Cognitive outcome after epilepsy surgery in children: A controlled longitudinal study

Epilepsy & Behavior 73 (2017) 23–30

Contents lists available at ScienceDirect

Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh

Cognitive outcome after epilepsy surgery in children: A controlled longitudinal study Valentina Sibilia a, Carmen Barba a, Tiziana Metitieri a, Giovanni Michelini b, Flavio Giordano c, Lorenzo Genitori c, Renzo Guerrini a,d,⁎ a

Pediatric Neurology Unit and Laboratories, Children's Hospital A. Meyer-University of Florence, Viale Pieraccini 24, 50139 Florence, Italy Department of Neuroscience -University of Parma, Via Volturno 39, 43125, Parma, Italy Pediatric Neurosurgery Unit, Children's Hospital A. Meyer-University of Florence, Viale Pieraccini 24, 50139 Florence, Italy d IRCCS Stella Maris, Viale del Tirreno 331, 56128 Calambrone Pisa, Italy b c

a r t i c l e

i n f o

Article history: Received 20 June 2016 Revised 3 March 2017 Accepted 4 March 2017 Available online xxxx Keywords: Epilepsy surgery Cognitive outcome Behavioral outcome

a b s t r a c t Objective: To analyze the determinants of cognitive outcome two years after surgery for drug-resistant epilepsy in a cohort of 31 children when compared to a control group of 14 surgical candidates who had yet to undergo surgery two years after the first neuropsychological assessment. Methods: Controlled longitudinal study including three evaluations of IQ (Intelligence Quotient) scores or GDQ (General Developmental Quotient) for each group depending on the patient's age: prior to surgery (T0), one year (T1) and two years (T2) after surgery for the surgical group; baseline (T0) and one year (T1) and 2 years (T2) after the first evaluation for the control-group. At follow-up, 25 children (80%) of the surgical group were seizure free, while seizure outcome was unsatisfactory in the remaining six (20%). To analyze language, visuomotor skills, memory, reading, visual attention, and behavior, we selected 11 school age children in the surgical group and nine controls. We reported performance prior to (T0) and one year after surgery (T1). Results: There was a significant correlation between earlier age at seizure onset and lower IQ/GDQ at T0 (r = 0.39; p = 0.03) in the overall cohort. IQ/GDQ scores did not significantly differ between the surgical and control groups when analyzed at T0 and T2. However, they evolved differently with an improved developmental trajectory becoming identifiable only in the surgical group (F1,31 = 5.33 p = 0.028; η2 = 0.15). There was also a significant increase of forward digit span (Z = 2.33; p = 0.02) and Rey recall scores (Z = 1.97; p = 0.049) in the surgical school age subgroup at T1 versus T0. Significance: We identified significantly different developmental trajectories in operated versus non- operated children with improved IQ/GDQ scores in operated children only. We also observed a significant increase of digit span scores and Rey recall scores a year after surgery. Further studies including larger samples with longer follow-ups are needed to confirm these preliminary findings. © 2017 Published by Elsevier Inc.

1. Introduction Epilepsy puts children at higher risk for cognitive, behavioral, and psychosocial impairment, compared to the population at large [1,2]. Several studies have demonstrated a high incidence of developmental delay in children with early onset drug-resistant epilepsy, high frequency of seizures, long disease duration, and chronic polytherapy [1–3]. The prognosis of childhood epilepsy is strictly linked to intractability [4,5]. A critical change in treating children with medically intractable epilepsy was the introduction of surgery. The primary goal of surgery in children

⁎ Corresponding author at: Neuroscience Department, Children's Hospital MeyerUniversity of Florence, Viale Pieraccini 24, 50139 Florence, Italy. E-mail address: [email protected] (R. Guerrini).

http://dx.doi.org/10.1016/j.yebeh.2017.03.001 1525-5050/© 2017 Published by Elsevier Inc.

is to achieve seizure control; however, the potentially added benefit for improved neurodevelopment [4] is very encouraging. Previous studies have attempted to identify pre-surgical, surgical or post-surgical variables associated with favorable cognitive and behavioral outcomes in children. Yet the multiple variables, which might be influential, are difficult to uncouple [6]. In particular, surgery at an early age with a consequent shorter duration of epilepsy has been associated with higher cognitive scores after hemispheric disconnections [7] and lobar resections [8,9]. The extent of epileptogenic lesions is an additional factor influencing cognitive development. Children with multilobar lesions are more likely to exhibit global deficits when compared to those with frontal or temporal lesions [2]. Regarding post-surgical variables some authors were unable to find any significant correlation between postoperative seizure outcome and cognitive outcome [2,10,11], while others did [9,12].

24

V. Sibilia et al. / Epilepsy & Behavior 73 (2017) 23–30

Additionally, studies evaluating possible changes of neuropsychological variables after surgery showed conflicting results. Vulnerability of verbal memory function to left temporal lobe resection has been reported [13,14], but has not been confirmed in other series [15,16]. Overall, stability or an improvement was reported more often than a decline in verbal and visual memory after temporal lobectomy [17]. Visual memory may be at risk after extra-temporal surgery in children according to some authors [18] while others reported no change or minor improvements [19,20]. In a cohort of 20 patients, Lendt et al., 1999 [21] documented no change in verbal fluency after surgery while others documented a decline in lexical retrieval after temporal lobe resection suggesting a functional organization comparable to adults [22]. No significant changes in reading achievements were found over the short and long term after epilepsy surgery [23–26]. Altogether, small sample size and heterogeneous etiologies preclude drawing firm conclusions on cognitive and behavioral changes after epilepsy surgery from the available studies [28]. The primary cause of the lack of conclusive evidence on postoperative cognitive outcome in children could possibly be due to flaws in the methodology, with lack of adequate controls being a major weakness in most epilepsy surgery studies. The controls in certain studies were comparable to the surgical group only in terms of age at onset of epilepsy and duration of followup; however not all patients in the control group were eligible for surgery [11,28] nor were they adequately described [28]. In other studies a healthy control cohort was recruited from regular schools [6]. In this study we analyzed the possible determinants of cognitive functioning two years after surgical treatment for drug-resistant seizures in a cohort of 31 children. Based on available literature, we identified age at seizure onset, seizure frequency, age at surgery, etiology and extent of the lesion as possible variables influencing post-surgical cognitive outcome [6]. We also evaluated a control group of 14 children with drug-resistant seizures potentially eligible for surgery, who, two years following the first neuropsychological evaluation, had not yet undergone surgery. Finally, we analyzed whether the IQ/GDQ and specific measures of reading, verbal/visual-spatial memory, naming and visual attention of children in the surgical group evolve differently from those of the control group. 2. Material and methods 2.1. Participants We studied 31 children (20 males/11 females; mean age at surgery 8.73, standard deviation [SD] 4.33; mean age of seizure onset 4.41, SD 3.66) undergoing surgical treatment at the Meyer Children's Hospital between 2007 and 2011 (‘surgical group’). Inclusion criteria for the surgical group were: a) drug-resistant seizures; b) at least two years followup post surgery; c) cognitive evaluations at one year and two years after surgery; d) no previous neurosurgery; e) cognitive skills and behavioral profile of the child allowing reliable testing. We used as a control group 14 children (‘control group’) with drug-resistant seizures (9 females/ 5 males, mean age at first evaluation 10.26, SD 3.3, mean age of seizure onset 4.89, SD 3.44) who were potential surgical candidates and were tested according to the same time scale of the study group, but had not been operated on two years after the first neuropsychological evaluation. Inclusion criteria for the control group were: a) drug-resistant seizures; b) at least two years of neuropsychological follow-up after surgery; c) no previous epilepsy surgery at the time of the second neuropsychological assessment; d) no previous brain surgery; e) cognitive skills and behavioral profile of the child allowing reliable testing. We excluded nine operated patients from the analysis, as we could not complete the neuropsychological testing at T0 due to very severe behavioral disorders or refusal to complete the assessment. Three other patients were excluded as they underwent palliative surgery (callosotomy).

Children in the control group were also eligible for ablative surgery but were not operated on at T2 for different reasons, including parental requests to try all available drugs before considering surgery, parental refusal of surgery or a prolonged waiting period for surgery due to additional non-invasive and invasive tests to assess the link between the epileptogenic zone and eloquent areas. The ethical committee of the Children's Hospital Meyer approved the study. Informed consent forms were obtained for all patients. 2.2. Procedures We performed three evaluations of IQ/GDQ scores for each group: prior to surgery (T0), one (T1) and two years (T2) after surgery for the surgical group and at baseline (T0) and one (T1) and two (T2) years after the first evaluation for the control group. Clinical features, IQ/GDQ scores and surgical variables of the surgical group are summarized in Table 1; clinical and IQ/GDQ data of the control group are summarized in Table 2. The significant improvements or worsening of IQ/GDQ scores (change in IQ greater than 7 points) were defined by excluding the practical effect of test-retest and for higher scores to differences of average stability coefficients across all ages, as reported in technical reference manuals of cognitive scales [29,30]. At T2, statistical analyses were performed only in the 37 children (23 surgical children, 14 control children) in whom a complete and reliable assessment was available (mean of missing IQ/GDQ scores at T0: 89.12, SD: 15.36, range: 57–105; mean of completed IQ/GDQ scores at T0: 74.34, SD: 19.96, range: 45–109). Incompleteness of cognitive and neuropsychological measures was determined by several factors such as, for example, poor cooperation or clinical constraints. Some children in the surgical group underwent neuropsychological assessments in other settings (i.e. rehabilitation clinics), especially if they lived in remote areas from the study center. The analysis did not include the results of these neuropsychological assessments in order to maintain the score's reliability. To analyze how specific neuropsychological variables evolved i.e. language, visuomotor skills, memory, reading, visual attention and behavior, we selected 11 school age children (mean age at T0: 11.89, SD: 2.30; range: 9.25–15.3) in the surgical group (‘surgical subgroup’) and nine (mean age at T0: 12.08, SD: 2.99; range: 8.25–15.5) in the control group (‘control subgroup’). All children within the subgroups were evaluated at T0 and T1 (but not at T2) in a 2 × 2 study. Subgroups were made up of school age children in order to minimize age influence and maximize comparability between the groups. Clinical variables and cognitive features of these 20 subjects are summarized in Tables 1 and 2. General cognitive abilities were tested using GMDS-ER [31] (Griffiths Mental Developmental Scales-Extended Revised) and Italian versions of the Wechsler Scales [32,33]. Non-verbal function was measured with the Leiter-r Scale [34]. Since a relatively rare population and its peculiar clinical features do not permit maximum homogeneity in a cognitive scale, we used a single psychometric measure (IQ/General Developmental Quotient). We evaluated language function using tasks of naming and fluency and assessed the ability to name drawings of objects using the Boston Naming Test [35]. To assess verbal and semantic fluency we used the word fluency test [36]. We used two tasks taken from Wechsler scales (digit span forward and digit span backwards) as part of the assessment of verbal and working memory. Incidental semantic memory was assessed presenting 20 animals names. Each child was requested to name the color of the animal but was not instructed to remember the animal itself. At the end of the incidental learning phase children were asked to report as many animals as they could. We used two tasks (word list reading and non-word list reading) taken from Battery for the evaluation of Dyslexia and Dysorthographia [37] to assess reading abilities. We used the Rey Figure [38] to assess drawing abilities in copying and visual memory.

V. Sibilia et al. / Epilepsy & Behavior 73 (2017) 23–30

25

Table 1 Description of the surgical group at T0 and at follow-up after surgery (T2). Children Age at baseline Age of seizure Age at Laterality Laterality of the IQ/GDQ Seizure Treatment Type of Etiology (months) onset (months) surgery of lesion seizure onset zone frequency T0 at T0 surgery

Seizure outcome Cognitive (Engel class) outcome

1a 2a 3 4a 5 6 7 8 9 10 11a 12a 13a 14a 15a 16a 17 18 19 20 21 22 23 24a 25 26 27 28 29 30a 31

IA IA IA IA ID ID IA ID IA IA IA IA II IA IA ID IC III IA IA IA IA IA IA IA IA III II III II IA

145 154 139 166 32 50 68 69 52 33 128 144 99 184 118 111 131 144 115 112 80 114 132 111 8 36 27 169 20 124 36

120 84 72 156 6 48 16 6 13 30 84 9 96 156 36 89 72 60 60 60 6 84 42 72 2 6 24 15 17 96 5

150 157 152 180 33 51 73 72 57 44 131 164 112 189 122 115 134 146 121 116 84 117 137 138 17 38 33 180 21 126 38

T T T T T F T F ML T T T P F F F T T ML T T F T T F T ML T T T ML

L R L R L R R L R L L L R R L R L L R L R R L R L R L R L L L

84 72 45 106 50 75 80 55 59 69 57 70 101 85 97 91 84 106 55 104 60 85 54 97 71 72 51 101 73 109 105

RD M RD M RD RD D W RD RD RD RD RD RD RD RD W W RD W RD W W M M RD W M M M RD

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

U UET ML U UET UET ML U ML U UET ML ML U U UET U U ML U U U UET U U U ML U U UET UL

Tumor Dysplasia Dysplasia Tumor Dysplasia Dysplasia Dysplasia Dysplasia Other Dysplasia Dysplasia Other Dysplasia Other Other Dysplasia Dysplasia Dysplasia Other Other Dysplasia Dysplasia Tumor Dysplasia Dysplasia Dysplasia Dysplasia Dysplasia Other Tumor Other

I I S I I S – S S I – S – – – – I S S S W W S – I I S S S W –

L left; R right; T temporal; P parietal; F frontal; U unilobar; UET unilobar extra-temporal; ML multilobar; M monthly; W weekly; D daily; RD repeated daily; 2 polytherapy; I improved (IQ N 7scores between T0 and T2); W worsened (IQ b 7scores between T0 and T2); S stable; other: post inflammatory or traumatic brain injury. a Children of surgical subgroup.

To assess visual attention, each child was asked to perform a task of the barrage of target numbers among the target distractors. We used a task from Wechsler scales (Digit Symbol) to measure the processing speed and assessed behavior by asking parents to complete the Child Behavior Check-List (CBCL 6–18) [39]. 2.3. Statistical analysis We analyzed data using the SPSS 18 statistical package. Due to the small sample size, we used nonparametric tests for comparisons

between groups and within subjects and parametric ANOVA to compare IQ in mixed designs. We used Mauchly's sphericity test to assess differences between variances of pairs of groups. For the analysis of specific neuropsychological variables, after verifying homogeneity based on the average age, we compared longitudinally the neuropsychological performances of surgical and control subgroups (N = 20). We assessed whether IQ and neuropsychological findings correlated with several variables (age at seizure onset, age at surgery, type of resection, etiology, localization and lateralization of the seizure onset zone,

Table 2 Description of the control group at T0 and at follow-up (T2). Children

Age at baseline (months)

Age of seizure onset (months)

Laterality of lesion

Laterality of the seizure onset zone

IQ/GDQ

Seizure frequency T0

Treatment at T0

Etiology

Seizure outcome

Engel classa

Cognitive outcome

1b 2b 3b 4 5 6b 7 8b 9 10 11b 12b 13b 14b

138 186 99 163 20 136 136 121 87 99 148 157 126 108

120 114 76 9 60 120 2 72 72 4 24 72 30 47

F T P F T T T F F T T T T F

L L L L R L R R R R L R R R

80 100 126 68 68 85 49 106 92 45 70 94 91 75

RD RD W RD M M M D RD RD M M D W

2 2 1 2 1 2 2 1 1 2 2 2 2 2

Other Dysplasia Tumor Other Dysplasia Other Dysplasia Dysplasia Dysplasia Other Other Dysplasia Other Other

M D W M M W D M D D D M M M

IIIB IVA IVB IIIA IVB IVC IVC IIIA IVA IVA IVC IVB IIIA IIIA

S S W S I S S S W S W S S S

L left; R right; T temporal; P parietal; F frontal; M monthly; W weekly; D daily; RD repeated daily; 1 monotherapy; 2 polytherapy; I improved (IQ N 7scores between T0 and T2); W worsened (IQ b 7scores between T0 and T2); S stable; other: post inflammatory or post traumatic brain injury. a Being equivalent to Engel class. b Children of control subgroup.

26

V. Sibilia et al. / Epilepsy & Behavior 73 (2017) 23–30

seizure frequency, and monotherapy versus polytherapy) at T0, T1 (for IQ/GDQ and neuropsychological variables) and T2 (for IQ/GDQ). We used the Mann-Whitney Test for comparisons between the groups and the Wilcoxon signed-rank test for intra-group comparisons. We report the Wilcoxon signed-rank tests using the normal approximation and the Z-statistic. We used the Wilcoxon signed-rank to assess separately in the surgical and the control subgroups differences between T0 and T1 values of specific neuropsychological variables (naming, phonemic and semantic fluency, digit span forward and backward, digit symbol, semantic memory, copying and visual memory abilities, visual attention, reading times and CBCL). Only within the surgical group differences between T0 and T1 values of the same neuropsychological variables were assessed separately in children with different etiology (focal cortical dysplasia and tumors), topography of the seizure onset zone (left and right lateralization) and therapy (monotherapy and polytherapy). 3. Results 3.1. Seizure outcome Two years after surgery, 25 children (80%) of the surgical group were in Engel class I [20 Engel class IA (64.5%), one in class IC (3.2%), four in class ID (12.9%)], three (9.6%) experienced yearly seizures (Engel class II) and three (9.6%) monthly seizures (Engel class III). In the control group, no child was seizure free two years after the first evaluation. Seven children experienced a reduction of seizure frequency: three from daily or repeated daily to monthly (equivalent to Engel Class IIIA), three from repeated daily to daily (equivalent to Engel Class IVA) and one from weekly to monthly (equivalent to Engel class IIIA). Three children worsened (equivalent to Engel Class IVC) and three remained unchanged (equivalent to Engel Class IVB). At T0, 31 children of the surgical group were in polytherapy, while in the control group four children were in monotherapy and 10 in polytherapy. At T1, 23 children in the surgical group were in polytherapy and eight were in monotherapy. All children of the control group were in polytherapy. At T2, 15 children from the surgical group were in monotherapy, six in polytherapy and two children had discontinued AEDs. All children from the control group were in polytherapy. 3.2. Neuropsychological assessment 3.2.1. IQ/GDQ scores We performed a 2 × 3 factorial ANOVA with one between-subject factor (group: surgical and control) and one within-subject factor (time: T0, T1 and T2). The effect of the interaction was significant (F2,70 = 4.91; p = 0.01; η2 = 0.12), with a slight increase of IQ/GDQ scores in the surgical group (mean score at T0 74.35, SD 19.97; mean score at T2 77.04, SD 21.66) and a slight decline in IQ/GDQ in the control group (mean score at T0 82.07, SD 21.8; mean score at T2 77.00, SD 20.8) (Fig. 1). 3.2.2. IQ/GDQ and age at seizure onset We observed a significant correlation between earlier age at seizure onset and lower IQ/GDQ at T0 (r = .391; p = 0.030) in both groups. 3.2.3. IQ/GDQ and surgery There was a significant difference between preoperative IQ/GDQ scores at T0 and type of resection (F2,20 = 4,26; p = .03; η2 = 0.30): patients with temporal lobe resections had IQ/GDQ scores significantly higher than those with multilobar or extra-temporal resections (p = 0.022). However, the type of resection did not significantly influence IQ/GDQ evolution. In patients undergoing temporal resections,

Fig. 1. IQ/GDQ scores evolved differently over time (between TO and T2) in surgery group and control group.

average IQ/GDQ scores increase from T0 (mean 81.57, SD 17.08) to T2 (mean 85.79, SD 19.50) was non-significant. Children with multilobar resections exhibited a slight, non-significant decline of IQ/GDQ from T0 (mean 56.00 SD 9.31) to T2 (mean 54.8, SD 12.2). 3.2.4. IQ/GDQ and etiology We found no correlation between etiology of epilepsy and IQ/GDQ scores in any evaluation of either group. 3.3. Specific neuropsychological variables For each independent variable (etiology, topography of the seizure onset zone and therapy), we performed Wilcoxon tests for all dependent variables (naming, phonemic and semantic fluency, digit span forward and backward, digit symbol, semantic memory, copying and visual memory scores, visual attention, reading times and CBCL). Mean z-scores of neuropsychological testing for both subgroups are summarized in Table 3. For digit symbols we used scaled scores. For the digit span we used the raw scores in order to differentiate the shortterm memory capacity on forward span from attentional capacity on backward span as well as eliminate the risk of losing any information [40]. One child in the surgical subgroup and one child from the control subgroup did not complete the reading tasks at T0 and T1 due to inability; two children from the control subgroup did not complete a visual attention task at T1 and four children in the same subgroup did not complete the reading tasks at T1, due to poor cooperation. No significant differences were noted comparing the surgical and control groups at baseline and at first year follow-up. When comparing performances of the surgical subgroup at T1 versus T0, we observed a significant increase of forward span scores (Z = 2.33; p = 0.02) (Fig. 2) and visuospatial memory (Z = 1.97; p = 0.049) (Fig. 3). In the same group, we observed a slight but insignificant increase of visuomotor integration speed at T1 (mean 9.45, SD 3.8) compared to T0 (mean 8.91, SD 2.7). Individual comparisons of pre-and post-surgical scores showed a normed score increase N 1SD on visual memory measures in 2 out of 3 patients (66.7%) with frontal lobe resections and in 1 out of 7 patients (14.3%) with temporal lobe resections, at T1. Increase of forward span at T1 occurred in 2 out of 3 patients (66.7%) with frontal lobe resections and in 4 out of 7 patients (57.2%) with temporal lobe resections. At T1, the surgical subgroup showed an improvement in externalizing, internalizing and total behavioral disorders, which did not reach statistical significance. 3.3.1. Etiology The analysis of the surgical subgroup detected a slight increase of forward span in patients with focal cortical dysplasia (Z = 1.73;

V. Sibilia et al. / Epilepsy & Behavior 73 (2017) 23–30

27

Table 3 Mean scores of neuropsychological tests for both subgroups. Time

Phonemic fluency Semantic fluency Rey copy Rey recall Naming Semantic memory Word list Non-word list Visual attention Span forward† Span backward† Digit symbol•

T0 T1 T0 T1 T0 T1 T0 T1 T0 T1 T0 T1 T0 T1 T0 T1 T0 T1 T0 T1 T0 T1 T0 T1

N

11 11 11 11 11 11 11 11 11 11 11 11 10 10 10 10 10 10 11 11 11 11 11 11

Surgical subgroup M

SD

−1.81 −1.26 −1.20 −1.24 −0.80 −0.81 −1.60 −0.96 −1.79 −1.76 −1.85 −1.02 −3.05 −3.01 −2.58 −2.06 −1.27 −1.36 3.91 4.55 2.91 3.27 8.91 9.45

1.48 0.98 1.31 0.70 1.35 1.68 1.26 0.99 1.17 1.01 1.16 0.90 4.18 3.29 3.41 2.45 1.62 1.29 1.14 1.29 0.83 1.10 3.63 3.35

Z

N

Control subgroup M

SD

9 9 9 9 9 9 9 9 9 9 9 9 8 4 8 4 9 7 9 9 9 9 9 9

−1.43 −1.78 −0.93 −1.13 −0.74 −0.96 −1.84 −1.99 −1.18 −1.51 −1.07 −0.78 −0.51 −0.70 −0.13 −0.22 −0.56 −1.27 4.00 3.89 3.22 3.11 7.33 8.00

1.33 1.83 0.85 1.21 2.44 1.84 1.78 1.41 2.07 2.06 1.68 1.51 0.92 0.48 1.13 1.31 1.24 1.47 0.87 1.05 1.30 1.05 4.55 4.41

p-Value

−1.231

0.218

.000

1.000

−.568

0.570

−1.970

0.049⁎

.000

1.000

−1.827

0.068

−.420

0.674

−1.535

0.125

−1.841

0.066

−2.333

0.020⁎

−1.414

0.157

−1.023

0.306

Z

p-Value

−0.676

.499

−0.762

.446

0.000

1.000

−0.423

.672

−0.339

.735

−0.315

.752

−0.184

.854

−1.000

.317

−0.712

.476

−1.000

.317

−0.577

.564

−0.682

.495

⁎ Significant improvement at T1 (p b 0.05). † Raw scores. • Scaled scores.

p = 0.083) at T1. In patients with tumors we observed a slight reduction in reading time (Z = 1.89; p = 0.059), although they remained seriously compromised. These findings did not reach statistical significance.

4. Discussion

3.3.3. Behavior At T1, we observed a significant decrease of internalizing (Z = −2.03; p = 0.042), externalizing (Z = −2.03; p = 0.043) and total behavioral disorders (Z = −2.02; p = 0.043) in the surgical group participants in polytherapy.

This was a controlled longitudinal study on neuropsychological and behavioral outcome after surgery for drug-resistant epilepsy in children. Our longitudinal comparative analysis revealed that the overall IQ/GDQ, evolved in a different manner with a significantly improved developmental trajectory becoming identifiable only in children who underwent surgery. Lacking an adequate control group was one of the major weaknesses in previous studies on postoperative seizure outcome in children [28]. Longitudinal comparisons had been carried out in previous studies [15,27] yet no significant changes emerged after a short follow-up. Nevertheless, surgery was not an option for the majority of the participants because of discordance in the EEG and SPECT or PET changes and MRI findings. In contrast, the control group in our study was made up of children eligible for surgery and representative of an equally complex population for clinical variables and neuropsychological profile at baseline. Although an overall gain of developmental skills would require a

Fig. 2. Children of the surgical subgroup (N = 11) performed better in a task involving verbal short-term memory at T1 (mean span forward: 3.91) compared at T0 (mean span backward: 4.55).

Fig. 3. Children of the surgical subgroup (N = 11) performed better in tasks involving visuospatial memory at T1 (Z = −1.84) compared to T0 (Z = − 0.96). Children of the control subgroup (N = 9) exhibited a slight reduction of their performances at T1 (Z = −1.99) compared at T0 (Z = −1.60).

3.3.2. Topography of the seizure onset zone At T1, children of the surgical subgroup with left lateralization exhibited significant performance increases compared to T0 both in phonemic fluency (Z = 2.03; p = 0.042) (Fig. 4) and semantic memory (Z = −2.04; p = 0.041) (Fig. 5). In children of the surgical subgroup with right lateralization, the increase of the span and semantic fluency did not reach statistical significance. In the same group, we detected a slight decline of visual attention, with no statistical significance.

28

V. Sibilia et al. / Epilepsy & Behavior 73 (2017) 23–30

Fig. 4. Children of the surgical subgroup (N = 11) with left lateralization exhibited a significant performance increase of phonemic fluency at T1 (Z = −1.33) compared to T0 (Z = −2.50).

lengthier post-surgical observation period in order to reach statistical significance and clinical relevance [10,41,42], our results suggest that even a relatively short follow-up detects different developmental trajectories. Unlike other studies [15,43], we found 80% of the surgical patients to be seizure free, whereas not one child in the control group was seizure free at T2. We can hypothesize that the higher percentage of seizure free patients in the surgical group compared to the control group at T2 might have influenced such a difference in developmental trajectory. In our study, IQ scores were missing in 26% of the operated group and in none of controls. However, the average IQ/GDQ scores of the operated children who did not complete follow-ups were higher than the average IQ/GDQ scores of the operated children who did complete them. Therefore, children with missing data differed from those who completed the follow-ups, in that they lived in different parts of Italy, and more importantly, were seizure-free. Since seizure freedom may have a positive impact on post-surgical cognitive scores [9,12], lack of data from children who did not complete the follow-ups might have negatively affected the estimates of the IQ/GDQ outcome in the surgical group. The longitudinal analysis of specific neuropsychological variables revealed that surgically treated children exhibited a significant increase of forward span and Rey recall scores one year after surgery compared to the control group. These findings might suggest an improvement in verbal and visuospatial memory, in line with a previous retrospective report [44]. Conversely, neither digits forward nor digits backward revealed significant changes 12 months after surgery in the series described by Meekes et al. [13]. Post-surgical improvement has been interpreted as the release of reserve capacities, which were suppressed

Fig. 5. Children of the surgical subgroup (N = 11) with left lateralization exhibited a significant performance increase of semantic memory at T1 (Z = −0.96) compared to T0 (Z = −2.56).

or disturbed by epilepsy [45]. In our sample a gain in test scores could be explained by seizure control and influence of improvements in children with frontal lesions, which were observed through qualitative analysis. The increase of visual memory scores can be partly due to improvements of the skills of planning and constructional praxis, which are typically impaired in patients with frontal lobe damage. We cannot exclude an influence of the higher number of subjects in polytherapy at T1 in the control group as compared to the surgical group (100% vs. 74%) on digit forward scores. In previous studies, polytherapy has indeed been associated with lower scores in working memory and processing speed [46].Finally, in our study surgical patients with left sided seizure onset zone showed a significant improvement in task of phonemic fluency and semantic memory one year after surgery. Specifically, we observed low scores (below average) at T0 and a significant improvement at T1. Conversely, patients with right lateralization showed no changes from T0 to T1, suggesting hemispheric specificity in performance recovery after surgery. Recently, post-operative decline in semantic memory has been reported in children after left temporal lobectomy [47], and left temporal lobe resections were associated with gains in phonemic fluency [48]. However, seemingly contradictory findings could be influenced by the specific tests used. In this study, we used a specific measurement of semantic memory, involving categorical knowledge, while the vocabulary subtest, used in previous reports, is a general measurement of word knowledge and verbal concept formation. Little is known regarding behavioral changes after epilepsy surgery [49,50]. Using the CBCL scale, we observed that both the surgical and control groups continued to experience elevated rates of behavioral problems at T1, although children of the surgical group, 64.5% of which were seizure free after operation, exhibited a non-significant improvement in clinical scales (internalizing, externalizing and total). This finding is not unexpected. Several studies reported elevated rates of behavioral problems in children with medically refractory epilepsy, with little or no changes after surgery [51,52]. On the other hand, Lendt et al. [53] reported a significant improvement in seizure free patients concerning internalizing, externalizing and attention scales, one year after surgery. There are also multiple factors likely contributing to behavioral problems, including anti-epileptic medications, seizures, interictal EEG activity and psychosocial factors [49,51]. Polytherapy has been associated with behavioral problems [54,55], but did not represent a significant factor according to Austin et al. [56]. We observed a significant decrease of internalizing, externalizing and total disorders one year after surgery in children on polytherapy. While ratings provided by parents are likely to offer the best estimation of children's behavior in the general population, in cases of successful surgical intervention changes of some psychosocial conditions (for example family stress) can influence observations. Indeed, although the CBCL scale is a measure of behavior, multiple informants are needed to assess emotional functioning in patients with epilepsy [57]. In our study we have also analyzed possible clinical determinants of cognitive outcome two years after surgical treatment for drug-resistant seizures. We observed a significant relationship between early age at seizure onset and low IQ/GDQ scores at T0. This finding is in line with some retrospective studies describing severe cognitive impairment in early onset epilepsies [3]. Freitag and Tuxhorn [2] suggested that children with multilobar cortical dysplasia had a greater global deficit than those with unilobar lesions. We observed that children undergoing unilobar resections exhibited higher IQ/GDQ scores at baseline compared to those undergoing multilobar resections, irrespective of etiology. In general, we could not find a major effect of etiology on IQ/GDQ either at baseline or at follow-up. Conversely, in Pulsifer's series [58], which included only patients who had hemispherectomy, etiology was the most significant predictor of cognitive skills at follow-up, with Rasmussen and vascular groups scoring significantly higher than the cortical dysplasia group in IQ scores and receptive language. These divergent results might be the

V. Sibilia et al. / Epilepsy & Behavior 73 (2017) 23–30

result of the small size and limited etiological variability in our sample: 55% of our patients had dysplasia, 33% had tumors and 11% had post inflammatory or post traumatic brain injuries. 5. Limitations of the study Due to the etiologic heterogeneity of our cohort we were able to evaluate only a fraction of children who were candidates for epilepsy surgery. Heterogeneity made it even more difficult to explore potential differences because of epilepsy-related variables and limited the choice of available neuropsychological measures. The age difference between the two groups prevented us from using the same psychometric tests for IQ/GDQ assessment. It could be argued that the loss or gain in IQ/GDQ scores in some children was the effect of having used different tasks rather than an actual improvement or impairment of cognitive functioning. When serial cognitive assessments are performed in children, the test-retest reliability, the practice effect and the child's maturation must be considered [42]. In order to minimize the risk that the increase in IQ might affect practice, we referred to differences of average stability coefficients across all ages obtained from test-retest, using a higher cut-off than those specified in the manuals. Furthermore, the control and surgical subgroups were agematched to minimize any possible influence the mean age difference might have, with a gain in term of homogeneity in neuropsychological measures. 6. Conclusions Despite the limitations (i.e. the relatively small sample size and short follow up), our controlled longitudinal study allowed us to scrutinize possible predictors of neuropsychological and behavioral outcomes after surgery. We could demonstrate a significant correlation between earlier age at seizure onset and lower IQ/GDQ at T0 in surgery and control groups as well as between preoperative IQ/GDQ scores and the type of resection. Except for verbal short-term memory and visuospatial memory, we observed no significant differences in specific neuropsychological variables one year after surgery. Moreover, we identified different developmental trajectories two years after surgery with improved IQ/DGQ scores only in children who underwent surgery. Finally, we observed a significant correlation between earlier age at seizure onset and lower IQ/GDQ at T0 in both groups, as well as between preoperative IQ/GDQ scores and type of resection. Further studies with larger samples are needed to specify these preliminary results. Funding This research was supported by the Italian Ministry of Health and Tuscany Region (RF-2010-2309954) ‘Reorganization of cortical function after surgery for lesional epilepsy in children’ to C.B and by the European Union Seventh Framework Programme FP7/2007–2013 under the project (DESIRE) (grant agreement 602531) to RG and the Pisa Foundation (Project 133/11) to RG. Declaration of interest Drs Guerrini and Barba received support from the Italian Ministry of Health and the Tuscan Region (RF-2010-2309956), the European Union Seventh Framework Programme FP7/2007–2013 under the project (DESIRE) (grant agreement 602531) and the Pisa Foundation (Project 133/11). There are no conflicts of interest, nor anything to disclose with any of the other authors.

29

References [1] Bailet LL, Turk WR. The impact of childhood epilepsy on neurocognitive and behavioral performance: a prospective longitudinal study. Epilepsia 2001;41:426–31. [2] Freitag H, Tuxhorn I. Cognitive function in preschool children after epilepsy surgery: rationale for early intervention. Epilepsia 2005;46:561–7. [3] Vasconcellos E, Wyllie E, Sullivan S, Standford L, Bulacio J, Kotagal P, et al. Mental retardation in pediatric candidates for epilepsy surgery: the role of early seizure onset. Epilepsia 2001;42:268–74. [4] Cross JH, Jayavar P, Nordli D, Deladande O, Duchowny M, Wieser HG, et al. Proposed criteria for referral and evaluation of children for epilepsy surgery: recommendations of the subcommission for pediatric epilepsy surgery. Epilepsia 2006;47: 952-9. [5] Kwan P, Arzimanoglou A, Berg AT, Brodie MJ, Allen Hauser W, Mathern G, et al. Definition of drug resistant epilepsy: consensus proposal by the ad hoc task force of the ILAE commission on therapeutic strategies. Epilepsia 2010;6:1069-77. [6] Van Schooneveld MM, Braun KP. Cognitive outcome after epilepsy surgery in children. Brain Dev 2013;35:721–9. [7] Jonas R, Nguyen S, Hu B, Asarnow RF, Lo Presti C, Curtiss S, et al. Cerebral hemispherectomy: hospital course, seizure, developmental, language, and motor outcomes. Neurology 2004;62:1712–21. [8] Westerveld M, Sass KJ, Chelune GJ, Jacques TS, Harkness W, Vargha-Khadem F. Temporal lobectomy in children: cognitive outcome. J Neurosurg 2000;92:24–30. [9] Bourgeois M, Crimmins DW, De Oliveira RS, Arzimanoglou A, Garnett M, Roujeau T. Surgical treatment of epilepsy in Sturge-Weber syndrome in children. J Neurosurg 2007;106:20–8. [10] D'Argenzio L, Colonnelli MC, Harrison S, Jacques TS, Harkness W, Vargha-Khadem F. Cognitive outcome after extratemporal epilepsy surgery in childhood. Epilepsia 2011;52:1966–72. [11] Skirrow C, Cross JH, Cormack F, Harkness W, Varga-Khadem F, Baldeweg T. Longterm intellectual outcome after temporal lobe surgery in childhood. Neurology 2011;76:1330–7. [12] Miranda C, Smith ML. Predictors of intelligence after temporal lobectomy in children with epilepsy. Epilepsy Behav 2001;2:13–9. [13] Meekes J, Braams O, Braun PJ, Jennekens-Schinkel A, Van Nieuwenhuizen O. Verbal memory after epilepsy surgery in childhood. Epilepsy Res 2013;107:146–55. [14] Jambaqué I, Dellatolas G, Fohlen M, Bulteau C, Watier L, Dorfmuller G, et al. Memory functions following surgery for temporal lobe epilepsy. Neuropsychologia 2007;45: 2850–62. [15] Skirrow C, Cross JH, Harrison S, Cormack F, Harkness W, Coleman R, et al. Temporal lobe surgery in childhood and neuroanatomical predictors of long-term declarative memory outcome. Brain 2015;138:80–93. [16] Helmstaedter C, Gleissner U, Zentner J, Elger CE. Functional plasticity after left anterior temporal lobectomy: reconstitution and compensation of verbal memory functions. Epilepsia 1998;39:399–406. [17] Lah S. Neuropsychological outcome following focal cortical removal for intractable epilepsy in children. Epilepsy Behav 2004;5:804-17. [18] Meekes J, Braams OB, Braun KPJ, Jennekens-Schinkel A, van Rijen PC, Alpherts WCJ, et al. Visual memory after epilepsy surgery in children: a standardized regressionbased analysis of group and individual outcomes. Epilepsy Behav 2014;36:57–67. [19] Lendt M, Gleissner U, Helmstaedtr C, Sassen R, Clussman H, Elger CE. Neuropsychological outcome in children after frontal lobe epilepsy surgery. Epilepsy Behav 2002; 3:51-9. [20] Mabbott DJ, Smith ML. Memory in children with temporal or extra-temporal excisions. Neuropsychologia 2003;41:995–1007. [21] Lendt M, Helmstaedter C, Elger CE. Pre- and postoperative neuropsychological profiles in children and adolescents with temporal lobe epilepsy. Epilepsia 1999;40: 1543-50. [22] Vega C, Brenner LA, Madsen J, Bourgeois B, Waber DP, Boyer K. Lexical retrieval preand posttemporal lobe epilepsy surgery in a pediatric sample. Epilepsy Behav 2015; 42:61–5. [23] Blanchette N, Smith ML. Language after temporal or frontal lobe surgery in children with epilepsy. Brain Cogn 2002;48:280-4. [24] Smith ML, Elliot IM, Lach L. Cognitive, psychosocial, and family function one year after pediatric epilepsy surgery. Epilepsia 2004;45:650-60. [25] Fastenau PS, Jianzhao S, Dunn DW, Austin JK. Academic underachievement among children with epilepsy: proportion exceeding psychometric criteria for learning disability and associated risk factors. J Learn Disabil 2008;41:195–207. [26] Puka K, Khattab M, Kerr EN, Smith ML. Academic achievement one year after resective epilepsy surgery in children. Epilepsy Behav 2015;43:117-21. [27] Puka K, Smith ML. Academic skills in the long term after epilepsy surgery in childhood. Epilepsy Behav 2016;62:97–103. [28] Téllez-Zenteno JF, Hernàndez Ronquillo L, Wiebe S. Surgical outcomes in lesional and non-lesional epilepsy: a systematic review and meta-analysis. Epilepsy Res 2010;89:310–8. [29] Wechsler D. Wisc-IV: technical and interpretive manual. London: Pearson; 2003. [30] Wechsler D. Wppsi-III: technical manual. Pearson: San Antonio; 2002. [31] Griffiths R. Griffiths mental developmental scales-extended revised. Giunti OS: Firenze; 2006. [32] Wechsler D. Wechsler preschool and primary scale of intelligence. Giunti OS: Firenze; 2008. [33] Wechsler D. Wechsler intelligence scale for children. Giunti OS: Firenze; 2013. [34] Roid GH, Miller LJ. Leiter international performance scale-revised. Giunti OS: Firenze; 1997. [35] Riva D, Nichelli F, Devoti M. Developmental aspects of verbal fluency and confrontation naming in children. Brain Lang 2000;71:267–84.

30

V. Sibilia et al. / Epilepsy & Behavior 73 (2017) 23–30

[36] Korkman M, Kirk U, Kemp S. NEPSY: a developmental neuropsychological assessment manual. San Antonio, TX: The Psychological Corporation; 1998. [37] Sartori G, Job R, Tressoldi E. DDE-2. Batteria per la valutazione della Dislessia e della Disortografia Evolutiva. Giunti OS: Firenze; 2010. [38] Rey A. Reattivo della figura Complessa. Organizzazioni Speciali: Florence; 1979. [39] Achenbach TM, Rescorla LA. Manual for the ASEBA school-age forms profiles. Burlington Vt USA: University of Vermont, Research Center for Children, Youth & Families; 2001. [40] Lezak MD, Howieson DB, Loring DW. Neuropsychological assessment. 4th ed. Oxford: University Press; 2004. [41] Gilliam F, Wyllie E, Kashden J, Faught E, Kotagal P, Bebin M, et al. Epilepsy surgery outcome: comprehensive assessment in children. Neurology 1997;48: 1368–74. [42] Viggedal G, Olsson I, Carlsson G, Rydenhag B, Uvebrant P. Intelligence two years after epilepsy surgery in children. Epilepsy Behav 2013;29:565–70. [43] Edelvik A, Rydenhag B, Olsson I, Flink R, Kumlien E, Kallen K, et al. Long-term outcomes of epilepsy surgery in Sweden: a national prospective and longitudinal study. Neurology 2013;81:1244-51. [44] Gleissner U, Sassen R, Schramm J, Elger CE, Helmstaedter C. Greater functional recovery after temporal lobe epilepsy surgery in children. Brain 2005;128:2822–9. [45] Helmstaedter C, Kurthen M, Lux S, Reuber M, Elger CE. Chronic epilepsy and cognition: a longitudinal study in temporal lobe epilepsy. Ann Neurol 2003;54: 425–32. [46] Reilly C, Atkinson P, Das KB, Chin RF, Aylett SE, Burch V, et al. Cognition in school-aged children with “active” epilepsy: a population-based study. J Clin Exp Neuropsychol 2015;37:429–38.

[47] Lah S, Smith ML. Verbal memory and literacy outcomes one year after pediatric temporal lobectomy: a retrospective cohort study. Epilepsy Behav 2015;44:225–33. [48] Sherman EM, Wiebe S, Fay-McClymont TB, Tellez-Zenteno J, Metcalfe A, HernandezRonquillo L. Neuropsychological outcomes after epilepsy surgery: systematic review and pooled estimates. Epilepsia 2011;52:857-69. [49] Rodenburg R, Stams G, Meijier A, Aldenkamp A, Dekovic M. Psychopathology in children with epilepsy: a meta-analysis. J Pediatr Psychol 2005;30:453–68. [50] Austin JK, Caplan R. Behavioral and psychiatric comorbities in pediatric epilepsy: toward an integrative model. Epilepsia 2007;48:1639–52. [51] Elliot I, Lach L, Kadis DS, Smith ML. Psychosocial outcomes in children two years after epilepsy surgery: has anything changed? Epilepsia 2008;49:634–41. [52] McLellan A. Psychopathology in children with epilepsy before and after temporal resection. Dev Med Child Neurol 2005;47:666–72. [53] Lendt M, Helmstaedter C, Kuczaty S, Schramm J, Egler CE. Behavioural disorders in children with epilepsy: early improvement after surgery. Neurol Neurosurg Psychiatry 2000;69:739–44. [54] Freillinger M, Reisel B, Reiter E, Zelenko M, Hauser E, Seidl R. Behavioral and emotional problems in children with epilepsy. J Child Neurol 2006;21:939–45. [55] Hermann BP, Seidenberg M. Childood epilepsies. Neuropsychological, psychological and intervention aspects. New York: Wiley; 1989. [56] Austin JK, Huster GA, Dunn DW, Risinger MW. Adolescents with active or inactive epilepsy or asthma: a comparison of quality of life. Epilepsia 1996;37:1228–38. [57] Puka K, Smith ML. Predictors of long-term quality of life after pediatric epilepsy surgery. Epilepsia 2015;56:873–81. [58] Pulsifer MB, Brandt J, Salorio CF, Vining EP, Carson BS, Freeman JM. The cognitive outcome of hemispherectomy in 71 children. Epilepsia 2004;45:243–54.