RCGP SCA Algorithm — UK Primary Care
First episode of loss of consciousness or seizure-like activity requires urgent assessment — do not manage in routine follow-up.
The clinical challenge is that 30–50% of patients presenting with "fits" do not have epilepsy — syncope and psychogenic non-epileptic attacks (PNEA) are commonly misdiagnosed as epilepsy, leading to unnecessary antiepileptic treatment. Conversely, cardiac syncope from arrhythmia accounts for 15% of sudden cardiac deaths in under-35s and is life-threatening if missed. The single most useful discriminator is a witness account — always obtain collateral history.
90% of the diagnosis comes from history. Always obtain a witness account — phone the witness if necessary.
Lateral tongue biting has a 96% specificity for epileptic seizure — it is the single most reliable feature distinguishing seizure from syncope. Prolonged post-ictal confusion (>5–10 minutes) is equally specific. The Edmonton Syncope Score and LOSS Score help risk-stratify cardiac vs non-cardiac syncope. DVLA notification is a legal requirement — failure to advise patients about driving restrictions after first seizure is a significant medico-legal risk.
Psychogenic non-epileptic attacks (PNEA) account for 20–30% of first presentations to epilepsy clinics — misdiagnosis leads to years of unnecessary antiepileptic drugs. EEG during an attack showing normal activity confirms PNEA. Cardiac syncope is lethal if missed: Brugada syndrome, long QT, and HOCM cause sudden cardiac death in young, apparently healthy people — the ECG is the key screening test.
Lying-to-standing BP is essential in any unexplained LOC — orthostatic hypotension is present in 20% of community-dwelling elderly and is frequently missed. An innocent-sounding ejection systolic murmur in a young patient with exertional syncope may indicate HOCM — the murmur increases on Valsalva (unlike aortic stenosis). Papilloedema in a patient with seizures mandates emergency CT before LP — lumbar puncture with raised ICP causes tonsillar herniation.
ECG is the most important immediate investigation — long QT syndrome causes 3,000 UK deaths per year, many in young people without prior symptoms. QTc measurement requires a calibrated measurement (automated ECG machines often miscalculate). NICE NG217 recommends MRI brain as the preferred imaging modality for new seizure — CT misses ~30% of epileptogenic lesions visible on MRI. EEG has false-negative rate of 50% in epilepsy — a normal EEG does NOT exclude epilepsy.
NICE NG217 is explicit: all patients with a first seizure must be seen by an epilepsy specialist within 2 weeks. Starting AEDs in primary care before specialist review is inappropriate — the diagnosis must be confirmed because 30–50% of "first seizures" are not epilepsy. DVLA rules for seizures are strict and legally binding — there is significant clinical and medico-legal risk in failing to advise patients. For vasovagal syncope, the DVLA publishes specific criteria — direct patients to dvla.gov.uk/medical-conditions.
Sodium valproate is contraindicated in women of childbearing potential unless there is a Pregnancy Prevention Programme in place (MHRA 2018) — it causes neural tube defects and neurodevelopmental delay in ~10% of in-utero exposed children. Lamotrigine is preferred for women. Physical counterpressure manoeuvres for vasovagal syncope are backed by RCT evidence (PC Trial) — they reduce syncopal episodes by 39% without medication. Every GP surgery should have buccal midazolam for seizure management.
SUDEP is the leading cause of death in young people with epilepsy — it kills approximately 1,000 UK people per year. Good seizure control (achieved in ~70% with correct medication) dramatically reduces SUDEP risk. Enzyme-inducing AEDs (carbamazepine, phenytoin, phenobarbital, topiramate, oxcarbazepine) reduce contraceptive pill efficacy by up to 50% — unintended pregnancy with ongoing teratogenic AED exposure is a significant preventable harm. Vasovagal syncope patient education about physical counterpressure techniques prevents ~40% of further episodes.
Carbamazepine causes hyponatraemia (SIADH) in 5–10% of patients — elderly patients are at particular risk of falls from dilutional hyponatraemia. Valproate causes hyperammonaemia in ~5% — presenting as encephalopathy. Annual review is the NICE standard for established epilepsy — seizure diaries should be reviewed, seizure frequency tracked, and medication step-up considered if breakthrough seizures occurring. Folic acid 5 mg (not the standard 400 mcg) is required in women with epilepsy on AEDs — enzyme-inducing AEDs deplete folate, and higher supplementation is needed to reduce neural tube defect risk.