality disorder was not significantly higher. The main reason may be that the diagnosis of personality disorders usually requires longer-term observations of behavioral patterns by mental health clinicians. Usually the diagnosis of personality disorder cannot be made until 18 years old because the personality is still developing before age 18 (55). It is possible that our study subjects were not old enough to be diagnosed with personality disorders during our observation period between 2008 to 2018. Further research is required to confirm the correlations of the etiology of emotional disturbances or conduct disorders with MLBW.
The use of a large and nationally representative population-based birth cohort with perinatal data allowed determination of risks without recall bias. This was a major strength of this study. The study carefully matched the control group, which enhanced the reliability of our findings. Nevertheless, the study has several limitations. First, the possibility of misclassifying psychiatric outcomes is a potential limitation inherent to population-based studies. Diagnosis may be influenced by multiple factors, eg. healthcare access, knowledge, education, family support, denial, or opportunity. Although these diagnoses were made by clinicians and may reflect actual observations, they were not obtained through psychiatric diagnostic interviews and may not be valid for research purposes (32). Second, despite our efforts to minimize differences in demographic and comorbidity variables when matching our study and comparison cohorts, there may still be potential confounders. Family history of mental illnesses, tobacco use, dietary habits, and psychosocial stressors are not available in the NHIRD and could not be controlled for. However, even with the past divide of nature versus nurture, more recent evidence has revealed the regulating roles of neuronal connectivity and neural plasticity on the control of behavioral changes and higher cognitive functions (56). Future establishment of a sibling comparison study may provide more insight. Third, although we used a large population-based sample representative of Taiwan’s population, these findings may have limited generalizability to other countries due to differing ethical considerations or national healthcare system. Fourth, despite a cohort design that enabled us to track our study population for 11 years (2008–2018), our follow-up period was not long enough to investigate incidences of mental illnesses emerging after puberty. Further long-term follow-up research using psychiatric diagnostic interviews to determine the occurrences of psychiatric disorders may help overcome these limitations.