Monique MJ van Schooneveld, Kees PJ Braun and J Helen Cross
Almost 30% of children with epilepsy continue to have troublesome seizures despite trying antiepileptic drugs (AEDs). For children with lesional focal epilepsy, uncontrolled by medical treatment (i.e. failure of, or disabling side effects from, two or three appropriate drugs), epilepsy surgery has increasingly become an important treatment option over the past 30 years (Ryvlin et al. 2014; Pestana Knight et al. 2015; Lamberink et al. 2015). It is now considered possible to offer surgery much earlier in the course of the epilepsy and to younger children. Furthermore, developmental delay is no longer a contraindication for epilepsy surgery (Cross et al. 2006; 2013; Ryvlin et al. 2014).
The most frequently used surgical techniques intended to abolish seizures and cure the epilepsy are functional hemispherectomy (hemispherotomy) and resective surgery. Hemispherectomy, a disconnection of one-half of the brain with limited resection, is indicated when seizures originate from, and the underlying pathology affects, an entire cerebral hemisphere. Children who may benefit from this procedure usually have a preexisting hemiparesis as a result of the severe contralateral hemispheric pathology, which has mostly been present from birth. A focal, tailored, resection or lobectomy is considered when the area of the brain thought to be responsible for the seizures is well circumscribed or confined to a single lobe. This in the majority of cases concerns the temporal lobe. Increasingly, however, children undergo epilepsy surgery for extratemporal lobe epilepsy. Resection or disconnection of more than one lobe is called multilobar or subhemispheric surgery. Additional types of surgery that are not intended to cure epilepsy but to improve seizure status (the so-called palliative procedures) are multiple subpial transections and corpus callosotomy. These techniques will not be discussed here in detail.
The beneficial effect of paediatric epilepsy surgery on seizure control in carefully selected candidates has been well established (Spencer and Huh 2008; Ryvlin et al. 2014). People with pharmacoresistant epilepsy, however, often have additional developmental problems, due to the underlying epileptogenic pathology, the frequent seizures themselves or the chronic use of (in most cases) multiple AEDs (Lah 2004), particularly in infancy and early childhood. Both the anticipated seizure freedom and the expected improvement of postoperative cognitive development contribute to the final decision to accept or reject a candidate for epilepsy surgery (Perry and Duchowny 2013).
Early surgical intervention—thus a shorter duration of active epilepsy—has been reported to correlate with a higher chance of reaching seizure freedom (Englot et al. 2013a, 2013b; Lamberink et al. 2015). It is not unexpected that early intervention also reduces the eventual severity of cognitive impairment. In addition, given the known cognitive side effects of AEDs (Park and Kwon 2008; Mula and Trimble 2009) and the improvement of cognitive function after discontinuation of AEDs in non-surgical cohorts (Aldenkamp et al. 1998; Hessen et al. 2006; Lossius et al. 2008), AED withdrawal, which is possible after anticipated curative surgery, increases the probability of postoperative cognitive improvement. In fact, it has convincingly been demonstrated that AED reduction improves postoperative intelligence and other domains of cognitive outcome (Skirrow et al. 2011; Meekes et al. 2013; van Schooneveld et al. 2013; Boshuisen et al. 2015). Therefore, the ultimate goal of epilepsy surgery in children should be not only to achieve freedom from seizures and enable AED discontinuation, but also to improve developmental capacities (Shields 2000; Cross 2002; van Schooneveld and Braun 2013). Postoperative outcome after childhood epilepsy surgery is therefore preferably expressed in terms of seizure freedom; cognitive outcome can be considered an equally important outcome measure (Spencer and Huh 2008; Ryvlin et al. 2014). Age-appropriate neuropsychological assessments are a mandatory aspect of the pre- and postsurgical evaluation (Cross et al. 2006).
This chapter provides an overview of what is known about neuropsychological assessment and cognitive outcome and its predictors after hemidisconnective or resective epilepsy surgery in children. Because presurgical cognitive functioning significantly correlates with postsurgical cognitive outcome, we included a section about cognitive (dys)function and its determinants in children with epilepsy, prior to surgical treatment. We summarize the current knowledge about different domains of cognitive outcome after surgery. As children with complex epilepsy and ongoing seizures have high rates of associated behaviour disorders (Davies et al. 2003), particularly those with focal epilepsy that are candidates for epilepsy surgery (McLellan et al. 2005; Colonnelli et al. 2012), a section on behaviour is also included.
The purpose of the assessment is to evaluate cognitive, behavioural and emotional consequences of the epilepsy and its treatment. It should preferably encompass all the domains of cognitive development: intelligence, memory and learning, language, visual spatial skills, processing speed, attention and executive functions (Lassonde et al. 2000; Battaglia et al. 2006). The neuropsychological assessment of infants and children, however, is complicated for a number of reasons. Given the extensive neurodevelopmental trajectory that children undergo, different periods of development require different skills, and not all cognitive functions can be reliably assessed in early childhood. Furthermore, there is a high incidence of cognitive disability in paediatric epilepsy surgery candidates. Accordingly, the test battery varies depending on the child’s age and skill level. Different tests used at different ages and varying levels of cognitive functioning are difficult to compare and hamper accurate longitudinal monitoring of functions over time. Intelligence—or in younger children indices of general cognitive development—is the most common measure of cognitive outcome after paediatric epilepsy surgery (Lah 2004). Less research has addressed specific cognitive functions, such as memory, language, visual spatial skills, attention, and executive functions after surgery in children. Assessments should be performed at regular intervals, before and at least once after surgery. In most studies, postsurgical follow-up intervals are limited to either 1 or 2 years, with only few reports describing the long-term cognitive outcome, or more than 5 years after paediatric epilepsy (Tellez-Zenteno et al. 2007; D’Argenzio et al. 2011; Skirrow et al. 2011; Viggedal et al. 2012; van Schooneveld et al. 2016). A more extensive follow-up may be required to establish the full developmental impact of surgery, as it has been suggested that epilepsy surgery, ending a shorter or longer time of severe epilepsy, needs an unusually long time to manifest cognitive improvement or recovery (Freitag and Tuxhorn 2005; Skirrow et al. 2011).
Although non-invasive tests—such as functional magnetic resonance imaging (fMRI) (Wang et al. 2012)—are increasingly used to lateralize or localize language functions, in children who are too young, impaired or uncooperative to reliably undergo fMRI, the neuropsychologist is invaluable in testing the child during the intracarotid amybarbital—the so-called Wada—test (anesthetizing one hemisphere), by presenting test material that the child has to recognize, name or read, as well as by testing comprehension of simple questions (Jansen et al. 2002). In contrast to speech lateralization, the assessment of memory function during the Wada test is less successful in children (Jones-Gotman et al. 2010). Other invasive procedures, such as functional mapping by means of cortical stimulation during intracranial monitoring with subdural grid or depth electrodes or during awake surgery, may be indicated to localize eloquent areas in relation to the epileptogenic zone, when alternative and non-invasive methods are inconclusive or when the epileptogenic focus lies in close proximity to what is believed to be functional cortex. The neuropsychologist has a critical role in testing eloquent functions during the cortical stimulations.
In the neuropsychological follow-up of epilepsy surgery patients, there are a large variety of tests in use. To compare the results of neuropsychological assessments in multicenter cohorts, there is a need for standardized core neuropsychological test batteries, which can be used across epilepsy surgical centers. However, although a core battery of tests can be valuable, a flexible battery with additional tests tailored to the clinical reference questions and the individual needs is especially important in the assessment of children, as tests need to be age sensitive and appropriate for the child’s development (Wilson et al. 2015).
Presurgical cognitive functioning in childhood epilepsy
Subnormal cognitive function (intelligence quotient [IQ] <80) is apparent in approximately 25% of all children with newly diagnosed epilepsy (Berg et al. 2008). Younger age at onset of epilepsy, a lesional cause, epileptic encephalopathy, and continuing AEDs have been reported to independently predict suboptimal cognitive functioning (Berg et al. 2008). Intellectual abilities may be particularly vulnerable to seizure-related disturbances during the period of rapid development in early childhood (Bjørnaes et al. 2001). In a group of children aged less than 3 years with refractory focal epilepsy due to cortical malformations, 89% had a presurgical IQ less than 85, with 64% of the children having impairments in the severe range (IQ <50) (Ramantani et al. 2013). Abnormal IQ levels, or even severe intellectual disability (IQ <50), have also been reported in children with temporal lobe epilepsy who present as candidates for surgical resection (Cormack et al. 2007; de Koning et al. 2009; Skirrow et al. 2011). Children with refractory extratemporal epilepsy, with lesions in frontal, parietal or occipital lobes, also frequently have preoperative learning impairments (D’Argenzio et al. 2011). Many studies have shown that 74%–88% of childhood hemispherectomy candidates have a severe cognitive delay (Devlin et al. 2003; Jonas et al. 2004; Boshuisen et al. 2010). Furthermore, case histories of infants and children who have undergone hemispherectomy have revealed that cognitive development had arrested or even regressed following the onset of epilepsy in 82% of these patients (van Schooneveld et al. 2011).
Infants and young children with frequent seizures are particularly at risk of developing a so-called epileptic encephalopathy, by definition implying that developmental arrest or regression can at least partially be attributed to the ongoing severe epileptic activity, and not merely to the underlying epileptogenic pathology (Berg et al. 2010). At a young age, an encephalopathic course may occur in association with any cause of epilepsy (Berg et al. 2010).
Postsurgical cognitive function in paediatric epilepsy
Cognitive profiles have most often been reported following temporal lobe resection or hemispherectomy. Studies on cognitive outcome after frontal, parietal or occipital resections are rather scarce, often including only small groups of children.
GENERAL MEASURES OF INTELLIGENCE AND DEVELOPMENT
In the relevant literature, cognitive functioning is usually defined as intelligence or development, with cognitive improvement being expressed as an increase of Mental Developmental Index (MDI) of 5 to 15, developmental quotient (DQ), or IQ points or as a shift from one to another predefined category of cognitive function, such as from below average to low average. Viewing case by case, a significant number of children (61% of the hemispherectomized and 70% of those who underwent a focal resection) have an almost unchanged cognitive level or remain in their presurgical category of functioning (for review, see van Schooneveld and Braun 2013). This would imply a maintained cognitive trajectory, because if children preserve their presurgical MDI/DQ/IQ level, they must continuously acquire new skills and knowledge over time, as IQ is measured relative to normal peers and is corrected for increasing age. However, many young children, particularly those who undergo hemispherectomy, are severely delayed, both before and after surgery. Test scores of these children usually lie in the lower end of the score distribution (MDI or IQ values < 55), where it is known that instruments do not measure accurately. This problem of inaccurate measurement in the lower tail of the scale distribution has consequences for longitudinal assessment, as some increase in raw scores of a developmental or intelligence test may go unnoticed after converting into standard scores. In those patients, a subtle progression of cognitive function can probably not be quantified and expressed in formal MDI, DQ or IQ scores. The frequency of cognitive improvement may therefore be underestimated in these often low-functioning children. Measurement of cognitive age, however, unveils and quantifies change. Even children with a ‘catastrophic’ course of their epilepsy, who had a very poor developmental progress with ongoing seizures, resume cognitive development after hemispherectomy, which was reflected by clear increases in cognitive age (Schooneveld et al. 2011). When pooling results from several studies, in only a minority of children, the standardized MDI, DQ or IQ values decreased significantly after epilepsy surgery: 10% of the hemispherectomy group and 11% of the focal resection group (van Schooneveld and Braun 2013). A decrease in IQ value may indicate a true loss of skills, but in children, it is possible that the conversion from raw to standard scores results in a lower IQ value, even in the presence of some positive development. Catch-up of development is noticed at least 2 times more frequently (20%–30%) after hemispherectomy or focal epilepsy surgery than a decrease of a similar amount of MDI/DQ/IQ points (van Schooneveld and Braun 2013). To truly appreciate the meaning of an unchanged or decreased MDI/DQ/IQ value, individual developmental trajectories should be assessed.
Multiple and often mutually dependent variables influence eventual intelligence or developmental outcome and change after paediatric surgery. Here, we report on the clinically most important presurgical, surgical, and postsurgical variables that may influence general cognitive outcome or change after epilepsy surgery in childhood. Children with developmental pathology who undergo hemispherectomy have relatively lower preoperative IQ/DQ scores but have most to gain from surgery, with higher chances of postsurgical cognitive improvement compared to those with acquired pathology (Pulsifer et al. 2004; Jonas et al. 2004). Actual achievement, however, is limited, possibly related to a more diffuse or bilateral nature of the underlying pathology. This is substantiated by the finding that a disturbed structural integrity of the remaining hemisphere, as defined by the presence of contralateral MRI abnormalities, affects cognitive functioning after hemispherectomy, both in terms of eventual IQ as the chance of significant IQ increases (Boshuisen et al. 2010).
Younger age at onset of epilepsy predicted poor IQ scores after extratemporal resections (D’Argenzio et al. 2011) and, more generally, after epilepsy surgery below the age of 15 years (Matsuzaka et al. 2001), but was not related to postoperative cognitive improvement. Shorter duration of epilepsy has been reported to predict both good (Jonas et al. 2004) and poor (Boshuisen et al. 2010) outcomes after hemispherectomy. It was related to good outcome in a cohort of children who underwent extratemporal surgery (D’Argenzio et al. 2011) and to more gain in IQ scores after surgery in preschool children (Freitag and Tuxhorn 2005). Younger age at surgery predicted poor eventual outcome after hemispherectomy, (Boshuisen et al. 2010) but was also associated with a larger positive change in cognitive scores after epilepsy surgery in infants and children (Bourgeois et al. 2007; Loddenkemper et al. 2007). These findings confirm several hypotheses and assumptions that are part of everyday clinical thinking and decision making. Early-onset epilepsy reflects severe epileptogenic pathology that, by itself, restricts future cognitive abilities. When hemispherectomy is indicated at very young age, the epileptic encephalopathy is severe, development has arrested or regressed, and surgical treatment can induce more pronounced cognitive improvement, although eventual cognitive function is limited. A shorter duration of active epilepsy limits the potentially irreversible effects of ongoing seizures and AEDs on brain function. This problem is independent of the underlying pathology. A higher presurgical level of cognitive functioning is significantly related to higher postsurgical cognition scores (Jonas et al. 2004; Loddenkemper et al. 2007, D’Argenzio et al. 2011), although cognitive improvement occurs less often (Loddenkemper et al. 2007), and IQ changes after temporal resection are only seen as late as 6 or more years after surgery (Skirrow et al. 2011). In general, children at the low end of the IQ spectrum improve more than those with a higher presurgical IQ/DQ level (Bjørnaes et al. 2001; Miranda and Smith 2001; Freitag and Tuxhorn 2005; Spencer and Huh 2008).