Included in this study were children with FXS (n = 158) with either the full mutation or mosaicism, and their mothers (n = 119) whom were seen between 2003 and 2009 at the MIND Institute. This study was approved by the Institutional Review Board of the University of California, Davis. The children with FXS in this cohort were divided into children (n = 61) who were born to mothers who had a history of autoimmune disease and children (n = 97) whose mothers did not have autoimmune disease.

The diagnosis of autoimmune disease in the mother was confirmed by a physician and these diseases include autoimmune thyroid disease, systemic lupus erythematosus (SLE), rheumatoid arthritis, multiple sclerosis (MS), Sjogren’s syndrome, psoriasis, optic neuritis, Raynaud’s phenomenon, or fibromyalgia, which has a postulated autoimmune component in its pathophysiology (Coffey et al. 2008; Di Franco et al. 2010; Greco et al. 2008; Staines 2004; Zhang et al. 2009). Sixty-one children with FXS with a mean age of 9.61 years (SD 5.59) were born to 41 mothers with a mean age of 40.54 years (SD 7.96) who had experienced autoimmune disease. They were compared to 97 children with FXS with a mean age of 9.41 years (SD 6.31), born to 78 mothers with a mean age of 38.62 years (SD 6.46) without a history of clinically defined autoimmune disease. Most mothers in each group [M + AI and M − AI, 23 (56.1%) vs. 59 (75.6%), respectively] had one child with FXS. There were 16 (39.0%) and 19 (24.4%) M + AI and M − AI, respectively, who had two children with FXS. Two (4.9%) M + AI had three children with FXS. Children with the FMR1 premutation born to these mothers were not included in the study. A description of both mothers and children with FXS are presented in Tables 1 and 2.Table 1

Characteristics of mothers with the fragile X premutation

All subjects underwent a full medical history and physical examination by a physician with expertise in fragile X-associated disorders (RJH) after informed consent was obtained. The medical history covered an extensive review of prenatal, perinatal and postnatal history and complications, developmental and behavioral variations and problems for each child including developmental delay or regression of development, tics (a sudden, repetitive, nonrhythmic, stereotyped motor movement or vocalization involving discrete muscle groups, such as eye blinking, shoulder shrugging, throat clearing, etc.; tics symptoms had to appear before 18 years of age and occur before taking stimulant medications if those with FXS also have co-morbid attention deficit/hyperactivity disorders), sleep problems (problems falling asleep or waking up at night), clinical seizures (seizures characteristics, onset, duration, frequency, any co-occurring loss of consciousness, other neurological deficit, associated symptoms, possible triggers if they could be defined, electrical activity recorded by an electroencephalography (EEG) testing, and anticonvulsant medications), other pertinent medical problems, review of systems, and medications. Autoimmune diseases/conditions in mothers were thoroughly obtained by self-report after questioning about all potential autoimmune disorders in the medical history mentioned above. Sometimes medical records (approximately 30%) of the mothers were available for review in those with autoimmune problems during the visits at our clinic; however, these data are not included in the analysis because it was not consistently obtained. To be documented as having an autoimmune disease in the mother or developmental, behavioral, and medical problems particularly seizures, tics, and sleep problems in children with FXS, the subjects had to have sought medical or pediatric professional help for their problems and to have been previously diagnosed and/or treated by a physician for each disease/condition (physician-documented clinical involvement). Each subject underwent a medical examination that included a head circumference. A large head or macrocephaly was defined as exceeding 2 standard deviations (SD) above normal when compared with typically developing children of the same age.

The diagnosis of ASD in those with FXS was made after the use of standardized measures including the Autism Diagnostic Observation Schedule (ADOS), the Autism Diagnostic Interview, Revised (ADI-R) and the Diagnostic and Statistical Manual of Mental Disorders (DSM IV) (American Psychiatric Association 2000; Le Couteur et al. 1989; Lord et al. 2000) followed by a team consensus discussion as previously described (Harris et al. 2008).

The developmental or cognitive abilities were assessed with age appropriate measures, including the Mullen Scales of Early Learning, the Wechsler Preschool and Primary Scale of Intelligence-Third Edition (WPPSI-III), the Wechsler Intelligence Scale for Children-Third Edition (WISC-III), the Wechsler Intelligence Scale for Children-Fourth Edition (WISC-IV), the Wechsler Abbreviated Scale of Intelligence (WASI), the Wechsler Adult Intelligence Scale-Third Edition (WAIS-III) or the Stanford-Binet Intelligence Scale, Fourth Edition. Children’s adaptive functioning was also assessed using the Vineland Adaptive Behavioral Scales, Second Edition (Vineland-II) by interviewing the caregivers (Mullen 1995; Sparrow et al. 2005; Thorndike et al. 1986; Wechsler 1991, 1997, 1999, 2002, 2003). Intellectual disability was defined as IQ less than 70.

Behavioral and emotional problems including externalizing behaviors (hyperactivity, aggression, and conduct problems), internalizing behaviors (anxiety, depression, and somatization), atypicality, withdrawal, attention problems, and adaptive skills (adaptability, social skills, leadership, activities of daily living, and functional communication) were completed by a limited percentage of parents for each group of children with FXS (autoimmune disease vs. no autoimmune disease, 72.1 vs. 54.6%) using Behavior Assessment System for Children (BASC) questionnaire (Reynolds and Kamphaus 1992, 2004). Behavior Symptoms Index (BSI), a measure of an overall level of problems consisting of the hyperactivity, aggression, depression, attention problems, atypicality and withdrawal scales, was also documented. The BASC parent rating scale yields T scores with a mean of 50 and a SD of 10. On the clinical scales, scores from 60 to 69 are in the “at-risk” range and scores of 70 and above indicate “clinically significant” problems. On the adaptive skill scales of the BASC, scores from 31 to 40 indicate “at-risk” problems and scores of 30 and lower are considered “clinically significant.” In this study, externalizing, internalizing behaviors, and BSI scores 60 or above were defined as having those problems accordingly, whereas deficits in adaptive skills were defined in which adaptive skills scores are 40 or lower. Table 3 lists the variables investigated statistically between both groups of children.Table 3

Neurodevelopmental and behavioral problems between both groups of children

A blood sample for measurement of methylation status, CGG repeat, and FMR1 mRNA levels were obtained from each subject (both mothers and children with FXS). FMR1 mRNA quantification, Southern Blot and PCR-based genotyping were performed as previously described (Tassone et al. 2000; Tassone et al. 2008).

Qualitative variables for which there were specified cut-offs including autoimmune disease in mothers, the diagnosis of ASD, intellectual disability, seizures, sleep problems, tics, macrocephaly, externalizing, internalizing behaviors, BSI and deficits in adaptive skills were rated dichotomously (presence or absence) as demonstrated in Table 3. Univariate analysis of continuous variables including age, CGG repeat size, FMR1 mRNA level, activation ratio (AR: the fraction of normal FMR1 allele as the active allele in female individuals with FXS), IQ, and adaptive functioning was based on t test. Fisher’s exact test for 2 × 2 contingency table analyses were used to assess the association between autoimmune status in the mothers of children with FXS and the diagnosis of ASD. Results are presented in terms of odds ratio (OR) and 95% confidence interval (95% CI). P value and 95% confidence interval (CI) relating to the primary hypotheses of whether there is an association of autoimmune disease in mothers to ASD in children with FXS was based on a logistic regression model using generalized estimating equations. This approach accounts for correlation within a family (i.e. mother–children correlation) in estimates of standard errors, although this did not make a material difference in the results. Also, models were explored with adjusting for mother’s age, child’s age, gender, number of offspring, whether a mother had a child before or after age 35, CGG repeat numbers, and mRNA levels. All covariates were not significant and adjusted OR estimates did not differ from OR estimates from models with only the main variable of interest—the presence of autoimmune disease. Thus, we reported this simple later model with main effect of interest. An exception is for the variable gender in the model of BSI outcome (as noted in Table 3; see “Results”). However, we did not adjust for multiple testing and the reason for this is that our original primary hypothesis relates to investigating the association between autoimmune disease in the mothers and ASD. Neurodevelopmental and behavioral outcomes, particularly the collection of BASC variables, were exploratory. BASC T scores were analyzed using general linear model adjusted for age and accounting for within-in family correlation and variation among subjects. All P reported are two-sided and the significance level is 0.05.

Characteristics of the mothers and their offspring with FXS are illustrated in Tables 1 and 2, respectively. There was no significant difference in mean age between M + AI and M − AI [40.54 (SD 7.96) vs. 38.62 (SD 6.46) years, P = 0.238]. There were no significant differences in molecular variables including CGG repeats, FMR1 mRNA levels, and AR between both groups of mothers. The two most common autoimmune disorders in the mothers were autoimmune thyroid diseases (27/41, 65.9%), and fibromyalgia (11/41, 26.8%). MS and Raynaud’s phenomenon were present at a similar rate in the mothers at 12.2% (5/41). Four (4/41, 9.8%) mothers reported to have been diagnosed with rheumatoid arthritis and three (3/41, 7.3%) mothers reported having SLE. Other reported autoimmune diseases included Sjogren syndrome (1/41), optic neuritis (1/41), and psoriasis for each disease (1/41). Of the 41 M + AI, there were 29 who presented with a single autoimmune disease, whereas a further 12 reported having two or more autoimmune diseases. Twenty (48.8%) mothers had autoimmune thyroid disease alone and 3 (7.3%) mothers had been diagnosed with fibromyalgia alone. In 23 (56.1%) mothers, the mean age of diagnosis with autoimmune disease was 30.26 (SD 9.45) years. The remainder of the group neither recalled the onset of the autoimmune disease nor the data for the onset of autoimmune disease in medical records were collected before the association between autoimmune disease in the mothers and the presence of ASD in their offspring with FXS was hypothesized.

Of the 41 mothers, 13 stated that their autoimmune disease occurred during their gestation. However, a precise ascertainment of the age of onset or diagnosis for each autoimmune disease based on self-report was very difficult due to the fact that the natural course of many autoimmune diseases can include a subclinical phase that does not meet stringent clinical diagnosis (Papi et al. 2007; Poppe and Glinoer 2003; Ulff-Moller et al. 2009).

There was significant difference neither in the mean age of children with FXS in the two groups [9.61 (SD 5.59) vs. 9.41 (SD 6.31) years, in children born of M + AI and M − AI, respectively, P = 0.836] nor in the type of the fragile X mutation (percentage of full mutation/mosaic = 78.7/21.3% vs. 79.4/20.6%, P = 0.876]. With regard to gender, there were fewer male children with FXS born to M + AI when compared with children with FXS of M − AI (60.7 vs. 78.4%, P = 0.016). Mean cognitive ability in both groups of children with FXS was in a mild range of intellectual disability and was not significantly different between groups [FSIQ = 62.75 (SD 17.77) vs. 60.80 (SD 17.55), in children born of M + AI and M − AI, respectively, P = 0.523]. Likewise, the adaptive behavior composite score documented by the Vineland between both groups of children with FXS was not significantly different [64.33 (SD 18.51) vs. 63.62 (SD 18.71), in children born of M + AI and M − AI, respectively, P = 0.847].

Of the 61 children with FXS born to M + AI, 38 (62.3%) children met the criteria for ASD. Twenty-two (36.07%) subjects had autism and 16 (26.23%) had PDD-NOS. Of the 32 mothers with children that had FXS and ASD, autoimmune thyroid disease was present in 22 (67.6%), fibromyalgia in 8 (25.0%), and Raynaud’s phenomenon in 4 (12.5%). Among children from M + AI, the odds ratio (OR) for FXS and ASD was 1.27 (95% CI 0.62–2.61, P = 0.5115). These data suggest that autoimmune disease in the mothers may not be a relevant risk factor for autism in combination with FXS. Other variables including intellectual disability, sleep problems, and macrocephaly were not different between both groups of children. However, the rate of tics was significantly higher in the group of children with FXS born to M + AI (15/58, 25.86%) compared with M − AI [10/94, 10.64%; OR = 2.94 (95% CI 1.19–7.24), p = 0.019].

Data about seizures were not available for three children. However, a higher rate of clinical seizures was observed in a group of children with FXS born to M + AI (9/60, 15.0%) when compared with children with FXS of M − AI [4/95, 4.21%; OR = 3.81 (95% CI 1.13–12.86), P = 0.031]. Details of the nine children with FXS who have seizures and were born to M + AI are as follows: age range, 4.17–24 years (median age 9 years); gender, 7 boys and 2 girls; FMR1 mutations, six have the full mutation and three have mosaicism in the FMR1 gene. Three have more than one seizures type including complex partial seizures, generalized seizures, gelastic seizures and absence seizures, four had generalized tonic-clonic convulsions alone, and two experienced petit mal or absence seizures alone. Seizures onset ranged from 17 months to 9 years with a median age of onset at 3 years. With regard to treatment of seizures, seven subjects reported to take long-term anticonvulsant medications. Four of these seven subjects took more than one anticonvulsants to control the seizures. Of these nine children with FXS born to M + AI who developed clinical seizures, eight mothers had autoimmune thyroid disease. There was no history of birth asphyxia or neonatal encephalopathy. The cognitive abilities of children with seizures were similar to the whole group of children with FXS.

Details of four children with FXS who had seizures born to M − AI are as follows: age range 3.63–17.75 years (median age 8.25 years), gender 3 boys and 1 girl, all have the full mutation in the FMR1 gene; one reported to have more than one seizure type including absence, complex partial seizures, and generalized seizures, one experienced absence seizures alone, one had partial seizures alone, and one subject was noted to have two episodes of simple febrile seizures but this child has never taken any medications. Seizures onset in this group ranged from 2.67 to 5 years with a median age of onset at 3.84 years. Abnormal EEG results were confirmed in all subjects in this group of children. With regard to anticonvulsant medications, one child has taken more than one medication to treat the seizures and seizures were well controlled by one medication in two subjects in this group.

Behavioral and emotional problems documented by BASC were completed in 97 (61.4%) children with FXS in our study. The children’s age, gender and type of the fragile X mutations were not significantly different between both groups (children of M + AI and M − AI). The rate of externalizing behaviors [36.36 vs. 16.98%; OR = 3.04 (95% CI 1.10–8.38), P = 0.0319] was increased in children with FXS born to M + AI. T scores of these behavioral and emotional problems between both groups of children are illustrated in Table 4. Children with FXS born to M + AI were likely to have higher T scores on aggression, internalizing behaviors, depression, withdrawal symptoms, and BSI when compared with those children with FXS born to M − AI, even though only withdrawal symptoms and BSI scale reached a significant level for documenting the problems.Table 4

Parent rating scale (T score) of behavioral and emotional problems documented by BASC between both groups of children