Dehydration confounds the assessment of brain atrophy
Abstract—Computerized brain volumetry has potential value for diagnosis and the follow-up evaluation of degenerative disorders. A potential pitfall of this method is the extent of physiologic variations in brain volume. The authors show that dehydration and rehydration can significantly change brain volume: lack of fluid intake for 16 hours decreased brain volume by 0.55% (SD, 0.69), and after rehydration total cerebral volume increased by 0.72% (SD, 0.21).
NEUROLOGY 2005;64:548 –550
T. Duning, MD*; S. Kloska, MD*; O. Steinsträter, PhD; H. Kugel, PhD; W. Heindel, MD; and S. Knecht, MD
Studies on neurodegeneration increasingly rely on computerized measurements of brain volume using MRI.1-3 With a spatial resolution of 1 mm, current MRI technique can detect volumetric changes as little as 1.5 to 3 mL or 0.2 to 0.13% of total brain volume.3,4 This precise quantitative information has potential for early disease diagnosis, evaluation of therapies, and study of disease severity.2,3,5 However, with improved analytical sensitivity, conventional assumptions need to be reconsidered. One such assumption is that brain volume does not have transient variation. Studies on patients with anorexia already revealed various effects of nutritional status on cerebral volume.6 Fluid loss and fluid intake affect body volume and could therefore also change brain volume. We tested to what extent water balance can influence measurements of brain volume.
Methods. Subjects comprised 20 clinically healthy volunteer adults of European descent (mean age, 27 years; SD, 2.8; range, 22 to 32 years; 9 men and 11 women). Subjects gave informed consent in accordance with institutional and federal rules. MR acquisition. Image data were obtained on a 3.0-T system (Gyroscan Intera T30, Philips Medical System, Best, The Netherlands) with a T1-weighted three-dimensional turbo field echo sequence, acquired matrix 256 205 160 over a field of view of 25.6 20.5 16 cm, reconstructed after zero filling to 512 410 320 cubic voxels with an edge length of 0.5 mm. Sequence parameters were repetition time (TR) 7.4 ms, echo time (TE) 3.4 ms, flip angle (FA) 9°, and an inversion recovery prepulse every 805 ms, resulting in a total scanning time of 11:01 minutes. We analyzed the three-dimensional data sets in 10 subjects (median age, 27 years; range, 24 to 31 years; 3 men) at three time points: 1) before and 2) after thirsting for 16 hours, and 3) 20 to 30 minutes after drinking 1.5 L of mineral water (figure 1). The subjects did not drink overnight, precisely from 8 PM to 10 AM the
next morning. They were allowed to eat food containing 500 mL of water in total. Strenuous exercising was also not allowed. Ten control subjects (median age, 28 years; range, 22 to 32 years; 6 men) were examined on two occasions to assess the effect of repositioning in the scanner. They were removed completely from the scanner table, and the prescanning procedure was repeated independently between the two scans. Image analysis. The new fully automated brain atrophy technique SIENA was applied to the three-dimensional T1-weighted images. The SIENA software is part of FSL-FMRIB’s software library (www.fmrib.ox.ac.uk/fsl). This method is a completely automated brain change analysis tool using longitudinal measurements and has been validated using control and patient data (scan rescan; three time points measured, compared time points 1 3 with time points 1 2 2 3). The reproducibility assessment techniques agree with the postulated error in brain volume changes (BVCs) of only 0.2%.4 To assess whether BVCs associated with hydration status are also detectable by conventional neuroradiologic reading, four representative 0.5-mm-thick slices were reconstructed in axial orientation for all subjects from the acquired T1-weighted high-resolution three-dimensional MR data sets. For every subject, the positions of the chosen four slices were equal. All four revealed ventricular space, two only the lateral ventricles (an example is shown in figure 2), and two additionally presented the third ventricle. Two readers (one experienced radiologist and one experienced neurologist) visually rated the images. Both readers were blinded to the hydration states. Images obtained before and after drinking were ranked for decreases in brain volume by a two alternative forced choice procedure. Statistical analysis. To evaluate whether the BVCs at different fluid intake levels differ significantly from zero, a one-sample t test was calculated. The differences between the two hydration states and repositioning next were analyzed using unpaired t tests. Significance levels for statistics were set at p 0.05. Because of the explorative character of the study design, Bonferroni adjustments were not considered appropriate.
*These authors contributed equally to this work. From the Departments of Neurology (Drs. Duning, Kloska, Steinsträter, and Knecht) and Clinical Radiology (Drs. Kloska, Kugel, and Heindel), University of Münster, Germany. Supported by grants from the German Federal Ministry of Education and Research (Mednet Atrial Fibrillation, S.K.) and the IZKF Muenster (NWG2, S.K.). Received December 18, 2003. Accepted in final form September 30, 2004. Address correspondence and reprint requests to Dr. Thomas Duning, University of Muenster, Department of Neurology, Albert-Schweitzer-Str. 22, 48129 Muenster, Germany; e-mail: email@example.com 548 Copyright © 2005 by AAN Enterprises, Inc.
Results. Lack of fluid intake for 16 hours and rehydration with 1.5 L of water led to BVCs significantly different from zero (after thirsting: t 2.48; mean BVC rate, 0.55%; SD, 0.69; after fluid intake: t 10.56; mean BVC rate, 0.72%; SD, 0.21), whereas repositioning did not. Even after exclusion of an outlier, the difference from zero after thirsting remained significant (see figure 3; t 3.16; mean BVC rate, 0.35%; SD, 0.33). Thirsting and fluid intake also yielded greater BVC compared with repositioning (repositioning vs thirsting: t 2.56; p 0.02; repositioning vs fluid intake: t 7.13; p 0.001). The variability of BVCs in the 10 control subjects caused by repositioning was 0.21% (SD, mean BVC rate 0.04; figure 3). The absolute weight loss after 16 hours of thirsting was 1.56 1.01% (mean SD). On conventional neuroradiologic reading of the recon-
Figure 1. Experimental design: three-dimensional MR data sets were acquired before and after thirsting for 16 hours and after drinking 1.5 L of mineral water.
structed high-resolution MRIs, the forced choice visual ranking corresponded to quantitative volumetry in 9 of 10 cases for both readers (see figure 2).
Discussion. During normal conditions, the water content of the human body fluctuates on the order of 2.5 L (i.e., ~3% of the total body weight). The osmotic gradient caused by thirsting leads to a shrinking of astrocytes, which are the key cells for water movement between cellular, vascular, and ventricular compartments of the brain.7,8 However, the brain possesses complex protective regulating mechanisms. Several independent systems, including angiotensin, vasopressin, atrial natriuretic peptide or brain natriuretic peptide, and so-called aquaporins (especially aquaporin-4), protect neuronal tissue against short-term fluid shifts.8 Our data demonstrate that these mechanisms are active but that the counter-regulation was not complete. A decrease in brain volume could still be detected by computerized methods and by visual inspection.
Given a total brain volume of 1,450 mL, decreases of ~0.60% would correspond to a change of ~ 8.7 mL in brain volume, and an increase of 0.72% after severe fluid intake leads to an increase of ~ 10.4 mL. Although the present study design may present extremes of fluctuations in body composition, our findings demonstrate that the hydration status can significantly affect the computed measurement of brain volume. In particular, we consider this important in patient conditions in which small changes of total brain volume are important diagnostic parameters (i.e., in neurodegenerative disease).2,3,9 Therefore, a comparable hydration status should be considered in longitudinal MR-based automated measurement of brain volume. In elderly persons, predominantly studied for neurodegenerative disease, the physiologic reserve of organic systems taking part in regulatory mechanisms is often impaired.10 Thus, further studies in patients with neurodegenerative disease or elderly persons have to clarify whether the reported findings are also valid in these subgroups.
Figure 2. Example of T1-weighted 0.5-mm MR scans ranked by forced choice for decreases in brain volume by two readers. Left, before drinking; right: after drinking 1.5 L of mineral water. Please note the encircled areas (left) that show subtle enlargements of sulci compared with the correspondent image after drinking (right).
February (1 of 2) 2005
1. Wang D, Doddrell DM. MR image-based measurement of rates of change in volumes of brain structures. Part I: method and validation. Magn Reson Imaging 2002;20:27– 40. 2. Chetelat G, Baron JC. Early diagnosis of Alzheimer’s disease: contribution of structural neuroimaging. Neuroimage 2003;18:525–541. 3. Fox NC, Freeborough PA. Brain atrophy progression measured from registered serial MRI: validation and application to Alzheimer’s disease. J Magn Reson Imaging 1997;7:1069 –1075. 4. Smith SM, Zhang Y, Jenkinson M, et al. Accurate, robust, and automated longitudinal and cross-sectional brain change analysis. Neuroimage 2002;17:479 – 489. 5. Frisoni GB. Structural imaging in the clinical diagnosis of Alzheimer’s disease: problems and tools. J Neurol Neurosurg Psychiatry 2001;70: 711–718. 6. Addolorato G, Taranto C, De Rossi G, Gasbarrini G. Neuroimaging of cerebral and cerebellar atrophy in anorexia nervosa. Psychiatry Res 1997;76:139 –141. 7. Stricker EM, Sved AF. Thirst. Nutrition 2000;16:821– 826. 8. Badaut J, Lasbennes F, Magistretti PJ, Regli L. Aquaporins in brain: distribution, physiology, and pathophysiology. J Cereb Blood Flow Metab 2002;22:367–378. 9. Enzinger C, Ropele S, Smith S, et al. Accelerated evolution of brain atrophy and “black holes” in MS patients with APOE-epsilon4. Ann Neurol 2004;55:563–569. 10. Allison SP, Lobo DN. Fluid and electrolytes in the elderly. Curr Opin Clin Nutr Metab Care 2004;7:27–33.
Figure 3. Dehydration and rehydration significantly change brain volume. (A) Effect of thirsting for 16 hours (change in volume between first and second MRI scan). (B) Effect of 1.5-L fluid intake (change in volume between first and third MRI scan). (C) Effect of repositioning (volume change between two scans without drinking or thirsting). The figure represents the volumetric changes at different fluid intake levels relative to the baseline measurement (first MRI scan). Each subject is represented by a single dot. After exclusion of one extreme value (marked by an asterisk), the mean decrease in cerebral volume remained significantly different from zero (mean BVC rate, 0.35%; SD, 0.33).
February (1 of 2) 2005