The objective of the study was to assess total cerebrospinal creatinine-kinase activity (CSF-CK) measurement in differential diagnosis of "metabolic" and organic causes of coma. The setting for the study was a tertiary care reference medical center and community general hospital. The design of the study was a series of consecutive patients with profound coma (Glasgow scale ratings between 3 and 6) as the presenting complaint to the emergency room and controls. Measurements and main results were as follows: CSF-CK was measured in 103 consecutive patients including 18 patients with metabolic causes of coma, 27 patients with organic causes of coma, 18 patients scheduled for elective orthopedic surgery with epidural anesthesia and 27 patients with compressive myelopathy and radiculopathy. CSF-CK activities were significantly different between groups (H = 29.48, p < 0.001, Kruskal-Wallis test), controls had a median of 0 mU/ml (range 0-16 mU/ml), patients with metabolic causes of coma had a median of 0 mU/ml (range 0-65 mU/ml), patients with compressive myelopathy or radiculopathy had a median of 19 mU/ml (range 0-80 mU/ml), and patients with organic causes of coma had a median of 20 mU/ml (range 0-400 mU/ml). The test sensitivity was 83% (95% confidence interval (CI 65-100%) specificity 62% (CI 43-80%) positive predictive value 60% (CI 49-79%) and negative predictive value 85% (CI 75-95%). The conclusion of the study was that the test is useful for ruling out metabolic causes of coma when CSF-CK activity is high (i.e., above 16 mU/ml).