Seizure-related cardiac repolarization abnormalities are associated with ictal hypoxemia

Abstract

Cardiac arrhythmias and respiratory disturbances have been proposed as likely causes for sudden unexpected death in epilepsy. Oxygen desaturation occurs in one-third of patients with localization-related epilepsy (LRE) undergoing inpatient video-electroencephalography (EEG) telemetry (VET) as part of their presurgical workup. Ictal-related oxygen desaturation is accompanied by hypercapnia. Both abnormal lengthening and shortening of the corrected QT interval (QTc) on electrocardiography (ECG) have been reported with seizures. QTc abnormalities are associated with increased risk of sudden cardiac death. We hypothesized that there may be an association between ictal hypoxemia and cardiac repolarization abnormalities. VET data from patients with refractory LRE were analyzed. Consecutive patients having at least one seizure with accompanying oxygen desaturation below 90% and artifact-free ECG data were selected. ECG during the 1 min prior to seizure onset (PRE) and during the ictal/postictal period with accompanying oxygen desaturation below 90% (DESAT) was analyzed. Consecutive QT and RR intervals were measured. In the same patients, DESAT seizures were compared with seizures without accompanying oxygen desaturation below 90% (NODESAT). For NODESAT seizures, QT and RR intervals for 2 min after seizure onset were measured. Thirty-seven DESAT seizures were analyzed in 17 patients with localization-related epilepsy. A total of 2,448 QT and RR intervals were analyzed during PRE. During DESAT, 1,554 QT and RR intervals were analyzed. Twelve of the 17 patients had at least one NODESAT seizure. A total of 19 NODESAT seizures were analyzed, including 1,558 QT and RR intervals during PRE and 3,408 QT and RR intervals during NODESAT. The odds ratio for an abnormally prolonged (>457 ms) QTcH (Hodges correction method) during DESAT relative to PRE was 10.64 (p < 0.0001). The odds ratio for an abnormally shortened (<372 ms) QTcH during DESAT relative to PRE was 1.65 (p < 0.0001). Seizure-related shortening and prolongation of QTc during DESAT were also observed when Fridericia correction of the QT was applied. During DESAT seizures, the mean range of QT values (QTr) (61.14 ms) was significantly different from that during PRE (44.43 ms) (p = 0.01). There was a significant association between DESAT QTr and oxygen saturation nadir (p = 0.025) and between DESAT QTr and duration of oxygen desaturation (p < 0.0001). Both QTcH prolongation and shortening also occurred with NODESAT seizures. A seizure-associated prolonged QTcH was more likely during DESAT than NODESAT, with an odds ratio of 4.30 (p < 0.0001). A seizure-associated shortened QTcH was more likely during DESAT than NODESAT with an odds ratio of 2.13 (p < 0.0001). We have shown that the likelihood of abnormal QTcH prolongation is increased 4.3-fold with seizures that are associated with oxygen desaturation when compared with seizures that are not accompanied with oxygen desaturation. The likelihood of abnormally shortened QTcH increases with seizures that are accompanied by oxygen desaturation with an odds ratio of 2.13 compared with that in seizures without desaturations. There is a significant association between the depth and duration of oxygen desaturation and QTr increase. These findings may be related to the pathophysiology of SUDEP.