Kari Modalsli Aaberg, MD a,b, Inger Johanne Bakken, PhD a, Morten I Lossius, MD, PhD b, Camilla Lund Søraas, MD, PhD a, Siri Eldevik Håberg, MD, PhD a, Camilla Stoltenberg, MD, PhD a,d, Pål Surén, MD, PhD a,b, Richard Chin, MD, PhD c
a The Norwegian Institute of Public Health, Oslo, Norway
b National Center for Epilepsy, Oslo University Hospital, Oslo, Norway
c Muir Maxwell Epilepsy Centre, University of Edinburgh, United Kingdom
d Department of Global Public Health and Community Care, University of Bergen, Norway
Address correspondence to
Kari Modalsli Aaberg
Norwegian Institute of Public Health
P.O. Box 4404 Nydalen
Comorbidity and Childhood Epilepsy
Financial Disclosure Statement
The authors have no financial relationships relevant to this article to disclose.
The study was funded by grants from the Research Council of Norway (grants 213699 and 221919) and the Regional Health Authority for South-East Norway (grant 2014057).
Conflict of Interest Statement
The authors have no conflicts of interest relevant to this article to disclose.
What`s Known on This Subject
Children with epilepsy are at risk of a wide range of comorbid disorders, but few risk estimates are available from population-based samples, and few studies have assessed a broad range of medical, neurological, developmental, psychiatric comorbidities in these children.
What This Study Adds
This nationwide registry-based study shows that children with epilepsy are at increased risk of almost all types of medical, neurological, developmental and psychiatric disorders. Elevated risks are observed across all types of epilepsy, even for presumably uncomplicated epilepsies.
Childhood epilepsy, comorbidity, epidemiology
ADHD Attention-deficit hyperactivity disorder
CI Confidence interval
CWE Children with epilepsy
CWE+ Complicated epilepsy: Epilepsy with additional recordings of neurological
and/or developmental disorders
CWE– Uncomplicated epilepsy: Epilepsy without additional recordings of
neurological or developmental disorders
ICD-10 International Classification of Diseases, 10th Revision
Contributors’ statement page Kari Modalsli Aaberg conceptualized and designed the study, conducted the data analyses, drafted the initial manuscript, and approved the final manuscript as submitted.
Inger Johanne Bakken contributed to the study design, the acquisition, analysis and interpretation of the data, and the drafting of the manuscript. She has approved the final manuscript as submitted.
Pål Surén and Richard Chin have contributed to the study design, the analysis and interpretation of the data, and the drafting of the manuscript. They have approved of the final manuscript as submitted.
Morten I Lossius, Camilla Lund-Søraas, Siri E Håberg and Camilla Stoltenberg have contributed to the analyses and interpretation of the data, reviewed and revised the manuscript, and approved of the final manuscript as submitted. Siri E Håberg also contributed to the acquisition of the data.
ABSTRACT Background and objectives
Children with epilepsy are at increased risk of other disorders and difficulties, preceding, co-occurring with, or following the diagnosis of epilepsy. Risk estimates vary, few studies are population-based, and few provide comprehensive assessments of comorbidities. We used nationwide registry data to describe frequencies of medical, neurological, developmental and psychiatric conditions occurring before and after children are diagnosed with childhood epilepsy.
Data were obtained from the Norwegian Patient Register which is an administrative database recording ICD-10 diagnoses from all government-funded specialist health services in Norway (outpatient consultations and hospitalizations). We included data from years 2008 through 2013 for all children born in Norway 1996-2013 (0-17 years of age at the end of follow-up). Children with epilepsy were compared to the general child population, adjusting for sex and age. We also compared children with complicated epilepsies, i.e., epilepsies with additional neurological and/or developmental disorders, to children with uncomplicated epilepsies.
The study population included 1,125,161 children. There were 6635 (0.6%) children with epilepsy. Nearly 80% of children with epilepsy had one or more comorbid disorders. All types of disorders were more frequent in children with epilepsy, with medical disorders recorded in 55%, additional neurological disorders in 41%, and developmental/psychiatric disorders in 43%. Children with complicated epilepsies had the highest overall levels of comorbidity, but the risk of medical and psychiatric comorbidities was also substantial among children with uncomplicated epilepsies.
The overall frequency of comorbid disease is high in children with epilepsy, including children with presumably uncomplicated epilepsies.
There is an increasing focus on comorbid disorders in people with epilepsy, and new definition proposals have sought to frame epilepsy as not just a seizure disorder, but as a disorder with a wide range of neurobiological, cognitive, psychological and social aspects.1
Comorbid disorders may share causes or risk factors with epilepsy, or even be the actual cause of epilepsy. They may also be consequences of seizures, epileptic activity or antiepileptic treatment.2-4 Consequently, these other conditions may precede, co-occur with, or follow the diagnosis of epilepsy.2 Comorbid disorders contribute to the disease burden experienced by patients and their families, and influence their quality of life and long-term outcome.2-11 Most studies of comorbidity in childhood epilepsy have focused on neurocognitive, behavioral, social and psychiatric disorders or difficulties.12-21 Knowledge about other medical comorbidities is limited, and only a couple of studies have investigated this in children specifically.22,23 Most studies of medical comorbidities have included only adults or subjects of all ages, and primarily reported findings in adults.24-34 Some studies have focused on specific diagnoses rather than an extensive range.35-38 For all categories of comorbid conditions, the prevalence estimates vary widely, depending on the study design, methodology and population under study.5,39,40 Only a few studies of children with epilepsy (hereafter referred to as CWE) have been able to compare with the general population.12,14,20,21,23,41-43 To improve our knowledge about comorbidities in CWE, we have used nationwide registry data to:
Estimate the proportions of medical, neurological, developmental and psychiatric disorders in CWE compared with the general child population.
Compare comorbidity patterns in CWE with and without additional neurological and/or developmental disorders.
Examine differences in comorbidity patterns by age and sex in CWE.
We have defined comorbidity as the co-occurrence of conditions in the same individual irrespective of temporal or causal relations, as proposed by Feinstein in 1970.44
POPULATION, PATIENTS AND METHODS Data and population
The study is based on the Norwegian Patient Register (NPR), an administrative database containing data from all hospitals and outpatient clinics owned and/or reimbursed by the Norwegian government. Reporting to the NPR is mandatory. Diagnoses are coded by physicians according to the International Classification of Diseases – 10th Revision (ICD-10).45 Individual-level research data are available from 2008 onwards.
We included data from 2008-2013 on children born 1996-2013 (0-17 years at the end of follow-up). Epilepsy was defined as having two or more registrations with ICD-10 codes G40 and/or G41. The NPR provided all ICD-10 codes registered in the study period in CWE, plus information about sex and year of birth. To enable comparisons with the general child population, the NPR provided an additional anonymized data file which included ICD-10 codes, sex and year of birth for all Norwegian children born in 1996-2013, including CWE. The total number of individuals in the population, by sex and year of birth, were obtained from Statistics Norway (www.ssb.no).
We included data from all government-owned health services. Private practices receiving government reimbursements are also obliged to report to the NPR, but reporting was incomplete before 2013, and we did not include data from private practices in our study. The NPR data from 2013 showed that 24 CWE only had epilepsy diagnoses recorded in private practices that year. Consequently, the number of missed diagnoses is low and unlikely to affect the overall estimates of comorbidity in CWE.
Our aim was to capture and categorize all comorbid conditions that were chronic or long-lasting and likely to have a significant effect on overall health and quality of life. Transient conditions (e.g. infectious episodes) and conditions unlikely to occur in children (e.g. varices) were excluded. Comorbid conditions were divided into three main categories – medical, neurological and developmental/psychiatric – and further subdivided into disease categories based on the ICD-10. The study was not designed to investigate causal relations, and some categories contain comorbid diagnoses that are potential causes of epilepsy, such as brain neoplasms and malformations, metabolic disorders, and chromosomal abnormalities.
The exact codes included in each category are listed in Supplementary Table 1. In general, the categories correspond to the individual ICD-10 blocks, but in some cases we added codes from the R block (“Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified”). Norwegian physicians often use these codes when diagnostic conclusions have not yet been made. For example, ICD-10 code R62 (“Lack of expected normal physiological development”) is often used to denote significant developmental delay, which is why we have included R62 in our definition of developmental disorders.
To study how comorbidity patterns vary by the complexity of the epilepsy, we divided CWE into two subgroups based on presence or absence of comorbid neurological and/or developmental disorders. Complicated epilepsy, hereafter referred to as CWE+, included CWE with any additional diagnoses of neurological disorders, intellectual disability (F70-79), autism or other disorders of psychological development (F80-89), and/or lack of expected normal physiological development (R62). Uncomplicated epilepsy, hereafter referred to as CWE–, included CWE with no such additional diagnoses.
We used SPSS version 22 (SPSS Inc.) and Stata version 13.1 (StataCorp). CWE were compared with the general child population by logistic regression analyses adjusted for sex and year of birth. Because of the large number of comparisons conducted, we used 99% confidence intervals (CIs) for the odds ratios (ORs).
The study was approved by the Regional Committee for Medical and Health Research Ethics for South-East Norway, ref. 2010/2583. The study data were derived from health registries mandated by Norwegian law, for which individual consent is not required.
The study population included 1,125,161 children, of whom 9215 were registered with a diagnosis of epilepsy. Of these 9215, 6635 had two or more recordings of G40/41 and were defined as having epilepsy. The proportions increased by age to a maximum of 0.9% in 14-year-olds (Figure 1). For the study population as a whole, the proportion of CWE was 0.6%. The mean age at end of follow-up was higher in CWE than in the general child population (11.7 versus 9.1 years), and the proportion of boys was higher (54.3% versus 51.2%).
Overall frequencies of comorbidity
Overall, 78.3% of CWE had one or more comorbid disorders recorded, whereas the similar proportion was 30.3% in the general child population. Multiple comorbidities were common, and 13.4% of CWE had diagnoses within all three main categories of comorbidity (medical, neurological and developmental/psychiatric).
All medical conditions were more frequent in CWE than in the general child population (Table 1). The most frequent, both in CWE and general child population, were gastrointestinal disorders (most commonly constipation and gastro-oesophageal reflux), which were recorded in 19.1% of CWE versus 5.4% in the general child population. Other frequent disease categories in CWE were congenital malformations outside of the central nervous system (18.2%), musculoskeletal disorders (15.3%), chronic lower respiratory disorders – mainly asthma (10.3%), and malnutrition and/or eating difficulties (10.1%). The largest relative increases in CWE (the highest ORs) were observed for visual impairments (OR=30.6), chromosomal abnormalities (OR=19.6), malnutrition and/or eating difficulties (OR=16.1), sleep disorders (OR=13.0), immune disorders (OR=8.6), nutritional deficiencies (OR=7.4) and metabolic disorders (OR=7.3).
For neurological disorders (Table 1), the most frequent in CWE were cerebral palsy (13.9%), headache conditions (6.6%) and congenital neurological malformations (6.5%). Both the absolute and the relative increases were large, with ORs >25 for all types of neurological disorders except headache conditions.
Developmental and/or psychiatric disorders were registered in 42.9% of CWE overall, compared to 6.6% in the general population (Table 1). The relative increase in CWE was particularly large for developmental disorders, with intellectual disability in 17.0% (OR=41.0), disorders of psychological development in 21.3% (OR=11.6), and unspecified developmental delay (R62) in 7.5% (OR=8.2). Of the disorders of psychological development, autism accounted for 7.8% (OR=10.7). Attention-deficit hyperactivity disorder (ADHD) was the most common psychiatric diagnosis in CWE, occurring in 12.1% (OR=5.4). There were few cases of comorbid anxiety or depression in CWE, probably because emotional disturbances in children are mostly coded under the ICD-10 section for childhood behavioural and emotional disorders (F90-98). A total of 10.5% of CWE had diagnoses from this section (OR=3.6).
Complicated versus uncomplicated epilepsy
Of the 6635 CWE, 3883 (58.5%) were defined as CWE+ and 2752 (41.5%) as CWE–. In general, all medical and psychiatric conditions were more frequent in CWE+ than in CWE–, and more frequent in CWE– than in the general child population (Table 2).
Medical disorders were recorded in 65.9% of CWE+ and 38.8% of CWE–. ORs for CWE+ versus CWE– are not included in Table 2, but the largest relative differences between CWE+ and CWE– were observed for visual impairments (OR=14.3), malnutrition/eating difficulties (OR=13.7) and chromosomal abnormalities (OR= 9.8). Psychiatric disorders were also common in both groups, with the highest proportions in CWE+.
Most of the medical conditions were more commonly registered in the young CWE, except some conditions that were fairly evenly distributed across age groups, such as neoplasms, endocrine, urinary and genital conditions, or increasing by age, such as obesity and musculoskeletal conditions. Neurological conditions were also more common in the youngest children, except for headache conditions, which increased by age, and cerebral palsy and cerebral neoplasms, which were fairly stable across age groups. These age differences probably reflect a genuine decline in medical and neurological comorbidity by age, but there is also likely to be some under-ascertainment of medical and neurological disorders among the older children, due to the lack of data prior to 2008.
The proportions of CWE with developmental disorders in total were relatively stable across ages. The unspecified developmental delay diagnoses (R62) appear to be replaced by specific diagnoses as the children grow older, as shown in Figure 2. For psychiatric diagnoses, the proportions increased by age.
Proportions of comorbid conditions in CWE distributed by sex are shown in Supplementary Table 2. We found no sex differences in CWE that were not apparent in the general child population as well. Both among CWE and other children, boys had an increased risk of most developmental disorders and ADHD, whereas girls had an increased risk of depression.
The aim of this study was to assess comorbidities in children with epilepsy. Our main findings do not pertain to specific disorders, but the overall burden of disease and the patterns of comorbidity observed. Nearly 80% of CWE had one or more comorbid disorders, and multiple comorbidities were common. These consisted not only of additional disorders affecting brain development and functioning, but a wide range of medical and psychiatric conditions. The increase in disease risk also included the presumably uncomplicated epilepsies.
The increase in risk of disorders originating in the brain, i.e., neurological, developmental and psychiatric conditions, is well known in CWE. Our findings for those conditions are in line with previous studies.12-14,21,22,46-48 Some of the comorbid neurological, chromosomal and metabolic disorders are likely to represent causes of epilepsy. The high proportions of medical disorders were a more surprising finding. Some of the specific medical diagnoses, such as asthma, have been investigated in other studies,22,23 and found to be increased in CWE. However, few previous studies have assessed a broad range of medical disorders in CWE, so we will focus on this in the remainder of the discussion.
Gastrointestinal disorders were the most frequent type of comorbid disorders. These were mostly treatable conditions such as constipation and gastrointestinal reflux occurring among younger CWE. More alarming were the high proportions of malnutrition and eating difficulties, which were found in 1 out of 6 CWE overall, and increased in both CWE+ and CWE–. Nutritional deficiencies were also considerably more frequent in CWE relative to the general child population. An increased risk of nutritional deficiencies in epilepsy has been demonstrated in a study that included subjects of all ages,31 and our findings support that this is a considerable problem in CWE. In many cases, particularly for CWE+, nutritional difficulties are likely to result from the underlying cause of epilepsy, such as chromosomal disorders or cerebral palsy, rather than the epilepsy itself.
Sleep disorders were registered in 2.9% of CWE, which is 13 times more often than in the general child population. Sleep disorders are often not recorded, especially in children who are not in regular contact with specialist health services. There is likely to be considerable under-reporting of sleep disorders in our data, and more so for children without epilepsy. Sleep disturbances may influence the outcome of the epilepsy, as well as the health-related quality of life and the psychological functioning of the patients,35,49-54 and it has been shown that diagnosing and improving sleep may have a positive impact in CWE.35,51-56 The relative risk of visual impairments in CWE was high, especially in CWE+. This has also been shown in other studies, with proportions ranging from 3.4% to 5%.31,48,57 Our proportion of 1.9% was lower, possibly because not all children needing glasses are examined by ophtalmologists, and thereby not captured by the registry.58 Although the absolute risk is low, the finding of a high relative risk of immune disorders in CWE is interesting, because the interaction between epilepsy and the immune system is currently debated, both with regards to the pathogenesis of epilepsy59-61 and antiepileptic treatment.8,62 It is known that antiepileptic drugs may affect the immune system,63,64 and immune-modulating therapies are used in treatment.65,66 Some rare forms of epilepsy are even caused by neuronal autoantibodies.62,65 In general, the disorders and frequencies of comorbid medical conditions in CWE were different from those observed in studies of adults with epilepsy.2,24-28 However, there is no reason to believe that the consequences are very different. Medical comorbidities have a considerable negative impact on health-related quality of life in adults with epilepsy.27,32 The findings of increased levels of comorbidity in CWE+ relative to CWE– is in concordance with previous studies,12,13,67,68 but it was surprising to find such high levels of medical and psychiatric disorders in CWE– relative to the general child population. Our definition of uncomplicated epilepsy excluded those with any type of additional neurological disorders or developmental delay, which was strict compared with other studies.12,13,68 Despite this, there was still a substantial increase in risk of most types of comorbidities in CWE–, indicating that uncomplicated epilepsies are often more complex than the term “uncomplicated” suggests. This general increase in comorbidity in CWE– may relate to various aspects of epilepsy: the seizures, the epileptic activity, the underlying cause, the treatment, and the burden of living with a chronic disease.
Strengths and limitations
The strengths of our study are the assessment of a broad range of comorbid disorders and the size and completeness of the data set. The large numbers allowed us to assess the risk of both rare and common disorders, and the inclusion of the whole child population largely eliminates selection bias, making our findings generalizable to other developed countries as well.
A limitation is the lack of validity data for the epilepsy diagnoses, as well as the other recorded diagnoses. However, previous studies of epilepsy in Norway, based on data from single counties, have found positive predictive values of 80%69 and 74%70 for registered diagnoses of epilepsy, indicating a relatively high quality of the data. Another study found a high quality of autism diagnoses in the NPR, with a positive predictive value of 94% (95% CI 79 to 99%).46 In this study, we chose to err on the side of caution by restricting the CWE definition to those with two or more recordings of epilepsy to avoid including children with erroneously recorded epilepsy diagnoses. This may have caused us to lose some true cases of epilepsy, but the overall proportion of epilepsy observed using our definition (0.6%) is in line with estimates from Norway and other countries, which supports our choice of restriction.48,57,67,69,71-74 Another limitation is that the proportions of some disorders tend to be underestimated in registry studies.58 Studies conducting in-person assessments of CWE indicate that psychiatric disorders are often under-diagnosed. For example, Reilly et al. found similar proportions to our study of previously diagnosed ADHD, autism, depression and anxiety, but the proportions increased considerably after in-person assessments had been conducted.19 Other in-person studies have found high proportions of psychiatric disorders in CWE, and that CWE have unmet psychiatric needs.17,18,75,76 Under-diagnosis of psychiatric disorders is likely to affect the general population, not just CWE. Obesity also appears to be substantially underdiagnosed compared with other studies, both in CWE and the general child population.23,77-79 On the other hand, the relative differences between CWE and the general child population may be somewhat inflated since CWE are in regular contact with specialist health services. This regular contact increases the chance of diagnosing any type of comorbid disorder, a type of ascertainment bias often referred to as Berkson’s bias.80 CWE are also likely to undergo thorough investigations of potential causes of epilepsy, which also increase the chances of discovering comorbid disorders. In general, our data are likely to be most accurate for conditions that are serious enough to warrant specialist treatment and follow-up, both in CWE and general population, while less severe conditions and difficulties, such as headache conditions and sleep disorders, are more likely to be captured in CWE than in the general child population.
The lack of data from the NPR prior to 2008 is also a limitation. As described in the Results section, we are likely to have missed some comorbid conditions among the oldest children. This would apply to conditions that children can grow out of, such as gastro-oesophageal reflux and asthma. A data set covering all years since birth for the whole study population would have provided better opportunities to study temporal relations between epilepsy and comorbid conditions.