How to assess prognosis after cardiac arrest and therapeutic hypothermia

Abstract

The prognosis of patients who are admitted in a comatose state following successful resuscitation after cardiac arrest remains uncertain. Although the introduction of therapeutic hypothermia (TH) and improvements in post-resuscitation care have significantly increased the number of patients who are discharged home with minimal brain damage, short-term assessment of neurological outcome remains a challenge. The need for early and accurate prognostic predictors is crucial, especially since sedation and TH may alter the neurological examination and delay the recovery of motor response for several days. The development of additional tools, including electrophysiological examinations (electroencephalography and somatosensory evoked potentials), neuroimaging and chemical biomarkers, may help to evaluate the extent of brain injury in these patients. Given the extensive literature existing on this topic and the confounding effects of TH on the strength of these tools in outcome prognostication after cardiac arrest, the aim of this narrative review is to provide a practical approach to post-anoxic brain injury when TH is used. We also discuss when and how these tools could be combined with the neurological examination in a multimodal approach to improve outcome prediction in this population.

Figure 1

Multimodal approach to assess prognosis in comatose survivors after cardiac arrest treated with hypothermia. The multimodal approach to assess prognosis in comatose survivors after cardiac arrest treated with hypothermia should ideally include neurological examination combined with the use of electroencephalography (EEG) and somatosensory evoked potentials (SSEPs), the measurement of biomarkers (neuron-specific enolase (NSE) and S-100β protein) and magnetic resonance imaging (MRI).

Figure 2

Electroencephalogram findings from resuscitated patients after cardiac arrest. (A) Electroencephalogram (EEG) recorded during therapeutic hypothermia, showing an example of continuous EEG: the patient had complete recovery of consciousness. (B) Burst-suppression findings during normothermia; the patient had concomitant myoclonus and bilateral absent N20 cortical responses to somatosensory evoked potentials, and eventually died. (C) Generalized periodic epileptiform discharges at 36 hours after hospital admission; we decided to withdraw care on day 5 because of persistent coma with posturing and absent pupillary reflexes.

Figure 3

Somatosensory evoked potentials. (A) In comatose survivors after cardiac arrest, somatosensory evoked potentials are elicited by transcutaneous electrical stimulation applied to the median nerve and then recorded at Erb’s point (N9), the cervical medulla (N13) and the controlateral cortex (N20). (B) Example of present N20 cortical response (C3’) in two comatose patients after cardiac arrest. (C) Example of absent N20 cortical response (C3’) in two comatose patients after cardiac arrest.

Figure 4

Diffusion-weighted magnetic resonance imaging. Diffusion-weighted magnetic resonance imaging scan of a 68-year-old man who suffered a ventricular fibrillation cardiac arrest with prolonged resuscitation. Diffuse cortical hyperintensities are observed, consistent with severe global anoxic injury.

Figure 5

Multimodal prognostication of coma after cardiac arrest and therapeutic hypothermia. Summary of the suggested timing after cardiac arrest of all available tools that are used to predict poor outcome or neurological recovery from coma. This algorithm suggests that poor prognosis or neurological recovery should be considered when specific findings are present but this does not necessarily mean that withdrawal of care should be initiated, because this approach has not been validated. Dashed lines, lack of strong evidence to support the suggestion. High biomarker levels did not relate to a specific cutoff value because of several limitations affecting their measurements and accuracy (see text). BS, burst suppression; BSR, brainstem reflexes; EEG, electroencephalogram; GPED, generalized periodic epileptiform discharge; M_1–2, absent motor response or posturing; MRI, magnetic resonance imaging; N20, cortical responses to somatosensory evoked potentials; NSE, neuron-specific enolase; SM, status myoclonus; SSEP, somatosensory evoked potential; uNCSE, unreactive nonconvulsive seizures; uSE, unreactive status epilepticus.