Paediatric epilepsy surgery: making the best of a tough situation

Abstract

This scientific commentary refers to ‘Temporal lobe surgery in childhood and neuroanatomical predictors of long-term declarative memory outcome’ by Skirrow et al. (10.1093/brain/awu313).

This scientific commentary refers to ‘Temporal lobe surgery in childhood and neuroanatomical predictors of long-term declarative memory outcome’ by Skirrow et al. (10.1093/brain/awu313).

Interest in resective surgery as a therapeutic option for uncontrolled paediatric epilepsy is increasing (Cross et al., 2006). Seizures that are left unchecked can have deleterious and largely irreversible effects on cognitive development, particularly if they occur during critical periods of neurodevelopment in the very young. However, surgery itself can lead to cognitive deficits if it entails resection of an area that had been supporting key functions. In adults, there is a large literature on the neuropsychological consequences of surgery and on the careful planning and extensive testing that is performed to guide the resection or sometimes the decision not to perform surgery if removal of eloquent cortex would result in unacceptable deficits (Langfitt and Wiebe, 2008). In this issue of Brain, Skirrow et al. (2014) expand the corresponding literature in paediatric populations by presenting an elegant analysis of the impact of surgery on memory functions in patients who underwent temporal lobe resections for childhood-onset epilepsy at Great Ormond Street Hospital (GOSH).

The GOSH study is remarkable for several reasons: the sophistication of its analysis of neuropsychological outcomes, the duration of time over which the patients were followed (on average, 9 years post-surgery), and the gains in function that were detected years after surgery. In striking contrast to adults in whom memory is often negatively affected by surgery—in one large series, a third of adult patients who had temporal lobectomies experienced significant (≥1 standard deviation) declines in verbal memory (Langfitt et al., 2007)—no postoperative memory deficits were detected in the GOSH series. Instead, and as previously reported, full-scale IQ improved in the paediatric patients by ∼10 points, although this took several years to occur (Skirrow et al., 2011). Memory improved as well, but the patterns of improvement were complex. Most strikingly, improvements were seen in memory functions usually subserved by the non-resected hemisphere: verbal memory gains following right resections and visual memory following left. While at first glance counterintuitive, this is consistent with evidence that there can be substantial reorganization and re-lateralization of function in the developing brain (Tivarus et al., 2012). The results of Skirrow et al. suggest that the healthy temporal lobe had been recruited to help maintain the functions of the seizure-occupied lobe, and that after surgery it was ‘released’ and allowed to return to its preferred tasks as function re-established itself in the operated hemisphere. Furthermore, larger memory gains were seen in association with greater residual hippocampal volumes and smaller resection of the temporal pole, depending on the type of memory tested. In children, recent efforts have focused so intently on stopping seizures before they do irreversible damage and on allowing the time window for developmental plasticity to be as large as possible that rather less attention has been paid to the extent of surgery and how it might interact with developmental plasticity. The findings of Skirrow et al. are novel and have important implications for planning resections so as to spare cognitive function in younger patients.

Perhaps equally important, earlier surgery and shorter duration of epilepsy were also associated with greater post-surgical gains. This is consistent with other studies of early life epilepsies that have focused upon more global measures of development and have demonstrated that the earlier surgery is performed, the better the overall post-surgical cognitive-developmental outcomes (Freitag and Tuxhorn, 2005).

Several further points raised by the Skirrow et al. study need to be considered. The average delay to surgery in the GOSH series was ∼10 years (average age at onset was ∼3 years and at surgery ∼13 years). By the time the children were evaluated for surgery, overall cognition was substantially impaired with presurgical IQ in the low 80s, a good standard deviation below population norms. Almost 20% of the children had abnormal language lateralization indices. Whether these deficits and abnormalities were present for reasons other than seizures and treatment is unknown; however, cognition has been shown to decline over the course of pharmacoresistant epilepsy both in children (Berg et al., 2004) and adults (Hermann et al., 2006). This suggests that damage was occurring in the years before surgery.

This 10-year delay is probably typical of the period in which these surgeries were performed (1992–2002) and typical of what occurs in many places without good access to a surgical centre. Surgery is generally viewed as a last resort therapy when all else has failed; in fact, the delay may be as much, if not more, due to attitudes about surgery than it is about access to surgical therapies. However, delays have consequences. How much does cognition deteriorate during the average 10-year delay? What if surgery were performed early enough for cognitive functions not to have had chance to reorganize/re-lateralize? Would that mean better cognitive outcomes? Better social outcomes? Could it have an impact on seizure outcomes post-surgery? Notably, studies in children and adults who underwent resective frontal lobe procedures suggest that a prolonged delay (>5 years) to surgery is associated with relapse many years later after patients have been seizure-free for extended periods of time (Simasathien et al., 2013).

The complexity of the findings, as well as the implications for planning resection in younger brains, are considerable. While all brains are capable of plasticity to some degree, we generally think of the greatest potential for plasticity as being in the first few years of life, with plasticity dropping off steeply in adolescence and early adulthood. Given that the average age at surgery was 13, this raises questions about whether the findings might have implications for older surgical patients, many of whom also have epilepsy of early childhood onset.

Perhaps most importantly, the findings of Skirrow and colleagues highlight the urgent need for earlier intervention and the importance of considering developmental plasticity and reorganization as part of efforts to spare and maximize post-surgical cognitive function.