Febrile convulsions increase risk of Tourette syndrome
Yi-Fang Tu
a,d, Cheng-Li Lin
b,c, Chih-Hao Lin
d, Chao-Chin Huang
a,
Fung-Chang Sung
b,c, Chia-Huang Kao
e,f,*
a Department of Pediatrics, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University,
Tainan, Taiwan
b Management Office for Health Data, China Medical University Hospital, Taichung, Taiwan
c Department of Public Health, China Medical University, Taichung, Taiwan
d Department of Emergency Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung
University, Tainan, Taiwan
e Graduate Institute of Clinical Medical Science and School of Medicine, College of Medicine, China Medical University,
Taichung, Taiwan
f Department of Nuclear Medicine and PET Center, China Medical University Hospital, Taichung, Taiwan
1. Introduction
Febrile convulsion (FC) is the most common seizures in infants and children between the ages of 6 months and 6 years. Accumulated incidence of FC is 2–5% of children by the age of 5 years, with a slight predominance in boys.1,2 Several large longterm prospective studies have shown that the majority of patients with FC have a benign outcome with normally developing intelligence and neurodevelopment.3,4 Only the subtype of FC that are prolonged or recurrent may be associated with an increased risk of selected cognitive deficits such as delayed vocabulary development or learning difficulties.4–7 However, none of these studies have mentioned an association between FC and Tourette syndrome (TS).
Tourette syndrome, a common neuropsychiatric disorder in children, is characterized by multiple brief, stereotypical, and nonrhythmic movements and sounds.8 These clinical symptoms of TS last at least 1 year and often waxes and wanes. The overall prevalence of TS is 0.77% in children, with a strong preponderance in boys.9 The mean age of TS onset is approximately 6–7 years, which is older than the onset of FC. Ninety-three percent of TS patients are symptomatic by age 10.9 By late adolescence or young adulthood, over one third of TS patients are virtually symptomfree, less than half have minimal to mild symptoms and less than a quarter have persistently moderate to severe symptoms.8 Most TS
children have normal intelligence but commonly associate with
other neuropsychiatric comorbidities, generally comprising attention-deficit hyperactivity disorder (ADHD) and obsessivecompulsive
disorder (OCD).9 Similar to FC, TS children are common in boys, have paroxysmal symptoms and can expect a clinical remission over time. Thus, we suspect these two diseases might have a clinical or consequent association. We used a national population data bank to explore it.
2. Methods 2.1. Data sources
We obtained data from the National Health Insurance Research Database (NHIRD) and the National Health Research Institutes (NHRI) for the Taiwanese National Health Insurance (NHI) program. The NHI program was launched in March 1995, and provides health care to 99% of the 23.74 million residents in Taiwan. The Longitudinal Health Insurance Database contained all of the original claims data of 1,000,000 individuals randomly sampled from the Registry for Beneficiaries of the NHIRD in 2000 (LHID2000), which maintained the registration data of every NHI program beneficiary during the period 1996–2000. There was no significant different in gender, age or health care costs between cohorts in LHID2000 and all insurance enrollees, as reported by the NHRI in Taiwan. The NHRI followed all of the aforementioned randomly sampled patients to the year 2010. Details of this population-based database have been described elsewhere.10 The patient identification numbers that were previously used for linking files were scrambled before the dataset was released for
research and administrative purposes. Limited data on sociodemographic status, such as sex, birth date, occupation, income
level for insurance fee estimates, and area of residence, were also made available. The accuracy and high validity of diagnosis in the NHIRD has been previously demonstrated.11,12 This study was exempted by the Institutional Review Board (CMU-REC-101-012). The International Classification of Disease, 9th Revision, Clinical Modification (ICD-9-CM) was used to identify health statuses. 2.2. Study design and patients
In this study, we selected infants and children aged 6 months to 6 years, who were diagnosed with FC (ICD-9-CM codes 780.3
excludes 780.33) between 1998 and 2010 as the study cohort. The primary purpose of this study was to determine the association between FC and TS. Hence, we excluded infants and children with a diagnosis of meningitis or encephalitis (ICD-9-CM codes 320-324, 047-049, 062-064), CNS malformation (ICD-9-CM codes 740-742; 330-337), and epilepsy (ICD-9-CM codes 345). The index date for infants and children with FC was the date of their first medical visit. For the comparison cohort, we used a simple random sampling method and selected 4 infants or children without a history of FC before the index date, who were excluded according to the same criteria as those of the study cohort. The FC patients and controls were matched for age (every 1 year), sex, urbanization level, parental occupation, and index year. Apart from common demographic factors, their associated comorbidities were also considered as potential confounding factors between FC and TS. These comorbidities included ADHD (ICD-9-CM codes 314), OCD (ICD-9-CM codes 300.3) and anxiety (ICD-9-CM codes 300.00). Taiwan launched a national health insurance (NHI) in 1995, operated by a single-buyer, the government. All insurance claims should be scrutinized by medical reimbursement specialists and peer review. According to the revised fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IVTR), TS may be diagnosed when a person exhibits both multiple motor and one or more vocal tics over the period of a year, with no more than three consecutive tic-free months.13 Person-years of follow-up were calculated for each infant and child until TS (ICD-9-CM codes 307.2) was diagnosed, death occurred, the infant or child withdrew from the insurance system, or until the end of 2010.
2.3. Variables of interest
The socio-demographic variables used in this study comprised age, sex, urbanization level, and parental occupation. White-collar workers were defined as people with occupations characterized by long indoor work hours such as institutional workers, and business and industrial administration personnel. Blue-collar workers were defined as people with occupations characterized by long outdoor work hours, such as fishermen, farmers and industrial laborers. Other occupations included primarily retired, unemployed, or low
income populations. The level of urbanization was divided into 4 levels based on the NHRI report (Level 1 was the highest level of urbanization and Level 7 was the lowest. Because few people lived in Levels 5–7, we grouped the least urbanized populations into Level 5). Another variable was the frequency of FC-related medical visits, which were counted when patients visited outpatient/ inpatient or emergency department, and had FC in one of the discharge diagnosis. The average frequency was counted per 5 years in the study period.
2.4. Statistical analysis
All analyses were performed using the statistical package SAS for Windows (Version 9.2, SAS Institute Inc., Carey, NC). The Kaplan–Meier method and the log-rank test were performed using R software (R Foundation for Statistical Computing, Vienna, Austria). A two-tailed P value <0.05 was considered statistically significant. Distributions of categorical and socio-demographic factors, including sex, urbanization level, and occupation, were compared between the FC and the non-FC cohorts. The differences were determined using the chi-square test or Fisher-exact test. The mean ages were measured and examined using a Student’s t-test. The follow-up time (in person-years) was used to estimate
incidence rates according to demographic status, urbanization level, and occupation. A Cox proportional hazard regression was used to compare the risk of developing FC-associated TS, in the FC cohort and non-FC cohort. The hazard ratio (HR) and 95%
confidence interval (CI) were estimated using the Cox model. Adjusted hazard ratios were also determined after adjusting for potential confounders. This model was also used to assess the association between TS and the frequency of FC-related medical visits. The Kaplan–Meier method was used to plot the cumulative incidence of TS, and a log-rank test was used to compare the cohorts.
3. Results
In total, 1586 FC patients from the NHIRD met our criteria as the FC cohort and 6344 non-FC patients matched based on age, sex, urbanization level and parental occupation were randomly sampled as a non-FC comparison cohort (Table 1). The mean age of the FC and non-FC cohorts were 2.50 _ 1.34 and 2.52 _ 1.36
years, respectively (Table 1). The FC cohort was associated higher incidence of the comorbidities, such as ADHD (P < 0.0001), OCD (P < 0.01) and anxiety (P < 0.0001).
In the FC cohort, TS occurred primarily in boys (9028/15,434, 58.5%), children living in urban areas (8766/15,434, 56.8%), and children with white-collar parents (9219/15,434, 59.7%; Table 2). The overall incidence of TS was higher in the FC cohort than in the non-FC cohort after adjusted for confounders of ADHD, OCD and
anxiety (28.5 vs 13.9 per 10,000 person-years; adjusted HR = 1.91, 95% CI = 1.32–2.75; Table 2). The Kaplan–Meier analysis showed
that, during the 13-year study period, the overall TS rate was 1.5% higher in the FC cohort than that in non-FC cohort (log-rank P < 0.001, Fig. 1A). Children with an onset of FC older than 1 year and boys in FC cohort had a significantly higher risk of developing TS than non-FC cohort, whereas the infants with an onset of FC before 1 year and girls in the FC cohort had a non-significant increase risk of TS (Table 2). The urbanization-specific analysis showed that the patients in the FC cohort who particularly lived in rural areas had a significantly highest risk of TS compared with those in the non-FC cohort (adjusted HR = 6.16, 95% CI = 1.74–21.8,
Table 2). The patients lived in highest urban areas were also under a significant risk to have TS. Moreover, the occupation-specific analysis showed that both patients with white-collar or blue-collar parents had a significantly higher risk of developing TS in the FC cohort than the non-FC cohort, especially the patients with bluecollar parents (adjusted HR = 3.38, 95% CI = 1.65–6.52, Table 2).
The patients with comorbidities in FC cohort did not increase the risk to have TS compared with patients with comorbidities in non-FC cohort (Table 2). Instead, patients without comorbidities in FC cohort had higher risk to have TS than patients without
comorbidities in non-FC cohort, especially the patients without ADHD (adjusted HR = 2.58, 95% CI = 1.72–3.86, Table 2). The overall frequencies of medical visits were 21.3 _ 15.5 times in FC cohort and 24.2 _ 21.6 times in non-FC cohort (P < 0.0001). The frequency of FC-related medical visits in FC cohort was 1.61 _ 29.7 times per one year. We further analyzed the FC-related medical visits in FC cohort and we found an association between the frequency of FC-related medical visits and TS occurrence (Table 3). Although the
risk of TS in patients with 1–2 times FC-related medical visits was similar to that in patients without any FC-related medical visit in non-FC cohort, the risk of TS rose from 0.89 to 16.0 (trend test P < 0.001) when the frequency of FC-related medical visits increased from 1 to 2 times to more than 4 times (Table 3). The adjusted HR for TS in related to FC-related medical visits was 1.02 (95% CI = 1.02–1.03) per one frequency increment. The same trend was noted in both genders. Kaplan–Meier analysis of probability free of TS during the 13-year study period in Fig. 1B revealed the more FC related medical visits, the lower probability free of TS. The overall TS rate was 19.5% higher for FC children with >4 medical visits than that in the non-FC cohort (P < 0.0001, Fig. 1B). 4. Discussion
In this nationwide study, we determined that children with FC in early childhood had an increased risk of developing TS later in life after adjusting the confounding effects of comorbidities, such as ADHD, OCD and anxiety. The overall TS rate in FC cohort was higher in boys, children who lived in urban areas and children with white-collar parents. However, boys, children who lived in rural area and children with blue-collar parents in FC cohort were risky to have TS. We also noted that higher frequencies of FC-related medical visits in FC cohort increased the risk of TS. This association between FC and TS has not been studied in the literature, and is considered for the first time herein.
In early era, FC was reported to have a significant risk of intellectual and behavioral problems based on studies from tertiary referral centers or discordant twin-pairs.14,15 However, data from large population-based studies later showed a favorable outcome in children with a history of FC. The Child Health and Education Study, a national population-based study performed in UK, demonstrated children who had febrile convulsions performed as well as other children in terms of their academic progress,
intellect and behavior at 10 years of age.7 Three other populationbased studies in USA, Taiwan and Netherlands also showed a
favorable outcome in FC children.3,4,16,17 Chang et al. even reported FC children had better control of distraction, attention and working memory.3,17 However, some contradictory data were also explored by Chang et al. FC children were noted to have higher impulsivity than their age-matched control subjects.17 High impulsivity is one
of the clinical characteristics of ADHD. Another population-based
study using NHRID in Taiwan documented that FC children increased risk of subsequent ADHD occurrence.18 In accordance
with these studies, we found FC children had higher incidence of problem behavior comorbidities, such as ADHD as well as OCD and anxiety, than non-FC children. However, FC children with these problem behavior comorbidities did not increase the risk of subsequence occurrence of TS.
Epidemiological data showed that the risk of developing epilepsy is only slightly enhanced after a single, short FC, but is heightened after prolonged or recurrent FC.19 It has been also shown that children with prolonged or recurrent FC perform more poorly in mental performance in hospital-based studies.20 Neuroimaging studies have elucidated the acute hippocampal edema and chronic hippocampal signal changes in T2-weighted magnetic resonance images, especially in patients with prolonged focal FC.21 Moreover, animal studies on prolonged or recurrent FC have demonstrated neuronal cytoskeleton changes in the hippocampus and have enhanced long-term hippocampal excitability in
infant rats.22,23 Taken together, prolong or recurrent FC has more impacts on the brain than simple, short FC. Our study using NHIRD data found that children with FC in early childhood had a 2-fold increased risk of developing TS later in life. Even though we cannot explore the effect of prolong FC on the risk of TS occurrence because of the lack of FC duration record in NHIRD, we found that patients who had more FC-related medical visits increased risk of developing TS. Although the numbers of FC-related medical visits
may not precisely equal to the exact frequency of FC, increased FCrelated medical visits may implicate recurrent FC. Therefore, our
data may support that recurrent FC correlated with increased risk of developing TS. Oppositely, our data also showed that patients with 1–2 times FC-related medical visits was similar to that in patients without any FC-related medical visit in non-FC cohort. It implies that children with a few FC or non-recurrent FC did not seem vulnerable.
It has been reported that infants whose FC occurred within the first year of life required more special schooling than did children whose FC occurred later in life.7 The onset of FC before 1 year of age
was reported as the significant risk factor for deficits in learning tasks, consolidation, and delayed recognition, which presented after 6 years of age.17 Our data showed that infants whose FC presented before 1 year of age had a higher risk of developing TS (adjusted HR = 3.26) than did children whose FC presented after 1 year (adjusted HR = 1.86) even though the HR in FC patients before 1 year of age is not statistically significant compared with non-FC cohort (Table 2). Thus, the onset of FC before 1 year of age seems to affect the neurocognitive development. The relationship between FC before 1 year of age and risk of TS needs further investigations. Some might argue that the increment of TS in FC patients may be due to the increment of medical care in FC patients. However, the overall frequencies of medical visits were actually significantly higher in non-FC cohort than FC cohort. Thus, this argument of medical surveillance biases in our study would be slighted.
The strengths of our study included its use of population-based data that are highly representative of the general population. However, certain limitations to our findings should be considered. First, the NHIRD does not contain detailed information regarding socioeconomic status, or the family history of systemic diseases, both of which may be risk factors for FC or TS. Second, the evidence derived from a retrospective cohort study is generally lower in statistical quality than that from randomized trials because of potential biases related to adjustments for confounding variables. Despite our meticulous study design and the control measures for confounding factors, bias resulting from unknown confounders may have affected our results. Third, all data in the NHIRD are anonymous. Relevant clinical variables, such as blood pressure, imaging results, pathology findings, and serum laboratory data were unavailable. Thus, we cannot explore other known risk factors of TS such as group A streptococcus infections or prenatal morbidity in our patients.24,25
In conclusion, FC may increase the risk of subsequent TS occurrence in children. Children who had frequent medical visits for FC were particularly vulnerable. Additional prospective studies should be designed to support our findings.
