A Prospective Study of the Requirement for and the Provision of Epilepsy Surgery in the United Kingdom


Address correspondence and reprint requests to Dr. S.D. Lhatoo at Department of Neurology, Institute of Clinical Neurosciences, Frenchay Hospital, Bristol BS6 7AB, U.K. E-mail slhatoo@aol.com


Summary:  Purpose: Of the 30,000 persons in whom epilepsy develops annually in the United Kingdom, in ∼6000 (20%), intractability develops. Some of these patients will be appropriate for epilepsy surgery. We aimed to estimate the number of patients who should be considered surgical candidates, by extrapolation from a population-based study of prognosis and the number who are receiving epilepsy surgery, by a survey of U.K. neurosurgeons.

Methods: We identified the number of patients who may eventually require surgery from a prospective cohort of patients with newly diagnosed epilepsy. We identified all U.K. neurosurgeons who had performed any epilepsy surgery in the past year. Each identified surgeon prospectively recorded the number and types of operations carried out for 6 months.

Results: Of newly diagnosed patients each year, 450 (1.5%) may eventually require surgery. Thirty-two respondents (22% of all U.K. neurosurgeons) reported that they performed epilepsy surgery. The 211 operations were carried out in the 6 months surveyed (422 operations annually or 13 per surgeon per year). Temporal lobe resection (77%) was the most common procedure.

Conclusions: Based on a prevalence of 5/1,000 persons with epilepsy, ≤4,500 patients in the U.K. require epilepsy surgery. Every year, 450 patients with newly diagnosed epilepsy who may eventually require surgery are added to this “surgical pool.” At the current annual rate of operations, a large number of refractory patients remain untreated. This is probably partly because many patients are not referred for specialist care and therefore remain underinvestigated.

The treatment of epilepsy has seen some remarkable advances in the last century, particularly in the pharmacologic arena. In the 20% of patients with medically refractory epilepsy (1), however, trials with successive antiepileptic drugs (AEDs) in many remain futile (2). For up to half of this number, epilepsy surgery offers a realistic means of achieving otherwise elusive seizure remission. Epilepsy surgery has been performed since the late nineteenth century, and in the last 10 years has assumed an important role in the treatment of chronic epilepsy. Improved neuroimaging techniques have undoubtedly contributed to the renewed interest in surgery (3–5), as have new surgical techniques. It is now well established that with careful preoperative evaluation, 60–70% of patients are rendered seizure free after temporal lobectomy for hippocampal sclerosis (6), and other techniques also offer a substantial chance of improvement (7,8).

Epidemiologic studies in the United Kingdom (U.K.) suggest that of the 30,000 persons in whom epilepsy develops de novo annually [estimated incidence, 50 per 100,000 in the general population (1)], ∼6,000 (20%) may become intractable (9). It has been suggested that ≤1,000 (3.33%) (7) of these patients with intractable epilepsy might benefit from epilepsy surgery, although this estimate was a calculation based on theoretic data, and more precise figures are unknown. A retrospective survey in 1991 estimated that ∼150 procedures were carried out in the U.K. each year, and that the number of patients requiring epilepsy surgery was far in excess of the number actually operated on (10). Since then, the numbers of centres and surgeons specializing in epilepsy care have increased worldwide, as has the frequency of procedures performed specifically to control epilepsy seizures (8). The extent to which this is reflected in epilepsy practice in the U.K. is not known.


The aims of the study were twofold: (a) to estimate the number of patients with newly diagnosed epilepsy who may eventually benefit from epilepsy surgery; and (b) to evaluate prospectively the total number and types of epilepsy procedures currently performed in adults and children in the U.K.


Analysis 1

The number of patients with newly diagnosed epilepsy who may eventually require epilepsy surgery was estimated from the U.K. National General Practice Study of epilepsy (NGPSE) (11). This is a prospective, general population-based cohort of 564 patients with definite epilepsy recruited over a 3-year period (1984–1987) from 275 urban and rural general practices throughout the U.K. These patients have now been followed up annually for ≥14 years, and remission from seizures (12), recurrence (13), prognosis (14), and mortality (15) were previously described in detail.

Analysis 2

All practicing consultant adult and pediatric neurosurgeons in the U.K. identified through the 1998–1999 Medical Register and the British Society of Neurological Surgeons (145 in total) were sent a postal questionnaire asking if they had performed any epilepsy surgery in the last year. Epilepsy surgery was defined as surgery undertaken primarily to control epilepsy rather than for the underlying cerebral lesion. Each identified surgeon was then asked to record the types and numbers of surgical procedures performed by themselves in a short questionnaire mailed to them at the end of every month for 6 months, commencing February 1, 2000. Where necessary, details and numbers of procedures were confirmed by telephone. Data were collected in a tabular form and made anonymous for surgeon, centre, and patient. Details of patients, presurgical assessments and investigations, and surgical outcomes were not sought.


Analysis 1

Median (25th, 75th centiles) follow-up of the NGPSE patients was 11.8 (10.6–12.7) years, equivalent to 11,400 person years. One hundred eleven (20%) of 564 patients with definite epilepsy had not achieved 5-year terminal remission; 31 (5%) of 564 patients with definite epilepsy had frequent (one or more per week) seizures; 22 (4%) of these patients had partial seizures with or without secondary generalization. Five [one, after subarachnoid hemorrhage seizures; one, alcohol-related seizures; one, Asperger syndrome; one, after intrauterine cytomegalovirus (CMV) infection–related brain damage; one, severe learning and physical disabilities of undetermined cause] were considered inappropriate for surgery. None of these patients was older than 60 years. Thus 17 of 564 patients could be considered appropriate for presurgical investigations [900 patients in the U.K. or 3% of an incident cohort (95% confidence intervals (CI), 1.6–4.4]. Assuming up to half of patients who undergo presurgical assessment (a figure based on practice at the National Hospital for Neurology and Neurosurgery, the hospital with the largest epilepsy surgical practice in the U.K.), the annual recurring need for surgery, based on a figure of 30,000 incident cases in the U.K., would be ∼1.5% (95% CI, 0.5–2.5) or 450 patients per year.

Analysis 2

Of the 145 neurosurgeons, 110 (76%) responded. We were confident from subsequent local enquiry that none of the nonresponders had an epilepsy surgery practice. Thirty-two consultant neurosurgeons in current practice who performed epilepsy surgery were identified (22% of 145 consultant neurosurgeons). Of a total of 192 monthly questionnaires sent out, 183 completed forms were returned, a response rate of 95%. A total of 211 surgical procedures for epilepsy was performed, suggesting an annual figure of 422 operations and a mean of 13 per surgeon per year (Table 1). In addition, 78 vagal nerve stimulator implants were carried out, suggesting an annual figure of 156 procedures.

Table 1.  Classification and numbers of epilepsy surgical procedures performed in the U.K. over a 6-month period
Type of surgical procedureNo.Annual no./%
Temporal lobe resection   
 Mesial temporal sclerosis10420850%
 Other lesions5511026%
Extratemporal resection   
Multiple subpial transection00 
Corpus callosotomy7143%
Other procedures   
 Vagal nerve stimulation78156100%

The commonest surgical procedures (Table 1) performed were temporal lobe resection for hippocampal sclerosis (50%) and temporal lobe resection for lesions other than hippocampal sclerosis (26%). Nonlesional resections were uncommonly performed, although nonlesional extratemporal lobe (invariably frontal lobe) resections (3%) accounted for twice as many nonlesional temporal lobe resections (1%). When all surgeries were considered together, vagal nerve stimulator implantation was the second most common procedure performed, accounting for 156 (27%) of all procedures. Both hemispherectomies (3%) and corpus callosotomies (3%) were infrequently performed. No radiosurgery or multiple subpial transections were performed during the study period.


Broad international consensus is now found that surgical procedures for chronic epilepsy are effective (6–8) in seizure control, and procedures such as anterior temporal lobectomy can lead to complete cessation of seizures in ≤80% of patients with temporal lobe epilepsy (6,7,16). A recent randomized, controlled trial of surgery for refractory temporal lobe epilepsy showed that at 1 year, >12 times as many patients were completely seizure free in the surgical group as compared with the medically treated group (17). Epilepsy is said to be the commonest serious neurologic disorder, and its lifetime prevalence in the general population is as high as 2–5%(18). The remaining 20% of patients will have chronic epilepsy refractory to medical treatment (1). The latter constitute a group that is difficult to manage, even with the plethora of newly available AEDs.

A total of 8,324 patients had undergone epilepsy surgery in 103 centers, compared with 3,446 patients in 44 centers worldwide between 1986 and 1990 alone at the second Palm Desert Conference Survey, a large increase (8). The numbers that require surgical procedures, however, are not precisely known, and estimates can be made from only epidemiologic parameters such as incidence, prevalence, mortality and prognosis, and remission from epilepsy. Before this study, no precise figures existed from the U.K.

It has been estimated that the U.K. has an epilepsy prevalence of 5/1,000 in the general population (1), of which 20% consist of patients who have at least one seizure a month (9). Of these patients with frequent seizures, ∼60% may have partial seizures (19,20) potentially amenable to “curative” surgery. In the U.K. (Fig. 1), this means that ≤4500 patients with epilepsy in the general population constitute a “surgical pool” who potentially require surgery. It is likely that in the U.K., only a minority of these patients will be those for whom previous epilepsy surgery failed (20–30% of those operated on). The annual incident population (50/100,000 or 30,000 in the U.K.) will contribute ∼450 patients with surgically remediable refractory partial epilepsy annually (1.5% of 30,000) to this “surgical pool.” The NGPSE began recruiting patients in 1984, before new AEDs were commonly used, and therefore the number of refractory patients may be lower, although few patients with refractory epilepsy achieve lasting remission with new AEDs. It must be emphasized that these figures are estimates based on epidemiologic studies.

Figure 1.

Epilepsy surgery requirements in the United Kingdom calculated from the incident and prevalent populations with epilepsy.

Figures from this study suggest the number of “curative” operations for epilepsy at 422 per year. Although this keeps up with the incident cases being added to the surgical pool, it does not address the backlog of patients in the prevalent surgical pool. Although thrice the number of procedures estimated in 1991 currently take place (10), it is still below that required to treat all patients in this surgical pool. The period of study represents the second half of the financial year, and it is possible that this may have resulted in an underestimation of numbers, although the effect of this is likely to have been slight.

Only nine more surgeons are currently performing epilepsy surgery than in 1991 (10). However, many surgical units carry out relatively small numbers of operations. Recent guidelines suggest that a minimum number of 25 operated-on cases per year is required to ensure sufficient experience and expertise (21). Few of the U.K. centers carry out this number, and the extent to which epilepsy surgery should be focused on fewer centers with higher throughputs is a question that the neurosurgical community should address.

The commonest procedure performed was temporal lobe resection for hippocampal sclerosis. This particular pathology accounts for a third of patients with refractory partial epilepsy (22) and up to half of all patients with temporal lobe epilepsy (23), although its true incidence is unknown. Nonlesional temporal lobe surgery is infrequently performed, accounting for only 1% of all procedures. This may be reflective both of the relative infrequency of such patients (23,24) and relatively poor outcome after surgery (6).

Palliative surgery with vagal nerve stimulator implants appears to be a frequently performed procedure in the U.K., accounting for a fourth of all procedures. Seizure freedom is rarely reported in patients with refractory seizures with this procedure, although seizure frequency may be significantly reduced by up to a third in some patients (25). In contrast, hemispherectomies are infrequently performed, presumably because of mortality rates that approach 30%, although in well-selected patients, this can be an effective procedure (26). Similarly corpus callosotomies account for only 3% of all procedures. This procedure is appropriate only in selected patients with multifocal tonic and atonic seizures, and although it can result in reduced seizure frequency in a significant number of patients, morbidity is high.

Thus although there has been an almost threefold increase in “curative” epilepsy surgery in the last decade, epidemiologic estimates suggest a significant treatment gap in the U.K. The cause of this gap remains unstudied, although it probably largely reflects lack of awareness of the effectiveness of “curative” epilepsy surgery among nonspecialist physicians, general practitioners, as well as patients. The large numbers of patients undergoing vagal nerve stimulator implants suggests that although many patients undergoing this procedure have been deemed unsuitable for resective surgery, an underemphasis may be placed on “curative” procedures and an overemphasis on palliative procedures. Clinical experience suggests that many patients remain underinvestigated, and community-based studies to address this issue would be welcome. Although the increase in surgical numbers is a welcome trend, we believe there should be an increased focus on patients with remediable lesional epilepsy.

Acknowledgment: We thank Ms. Jane Sweetland for her help with the preparation of this manuscript.