The central finding in this study is that alleviation of hypertension and reduction of BP to normotensive levels failed to prevent grey matter loss over one year of treatment. As noted previously, grey matter morphology differences related to hypertension appear to differ from those related to normal aging, e.g. [7
]. It is possible that high BP induces central hemodynamic effects contributing to atrophy of particular grey matter areas. Based on this inference, we expected that reducing BP over the course of a year might arrest or even diminish grey matter loss. Our findings suggest, however, that some aspect of essential hypertension is continuing to influence brain morphology independent of the level of peripheral BP. Of note, another index of brain morphology, cortical thickness, was recently observed to correlate negatively with blood pressure in pre-hypertensive individuals [52
]. More generally, a recent review of structural neuroimaging studies of the aging brain [53
] concluded that treatment of hypertension reduces, but does not eliminate effects of aging on the brain. As summarized in that review, compared to normotensives, treated hypertensives have more white matter hyperintensities, prefrontal volume shrinkage and reduced hippocampal volume [11
We must emphasize though that we only studied two medications over a year of treatment. The literature is not specific enough to determine whether areas specifically related to hypertension show continued or arrested atrophy with longer term treatment. It is possible that early intervention as well as longer duration of intervention would maintain grey matter morphology among treated hypertensive individuals. Alternately, it is conceivable that medication rather than a hypertensive process affects grey matter morphology. We cannot discount this possibility without a placebo control arm to our study, but doubt on this hypothesis is cast by our finding that two medication groups did not differ in their morphology changes despite differences in the mechanism of action of the two medications.
We observed volume loss in nine brain regions over the course of a year of treatment lowering blood pressure. The grey matter regions were chosen for this study specifically because of prior cross-sectional findings of their smaller volume in these areas in hypertensives compared to normotensive controls. Shrinkage of specific grey matter regions after a year of effective anti-hypertensive treatment is in accord with the cross-sectional findings in similar brain areas. The validity of our results rests primarily on this longitudinal verification of cross-sectional results. In the hypertensive group, in seven of the nine examined regions, we observed a mean one-year loss that exceeded shrinkage in the normotensive sample.
Caution is required in interpreting the statistical inferences given the variability of the results in the small comparison sample. A portion of the variability may be due to the use of different scanners at different sites in the comparison sample drawn from the ADNI study. Despite this variability, the utility of the comparison sample was supported by the typical age-related changes in grey matter volume observed. Changes over a year in the percentage of grey matter relative to intracranial volume were similar to those reported in the literature. These changes were the same in women as estimates obtained in large samples by Taki [57
] and slightly higher, but well within the range of values for men. Importantly the results do not depend entirely on the comparison sample, as the larger hypertensive sample showed decreases in grey matter volume across all of the nine areas examined, albeit varying in degree of decrease. Exploratory analyses suggested that longitudinal changes in treated hypertensives was largely confined to the areas defined a priori, although gyrus rectus was identified as decreasing more in hypertensives relative to controls in these exploratory analyses. The gyrus rectus has not been well studied specifically in hypertensives, but its grey matter has been shown to decrease with age [58
] and to correlate with diagnosis of unipolar depression [59
], age-related differences in the olfaction [60
], and performance on executive function tasks [61
Although some of the examined regions were identified in prior work as hemisphere- or gender-specific, our conservative strategy was to test areas bilaterally and across genders in an attempt to avoid spurious findings. Despite this, we did find accentuated grey matter loss among hypertensives in BA 21, BA28, and thalamus extending prior findings [8
]. Follow-up analyses showed that the loss in BA21 was greater in women than in men. Medial frontal and hippocampal areas showed directional effects, but were not separately significant discriminators between hypertensives and controls [6
], although our follow-up analysis showed hypertensive women showing hippocampal loss that was greater than that among men. Occipital and inferior parietal areas showed no losses among hypertensives relative to controls over the year, despite some evidence of such loss longitudinally in occipital cortex [7
] and cross-sectionally in inferior parietal cortex [13
]. We have no explanation for this. Aging has typically been shown to influence the occipital areas less than the other areas we examined, although the parietal lobe is typically shown to be influenced by age [62
]. An increase in cerebellar cerebrospinal fluid with hypertension was previously reported by Strassburger et al. [15
], but we did not replicate this finding and found no substantial differences in white matter or cerebrospinal fluid volumes over the year of treatment. As noted earlier, functional correlations of grey matter change have not been widely investigated. We examined this question but found no strong correlations between neuropsychological function and the decrease in grey matter volume over a year. Neuropsychological deficits related to hypertension are small and the current sample size as well as the relatively short (one year) time frame may have obscured any ongoing functional changes related to grey matter volume loss.
These analyses driven by prior literature were supplemented by a post hoc examination of BP and grey matter volume in the current sample. Using pretreatment and control participant BP we found relations between systolic BP and superior orbital frontal cortex, lateral orbital cortex, gyrus rectus, cerebellum, and BA 11. Note that areas detected did not include all the areas assessed in our longitudinal analysis; more information is clearly needed on the time course of morphological changes. Slow, incremental change would not have been detected in our one year study. We did not detect interactions with age or gender (overall), but it is possible that morphological changes are accentuated in older sample or cluster around menopausal periods in women.
Mechanisms accounting for the precise pattern of grey matter change in hypertension are unclear. General changes in brain morphology have been related to small vessel arteriosclerosis and lipohyalinosis as well as rarefaction, white matter lesions, oxidative stress, neuronal shrinkage, apoptosis, local ischemia, and beta amyloid deposition [65
]. It remains unclear how hypertension accelerates these factors, as these reviews indicate, and it is equally unclear at present why these mechanisms would be specifically active in the regions identified in our study. Areas showing specific loss with hypertension were subsumed within larger regions typically associated with aging, e.g., temporal lobe. Our analyses directed at areas know to show changes with aging suggested accelerated aging in the temporal lobe of hypertensives although other areas seemed to show equivalent aging effects between hypertensive and control participants. Given this, we cannot determine at this point whether hypertension accelerates processes related to normal aging in specific areas or introduces new factors independent of normal aging process. Overall, temporal and thalamic changes appeared the most robust in the current sample. Future work with fine-grained analyses will be required to clarify the possible mechanisms.
Our study has several limitations. The most prominent is we were unable to find closer matching between the groups. The members of a small sample obtained form LONI ADNI database differed from our hypertensive participants in age and BMI. This comparison provides only initial data suggesting the observed losses in the hypertensive patients were not due to normal aging. This limitation is emphasized by the gains in grey matter volume over a year apparent in the mean data from the small comparison group; such an effect is possible, but larger sample sizes would be required to decrease the likelihood of variation due to statistical and methodological ‘noise’. The ADNI sample did, however, appear to represent normal aging appropriately. Moreover, because hypertensive group was younger, there was an a priori lesser likelihood of finding brain shrinkage in that group rather than in older controls. This reduces a threat to validity of our conclusions. A strength of this study is that the results were not confounded by prior pharmacological treatment. However, we observed the reduction of BP with only two of the many possible medications for treating BP. Thus, generalization to patients who were exposed to various pharmacological regimens and those who were treated with different anti-hypertensive agents must be made with caution.
In summary, in cortical regions previously identified as vulnerable to hypertension, newly treated hypertensive individuals continue to show grey matter shrinkage over a year of successful treatment. The continued atrophy in the presence of successful reduction of BP suggests that essential hypertension may directly affect the brain via factors that are not mediated by increased BP. This observation contributes to accumulating research suggesting that essential hypertension from its onset may be a disease of the brain as well as of the vasculature [69