Built on West Midlands Neonatal Guidelines 2025–28 (Seizures) · ILAE neonatal seizure classification · WHO/AAP · Baylor Guidelines for Acute Care of the Neonate
Neonatal seizures are the commonest overt sign of neurological dysfunction in the newborn and are almost always symptomatic of an underlying, often treatable, cause rather than idiopathic epilepsy. Two features make them distinctive: many are subtle or purely electrographic (electroclinical dissociation), so EEG/aEEG is central to diagnosis; and the priority is to find and treat the cause (especially hypoglycaemia, electrolyte disturbance, infection, HIE, and stroke) as much as to give an anticonvulsant.
Prolonged and repetitive seizures may worsen brain injury, and both under-treatment (missing electrographic seizures) and over-treatment (sedating a jittery baby or a non-seizing baby) cause harm. A structured approach — recognise, correct reversible causes, confirm with EEG, treat with a stepwise drug ladder, and diagnose the cause — gives the best outcome.
A neonatal seizure is a paroxysmal, abnormal, stereotyped alteration of neurological function (motor, behavioural, and/or autonomic) with a corresponding electrographic correlate. The ILAE framework emphasises that neonatal seizures are, by definition, confirmed or strongly supported by EEG.
Neonatal seizures are often subtle; recognise the range of motor and autonomic manifestations.
| Type | Features | Note |
|---|---|---|
| Subtle | Eye deviation/blinking/staring; oral–buccal–lingual movements (chewing, sucking, lip-smacking); "boxing/cycling/pedalling" limb movements; apnoea. | Most common; easily missed — EEG confirms. |
| Clonic | Rhythmic jerking, focal (one limb) or multifocal. | Often has a clear EEG correlate. |
| Tonic | Sustained posturing, focal or generalised. | Generalised tonic events may lack EEG correlate — confirm. |
| Myoclonic | Rapid, single or multiple jerks. | Distinguish from benign sleep myoclonus. |
| Autonomic | Apnoea, tachycardia, blood-pressure/pupil changes. | Consider seizure in unexplained recurrent apnoea. |
A frequent bedside dilemma — these features reliably separate the two.
| Sign | Jitteriness | Seizure |
|---|---|---|
| Stimulus-provoked | Yes | No |
| Movement | Rapid, oscillatory tremor | Clonic/tonic |
| Stops when limb held | Yes | No |
| Conscious state | Awake or asleep, normal | Often altered |
| Eye deviation | No | Yes |
| Agent and rank | Loading and maintenance | Targets and cautions |
|---|---|---|
| Phenobarbital — first line (very low certainty evidence, strong recommendation) | 20 mg/kg IV at 1–2 mg/kg/min, then up to two further 10 mg/kg doses to a total of 40 mg/kg | Target level 20–40 mcg/mL. Respiratory depression may force intubation. |
| Midazolam — second line (very low certainty, weak) | 0.15 mg/kg IV bolus, a further 0.1–0.15 mg/kg after 5–10 minutes, then an infusion at 0.1–0.4 mg/kg/h if the bolus does not abort the seizure | Preferred in the sick, already intubated neonate. Respiratory depression may necessitate intubation. |
| Fosphenytoin — second line (very low certainty, weak) | 20 mg/kg IV at 0.5–1 mg/kg/min; level 2 hours after the load | Total phenytoin 15–20 mcg/mL. Because protein binding differs qualitatively and quantitatively in neonates, adult correction equations do not apply — send a free level, target 1–2 mcg/mL. Hypotension and arrhythmia occur; avoid in cardiac patients. |
| Levetiracetam — third line | 60 mg/kg IV at 2–5 mcg/kg/min; maintenance 20–60 mg/kg/day divided three times daily | The trials conflict. Gowda 2019 (100 neonates, clinical seizures) found cessation in 86% with levetiracetam against 62% with phenobarbital, adverse events only in the phenobarbital arm — but without EEG, drug levels or follow-up. Sharpe 2020, blinded and multicentre with electrographic endpoints, found 80% seizure-free at 24 hours on phenobarbital against 28% on levetiracetam. Sharpe generalizes better because it captured subclinical seizures — which is why phenobarbital stays first line. |
| Pyridoxine (very low certainty, strong recommendation) | 1–5 IV doses of 100 mg, then oral 15–30 mg/kg/day, up to 200 mg/day | For recurrent neonatal seizures unresponsive to conventional agents. Risk of respiratory failure with the IV dose. Continue until biochemical or genetic testing excludes pyridoxine-dependent epilepsy, then stop — long-term use causes peripheral neuropathy. |
Two Cochrane reviews questioned treating every seizure: the 2001 review (updated 2004) found "little evidence from randomized controlled trials to support the use of any anticonvulsant in the neonatal period", and the 2007 review concluded that medication after perinatal asphyxia "cannot be recommended for routine clinical practice, other than in the treatment of prolonged or frequent clinical seizures". Animal models show some of these agents produce widespread neuronal apoptosis. Since aetiology may matter more than the seizures themselves for outcome, choose who to treat carefully. Why phenytoin is not first line despite equal efficacy in head-to-head studies: a narrow therapeutic range demanding frequent levels, poor oral absorption in the newborn, and cardiotoxicity concerns with fosphenytoin.
Stopping: Glass and colleagues found that where medication was successfully discontinued there was no adverse effect on functional neurodevelopment at 24 months, no association between treatment duration and later epilepsy or gross motor function, and that prolonged treatment is unnecessary for most infants. Consider routine discontinuation once acute symptomatic seizures resolve; where it cannot be stopped, keep to one agent for the shortest possible time. Consult neurology for every neonatal seizure, and the epilepsy service for refractory seizures or status.
Correct reversible causes first, then use a stepwise anticonvulsant ladder. Confirm all doses against the Neonatal Formulary.
| Agent | Indication | Dose (verify) | Cautions / monitoring |
|---|---|---|---|
| Glucose 10% | Hypoglycaemic seizure | 2.5–5 mL/kg IV bolus, then maintenance infusion. | Take hypoglycaemia screen first where possible; recheck glucose. |
| Calcium gluconate 10% | Hypocalcaemic seizure | 0.5 mL/kg IV over 5–10 min. | ECG monitoring; severe tissue injury if extravasated. |
| Magnesium sulphate | Hypomagnesaemic (and refractory hypocalcaemic) seizure | 100 mg/kg IV or deep IM. | Monitor BP, tone, respiration. |
| Phenobarbital | First-line anticonvulsant | Load 20 mg/kg IV; further 10 mg/kg increments to ~40 mg/kg per protocol. | Sedation, respiratory depression, hypotension; monitor levels. |
| Levetiracetam | Second-line | Per Formulary (commonly 20–40+ mg/kg IV load). | Generally well tolerated; efficacy may be lower than phenobarbital. |
| Phenytoin / fosphenytoin | Second-line | Load ~20 mg/kg (PE for fosphenytoin) per Formulary. | Cardiac monitoring; extravasation risk with phenytoin. |
| Midazolam | Refractory seizures | Infusion per protocol; buccal/intranasal if no IV and metabolic causes excluded. | Sedation, hypotension, respiratory depression. |
| Pyridoxine | Refractory/unexplained (diagnostic-therapeutic) | 50–100 mg IV, ideally during EEG. | Apnoea/hypotension possible — give with monitoring/support ready. |
| Aciclovir | Empirical HSV cover pending exclusion | Per Formulary. | Renal function/hydration. |
Confirm seizure control electrographically, not only clinically.
| Parameter | Frequency | Target | Action if abnormal |
|---|---|---|---|
| EEG / aEEG | Continuous where available | No electrographic seizures | Escalate anticonvulsant ladder; treat electrographic seizures. |
| Glucose / Ca / Mg / electrolytes | Early & serial | Normal ranges | Correct promptly — often stops seizures. |
| Cardiorespiratory | Continuous | Stable; no apnoea | Support ventilation; anticonvulsants cause depression. |
| Drug levels (phenobarbital/phenytoin) | If on maintenance | Therapeutic range | Adjust dose; watch toxicity. |
| Cause work-up (CSF, MRI, metabolic) | As indicated | Diagnosis established | Direct specific therapy (e.g., aciclovir, pyridoxine). |
The recurring errors in neonatal seizure care.
| Mistake | Why it harms | Better practice |
|---|---|---|
| Treating jitteriness as seizures. | Unnecessary sedation, missed real diagnosis. | Apply jitteriness-vs-seizure features; confirm with EEG. |
| Not checking/treating glucose first. | Misses the fastest reversible cause. | Immediate glucose; correct electrolytes. |
| Relying on clinical signs alone. | Electrographic-only seizures missed or over-called. | Use EEG/aEEG for diagnosis and to gauge response. |
| Adding a second drug before maximising the first. | Sub-therapeutic dosing. | Load to the maximum of one agent before the next. |
| Forgetting HSV/meningitis or pyridoxine-dependency. | Untreated, dangerous causes. | Empirical aciclovir/antibiotics; pyridoxine trial if refractory. |
| Prolonged maintenance without indication. | Unnecessary drug exposure. | Keep duration short; reassess need. |