The Epidemiology of Epilepsy and Learning Disability

Abstract

Epilepsy, the most prevalent serious neurological disorder, afflicts 0.5% of the general population (1). The prevalence of learning disabilities (LDs) in the general population is approximately similar (2). Both groups of conditions share a common heritage of heterogeneity, high prevalence figures, a dearth of care expertise outside major centres or institutions, and an unfortunate degree of stigmatization. Up to a fourth of patients with epilepsy are said to have LDs, and conversely, up to half of all patients with LDs are said to have seizure disorders, thus rendering this particular combination an especially common one. The paucity of authoritative studies on prognosis, mortality, and treatment of this special group belies this, however, and is all the more surprising. This encourages a substantial lacune in current knowledge of the epilepsies, and these deficiencies must be addressed appropriately. Epidemiology is the study of the dynamics of a condition in a population, and the “patient” therefore represents the community rather than an individual. A sine qua non is that data must be derived or collected from an unselected population. In this regard, there is little comprehensive information on the epidemiology of epilepsy in LDs and LDs in epilepsy. As with many epidemiologic studies in epilepsy, lack of standardization in methods used and inherent population biases have plagued estimates of prevalence of epilepsy in patients with LDs and vice versa. There are no studies of incidence, and most large-scale prevalence studies of epilepsy have seldom addressed or defined subpopulations with additional LDs. Both LDs and epilepsy have wide heterogeneity, and this should be addressed in any comprehensive study. In epidemiologic studies of LDs, true prevalence rates of epilepsy are difficult to ascertain because of the variability in screening instruments as well as the lack of standardized definitions and classifications of epilepsy. Adults are underrepresented in most such research, and the vast majority of published data pertain to children. Studies of mortality in patients with epilepsy and LDs are more readily undertaken in institutionalised patients and probably overestimate such deaths in the community. The prevalence of epilepsy varies in direct proportion to the severity of the LD. Community studies have indicated prevalence rates ranging from of 6% amongst children with mild LDs (IQ, 50–70) (3) to 24% in severe LD (IQ, <50) (4) and 50% in those with profound LDs (IQ, <20) (5). A reasonable estimate of the prevalence of epilepsy in patients with moderate to severe LDs probably ranges between 1 in 4 and 1 in 5 such patients (4–6). Some studies have addressed epidemiologic issues in individual syndromes and etiologies. In patients with cerebral palsy and LDs, the cumulative risk for developing epilepsy has been estimated at 28, 31, and 38% at ages 5, 10, and 22 years (7). Of children with Down's syndrome, 1.4% are said to have epilepsy (8), and the prevalence increases substantially in those older than 35 years (12.2%) (9). A much larger proportion of patients with fragile X syndrome have epilepsy, with prevalence rates approaching 25%(10). Both tuberous sclerosis (>60%) (11) and the Sturge–Weber syndrome (>70%) (12) have very high prevalence rates of epilepsy, although not all have LDs. Up to 90% of patients with Lennox–Gastaut syndrome or West's syndrome may have LDs (13). In several conditions, epilepsy and LDs are common associations. These range from cerebral palsy to rare conditions such as cortical dysplasias and inherited metabolic defects. Some of these associations are listed in Table 1. The prognosis for seizure control in patients with epilepsy who have LDs is probably poorer than that in those without LDs (14). No recent studies, however, specifically address the problem of seizures in populations with LDs, although observations have been made in the context of individual syndromes and etiologies. In cerebral palsy, only 13% of 531 patients with epilepsy in a large cohort achieved 2-year remission from seizures. Encouragingly, the same study found that drug discontinuation in these patients was associated with lasting remission in 60%. LDs however, did not specifically appear to influence remission (15). In Down's syndrome, prognosis for seizure control may depend on the etiology of seizures; patients with cardiovascular etiologies and idiopathic seizures generally did well, those with central nervous system (CNS) infective seizures did less well, and those with neonatal hypoxic–ischemic brain damage did least well (16). Patients with seizures in fragile X tend to have infrequent attacks and are mostly well controlled with standard therapy (10). In the neurocutaneous syndromes, seizure control can be particularly difficult. Infantile spasms occur in 40–60% of patients with tuberous sclerosis (17,18). In patients with the Sturge–Weber syndrome also, the prognosis for satisfactory seizure control is poor. In the Lennox–Gastaut syndrome, complete seizure freedom is rare. Not all patients maintain the full profile of the syndrome, but frequent disabling seizures persist in the majority. The main predictors of poor prognosis appear to be symptomatic seizures, infantile spasms, early age at onset of seizures, frequent seizures, repeated status epilepticus, long periods of worsening, and EEG abnormalities such as constant slow background activity and localized changes. Mortality in patients with epilepsy is higher than that in the general population, and subnormal mentation has traditionally been identified as a significant adverse prognostic factor (19). Again, however, few recent studies have examined this in patients with LDs. In a Swedish study, 1,478 patients with LDs, 296 of whom had epilepsy, were followed up for 7 years. Standardized mortality ratios (SMRs) in patients with LDs and epilepsy were substantially higher [SMR, 5.0 (95% CI, 3.3–7.5)] than those in patients with LDs alone [SMR, 1.6 (95% CI, 1.3–2.0)] and highest in patients who had LDs, epilepsy, and cerebral palsy [SMR, 5.8 (95% CI, 3.4–9.7)](20). Mortality was especially significant with generalised seizures and was related to seizure frequency in the year preceding death. Only 7% of deaths in patients with epilepsy and LDs appeared to be directly seizure related, and the commonest cause of death otherwise was pneumonia. In patients with LDs and very severe epilepsy, direct seizure-related deaths are much more frequent, and sudden unexpected death in epilepsy (SUDEP) is probably the commonest category. In a study of 310 patients with epilepsy and severe LDs followed up for a total of 4,135 person years, the overall SMR was 15.9 (95% CI, 10.6–23.0), and 70% of the 28 recorded deaths were seizure related (21). The incidence of SUDEP was estimated at 1:295 per year compared with 1:1,000 per year in epilepsy patients in the general population. This suggests that patients with LDs and severe epilepsy are particularly at risk of sudden death. In specific syndromes such as in the progressive myoclonic epilepsies, mortality rates can be very high, and prognosis, generally dismal. Few properly designed trials have assessed individual drugs in the treatment of epilepsy in LDs. Current treatment practice is based largely on retrospective studies, open trials, and center-based treatment policies. This approach is easily open to criticism and does not address the specific difficulties and characteristics of patients with LDs. There is some evidence, however, that, as in the general population, there is a trend toward modernizing and rationalizing therapy. The use of phenobarbitone (PB) in developed countries, for example, appears to have declined, and the majority of patients with LDs are receiving monotherapy. Lack of standardization in many studies limits the usefulness of their conclusions. No specific drugs are used in the treatment of seizure disorders in patients with LDs, and, as in the current treatment of epilepsy, choice of drug is dictated largely by seizure types. There are no randomised, controlled trials of drug treatment in large populations with LDs, although there are small open studies that have looked at newer anticonvulsant use in LD. An uncontrolled study of add-on vigabatrin (VGB) use in LD showed ≥50% seizure reduction in 42, 25, and 22% of 36 patients studied at 3 months, 2 years, and 5 years, respectively (22). In a few specific syndromes, some specific drugs are said to have moderate to high efficacy. Studies have shown that VGB, for example, significantly reduces seizures in patients with West's syndrome, both as monotherapy (23,24) and as add-on. In one study of 40 newly diagnosed children, patients were randomised to VGB monotherapy or placebo for 5 days, and then all were treated openly for 24 weeks. At the end of the study, 38% were seizure free (23). Both lamotrigine (LTG) and topiramate (TPM) are thought to be effective in Lennox–Gastaut syndrome. In a double-blind, placebo-controlled study of 169 patients with Lennox–Gastaut syndrome, 33% of patients treated with LTG had ≥50% seizure reduction compared with 16% in the placebo group (25). A trial of TPM in 98 patients showed similar results (26). Discontinuation of medication is an issue that is similarly understudied, and little information is available on the matter. In one study of 50 patients with LDs followed up for 8 years, patients in 2-year terminal seizure remission were discontinued from medication (27). In half, seizures recurred, most during the first 3 years after discontinuation. A small lifetime number of seizures, absence of gross neurologic abnormalities, low drug levels at termination of therapy, and normal EEGs before and after discontinuation of medication appeared to predict successful outcome. In summary, the association between epilepsy and LDs, both very prevalent conditions, is very common. LDs, by definition, have a childhood or adolescent onset, and a variety of seizure disorders commonly arise in these age groups. Despite high prevalence and the significant morbidity and mortality associated with these conditions, a surprising dearth of literature is focussed on prognosis, treatment, and outcomes in populations with LD. Most of the existing drug treatments for seizures in LDs, for example, are derived from studies on populations without LDs. A few studies that provide specific information on individual syndromes such as Lennox–Gastaut and infantile spasms exist, although these patients constitute but a small proportion of patients with LDs. Given the substantial implications to health professionals and health care planning, these deficiencies should be addressed. There is still a requirement for properly designed prospective incidence studies of patients with LDs developing epilepsy. In this context, it is important to emphasise that cross-sectional studies, for a range of reasons, are unsuitable and unlikely to add significantly to current knowledge. There also is a need for well-designed case–control studies analyzing etiologies and risk factors for seizures in LDs. With the advent of sophisticated neuroimaging, genetic tests, and other investigational techniques, these have become more feasible than ever. Long-term studies of prognosis using standardized epidemiologic methods are required, given the current lacunae in our understanding of seizure remission, recurrence, morbidity, and mortality in patients with epilepsy and LD. In particular, the issues surrounding SUDEP should be clarified, as patients with refractory epilepsy and LD appear to be particularly at risk of this phenomenon.