Spinal Arteriolosclerosis Is Common in Older Adults and Associated With Parkinsonism
Background and Purpose—There are few studies of spinal microvascular pathologies in older adults. We characterized spinal cord microvascular pathologies and examined their associations with other spinal and brain postmortem indices and parkinsonism in older adults.
Methods—We documented 3 features of microvascular pathologies in spinal cord and brain specimens from 165 deceased older participants. We also measured spinal white matter pallor. Parkinsonian signs were assessed with a modified version of the motor section of the Unified Parkinson’s Disease Rating Scale. We examined the associations of spinal arteriolosclerosis with other spinal and brain postmortem indices and parkinsonism proximate to death using regression models which controlled for age and sex.
Results—Microinfarcts and cerebral amyloid angiopathy were not observed within the spinal cord parenchyma. Spinal arteriolosclerosis was observed at all spinal levels (C7, T7, L4, S4) examined and was more severe posteriorly than anteriorly (posterior: 4.3, SD=0.72 versus anterior: 3.9, SD=0.74; t=14.58; P<0.001). Arteriolosclerosis was more severe in the spinal cord than in the brain (cord: 4.10, SD=0.70; brain: 3.5, SD=0.98; t=10.39; P<0.001). The severity of spinal arteriolosclerosis was associated with spinal white matter pallor (r=0.47; P<0.001). Spinal arteriolosclerosis accounted for ≈3% of the variation in parkinsonism in models controlling for age, sex, brain arteriolosclerosis, and cerebrovascular disease pathologies. Further models showed that the association of spinal arteriolosclerosis and parkinsonism was not mediated via spinal white matter pallor.
Conclusions—Although the regional distribution of microvascular pathologies varies within the central nervous system, spinal arteriolosclerosis is common and may contribute to the severity of spinal white matter pallor and parkinsonism in older adults.
There is increasing recognition that microvascular brain pathologies may contribute to late-life motor impairment.1,2 To date, antemortem imaging and autopsy studies of microvascular pathologies in older adults have focused on the brain, whereas the spinal cord has been virtually unstudied.3 Thus, it is unknown to what degree microvascular pathologies which are common in the brain occur in the spinal cord of older adults.4
To fill these gaps in our knowledge about the spinal cord, we assessed 3 microvascular pathologies including microinfarcts, cerebral amyloid angiopathy (CAA), and arteriolosclerosis in spinal specimens from well-characterized older individuals participating in the Rush Memory and Aging Project.5 First, we examined the distribution of microvascular pathologies in spinal cord. Next, we examined the intercorrelation of spinal and brain arteriolosclerosis, then its associations with other indices of cerebrovascular disease and degenerative brain pathologies.
Next, we sought to determine whether the spinal arteriolosclerosis is associated with potential deleterious tissue or clinical consequences. First, we examined its association with spinal white matter pallor. Our previous work has shown that several indices of brain pathology when considered together only account for a small minority of the variance of parkinsonism.6,7 Motor control networks in the cortical and subcortical brain regions extend to networks located within the spinal cord, and damage to either or both can impair motor function.8,9 So in further analyses, we examined whether spinal arteriolosclerosis was independently associated with the severity of parkinsonism proximate to death.6
Participants were from the Memory and Aging Project approved by the Institutional Review Board of Rush University Medical Center. Each subject signed an informed consent for annual examination and donation of brain and spinal cord at the time of death.5
MAP (Memory and Aging Project) began in 1997 and its structured postmortem examination was limited to the brain. The current study prospectively collecting spinal microvascular pathologies from decedents undergoing autopsy was added in 2012. Decedents in the current study were older at baseline and had more severe parkinsonism and lower cognitive function compared with living MAP participants. At baseline, these 2 groups did not differ with respect to sex, years of education or history of the number of vascular diseases and vascular risk factors (results not shown).
Parkinsonism and Other Clinical Covariates
A uniform structured clinical evaluation was performed each year and included a 26-item modified version of the motor portion of the Unified Parkinson’s Disease Rating Scale and cognitive testing.6 Date of birth, sex, vascular diseases, and risk factors were collected through a participant interview (Methods in the online-only Data Supplement).5
A uniform gross and microscopic postmortem examination was conducted to collect brain indices of macroinfarcts, atherosclerosis, microvascular pathologies, and neurodegenerative pathologies including Alzheimer’s disease pathology, Lewy body disease pathology, and nigral neuronal loss as previously described (Methods in the online-only Data Supplement). Postmortem indices described below were collected from 4 levels of the spinal cord (C5/6, T7, L4/5, and S4/5). Neuropathologic evaluation was performed blinded to the clinical data, reducing the potential for bias.
Hematoxylin and eosin stained 6 µm sections were used to identify microscopic infarcts in midfrontal, midtemporal cortex, anterior cingulate, inferior parietal, entorhinal and occipital cortices, hippocampus, anterior basal ganglia, anterior thalamus, cerebellum, substantia nigra, and 4 levels of the spinal cord. Age and location was recorded for each microscopic infarct. Only chronic microinfarcts were included in our analyses.10
Cerebral Amyloid Angiopathy
CAA pathology was assessed in 4 levels of the spinal cord and 4 brain regions: midfrontal, midtemporal, inferior parietal, and calcarine cortices. The presence of CAA was assessed by immunohistochemistry using monoclonal anti-human amyloid-β, 4G8 (1:9000; Covance Labs, Madison, WI). For each region, meningeal and parenchymal vessels were assessed for amyloid deposition and scored from 0 to 4 for the degree of circumferential deposition and the number of vessels involved.11 We also documented the presence or absence of capillary CAA and double barreling in all regions examined.11
We evaluated the small arterioles within the spinal cord and brain with a semiquantitative grading system from 0 (none) to 6 (severe) as previously described.10 Figure 1 shows the spectrum of arteriolosclerosis observed within the spinal cord, and Figure 2 shows the similarity of arteriolosclerosis in the spinal cord and brain. We examined vessels within 4 levels of the spinal cord, in the white matter of the anterior and posterior watershed regions and the anterior basal ganglia.10 The weighted κ statistics for interrater agreement for assessment of the severity of arteriolosclerosis was 0.47 reflecting moderate agreement.
Spinal White Matter Pallor
At each of the 4 levels of the spinal cord, a semiquantitative scale from 0 (none) to 6 (severe) was used to assess the degree of pallor of the posterior column (gracile fasciculi) and the lateral corticospinal tracts bilaterally. In the cervical cord, pallor of the cuneatus was also recorded. These individual measures were used to summarize pallor for each cord level. Figure 3 illustrates the locations of the spinal regions in which white matter pallor was measured and also illustrates the spectrum of white matter pallor observed in the cervical spinal cord. The weighted κ statistic for interrater agreement for assessment of the severity of white matter pallor was 0.86 reflecting substantial agreement.
We examined the bivariate associations of spinal arteriolosclerosis and other spinal and brain postmortem indices and vascular disease and risk factors. The global parkinsonian sign score had a positively skewed distribution and was subjected to a square root transformation; the transformed scores were used in all analyses. We used a series of regression models to document the association of postmortem indices with the severity of parkinsonian proximate to death and included terms for age, sex, and for the time from the last valid examination proximate to death. To quantify interrater reliability, weighted κ assigned partial weight (0.9) to adjacent raters. All analyses were performed using SAS/STAT software Version 9.3 (SAS Institute Inc, Cary, NC) on a Hewlett Packard ProLiant ML350 server running LINUX.12 Statistical significance is recognized when P<0.05.
Microvascular Pathologies in the Spinal Cord
There were 165 individuals included in these analyses and their clinical characteristics proximate to death and their postmortem indices are included in Table 1.
Although microinfarcts are common in the brain (Table 1), none were observed in the spinal cord.
CAA was common in the brain (Table 1) and was documented in a minority of meningeal vessels (cervical 9/160, 6%; thoracic 7/152, 5%; lumbar 7/158, 4%; sacral 7/140, 5%). Parenchymal, double-barreling, and capillary CAA was not observed in the spinal cord.
Arteriolosclerosis was common in the brain (Table 1) and was observed at all 4 levels of the spinal cord. The severity in the spinal cord ranged from 2 to 6 with an average value of 4.17 (SD=0.83). In >50% of participants, the severity of anterior and posterior spinal arteriolosclerosis was rated the same. When the ratings differed, posterior spinal arteriolosclerosis was almost always more severe (Table I in the online-only Data Supplement). These findings and further analyses support combining severity measures into a summary spinal measure (Methods in the online-only Data Supplement). The average severity for spinal arteriolosclerosis based on all sites examined was 4.10 (SD=0.70). Spinal arteriolosclerosis was not associated with age of death (r=0.12; P=0.131) or sex (t163=0.04; P=0.967).
Associations of Spinal Arteriolosclerosis With Other Cerebrovascular and Degenerative Indices of Brain Pathology
Spinal and brain arteriolosclerosis showed a moderate association (r=0.23; P=0.004), but spinal arteriolosclerosis was more severe than brain arteriolosclerosis (spinal: 4.1, SD=0.70 versus brain: 3.5, SD=0.98; t=10.38; P<0.001). Figure I in the online-only Data Supplement contrasts the heterogeneity of arteriolosclerosis in spinal cord and brain.
In contrast, spinal arteriolosclerosis was not related to other cerebrovascular pathologies (macroinfarcts, microinfarcts, atherosclerosis, CAA) or to indices of other degenerative brain pathologies (Alzheimer’s disease pathology, Lewy bodies, and nigral neuronal loss; all P>0.200; Table II in the online-only Data Supplement).
Clinical Risk Factors, Spinal, and Brain Arteriolosclerosis
Brain arteriolosclerosis was not associated with clinical vascular diseases and risk factors, although spinal arteriolosclerosis was associated with higher BMI and smoking (Table III in the online-only Data Supplement).
Spinal Arteriolosclerosis and Spinal White Pallor
At all 4 levels of the spinal cord, white matter pallor was more severe in the posterior column when compared with the lateral corticospinal tract. In the cervical region within the posterior column, the cuneatus was intermediate in severity between the gracilis and lateral corticospinal tract at the same level (Table IV in the online-only Data Supplement).
Mean white matter pallor ranged from 2.7 at the sacral level to 3.7 in the lumbar cord and was intermediate for the cervical and thoracic levels. Further analyses supported combining severity assessments from all 4 spinal cord levels into a composite white matter pallor severity measure (Methods in the online-only Data Supplement). The average white matter pallor for all 4 spinal cord levels was 3.4 (SD=0.61) and was not associated with age at death (r=0.08; P=0.303) or sex (t163=−0.65; P=0.520). A scatter plot (Figure 4A) illustrates that the severity of spinal arteriolosclerosis was related to the severity of white matter pallor (r=0.47; P<0.001). Although structural spinal abnormalities can cause spinal damage, spinal arteriolosclerosis was not related to the history of cervical radiculopathy (r=0.08; P=0.328).
Spinal Arteriolosclerosis and Parkinsonism Proximate to Death
The association of spinal arteriolosclerosis and level of clinical parkinsonism proximate to death is illustrated in a scatter plot in Figure 4B. Spinal arteriolosclerosis accounted for an additional 3% of the variance of parkinsonism when controlling for age, sex, and time from the last examination to death (Table 2, model A). In contrast, brain arteriolosclerosis was not associated with parkinsonism in a model alone. (Table 2, model B). In contrast, spinal arteriolosclerosis was independently associated with parkinsonism in models which included terms for brain arteriolosclerosis (Table 2, model C), other cerebrovascular disease pathologies (Table 2, model D), and other degenerative brain pathologies (Table 2, model E).
The pathological basis for parkinsonism in older adults without a clinical diagnosis of Parkinson’s disease (PD) may vary from individual’s with a clinical diagnosis of PD.6 In a sensitivity analysis, we repeated the analyses described above to establish that spinal arteriolosclerosis remained associated with the severity of parkinsonism when we excluded 6 cases with a diagnosis of clinical PD proximate to death (Table V, models A through E in the online-only Data Supplement).
Previous antemortem brain imaging studies have suggested that small-vessel disease manifested by white matter brain abnormalities is associated with late-life motor impairments.13 Thus, the association of spinal arteriolosclerosis with clinical parkinsonism could be attenuated when a term for spinal white pallor is included in the same model. In a regression model which controlled for age and sex, spinal white matter pallor was not associated with parkinsonism proximate to death (estimate, 0.220; SE, 0.174; P=0.207). Spinal arteriolosclerosis remained associated with parkinsonism in a model including a term for spinal white matter pallor (estimate, 0.347; SE, 0.168; P=0.040).
Microvascular pathologies in the spinal cord have been almost entirely unstudied. We documented indices of microvascular pathologies in spinal cord and brain specimens from 165 deceased older participants with motor assessments proximate to death. Microinfarcts and CAA were common in brain but were not observed within the spinal cord parenchyma. In contrast, arteriolosclerosis was common at all levels of the spinal cord and associated with brain arteriolosclerosis. Support for the deleterious effects of spinal arteriolosclerosis is suggested by its association with the severity of white matter pallor and its independent association with the severity of parkinsonism proximate to death. These data suggest that although the regional distribution of microvascular pathologies may vary within the central nervous system (CNS), spinal arteriolosclerosis is common and may contribute to the severity of spinal white matter pallor and parkinsonism in older adults.
The increased recognition of the importance of microvascular pathologies and late-life motor impairment in older adults derives almost exclusively from brain imaging and postmortem studies of the brain, whereas studies of the spinal cord of well-characterized older adults are lacking.1,14–16 There are a handful of previous studies which have examined postmortem indices of spinal cord in older adults, but these examined a small numbers of cases or have generally focused on larger arteries such as the anterior spinal artery or meningeal arteries.3,17,18 The current study provides novel data about the accumulation of microvascular pathologies in spinal cords of older adults and lends support to several observations made in past smaller autopsy studies. Whereas microinfarcts and CAA were common in the brains of older adults,6,7,10 microinfarcts were not observed within the spinal cord segments which we examined. Thus, although an individual might have clinical vascular diseases and risk factors and documented brain microinfarcts, no spinal microinfarcts were observed as noted in previous smaller studies.17 Moreover, CAA, which is common in brains from older adults, did not affect intramedullary spinal vessels but was noted in 5% of meningeal vessels.
In contrast to microinfarcts and CAA, we observed arteriolosclerosis in both the brain and at all levels of the spinal cord. In fact, the severity of small-vessel disease in the same individual seemed more severe in the spinal cord than in the brain. Although it can be difficult to conclusively differentiate between arterioles and venules, both are important components of the microvascular network underlying the blood–brain/spinal barrier.19,20 The dissociation between the presence of spinal arteriolosclerosis and the absence of spinal microinfarcts underscores the necessity for further work to explicate alternative mechanisms through which arteriolosclerosis might contribute to age-related spinal cord changes including but not limited to neuronal loss and demyelination.3,17,18 Some have suggested that the lack of spinal microinfarcts derives from the extensive collaterals present in the spinal cord when compared with subcortical brain regions. Yet, this suggestion does not account for the lack of microinfarcts in known spinal watershed regions. The absence of microinfarcts in the spinal cord despite arteriolosclerosis suggests that microinfarcts are not the inevitable consequence of arteriolosclerosis and underscore important gaps in our understanding about the pathogenesis of microinfarcts within the CNS.14 Nonetheless, in the absence of microinfarcts, that is, frank spinal tissue damage, other local mechanisms leading to spinal cord degeneration such as oxidative stress, inflammation or other nonischemic mechanisms, and spinal changes secondary to brain lesions will need to be explored.21–24 The current data suggest that microvascular pathologies within the CNS are heterogeneous, and further studies are needed to elucidate which factors account for the observed regional differences.3,17,18,25
Documenting spinal arteriolosclerosis does not inform on its tissue or clinical consequences. The severity of spinal arteriolosclerosis was associated with the degree of spinal white matter pallor. Posterior column demyelination documented in some previous reports has been considered secondary because of axonal or regional nerve root loss.17 The current study focused on microvascular pathologies, and it will be important to assess the loss of axons and ganglion cells to determine whether the pallor which was observed is primary or secondary demyelination.18 In the absence of microinfarcts, nonischemic mechanisms which may be mediated via extracellular matrix inflammation may underlie the association of spinal arteriolosclerosis and white matter pallor.24
Further evidence suggesting that spinal arteriolosclerosis may have a deleterious effect on health is the finding in the current analyses that it is independently associated with the severity of parkinsonism proximate to death. This finding was robust and remained significant when controlling for brain arteriolosclerosis, several vascular brain pathologies, and spinal white matter pallor. Although spinal arteriolosclerosis was associated with both spinal white matter pallor and parkinsonism, spinal white matter pallor was not associated with parkinsonism proximate to death. These results suggest that other unmeasured cellular elements or factors may link spinal arteriolosclerosis with the severity of parkinsonism proximate to death. These data also suggest that there may be several mechanisms through which spinal arteriolosclerosis contributes to tissue damage and adverse health consequences. These findings need to be replicated by others, and longitudinal studies are needed to determine the directionality of these cross-sectional associations.
In the current study, spinal arteriolosclerosis alone accounted for ≈3% of the variance of parkinsonism when controlling for brain arteriolosclerosis, other cerebrovascular indices, and multiple degenerative brain pathologies. To appreciate these findings, it is worth noting that our previous work has shown that together multiple brain pathologies accounted for <10% of the variance of parkinsonism proximate to death in older adults without Parkinson’s disease.6,7,10 Because motor control systems extend beyond the brain through the CNS to the spinal cord, the additional variance of parkinsonism accounted for by spinal arteriolosclerosis in the current study is not surprising. Moreover, it suggests that other motor-related structures within and outside the CNS such as cerebellum or muscle are likely to make separate contributions to the severity of parkinsonism in older adults.7,26 The current spinal data extend previous brain imaging and autopsy studies by providing evidence that indices of vascular pathology in both spinal cord and brain may contribute to parkinsonism in many older adults.7,10,27 Thus, the clinical phenotype of parkinsonism in older adults is not simply an early manifestation of PD, but this common clinical phenotype derives from diverse neuropathologies including Alzheimer’s disease, PD, and CNS cerebrovascular pathologies.6,28–31 Thus, like clinical Alzheimer’s disease dementia, the phenotype of parkinsonism in older adults may be commonly caused by mixed neuropathologies.16 Finally, the finding that spinal arteriolosclerosis may be a common unrecognized cause of parkinsonism in older adults suggests that the health burden of CNS microvascular pathologies may be underestimated in our aging population.1,32 Approaches which allow for antemortem identification and treatment of at risk individuals are needed to prevent late-life motor impairment.2,14
There are several limitations to the current study. The cases we studied are from a selected cohort. It will be important to investigate these findings in more diverse cohorts. Although we examined 3 features of microvascular pathologies which can only be assessed via histopathology, several other features including expanded Virchow–Robin spaces and microbleeds were not systematically assessed. Modest interrater agreement for severity of arteriolosclerosis is a limitation. Identifying unmeasured confounding pathologies is always a concern in observational studies. Although spinal white matter pallor was assessed and did not link the association of spinal arteriolosclerosis with parkinsonism, further work is needed to evaluate whether the preferential posterior column involvement may manifest as sensory loss that might contribute to other clinical impairments such as urinary incontinence, balance, or increased motor variability in older adults. Moreover, spinal white matter assessments in this study were not compared with rostral white matter measures in the brain and brain stem or with musculoskeletal or clinical measures of cervical myelopathy. More precise measures to quantify white matter tissue integrity and tractography are needed to more fully assess the tissue and clinical consequences of spinal arteriolosclerosis.
There are several strengths to the study, including the community-based cohort with large numbers of women and men coming to autopsy following high rates of clinical follow-up and high autopsy rates. Uniform structured clinical procedures were used that included a detailed assessment of parkinsonian signs that have been widely used in other studies.
We thank all the participants in the Memory and Aging Project. We also thank the staff of the Rush Alzheimer’s Disease Center.
Drs Buchman and Nag are supported by grants from the National Institutes of Health; Dr VanderHorst by National Institutes of Health and Judy Goldberg Foundation; Dr Leurgans by National Institutes of Health/National Institute of Aging; Dr Schneider by National Institutes of Health and Advisory Board-Eli Lily; Dr Bennett by National Institutes of Health, has received honoraria for nonindustry sponsored lectures, and has served as a paid consultant to Danone, Inc, Wilmar Schwabe GmbH & Co, Eli Lilly, Inc, Schlesinger Associates, and the Gerson Lehrman Group. The other author reports no conflicts.
Sources of Funding
This work supported by the NIH grants R01AG17917, R01NS708009, R01AG43379 the Illinois Department of Public Health, and the Robert C. Borwell Endowment Fund.
Guest Editor for this article was Ralph Sacco, MD.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.117.017643/-/DC1.
- Received April 7, 2017.
- Revision received August 9, 2017.
- Accepted August 14, 2017.
- © 2017 American Heart Association, Inc.
- Gorelick PB,
- Scuteri A,
- Black SE,
- Decarli C,
- Greenberg SM,
- Iadecola C,
- et al
- Jellinger K
- Buchman AS,
- Leurgans SE,
- Nag S,
- Bennett DA,
- Schneider JA
- 12.↵SAS Institute Inc. SAS/STAT® User's Guide, Version 9.3 [Computer Program]. Version 9.3. Cary, NC: SAS Institute Inc; 2012.
- Charidimou A,
- Boulouis G,
- Gurol ME,
- Ayata C,
- Bacskai BJ,
- Frosch MP,
- et al
- Kapasi A,
- DeCarli C,
- Schneider JA
- Wilhelm I,
- Nyúl-Tóth Á,
- Suciu M,
- Hermenean A,
- Krizbai IA
- Simpson JE,
- Hosny O,
- Wharton SB,
- Heath PR,
- Holden H,
- Fernando MS,
- et al
- Garbuzova-Davis S,
- Haller E,
- Tajiri N,
- Thomson A,
- Barretta J,
- Williams SN,
- et al
- Rosenberg GA
- Hashizume Y,
- Yoshida M,
- Wang Y,
- Kume A,
- Kameyama T,
- Ando T,
- et al
- Buchman AS,
- Leurgans SE,
- Yu L,
- Wilson RS,
- Lim AS,
- James BD,
- et al
- Korczyn AD
- Gouw AA,
- Seewann A,
- van der Flier WM,
- Barkhof F,
- Rozemuller AM,
- Scheltens P,
- et al