Does Therapeutic Hypothermia Affect Predictive Value of Somatosensory Evoked Potentials after CPR?

Abstract & Commentary

By Elayna Rubens, MD. Dr. Rubens is Assistant Professor of Neurology at Weill Cornell Medical College. Dr. Rubens reports no financial relationships relevant to this field of study.

Synopsis: Bilateral absence of N20 responses in the setting of therapeutic hypothermia does not preclude neurologic recovery in comatose survivors of cardiac arrest.

Source: Leithner C, et al. Does hypothermia influence the predictive value of bilateral absent N20 after cardiac arrest? Neurology 2010;74:965-969.

Predicting neurologic recovery in comatose survivors of cardiac arrest is a challenge faced routinely by neurologists in the inpatient setting. To date, therapeutic hypothermia (TH), where the patient's core temperature is maintained at 32-34 °C for 12-24 hours following resuscitation, is the only treatment modality that has been shown to improve neurologic outcome in this population. Because the currently accepted prognostic indicators of outcome after cardiac arrest were validated (and subsequently outlined in the 2006 AAN practice parameter) prior to the widespread use of TH, the accuracy of these indicators in patients treated with hypothermia is not yet clear. Of the established predictors, bilateral absence of the cortical N20 response in median nerve somatosensory evoked potential (SSEP) testing is considered to be the most reliable early indicator of a poor neurologic outcome and, in one small study, appeared equally useful in the setting of hypothermia. Prompted by a case of a patient who had a favorable neurologic recovery after TH despite an initially absent N20 bilaterally, Leithner and colleagues set out to examine the predictive value of bilateral absent N20 in the setting of therapeutic hypothermia.

In this study, the authors retrospectively analyzed the records of 185 consecutive patients treated with hypothermia following cardiac arrest. Of 185 patients, 112 had SSEP testing. Testing was performed more than 24 hours after resuscitation using a standard technique. The N20 responses were categorized as: absent, pathologic (prolonged latency and/or reduced amplitude of the cortical response), or normal. Using a clinical database, baseline and follow up information regarding neurologic outcome was obtained. Outcomes were determined at the time of the ICU discharge and assessed using the Pittsburgh cerebral performance category (CPC).

The findings revealed that N20 was absent in 36 (32%) patients, pathologic in 22 (20%) patients, and normal in 54 (48%) patients. Of the 36 patients with bilaterally absent N20 responses, 35 (97%) had poor outcome (CPC 4 or 5). One patient with initially absent N20 response three days after cardiac arrest (after normothermia) had an excellent outcome (CPC1) with subsequent recovery of the N20 response at 18 month follow up testing. In this patient, the peripheral (N9) and cervical (N13) responses were significantly delayed on both initial and follow up testing. Though the peripheral response delay alone is unlikely to explain the initial absence of the cortical response, it does indicate a coexisting peripheral abnormality, which the authors suggest was due to alcoholic polyneuropathy and reduced extremity temperature at the time of the initial testing. The authors also identified another case in which the cortical response amplitudes were severely diminished such that the N20 was nearly absent three days after resuscitation (during normothermia). This patient also had a favorable outcome (CPC1) and recovery of normal N20 response amplitude nine days after cardiac arrest.

The authors conclude that their results reaffirm the high negative predictive value of bilaterally absent N20 responses in comatose survivors of cardiac arrest in the setting of TH. However, the identification of two cases of absent, or nearly absent, N20 responses after cardiac arrest treated with hypothermia with subsequent recovery of both neurologic functioning and cortical somatosensory evoked responses suggests that the certainty of this prediction may be diminished in patients treated with hypothermia. The authors propose that hypothermia may allow for a delayed functional recovery of somatosensory evoked responses well beyond the established one to three day period post resuscitation during which the test is typically performed.


This study highlights the importance of rigorous re-evaluation of the standard indicators of neurologic prognosis following cardiopulmonary resuscitation in patients undergoing therapeutic hypothermia. Inherent in hypothermia treatment is the use of sedative and paralytic agents which hinder clinical assessment and therefore increase dependence on neurophysiologic parameters, particularly SSEP. Though this is a small study with a bias towards identification of false positive results, it suggests that absent N20 may not be as reliable early in the course of TH and that meaningful recovery of cortical responses may occur later than previously expected in these patients. Further prospective studies designed to establish false positive rates and appropriate timing of evoked potential testing are needed to guide decision making in patients treated with induced hypothermia. Until then, bilateral absence of the N20 response in these patients should be interpreted cautiously.