OUP user menu

The misdiagnosis of epilepsy and the management of refractory epilepsy in a specialist clinic

D. Smith, B.A. Defalla, D.W. Chadwick
DOI: http://dx.doi.org/10.1093/qjmed/92.1.15 15-23 First published online: 1 January 1999


We assessed the frequency, causes and consequences of erroneous diagnosis of epilepsy, and the outcome of patients referred with `refractory epilepsy', by retrospective analysis of the case records of 324 patients. The sample was divided into those exposed to anti-epileptic drugs (n=184), of whom 92 were said to have refractory seizures, and those who had not received treatment (n=140). The latter group is reported elsewhere. The overall misdiagnosis rate was 26.1% (46/184), with incomplete history-taking and misinterpretation of the EEG equally responsible. Side-effects were reported by 19/40, while unnecessary driving restrictions and employment difficulties were encountered by 12/33 and 5/33, respectively. Of those labelled `refractory epilepsy', 12 did not have epilepsy. Sixteen were rendered seizure-free and 25 significantly improved by the optimal use of anti-epileptic drugs or surgery. Diagnostic and management services for patients with suspected and established epilepsy are suboptimal, with psychological and socio-economic consequences for individual patients. The resulting economic burden on the health and welfare services is probably substantial.


While the true epidemiology of the epilepsies is not known, the available data indicates an annual incidence of 40–200/105 and a prevalence of 0.5–1%.1 A community-based study in the UK estimates that, at any one time, there are 350 000 people with active epilepsy and 30 000 new cases per year.2 While the overall prognosis is favourable, 20–30% of people suffer a chronic disabling condition.3 It has been estimated that the annual cost of epilepsy is £1.8 billion with indirect costs, mainly loss of productivity, far outweighing direct costs and with the bulk of these resources consumed by those with refractory epilepsy.4

Epilepsy is not a diagnosis per se but a symptom of a disorder of the cerebral hemispheres with a diverse aetiology which is principally determined by the age at which habitual seizures start. A complete diagnosis of epilepsy requires differentiation of seizures from other causes of altered consciousness/behaviour, differentiation of spontaneous unprovoked seizures (epilepsy) from provoked seizures (e.g. reflex anoxia), classification of seizures5 and epilepsy6 and determination of the underlying cause (Box 1). The classification of epilepsy6 is determined by seizure type(s), the age of onset of seizures, EEG characteristics and associated motor or mental handicaps and is useful in determining the need for investigations, predicting prognosis and choosing treatment. However outside of specialist centres, this syndromic approach is rarely used, and, therefore incomplete diagnosis leading to inappropriate treatment is well-documented within the specialist literature7 and, more recently, within the general medical literature.8

Much more common than epilepsy is the presentation with non-specific `funny turns' to general practitioners, general physicians, paediatricians and specialists. These patients are often referred for exclusion of a possible diagnosis of epilepsy. Whilst a delay in the diagnosis of genuine epilepsy rarely causes problems, a false-positive diagnosis may have severe psychological and socio-economic consequences for the patient and economic implications for the NHS and welfare services. Despite the widespread recognition of this problem within neurological and paediatric neurological circles thorough Medline Searches identified the only two articles9,10 published within specialist literature, which address the issue of misdiagnosis of epilepsy.

Improvement of the diagnosis of `funny turns' and services for patients with epilepsy are the principal aims of the first Epilepsy Task Force document11 which highlights the benefits of multidisciplinary epilepsy clinics. There is evidence that these issues have been recognized by the `new NHS'. On 17 January 1995, John Bowis, then parliamentary under-secretary of state for health, announced in parliament a programme of coordinated initiatives on epilepsy. Subsequently EL(95)12012 recommended that Health Authorities, GP Fund Holders, the Primary Care Teams and NHS Trusts should `work towards best practice in commissioning, contracting, and service delivery for epilepsy', and these issues were discussed at a NAHAT-sponsored conference entitled `Contracting for Quality Epilepsy Services' in May, 1996.

The last few years have witnessed a long overdue expansion of the neurology consultant grade. However this expansion has not been uniform and, therefore, accessibility to neurologists varies widely throughout the UK. Furthermore, <5% of neurologists have a particular expertise in epilepsy which would permit their taking a `lead role' in a comprehensive epilepsy service. Thus most people with established or suspected epilepsy will continue to be managed in general medical or paediatric clinics. With these issues in mind, this study was designed to assess the frequency of, factors contributing to, and consequences of an erroneous diagnosis of epilepsy, and to evaluate the effectiveness of a specialist clinic in terms of the outcome of patients referred with `refractory epilepsy'.


We conducted a retrospective analysis of the case-records of 324 patients referred to a single consultant (DWC) during a 12-month period. Although patients were seen in three different clinic settings (regional epilepsy clinic, centre-based general neurology clinic, DGH-based general neurology clinic) a policy of open access to these clinics was adopted. The source and reason for referral was established in all cases. The whole group was sub-divided into patients with `epilepsy for management' (n=184) and `funny turns for diagnosis' (n=140)13 based on whether or not they had been exposed to anti-epileptic drugs. Within the former group reasons for referral included: seizure-free for possible withdrawal of drugs, `refractory epilepsy', rationalization of polytherapy, suspected drug toxicity, diagnostic uncertainty and `general advice'.

Epilepsy is a clinical diagnosis based on an individuals account of their symptoms and, crucially, an eye-witness account of events. It is accepted practice that treatment should not be commenced until the diagnosis is made with certainty. Accordingly, DWC considered that a misdiagnosis had occurred when the history indicated an alternative diagnosis, or there was insufficient clinical evidence to support the diagnosis of epilepsy. In eight cases the alternative diagnosis was confirmed by observing attacks (pseuodoseizures), without electrical correlate, during ambulatory EEG monitoring.

Patients were subsequently placed, according to specialist opinion, into the following categories; epilepsy, syncope, psychogenic attacks (pseudoseizures, panic attacks), other specific diagnosis (e.g. migraine), mixed diagnoses (e.g. epilepsy plus non-epileptic attacks) and no firm conclusion. Data is presented regarding investigations used in establishing diagnosis, the reasons, sources and consequences to the patient and the health service of a misdiagnosis of epilepsy, and the outcome of patients referred with a diagnosis of refractory epilepsy.


Misdiagnosis of epilepsy

The overall misdiagnosis rate was 46/184 (26.1%). In some patients who were referred with `refractory epilepsy', credit should be given to referring doctors, often general practitioners, for questioning a previously made diagnosis (Table 1). No one is immune to this problem (Table 2) but, in this study, where a neurologist made the erroneous diagnosis, a second opinion was arranged because, after review, the initial diagnosis was brought into question. The referring doctor commenced drug therapy in 26/46 cases, but in 50% of these patients the reason for referral was `diagnostic uncertainty' suggesting that therapeutic trials of anti-epileptic drugs remain commonplace (Table 3).

View this table:
Table 1

Diagnosis according to reason for referral

Reason for referralDiagnosis
EpilepsyNot epilepsyTotal
AED, anti-epileptic drugs.
AED withdrawal606
`Refractory epilepsy'821294
Rationalize polytherapy404
AED toxicity617
Diagnostic uncertainty32427
View this table:
Table 2

Sources of erroneous diagnosis

View this table:
Table 3

Erroneous diagnosis according to reason for referral (where diagnosis was made and treatment commenced by doctor making referral)

Referring doctorReason for referral
`Refractory epilepsy'Diagnostic uncertaintyAdviceTotal

Syncope and non-epileptic attacks of psychogenic origin were the conditions most commonly mistaken for epilepsy. Others included migraine, benign paroxysmal positional vertigo and hypnic jerks. Within the mixed diagnoses group, there were two patients with epilepsy in remission who had developed `a new form of attack' which had not responded to increasing doses of anti-epileptic drugs (Table 4).

View this table:
Table 4

Distribution of conditions misdiagnosed as epilepsy

Actual diagnosisn
NEA, non-epileptic attacks.
Mixed diagnoses6
No firm conclusion1

Incomplete history-taking appeared to be the principal cause for misdiagnosis. The commonest problems were the absence of an eye-witness description of attacks, misinterpretation of motor phenomena associated with syncope, and lack of recognition of the significance of extensive past medical and psychiatric histories in patients with pseuodoseizures. However misinterpretation of normal phenomena,14 e.g. 14 and 6 cps positive spikes, frontal delta on over-breathing or over-interpretation or non-specific features e.g. focal temporal slow waves, with or without sharper components, was a contributory factor in 19/46 cases.

Consequences of a misdiagnosis of epilepsy

Most patients received only one drug and 20 patients, many of whom experienced side-effects, had stopped treatment by the time of neurological consultation. However patients with psychogenic attacks are often exposed to dual or polytherapy for prolonged periods (Table 5). Symptoms of dose-related neurotoxicity were common and in addition, two patients had hypersensitivity reactions to carbamazepine, one had gynaecomastia, and two developed unsightly gingival hypertrophy on phenytoin (Table 6).

View this table:
Table 5

Consequences of misdiagnosis: exposure to anti-epileptic drugs

DiagnosisNumber of AEDsTotal
NEA, non-epileptic attacks.
Mixed diagnoses3126
No firm conclusion1001
View this table:
Table 6

Consequences of misdiagnosis: anti-epileptic drug toxicity

Type of toxicityn
No side-effects21
Not yet contacted6

Whilst it is usually easy to stop the anti-epileptic drugs (Table 7), it is much more difficult to reverse the negative social implications of a misdiagnosis of epilepsy. Thirty-three patients were contactable by telephone; of the 14 who possessed a driving licence at the time of diagnosis, 12 had driving interrupted unnecessarily, with five losing their licence temporarily (Table 8a). Furthermore, three patients lost full-time jobs, one was demoted and one had to refuse a job which involved driving (Table 8b).

View this table:
Table 7

Stopping AED after misdiagnosis of epilepsy

DiagnosisYesNoNot applicableTotal
AED, anti-epileptic drugs; NEA, non-epileptic attacks.
No firm conclusion1001
View this table:
Table 8a

Consequences of misdiagnosis: driving

No licence19
Loss of licence (appropriate)1
Loss of licence (unnecessary)5
Stop driving temporarily5
Stop driving permanently2
Not told to stop driving1
Not yet contacted13
View this table:
Table 8b

Consequences of misdiagnosis: Gainful employment

No job18
Lost job3
Job under threat1
Job refused1
Did not lose job10
Not contactable13

Method of establishing correct diagnosis

More than 50% of those with psychogenic attacks underwent in-patient investigation to `refute the original diagnosis'(Table 9a). In contrast, the vast majority (89%) of patients referred directly to a specialist are diagnosed on clinical grounds only, irrespective of the eventual diagnosis (Table 9b).

View this table:
Table 9a

Consequences of misdiagnosis: investigations required to establish the correct diagnosis

DiagnosisClinicalIEEG +imagingIEEG +AEEG +imagingImaging onlyTotal
NEA, non-epileptic attacks; IEEG, interictal EEG (out-patient); AEEG, ambulatory EEG (in-patient); NFC, no firm conclusion.
View this table:
Table 9b

Patients referred de novo: investigations required to make the diagnosis

DiagnosisClinicalIEEG +imagingIEEG +AEEG +imagingImaging onlyTotal
Abbreviations as Table 9a.

Outcome of patients referred with `refractory epilepsy'

Of the 94 patients referred with `refractory epilepsy', 12 (13%) had pseudoseizures, and, of the remainder, 16/82 (19%) were seizure-free (Table 10). Only 7 of these 16 patients (five with temporal lobe epilepsy treated surgically and two with valproate-resistant idiopathic generalized epilepsy responding to lamotrigine) had genuinely refractory epilepsy. The main cause of `pseudorefractoriness', was failure to recognize idiopathic generalized epilepsy, usually juvenile myoclonic epilepsy,8 such that these patients had never received the drug of first choice, sodium valproate. Furthermore two patients, with long-standing partial seizures, had never received an adequate dose of any drug! Many of the patients in the `improved' category had responded to newer anti-epileptic drugs.

View this table:
Table 10

Outcome of patients referred with `refractory epilepsy'

Total referred94
Not epilepsy12
Not known7


Misdiagnosis of epilepsy

`Funny turns' are a common reason for consulting a doctor. Most people see their general practitioner who, if uncertain of the diagnosis, will usually refer to a local physician.15 The patient may be seen by anyone from SHO to consultant level but he/she is unlikely to have received any formal training in neurology or epilepsy. Whether indicated or not, EEGs are often requested16 and tend to be reported by other doctors who receive limited clinical information about the patient and who, sometimes, have had neither neurological nor neurophysiological training. It is not surprising, therefore, that misdiagnoses or incomplete diagnoses occur, as indicated by the results of this study.

In this study, the misdiagnosis of epilepsy in adults was approximately 25%. It could be argued that our definition of `exposure to anti-epileptic drugs' artificially overestimates the misdiagnosis rate, but our findings are in keeping with those from other more selected populations.9

No one is immune to making a false diagnosis of epilepsy. Indeed, in addition to the four cases referred by neurologists, two of the authors (DWC, DS), with a specific interest in epilepsy, would readily concede that they have repeatedly made this mistake. However this is much less likely to happen where the following principal is carefully applied: if in doubt do not diagnose epilepsy but simply await the passage of time before coming to a firm conclusion. During this time, the patient should be asked to keep a careful record of the circumstances of their attacks and further eye-witness descriptions should be obtained. A therapeutic trial of anti-epileptic drugs is recommended in the following circumstances only: if the features of attacks support a diagnosis of epilepsy; unwitnessed events associated with tongue biting, incontinence and amnesia or unprovoked, stereotyped funny turns with a tendency to occur in clusters; and assuming that the response to treatment is carefully monitored by the prescribing physician.

The commonest condition misdiagnosed as epilepsy was syncope. While the common features of syncope (Box 2) are well-recognized, these are often missed unless the patient/eye-witness is directly questioned. Furthermore, and contrary to popular conception, motor phenomena (a brief tonic phase and myoclonic jerks) are extremely common17 and should not be mistaken for convulsive activity. If a `convulsion' has been witnessed, ascertainment of the precise circumstances will usually reveal that the patient has been held in an upright position after losing consciousness, and has sustained a reflex anoxic seizure.18 These events are symptomatic of cerebral hypoxia, not indicative of an inherent predisposition to seizures, do not require treatment and have no implications for an individual's eligibility to drive. The management of syncope involves counselling about avoidance of precipitating factors and about postural manoeuvres which can prevent loss of consciousness when pre-syncopal symptoms occur. Onset in middle age/elderly, absence or brief premonitary symptoms and other cardiac symptoms or signs demand referral to a cardiologist.19

If syncope is suspected, an EEG should not be requested because this much more likely to reveal normal phenomena or non-specific abnormalities than specific epileptiform discharges.16 Misinterpretation of minor abnormalities produces misleading reports which coerce the unwary into prescribing anti-epileptic drugs. If in doubt one should `track down' eye-witnesses and ask patients to keep a record of the date and circumstances of subsequent events.

The consequences of a misdiagnosis of epilepsy

An erroneous diagnosis of epilepsy has physical, psychosocial21 and socioeconomic consequences for the patient, and economic implications for the health and welfare services.

Patients with newly diagnosed and misdiagnosed epilepsy may be intolerant of anti-epileptic drugs, especially if these are introduced too quickly. It is not surprising that 13/46 misdiagnosed patients reported symptoms of dose-related neurotoxicity (diplopia, ataxia, nausea, sedation). Two patients developed acute allergic reactions to carbamazepine. While these are usually mild, fatal hypersensitivity reactions to anti-epileptic drugs are well-documented.22 Prolonged use of phenytoin causes gingival hypertrophy in one third of patients.23 Whilst this recedes with cessation of treatment, one of the two patients who reported this problem required a complete dental clearance. In this study many of these patients were women of child-bearing age but, fortunately, as far as we are aware, no drug-induced congenital malformations occurred. However the authors have experience of women with pseudoseizures who have had children with probable AED-related teratogenicity.

Whilst it is relatively easy to stop the drugs, it is very difficult to reverse the negative psychological and socio-economic consequences of the misdiagnosis of epilepsy. In this study, 12 people had their driving unnecessarily curtailed and four either lost or had to refuse gainful employment. When the diagnosis is changed, a patient may resume driving immediately (if they had not previously informed the DVLA) or within months if their licence has been submitted. However it is extremely unlikely that he/she will regain their `old job' or, in times of high unemployment, easily obtain a `new job'. Consequently these individuals are in the invidious position of having lost their job unnecessarily and being dependant on state benefits to support themselves and their families.

Wrongly-diagnosed epilepsy has economic implications for the health service. In addition to the cost of previous consultations and unnecessary investigations, 13/46 patients required out-patient tests and 8/19 patients with psychogenic attacks required in-patient assessment to help to refute the original diagnosis. In contrast, 89% of patients, referred directly to the specialist, were diagnosed on clinical grounds only, irrespective of the eventual diagnosis, with only 3/140 of the total and 2/17 with psychogenic attacks requiring admission to hospital.

Overall magnitude of the problem

It is very difficult to estimate accurately the absolute number of misdiagnosed cases. Assuming 20% of the accepted annual incidence of 50/105 produces a figure of 6000 per year. Remarkably, if our figures are representative of the approximately 150 practising neurologists in 1991, this produces the same total. Even if these are over-estimates, there seems little doubt that thousands of people receive an erroneous diagnosis of epilepsy each year. If one applied the same rationale to the prevalence data then, at any one point in time, up to 70 000 people may be wrongly diagnosed. Although the diagnosis may be rapidly reversed in the majority of patients, a significant proportion, 12/46 in this study, are considered to have refractory epilepsy. Accordingly, it is safe to assume that tens of thousands of individuals carry this stigmatizing label inappropriately. Furthermore, if one applies the figures from Cockerill et al. on cost of illness4 to only those misdiagnosed patients thought to have refractory epilepsy, the cost of misdiagnosis is conservatively estimated at £125m per annum.

Outcome of patients referred with `refractory epilepsy'

This subgroup of patients is particularly interesting. Firstly, 13% had pseudoseizures,20 which are physical symptoms of psychological origin that, in most cases, are unconsciously produced for unconscious gain. The typical features of pseudoseizures appear in Box 3; the variability of attacks and resistance to examination, especially eye-opening, are particularly important features. About 90% of these patients are female and a `role model', either a family history of epilepsy or experience of epilepsy in a paramedical occupation is often present. Patient or family pressure to `do tests' or prescribe drugs should be resisted. Management involves frank, but not confrontational discussion, cessation of anti-epileptic drugs and referral for counselling.

Secondly, some patients with benign epilepsy syndromes fail to achieve remission. Factors contributing to `pseudorefractoriness' of genuine epilepsy include the suboptimal use of conventional AEDs,24 poor compliance25 and, in particular, failure to classify the epilepsy.6 Recognition of idiopathic generalized epilepsies,26 especially juvenile myoclonic epilepsy27 (Box 4), is particularly important. This syndrome accounts for 6–8% of all epilepsy, and presents to both paediatricians and adult physicians. The myoclonic jerks are often misinterpreted as focal motor seizures with consequent prescription of carbamazepine, to which this condition responds poorly.8 Furthermore, whilst the prognosis is favourable, relapse off treatment is almost inevitable, and failure to recognize this condition often leads to inappropriate cessation of treatment. In this study, six patients with unrecognized idiopathic generalized epilepsy became seizure-free with the introduction of valproate.

Box 4.

Juvenile myoclonic epilepsy

Thirdly, widespread lack of awareness of therapeutic developments can deny patients with genuinely difficult epilepsy early access to optimal management. In this study, two patients with valproate-resistant idiopathic generalized epilepsy responded to the addition of lamotrigine, a novel anti-epileptic drug with a broad spectrum of efficacy.28

Finally that surgical treatment (Box 5) is a realistic treatment option for drug-resistant temporal lobe epilepsy is not widely recognized. Mesial temporal sclerosis (Box 6), the commonest pathology in surgical series, is readily identified by volumetric MRI,29 and is associated with an excellent post-operative outcome. Epidemiological estimates30 suggest the existence of a reservoir of 16 000 cases with 600–1200 new cases presenting each year. However, <400 of these patients are treated surgically each year. There are many reasons for this `treatment gap', but failure to recognize and refer is certainly an important factor. In this study, four patients were seizure-free post-operatively, and five others are undergoing investigation.

Box 6.

Mesial temporal sclerosis


The misdiagnosis of syncope and psychogenic attacks as epilepsy and the inadequate management of of genuine epilepsy are common problems which have negative psychological and socio-economic consequences for patients and economic implications for the health and welfare services. The main reasons for misdiagnosis are incomplete history-taking and misinterpretation of the interictal EEG, while lack of awareness of the classification of epilepsy is primarily responsible for suboptimal management. The ideal solution involves a massive expansion of specialist services which, presently, faces insurmountable problems. Furthermore the structure of post-graduate training of hospital doctors will not have a positive impact on these issues. However, in the short to medium term, the realistic expansion of neurology, and epilepsy services, the development of close links between these services and local primary health-care services, pragmatic audit projects and undergraduate neurology courses have the potential to improve the management of these common clinical problems.


  • Address correspondence to Dr D. Smith, Walton Centre for Neurology and Neurosurgery, Lower Lane, Fazakerley, Liverpool L9 7LJ


View Abstract