Face to face - Epilepsy
A GP quizzes an expert to take a medical issue beyond the textbook
GP Dr Bill Laughey talks to consultant neurologist
Dr Giles Elrington
It is better to wait and keep an open mind, than to wrongly diagnose epilepsy
There is rarely any hurry to start treatment
Don't rely solely on the EEG
Read the driving regulations on the DVLA's website and download them to your PC:
Facts on sudden unexplained death in epilepsy (SUDEP)
How common is epilepsy and who gets it?
Prevalence is one in 200 and incidence is one in 2,000. Lifetime (80 years) risk of seizure is one in 10. Onset is commonest in the youngest and the oldest patients, though no age is exempt. Both sexes are equally affected. Socio-economic deprivation increases the risk of epilepsy and sub-optimal treatment.
Is there a simple way to classify epilepsy?
Primary epilepsy occurs in otherwise normal brains: this usually presents in teens or early adult life, is relatively easy to treat, and has a favourable prognosis.
Focal, or localisation-related, epilepsy is the result of brain disease: the prognosis may be that of the underlying disease (for example brain tumour or learning difficulties), though even when the cause is benign, focal seizures are typically more challenging to abolish than primary seizures.
For practical purposes, neurologists distinguish three groups of patients:
· Attacks which may or may not be epilepsy, who have no diagnosis or treatment.
· Diagnosis (correct or erroneous) of epilepsy in patients who are disabled by attacks or by the effects of treatment.
· Diagnosis (correct or erroneous) of epilepsy who are not disabled.
Groups one and two seek help, not always appropriately: presenting to A&E with chronic epilepsy, rather than coming to the outpatient clinic. Group three may not seek help but care should still be proactive, not just repeat prescriptions.
Epilepsy is rarely a 'snap diagnosis'
How do you make the diagnosis and what are the common diagnostic pitfalls?
One of the keys to diagnosis is a witness's description of the attacks. No investigation is helpful where the nature of the attack remains in doubt. A patient's or witness's conviction that an attack is, or is not, epilepsy is of no diagnostic value. The commonest error in over-diagnosis arises in patients with syncope (fainting) which can cause jerking, twitching, tongue biting and urinary incontinence. Cardiac and psychiatric disease is commonly misdiagnosed as epilepsy.
Under-diagnosis errors arise when people mistakenly believe that epilepsy is manifest only by tonic/clonic convulsions, overlooking simple or complex partial seizures, and myoclonus.
Another key to diagnosis is a history of stereotyped attacks with a postdrome of altered cognition or consciousness. The most reliable diagnoses of epilepsy are made over time: it is rarely a 'snap diagnosis'.
Investigations are usually unhelpful if the clinical diagnosis is insecure. The electroencephalogram is diagnostic only when an attack is recorded on the EEG which is rarely achieved in routine practice. Between attacks, the (interictal) EEG is normal in 50 per cent of people with epilepsy, and is mildly abnormal in 20 per cent of people without epilepsy. 1-2 per cent of people without epilepsy have frank epileptiform discharges on the EEG. A brain scan MRI is almost always better than CT is normal in the large majority of people with epilepsy, but is mandatory in all cases of possible epilepsy, except where there is a clear diagnosis of a primary epilepsy syndrome. Therefore a normal EEG and MRI do not rule out epilepsy; and an abnormal interictal EEG or MRI do not necessarily prove a diagnosis of epilepsy.
How do you differentiate fits from fainting,
and true fits from pseudo fits?
With difficulty! We should all expect to get this wrong sometimes. This is partly a semantic question, properly asked as the differentiation of epileptic fits from non-epileptic attacks including syncope.
Remember that 'pseudo fits' really do occur (nothing pseudo about them) and are a big problem for the patient, their family and for healthcare providers half of 'status epilepticus' in A&E is not epilepsy. The simple answer is that differentiation can be achieved reliably only by the ictal EEG. In practice, it is done by pattern recognition and the continuing awareness that not all that appears to be epilepsy is epilepsy.
Initiating and continuing therapy
When do you decide to start epilepsy treatment? What are the most commonly used drugs and how do you monitor them?
Valproate is most commonly used for primary epilepsy, carbamazepine for focal epilepsy. Some experts prefer lamotrigine for both because it has fewer side-effects and interactions, may possibly be less unsafe in pregnancy and, unlike carbamazepine, does not interact with the contraceptive pill.
The right dose is that which minimises or abolishes seizures with no side-effects. The best monitoring is clinical assessment of the patient; it can be helpful to correlate actual with expected prescriptions. Drug levels are rarely helpful. EEG monitoring is not helpful. No 'routine' blood tests are necessary.
Treatment is conventionally started after the second attack, though if attacks occur less than once in six months delayed treatment may be preferable. Also, do not start treatment unless the clinical diagnosis is secure. Ideally the initiation of anti-epilepsy treatment should be made in secondary care by an appropriate specialist, who will also be responsible for follow-up. In neurological practice it is regrettably common to see patients with insecure labels of epilepsy who have been started on inappropriate treatment (for example, phenytoin) by A&E doctors or GPs who do not want to follow the patient, so do not see the problems. If waiting time for a neurological opinion is so long that treatment must start in primary care, then please at least telephone the neurologist before prescribing.
Many patients, especially those who are well controlled, are lost to specialist review. Quality points for epilepsy in the new contact may well bring some of these patients out of the woodwork.
When can I consider stopping treatment?
After two seizure-free years, treatment cessation may be considered. driving and pregnancy often influence the decision. The DVLA recommends abstaining from driving during the period of weaning-off medication and for six months after the day that medication is stopped.
Some seizure types, such as juvenile myoclonic epilepsy a common disease presenting with myoclonus, and then generalised convulsions in adolescence usually require lifelong treatment.
Patients mistakenly think they are seizure-free if they are experiencing the typical warning symptoms of focal epilepsy without going on to experience the full tonic clonic seizure. These patients still have epilepsy.
There is an issue with primidone because it will be discontinued in the foreseeable future. This is a particularly difficult drug to withdraw and needs special consideration.
Incidentally, the myoclonus in this syndrome is not simply the odd myoclonic jerk that most of us can experience before going to sleep, it typically occurs in the mornings while awake. The dicision is complex and best made in secondary care.
Practical advice in epilepsy
What are the key facts to know about fits and driving?
Drivers must report their epilepsy to the DVLA. No one may drive within a year of an epileptic seizure including partial seizures ('auras') and awake myoclonus unless for at least three years attacks have begun only while asleep. The rules are stricter for vocational drivers, including HGV and PSV drivers.
Remember that once a patient is seizure free, and legally resumes driving, there is no guarantee that seizures will remain in remission 'do not mortgage yourself to a driving licence'. By this I mean that patients with a history of epilepsy are well advised not to plan careers that absolutely depend on them continuing to hold a driving licence.
What about pregnant women, or women trying to conceive?
All potentially pregnant women on anti-epilepsy drugs should take folic acid 5mg daily. All anti-epilepsy drugs can harm the unborn baby, probably doubling the malformation to 1:15. But this risk is greater for valproate, which can cause neural tube defects (spina bifida) in up to 10 per cent and may be associated with educational disadvantage.
The risk increases if more than one drug is used.
Many anti-epilepsy drugs such as carbamazepine, phenytoin, phenobarbitone and topiramate are enzyme inducers, reducing the efficacy of all oral contraceptives (but not of injectable progestogens).
What should GPs know about sudden unexplained death in epilepsy?
People with epilepsy have an increased death risk; this is called 'sudden unexplained death in epilepsy' (SUDEP). This is not always associated with an obvious seizure: the patient is simply found dead with no bitten tongue, no incontinence. SUDEP probably occurs early in a seizure, caused by cardiac arrest as an autonomic seizure discharge. The risk is highest in those with severe intractable seizures; among institutionalised people with epilepsy (who often have substantial co-morbidity) the risk may be as high as 1:80 per year; a more reassuring figure is 1:10,000 major seizures remember the average healthy adult has a death risk from all causes of about 1:1,000 a year. SUDEP is slightly commoner in young men, probably because of their profound cardiovascular reflexes. Polypharmacy, alcohol abuse, and poor adherence to medication are the other risk factors. Partial seizures carry a low, probably zero, risk of SUDEP. SUDEP is not the same as traumatic death during an epileptic seizure, such as drowning in the bath, which is less common than SUDEP.
The prevention of SUDEP is presumably the optimal control of epilepsy, though SUDEP is so rare that this is not evidence based. It is arguable that first aid during a seizure might prevent SUDEP. Appropriate concern about SUDEP might discourage people with epilepsy from being alone; conversely, overprotection, overtreatment, poor compliance, and alcoholism are very much larger problems than SUDEP in clinical practice.
A typical district general hospital has a population of a couple of thousand people with epilepsy among whom about one a year die from SUDEP.