ORIGINAL ARTICLE
Association between white matter hyperintensity and medial temporal atrophy at various stages of Alzheimer’s disease N. Kandiaha,b, R. J. Chandera, A. Nga, M. C. Wena, A. R. Ceninaa and P. N. Assamc,d a
Department of Neurology, National Neuroscience Institute, Singapore; bDuke NUS Graduate Medical School Singapore, Singapore; c Centre for Quantitative Medicine, Duke NUS Graduate Medical School, Singapore; and dSingapore Clinical Research Institute, Singapore, Singapore
EUROPEAN JOURNAL OF NEUROLOGY
Keywords:
Alzheimer’s disease, medial temporal atrophy, white matter hyperintensity Received 11 April 2014 Accepted 2 July 2014 European Journal of Neurology 2015, 22: 150–155 doi:10.1111/ene.12546
Background and purpose: Whilst there is evidence implicating small vessel cerebrovascular disease in the pathogenesis of Alzheimer’s disease (AD), its specific contribution to the pathophysiology of AD remains unclear. The burden of small vessel cerebrovascular disease visualized as white matter hyperintensity (WMH) and its association with medial temporal atrophy (MTA) at different stages of AD was studied. Methods: One hundred and sixty-five cognitively normal (CN) community controls, 103 mild cognitive impairment (MCI) patients, 141 mild AD patients and 68 moderate severe AD patients were studied. Clinical, cognitive and risk factor data were collected, and WMH and MTA were quantified by trained raters. The Jonckheere Terpstra test for ordered alternatives was used to study the association between WMH and MTA in different stages of AD. Results: The burden of total WMH increased significantly with increasing severity of AD, even after correcting for confounders. The proportion of CN, MCI, mild AD and moderate severe AD subjects with severe burden of WMH was 6.7%, 9.7%, 28.4%, and 39.7%, respectively. A strong positive association between WMH severity and MTA was evident amongst MCI (P = 0.011) and mild AD (P = 0.003) subjects, but not in CN (P = 0.953) and moderate severe AD subjects (P = 0.301). Conclusions: The burden of WMH increased significantly from the stage of CN to MCI to AD. The association between WMH and MTA was greatest at the stage of MCI and mild AD. This has implications on the strategy to slow the progression of AD, where measures to reduce WMH, including control of vascular risk factors, need to be optimized at the stage of MCI and mild AD.
Introduction Alzheimer’s disease (AD) is the most common cause of dementia worldwide, with global prevalence estimated at 3.9% in people older than 60 years [1]. However, AD pathology often coexists with other neurodegenerative and vascular pathology [2]. Clinical studies have reported the burden of cerebrovascular disease to be higher in AD compared with elderly controls [3]. Autopsy series suggest that prevalence of Correspondence: N. Kandiah, Department of Neurology, National Neuroscience Institute, 11 Jalan Tan Tock Seng, Singapore 308433, Singapore (tel.: +65 6357 7199; fax: +65 6357 7137; e-mail: [email protected]).
150
vascular pathology ranges from 8% to 35% [4,5]. Whilst the contribution of cerebral large vessel disease to dementia by means of multi-infarct vascular dementia has been widely studied, studies investigating the contribution of the more prevalent small vessel cerebrovascular disease (svCVD) to the pathogenesis of AD remains inadequate [6,7]. With the rising prevalence of AD and failure of antiamyloid compounds in achieving primary end-points in AD clinical trials, there is an urgent need to explore other biological factors that can delay AD onset and progression [8]. In this regard, the role of svCVD in AD warrants further clarification. Whilst some studies have demonstrated that svCVD accelerates the progression
© 2014 EAN
WHITE MATTER HYPERINTENSITY IN ALZHEIMER’S DISEASE
of mild cognitive impairment (MCI) to dementia, others show that svCVD has no influence on cognition or AD disease progression [9–13]. These conflicting findings on the contribution of svCVD to AD pathogenesis may be related to the non-uniform influence of svCVD at different stages of AD pathogenesis. White matter hyperintensity (WMH) on magnetic resonance imaging (MRI) has been widely accepted as a surrogate marker for svCVD and pathological studies have demonstrated a definite relationship between WMH and arteriolosclerosis [14]. A recent study employing positron emission tomography with fluorodeoxyglucose F18 (FDG PET) demonstrated that, amongst subjects with MCI who converted to AD, WMH correlated with frontal hypometabolism whilst hippocampal atrophy correlated with parietooccipital hypometabolism, suggesting different but synergistic pathways of svCVD and amyloid pathology in the pathogenesis of AD [15]. It is important to study the prevalence of svCVD at different stages of AD. Elucidating the effect of svCVD at different stages of AD would allow more targeted management of the disease [8,16]. To further understand the contribution of svCVD in AD, the prevalence of WMH and the effect of WMH on medial temporal atrophy (MTA) in the different stages of the AD continuum were examined.
Methods Study sample
Patients from a tertiary dementia centre diagnosed with MCI and AD between 2011 and 2013 and cognitively normal (CN) elderly subjects from the Singapore Longitudinal Aging Brain Study (S-LABS) were evaluated [17]. Diagnosis of AD was based on the Diagnostic and Statistical Manual, 4th edition text revision [18], and the National Institute of Neurologic, Communicative Disorders and Stroke – Alzheimer’s Disease and Related Disorders Association criteria [19]. The Clinical Dementia Rating (CDR) scale was used to categorize patients into mild (CDR 1) and moderate severe (CDR 2 3) groups [20]. These criteria were shown to be highly reliable in the clinical diagnosis of AD [16]. MCI was diagnosed based on Petersen’s criteria [21]. These patients reported subjective memory complaints, and their cognitive symptoms were corroborated by a reliable caregiver. They also had impaired memory function for age and education, a CDR of 0.5, and preserved activities of daily living. All diagnoses of MCI and AD were reached by a consensus amongst a clinical team. CN control subjects were recruited
© 2014 EAN
151
from the S-LABS, a community based, convenience sample, longitudinal neuroimaging study involving over 200 cognitively healthy older Singaporeans aged above 55 years. CNs were defined with a CDR score of 0, a Mini-Mental State Examination (MMSE) score >27 and a 15-item Geriatric Depression Scale score 8 on the modified Fazekas scale. The use of coronal T1weighted sequences for the evaluation of MTA and axial T2-weighted sequences for the evaluation of WMH has been recommended for the clinical rating of severity of AD and svCVD [24]. All MRI scoring was performed independently by two experienced raters blinded to the clinical diagnosis. The initial scoring of WMH had high inter-rater agreement, with 85% agreement between the raters and a j of 0.82, and the quantification of MTA had modest inter-rater agreement, with 76% agreement between the raters and a j of 0.71. Any difference in rating scores between the two raters was subsequently resolved by consensus.
152
N. KANDIAH ET AL.
Statistical methods
Demographic, vascular risk factor and neuroimaging parameters were summarized by subject groups, using mean and standard deviation for continuous variables and frequency and percentages for categorical variables. In the univariate analysis, differences between subject groups were analysed by chi-squared test for categorical outcomes and one-way analysis of variance or Kruskal–Wallis test for continuous outcomes. The Jonckheere Terpstra test for ordered alternatives was used to evaluate whether the neuroimaging parameters increase significantly with increasing severity of cognitive impairment across CN, MCI, mild AD and moderate severe AD. For multivariate analysis, ordered logistic regression analysis was used to look at WMH differences across stages of AD whilst correcting for potential confounding variables. WMH was later stratified into three stages of severity: low (total WMH 0–4), medium (total WMH 5–8) and high (total WMH 9–12) burden of WMH. The Jonckheere Terpstra test was performed to study the association between stages of WMH severity and MTA at different stages of AD. All statistical analyses were performed using Stata version 10.1 (StataCorp, College Station, TX, USA) statistical software. All tests were two-tailed and were conducted at the 5% level of significance.
Results From a total of 349 subjects in the S-LABS study, 101 subjects without MRI and 83 subjects with MMSE ≤27 were excluded, leaving 165 elderly CN subjects. Between August 2011 and August 2013, the dementia database included 156 subjects with MCI, 165 subjects with mild AD and 88 subjects with moderate severe AD. Fifty-three MCI subjects (45 without MRI and eight with CDR >0.5), 24 mild AD subjects (all without MRI) and 20 subjects with moderate severe AD (14 without MRI and six without MMSE scores) were excluded. The remaining 477 subjects, consisting of 203 (45.2%) males, with a mean age of 69.8 (SD 7.8) years and mean education of 8.8 (SD 5.1) years, were studied. The cohort comprised 165 (34.6%) elderly CN subjects, 103 (21.6%) subjects with MCI, 141 (29.6%) subjects with mild AD and 68 (14.3%) subjects with moderate severe AD. Patients with AD were significantly older than the MCI and CN subjects. The prevalence of diabetes mellitus and hyperlipidaemia were significantly higher in the mild AD group compared with CN subjects. There were no significant differences in the prevalence of hypertension between groups. The mean MMSE
score significantly decreased from CN to MCI to AD (Table 1). With increasing AD stage, a corresponding significant increase in total WMH, PVH and DSC was demonstrated (Table 1). The burden of severe WMH, defined as total WMH ≥8, amongst CN and MCI subjects was low at 6.7% and 9.7%, respectively, but increased to 28.4% at the stage of mild AD and to 39.7% at the stage of severe AD. Total MTA scores increased significantly with increasing severity of cognitive impairment. Age, gender, education, diabetes mellitus, hypertension, hyperlipidaemia and total MTA were identified in the univariate analysis as potential confounders to be corrected for in multivariate analysis. It was found that total PVH, total WMH and classification of severe WMH were found to remain statistically significant after correction (Table 1). The relationship between mean MTA and WMH severity was also studied in the clinical stages of the AD continuum. A positive and significant association was observed between mean MTA and WMH severity in MCI [mean MTA in low WMH burden, 3.08 (SD 1.41) vs. medium WMH burden, 3.49 (SD 1.90) vs. high WMH burden, 4.60 (SD 1.90); P = 0.011; Fig. 1b] and mild AD [3.80 (SD 1.45) vs. 4.30 (SD 1.58) vs. 4.80 (SD 1.64); P = 0.003; Fig. 1c] subjects. This association was not observed in CN [3.68 (SD 1.68) vs. 3.03 (SD 1.90) vs. 3.73 (SD 1.49); P = 0.953; Fig. 1a] and moderate severe AD [5.29 (SD 2.54) vs. 6.72 (SD 1.90) vs. 6.00 (SD 1.80); P = 0.301; Fig. 1d) subjects.
Discussion The findings from this study demonstrate a high burden of WMH amongst subjects with AD. The burden of severe WMH increased significantly across the continuum of MCI to mild AD to moderate severe AD. The association between WMH and MTA was significant at the stage of MCI and mild AD but not at the stage of moderate severe AD, independent of age and vascular risk factors. The finding of significant association between WMH and MTA only in the earlier stages is novel and may suggest that svCVD may have a synergistic effect with amyloid-tau pathology early in the course of AD pathogenesis. This finding has been echoed in the aforementioned FDG PET study [15]. Another study employing amyloid PET demonstrated that, amongst subjects with MCI, both WMH and amyloid-positive status at baseline conferred risk for future diagnosis of AD [25]. These two PET studies support the hypothesis that silent CVD and
© 2014 EAN
WHITE MATTER HYPERINTENSITY IN ALZHEIMER’S DISEASE
153
Table 1 Demographic, clinical and MRI findings on CN, MCI, mild AD and moderate severe AD subjects
Age (years) Gender (males) Education (years) Diabetes mellitus Hypertension Hyperlipidaemia MMSE Total MTA Total PVH Total DSC Total WMH Severe WMH
Mean (SD) N (%) Mean (SD) N (%) N (%) N (%) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) N (%)
CN (n = 165)
MCI (n = 103)
Mild AD (n = 141)
Moderate severe AD (n = 68)
Univariate P value
Multivariate P value
67.66 77 11.80 25 95 75 28.86 3.42 2.22 2.41 4.64 11
67.83 41 9.35 13 36 27 26.65 3.41 2.30 2.70 5.00 10
71.38 62 6.57 39 81 78 22.63 4.28 3.60 3.02 6.62 40
74.62 23 4.21 13 29 24 11.79 6.02 3.82 3.40 7.22 27
0.0001b,c 0.040 0.0001a,b,c 0.004b 0.185 0.004b 0.0001a,b,c
Association between white matter hyperintensity and medial temporal atrophy at various stages of Alzheimer’s disease N. Kandiaha,b, R. J. Chandera, A. Nga, M. C. Wena, A. R. Ceninaa and P. N. Assamc,d a
Department of Neurology, National Neuroscience Institute, Singapore; bDuke NUS Graduate Medical School Singapore, Singapore; c Centre for Quantitative Medicine, Duke NUS Graduate Medical School, Singapore; and dSingapore Clinical Research Institute, Singapore, Singapore
EUROPEAN JOURNAL OF NEUROLOGY
Keywords:
Alzheimer’s disease, medial temporal atrophy, white matter hyperintensity Received 11 April 2014 Accepted 2 July 2014 European Journal of Neurology 2015, 22: 150–155 doi:10.1111/ene.12546
Background and purpose: Whilst there is evidence implicating small vessel cerebrovascular disease in the pathogenesis of Alzheimer’s disease (AD), its specific contribution to the pathophysiology of AD remains unclear. The burden of small vessel cerebrovascular disease visualized as white matter hyperintensity (WMH) and its association with medial temporal atrophy (MTA) at different stages of AD was studied. Methods: One hundred and sixty-five cognitively normal (CN) community controls, 103 mild cognitive impairment (MCI) patients, 141 mild AD patients and 68 moderate severe AD patients were studied. Clinical, cognitive and risk factor data were collected, and WMH and MTA were quantified by trained raters. The Jonckheere Terpstra test for ordered alternatives was used to study the association between WMH and MTA in different stages of AD. Results: The burden of total WMH increased significantly with increasing severity of AD, even after correcting for confounders. The proportion of CN, MCI, mild AD and moderate severe AD subjects with severe burden of WMH was 6.7%, 9.7%, 28.4%, and 39.7%, respectively. A strong positive association between WMH severity and MTA was evident amongst MCI (P = 0.011) and mild AD (P = 0.003) subjects, but not in CN (P = 0.953) and moderate severe AD subjects (P = 0.301). Conclusions: The burden of WMH increased significantly from the stage of CN to MCI to AD. The association between WMH and MTA was greatest at the stage of MCI and mild AD. This has implications on the strategy to slow the progression of AD, where measures to reduce WMH, including control of vascular risk factors, need to be optimized at the stage of MCI and mild AD.
Introduction Alzheimer’s disease (AD) is the most common cause of dementia worldwide, with global prevalence estimated at 3.9% in people older than 60 years [1]. However, AD pathology often coexists with other neurodegenerative and vascular pathology [2]. Clinical studies have reported the burden of cerebrovascular disease to be higher in AD compared with elderly controls [3]. Autopsy series suggest that prevalence of Correspondence: N. Kandiah, Department of Neurology, National Neuroscience Institute, 11 Jalan Tan Tock Seng, Singapore 308433, Singapore (tel.: +65 6357 7199; fax: +65 6357 7137; e-mail: [email protected]).
150
vascular pathology ranges from 8% to 35% [4,5]. Whilst the contribution of cerebral large vessel disease to dementia by means of multi-infarct vascular dementia has been widely studied, studies investigating the contribution of the more prevalent small vessel cerebrovascular disease (svCVD) to the pathogenesis of AD remains inadequate [6,7]. With the rising prevalence of AD and failure of antiamyloid compounds in achieving primary end-points in AD clinical trials, there is an urgent need to explore other biological factors that can delay AD onset and progression [8]. In this regard, the role of svCVD in AD warrants further clarification. Whilst some studies have demonstrated that svCVD accelerates the progression
© 2014 EAN
WHITE MATTER HYPERINTENSITY IN ALZHEIMER’S DISEASE
of mild cognitive impairment (MCI) to dementia, others show that svCVD has no influence on cognition or AD disease progression [9–13]. These conflicting findings on the contribution of svCVD to AD pathogenesis may be related to the non-uniform influence of svCVD at different stages of AD pathogenesis. White matter hyperintensity (WMH) on magnetic resonance imaging (MRI) has been widely accepted as a surrogate marker for svCVD and pathological studies have demonstrated a definite relationship between WMH and arteriolosclerosis [14]. A recent study employing positron emission tomography with fluorodeoxyglucose F18 (FDG PET) demonstrated that, amongst subjects with MCI who converted to AD, WMH correlated with frontal hypometabolism whilst hippocampal atrophy correlated with parietooccipital hypometabolism, suggesting different but synergistic pathways of svCVD and amyloid pathology in the pathogenesis of AD [15]. It is important to study the prevalence of svCVD at different stages of AD. Elucidating the effect of svCVD at different stages of AD would allow more targeted management of the disease [8,16]. To further understand the contribution of svCVD in AD, the prevalence of WMH and the effect of WMH on medial temporal atrophy (MTA) in the different stages of the AD continuum were examined.
Methods Study sample
Patients from a tertiary dementia centre diagnosed with MCI and AD between 2011 and 2013 and cognitively normal (CN) elderly subjects from the Singapore Longitudinal Aging Brain Study (S-LABS) were evaluated [17]. Diagnosis of AD was based on the Diagnostic and Statistical Manual, 4th edition text revision [18], and the National Institute of Neurologic, Communicative Disorders and Stroke – Alzheimer’s Disease and Related Disorders Association criteria [19]. The Clinical Dementia Rating (CDR) scale was used to categorize patients into mild (CDR 1) and moderate severe (CDR 2 3) groups [20]. These criteria were shown to be highly reliable in the clinical diagnosis of AD [16]. MCI was diagnosed based on Petersen’s criteria [21]. These patients reported subjective memory complaints, and their cognitive symptoms were corroborated by a reliable caregiver. They also had impaired memory function for age and education, a CDR of 0.5, and preserved activities of daily living. All diagnoses of MCI and AD were reached by a consensus amongst a clinical team. CN control subjects were recruited
© 2014 EAN
151
from the S-LABS, a community based, convenience sample, longitudinal neuroimaging study involving over 200 cognitively healthy older Singaporeans aged above 55 years. CNs were defined with a CDR score of 0, a Mini-Mental State Examination (MMSE) score >27 and a 15-item Geriatric Depression Scale score 8 on the modified Fazekas scale. The use of coronal T1weighted sequences for the evaluation of MTA and axial T2-weighted sequences for the evaluation of WMH has been recommended for the clinical rating of severity of AD and svCVD [24]. All MRI scoring was performed independently by two experienced raters blinded to the clinical diagnosis. The initial scoring of WMH had high inter-rater agreement, with 85% agreement between the raters and a j of 0.82, and the quantification of MTA had modest inter-rater agreement, with 76% agreement between the raters and a j of 0.71. Any difference in rating scores between the two raters was subsequently resolved by consensus.
152
N. KANDIAH ET AL.
Statistical methods
Demographic, vascular risk factor and neuroimaging parameters were summarized by subject groups, using mean and standard deviation for continuous variables and frequency and percentages for categorical variables. In the univariate analysis, differences between subject groups were analysed by chi-squared test for categorical outcomes and one-way analysis of variance or Kruskal–Wallis test for continuous outcomes. The Jonckheere Terpstra test for ordered alternatives was used to evaluate whether the neuroimaging parameters increase significantly with increasing severity of cognitive impairment across CN, MCI, mild AD and moderate severe AD. For multivariate analysis, ordered logistic regression analysis was used to look at WMH differences across stages of AD whilst correcting for potential confounding variables. WMH was later stratified into three stages of severity: low (total WMH 0–4), medium (total WMH 5–8) and high (total WMH 9–12) burden of WMH. The Jonckheere Terpstra test was performed to study the association between stages of WMH severity and MTA at different stages of AD. All statistical analyses were performed using Stata version 10.1 (StataCorp, College Station, TX, USA) statistical software. All tests were two-tailed and were conducted at the 5% level of significance.
Results From a total of 349 subjects in the S-LABS study, 101 subjects without MRI and 83 subjects with MMSE ≤27 were excluded, leaving 165 elderly CN subjects. Between August 2011 and August 2013, the dementia database included 156 subjects with MCI, 165 subjects with mild AD and 88 subjects with moderate severe AD. Fifty-three MCI subjects (45 without MRI and eight with CDR >0.5), 24 mild AD subjects (all without MRI) and 20 subjects with moderate severe AD (14 without MRI and six without MMSE scores) were excluded. The remaining 477 subjects, consisting of 203 (45.2%) males, with a mean age of 69.8 (SD 7.8) years and mean education of 8.8 (SD 5.1) years, were studied. The cohort comprised 165 (34.6%) elderly CN subjects, 103 (21.6%) subjects with MCI, 141 (29.6%) subjects with mild AD and 68 (14.3%) subjects with moderate severe AD. Patients with AD were significantly older than the MCI and CN subjects. The prevalence of diabetes mellitus and hyperlipidaemia were significantly higher in the mild AD group compared with CN subjects. There were no significant differences in the prevalence of hypertension between groups. The mean MMSE
score significantly decreased from CN to MCI to AD (Table 1). With increasing AD stage, a corresponding significant increase in total WMH, PVH and DSC was demonstrated (Table 1). The burden of severe WMH, defined as total WMH ≥8, amongst CN and MCI subjects was low at 6.7% and 9.7%, respectively, but increased to 28.4% at the stage of mild AD and to 39.7% at the stage of severe AD. Total MTA scores increased significantly with increasing severity of cognitive impairment. Age, gender, education, diabetes mellitus, hypertension, hyperlipidaemia and total MTA were identified in the univariate analysis as potential confounders to be corrected for in multivariate analysis. It was found that total PVH, total WMH and classification of severe WMH were found to remain statistically significant after correction (Table 1). The relationship between mean MTA and WMH severity was also studied in the clinical stages of the AD continuum. A positive and significant association was observed between mean MTA and WMH severity in MCI [mean MTA in low WMH burden, 3.08 (SD 1.41) vs. medium WMH burden, 3.49 (SD 1.90) vs. high WMH burden, 4.60 (SD 1.90); P = 0.011; Fig. 1b] and mild AD [3.80 (SD 1.45) vs. 4.30 (SD 1.58) vs. 4.80 (SD 1.64); P = 0.003; Fig. 1c] subjects. This association was not observed in CN [3.68 (SD 1.68) vs. 3.03 (SD 1.90) vs. 3.73 (SD 1.49); P = 0.953; Fig. 1a] and moderate severe AD [5.29 (SD 2.54) vs. 6.72 (SD 1.90) vs. 6.00 (SD 1.80); P = 0.301; Fig. 1d) subjects.
Discussion The findings from this study demonstrate a high burden of WMH amongst subjects with AD. The burden of severe WMH increased significantly across the continuum of MCI to mild AD to moderate severe AD. The association between WMH and MTA was significant at the stage of MCI and mild AD but not at the stage of moderate severe AD, independent of age and vascular risk factors. The finding of significant association between WMH and MTA only in the earlier stages is novel and may suggest that svCVD may have a synergistic effect with amyloid-tau pathology early in the course of AD pathogenesis. This finding has been echoed in the aforementioned FDG PET study [15]. Another study employing amyloid PET demonstrated that, amongst subjects with MCI, both WMH and amyloid-positive status at baseline conferred risk for future diagnosis of AD [25]. These two PET studies support the hypothesis that silent CVD and
© 2014 EAN
WHITE MATTER HYPERINTENSITY IN ALZHEIMER’S DISEASE
153
Table 1 Demographic, clinical and MRI findings on CN, MCI, mild AD and moderate severe AD subjects
Age (years) Gender (males) Education (years) Diabetes mellitus Hypertension Hyperlipidaemia MMSE Total MTA Total PVH Total DSC Total WMH Severe WMH
Mean (SD) N (%) Mean (SD) N (%) N (%) N (%) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) N (%)
CN (n = 165)
MCI (n = 103)
Mild AD (n = 141)
Moderate severe AD (n = 68)
Univariate P value
Multivariate P value
67.66 77 11.80 25 95 75 28.86 3.42 2.22 2.41 4.64 11
67.83 41 9.35 13 36 27 26.65 3.41 2.30 2.70 5.00 10
71.38 62 6.57 39 81 78 22.63 4.28 3.60 3.02 6.62 40
74.62 23 4.21 13 29 24 11.79 6.02 3.82 3.40 7.22 27
0.0001b,c 0.040 0.0001a,b,c 0.004b 0.185 0.004b 0.0001a,b,c
