Demographic, clinical, biomarker, and neuropathological correlates of posterior cortical atrophy: an international cohort study and individual participant data meta-analysis

Abstract

Background: Posterior cortical atrophy is a rare syndrome characterised by early, prominent, and progressive impairment in visuoperceptual and visuospatial processing. The disorder has been associated with underlying neuropathological features of Alzheimer's disease, but large-scale biomarker and neuropathological studies are scarce. We aimed to describe demographic, clinical, biomarker, and neuropathological correlates of posterior cortical atrophy in a large international cohort.

Methods: We searched PubMed between database inception and Aug 1, 2021, for all published research studies on posterior cortical atrophy and related terms. We identified research centres from these studies and requested deidentified, individual participant data (published and unpublished) that had been obtained at the first diagnostic visit from the corresponding authors of the studies or heads of the research centres. Inclusion criteria were a clinical diagnosis of posterior cortical atrophy as defined by the local centre and availability of Alzheimer's disease biomarkers (PET or CSF), or a diagnosis made at autopsy. Not all individuals with posterior cortical atrophy fulfilled consensus criteria, being diagnosed using centre-specific procedures or before development of consensus criteria. We obtained demographic, clinical, biofluid, neuroimaging, and neuropathological data. Mean values for continuous variables were combined using the inverse variance meta-analysis method; only research centres with more than one participant for a variable were included. Pooled proportions were calculated for binary variables using a restricted maximum likelihood model. Heterogeneity was quantified using I².

Findings: We identified 55 research centres from 1353 papers, with 29 centres responding to our request. An additional seven centres were recruited by advertising via the Alzheimer's Association. We obtained data for 1092 individuals who were evaluated at 36 research centres in 16 countries, the other sites having not responded to our initial invitation to participate to the study. Mean age at symptom onset was 59·4 years (95% CI 58·9-59·8; I²=77%), 60% (56-64; I²=35%) were women, and 80% (72-89; I²=98%) presented with posterior cortical atrophy pure syndrome. Amyloid β in CSF (536 participants from 28 centres) was positive in 81% (95% CI 75-87; I²=78%), whereas phosphorylated tau in CSF (503 participants from 29 centres) was positive in 65% (56-75; I²=87%). Amyloid-PET (299 participants from 24 centres) was positive in 94% (95% CI 90-97; I²=15%), whereas tau-PET (170 participants from 13 centres) was positive in 97% (93-100; I²=12%). At autopsy (145 participants from 13 centres), the most frequent neuropathological diagnosis was Alzheimer's disease (94%, 95% CI 90-97; I²=0%), with common co-pathologies of cerebral amyloid angiopathy (71%, 54-88; I²=89%), Lewy body disease (44%, 25-62; I²=77%), and cerebrovascular injury (42%, 24-60; I²=88%).

Interpretation: These data indicate that posterior cortical atrophy typically presents as a pure, young-onset dementia syndrome that is highly specific for underlying Alzheimer's disease pathology. Further work is needed to understand what drives cognitive vulnerability and progression rates by investigating the contribution of sex, genetics, premorbid cognitive strengths and weaknesses, and brain network integrity.

Funding: None.

Figure 1

Frequencies of core features and involvement of additional cognitive or neuropsychiatric domains at the first diagnostic visit. Biomarker findings are reported in figure 2 (site-level data are in the appendix pp 20–27). When reported, CSF amyloid β was positive (ie, in the range consistent with underlying brain amyloid or tau deposition) in 81% (95% CI 75–87) of participants and CSF p-tau was positive in 65% (56–75), but findings were heterogeneous across research centres (I2 >75%). Amyloid-PET and tau-PET were positive for most participants (amyloid-PET, 94% [95% CI 90–97]; tau-PET 97% [93–100]), and heterogeneity statistics were low (I2 ≤15%). Predominant posterior cortical atrophy on MRI was found in 85% (79–91) of participants, and predominant posterior [18F]FDG-PET hypometabolism was reported for 97% (95–98). DaT-SPECT scan results were reported in a small subsample (72 participants from 15 research centres) and showed evidence of nigrostriatal loss in 51% (95% CI 33–69) of participants.

Figure 2

Frequency of biomarker abnormality. To assess associations between Alzheimer’s disease biomarker results and clinical variables, 689 participants from 30 research centres who had either amyloid-PET or CSF amyloid biomarker data (or both) were classified as amyloidpositive when at least one of the biomarkers was positive, and as amyloid-negative when both markers (or the only available marker) were negative. Group comparisons are shown in table 2. Patients who were amyloid-positive were more likely to have posterior cortical atrophy pure than were amyloid-negative patients (95% vs 81%, p<0·0001). Age at symptom onset and age at diagnosis, sex, and other clinical features did not differ by amyloid status.

Figure 3

Frequency of neuropathological findings in the sample. Only ten participants from six research centres had a primary neuropathological diagnosis that was not Alzheimer’s disease (table 3). Four had a primary neuropathological diagnosis of Lewy body disease, all of whom also had significant levels of Alzheimer’s disease neuropathological changes. Other primary diagnoses were frontotemporal lobar degeneration with non-Alzheimer’s disease tauopathy (n=3, two with corticobasal degeneration and one with Pick’s disease) or TDP-43 type A (n=1, due to a pathogenic granulin [GRN] mutation), and brain infarct with minimal comorbid Alzheimer’s disease neuropathology (n=2; both cases had late age at onset [ages 88 years and 90 years] and were from the same research centre).