Surgical Treatment for Epilepsy: Title and subTitle BreakToo Little, Too Late?

Epilepsy is one of the most common serious neurological conditions. According to a study by the World Health Organization, epilepsy accounts for 1% of the global burden of disease based on disability-adjusted life-years (DALYs), productive years lost due to disability or premature death.¹ This is equivalent to lung cancer in men and breast cancer in women. Among primary disorders of the brain, epilepsy ranks with depression and other affective disorders, Alzheimer disease and other dementias, and substance abuse. Of the world's population, 0.5% to 1.0% has active epilepsy, and pharmacotherapy is unsuccessful in controlling seizures in 20% to 40% of patients.² In the United States, 80% of the cost of epilepsy is attributable to patients with medically intractable seizures.³ Temporal lobe epilepsy is the most common cause of pharmacoresistant seizures⁴ and may constitute half or more of the patients in the United States with medically intractable epilepsy.⁵ On the other hand, temporal lobe epilepsy is the form of epilepsy most easily and effectively treated with surgery; 60% to 90% of patients can expect to become free of disabling seizures postoperatively.^{6 - 8} Appropriately applied surgical treatment, therefore, is crucial for reducing the health burden represented by epilepsy.

Surgical treatment for epilepsy, however, is arguably the most underused of all effective therapeutic interventions. It is estimated that more than 100 000 patients in the United States with temporal lobe epilepsy may be candidates for surgical treatment, although fewer than 2000 received therapeutic surgical resections in 1990.⁹ A 2001 Western Ontario randomized controlled trial involving patients with temporal lobe epilepsy confirmed many published surgical series findings that surgical treatment is more likely to eliminate disabling seizures and improve quality of life than pharmacotherapy.⁸ As a result of this finding and a literature review, the American Academy of Neurology issued a practice parameter in 2003 recommending surgery as the treatment of choice for medically intractable temporal lobe epilepsy.⁷

Although no evidence exists to substantiate whether the Western Ontario trial and the American Academy of Neurology practice parameter has increased surgical treatment for temporal lobe epilepsy in the United States, unpublished data from the National Association of Epilepsy Centers indicate that the number of anterior temporal resection for epilepsy performed in the United States is not appreciably greater in 2008 than it was in 1990 (Robert J. Gumnit, MD, president, National Association of Epilepsy Centers, written communication, November 2, 2008). Furthermore, those who eventually undergo surgery are not referred until an average of 22 years after onset.¹⁰ Not only are most surgical candidates never referred for surgery but those who are often undergo surgery too late to prevent irreversible disabling psychological and social consequences of epilepsy.¹¹ In theory, for many of these patients early surgical intervention could prevent a lifetime of disability, as well as premature death.¹²

A clinical trial to establish the putative positive effect of surgical intervention on the mortality rate associated with medically intractable temporal lobe epilepsy would require following up a large population of patients for many years, making such a trial prohibitively expensive. Consequently, Choi and colleagues¹³ used a decision analytic model to quantitatively predict survival following surgical therapy compared with continued medical management of patients with temporal lobe epilepsy and report their results in this issue of JAMA. Their data indicate that for a 35-year-old patient, surgical therapy would be associated with an increase in life expectancy of 5.0 years and 7.5 in quality-adjusted life-years. These results should provide a further incentive for physicians to refer potential surgical candidates to epilepsy surgery centers but most likely underestimate the benefit, given that the authors chose age 35 years for their analysis based on the current practice of performing surgery more than 20 years after the onset of epilepsy. The authors also acknowledge that a major limitation in their model is the lack of data on the natural history of pharmacoresistant temporal lobe epilepsy. This knowledge deficit could also account for the fact that surgical treatment is still practiced too little and too late.

The inability to reliably identify medical intractability early in the course of temporal lobe epilepsy is perhaps the major reason for the long delay in patient referral to epilepsy surgery centers. With the introduction of many new antiepileptic drugs over the past 2 decades, it appears that the duration between epilepsy onset and referral for surgical treatment may have actually increased; in 1986, only 12.5% of patients who had undergone temporal lobe surgery at the University of California-Los Angeles Center for Health Sciences had epilepsy for more than 20 years.¹⁴ It would now be impractical to prove that any given patient with epilepsy has seizures that are refractory to all available medications in every possible combination. The concept of medically intractable epilepsy, therefore, no longer has practical meaning for the clinician and has been replaced by the concept of surgically remediable epilepsy syndromes.¹⁵ Temporal lobe epilepsy is the prototype of a surgically remediable epilepsy syndrome, a condition with a well-defined pathophysiology and anatomical substrate, a known poor response to medical therapy, and an excellent response to a standardized surgical intervention based on noninvasive evaluation. Early surgical intervention for these syndromes offers the best opportunity to prevent the development of irreversible disabling psychological and social consequences of recurrent seizures and premature death.

The remaining obstacle to effective early application of surgical therapy for temporal lobe epilepsy, therefore, is insufficient information regarding the natural history of this disorder to permit reliable early prediction of continued pharmacoresistance in individual patients. Although one study has suggested that less than 3% of patients with epilepsy will ultimately become seizure free after 2 trials of antiepileptic drugs have failed due to inefficacy and not intolerance,¹⁶ others have since documented up to a 15% to 20% response to continued pharmacotherapy in this population.^{17 - 20} A large multicenter observational study would be needed to determine the actual percentage of patients with temporal lobe epilepsy who will remain pharmacoresistant after failure of 2 antiepileptic drugs and to identify factors that could reliably predict persistent pharmacoresistance.

In the meantime, given the complexity of issues involved in the consideration of referral for epilepsy surgery, the uncertainty of the persistence of pharmacoresistance and the critical importance of early intervention, what are primary care physicians and general neurologists to do? It is perhaps simpler to put thoughts of surgical therapy aside initially and begin by distinguishing relatively benign epilepsies from serious epilepsies. The former are easily treated with the first 1 or 2 antiepileptic drugs tried, resulting in no disabling seizures and no adverse effects. The more serious epilepsies are those in which epileptic seizures or adverse effects continue and compromise work, school, interpersonal relationships, or quality of life, despite adequate trials of a few appropriate antiepileptic drugs at maximum tolerable doses. Ideally, such patients should be referred for second opinions to epilepsy centers, where epilepsy specialists would then evaluate the surgical therapy option.

Referral to an epilepsy center, particularly if early surgical therapy might be appropriate, is most important for patients with surgically remediable syndromes such as temporal lobe epilepsy, as well as those with focal seizures due to localized resectable neocortical lesions, and infants and young children with severe epilepsies caused by more diffuse lesions that are limited to 1 hemisphere. In most cases, these diagnoses can be suspected based on magnetic resonance imaging scans and seizure semiology, even when interictal electroencephalographic abnormalities are inconsistent. Early in the course of temporal lobe epilepsy, however, the magnetic resonance imaging scan results can be normal and more detailed testing at an epilepsy center—including long-term video-electroencephalographic monitoring, positron emission tomography, ictal single-photon emission computed tomography, magnetoencephalography, magnetic resonance spectroscopy, and occasionally intracranial recording—may be necessary to determine whether an anteromesial temporal resection would most likely render a patient free of disabling seizures.

Patients with medically refractory epilepsy who do not meet criteria for diagnosis of a surgically remediable syndrome, as defined herein, might still benefit from surgical therapy. For these patients, the presurgical evaluation almost always requires intracranial monitoring, which increases the risk and cost, while the likelihood of a seizure-free result can be only 50% or less.²¹ The decision to proceed with surgery in this situation can be difficult, as illustrated by the Clinical Crossroads discussion by Schomer and Black,²² also published in this issue of JAMA. One of the authors' recommendations for this patient is vagus nerve stimulation, given the patient's lack of a focal lesion and the potential risks of surgery. Although vagus nerve stimulation can reduce seizure frequency and severity, it rarely results in seizure freedom. Further testing would be indicated to better determine the patient's actual risk from a resective procedure, which could offer a much greater chance of a seizure-free outcome. Ultimately, the patient will need to decide whether the risks are worth the possible benefits, a significant challenge in the face of limited data.

Surgical treatment for epilepsy remains underused, and when surgery is performed, it is often delayed until it is too late to reverse associated severe disabilities. The study by Choi et al¹³ underscores, but most likely underestimates, the benefits of surgical therapy for temporal lobe epilepsy. More information is needed to better understand the natural history of intractable temporal lobe epilepsy and to identify factors that reliably predict which patients will continue to have persistent seizures and thereby enable an accurate prediction of surgical benefit early in the course of antiepileptic drug treatment, before the disabling psychological and social consequences of recurrent seizures become irreversible. In the meantime, however, the best approach to facilitate early surgical treatment, when appropriate, would be for primary care physicians and general neurologists to refer patients who continue to have disabling seizures after a few antiepileptic drugs have failed to an epilepsy center for additional evaluation. Epileptologists at these centers are specifically trained to carry out the testing necessary to determine whether early surgical therapy is a viable option.