Association between SCD‐Plus features and GDS factors in subjective cognitive decline and healthy controls in the studies DELCODE and SILCODE. (31st December 2021)

• Record Type: Journal Article
• Authors: Hu, Xiaochen; Chen, Ji; Jiang, Xueyan; Daamen, Marcel; Wang, Xiaoqi; Lee, Hwee Ling; Spottke, Annika; Düzel, Emrah; Peters, Oliver; Buerger, Katharina; Teipel, Stefan J.; Laske, Christoph; Boecker, Henning; Wagner, Michael; Han, Ying; Jessen, Frank
• Is Part Of: Alzheimer's & dementia. Volume 17(2021)Supplement 6
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-12-31
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.053619
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: A previous study showed a subjective cognitive decline (SCD) related factorial score derived from the geriatric depression scale (GDS, 30 items) to predict cognitive decline within a group of healthy elderlies in interaction with amyloid status[1]. On the other hand, several features associated with SCD (SCD plus), including subjective decline in the memory domain, concerns about cognitive decline, onset within five years and over 60 years old, and cognitive performance worse than other people, have been proposed to increase the likelihood of preclinical AD in subjects with SCD[2]. Here, we test the association between SCD plus features and GDS factorial scores in two independent cohorts from Germany and China. Method: The 15‐items GDS(GDS15) and the SCD plus score, which was assessed with the SCD‐interview[3] and which summarizes SCD plus features, were obtained from two cohorts, the DZNE‐Longitudinal Cognitive Impairment and Dementia Study (DELCODE, 230HC/361SCD), and the Sino Longitudinal Cognitive Impairment and Dementia Study(SILCODE, 99HC/135SCD). Non‐negative matrix factorization[4, 5] was applied to the GDS items from the two cohorts separately with the number of factors ranging from 2 to 9. The optimal number of factors, in terms of stability and generalizability, was identified by using cross‐validation in 5.000 split‐half analyses in each cohort. Partial correlations between SCD plus scores and individual loadings on each factor derived from the optimal factor structure were calculated, controlling for age, sex, education years, and the loadings on other factors. Result: For both cohorts, the three‐factorial‐solution revealed the best generalizability (lowest out‐of‐sample increased reconstruction error of individual GDS scores), as well as good stability of item‐to‐factor assignment. Partial correlations revealed significant association of SCD plus score with two factors in the DELCODE cohort (positive mood:r=‐0.20, p=1x10 ‐6 ;cognition/anxiety:r=0.41, p=8.1x10 ‐25 ), and with two factors in the SILCODE cohort (positive mood:r=‐0.18, p=0.007; memory problem:r=0.51, p=2.6 x 10 ‐16 )(figure 1). Conclusion: The present work corroborate the previous finding [1] that a SCD related factor can be derived from GDS. Furthermore, we demonstrated that the individual loadings on this factor are strongly associated with SCD plus features in two independent cohorts. Interestingly, the item assigned to the SCD related factor are different between two cohorts.