Intended for healthcare professionals


Epilepsy: getting the diagnosis right

BMJ 1997; 314 doi: (Published 18 January 1997) Cite this as: BMJ 1997;314:158

All that convulses is not epilepsy

  1. J W A S Sander, Senior lecturer,a,
  2. M F O'Donoghue, Research fellowa
  1. a Institute of Neurology, National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3BG

    Epilepsy may present with a variety of symptoms, and other conditions may mimic its manifestations. The diagnosis is almost always based solely on the clinical history. It is therefore not surprising that diagnostic accuracy remains a major problem.1 About a fifth of patients referred to specialist units with “intractable epilepsy” are found, on further assessment, not to have epilepsy.2 It is also common for patients to have symptoms for months or even years before epilepsy is diagnosed. Thus, it is important to be aware of both the heterogeneous and sometimes subtle forms of epilepsy and of the alternative diagnoses.3

    The differential diagnosis of epilepsy includes all causes of transient loss of awareness, falls, paroxysmal sensory-motor phenomena, and generalised convulsive movements which are the most common presenting symptom of epilepsy.3 Tonic-clonic seizures (“a convulsion”) start with sudden loss of awareness, a guttural cry, generalised stiffening of the body and limbs, followed by rhythmic jerking of the limbs, often associated with tongue biting and urinary incontinence. The convulsive movements usually last for at most one to two minutes, and, as the attack proceeds, the jerking slows in frequency and increases in amplitude. There is often cyanosis and irregular breathing followed by confusion, headache, and drowsiness. Tonic-clonic seizures may sometimes be preceded by myoclonic jerks in idiopathic generalized epilepsy or by a simple partial seizure (“aura”) in partial epilepsy. When all or most of these features are reported there is little room for diagnostic confusion.3

    However, other conditions may present with similar phenomena. If misinterpreted, these can lead to unnecessary treatment and social and occupational handicap. The most common sources of confusion are vasovagal syncope and non-epileptic attacks of a psychological origin. Syncope is often misdiagnosed, as it may be accompanied by brief stiffening or jerking of the extremities, and consequently is liable to be reported as a convulsion by witnesses. A video study of syncope induced in healthy volunteers has shown that multifocal and generalised myoclonic jerks are common in syncope.4 However, syncope can usually be correctly identified by the presence of precipitating factors and prodromal symptoms. Syncope often occurs on prolonged standing or when rising quickly, particularly if associated with peripheral vasodilatation. Syncope is unusual when recumbent, unless it is of cardiac origin. Fright, painful stimuli, cough, and micturition (particularly in older people) may also be triggers. Syncopal attacks are preceded by a feeling of lightheadedness, dizziness, nausea, ringing in the ears, and the vision “going grey”–features that are rare in epilepsy. Incontinence is rare, and recovery of consciousness usually occurs within a minute.3

    Non-epileptic attack disorder may be characterised by semi-purposeful thrashing of all four limbs that waxes and wanes in intensity over many minutes, and some patients exhibit prominent pelvic movements and back arching, often with evidence of retained awareness.2 Recovery is variable and may be much quicker than expected from the duration of the attack.3

    In this week's BMJ, McCrory and colleagues describe what they call “concussive convulsion,” (p 171) another potential pitfall in the diagnosis of epilepsy.5 Convulsions that occur within seconds of an impact to the head have been widely assumed to represent a form of post-traumatic epileptic seizure, but McCrory and colleagues suggest that these are a non-epileptic phenomenon. Studying a series of 22 well documented attacks, some captured on video, that occurred after minor head trauma during Australian football, the authors were able to ascertain the benign nature of these attacks. The convulsions occurred within two seconds of the impact and resembled tonic-clonic seizures. The convulsions were usually brief, but some lasted for over two minutes. Recovery was quick; indeed, in two cases the players were alert and oriented within seconds of the convulsive event. Of particular importance is that, after a mean follow up of 3.5 years, no subjects developed epilepsy. This corroborates Jennett's observation over 20 years ago that seizures confined to the time of the head injury are not associated with subsequent epilepsy.6

    The clinical characteristics of concussive convulsions seem to differ somewhat from those after syncope. There is a more prominent tonic phase, and the attack lasts longer. Electroencephalography shows that syncopal convulsions are non-epileptic, but it remains to be established whether these events arise as a result of an epileptic discharge. McCrory et al argue that they do not. Alternatively, we suggest that they represent an acute symptomatic but benign seizure. Pathophysiology notwithstanding, the observations provide helpful prognostic information. Concussive convulsions should be distinguished from seizures that occur within the first week of head injury rather than instantly after impact. These carry a 25% risk of later epilepsy.6

    Every effort should be made to reach a firm diagnosis in cases of possible epilepsy. If doubt remains after the first event it is usually wise to await further events and reach a secure diagnosis, rather than initiate anti-epileptic treatment prematurely.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.