Systematic review of early prediction of poor outcome in anoxic-ischaemic coma.

BACKGROUND: Studies to assess the prognostic value of early neurological and neurophysiological findings in patients with anoxic-ischaemic coma have not led to precise, generally accepted, prognostic rules. We did a systematic review of the relevant literature to assess whether such rules could be derived from the combined results of these studies.

METHODS: From Medline and Embase databases we selected studies concerning patients older than 10 years with anoxic-ischaemic coma in which findings from early neurological examination, electroencephalogram (EEG), or somatosensory evoked potentials (SSEP) were related to poor outcome--defined as death or survival in a vegetative state. We selected variables with a specificity of 100% for poor outcome in all studies, and expressed the overall prognostic accuracy of these variables as pooled positive-likelihood ratios and as 95% CIs of the pooled false-positive test rates.

FINDINGS: In 33 studies, 14 prognostic variables were studied, three of which had a specificity of 100%: absence of pupillary light reflexes on day 3 (pooled positive-likelihood ratio 10.5 [95% CI 2.1-52.4]; 95% CI pooled false-positive test rate 0-11.9%); absent motor response to pain on day 3 (16.8 [3.4-84.1]; 0-6.7%); and bilateral absence of early cortical SSEP within the first week (12.0 [5.3-27.6]; 0-2.0%). EEG recordings with an isoelectric or burst-suppression pattern had a specificity of 100% in five of six relevant studies (pooled positive-likelihood ratio 9.0 [2.5-33.1]; 95%CI pooled false-positive test rate 0.2-5.9%). These characteristics were present in 19%, 31%, 33%, and 33% of pooled patient populations, respectively. For the 11 SSEP studies, results did not significantly differ between studies in which the treating physicians were or were not masked from the test result, prospective and retrospective studies, studies with short and long follow-up periods, and studies with high or low overall poor outcome.

INTERPRETATION: SSEP has the smallest CI of its pooled positive-likelihood ratio and its pooled false-positive test rate. Because evoked potentials are also the least susceptible to metabolic changes and drugs, recording of SSEP is the most useful method to predict poor outcome.