Condition
Change during TH
EEG features during rewarming
Applications of EEG
EEG prognosis
HIE
Delay of onset of normal SWC
Decrease seizure activity
Seizure de novo
Seizure identification
Prognostication
Normal EEG background at any moment
Abnormal EEG background at 36 h
Presence of sleep–wake cycle
Cardiac arrest
Unknown
Stability of background
Seizures de novo
Seizure identification
Prognostication
Reactivity
Continuity
ECMO
Stable background
No modification
Seizure identification
Unknown
Status Epilepticus
Attenuation of ictal discharges
No consistent pattern
Seizure identification
Unknown
Low-amplitude burst suppression
ALF
Unknown
Unknown
Seizure identification
Unknown
Mechanisms of Therapeutic Hypothermia and Impact on EEG
Hypothermia following central nervous system (CNS) injury or ischemia may offer neuroprotection by targeting multiple underlying mechanisms involved in the pathologic effects of ischemia–reperfusion injury [5]. The acute ischemic phase initiates a toxic neuroexcitatory cascade with a release of excitatory amino acids and glutamate. A relative deficit of ATP leads to mitochondrial dysfunction and depolarization of neuronal cell membranes. The resulting high concentration of extracellular glutamate exposes neurons to a hyperexcitable state consisting of stimulation of NMDA receptors, influx of calcium, extracellular acidosis, increased synthesis of nitric oxide (NO), and reactive oxygen species (ROS). All these mechanisms result in neuronal injury and death. Hypothermia modulates the release of excitatory amino acids and reduces the oxidative and nitrosative stress. Hypothermia also significantly reduces cerebral metabolic rate and cerebral blow flow (CBF), thereby mitigating the injurious effects of the hyperemic response to cerebral ischemia. In the subacute state following cerebral injury, a CNS inflammatory response is initiated in response to release of proinflammatory mediators from ischemic tissue with activation of microglial cells and migration of systemic inflammatory cells. Hypothermia attenuates both the pro- and anti-inflammatory response to ischemia. Finally, hypothermia helps to maintain the integrity of the blood–brain barrier by inhibition of metalloproteinase activation, reduction of NO expression, and reduction of aquaporin-4 expression.
Hypothermia induces multiple physiological changes in neonates and children indirectly affecting CBF [6]. The cardiovascular system responds to the decrease in metabolic demand with bradycardia, decreased cardiac output, and global sympathetic tone. Despite these physiological variations, mild and moderate hypothermia does not impact hemodynamic stability. Hypothermia induces a reduction in cerebral and whole-body metabolism, prompting a reduction in CO2 production and altering glucose metabolism, with a risk of neonatal hyperglycemia. Global CBF decreases by about 5 % for every °C reduction of body temperature.
The reduction in cerebral metabolism and CBF produced by hypothermia influence the EEG voltage. Animal and neonatal studies using amplitude-integrated electroencephalogram (aEEG) have demonstrated a stable voltage despite mild hypothermia up to 34 °C. In adult studies, within a few minutes after initiation of cooling, periodic, unilateral, bilateral, or dyssynchronous complexes may occur, superimposed on a continuous background [7]. With colder temperatures (16–33 °C), a reduction in the amplitude of the electrical activity can be seen, and eventually a burst suppression pattern is observed. Finally, electrocerebral silence occurs within an hour of cooling, with temperature ranging from 2.5 to 27.2 °C. In neonates, hypothermia also modifies the sleep pattern, producing a reduction in the time spent in deep sleep and a delay in the onset of normal sleep–wake cycling.
Patients undergoing TH often receive analgesia, sedation, and neuromuscular blockade that may confound the clinical exam. The use of these medications also influences EEG activity [8]. Multiple EEG changes have been described in adults and children, and the EEG should be interpreted with this confounding factor in mind. Increased beta activity resulting in a wider field distribution and persistence of beta activity is often seen with the use of GABA agonists (barbiturates, benzodiazepines). Children are more susceptible than adults to this accentuation of beta activity. Background slowing with decreased amplitude and frequency of the alpha rhythm is related to many sedative agents and anticonvulsants. Opioid use in neonates has been reported to correlate with excessive spike and sharp transients, periods of background attenuation, and suppression and lack of trace reactivity.
Neonates
Animal Studies
In a preterm fetal sheep model mimicking preterm brain injury [9], umbilical occlusion was followed by immediate suppression of EEG intensity, remaining significantly suppressed during and after reperfusion. TH produced more pronounced suppression of EEG activity, lasting 9 h after occlusion.
Use of EEG in Perinatal Asphyxia
Perinatal HIE associated with intrapartum asphyxia is a common cause of permanent disability and death in neonates. Moderate TH is an established, evidence-based therapy that improves the neurological outcome of a specific subpopulation of HIE patients. TH significantly reduces death or disability at 18 months, without any adverse events, in infants with moderate HIE [10]. In 2013, a Cochrane review identified 11 randomized-controlled trials of TH, involving 1505 late preterm and term infants with moderate and severe encephalopathy in the context of birth asphyxia. The entry criteria for TH for infants with HIE comprise gestational age ≥36 weeks and ≤6 h of age, Apgar score ≤5 at 10 min after birth or continued need for resuscitation 10 min after birth or pH <7.00 or base deficit ≥16 mmol/L within 60 min after birth, and moderate or severe encephalopathy on clinical examination. Infants receiving preferential head cooling must also demonstrate an abnormal background activity for at least 30 min or seizures detected by aEEG. Many different protocols exist, but TH for the neonate with HIE involves cooling to 32–34 °C within the first 6 h of life for 72 h, with gradual rewarming over the next 24 h. The method of cooling varies according to the institution (whole-body cooling with a blanket versus selected head cooling with a cooling cap).
The American Clinical Neurophysiology Society recommends that all neonates at risk for brain injury undergo continuous video EEG (vEEG) monitoring to assess for the presence of electrographic seizures and to evaluate suspicious clinical events [11]. Continuous bedside vEEG or aEEG are used in neonates suffering from HIE for many purposes: assessment of indication for neuroprotection with TH, monitoring of seizure activity, and evaluation of response to treatment with antiseizure medications and to help determine prognosis. Multichannel vEEG study is the gold standard for accurate identification and quantification of seizures and analysis of background activity in this population. Most patients undergoing TH are monitored with aEEG, continuous vEEG, or periodic vEEG throughout the process of hypothermia and rewarming.
Detection of Seizures Following Perinatal Asphyxia and Effects of Hypothermia
HIE is a common cause of neonatal seizures, which are an important predictor of neurological morbidities and mortality in this population. Prior to the era of TH, the incidence of seizures in neonates with moderate to severe HIE approached 90 % [12]. Despite the neuroprotective and anticonvulsant effects of TH, electrographic seizures are reported in about 50 % of infants undergoing TH following birth asphyxia, and SE occurs in up to 25 % of this population. TH may reduce the risk for electrographic seizures in neonates with HIE [12, 13].
Within the interval from birth to initiation of EEG monitoring (usually a few hours), about half of moderate to severe HIE patients have clinical events suspicious for seizure activity and most of these patients are treated with phenobarbital initially (78 %) [13]. Half of these patients continue to have further electrographic seizures while monitored by continuous vEEG. Interestingly, electrographic seizure onset while monitored by vEEG occurs at a median time of 13–18 h, but a small number of patients are in electrographic SE at the onset of recording [13]. Seizures may occur de novo during rewarming, but the reported incidence is variable. A 3-center observational study of 90 term neonates treated with whole-body TH for HIE who underwent continuous vEEG monitoring on the first day of life and continued for >24 h identified electrographic seizures in 48 % of cases [14]. Electrographic seizure onset occurred at a median of 19.9 h of life. Notably, treatment with phenobarbital prior to vEEG recording was not associated with risk for electrographic seizures, and only 4 % of the study population had seizure onset de novo during rewarming. The EEG of a 3-day-old infant with HIE and multiple focal electrographic seizures during rewarming is shown in Fig. 1.
Fig. 1
EEG of a 3-day-old infant, 39 postconceptional age, with hypoxic ischemic encephalopathy (HIE). The study was performed during rewarming. There are multiple focal electrographic seizures with onset from the left occipital region; excess multifocal sharp transients for age, maximal in the occipital regions; and excessively discontinuous background for conceptual age. These findings are suggestive of diffuse cerebral dysfunction and multifocal potentially epileptogenic foci with the dominant focus in the left occipital region
Neonates with HIE undergoing TH have a lower seizure burden than normothermic neonates. Seizure burden is defined in this context as the total duration of electrographic seizures in seconds. For a similar population in terms of patient characteristics and HIE severity, cooled neonates have a lower seizure burden, but both groups develop a comparable number of seizure events. The mean duration of seizures and prevalence of SE are similar with regard to antiseizure medication before and after hypothermia initiation [12, 13]. Newborns with moderate HIE show a greater reduction in seizure burden with TH than those with severe HIE [12]. The mechanisms producing a decrease in seizure activity during TH and occurrence of seizure during rewarming may be multifactorial. TH may delay the evolution of CNS injury leading to a delay in onset of seizures. TH may also suppress seizure activity, and consequently, seizures occur later during the return to normothermia. The use of TH for HIE has lead to a change in the duration of EEG monitoring. Normothermic neonates were previously monitored for the first 2 days of life, but infants undergoing TH are now often monitored through hypothermia and rewarming (approximately 4 days) [13].
An abnormal vEEG background is a risk factor for seizures following HIE [14]. An excessively discontinuous vEEG background at the onset of monitoring is associated with a 70 % risk of seizures. A severely abnormal, but not discontinuous, and normal vEEG background are associated with a 63 % and 12 % risk of seizures respectively. No perinatal clinical variable predicts the risk of electrographic seizures. Accordingly, vEEG monitoring remains an essential tool to assess seizure activity during TH in the asphyxiated newborn.
The relationship between seizure burden and the severity of CNS injury remains unclear, but it is possible that reducing the amount and duration of seizures with TH improves neonatal outcome. Seizure burden may reflect the severity of CNS injury or may contribute to and exacerbate underlying injury. Clinical recognition of seizures in this specific population is known to underestimate the real seizure incidence, as many neonatal seizures are subclinical. Accordingly, continuous vEEG monitoring allows prompt recognition and treatment of seizures. Importantly, there is no consensus regarding the threshold for treatment of electrographic seizures.
Use of Amplitude-Integrated EEG in Perinatal Asphyxia
Multichannel vEEG remains the gold standard for background analysis and seizure detection, but its use in the neonatal intensive care unit is limited by availability and by the expertise required for application and interpretation. Amplitude-integrated EEG is a simplified alternative method allowing brain activity monitoring with 3–5 electrodes attached to the scalp. The electrode application is simpler and aEEG interpretation is reliable for neonatologists and intensivists trained in its use. In a study comparing the use of continuous vEEG and aEEG in infants at risk for seizures, single-channel aEEG identified seizures in 56 % of cases. The ability to detect seizures using aEEG correlated with the duration of seizure, aEEG background, and skills of the interpreter [15]. Nevertheless, the sensitivity of aEEG to detect seizures during TH and following perinatal asphyxia remains controversial.
Amplitude-integrated EEG may be used as a tool to help in the selection of neonates for TH. In a study examining the predictive value of early aEEG (within the first 6–9 h) and clinical staging of encephalopathy among children with HIE, the strongest variables associated with a poor neurological outcome were a neurological exam compatible with severe HIE, followed by an abnormal aEEG (burst suppression, continuous low voltage or flat voltage) and a moderate HIE assessed by neurological examination [16]. Therefore, aEEG may be helpful for the selection of children with moderate HIE, as their early aEEG is more sensitive than the clinical examination for prediction of poor outcome in this population. This finding is important for this specific population, as infants with moderate HIE derive the greatest benefit from TH.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
