Influence of Different Diagnostic Criteria on Alzheimer Disease Clinical Research
Andrei Bieger,Wagner S Brum,Wyllians V Borelli,Joseph Therriault,Marco A De Bastiani,Amanda G Moreira,Andrea L Benedet,João Pedro Ferrari-Souza,Jaderson C Da Costa,Diogo O Souza,Raphael M Castilhos,Artur Francisco Schumacher Schuh,Marcia L Fagundes Chaves,Michael Schöll,Henrik Zetterberg,Kaj Blennow,Tharick A Pascoal,Serge Gauthier,Pedro Rosa-Neto,Lucas P Schilling,Eduardo R Zimmer,Alzheimer's Disease Neuroimaging Initiative
DOI: https://doi.org/10.1212/WNL.0000000000209753
IF: 9.9
2024-09-10
Neurology
Abstract:Background and objectives: Updates in Alzheimer disease (AD) diagnostic guidelines by the National Institute on Aging-Alzheimer's Association (NIA-AA) and the International Working Group (IWG) over the past 11 years may affect clinical diagnoses. We assessed how these guidelines affect clinical AD diagnosis in a cohort of cognitively unimpaired (CU) and cognitively impaired (CI) individuals. Methods: We applied clinical and biomarker data in algorithms to classify individuals from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort according to the following diagnostic guidelines for AD: 2011 NIA-AA, 2016 IWG-2, 2018 NIA-AA, and 2021 IWG-3, assigning the following generic diagnostic labels: (1) not AD (nAD), (2) increased risk of developing AD (irAD), and (3) AD. Diagnostic labels were compared according to their frequency, convergence across guidelines, biomarker profiles, and prognostic value. We also evaluated the diagnostic discordance among the criteria. Results: A total of 1,195 individuals (mean age 73.2 ± 7.2 years, mean education 16.1 ± 2.7, 44.0% female) presented different repartitions of diagnostic labels according to the 2011 NIA-AA (nAD = 37.8%, irAD = 23.0%, AD = 39.2%), 2016 IWG-2 (nAD = 37.7%, irAD = 28.7%, AD = 33.6%), 2018 NIA-AA (nAD = 40.7%, irAD = 9.3%, AD = 50.0%), and 2021 IWG-3 (nAD = 51.2%, irAD = 8.4%, AD = 48.3%) frameworks. Discordant diagnoses across all guidelines were found in 512 participants (42.8%) (138 [91.4%] occurring in only β-amyloid [CU 65.4%, CI 34.6%] and 191 [78.6%] in only tau-positive [CU 71.7%, CI 28.3%] individuals). Differences in predicting cognitive impairment between nAD and irAD groups were observed with the 2011 NIA-AA (hazard ratio [HR] 2.21, 95% CI 1.34-3.65, p = 0.002), 2016 IWG-2 (HR 2.81, 95% CI 1.59-4.96, p < 0.000), and 2021 IWG-3 (HR 3.61, 95% CI 2.09-6.23, p < 0.000), but not with 2018 NIA-AA (HR 1.69, 95% CI 0.87-3.28, p = 0.115). Discussion: Over 42% of the studied population presented discordant diagnoses when using the different examined AD criteria, mostly in individuals with a single positive biomarker. Except for 2018 NIA-AA, all guidelines identified asymptomatic individuals at risk of cognitive impairment. Our findings highlight important differences between the guidelines, emphasizing the necessity for updated criteria with enhanced staging metrics, considering clinical, research, therapeutic, and trial design aspects.