OBJECTIVE: Hyperglycemia occurring after head injury is associated with poor neurological outcome. We tested the hypothesis that blood glucose levels are associated with brain tissue pH (pH(b)) and that the correction of hyperglycemia would result in an improvement in pH(b). METHODS: This is a retrospective analysis of a prospectively collected database. Thirty-four patients in a tertiary care neuroscience critical care unit with major traumatic brain injury underwent pH(b) monitoring. RESULTS: A total of 428 glucose measurements were recorded during pH(b) monitoring. Mean glucose level was 7.1 mmol/L (range, 2.8-21.7 mmol/L) and median (interquartile range) pH(b) was 7.11 mmol/L (7.00-7.19 mmol/L). To account for the correlated, unbalanced nature of the data, a linear generalized estimating equation model was created. This model predicted that for each 1 mmol/L increase in blood glucose, pH(b) changed by -0.011 mmol/L (95% confidence interval, -0.016 to -0.005 mmol/L; P < 0.001). This relationship remained significant in a multivariable model that included cerebral perfusion pressure, brain tissue oxygen and carbon dioxide tension, and brain temperature. Twenty-one episodes of significant hyperglycemia (>or=11.1 mmol/L) treated with intravenous insulin were identified. Insulin therapy significantly reduced blood glucose concentration from a median (interquartile range) of 11.9 mmol/L (range, 11.4-13.6 mmol/L) to 8.8 mmol/L (range, 7.3-9.6 mmol/L; P < 0.001). Baseline pH(b) was not significantly different from pH(b) associated with the subsequent glucose reading of less than 11.1 mmol/L (P = 0.29), but there was a suggestion of improvement if the change in blood glucose was large. CONCLUSION: Blood glucose is associated with brain tissue acidosis in patients with major head injury. Prospective studies are required to confirm these results and to determine whether treatment of hyperglycemia improves outcome.