Real time estimation of brain water content in comatose patients
Version of Record online: 22 AUG 2012
Copyright © 2012 American Neurological Association
Annals of Neurology
Volume 72, Issue 3, pages 344–350, September 2012
How to Cite
Ko, S.-B., Choi, H. A., Parikh, G., Schmidt, J. M., Lee, K., Badjatia, N., Claassen, J., Connolly, E. S. and Mayer, S. A. (2012), Real time estimation of brain water content in comatose patients. Ann Neurol., 72: 344–350. doi: 10.1002/ana.23619
- Issue online: 2 OCT 2012
- Version of Record online: 22 AUG 2012
- Accepted manuscript online: 14 APR 2012 08:27AM EST
- Manuscript Accepted: 6 APR 2012
- Manuscript Revised: 27 MAR 2012
- Manuscript Received: 26 MAY 2011
Although brain swelling is an important cause of neurological deterioration, real time measurement of brain edema does not currently exist. Because thermal conductivity is proportional to percentage water content, we used the thermal conductivity constant to estimate brain water content (BWC).
Between June 2008 and November 2010, 36 comatose brain-injured patients underwent cerebral blood flow monitoring using a thermal diffusion probe in our neurocritical care unit. BWC was estimated hourly utilizing the measured thermal conductivity and the known temperature-adjusted thermal conductivity of water. In vitro experiments were performed to validate this formula using agar, glycerol, and water mixtures with different water content.
Thermal conductivity was highly correlated (R2 = 0.99) and estimated water content was well correlated with actual water content (mean difference, 0.58%) in the in vitro preparations. The majority of the 36 patients (median age, 57 years; 44% female) had subarachnoid hemorrhage (n = 14) or cardiac arrest (n = 9). Initial BWC at the time of monitoring ranged from 67.3 to 85.5%. Brain regions appearing edematous on computed tomography showed higher estimated BWC than normal-appearing brain regions (79.1 vs 70.2%; p < 0.01). Bolus osmotherapy (20% mannitol or 23.4% hypertonic saline) decreased BWC from 77.2 ± 0.7% (mean ± standard error) at baseline to 76.1 ± 0.5% at 1 hour, 76.5 ± 0.3% at 2 hours, and 76.7± 0.2% at 3 hours (all p ≤ 0.03).
Real time monitoring of BWC is feasible using thermal conductivity. Further studies are needed to confirm the clinical utility of this technique. ANN NEUROL 2012;72:344–350