© 2005 American Thoracic Society
OSA Brain Morphology DifferencesMagnitude of Loss Approximates Age-related EffectsTo the Editor:O'Donoghue and colleagues report no major brain morphology differences in 25 severe OSA patients (1), and imply that these findings contradict the results reported in our earlier paper (2). However, their findings do not contradict the earlier study; the absence of detected differences in brain morphology in the paper by O'Donoghue and coworkers is explained by higher threshold levels and exclusion of age and handedness as independent variables. The combined message is that gray matter loss in OSA is not extreme, but occurs with a magnitude approximating age-related effects. We had set a statistical threshold that allowed demonstration of known age effects, and found that regions of gray matter loss correlated with severity of OSA (2). O'Donoghue and coworkers used a higher threshold and found no effect. We reanalyzed our data using the same higher threshold and also saw no effect. Large differences (at least 17% in some regions) were required in the O'Donoghue study for statistical significance. The maximum effect in our 2002 study was 18%; in many regions, the minimum detectable change in the O'Donoghue study was likely greater than the detected change in our earlier study (2). O'Donoghue and coworkers also present results using the same threshold level as the Macey paper without replicating the earlier findings. This discrepancy likely stems from their use of a two-sample t test without accounting for age and handedness. Reanalyzing our data with a similar test illustrates the critical difference that including these variables exerts on the results (Figure 1). The t test relies on a difference in group means; if there is age-related overlap (e.g., older control subjects have lower gray matter volume than younger patients with OSA), the test is weakened. Age-matching does not correct for this issue. Our approach in the 2002 paper was an analysis of covariance, with age and handedness (both of which had significant effects) as covariates.
The animal literature shows that intermittent hypoxia, a principal feature accompanying OSA, results in severe damage to hippocampal, frontal cortex and cerebellar structures (3, 4), providing a basis for the volume loss we detected in these regions. Metabolic injury also occurs in the hippocampus of patients with OSA (5). To summarize, OSA is not associated with extreme gray matter loss, but volumetric changes in gray matter occur; these changes are localized, and of a magnitude that approximates age effects.
David Geffen School of Medicine at UCLA, Los Angeles, California FOOTNOTES Conflict of Interest Statement: Neither of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
