Full-Length Original Research
Cardiopulmonary complications during pediatric seizures: A prelude to understanding SUDEP
Drs. Singh and Katz contributed equally to this work and are the co-first authors in this study.
Drs. Pavlova and Kothare contributed equally to this work and are the co-last authors in this study.
Sudden unexpected death in epilepsy (SUDEP) is an important, unexplained cause of death in epilepsy. Role of cardiopulmonary abnormalities in the pathophysiology of SUDEP is unclear in the pediatric population. Our objective was to assess cardiopulmonary abnormalities during epileptic seizures in children, with the long-term goal of identifying potential mechanisms of SUDEP.
We prospectively recorded cardiopulmonary functions using pulse-oximetry, electrocardiography (ECG), and respiratory inductance plethysmography (RIP). Logistic regression was used to evaluate association of cardiorespiratory findings with seizure characteristics and demographics.
We recorded 101 seizures in 26 children (average age 3.9 years). RIP provided analyzable data in 78% and pulse-oximetry in 63% seizures. Ictal central apnea was more prevalent in patients with younger age (p = 0.01), temporal lobe (p < 0.001), left-sided (p < 0.01), symptomatic generalized (p = 0.01), longer duration seizures (p < 0.0002), desaturation (p < 0.0001), ictal bradycardia (p < 0.05), and more antiepileptic drugs (AEDs; p < 0.01), and was less prevalent in frontal lobe seizures (p < 0.01). Ictal bradypnea was more prevalent in left-sided (p < 0.05), symptomatic generalized seizures (p < 0.01), and in brain magnetic resonance imaging (MRI) lesions (p < 0.1). Ictal tachypnea was more prevalent in older-age (p = 0.01), female gender (p = 0.05), frontal lobe (p < 0.05), right-sided seizures (p < 0.001), fewer AEDs (p < 0.01), and less prevalent in lesional (p < 0.05) and symptomatic generalized seizures (p < 0.05). Ictal bradycardia was more prevalent in male patients (p < 0.05) longer duration seizures (p < 0.05), desaturation (p = 0.001), and more AEDs (p < 0.05), and was less prevalent in frontal lobe seizures (p = 0.01). Ictal and postictal bradycardia were directly associated (p < 0.05). Desaturation was more prevalent in longer-duration seizures (p < 0.0001), ictal apnea (p < 0.0001), ictal bradycardia (p = 0.001), and more AEDs (p = 0.001).
Potentially life-threatening cardiopulmonary abnormalities such as bradycardia, apnea, and hypoxemia in pediatric epileptic seizures are associated with predictable patient and seizure characteristics, including seizure subtype and duration.
Sudden unexpected death in epilepsy (SUDEP) is an important cause of death in epilepsy (Nilsson et al., 2003) and has an incidence ranging from 0.09 to 9 per 1,000 patient-years (Walczak et al., 2001). Studies have suggested that higher seizure frequency, antiepileptic drugs (AEDs), seizure duration, childhood-onset, male gender, and symptomatic epilepsy are associated with SUDEP (Walczak et al., 2001). Cardiac abnormalities (Espinosa et al., 2009), respiratory abnormalities (Johnston et al., 1995), and a combination of both (Walczak, 2003; Langan et al., 2005) are postulated to cause SUDEP. Age affects the rate and severity of respiratory abnormalities (Young et al., 1993; Pavlova et al., 2008). SUDEP most commonly occurs during sleep (Langan et al., 2005; Nashef et al., 2007; Lamberts et al., 2012). Sleep may lead to potentiation of epileptiform activity (Pavlova et al., 2008), thus sleep stage and sleep/wake state may influence the seizure likelihood (Herman et al., 2001; Langan et al., 2005).
There have been a few prospective studies addressing breathing abnormalities during seizures in children, although several studies have addressed this issue in adults (see Table 2 for details). Most pediatric studies, however, were either descriptive (Hewertson et al., 1994, 1996; O'Regan & Brown, 2005)—showing presence of apneic events in apparent life-threatening events (ALTEs), ictal- and postictal tachypnea, apnea, respiratory pauses and hypoxemia in focal seizures, or included anecdotal reports or small case-series of central apnea during seizures (Southall et al., 1987; Singh et al., 1993). Some pediatric studies focused on cardiac function only–demonstrating association of ictal tachycardia, bradycardia, and desaturation with several seizure characteristics (Moseley et al., 2010, 2011). In adult seizure studies using long-term video–electroencephalography (EEG) monitoring, respiratory abnormalities were seen in majority of patients, but these were confirmed by visual observation (Bateman et al., 2010; Lhatoo et al., 2010).
Therefore, the interaction between seizures, sleep, and cardiopulmonary function is an important, yet understudied, area and there is an urgent need to characterize these findings in children. Consequently, our objective was to prospectively investigate whether seizures provoke systematic and predictable cardiopulmonary abnormalities related to patient age, seizure characteristics, severity, and sleep/wake state.
The study was approved by the Children Hospital Boston's Institutional Review Board, and all enrolled patients or their legal guardians provided signed, informed consent to participate. The study population was drawn from patients admitted to the hospital's video-EEG long-term monitoring unit between September 2010 and August 2011, either to establish the diagnosis of epilepsy or to localize the seizure focus. All patients had 50% AED reduction on day 1 and complete withholding of AEDs on day 2. Inclusion criteria were age 1–21 years (inclusive). Patients with nonepileptic events were excluded. One hundred forty-five patients met the eligibility criteria and 29 of those agreed to participate—the rest did not agree due to the prospects of having respiratory belts attached in addition to the EEG electrodes throughout the admission.
Respiratory inductance plethysmography (RIP) (Iber et al., 2007) was used to monitor chest and abdominal excursion during breathing. RIP consists of an elastic belt placed over each of the thorax and abdomen. Breathing results in cross-sectional thoracic and abdominal area changes recorded as continuous tracings. The thorax and abdominal signals can be summed to yield a measure related to tidal volume. RIP can therefore be used to distinguish obstructive from central apneas. Central apnea was defined two or more missed effortless breaths, tachypnea/bradypnea as 10% change in respiratory rate from baseline for two or more breaths. Respiratory analysis was performed by a coauthor (EK). ProTech zRIP belts were used (Pro-Tech Diagnostics, Pittsburgh, PA, U.S.A.).
Finger pulse-oximetry sensors (Masimo-Rad-57; Masimo-Corp, Irvine, CA, U.S.A.) were used for oxygenation. Desaturation was defined as ≥3% decrease from baseline SaO2 levels, or SaO2 value <92%. Standard single-lead electrocardiography (ECG) channel with chest electrodes was used for ECG. Tachycardia was defined as heart rate >100/min, and bradycardia <60/min.
Continuous video-EEG monitoring (XLTEC-EEG-EMU40-unit; Natus-Medical, San Carlos, CA, U.S.A.): Scalp-EEG recordings were performed using the 10–20 system of electrode placement. Continuously monitored video recordings were performed with digital closed-circuit video-cameras. EEG-data were analyzed by a board-certified epileptologist (SK) based on occurrence of seizures during day or night, relationship to sleep/awake, and lobar localization. Seizure duration was determined by marking seizure onset and end on the EEG. Seizures were classified using the 1981 and 1989 International League Against Epilepsy (ILAE) guidelines (ILAE, 1981, 1989).
Respiratory rate, oxygen saturation, heart rate, and central/obstructive apnea were assessed. Age, seizure localization, seizure type, generalization, sleep/awake state at seizure onset, lung disease (obstructive/restrictive), and MRI abnormalities were recorded. Bivariate associations between outcome and predictor variable were accessed through logistic regression. Significant association was reported if regression coefficient significantly differed from zero. Wald-type tests were used to determine significance. Odds ratios (OR) with p-values were reported. Exact logistic regression was employed in cases of sparse number of events in a stratum. Adjustment for multiple comparisons was done by Bonferroni correction, and significant results after the correction are indicated in Table 1. Statistical analysis was done with SAS 9.3 (SAS-Institute, Cary, NC, U.S.A.).
Table 1. Relationships between cardiopulmonary disturbances and patient/seizure characteristicsa
|Age|| 0.85 (0.74–0.97) || 0.01 || 1.19 (1.04–1.37) || 0.01 ||1.03 (0.88–1.20)||>0.1|| 0.89 (0.77–1.02) || <0.1 || 1.12 (0.98–1.27) || <0.1 ||0.96 (0.85–1.08)||>0.1|
|Gender (female vs. male)|| 0.39 (0.14–1.11) || <0.1 || 2.79 (0.99–7.82) || 0.05 ||0.44 (0.09–2.18)||>0.1|| 0.21 (0.05–0.84) || <0.05 || 1.99 (0.61–10.48) || 0.001c ||0.45 (0.16–1.24)||>0.1|
|Generalization (yes vs. no)||1.11 (0.46–2.70)||>0.1||0.97 (0.40–2.39)||>0.1||0.94 (0.29–3.10)||>0.1||0.45 (0.14–1.49)||>0.1||1.08 (0.37–3.10)||>0.1||0.91 (0.33–2.49)||>0.1|
|Localization (temporal vs. frontal)|| 8.04 (2.47–26.19) || <0.001c || 0.22 (0.07–0.68) || <0.01 || 4.89 (0.90–26.76) || <0.1 ||1.95 (0.37–10.31)||>0.1||0.80 (0.21–3.09)||>0.1||3.36 (0.87–12.93)||>0.1|
|Localization (temporal vs. others)||1.38 (0.39–4.91)||>0.1||0.91 (0.25–3.31)||>0.1||0.95 (0.24–3.72)||>0.1||0.30 (0.06–1.52)||>0.1||2.67 (0.59–12.04)||>0.1||0.86 (0.20–3.66)||>0.1|
|Localization (frontal vs. others)|| 0.17 (0.05–0.58) || <0.005c || 4.08 (1.24–13.43) || <0.05 || 0.19 (0.03–1.12) || <0.1 || 0.15 (0.04–0.65) || 0.01 || 3.33 (0.94–11.81) || <0.1 ||0.26 (0.08–8.41)||>0.1|
|Lateralization (left vs. right)|| 2.91 (0.95–8.99) || <0.1 || 0.12 (0.04–0.41) || <0.001c || 10.0 (1.15–86.74) || <0.05 ||1.03 (0.24–4.50)||>0.1||1.29 (0.33–5.02)||>0.1||1.94 (0.55–6.94)||>0.1|
|MRI lesions (yes vs. no)||1.81 (0.66–4.93)||>0.1|| 0.34 (0.12–0.94) || <0.05 || 6.0 (0.73–49.30) || <0.1 ||1.87 (0.51–6.86)||>0.1|| 0.19 (0.05–0.67) || 0.01 ||2.08 (0.71–6.08)||>0.1|
|Log duration|| 3.51 (1.82–6.80) || <0.0005c ||1.30 (0.84–2.01)||>0.1||1.20 (0.68–2.12)||>0.1|| 2.20 (1.08–4.49) || <0.05 ||0.90 (0.55–1.47)||>0.1|| 9.35 (3.30–26.53) || <0.0001c |
|Seizure type (complex partial vs. primary generalized)|| 3.45 (0.86–13.87) || <0.1 ||0.77 (0.23–2.58)||>0.1||2.79||>0.1b||0.56 (0.14–2.30)||>0.1||0.94 (0.26–3.39)||>0.1||1.39 (0.39–4.94)||>0.1|
|Seizure type (complex partial vs. symptomatic generalized)|| 0.15 (0.02–1.28) || <0.1 || 9.44 || <0.05b || 0.1 (0.02–0.49) || 0.005 ||0.25 (0.04–1.48)||>0.1||3.0 (0.49–18.36)||>0.1||0.29 (0.05–1.60)||>0.1|
|Seizure type (primary generalized vs. symptomatic generalized)|| 0.04 (0.004–0.05) || 0.01 || 10.28 || <0.05b || 0.05 || 0.005b ||044 (0.06–3.24)||>0.1||3.20 (0.42–24.41)||>0.1||0.21 (0.03–1.47)||>0.1|
|Sleep state (awake vs. sleep)||1.23 (0.51–2.99)||>0.1||0.60 (0.24–1.48)||>0.1||0.78 (0.24–2.58)||>0.1||1.14 (0.35–3.76)||>0.1||0.63 (0.22–1.82)||>0.1||0.79 (0.29–2.13)||>0.1|
|Desaturation (yes vs. no)|| 55.00 (9.07–333.43) || <0.0001c || 0.31 (0.09–1.11) || <0.1 || 7.67 (0.85–69.54) || <0.1 || 33.38 (3.97–280.81) || 0.001c ||0.49 (0.17–1.39)||>0.1|| || |
|Ictal apnea (yes vs. no)|| || || || || || || 13.85 (1.56–122.58) || <0.05 ||1.08 (0.32–3.64)||>0.1|| 55.00 (9.07–333.43) || <0.0001c |
|Ictal bradycardia (yes vs. no)|| 13.85 (1.56–122.58) || <0.05 ||0.50 (0.09–2.77)||>0.1||2.50 (0.36–17.60)||>0.1|| || || || || 33.38 (3.97–280.81) || 0.001c |
|Postictal bradycardia (yes vs. no)|| || || || || || || 7.99 (1.28–50.02) || <0.05 || || ||2.55 (0.43–15.20)||>0.1|
|Ictal tachycardia (yes vs. no)||1.08 (0.32–3.64)||>0.1||1.97 (0.50–7.82)||>0.1||0.44 (0.07–2.92)||>0.1|| || || || ||0.49 (0.17–1.39)||>0.1|
|Drugs|| 1.80 (1.19–2.74) || <0.01 || 0.56 (0.37–0.84) || <0.01 || 1.93 (0.96–3.90) || <0.1 || 1.90 (1.01–3.55) || <0.05 || 0.56 (0.35–0.90) || 0.01 || 2.34 (1.39–3.94) || 0.001c |
Twenty-six patients had 101 epileptic seizures (average 3.9, range 1–10, median 3). Eleven patients were girls. Average age was 10.7 years (range 2–20). One patient took no AEDs, five took one, and all others took two or more AEDs. Fifteen patients had MRI lesions, and eight did not. Three patients had no MRI data. Patient and seizure characteristics are summarized in Table S1.
Mean EEG-seizure duration was 86.7 s (range 2–1271). Forty-seven seizures arose from frontal lobe, 29 from temporal lobe, and 25 from other epileptogenic regions. Seventy-two seizures were complex partial, 20 primary generalized, and nine symptomatic generalized. Forty-three seizures occurred from sleep. Evolution into generalized tonic–clonic seizures was seen in 56 seizures. Eighteen generalized tonic–clonic seizures occurred in the sleep state, and 38 occurred in the awake state. Thirty-two complex-partial seizures, 10 primary generalized seizures, and one symptomatic-generalized seizure occurred in sleep.
Respiratory changes during seizures
Seventy-eight seizures provided good RIP data. Thirty-nine seizures were associated with ictal central apnea, 34 with ictal tachypnea, and 13 with ictal bradypnea. No patients had ictal obstructive sleep apnea.
Average age of patients with ictal apnea was 9.92 years, compared to 12.2 years in patients without ictal apnea. Odds of ictal apnea decreased by 0.85 for every unit increase in age (OR = 0.85, p = 0.01).
Seventy-two percent of temporal lobe seizures compared to 24% of frontal lobe seizures had ictal apnea. Odds of ictal apnea increased by 8.04 in temporal lobe seizures (OR = 8.04, p = 0.0005).
Eighty-eight percent with symptomatic generalized seizures had ictal apnea, compared to 23% with primary generalized and 51% with complex partial seizures. Odds of ictal apnea decreased by 0.04 in primary generalized compared to symptomatic generalized seizures (OR = 0.04, p = 0.01). In complex partial compared to primary generalized and symptomatic generalized seizures, the odds of ictal apnea were of borderline significance (OR = 3.45, p < 0.1) and (OR = 0.15, p < 0.1).
Seizure duration in patients with ictal apnea was 134 s, compared to 55 s in patients without ictal apnea. Odds of ictal apnea increased by 3.51 for every unit increase in log seizure duration (OR = 3.51, p < 0.0005).
Ninety-two percent seizures with desaturation had ictal apnea, compared to 17% seizures without desaturation. Odds of ictal apnea increased by 55 if patient had desaturation (OR = 55, p < 0.0001). Ninety percent seizures with bradycardia compared to 39% without bradycardia had ictal apnea. Odds of ictal apnea increased by 13.85 if patient had ictal bradycardia (OR = 13.85, p < 0.05).
Patients with ictal apnea took on an average 3.2 AEDs compared to 2.3 AEDs without ictal apnea. Odds of ictal apnea increased by 1.80 for every additional AED (OR = 1.80, p < 0.01).
Ictal apnea had a borderline association with gender: 33% girls compared to 56% boys had ictal apnea (OR = 0.39, p < 0.1) and with seizure lateralization: 71% left-sided compared to 46% right-sided seizures had ictal apnea (OR = 2.91, p < 0.1).
Thus, chances of ictal apnea increased if the patient was younger, had temporal lobe, longer duration, symptomatic generalized seizure, had desaturation, ictal bradycardia, and took more AEDs.
Twenty-nine percent left-sided compared to 4% right-sided seizures had ictal bradypnea. Odds of ictal bradypnea increased by 10 in left-sided seizures (OR = 10, p < 0.05). Sixty-three percent symptomatic generalized seizures had ictal bradypnea compared to 14% complex partial and 0% primary generalized seizures. Odds of ictal bradypnea decreased by 0.1 in complex partial compared to symptomatic generalized seizures (OR = 0.1, p = 0.005) and by 0.05 in primary generalized as compared to symptomatic generalized seizures (OR = 0.05, p = 0.005).
There were several borderline associations: 24% temporal lobe seizures compared to 6% frontal lobe seizures had ictal bradypnea (OR = 4.89, p < 0.1); 22% seizures with brain MRI lesions compared to 5% without lesions had ictal bradypnea (OR = 6.0, p < 0.1); 25% seizures with desaturation compared to 4% without desaturation had ictal bradypnea (OR = 7.67, p < 0.1), patients with ictal bradypnea took on an average 3.4 AEDs compared to 2.6 AEDs in patients without ictal bradypnea (OR = 1.93, p < 0.1).
Thus, chances of ictal bradypnea increased in left-sided, symptomatic generalized seizures.
Average age of patients with ictal tachypnea was 12.41 years, compared to 10 years in patients without ictal tachypnea. Odds of ictal tachypnea increased by 1.19 for every unit increase in age (OR = 1.19, p = 0.01). Sixty-two percent seizures in girls compared to 37% in boys had ictal tachypnea. Odds of ictal tachypnea increased by 2.79 in girls (OR = 2.79, p = 0.05).
Sixty-four percent frontal lobe seizures had ictal tachypnea compared to 28% temporal lobe seizures and 30% seizures originating elsewhere (i.e., not temporal/frontal). Odds of ictal tachypnea decreased by 0.22 in temporal compared to frontal lobe seizures (OR = 0.22, p < 0.01) and increased by 4.08 in frontal lobe compared to seizures originating elsewhere (OR = 4.08, p < 0.05).
Fifty-four percent primary generalized seizures had ictal tachypnea, compared to 47% complex partial and 0% symptomatic generalized seizures. Odds of ictal tachypnea increased by 9.4 in complex partial compared to symptomatic generalized seizures (OR = 9.4, p < 0.05), and by 10.3 in primary generalized compared to symptomatic generalized seizures (OR = 10.3, p < 0.05).
Sixty-nine percent right-sided compared to 21% left-sided seizures had ictal tachypnea. Odds of ictal tachypnea decreased by 0.12 in left-sided seizures (OR = 0.12, p < 0.001). Thirty-seven percent seizures with brain MRI lesions compared to 64% nonlesional seizures had ictal tachypnea. Odds of ictal tachypnea decreased by 0.34 with MRI lesions (OR = 0.34, p < 0.05).
There was a borderline-inverse association between ictal tachypnea and desaturation: 21% (5/24) seizures with desaturation compared to 46% (11/24) seizures without desaturation (OR = 0.31, p < 0.1) had ictal tachypnea.
Patients with ictal tachypnea took on an average 2.3 AEDs compared to 3.1 AEDs in patients without ictal tachypnea. Odds of ictal tachypnea decreased by 0.56 for every additional AED (OR = 0.56, p = 0.005).
Therefore, chances of ictal tachypnea increased if the patient was older, was a girl, in frontal lobe, right-sided seizures, and decreased in symptomatic generalized seizures, MRI lesions, and took more AEDs.
Desaturation and heart-rate during seizures
Sixty-three seizures had good pulse-oximetry and ECG signal. Thirty-one seizures had desaturation, 34 had ictal tachycardia, 15 had ictal bradycardia, and six had postictal bradycardia.
Eleven percent seizures in girls compared to 37% in boys had ictal bradycardia. Odds of ictal bradycardia decreased by 0.21 in girls (OR = 0.21, p < 0.05). Average seizure duration in patients with ictal bradycardia was 83 s, compared to 61 s in patients without bradycardia. Odds of ictal bradycardia increased by 2.20 for every unit increase in log seizure duration (OR = 2.20, p < 0.05).
Thirteen percent frontal lobe seizures compared to 50% originating elsewhere had ictal bradycardia. Odds of ictal bradycardia decreased by 0.15 in frontal lobe seizures (OR = 0.15, p = 0.01). Fifty-two percent seizures with desaturation compared to 3% without desaturation had ictal bradycardia. Odds of ictal bradycardia increased by 33.4 with desaturation (OR = 33.38, p = 0.001). Patients with ictal bradycardia took on an average 3.5 AEDs compared to 2.6 AEDs in patients without bradycardia. Odds of ictal bradycardia increased by 1.90 for each additional AED (OR 1.90, p < 0.05).
Ictal bradycardia had a borderline association with age: average-age of patients with ictal bradycardia was 8.5 years as compared to 10.7 years without ictal bradycardia (OR = 0.89, p < 0.1).
Association of seizure lateralization with ictal bradycardia was not significant (p > 0.1).
Therefore, chances of ictal bradycardia increased if patient was a boy, had longer seizure duration, desaturation, and more AEDs, and decreased in frontal lobe seizures.
Eighty-one percent of seizures in girls compared to 38% in boys had ictal tachycardia. The odds of ictal tachycardia increased by 1.99 in girls (OR = 1.99, p = 0.001).
Forty-four percent of seizures with ictal tachycardia had brain MRI lesions compared to 81% seizures without lesions. Odds of ictal tachypnea decreased by 0.19 with brain lesions (OR = 0.19, p = 0.01). Patients with ictal tachycardia took on an average 2.5 AEDs compared to 3.4 AEDs in patients without ictal tachycardia. Odds of ictal tachycardia decreased by 0.56 for every additional AED (OR = 0.56, p < 0.05).
Ictal tachycardia had a borderline association with age: average age of patients with ictal tachycardia was 11 years compared to 9 years in patients without ictal tachycardia (OR = 1.12, p < 0.1). Association of seizure lateralization with ictal tachycardia was not significant (p > 0.1).
Therefore, chances of ictal tachycardia increased if patient was a girl, and decreased with brain MRI lesions and more AEDs.
Sixty-seven percent seizures with postictal bradycardia compared to 20% without postictal bradycardia had ictal bradycardia. Odds of postictal bradycardia increased by 7.92 if patients had ictal bradycardia (OR = 7.92, p < 0.05). Therefore, chances of postictal bradycardia increased if patient had ictal bradycardia too.
Patients who desaturated had average seizure duration of 108 s, compared to 37 s in patients who did not. Odds of desaturation increased by 2.58 for every unit increase in log seizure duration (OR = 2.58, p < 0.0001). Patients with desaturation took on an average 3.5 AEDs compared to 2.3 AEDs in patients without desaturation. Odds of desaturation increased by 2.34 for every additional AED (OR = 2.34, p = 0.001). There was strong association between desaturation and ictal apnea (OR = 55, p < 0.0001), between desaturation and ictal bradycardia (OR = 33.38, p = 0.001).
Desaturation had a borderline association with ictal tachypnea: 21% seizures with desaturation compared to 46% without desaturation had ictal tachypnea (OR = 0.60, p < 0.1).
It was often difficult to assess oxygen saturation during a seizure because of significant movement artifacts impairing accurate recording of the pulse oximeter during the seizure. The results here reflect postictal desaturation.
Therefore, chances of desaturation increased with longer seizure duration, ictal apnea, ictal bradycardia, took more AEDs.
Seizure generalization and sleep/awake state
Sixty-eight percent of frontal lobe seizures secondarily generalized in awake state, whereas 32% of frontal lobe seizures did not secondarily generalize in awake state. In contrast, only 32% of frontal lobe seizures secondarily generalize in sleep. Odds of seizure generalization were increased by 4.62 in awake state for frontal lobe seizures (OR 4.62, 95% confidence interval [CI] 1.26–16.92, p = 0.02). However, no such relationship was observed between sleep/awake state and seizure generalization in temporal or occipital lobe seizures (p > 0.1).
Therefore, chances of secondary generalized tonic–clonic seizures increased in awake state for frontal lobe seizures.
Our study found several associations between seizure characteristics and cardiopulmonary disturbances during seizures (Table 1). Overall, most prominent ictal distress with decrease of respiration or heart rate (apnea, bradypnea, bradycardia, and desaturation) was seen in younger boys with symptomatic generalized, or left temporal and longer-duration seizures, who were taking multiple AEDs. Ictal tachypnea and tachycardia occurred more in children with older age and nonlesional brain MRIs, whereas ictal bradypnea occurred at younger age with lesional brain MRIs. Ictal apnea, bradycardia, and desaturation were positively associated, and ictal tachypnea and tachycardia inversely associated with number of AEDs. Postictal bradycardia was associated with ictal bradycardia. Patients with ictal apnea, ictal bradypnea, and ictal bradycardia were more likely, and patients with ictal tachypnea less likely to have accompanying desaturation.
Part of our findings (apnea, bradypnea, bradycardia, and desaturation seen in younger boys with symptomatic-generalized or left temporal and longer duration seizures and ictal tachypnea and tachycardia seen in older children) are similar to a 2006 retrospective study that reported that apnea was more common in younger children (Fogarasi et al., 2006). We hypothesize that the reason behind this might be the transient tachypnea that occurs at the beginning of the seizure, which leads to CO2 washout thus leading to a central apnea. Children (2–3 torr) and infants (1–2 torr) have lower apneic threshold and higher arousal threshold than adults (3–4 torr)—thus the same degree of tachypnea might be more likely to cause a central apnea, with subsequent bradycardia and desaturation in younger children (Busby et al., 1994; Rowley et al., 2006).
An important finding of our study was that symptomatic-generalized, longer duration, temporal lobe seizures occurring in younger children were at an increased risk of apnea and desaturation. This finding goes in hand with previously published studies where the risk of SUDEP was highest in symptomatic generalized, longer duration, temporal lobe seizures (Sillanpaa & Shinnar, 2010). Our findings need to be prospectively tracked for occurrence of SUDEP in this population.
Our study also found that evolution to generalized tonic–clonic seizures occurred in awake state with frontal lobe seizures, but no such relationship was found in temporal or occipital seizures. The findings of frontal lobe generalizing in awake state are in part similar to one previous study (Bazil & Walczak, 1997), which found that temporal lobe seizures were more likely to generalize in sleep while frontal lobe seizures do not. Possible reasons that frontal lobe seizures do not generalize during sleep, even though they occur more often in sleep, may be related to those structures necessary for early propagation of frontal lobe seizures not being easily activated by sleep (Quesney et al., 1995). The lack of sleep inducing secondary generalization that was not observed in our series in temporal lobe seizures may be related to the inpatient study setting and disturbed sleep/wake cycles due to hospital noise/lights (Hinds et al., 2007). Secondary generalization in temporal lobe seizures in sleep puts these patients at higher risk for SUDEP (Sillanpaa & Shinnar, 2010).
Epileptic cardiopulmonary abnormalities published previously
Cardiac abnormalities in our study add to previous findings. Tachycardia has been reported in both children and adults with three or more AEDs, generalized seizures, normal brain MRI; ictal bradycardia with seizure clustering and higher seizure frequency, and ictal desaturation with normal brain MRI and longer seizure duration (Hewertson et al., 1996; Moseley et al., 2011). In adults, preictal tachycardia has been shown to be associated with secondary generalization, ictal and postictal tachycardia, postictal heart rate variability (HRV), and abnormal QTc-shortening with secondary generalization (Nilsen et al., 2010; Surges et al., 2010b). Although our study showed a correlation of desaturation accompanying ictal apnea, prior research has shown that patients may not have apnea accompanying a seizure, and yet continue to have desaturation, which may be due to ictal hypoventilation or intrapulmonary shunting (Seyal et al., 2010). Our study presents data on respiratory abnormalities (relationship of apnea, bradypnea, tachypnea, desaturation to age, seizure type, and laterality in pediatric population) some of which have not been previously reported. In children, hypoxemia, desaturation, apnea, and sinus tachycardia have been shown to be present in several seizures (Hewertson et al., 1996), and seizures with apparent life-threatening events were associated with hypoxemia and apnea (Hewertson et al., 1994). These studies were, however, descriptive and used nasal airflow thermistors and volume expansion capsules, unlike RIP used in our study (which is far more sensitive). Focal seizures have been shown to be associated with tachypnea, apnea, respiratory pauses, and hypoxemia (O'Regan & Brown, 2005). However, that study used chest electrodes for respiratory monitoring and did not report associations with patient characteristics. Ictal hypoxemia was associated with seizure generalization and AEDs (Moseley et al., 2010), but no respiratory abnormalities were recorded in that study. In adults, desaturation was reported to be associated with seizure-localization, lateralization, and gender (Bateman et al., 2008). Table 2 summarizes studies looking into cardiopulmonary abnormalities in epileptic seizures.
Table 2. Summary of prior research looking at cardiopulmonary abnormalities in the setting of seizures
|1987||Southall et al. (1987)||Apnea and desaturation and complex partial seizures||Case report||Pediatric||Apnea and severe hypoxemia in a patient with complex partial seizures|
|1994||Hewertson et al. (1994)||Seizure-induced hypoxemia and apparent life-threatening events (ALTEs)||Prospective||Pediatric||Apneic events associated with ALTEs, and concluded that seizures with ALTEs can be associated with “potentially life-threatening episodes of severe hypoxemia and apnea, despite normal EEG between events”|
|1996||Hewertson et al. (1996)||Hypoxemia, apnea, ECG changes and epileptic seizures||Prospective||Pediatric||Hypoxemia/desaturation, apnea and sinus tachycardia present in several seizures, but no attempt made to find significant associations|
|2005||O'Regan and Brown (2005)||Cardiac and respiratory function abnormalities with seizures||Prospective||Pediatric||Focal seizures associated with tachypnea, apnea, respiratory pauses, and hypoxemia. No attempt made to find significant associations|
|2010||Moseley et al. (2010)||ECG and oximetry changes and seizures||Retrospective||Pediatric||Desaturation associated with seizure generalization and AEDs|
|2011||Moseley et al. (2011)||ECG and oximetry changes and seizures||Prospective||Adult and pediatric||Ictal tachycardia associated with AEDs, seizure generalization; ictal bradycardia associated with seizure clustering and high frequency; desaturation associated with normal brain MRI, longer seizure duration and with ictal tachycardia|
|2008||Bateman et al. (2008)||Ictal hypoxemia in localization-related epilepsy||Prospective||Adult||Desaturation associated with seizure localization, seizure lateralization, seizure duration, contralateral spread, gender|
|2009||Tezer et al. (2009)||Respiratory abnormalities and seizures||Case-report||Adult||Apneas associated with EEG seizure pattern|
|2009||Surges et al. (2009)||Interictal heart rate variability with SUDEP||Prospective||Adult||No association of heart rate variability with SUDEP|
|2010||Nilsen et al. (2010)||ECG changes and secondary generalization||Prospective||Adult||Preictal tachycardia associated with secondary generalization|
|2010||Seyal et al. (2010)||Respiratory changes with seizures in localization-related epilepsy||Prospective||Adult||Severe and prolonged increase in ETCO2 associated with seizures|
|2010||Surges et al. (2010a)||Abnormal cardiac repolarization and other ECG predictors with SUDEP||Case–control study||Adult||No association of pathologic cardiac repolarization or other ECG changes with SUDEP|
|2010||Surges et al. (2010b)||ECG abnormalities with generalized seizures||Prospective||Adult||Ictal and postictal tachycardia, and postictal heart rate variability, and abnormal QTc shortening associated with secondary generalization|
|2011||Lanz et al. (2011)||Cardiac asystole with seizures||Prospective||Adult||Temporal lobe epilepsy associated with higher risk for asystole|
|2012||Yildiz et al. (2012)||Sleep-awake state and seizures||Prospective||Adult||Awakening with temporal lobe and frontal lobe epilepsy|
|2012||Seyal et al. (2012)||Respiratory abnormalities and postictal EEG suppression||Prospective||Adult||Postictal EEG suppression not associated with postictal central apnea but likely related to severity of seizure related pulmonary dysfunction|
Other risk factors for SUDEP
Several studies reported right/left laterality for ictal tachycardia and bradycardia, but we and others could not detect any association (Sevcencu & Struijk, 2010). Studies have also looked into whether postictal generalized EEG suppression (PGES) is a SUDEP risk factor (Lhatoo et al., 2010). Our study had only one patient with PGES, so statistical analysis was not possible.
Brainstem serotonin-receptor abnormalities may lead to impaired postictal arousal, thereby leading to postictal hypoventilation and SUDEP in some cases. This relationship of SUDEP and serotonin receptors is similar to that of 5-HT receptors and sudden infant death syndrome (SIDS) (Duncan et al., 2010). It is unclear whether patients with apneas and desaturation in our study would have 5-HT abnormalities, if tissue analysis were available. Except for a relationship between seizure generalization and awake-state in frontal lobe seizures, our study did not find any other association of cardiorespiratory abnormalities in relation to the sleep-wake cycle. This may again be due to the in-patient study-setting and disturbed sleep/wake cycles due to hospital noise/lights (Hinds et al., 2007).
Our findings need to be interpreted in the setting of data acquisition. Although we recorded 101 seizures, they came from a relatively small study population. Because of a relatively small sample size, analysis of patients rather than seizures (as done by us in this study) or age-adjusted analysis for tachycardia/bradycardia was not feasible. End tidal CO2 levels were not specifically measured, so the role of hyperventilation and hypoventilation was determined from the RIP tracings. Serum AED levels were not measured, hence significance of subtherapeutic AED levels and use of rescue medications also cannot be assessed. Pulse oximetry data was missing in some patients because the sensor lost contact.
Our study objectively and prospectively recorded cardiopulmonary disturbances and associated risk factors in pediatric epileptic seizures using video-EEG, pulse oximetry, and RIP. We report that age, gender, seizure duration, localization, lateralization, and AEDs were associated with risk of cardiopulmonary abnormalities, with adequate data collected in more than two thirds of the patients.
This prospective study provides preliminary data and a starting-point for a more extensive study with a larger sample size, using multiple site pulse oximetry (legs, hands, and so on) and analysis of other variables such as QTc changes and heart rate variability. Analysis of seizure characteristics in relation to dynamic patterns of respiratory and cardiac signals may provide additional insight into the pathophysiology of SUDEP (Rejdak et al., 2011).
This study was funded with the support of a grant by the Harvard Catalyst (Grant no. UL1 RR 025758) in 2010, awarded to Drs. Kothare & Pavlova. Dr. Kothare interprets video-EEGs and routine EEG in the Division of Clinical Neurophysiology at Children's Hospital Boston; and has received research support from Eisai Inc. and the NIH. Dr. Loddenkemper receives funding from the Epilepsy Foundation of America, and serves on the Laboratory Accreditation Board for Long Term (Epilepsy and ICU) Monitoring (ABRET). Dr. Pavlova is also the co-PI of the same Harvard catalyst grant that funded the adult component to the study.
None of the other authors have any conflict of interest to report. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.