Nelson, Peter T. MD, PhD; Kukull, Walter A. PhD; Frosch, Matthew P. MD, PhD
From the Department of Pathology and Division of Neuropathology (PTN), University of Kentucky Medical Center and Sanders-Brown Center on Aging, University of Kentucky, Lexington, Kentucky; Department of Epidemiology (WAK), University of Washington, Seattle, Washington; and C.S. Kubik Laboratory for Neuropathology (MPF), Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts.
Send correspondence and reprint requests to: Peter T. Nelson, MD, PhD, Department of Pathology, Division of Neuropathology and the Sanders-Brown Center on Aging, Rm. 311, Sanders-Brown Building, 800 S. Limestone, University of Kentucky, Lexington, KY 40536-0230; E-mail: firstname.lastname@example.org
This study was supported by Grant Nos. R01 NS061933, K08 NS050110, P30 AG028383, P50 AG005134, and U01 AG016976 from the National Institutes of Health and Grant No. NIRG (89917) from Alzheimer's Association.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.jneuropath.com).
The most recent consensus guidelines for the neuropathologic diagnosis of Alzheimer disease (AD) were the National Institute on Aging and Reagan Institute (NIA-RI) recommendations that were published in 1997 (1). These guidelines stated that neuropathology is an absolute requirement "to provide an estimate of the likelihood that Alzheimer's disease pathological changes underlie dementia" (1). This consensus report also concluded that more modifications may be necessary, and that it is important to "validate and refine the procedures recommended above" (1). In the years since these guidelines were published, clinical diagnostic arsenals have expanded dramatically with improvements in neuroimaging and cerebrospinal fluid tests. Despite these advancements, the need for validating and refining neuropathologic practices, which remain the criterion standard for the diagnosis of AD and other neurodegenerative diseases, have not diminished.
There are 2 distinct "pathologic hallmarks" of AD: neurofibrillary tangles (NFTs) and neuritic amyloid plaques (NPs). Neurofibrillary tangles develop intracellularly and are composed of filamentous tau protein polymers. The severity of NFT pathology is graded on a 0-to-6 scale (using Roman numerals 0-VI by convention) according to "Braak stages," which pertain to the spread of NFTs in the brain (2). In contrast, NPs are extracellular amyloid deposits surrounded by argyrophilic degenerating neurites. The severity of NP pathology is scored according to a distinct metric named after the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) (3). The CERAD scoring system is a 4-tiered scale representing neocortical NP density. Neither diffuse amyloid plaques nor cerebral amyloid angiopathy are relevant to current AD pathologic diagnostic criteria.
According to Braak staging and the CERAD scale, cases are parsed into "low," "intermediate," and "high" likelihood that clinical dementia is due to AD. "Low" likelihood corresponds to CERAD "infrequent," Braak Stages 0 to II. "Intermediate likelihood" corresponds to CERAD "moderate," Braak Stages III or IV. "High likelihood" corresponds to CERAD "frequent," Braak Stages V or VI. Combinations of pairings of plaques and tangles off the diagonal are not clearly addressed in the NIA-RI criteria (Fig. 1).
Most human brains can be categorized readily among the NIA-RI diagnostic groups. However, a surprisingly high fraction of cases (∼18% of persons with antemortem diagnosis of dementia and without strong concomitant pathologies in the present study) fall outside of the NIA-RI rubric. To study these diagnostically problematic cases, we used the National Alzheimer's Coordinating Center (NACC) Registry database (4). This data set includes annotated clinical and pathologic information from 30 different Alzheimer Disease Centers (ADC) for thousands of individuals who had come to autopsy. The goal of the study was to track how neuropathologists diagnose AD using the CERAD and Braak staging criteria. We also sought to address whether the documented cognitive impairment in patients could provide some new information to help guide diagnostic neuropathologic practice.
MATERIALS AND METHODS
The NACC Registry contains data obtained from 30 different ADCs (4), the institutions from which cases derived are listed in Table, Supplemental Digital Content 1 (http://links.lww.com/NEN/A95). The data include 3,501 cases for the cohort with dementia and 3,822 cases for the cohort used to assess averaged final Mini-Mental State Examination (MMSE) scores. Many of the subjects had been followed with annual neurologic examinations up to the point of death; average intervals between the final MMSE scores and death were less than 1.5 years and a median interval less than 1 year.
In addition to previously published exclusion criteria for NACC data (5, 6), additional exclusion criteria were imposed for this study. Cases with death before 1999, no data regarding education level, no neuropathology, or clinical history of prion disease, synucleinopathy, triplet repeat disease, brain cancer, or frontotemporal dementia that might explain a dementia syndrome with early death were excluded. Demographic and neuropathologic data on cases from the NACC Registry that were included are summarized in Table 1.
We analyzed 2 separate case groups. The first cohort (n = 1,672; "inclusive criteria"; Table 1, left column) was used to test the diagnostic practices of ADC neuropathologists; all of these patients had the antemortem diagnosis of dementia. This criterion was necessary because the NIA-RI criteria were developed exclusively for this clinical context (1). The second cohort (n = 1,350, "high-stringency criteria"; Table 1, right column) was used for estimating the associative impact of AD-type pathology. There was no selection on the basis of antemortem dementia diagnosis for this group because we wanted to correlate outcomes along the entire continuum of cognitive impairment. More rigorous exclusion criteria were used for concomitant neuropathology for this group, however. Only patients with no known stroke history and with neuropathologic designation of "no Lewy bodies" (to exclude the possibility of bias of synucleinopathies) were included to isolate the specific associative effects of AD-type pathology. Neocortical Lewy bodies were determined by indication of "diffuse" or "intermediate" cortical Lewy body pathology using the NACC registry based on consensus diagnostic features (7). Statistical analyses were performed using 2-tailed Student t-tests.
To determine the diagnoses given by the ADC neuropathologists for technically unclassifiable cases, separate analyses were performed on cases that had been given the diagnoses of "high likelihood," "intermediate likelihood," and "low likelihood" of AD changes. The cases of primary interest were Braak Stages III/IV and "frequent" NPs as specified by CERAD ("plaque-intensive" cases), and those with Braak Stages V/VI and "moderate" NPs by CERAD ("tangle-intensive" cases). Statistical analyses were performed using 2-tailed Student t-tests.
The Braak staging and CERAD assessments of cases in the NACC Registry from demented subjects are shown in Table 2. Of the 1,672 cases evaluated, 1,378 (82.4%) fell into diagnostic "boxes" that are within the rubric of the consensus recommendations (Table 2; Fig. 1). Thus, approximately 18% of cases fell outside the NIA-RI diagnostic categories (i.e. "off-diagonal"). Table 2 also reveals that subjects coming to autopsy were skewed toward high burdens of both plaques and tangles.
The cases of interest, that is, Braak Stages III/IV and "frequent" NPs as specified by CERAD (plaque-intensive cases), and those with Braak Stages V/VI and "moderate" NPs by CERAD (tangle-intensive cases; corresponding to red shaded areas in Fig. 1) comprised 9.4% of the cases overall. This was a higher percentage than those meeting the intermediate likelihood criteria, which represented only 6.0% of cases in the database (partly due to restriction to demented patients).
Cases that had been given diagnoses of high likelihood (Table 3A), intermediate likelihood (Table 3B), and low likelihood (Table 3C) of AD changes showed a high concordance between the NIA-Reagan recommendations and neuropathologic criteria as illustrated in Figure 1. Cases that matched with the consensus recommendations are given the appropriate diagnoses with great consistency: 93.1% for intermediate-likelihood cases and 97.7% of high-likelihood cases are appropriate between the Braak, CERAD, and NIA-RI diagnoses (Table 4).
For cases outside the NIA-RI rubric, the consistency in terms of diagnostic categories was relatively low, but there are some notable tendencies. Persons with tangle-intensive pathology are far more likely to be diagnosed as high likelihood for AD than plaque-intensive pathology (56.2% vs 22.4%; Table 4). Conversely, 70.6% of plaque-intensive cases were designated intermediate likelihood via NIA-RI as opposed to 32.9% for tangle-intensive cases (Table 4).
Because the cases derived from different research centers, we sought to understand how an individual ADC might affect the overall result. The ADCs are listed in Table, Supplemental Digital Content 2 (http://links.lww.com/NEN/A96) according to the number of cases in the NACC Registry. Because no individual ADC comprised more than 20% of the cases in any particular category, it is unlikely that the diagnostic tendencies of an individual ADC drove the overall result.
To assess whether plaque-intensive or tangle-intensive neuropathologic features had more effect on cognition, we evaluated average final MMSE scores of those cases. For these analyses, 2 important factors were relevant to the inclusion criteria. First, we could not restrict the analyses to only those individuals with the antemortem diagnosis of dementia because we were trying to understand the clinicopathologic correlation of the diagnostic categories. Thus, if 1 category had "nondemented" patients, that would be quite relevant, and we did not wish to bias the results by excluding them a priori. Second, because we were interested in understanding the associative impact of AD-type pathology for this group, we had more rigorous exclusion criteria for "mixed" pathology such as excluding cases with stroke or possible synucleinopathies. The results of the averaged final MMSE scores are shown in Table 5. Results of 2-tailed Student t-tests to examine differences among mean final MMSE scores are shown in Table 6. Further tests were performed to evaluate final MMSE score distributions (rather than means); these did not provide additional information and are not shown.
Using NACC Registry data, we addressed some of the issues pertinent to cases that fall outside of the explicit recommendations of the NIA-RI Working Group. We found that approximately 18% of patients with antemortem dementia diagnoses fall outside of NIA-RI explicit categories, and almost 10% were either plaque-intensive or tangle-intensive high-pathology cases. Alzheimer Disease Centers neuropathologists tended to diagnose plaque-intensive cases as being intermediate likelihood for AD, whereas tangle-intensive cases tend to be placed in the high-likelihood category. These diagnostic tendencies are partially supported by an analysis of averaged final MMSE scores.
National Alzheimer's Coordinating Center Registry data derive from 30 different ADCs with differing demographics and recruitment criteria (5). As with all large databases, there is presumably some error rate in the classification schemes due to human and/or technical factors. For example, there is variability in the standards of AD severity scoring (e.g. Braak staging) among pathologists (8, 9). We also had to exclude many of the dementia cases in the NACC Registry data set due to concomitant pathologies other than AD (e.g. synucleinopathies), and only individuals who died after 1999 were included to ensure that current practices were being tested. It would be optimal to understand how these case categories are valid even in cases with "mixed" pathologies. Furthermore, the NACC registry data are generally derived from samples of convenience that may introduce some bias in case selection (10, 11). This is at least partly reflected by the large skew in the data toward advanced clinical and pathologic AD. Naturally, these data are also somewhat skewed by the fact that autopsy series tend to identify the latest stages in the disease. Finally, there are inevitably differences in the neuropathologic methods and practices among ADCs because there are no current consensus recommendations on how to process tissue for NPs and NFTs (12, 13). Despite these caveats, the number of cases included was quite large, and post hoc studies that included less stringent inclusion criteria had essentially similar results (not shown).
The first aim of this study was to query how ADC neuropathologists diagnose dementia cases in which CERAD and Braak stage parameters fall outside the current consensus guidelines. Tangle-intensive cases were more likely than plaque-intensive cases to be designated as high likelihood for AD. The cumulative tendencies of ADC neuropathologists may reflect the fact that clinicopathologic studies have provided stronger support for the pathologic impact of NFTs than amyloid plaques, including NPs (14-16). There are, however, compelling data in the literature to support the hypothesis that NPs also contribute, albeit to a lesser extent, to cognitive impairment in aged individuals (11, 17, 18). Addressing these issues is beyond the scope of the current study, however. In contrast to the case categories outside of the NIA-RI recommendations, the cases where the Braak stage and CERAD score are inside the guidelines the ADC neuropathologists were well greater than 90% consistent in their final diagnoses. These data seem to emphasize the importance of consensus guidelines such as NIA-RI.
A second goal of this study was to determine whether clinicopathologic correlation could provide insights into the global cognitive impairment (as quantified with final MMSE scores) associated with the various pathologic diagnostic categories. Presumably, this could be used to provide some guidance regarding the diagnostic implications of those categories. To address this question, inclusion criteria were adjusted to exclude mixed pathologies but to not limit the evaluation to clinically demented individuals. These results offer incomplete support for the prevalent practice of designating tangle-intensive cases as high likelihood for AD. More specifically, the average final MMSE scores for the tangle-intensive cases as a group are not lower than those of the plaque-intensive cases (there was a trend in that direction, but it was not statistically significant). The final MMSE scores for the tangle-intensive cases were indeed lower than the final MMSE scores for the intermediate-likelihood group. Furthermore, the tangle-intensive cases that are Braak Stage VI, and which comprise 1.3% of the demented patient cohort, seem to have final MMSE scores that approximate severe AD. This result is not unexpected because there is a large difference in the cognitive status in individuals without concomitant pathologies comparing Braak Stage V and VI patients (15, 19).
There have been previous descriptions of cases that fall outside the NIA-RI recommendations. For example, it was suggested that cases we designated as plaque-intensive and tangle-intensive cases all belong in the category of intermediate likelihood for AD (20). Cases have been described with moderate NFT pathology but very minimal plaque pathology, and these may not belong on the AD continuum (21). Nonetheless, there are many unanswered questions. Current neuropathologic methods are oriented toward making a fixed diagnosis in subjects at various stages of a disease. It will be a challenge to integrate future neuropathologic rubrics with the neuropsychologic testing, neuroimaging, and cerebrospinal fluid tests for earlier diagnosis of the disease. Diagnostic modalities are also challenged to account for the substantial numbers of "mixed pathology" cases that were excluded in the current study. Definitive recommendations that surmount these challenges, or that at least represent a diagnostic standard for the field, await future consensus guidelines. These guidelines may help integrate the growing literature regarding the manifestations of AD in the earlier stages of the disease.
The authors thank the patients and their families who participated in the studies. This study grew out of a discussion, chaired by Matthew P. Frosch, at the ADC NP Core Meeting, and the stimulating input by our colleagues was appreciated. The authors also thank Ms. Erin L. Abner, MPH, for advice in statistics.
1. Consensus recommendations for the postmortem diagnosis of Alzheimer's disease. The National Institute on Aging, and Reagan Institute Working Group on Diagnostic Criteria for the Neuropathological Assessment of Alzheimer's Disease. Neurobiol Aging 1997;18:S1-2
2. Braak H, Braak E. Stageing of Alzheimer's disease-related neurofibrillary changes. Neurobiol Aging 1995;16:271-78; discussion 8-84
3. Mirra SS. The CERAD neuropathology protocol and consensus recommendations for the postmortem diagnosis of Alzheimer's disease: A commentary. Neurobiol Aging 1997;18:S91-94
4. Beekly DL, Ramos EM, van Belle G, et al. The National Alzheimer's Coordinating Center (NACC) Database: An Alzheimer disease database. Alzheimer Dis Assoc Disord 2004;18:270-77
5. Nelson PT, Jicha GA, Kryscio RJ, et al. Low sensitivity in clinical diagnoses of dementia with Lewy bodies. J Neurol 2010;257:359-66
6. Nelson PT, Kryscio RJ, Jicha GA, et al. Relative preservation of MMSE scores in autopsy-proven dementia with Lewy bodies. Neurology 2009;73:1127-33
7. McKeith IG, Dickson DW, Lowe J, et al. Diagnosis and management of dementia with Lewy bodies: Third report of the DLB Consortium. Neurology 2005;65:1863-72
8. Halliday G, Ng T, Rodriguez M, Harding A, et al. Consensus neuropathological diagnosis of common dementia syndromes: Testing and standardising the use of multiple diagnostic criteria. Acta Neuropathol 2002;104:72-78
9. Alafuzoff I, Arzberger T, Al-Sarraj S, et al. Staging of neurofibrillary pathology in Alzheimer's disease: A study of the BrainNet Europe Consortium. Brain Pathol 2008;18:484-96
10. Schneider JA, Aggarwal NT, Barnes L, et al. The neuropathology of older persons with and without dementia from community versus clinic cohorts. J Alzheimers Dis 2009. Epub ahead of print
11. Nelson PT, Abner EL, Schmitt FA, et al. Modeling the association between 43 different clinical and pathological variables and the severity of cognitive impairment in a large autopsy cohort of elderly persons. Brain Pathol November 19, 2008. [Epub ahead of print]
12. Alafuzoff I, Pikkarainen M, Al-Sarraj S, et al. Interlaboratory comparison of assessments of Alzheimer disease-related lesions: A study of the BrainNet Europe Consortium. J Neuropathol Exp Neurol 2006;65:740-57
13. Alafuzoff I, Thal DR, Arzberger T, et al. Assessment of beta-amyloid deposits in human brain: A study of the BrainNet Europe Consortium. Acta Neuropathol 2009;117:309-20
14. Arriagada PV, Growdon JH, Hedley-Whyte ET, et al. Neurofibrillary tangles but not senile plaques parallel duration and severity of Alzheimer's disease. Neurology 1992;42:631-39
15. Nelson PT, Braak H, Markesbery WR. Neuropathology and cognitive impairment in Alzheimer disease: A complex but coherent relationship. J Neuropathol Exp Neurol 2009;68:1-14
16. Ingelsson M, Fukumoto H, Newell KL, et al. Early Aβ accumulation and progressive synaptic loss, gliosis, and tangle formation in AD brain. Neurology 2004;62:925-31
17. Nagy Z, Esiri MM, Jobst KA, et al. Relative roles of plaques and tangles in the dementia of Alzheimer's disease: Correlations using three sets of neuropathological criteria. Dementia (Basel) 1995;6:21-31
18. Tiraboschi P, Hansen LA, Thal LJ, et al. The importance of neuritic plaques and tangles to the development and evolution of AD. Neurology 2004;62:1984-89
19. Whitwell JL, Josephs KA, Murray ME, et al. MRI correlates of neurofibrillary tangle pathology at autopsy: A voxel-based morphometry study. Neurology 2008;71:743-49
20. Jellinger KA. Criteria for the neuropathological diagnosis of dementing disorders: Routes out of the swamp? Acta Neuropathol 2009;117:101-10
21. Nelson PT, Abner EL, Schmitt FA, et al. Brains with medial temporal lobe neurofibrillary tangles but no neuritic amyloid plaques are a diagnostic dilemma but may have pathogenetic aspects distinct from Alzheimer disease. J Neuropathol Exp Neurology 2009;68:774-84