Prediction of poor outcome within the first 3 days of postanoxic coma

Objective: To determine the optimal timing of somatosensory evoked potential (SSEP) recordings and the additional value of clinical and biochemical variables for the prediction of poor outcome in patients who remain comatose after cardiopulmonary resuscitation (CPR). Methods: A prospective cohort study was conducted in 32 intensive care units including adult patients still unconscious 24 hours after CPR. Clinical, neurophysiologic, and biochemical variables were recorded 24, 48, and 72 hours after CPR and related to death or persisting unconsciousness after 1 month. Results: Of 407 included patients, 356 (87%) had a poor outcome. In 301 of 305 patients unconscious at 72 hours, at least one SSEP was recorded, and in 136 (45%), at least one recording showed bilateral absence of N20. All these patients had a poor outcome (95% CI of false positive rate 0 to 3%), irrespective of the timing of SSEP. In the same 305 patients, neuron-specific enolase (NSE) was determined at least once in 231, and all 138 (60%) with a value 33 g/L at any time had a poor outcome (95% CI of false positive rate 0 to 3%). The test results of SSEP and NSE overlapped only partially. The performance of all clinical tests was inferior to SSEP and NSE testing, with lower prevalences of abnormal test results and wider 95% CI of false positive rates. Conclusion: Poor outcome in postanoxic coma can be reliably predicted with somatosensory evoked potentials and neuron-specific enolase as early as 24 hours after cardiopulmonary resuscitation in a substantial number of patients. NEUROLOGY 2006;66:62–68 The prediction of poor outcome in patients who remain unconscious after cardiopulmonary resuscitation (CPR) has recently been addressed in a number of systematic reviews. A positive predictive value of 100% has been demonstrated for the absence of early cortical responses (N20) of the somatosensory evoked potentials (SSEPs) in the first week after CPR,1-3 and similar values were found for the absence of pupillary and corneal reflexes and of any motor response 3 days after the hypoxic–ischemic insult.1,4 The usefulness of biochemical markers of brain damage in serum or CSF remained uncertain, because of the small number of patients in most studies and methodologic flaws in some.5 With these reviews, a number of questions could not be answered. 1) What is the earliest time when SSEP results may be considered reliable? 2) Are the predictive values of separate variables additive? 3) Can biochemical markers of anoxic–ischemic brain damage contribute to the prediction of poor outcome? We designed our study (Prognosis in Postanoxic Coma) to address these questions. Methods. From January 2000 to May 2003, we performed a multicenter prospective cohort study to correlate early clinical, neurophysiologic, and biochemical findings with clinical outcome. To study prediction in regular clinical practice, we chose a pragmatic design, with various types of hospitals and without centralized assessment of the neurophysiologic tests. CSF studies were not feasible with this design. Patients admitted to the intensive care units of 32 Dutch hospitals (13 teaching hospitals and 19 nonteaching hospitals) were included. The study was approved by the ethical review boards of all participating hospitals. Consecutive patients with CPR for primary or secondary circulatory arrest, persisting coma 24 hours after CPR, age 18 years or older, and informed consent from a legal representative were included. We defined coma as no eye opening to external stimuli, motor response to pain flexion or worse, and no speech. Exclusion criteria were a life expectancy of no more than several months caused by pre-existent disease, death by brain criteria after 24 hours, and concomitant head injury. Timing of assessments. Baseline characteristics registered were gender, age, medical history, prearrest level of functioning, cause of the arrest, location of arrest, cardiac rhythm before CPR, time between arrest and initiation of CPR, and duration of CPR. All patients underwent standard assessments 24, 48, and 72 hours ( 4 hours) after CPR, including neurologic examination, median nerve SSEP, and blood sampling at all time points and EEG at 72 hours. For practical reasons, SSEP recording was not Additional material related to this article can be found on the Neurology Web site. Go to www.neurology.org and scroll down the Table of