Introduction
Neurodevelopmental disorder (NDD) is a term used in the diagnostic and statistical manual of mental disorders, fifth edition (AmericanPsychiatricAssociation
2013) to describe a range of disorders with symptom onset during childhood. NDDs include—among other disorders—Autism spectrum disorder (ASD), which is characterized by impairment in social communication and inflexibility in interests, behavior and imagination (Lai et al.
2014), attention-deficit/hyperactivity disorder (ADHD), defined as problems with short attention span, impulsivity and/or hyperactivity (Pelham et al.
2005); and learning disorders (LD), most frequently displayed through symptoms of poor reading and/or writing ability (AmericanPsychiatricAssociation
2013).
The different NDDs share similar cognitive and behavioral processes, and frequently overlap with each other (and with other mental health disorders), which makes a dimensional approach to disorder classification useful (Anckarsater
2010). With this in mind, ESSENCE (early symptomatic syndromes eliciting neurodevelopmental clinical examinations) was introduced (Carlsson et al.
2013; Gillberg
2010) for the purpose of highlighting the significant amount of overlap across NDDs and the similarities of their symptoms mainly during young ages.
Recent research has shown a rise in the number of children clinically diagnosed with NDDs (Boyle et al.
2011) and that these children often have coexisting physical problems (Schieve et al.
2012). Until now, most studies have focused on physical problems co-occurring with either ASD or ADHD, such as the association between autism and epilepsy, which has been documented in many studies (Amiet et al.
2008; Jokiranta et al.
2014; Reilly et al.
2014). Besides epilepsy, the increased prevalence of gastrointestinal (GI) problems (Chaidez et al.
2014; McElhanon et al.
2014; Valicenti-McDermott et al.
2006), eating problems (Rastam et al.
2013), diarrhea and constipation (Chaidez et al.
2014), headache (Parisi et al.
2014) and asthma (Kotey et al.
2014) has been shown in children with ASD. A potential link between celiac disease and ASD has been investigated, but in most studies no such association has been found (Batista et al.
2012; Black et al.
2002; Lau et al.
2013). A case-control study reported elevated prevalence of specific immune-related comorbidities in ASD compared to controls; allergy, various autoimmune diseases and psoriasis (Zerbo et al.
2015). With regards to a link between type 1 diabetes mellitus (T1DM) and ASD, the results are diverse. One study from Finland could not find any increased occurrence of ASD among 5178 children with T1DM (Harjutsalo and Tuomilehto
2006), whereas a retrospective descriptive US study in young adults with ASD found a small but statistically significant increase in the prevalence of T1DM in ASD patients compared to general hospital patients (Kohane et al.
2012). A 35% increased risk of children with ASD to have asthma as compared to children in the general population was established in a large cross-sectional study in the United States (Kotey et al.
2014).
Some studies have shown increased prevalence of epilepsy (Cohen et al.
2013), migraine (Fasmer et al.
2011) and asthma (Fasmer et al.
2011) in patients diagnosed with ADHD, and vice versa increased prevalence of ADHD in clinical populations of children with somatic problems. For example, in children with asthma a significantly higher risk of developing ADHD during the school years has been detected compared to children without asthma (Chen et al.
2013). A recent Swedish study found that presence of ADHD made treatment of T1DM in teenagers a challenge (Lindblad et al.
2017) and a large epidemiological study found ADHD to be almost five times as common in children with enuresis as compared to children without (von Gontard et al.
2011). Using Norwegian national registries, a cross-sectional study found a significant association between ADHD and psoriasis, especially in females (Hegvik et al.
2018). A recent review on somatic comorbidity among adults with ADHD found that asthma is a well-documented comorbidity among adults with ADHD while the number of studies looking at the associations between ADHD and epilepsy, migraine, GI disorders and enuresis are still very few, ranging from one to three studies each (Instanes et al.
2018). A recent familial association case-control study of 117 children and adolescents with DSM-5 ADHD reported that a headache disorder diagnosis was common for both patients and healthy controls (59.0% vs. 37.8%), and migraine was found in 26.0% of ADHD cases and 9.9% of healthy controls (Kutuk et al.
2018). A positive association between the prescription of anti-migraine and ADHD medications and higher prevalence of migraine was found among adults with ADHD in Norway (Fasmer et al.
2012). One clinical study reported an overrepresentation of ADHD symptoms in patients with celiac disease (Niederhofer and Pittschieler
2006) and another study of patients with ADHD found an overrepresentation of celiac disease identified by the presence of celiac disease specific antibodies (Niederhofer
2011). In contrast, a large controlled study could not confirm the association between celiac disease and ADHD in children (Gungor et al.
2013).
There are few studies investigating physical health issues in children with LD or to what extent LD may exist in children with physical diseases (Schieve et al.
2012). One population-based study in Taiwan showed that the most common comorbid NDD among children with epilepsy was LD with a prevalence of 13% (Chiang and Cheng
2014). A Swedish study failed to show that children with T1DM had impaired academic ability (McCarthy et al.
2002).
In 2011, a population-based study was published investigating coexisting medical conditions in children with diagnoses of autism and/or ADHD (Schieve et al.
2012). This study found a moderate to high increase in the rate of most medical conditions examined in children with these NDDs. Phenotypical overlaps cannot be interpreted as causation or shared etiology per se, but a number of medical or genetic syndromes are known to cause physical problems and NDDs alike. Children with tuberous sclerosis have higher prevalence of ASD and learning disabilities (Mitchell et al.
2017) and studies have shown an association between early exposure of cranial radiation or chemotherapy and later development of neurocognitive problems (Anderson and Kunin-Batson
2009; Edelstein et al.
2011), but still the field of NDD in cancer patients is relatively unexplored. Even though comorbidity of NDDs is the rule rather than the exception (Gillberg
2010), no study has of yet specifically investigated the physical health of children with multiple overlapping NDDs, nor does a comprehensive overview of these associations exist.
In the present study, we used a Swedish nationwide general population cohort of twins to examine (a) the relationship between any “single” NDD and a variety of physical problems (epilepsy, migraine, asthma, cancer, diabetes, psoriasis, lactose intolerance, celiac disease, diarrhea, constipation, daytime enuresis, encopresis) and (b) if various constellations of NDDs are differently associated with these defined physical problems. The study provides a map of phenotypical associations between NDDs and physical problems, but does not assess underlying etiological factors, such as gestational age, birth weight or genetic mechanisms.
Discussion
In this study, we found increased rates of physical problems in Swedish twins screening positive for NDDs as compared to controls. In particular, epilepsy, migraine, asthma, diarrhea, constipation, daytime enuresis and encopresis were overrepresented in children who were screen-positive for NDDs. These results are generally in line with previous studies finding increased rates of physical problems in children with NDDs (Amiet et al.
2008; Batista et al.
2012; Kohane et al.
2012; McKeown et al.
2013; Schieve et al.
2012). Our study advances the field showing that children with different constellations of NDDs more often had coexisting physical problems as compared to children with a single NDD. Moreover, our study indicates that some specific combinations carry higher risk of specific physical problems, which has been difficult to assess by the existing literature that has mostly examined the relationship between physical problems and specific NDDs.
Recently, the well-documented association between childhood epilepsy and ASD has been confirmed using the same twin study population as the present study (Gillberg et al.
2017). However, our results indicate that children with LD and especially children with both ASD and LD have the highest prevalence of epilepsy (OR 53), and that as many as 1/5 of them may be affected in childhood. The association between ASD and epilepsy may be partially mediated by the presence of LD or intellectual disability, also supported by previous research (Amiet et al.
2008).
Migraine and ASD include some symptomatic similarities, e.g. many children with ASD have sensory hypersensitivity (Sullivan et al.
2014), many migraine patients report symptoms such as photophobia during headache attacks (Hansen et al.
2015), and a recent study provides some preliminary evidence for a link between migraine, sensory hypersensitivity and anxiety in ASD (Sullivan et al.
2014). In the present study we found significant increase in the prevalence of migraine in children with ASD, ADHD and their combination as compared to the prevalence found in children without any NDD. It has been reported that some anti-migraine medications (which act through increasing serotonergic activity) can improve repetitive behavioral symptoms of autism (Hollander et al.
2000). Considering the overlap of migraine with ADHD, it has previously been suggested that migraine is not comorbid with ADHD in children, but with their hyperactive-impulsive behaviors (Arruda et al.
2010). In adult ADHD patients, the risk of migraine is increased by about 70% (Fasmer, Halmoy, Oedegaard, et al., 2011) which is a very similar increase to what was found here (OR 1.8). Also, the prescription pattern for anti-migraine medications and drugs for the treatment of ADHD in adults seems to indicate a link between migraine and ADHD (Fasmer et al.
2012).
Asthma was found to be very common among children with NDDs in our study, being most prevalent in children with both ADHD and LD where nearly one-third reported asthma. An American national survey on children below the age of 17 with a diagnosis of asthma, found, in line with our results, that asthma was independently associated with ADHD and LD (Blackman and Gurka
2007). Moreover, a study by Chen and co-workers indicated that asthma in early life among Taiwanese children predispose for ADHD later (Chen et al.
2013). The converse may also be true as studies indicate that ADHD is affected by regulation of the autonomic nervous system (Ward et al.
2015).
The association between cancer and central nervous system disorders has been explored in several population-based studies (Crawley et al.
2016; Driver et al.
2012; Ji et al.
2013). In our study, 1.3% of children with “any NDD” (as compared to 0.3% of children from the comparison group) had a parent-reported cancer history. Our result should be interpreted with caution as the number of individuals reporting cancer was small and that only living children were included. The link between cancer and NDD might speculatively be mediated through the various cancer treatments such as cranial radiation (Edelstein et al.
2011), even though we did not specifically monitor such treatments. A review by Anderson and co-workers indicated that even chemotherapy alone (without CNS irradiation) during childhood may have subtle long-term effects on attention and executive functioning later in life (Anderson and Kunin-Batson
2009). There may also be shared etiological factors, such as increased prevalence of mutations in PTEN (phosphatase and tensin homolog) that has been observed in ASD (McBride et al.
2010). PTEN is normally involved in regulating signaling pathways in cell proliferation and apoptosis (Worby and Dixon
2014) and mutations in PTEN could therefore theoretically increase the risk of cancer. In addition, it is also well known that many children with tuberous sclerosis (a proliferative genetic condition) develop ASD and/or have LD and a range of other developmental problems often conceptualized as “tuberous sclerosis-associated neuropsychiatric disorders” (de Vries et al.
2015). Similarly, in our study we found increased prevalence of cancer among children with ASD (1.1%) or LD (1.4%).
As previously mentioned, research has shown a link between maternal inflammatory conditions during pregnancy and offspring risk for NDDs (Ginsberg et al.
2017), both in mouse models (Mottahedin et al.
2017) and in clinical settings (Goines et al.
2011; Jiang et al.
2016). One study showed an association between maternal and paternal history of T1DM and ASD (Atladottir et al.
2009), while other studies indicate a possible link between autism and T1DM where 0.9% of children with T1DM in a tertiary hospital were also diagnosed with autism (Freeman et al.
2005). In our study, 0.4% of the comparison group (consisting of 22,028 children) screened positive for diabetes, and the prevalence of diabetes was significantly increased in children with NDD, especially ADHD. Studies have also shown significantly higher occurrence of autoimmune diseases (mainly hypothyroidism and rheumatic fever) in relatives of children with ASD, even higher than among family members to children with autoimmune disease (Sweeten et al.
2003). In a Swedish nationwide register-based cohort study in Sweden (Butwicka et al.
2017), children with biopsy-verified diagnosis of celiac disease had a 1.4-fold greater risk for mood disorders, anxiety disorders, eating disorders, behavioral disorders, ADHD, ASD, and intellectual disability compared to the general population. In contrast, siblings of celiac disease probands were at no increased risk for any of the investigated psychiatric disorders. These results are in alignment with our findings that suggest a possible link between autoimmune disorders such as celiac disease and NDDs.
Our findings included the strong positive association found between NDDs and GI problems, that is in line with a recent review on the association between ASD and gut microbiota (Li et al.
2017). Research indicates that the gut is physiologically related to the brain through the vagal nerve and the enteric nervous system, and that increases in the sympathetic tone, such as stress or anxiety, may decrease bowel movement in humans. Previous research has found a partial relationship between GI problems and higher level of anxiety (Mazefsky et al.
2014), sensory hyperactivity among children with NDDs (Mazurek et al.
2013), as well as increased heart rate variability and GI symptoms in children with ASD compared to controls (Ferguson et al.
2016). Whether these hypotheses can be extended to ADHD and/or LD remains to be better elucidated, but some results such as those presented here indicate that they probably can be.
The prevalence of daytime enuresis and encopresis in the present study was up to 50% in the NDD groups. The treatment of enuresis and encopresis in children with ASD, with or without LD, is very important (von Gontard et al.
2015). It has been shown that multimodal and individualized treatment in children with ADHD can treat the voiding symptoms (Kaye and Palmer
2010). One study indicated that children with ADHD and enuresis are less likely than children without ADHD to be brought to medical attention for their voiding problems (Shreeram et al.
2009). It is important to actively ask about incontinence because many children and/or their families may not easily report such problems. Many children with LD or autism also have impaired communicative abilities, which may further contribute to the high prevalence of enuresis and encopresis that we found in our study. We know that daytime enuresis and encopresis are common among school children with a prevalence of daytime enuresis among Swedish 10 year-olds estimated to be around 4%, while that for fecal incontinence/soiling is 7% in boys and 4% in girls (based on parental questionnaires), (Soderstrom et al.
2004). The prevalence of enuresis and encopresis reported in previous studies are highly variable due to different definitions and ascertainment (Peeters et al.
2013; Simonoff et al.
2008). Estimated prevalences of enuresis range from 5 to 43% in children with ADHD (Gor et al.
2012; Niemczyk et al.
2015; Robson et al.
1997; Shreeram et al.
2009) and from 11 to 55% among children with ASD (Gor et al.
2012; Simonoff et al.
2008; Taira et al.
1998). One study reported that almost one-third of pediatric patients with functional defecation disorder in a specialized outpatient setting had concomitant ASD symptoms (Peeters et al.
2013). Compared to prevalence estimates shown so far, our results are generally in line with earlier research and also indicate an increased risk for incontinence if the child has several NDDs.