Study participants

Ninety-eight participants were included in analyses and 45/98 (46%) were female. Average age at time of enrollment was 8.8 (4.6) years, with a range of 3–21 years. At baseline, mean VABS-II composite score was 51.2 (13.9), and best estimate IQ was 26.2 (17.8). There were 26 participants with Class 1 deletions, 53 with Class 2 deletions, and 19 with sequence variants.

Seizure characteristics

Through the course of the study, 42% (41/98) of the cohort reported a history of any seizure type, including generalized seizures (27%, 26/98); focal seizures (20%, 20/98), or unclassified seizure type (13%, 13/98) (Table 1). Two participants reported febrile seizures only, which were included in any seizure type and unclassified seizure type. Prevalence of specific generalized and focal seizure subtypes are shown in Table 1. Caregivers of 19 participants reported a formal epilepsy diagnosis; seizure subtypes for participants with a reported epilepsy diagnosis are located in Supplemental Tables 2 and 3.

Table 1 Seizure types ever experience by participants in the cohort

There were 12 participants with generalized seizures only, five with focal seizures only, and eight with unclassified seizures only; 13 participants had both generalized and focal seizures. Two participants had generalized, focal, and unclassified seizures reported, and one participant had generalized and unspecified seizures reported. True count and prevalence of generalized and focal seizures may be different as unspecified seizures can be either focal or generalized. For participants who reported the date of seizure onset, average age of onset of each type of seizure is displayed in Table 2 and Fig. 1.

Table 2 Age of onset of seizures for those with this data availableFig. 1

Seizure onset. Legend: Histogram of the age of seizure onset (years). The Y axis is the proportion of individuals. Bin width is 1 year

Association of seizure prevalence with adaptive functioning and cognitive ability

At baseline, presence of any seizures, generalized seizures, and focal seizures were all significant predictors of the VABS-II composite standard score. The composite score was 8.73 [-13.3- -4.2] points lower in individuals with any seizures compared to those without seizures, while controlling for age, ASD diagnosis, and genetic grouping (p = 0.0003) (Fig. 2). The presence of focal seizures had a similar impact, though with more variability, with a composite score 8.07 [-14.8- -1.4] points lower in those with focal seizures, while controlling for age, ASD diagnosis, and genetic grouping (p = 0.019). Generalized seizures had a slightly greater impact on the composite score, where those with generalized seizures had a score 9.90 [-15.6- -4.2] points lower than individuals without generalized seizures, while controlling for the same covariates (p = 0.0008). Bonferonni corrections for multiple comparisons would provide an alpha of 0.017, leaving the categories of any seizures and generalized seizures remaining statistically significant and the category of focal seizures just above the threshold for statistical significance.

Fig. 2

Effect of seizures on Vineland Adaptive Behavior Composite. Legend: Multiple linear regression including presence of seizures at baseline with VABS-II Adaptive Behavior Composite. Blue scatterpoints and line represent individuals without seizures (left), without generalized seizures (middle), or without focal seizures (right). Red scatterpoints and line represent individuals with any seizures (left), generalized seizures (middle), or focal seizures (right). The X axis is age in years. Downward slope indicates that skills are not increasing at a rate comparable with the general population not that skills are declining

Similarly, presence of any seizures, generalized seizures, and focal seizures at baseline were all significant predictors of best estimate full-scale IQ. Best estimate FSIQ was 8.74 [-14.6- -2.9] points lower in individuals with any seizures compared to those without seizures, while controlling for age, ASD diagnosis, and genetic grouping (p = 0.004) (Fig. 3). Best estimate FSIQ was 10.14 [-18.6- -1.7] points lower in those with focal seizures compared to individuals without focal seizures, while controlling for age, ASD diagnosis, and genetic grouping (p = 0.025), again showing wider variability. Generalized seizures had the greatest impact on FSIQ where those with seizures had a score 11.33 [-18.5- -4.13] points lower than individuals without generalized seizures, while controlling for the same covariates (p = 0.002). Bonferonni corrections for multiple comparisons would provide an alpha of 0.017, leaving any seizure and generalized seizures significant and focal seizures just over the significance value.

Fig. 3

Association between occurrence of seizures and best estimate full scale IQ. Legend: Multiple linear regression of seizure occurrence at baseline and the best estimate full scale IQ. Blue scatterpoints and line represent individuals without any seizures (left), generalized seizures (middle), or focal seizures (right). Red scatterpoints and line represent individuals with seizures (left), generalized seizures (middle), or focal seizures (right). The X axis is age in years. Downward slope indicates that skills are not increasing at a rate comparable with the general population not that skills are declining

When assessing cognitive and adaptive scores within participants with seizures across either one (generalized or focal) or two (generalized and focal) seizure types at baseline, participants with two seizure types tended to have lower, but not significantly lower, scores, compared to participants with only one seizure type. Using the same model with age, genetics, and ASD held constant, participants with both generalized and focal seizures at baseline had an Adaptive Behavior Composite score 5.30 points lower [-14.7 – 4.1] (p = 0.25), and an FSIQ points 7.79 lower [-18.3 – 2.7] (p = 0.14) than individuals with only focal or generalized seizures.

Genotype–phenotype associations

No differences were found between genetic subgroups and the presence of seizures or any specific type of seizure at any timepoint. Eight of 26 (31%) subjects with a Class 1 deletion, 26/53 (49%) of subjects with a Class 2 deletion, and 7/19 (37%) with a sequence variant reported any seizure type (p = 0.27). 3/26 (12%) with a Class 1 deletion, 17/53 (32%) with a Class 2 deletion, and 6/19 (32%) with a sequence variant reported generalized seizures (p = 0.13). Five of 26 (19%), 11/53 (21%), and 4/19 (21%) of the Class 1, Class 2, and sequence variant group reported focal seizures, respectively (p = 0.98).

Among participants with chromosomal deletions (n = 79), those who had any seizure type had a median of 71 (IQR: 68) genes deleted as compared to individuals without seizures who had a median of 25 (41) genes deleted (p = 0.002) (Fig. 4). Participants with chromosomal deletions and generalized seizures had median of 72 (43) genes deleted as compared 31 (45) for those without generalized seizures (p = 0.0005). No significant differences were found in the presence of focal seizures, where subjects with chromosomal deletions and seizures had a median of 89 (89) genes deleted compared to 39 (49) in individuals without focal seizures. The differences in any seizures and generalized seizures were significant after Bonferonni corrections, with an alpha of 0.017.

Fig. 4

Number of genes included in deletions of participants with and without seizures. Legend: Boxplots representing the number of genes deleted in participants’ deletions for those with chromosomal deletions and without seizures (blue) or with seizures (red). The minimum and maximum values are represented with tails, the box represents the interquartile range, the bold line represents the median, and the dashed line represents the mean genes deleted per group. Asterisks represent significance level, ** p <  = 0.01, *** p <  = 0.001, ns p > 0.05

Individuals with PMS who have a ring chromosome had a similar seizure prevalence (2/6, 33%) compared to the rest of the cohort. These two participants both had focal and generalized seizures.

Clinical associations

The presence of any seizures at any timepoint was not associated with any demographic, pre- or neonatal complications, early infancy features, or medical features (Table 3).

Table 3 Prevalence of clinical features in participants with and without seizures

Developmental regression was not significantly associated with any seizure type or focal seizures at any timepoint, but was associated with generalized seizures (p = 0.003). Of the 26 subjects with generalized seizures, 18 (69%) of them had a reported developmental regression. Of the 72 participants without generalized seizures, 24 (33%) had a reported regression. For participants who reported both age of first seizure and age of regression, 13/16 (81%) had a reported regression prior to a seizure while the remaining 3/16 (19%) had a reported regression after a first seizure.

Medication and seizure management

Participants used multiple different ASMs. The three most commonly used ASMs were levetiracetam, valproic acid, and lamotrigine (Table 4). Of those who ever used ASMs (n = 25), over half (n = 15) used only one ASM, while 10 participants used two or more ASMs, and one participant used 7 different ASMs. Use of as needed diazepam and midazolam were not included in ASM counts. One participant had a corpus callosotomy for refractory epilepsy, five years later the patient had a placement of a vagus nerve stimulator, which was replaced four years after that.

Table 4 Anti-seizure medications utilization