The 2 key findings of this study were that WMH accrual largely amounts to growth of existing lesions rather than development of new lesions and that greater cognitive decline is associated with greater lesion growth but not with greater emergence of new lesions. We discuss the significance of each finding in turn.
Approximately 80% of incident WMH volume over time was due to growth of existing lesions. This finding suggests that the bulk of aging-associated WM degeneration may represent a continuous process of gradually spreading injury rather than a series of abrupt and isolated anatomic events. Our previous study suggested that WMHs are surrounded by regions of milder WM injury26
; the current study and others27
extend this finding by suggesting that this same mildly injured tissue is differentially vulnerable to subsequent worsening of the injury to such a severity that it is labeled as WMH. The histopathologic origins of age-related WM degeneration include gliosis, degeneration in myelinated axons, and, importantly, small vessel changes whose linear effects on WM integrity may lead to the emergence/development of WMH. Hypertension, diabetes, and other small-vessel disease risk factors, from their earliest phases in life, are likely to be accompanied by progressive and subtle cerebral WM degeneration and whose long-term expression is WMH.28,29
This finding has relevance for the clinical interpretation of WMHs, which should be considered merely the most extreme foci of much more diffuse and incipient brain injury. In addition, the vulnerable surround of existing WMHs may represent a novel treatment target that, if salvaged, might modify the time course of progressive WM degeneration and its cognitive consequences.
The finding that WMH extension paralleled episodic memory and executive function losses supports a model in which WMHs injure cortical connections critical to higher-order cognition30
and WMH extension further exacerbates the same injury. Results from previous cross-sectional studies linked the presence of WMH in specific white matter tracts to poorer executive function and episodic memory.5,30
Our finding further reinforces the importance of understanding the gradual spatial spreading of WM degeneration as the mode of aging or vascular disease-related WM injury that carries the most substantial cognitive consequences. Our findings also question whether disruption of connection between gray matter regions may provide complementary or better information for gray matter volumes in certain regions to explain worse cognitive performances in specific domains. Further validation of this model, for example, by linking WMH growth in specific WM tracts to tract-specific progressive cognitive losses is required.
In contrast to prior observations that different categories of WMHs carry different cognitive repercussions,31
our findings suggest that emergent WMHs might not be immediately relevant to changes in cognitive performance. Emergent WMHs increase in number with increases in baseline WMH burden but are not strongly associated with age and clinical diagnosis, 2 traditionally strong predictors of WMH accrual6
; this observation suggests that emergent WMH may involve different pathologic substrates than the far more numerous extending WMHs.7
However, the emergent WMHs constitute a very small proportion of the volume of new WMH, which may provide insufficient power to detect associations with change in cognitive measures. Further studies with greater representation of emerging WMHs are needed to determine this conclusively.
In the present study, age adjustment mitigated the significant effect of baseline WMH, but not of WMH extension, on the rates of change in both cognitive scores. Previous studies suggested that because brain injury (such as WMHs) become more common with age, the correlation between age and cognition, which is the basis of age corrections, is partly due to an underlying effect of brain injury on cognition.17,24
Thus, to correct for “age” is to correct for both age and brain injury, and this correction reduces the ability to associate brain injury with cognition. Therefore, whereas adjustments for age might be justified in the context of extremely healthy brains, such adjustments among populations representing substantial aging-related lesions may have lessened justification. Although age adjustment remains controversial, we emphasize that the key findings of predominant WMH extension and relation of WMH extension to cognitive change were observed irrespective of this methodologic detail. Because the theoretical and clinical implications of age adjustment are not entirely clear, additional investigations are needed to clarify this question.
These results have, nonetheless, to be considered with some caution with regard to the size sample of clinical groups, notably the AD group. That annual episodic memory change did not differ according to clinical group is consistent with findings of 2 previous studies that included larger number of individuals.11,25
These studies suggested that boundaries for differentiating normal cognition, MCI, and dementia are by nature arbitrary cut points in a continuum of pathologic changes, and none of the 3 diagnoses may strictly guarantee a particular predetermined level of cognitive ability in a particular domain.
The present study provides new evidence of a continuous process of WM degeneration over time that parallels concurrent cognitive decline. Additional studies are needed to clarify the distinct pathologic substrates and consequences of growing and emerging WMHs.