10.3 Neonatal Seizures and CNS Infection
Key Takeaways
- Classify events as subtle, clonic, myoclonic, or tonic, and separate jitteriness—stimulus-sensitive, stops with restraint, no gaze deviation—from seizure.
- Check glucose, ionized calcium, sodium, and a metabolic screen while you treat; hypoxic-ischemic encephalopathy is a common cause, but its full neuroprotection pathway is the next chapter.
- Phenobarbital is first-line conceptually for recurrent neonatal seizures after the metabolic fix; conventional EEG is the gold standard, and amplitude-integrated EEG is a limited bedside adjunct that can miss short or low-voltage events.
- Early meningitis is still group B Streptococcus and Escherichia coli (plus Listeria); later think hospital flora and Candida. Congenital cytomegalovirus and toxoplasmosis add calcifications, hearing or eye disease, and sometimes hydrocephalus.
- Start intravenous acyclovir as soon as neonatal herpes simplex is plausible—do not wait for vesicles or polymerase-chain-reaction results—and involve a ventricular drain when ventriculitis or obstructive hydrocephalus complicates infection.
10.3 Neonatal Seizures and CNS Infection
Quick Answer: Treat the glucose and calcium first, then treat the seizure. Subtle mouthing, eye deviation, pedaling, or unexplained apnea may be seizure and needs EEG confirmation. Phenobarbital is the usual first-line antiseizure medicine after the metabolic fix. If herpes simplex is on the list, start intravenous acyclovir without waiting for polymerase-chain-reaction results or a delayed lumbar puncture.
What this section does and does not cover
The Neonatal CCRN Test Plan lists infection, seizures, and (separately) ischemic insult, hemorrhage, and state dysregulation. This section teaches seizure recognition, first-line treatment concepts, and central-nervous-system infection, including meningitis, ventriculitis, and the congenital infections that declare in the NICU. Hypoxic-ischemic encephalopathy (HIE) is named here as a leading cause of seizures. The cooling protocol, staging, and white-matter injury teaching belong in the next chapter, as do germinal-matrix or intraventricular hemorrhage and the pain or sedation exam. Do not import those protocols into this page.
Independent practice for this exam is at /practice/ccrn-neonatal.
Seizure types at the warmer
Neonates do not have tidy generalized tonic-clonic convulsions as their usual first show. The older Volpe bedside list still organizes NICU teaching: subtle, clonic, myoclonic, and tonic. Newer International League Against Epilepsy language maps many “subtle” events to automatisms, autonomic change, or sequential features, and it reminds you that a large share of neonatal seizures are electrographic-only—no clinical sign at all, especially after the first antiseizure dose uncouples movement from the EEG.
Subtle seizures are the events staff argue about: tonic eye deviation, repetitive blinking, chewing or lip-smacking, swimming or pedaling, bicycling, or apnea with color change that is not otherwise explained. They are common and easy to miss. They are also easy to over-call. That is why EEG exists.
Clonic seizures are rhythmic, repetitive jerking of a limb or one side of the face, focal or multifocal. They migrate. They do not stop when you hold the limb. Focal clonic activity has a strong correlation with an EEG seizure and is the clinical pattern you may treat as probable seizure even while leads are being placed.
Myoclonic jerks are lightning-fast. They may be epileptic or not. Benign neonatal sleep myoclonus occurs in sleep, stops on arousal, and has a normal EEG. Epileptic myoclonus persists across states and has an EEG correlate. Do not load phenobarbital for sleep myoclonus you can extinguish by waking the infant.
Tonic events are stiffening—focal tonic posturing is more often epileptic; generalized tonic stiffening in a very preterm infant is often nonepileptic (decerebrate-like) and may lack an ictal EEG. Treat the infant, not the label, but do not assume every stiff preterm is in status.
Jitteriness is the main mimic. It is stimulus-sensitive, often stops when you flex or hold the limb, has no abnormal gaze, and lacks autonomic change. Hypoglycemia and drug withdrawal can produce jitteriness; they can also produce true seizures. If you are uncertain, check glucose and get EEG rather than debating the video in the hallway.
| Pattern | What you see | EEG likelihood |
|---|---|---|
| Subtle / automatism | Eyes, mouth, pedaling, unexplained apnea | Needs EEG; many are nonepileptic or electrographic-only |
| Clonic | Rhythmic jerking that ignores restraint | High; treat as probable seizure |
| Myoclonic | Lightning jerks | Epileptic if they persist across state |
| Tonic | Stiffening | Focal more reliable; generalized in preterms often nonepileptic |
| Jitteriness | Stimulus-sensitive tremor, stops with restraint | Not seizure |
Causes you must not miss
Build the cause list in the order that saves brain. Hypoglycemia is a point-of-care test, not a “later lab.” Symptomatic infants and most values under about 40–45 mg/dL are treated now; the full glucose chapter lives with endocrine content. Hypocalcemia (low ionized calcium) produces clonic or subtle events, especially in the infant of a diabetic mother, the preterm infant after a citrate load, or DiGeorge/22q11 physiology. Hyponatremia, hypomagnesemia, and severe hypernatremia belong on the same panel.
HIE is the most common cause of seizures in term infants who had a sentinel hypoxic event. Mentally tag it, start the metabolic and infection workup, and leave cooling criteria and amplitude-integrated EEG background grading to the next chapter. Infection—bacterial meningitis, herpes simplex encephalitis, congenital viral disease—can be the cause in a well-appearing infant on day 10 as easily as in a shocked infant on day 1. Intracranial hemorrhage and perinatal stroke cause focal clonic seizures; imaging follows stabilization. Inborn errors of metabolism declare as unexplained seizures plus encephalopathy, metabolic acidosis, hyperammonemia, or a peculiar odor; the endocrine/metabolic chapter owns the pathway, but you still draw ammonia, lactate, and newborn-screen status before you decide the events are “just HIE.” Drug withdrawal more often causes jitteriness; seizures still need another explanation until proven otherwise.
Worked scenario. A 39-week infant has repetitive right-arm clonic jerks at 12 hours of life after a difficult vacuum delivery. Glucose is 28 mg/dL. You give dextrose and the jerks stop. That was a metabolic seizure. You still examine for trauma and consider EEG if events return after euglycemia. You do not load a full phenobarbital course as the first move when the glucose is 28 mg/dL and the room has a dextrose syringe.
Phenobarbital, EEG, and aEEG
After you correct what is correctable, phenobarbital remains first-line conceptually for recurrent neonatal seizures in ILAE-linked consensus teaching. A typical intravenous load is on the order of 20 mg/kg, with a possible second load if events continue, then a maintenance plan if the team keeps the drug. Levetiracetam, fosphenytoin, and midazolam appear as next agents; they are not a reason to skip the metabolic labs. Phenobarbital sedates and can uncouple clinical signs from ongoing electrographic seizures—another reason monitoring matters.
Conventional video-EEG is the gold-standard diagnostic test. Amplitude-integrated EEG (aEEG) is a compressed bedside trend that can show prolonged seizures and background voltage. It misses brief or low-amplitude events and is not a license to skip conventional EEG when seizures are the question. Continue monitoring until you capture the event or until a unit-defined seizure-free window (often about 24 hours) or, in cooled infants, through rewarming—the cooling details wait for the next chapter.
Meningitis, ventriculitis, and congenital infection
Bacterial meningitis in the first days of life is still group B Streptococcus and Escherichia coli, with Listeria monocytogenes in the classic early-onset trio. Late-onset disease adds coagulase-negative staphylococci, Staphylococcus aureus, other gram-negatives, and Candida in the extremely preterm infant with central lines and broad prior antibiotics. The infant may have fever or hypothermia, lethargy, a full fontanelle, seizures, or a septic shock picture without a stiff neck—neonates do not advertise meningitis with meningismus.
Ventriculitis is infection of the ventricular lining, often in an infant with a reservoir, external drain, or shunt, or after gram-negative or candidal meningitis. Cerebrospinal fluid from a ventricular tap may show organisms when the lumbar fluid is less impressive. Treatment is systemic antimicrobials plus device management; some teams instill antibiotics into the ventricle under neurosurgical control. An external ventricular drain or reservoir appears when hydrocephalus from infection or debris obstructs outflow, when the abdomen cannot accept a shunt, or when infected hardware must come out. Nursing a drain means a closed, leveled system, exit-site care, and a ban on casual sampling.
Herpes simplex virus (HSV) produces skin-eye-mouth disease, central-nervous-system disease, or disseminated disease. Vesicles help but are often absent when the brain is involved. Seizures, temporal or multifocal EEG change, and cerebrospinal-fluid pleocytosis are enough to start treatment. High-dose intravenous acyclovir starts as soon as neonatal HSV is plausible. Do not wait for polymerase-chain-reaction confirmation, do not wait for vesicles, and do not hold the drug because the lumbar puncture is deferred for instability or thrombocytopenia. Confirmed central-nervous-system HSV is a long course (commonly 21 days) with a repeat cerebrospinal-fluid HSV polymerase-chain-reaction near the end of therapy.
Cytomegalovirus (CMV) is usually congenital in the NICU story: intrauterine growth restriction, thrombocytopenia, hepatitis, chorioretinitis, periventricular calcifications, and later sensorineural hearing loss. Seizures can occur. Treatment decisions (valganciclovir or ganciclovir) are specialist-led; your job is to recognize the constellation and protect hearing follow-up.
Toxoplasmosis still teaches the triad of chorioretinitis, intracranial calcifications, and hydrocephalus. Seizures and a large head bring these infants to you. Treatment is a prolonged antiparasitic combination directed by infectious disease, not a single NICU antibiotic night.
Candida reaches the central nervous system from candidemia. Endophthalmitis and renal fungus balls travel with it. Amphotericin-based therapy and device removal are the usual backbone; lumbar puncture is part of the disseminated-Candida evaluation in the neonate, not an optional extra.
Lumbar puncture without delaying treatment
A lumbar puncture obtains cell count, protein, glucose, culture, and multiplex or HSV polymerase chain reaction. Hold or delay the needle if the infant is too unstable, if there is a skin infection over the site, or if platelets or coagulopathy make spinal hematoma likely—but do not delay antibiotics or acyclovir for a perfect specimen. Treat, stabilize, then tap when you can. Imaging first is appropriate when a mass lesion or severe obstructive hydrocephalus is likely; it is not a reason to withhold acyclovir for six hours of transport logistics.
Worked scenario. A 12-day-old term infant has focal seizures and a temperature of 38.4 °C. There are no vesicles. The first lumbar-puncture attempt is traumatic and incomplete. You still start ampicillin, gentamicin or a third-generation cephalosporin per local early- versus late-onset rules, and intravenous acyclovir. You retry the tap or use interventional help when safe. Waiting overnight for “a cleaner sample” is the exam-fail move.
| Pathogen | Typical NICU story | First treatment concept |
|---|---|---|
| Group B Streptococcus | Early or late meningitis, seizures | Ampicillin-based regimen, then refine |
| Escherichia coli | Early gram-negative meningitis | Ampicillin plus gentamicin; watch resistance |
| HSV | Seizures ± vesicles, day 5–21 | Immediate IV acyclovir |
| CMV | IUGR, calcifications, hearing risk | Specialist antivirals, hearing plan |
| Toxoplasma | Calcifications, chorioretinitis, hydrocephalus | Prolonged antiparasitic course |
| Candida | ELBW, lines, candidemia | Amphotericin, device removal, LP |
- Glucose and ionized calcium are seizure vital signs.
- EEG confirms; aEEG screens and can miss.
- Acyclovir is a now drug when HSV is possible.
- A ventricular drain is a device with infection and leveling rules, not a casual sampling port.
- HIE, hemorrhage, and pain scales continue in the next chapter—use this section to recognize the event and start the first-hour workup.
Which statement about neonatal seizure semiology is most accurate for bedside screening?
After glucose and electrolytes are corrected, a neonate has recurrent electrographic seizures. Which first-line antiseizure approach matches current consensus teaching?
A 12-day-old infant has seizures and cerebrospinal-fluid pleocytosis. Surface cultures and lumbar puncture are in process. What is the correct antiviral decision?