A national survey of caregivers of individuals with fragile X syndrome addressed characteristics of epilepsy and co-occurring conditions. Of the 1,394 individuals (1,090 males and 304 females) with the full mutation, 14% of males and 6% of females reported seizures. Seizures were more often partial, began between ages 4 and 10 years, and were infrequent and easily treated. Similar characteristics and patterns were seen in medical chart review data from a large clinic cohort of patients with fragile X syndrome. National survey data showed that autism was significantly associated with seizures as a co-occurring condition. Although seizures in fragile X syndrome are typically not severe and easily treated with medications, they appear to be associated with developmental–behavioral comorbidity that impacts function.
Fragile X syndrome is the most common known inherited cause of intellectual disability, learning disability, and autism, with an estimated frequency of about 1∶2,500 (P. Hagerman, 2008). Individuals with fragile X syndrome display variable physical features, such as large ears, long face, macrocephaly, and macroorchidism, and variable level of functioning, with a fairly stereotyped cognitive profile and behavioral and personality features characterized by hyperactivity, anxiety, and socialization difficulties. Certain medical problems are more common in fragile X syndrome, such as strabismus, presbyopia, frequent otitis media, mitral valve prolapse, and seizures (Berry-Kravis, Grossman, Crnic, & Greenough, 2002; R. Hagerman et al., 2009).
Fragile X syndrome results from a trinucleotide repeat (CGG) expansion mutation (Verkirk et al., 1991) of greater than 200 repeats (full mutation) in the promoter of the FMR1 (fragile X mental retardation 1) gene, which leads to transcriptional silencing of FMR1 and loss or significant reduction of expression of the gene product, FMRP (fragile X mental retardation protein) (Devys, Lutz, Rouyer, Bellocq, & Mandel, 1993). FMRP is an RNA-binding protein that appears to function as a dendritic translational repressor (Bagni & Greenough; 2005; Grossman et al., 2006) that modulates receptor (including group I metabotropic glutamate [mGluR1 and mGluR5] and muscarinic acetylcholine receptor [Volk, Pfeiffer, Gibson, & Huber, 2007]) activated dendritic protein synthesis (Weiler et al., 2004), contributing to dendritic morphological maturation (Comery et al., 1997) and synaptic plasticity. Thus, an excess of immature dendritic spines is found in brain from humans with fragile X syndrome and from the fmr1 knockout mouse model (Bagni & Greenough, 2005). Deficits in cortical (Li, Pelletier, Velazquez, & Carlen, 2002) and hippocampal (Lauterborn et al., 2007) long-term potentiation (LTP) as well as enhanced hippocampal mGluR-mediated long-term depression (LTD) (Huber, Gallagher, Warren, & Bear, 2002) have been demonstrated in the fmr1 knock out as well. Accordingly, these synaptic abnormalities result in abnormal epileptiform discharges and a high frequency of audiogenic seizures in the fmr1 knock out (Chuang et al., 2005; Musumeci et al., 2000), corresponding to the increased risk of seizures in humans with fragile X syndrome (P. Hagerman & Stafstrom, 2009)
Previous researchers have described seizures in 10 to 40% of males with fragile X syndrome (Bailey, Raspa, Olmsted, & Holiday, 2008; Berry-Kravis, 2002; Klugger, Bohm, Laub, & Waldenmaier, 1996; Musumeci et al., 1991, 1999; Sabaratman, Vroegop, & Gangadharan, 2001), with a frequency of 13 to 18% in larger cohorts with less referral bias (Bailey et al., 2008; Berry-Kravis, 2002; Musumeci et al., 1999). Many children with fragile X syndrome also have abnormal EEGs without overt epileptic seizures, frequently with a pattern of centrotemporal spikes (similar to benign focal epilepsy of childhood) (Berry-Kravis, 2002; Musumeci et al., 1991; Rees et al., 1993). Complex partial seizures are most common in fragile X syndrome (Berry-Kravis, 2002; Musumeci et al., 1999), although simple partial and generalized tonic–clonic seizures may occur. Seizures are reported to be easily controlled (Berry-Kravis, 2002) and resolve during childhood in the majority of individuals with fragile X syndrome.
Despite reports on prevalence and types of seizures in fragile X syndrome, there has been a paucity of studies evaluating associations between seizures and level of function or other co-morbidities in fragile X syndrome. In a recent study of a small cohort, Garcia-Nonell et al. (2008) found a trend towards an increased rate of seizures in individuals with fragile X syndrome also diagnosed with autism. In the present report, we used data from the National Fragile X Survey, which was administered to families of a proband with fragile X syndrome, to investigate characteristics of seizures in fragile X syndrome and determine whether the presence of seizures is associated with other co-occurring conditions. Recorded data on seizure characteristics from a large clinic cohort with fragile X syndrome were also analyzed to help evaluate accuracy of parent report of data in the national survey.
Participants and Recruitment
Our study was part of a larger survey (see Bailey, Raspa, & Olmsted, 2010) that was used to assess the needs of families with at least one child who was a carrier of fragile X syndrome or had the full mutation. Study participants were recruited in two phases. The survey investigators partnered with three national foundations (National Fragile X Foundation, FRAXA Research Foundation, and Conquer Fragile X Foundation), researchers, and clinicians who sent study materials (a letter and brochure) to families on their mailing lists. Approximately 6 months later, families who enrolled in the first phase of recruitment were contacted and asked to complete a national survey in the second phase of the study. Families completed the survey between March and June 2008.
A total of 1,250 families enrolled and 1,075 families (86%) subsequently completed the full survey. Here we focus on a subset of 168 families who had at least one child or family member with the full mutation and who had been diagnosed with or treated for seizures. Demographic data for the cohort is shown in Table 1. Families were predominantly Caucasian. Among the respondents, 88% were mothers; 10%, fathers; and 2%, other family members. The majority were married and had a 4-year college or graduate degree. Overall, 66% of the respondents were employed, and there was a wide range of income level. The families had 1,394 individuals with an FMR1 full mutation (1,090 males and 304 females). In addition, there were 235 premutation carriers, 525 family members with no FMR1 and 518 family members with undetermined genetic status.
Fragile X Clinic Comparison Group
We conducted chart review using a clinical database that included all patients with mutation-confirmed fragile X syndrome seen and followed at the Fragile X Clinic at Rush University Medical Center from 1992 to 2008. Because questions regarding seizures and EEG history are part of the standard intake questions for all new and follow-up patients seen at the clinic, current data on seizure history were available from virtually all patients. Seizure data were entered into a database after review of the clinical information by the first author, who is a neurologist. This database study was approved by the Rush University Medical Center Institutional Review Board.
Instruments and Procedure
We generated National Fragile X Survey items through literature review and discussion with clinician experts in the fragile X field, with a goal of identifying characteristics of seizures, functional status, and potential comorbid problems. Before the survey was implemented nationally, a small group of parents of children with fragile X syndrome completed the survey and provided feedback to the authors regarding clarity of items. Minor revisions were made on the basis of this feedback. Families were invited to enroll and complete the national survey either online (80%) or over the phone with a trained interviewer.
During enrollment families provided demographic information as well as information about each child in the family (e.g., date of birth, gender, genetic status). Families were asked to rate a variety of characteristics for each child, including their (a) overall thinking, reasoning, and learning ability; (b) quality of life; and (c) overall health. For each of these items, families were asked, “How would you describe (child's)…?” Families were not specifically instructed to compare to another child of the same age or another child with fragile X syndrome. Before further analysis, we recoded these variables as dichotomous to account for low frequencies for some of the response categories. The overall thinking, reasoning, and learning ability item was answered on a 4-point scale (1 = poor, 2 = fair, 3 = good, 4 = very good) that was recoded as a dichotomous variable, where 1 = poor or fair and 2 = good or very good. The quality of life and health items were answered on a 5-point scale (1 = poor, 2 = fair, 3 = good, 4 = very good, 5 = excellent) that we recoded as dichotomous variables, where 1 = poor or fair and 2 = good to excellent. A dichotomous rating for the child's communication skills was created based on the family's response to whether (a) the child uses single words or signs or (b) uses 2- or 3-word sentences; the child's reading skills ratings were based on response to whether the child can read a picture book or simple story fairly or very well. In addition, families indicated whether each child had been diagnosed or treated for developmental delay or any of seven additional co-occurring conditions (attention problems, hyperactivity, aggressiveness, self-injury, autism, anxiety, depression). Families were asked to respond either yes or no to the list of eight co-occurring conditions. There were no qualifiers for these items, with the exception of the self-injury item, which provided an example (e.g., hitting head or biting hand).
When families later completed the national survey, we presented them with an additional set of items concerned with the child's experiences with seizures. Questions were focused on age of onset, number of seizures experienced, amount of time since last seizure, type and severity of seizure, and whether the child was currently taking any seizure medications. Only data from families currently living in the United States were included in the analysis.
For all analyses, we used the Statistical Analysis System (SAS, Version 9). Percentages were calculated for most analyses, including the demographic information and descriptive data on seizures. A second set of analyses were conducted to enable us to compare children with and those without seizures, both with the full mutation of fragile X syndrome, using a matched pair design. We matched children using exact matches on categorical variables (gender, family income) and within-range matches on continuous variables (age). The matched pair analysis was chosen to stratify by gender and adjust for the confounding factors of age and an economic factor. We calculated cross-tabulation frequency tables to compare the matched pairs on a number of variables, including presence of co-occurring conditions, communication skills, overall thinking, reasoning and learning ability, quality of life, and overall health. We used McNemar's test for 2 × 2 tables (Fisher & van Belle, 1993). In addition, we used logistic regression models to determine whether treatment or diagnosis of seizures were associated with the same variables included in the matched pair analysis while we controlled for child's age and family income and stratified by gender (Kleinbaum & Klein, 2002).
Prevalence of Seizures
Of the 1,394 individuals with an FMR1 full mutation who were included in the national survey (1,090 males and 304 females), respondents reported that 173 (12%) had seizures: 154 males (14%) and 19 females (6%). Individuals with a seizure history ranged in age from less than 1 to 55, whereas individuals who did not have seizures ranged in age from less than 1 to 62 years. Because both cohorts in this report included very young children who may not yet have had a seizure but may in the future, the seizure frequencies reported may be a bit low. This is substantiated by the finding that higher percentages of individuals with fragile X syndrome were reported to have seizures across four older age categories and the 18% frequency of seizures observed in the National Fragile X Survey enrollees with fragile X syndrome over age 5 (Bailey et al., 2008).
Seizure Characteristics and Response to Treatment
Data on seizure characteristics and response to treatment are shown in Table 2. The most common age of seizure onset was in young and mid-childhood between 4 and 10 years of age (53% of males and 32% of females), whereas only 8% of males and 21% of females reported seizure onset at age 11 or older. The majority of subjects had infrequent seizures, with the last seizure occurring 6 or more months ago in 69% of males and 89% of females, and a seizure occurring in the past month in only 13% of males and 0% of females. Seizures were most commonly considered mild or moderate in severity. The total number of seizures experienced varied widely, from 1 to more than 30, and could occur commonly in both sleep and wakefulness. Partial seizures were the most common type, although about one third of the subjects had what was thought to represent only generalized seizures. Most subjects were on zero or one medication, suggesting they were well-controlled or no longer had active seizures. Only 15% of males and no females were treated with more than one anticonvulsant. The majority of males (75%) and all females treated with medication for seizures reported that medication was effective for controlling their seizures.
Association Analyses for Seizures and Co-Occurring Conditions By Group
Co-occurring symptoms and conditions were frequent in individuals with fragile X syndrome reported to have seizures (Table 3). We employed matched-pair analysis to determine whether these problems were actually more frequent in the group with fragile X syndrome and seizures than in a matched control group with fragile X syndrome and an FMR1 full mutation but no seizures. This analysis revealed that males with fragile X syndrome who manifested seizures at some time were more likely to have a co-diagnosis of autism and symptoms of anxiety than those who never had seizures (Table 4). Further, aggressiveness, poor speech (nonverbal or minimally verbal), and fair–poor overall health were increased in the group with seizures. When we applied a Bonferroni correction for multiple comparisons, autism was the only characteristic that remained significantly more prevalent in the seizure group. For females, the low total number of individuals who had seizures limited significance of the analysis; however, as in males, autism appeared to be more frequent in the group with seizures (Table 4).
All the variables that were significant in the matched pair analysis were also found to be significant in the unmatched logistic regression analysis. In addition, for both males and females, ps = .02 and .04, respectively, the co-diagnosis of attention problems was found to be significant. The odds ratio (OR) for males was 1.91, with a 95% confidence interval (CI) (1.12, 3.25), indicating that males with seizures were almost two times more likely to have attention problems when compared to males without seizures. Due to the small sample size, the female OR of 3.77 has a very wide CI (1.05, 13.49) and should be interpreted with caution. The codiagnosis of self-injury had a significant OR for males, OR = 1.47, CI = 1.02, 2.14, p = .04. Females had a significant OR for the codiagnosis of depression, OR = 3.77, CI = 1.05, 13.49, p = .04, and fair–poor overall health, OR = 8.62, CI = 2.19, 33.98, p = .002. Additional results are available upon request.
Comparison of Survey Data on Seizures With Physician-Generated Clinic Data
In order to corroborate and lend clinical validity to the survey results, we compared physician-generated data (in categories available) from the Rush University Medical Center Fragile X Clinic with the survey data generated from parent report (Table 2). Clinic data were available from 352 individuals with fragile X syndrome; 39 of 268 males (14.5%) and 7 of 84 females (8.3%) had seizures, percentages very similar to those observed in the survey data. Age range of those with seizures at the time of chart review was 4 to 41 years, and for those without seizures, it was 1 to 57 years. Age of seizure onset was distributed similarly in the clinic cohort and survey data (Table 2), with the most common age of seizure onset in young and mid-childhood between age 4 and 10 years (59% of males and 42% of females), and only 23% of males and 29% of females had seizure onset at age 11 or older. As in the survey, focal seizures were most common, with a similar fraction (38% of males) exhibiting generalized seizures only. EEG data were available for 31 individuals from the clinic cohort and included focal sharp/spike discharges in centrotemporal (5), temporo-occipital (6), and frontal (2) areas, focal rhythmic frontal slow waves (2), generalized discharges (2), unspecified epileptic discharges (5), and normal findings/generalized slowing (9). Also similar to survey data, the vast majority of individuals in the clinic cohort had well-controlled seizures, with 90% of males and all females having had their last seizure over 6 months previously, no patients on more than one antiepileptic medication, and most patients (82% of males, 71% of females) treated with medication for seizures showing excellent effectiveness of their antiepileptic medication (seizure-free for over one year). Patients in this cohort were most commonly treated with valproic acid, lamotrigine, levetiracetam, and oxcarbazepine. The majority of both males and females were off medication because they had less than two seizures, had very infrequent seizures (several episodes separated by years), or because seizures had remitted. The average age of seizure remission was 9.5 (range = 3 to 23) and 5.5 (range 5 to 6) years for males and females, respectively.
This study represents the largest cohort of individuals with fragile X syndrome and seizures reported to date. It provides data on seizure characteristics and co-morbid diagnoses in patients with fragile X syndrome and seizures, not previously identified. The frequency of seizures in the National Survey Fragile X syndrome cohort for males (14%) is similar to that for males in the Rush University Medical Center Fragile X Clinic fragile X syndrome clinic cohort (14.5%). In a smaller recent study, Garcia-Nonell et al. (2008) reported a slightly higher seizure frequency (22% overall, 12% in those wihout autism and 28% in those with autism) for males with fragile X syndrome, but their study involved only 90 subjects, and seizure frequencies are more likely to be influenced by ascertainment bias and cohort effects in this and other smaller studies. The rates of reported autism in full mutation males in the survey (43%) are somewhat lower than those in Garcia-Nonell et al.'s study (63%), which might partly explain observed differences in seizure rates; however. the rate of seizures in the survey was 10% in those without autism and 21% in those with autism, suggesting that rates of seizures in all groups were lower in the survey cohort. The frequencies of seizures observed in the two male fragile X cohorts reported here are similar to those reported in the two largest prior studies from fragile X clinic populations, 18% and 13% (Berry-Kravis, 2002; Musumeci et al., 1999). Further, the survey cohort frequency is ascertained from a large survey that is less likely to be subject to ascertainment bias than would a nationally known clinic where problem cases are referred. Finally, the frequencies from the two large cohorts in this report are consistent to a large degree, despite being obtained via different methodology, supporting the accuracy of these data.
The characteristics and patterns of seizures in fragile X syndrome identified in this report are similar to those reported in prior studies (Berry-Kravis, 2002; Musumeci et al., 1999), with most frequent onset in mid-childhood, predominance of focal seizures, and a natural history including infrequent seizures, mild manifestations, and good effectiveness of anticonvulsants. Indeed, in about 40% of males and 50% of females participating in the survey, anticonvulsants were either not required or had been discontinued, presumably because the person no longer had seizures. Overall, the data suggest that seizures are rarely intractable and typically are easily controlled by aniticonvulsants in individuals with fragile X syndrome, although a small percentage of more severe cases required multiple medications or showed poor responsiveness to medications. Of the females with fragile X syndrome and seizures, all showed good medication responsiveness, and none was on more than one antiepileptic medication, suggesting that, as expected, seizures are an even less significant problem for females with fragile X syndrome. The concordance of seizure frequency, seizure type, typical age of onset, infrequent seizures, and good control with anticonvulsants between the survey data and Rush University Medical Center Fragile X Clinic data helps validate the seizure data derived from parent report in the survey as an accurate reflection of epilepsy characteristics in fragile X syndrome.
Although in the National Fragile X Survey, data were not collected on particular medications, the medications utilized for successful seizure control in the Rush University Medical Center Fragile X Clinic were associated with less cognitive and daytime sedation side effects, so as not to aggravate functional deficits in patients with fragile X syndrome. Given that there is a wide range of anticonvulsants with good safety profiles now available, good seizure control without substantial side effects should be possible in most patients with fragile X syndrome, even if trials of several antiepileptics are needed to achieve this goal.
The finding of an increased rate of autism in fragile X syndrome participants with seizures is consistent with the one available prior report indicating an increased incidence of seizures in the fragile X syndrome group with autism versus those without autism (Garcia-Nonell et al., 2008). Together, these studies suggest an association between seizures and autism in fragile X syndrome, likely promoted by a common problem with CNS connectivity. Aggressive and self-injurious behavior, anxiety, attention problems, and poor speech are features typically seen in individuals with both fragile X syndrome and autism and, thus, the weaker association (significant only without Bonferroni correction or only in the unmatched logistic regression analysis) of these features with seizures in fragile X syndrome would seem likely to be driven by the autism–seizure association.
Epilepsy is thought to occur in about a third of persons with idiopathic autism, with quoted frequencies ranging between 5 and 45% (Spence & Schneider, 2009). As observed in the fragile X syndrome cohort presented here, there is no one primary seizure type in idiopathic autism, and individuals can present with complex partial, absence, and generalized tonic–clonic episodes. As reported in fragile X syndrome (Berry-Kravis, 2002), epileptiform EEGs are found frequently in idiopathic autism, even in the absence of clinical seizures (Spence & Schneider, 2009). Age of seizure onset in idiopathic autism has two peaks, in early childhood and in adolescence, in contrast with our finding of peak onset in mid-childhood (ages 4 to 10) in fragile X syndrome, and reasons for this difference are unclear.
The association between an autism diagnosis and seizures in the National Fragile X Survey raises the question of whether seizures are causative of increased dysfunction in fragile X syndrome, resulting in the autistic phenotype. Alternatively, seizures and autistic features may just be common manifestations of the same underlying synaptic dysfunction, and the presence of both is reflective of a more severe level of dysfunction, perhaps in particular areas of the brain. In fact, seizures have been found to be more prevalent in children with Down syndrome and autism compared with those who do not have autism, and Malloy et al. (2009) hypothesized that this finding relates to more severely compromised neural connectivity as a common risk for both problems. Although this topic is controversial, there is currently no evidence that seizures or epileptic discharges on EEG cause autism or worsen the clinical course or brain connectivity (Deonna & Roulet, 2006; Levisohn 2007; Tharp 2004). There is also no evidence outside of anecdotal reports and uncontrolled series showing that treatment of seizures or EEG abnormalities in autism (Spence & Schneider, 2009; Tharp, 2004) change the clinical course of the disorder. In fact, the single report of a controlled trial of an anticonvulsant (lamotrigine) in autistic disorder did not show any behavioral improvement in the treated group and showed a high incidence of placebo effects (Belsito, Law, Kirk. Landa, & Zimmerman, 2001). Valproic acid did not improve learning or behavior in a placebo-controlled crossover trial in children with developmental problems and epileptiform bursts but no seizures (Ronen, Richards, Cunningham, Secord, & Rosenbloom, 2000). Because autistic features are often well-established by age 4, and the majority of individuals with fragile X syndrome develop seizures at age 4 or older, the autism likely predates the seizures in many cases, suggesting a common underlying etiology but not causality. Further research is needed to fully elucidate relationships between epilepsy and autism. In any case, it is important to treat clinical seizures when they occur to prevent complications. Current recommendations are that anticonvulsant treatment be implemented according to standard practice guidelines, which would typically dictate treatment after the second clinical seizure or after a single prolonged seizure.
It is interesting that thinking and learning function does not seem to be associated with seizures in males with fragile X syndrome, given the language and autism association. This suggests that either lay respondents to the survey may exhibit so much variability in perception of severity of thinking and learning ability in their relative with fragile X syndrome that the assessment of the relationship with seizures is rendered invalid or that overall ability may not be impacted by the same type or location of synaptic dysfunction producing seizures and autistic traits. The latter explanation would be supported by the lack of association between seizures and ability to read, a relatively specific achievement that caregivers should be able to accurately report. Further, although intellectual disability is associated with seizures in autism in many but not all studies, the higher than population rates of epilepsy in autism with normal IQ supports the notion that brain patterns that give rise to the autism itself also promote epilepsy.
Limitations of this study include those intrinsic to the National Fragile X Survey in general, delineated in the introductory paper for this series (Bailey, Raspa, & Olmsted, 2010). In addition, results from participants who responded on the phone were not compared with those who responded online. There may be discrepancies between these two groups. Particularly for the seizure study, however, we note that the results were similar to those obtained from the Rush University Fragile X Clinic medical records and neurologist review, with respect to seizure frequency, type, and characteristics, suggesting that family report is fairly accurate for these seizure parameters in the national survey. The comparison between the clinic and survey data may be limited somewhat by representation of some individuals in both data sets, although this is not likely a substantial portion of either dataset. For both the survey and clinic cohort, measured seizure frequencies may be biased toward lower numbers, due to inclusion of very young children who have not yet had seizures that will eventually occur.
A major strength of this study is the size of the cohort evaluated, as it would be difficult for any one fragile X clinic to amass such a large data set for seizure record review. The ability to do a well-matched controlled analysis of comorbid problems in fragile X syndrome groups with and without seizures is also an important strength of this study.
In conclusion, this study provides useful data for clinicians by characterizing seizures in fragile X syndrome and in guiding families with regards to prognosis of seizures in this syndrome. Although Garcia-Nonell et al. (2008) reported a nonsignificant trend toward increased seizures in males with fragile X syndrome and autism compared with males with fragile X syndrome without autism, the data presented here are the first to demonstrate a significantly increased prevalence of autism in males with fragile X syndrome and seizures when compared to those without seizures. Understanding of the relationship between seizures and autism is important for professionals who assist with clinical management of patients with fragile X syndrome. Further, the association between seizures and diagnosis of autistic disorder in fragile X syndrome is intriguing in that it may provide clues to linked patterns of clinical involvement that represent final common pathways of synaptic dysfunction in fragile X syndrome. Finally, our results provide information relevant to the design of future trials of new treatments, such as GABA agonists and mGluR negative modulators, targeted to treat seizures in fragile X syndrome based on synaptic mechanisms in the underlying genetic disorder (Berry-Kravis et al., 2009; P. Hagerman & Stafstrom, 2009).
Preparation of this article was supported in part by the Centers for Disease Control and Prevention (CDC) and the Association for Prevention Teaching and Research (APTR) Cooperative Agreement No. U50/CCU300860, Project TS-1380. The findings and conclusions in this publication are those of the authors and do not necessarily represent the views of CDC or APTR. Portions of the work were also supported by a grant from the Illinois-Eastern Iowa Kiwanis Spastic Paralysis and Related Disorders Foundation to the first author. A portion of the data was presented at the annual Child Neurology Society meeting in 2009, Louisville, KY, and published in abstract form in an Annals of Neurology supplement for the meeting.
Editor-in-charge: Leonard Abbeduto