By Vibhu Sharma, MD, MS

Attending Physician, Division of Pulmonary and Critical Care Medicine, John H. Stroger Hospital of Cook County; Assistant Professor of Medicine, Rush University Medical Center, Chicago

Dr. Sharma discloses no financial relationships relevant to this field of study.

SYNOPSIS: In this retrospective single-center study, late awakening after cardiac arrest was associated with a higher rate of unfavorable outcomes immediately after awakening and at three months.

SOURCE: Rey A, et al. Late awakening in survivors of postanoxic coma: Early neurophysiologic predictors and association with ICU and long-term neurologic recovery. Crit Care Med 2019;47:85-92.

Rey et al conducted a retrospective study of consecutive post-cardiac arrest (CA) comatose adult patients. All were admitted to the ICU between 2009 and 2016 and underwent standard post-arrest care, including targeted temperature management (TTM) for 24 hours using surface cooling to 33° C (until December 2014) or 36° C (after December 2014). All patients were sedated for the first 24 hours with either a midazolam drip (first choice) or propofol; a fentanyl drip was administered for analgesia. Opiates and sedatives were discontinued after the initial period of TTM. All patients were mechanically ventilated; mean arterial pressure (MAP) was maintained at > 70 mmHg using vasopressors, if necessary. It is unclear what proportion of these patients experienced CA in the community and how many episodes occurred in the hospital.

Certified neurologists performed a daily standardized exam to assess neurologic recovery, including Glasgow Coma Scale motor response (GCS-M) to pain and brainstem reflexes (pupillary reaction to light, corneal reflexes, and oculocephalic reflexes). The absence of pupillary and corneal reflexes or myoclonus within 72 hours were considered poor prognostic markers. A standard 10-20 lead montage video-electroencephalography (EEG) was performed on day 1 (during TTM and sedation) and then on day 2 after cessation of TTM, off sedation. EEG patterns in response to painful stimuli also were assessed by certified neurophysiologists on days 1 and 2. Absent normothermic EEG and absent bilateral somatosensory evoked potentials (SSEPs) were considered poor prognostic signs. Irreversible cerebral damage within 72 hours was defined as occurring if at least two of the following signs were observed: absent brainstem reflexes, myoclonus, nonreactive EEG, or absent SSEP bilaterally. Serum neuron-specific enolase (NSE) levels were drawn on days 1 and 2 after cardiac arrest. NSE levels > 33 ng/mL at any time within seven days of CA predict persistent coma with a high degree of specificity.1 Overall, 228 of 402 patients in the cohort awoke; the rest never regained consciousness and died. One hundred and fifty of the 228 who awoke did so early. Late awakening occurred in 78 of the 228 who awoke, with a median time to awakening of five days (range, five to 23 days). Late awakening was associated with more absent motor responses to painful stimuli and absent brainstem reflexes by day 3 after CA and discontinuous EEG on days 1 and 2 after CA.

Regarding prediction of late awakening, the highest odds were associated with findings of day 2 discontinuous EEG (odds ratio [OR], 3.459), followed by day 3 absence of brainstem reflexes (OR, 3.352), day 3 absent motor response (OR, 3.35), and the use of midazolam alone as the sedating agent during TTM (OR, 1.71). Time to return of spontaneous circulation after CA and day 1 Sequential Organ Failure Assessment score also were significantly associated with late awakening but the association was less robust. TTM strategy (33°C vs. 36°C) did not influence outcomes with respect to awakening.

Despite early unfavorable neurophysiologic signs, 73% of patients with late awakening eventually exhibited good neurologic recovery defined as Cerebral Performance Category (CPC) 1 and 2 outcomes (“Normal life” and “Disabled but independent,” respectively). Nevertheless, in this cohort, late awakening was associated with a higher likelihood of CPC categories 3-5 (“Conscious but dependent on others,” “Coma or vegetative state,” and “Brain death,” respectively).


The results of this study do not apply to patients with catastrophic neurologic injury. Predicting who might awaken after a cardiac arrest can be frustrating because most post-CA patients look “the same.” The most important concern for families in this setting may be when and if their loved one will awaken. The relevant finding of interest from the perspective of clinicians and families in this study is that prognostication remains a moving target.

For example, when queried early in admission during a family meeting, it may be reasonable to offer prognostication based on a 48-hour timeline from the point sedation is discontinued. Based on this single-center study, the longer it takes to awaken, the more likely it is that delirium will occur and the longer the patient will be in the ICU. Those awakening late are more likely to be severely disabled or die within three months. When and how much neurologic recovery will occur cannot be predicted with certainty, but early awakening (responding to pain and following commands within 48 hours of sedation discontinuation) makes meaningful recovery more likely.

The single most important recommendation for clinicians is to avoid the use of midazolam for sedation in the CA patient. Clinical findings within 72 hours of CA that convey a poor outcome or nonsurvivability include absent pupillary and corneal reflexes, myoclonus, nonreactive EEG background, and absent bilateral SSEP. In scenarios with disparate neurophysiologic findings, NSE levels may complement prognostication, with higher levels predicting a worse outcome. Prediction of outcomes early after CA remains a daunting task. Caution is warranted in prognostication despite early adverse results on clinical and neurophysiological testing. Sedation using midazolam ought to be avoided in the immediate post-CA phase of ICU care and is the single most modifiable factor influencing time to awakening.


  1. Fogel W, et al. Serum neuron-specific enolase as early predictor of outcome after cardiac arrest. Crit Care Med 1997;25:1133-1138.