Refine
Has Fulltext
- yes (2)
Is part of the Bibliography
- yes (2)
Year of publication
- 2023 (2) (remove)
Document Type
- Journal article (2)
Language
- English (2) (remove)
Keywords
- chronic heart failure (1)
- cluster analysis (1)
- cognitive impairment (1)
- comparability (1)
- disease score (1)
- disease severity (1)
- glial fibrillary acidic protein (1)
- intensity of attention (1)
- normal values (1)
- reference data (1)
Institute
- Medizinische Klinik und Poliklinik II (2) (remove)
Sonstige beteiligte Institutionen
Background
Cognitive impairment is a major comorbidity in patients with chronic heart failure (HF) with a wide range of phenotypes. In this study, we aimed to identify and compare different clusters of cognitive deficits.
Methods
The prospective cohort study “Cognition.Matters-HF” recruited 147 chronic HF patients (aged 64.5 ± 10.8 years; 16.2% female) of any etiology. All patients underwent extensive neuropsychological testing. We performed a hierarchical cluster analysis of the cognitive domains, such as intensity of attention, visual/verbal memory, and executive function. Generated clusters were compared exploratively with respect to the results of cardiological, neurological, and neuroradiological examinations without correction for multiple testing.
Results
Dendrogram and the scree plot suggested three distinct cognitive profiles: In the first cluster, 42 patients (28.6%) performed without any deficits in all domains. Exclusively, the intensity of attention deficits was seen in the second cluster, including 55 patients (37.4%). A third cluster with 50 patients (34.0%) was characterized by deficits in all cognitive domains. Age (p = 0.163) and typical clinical markers of chronic HF, such as ejection fraction (p = 0.222), 6-min walking test distance (p = 0.138), NT-proBNP (p = 0.364), and New York Heart Association class (p = 0.868) did not differ between clusters. However, we observed that women (p = 0.012) and patients with previous cardiac valve surgery (p = 0.005) prevailed in the “global deficits” cluster and the “no deficits” group had a lower prevalence of underlying arterial hypertension (p = 0.029). Total brain volume (p = 0.017) was smaller in the global deficit cluster, and serum levels of glial fibrillary acidic protein were increased (p = 0.048).
Conclusion
Apart from cognitively healthy and globally impaired HF patients, we identified a group with deficits only in the intensity of attention. Women and patients with previous cardiac valve surgery are at risk for global cognitive impairment when suffering HF and could benefit from special multimodal treatment addressing the psychosocial condition.
We assume that a specific health constraint, e.g., a certain aspect of bodily function or quality of life that is measured by a variable X, is absent (or irrelevant) in a healthy reference population (Ref0), and it is materially present and precisely measured in a diseased reference population (Ref1). We further assume that some amount of this constraint of interest is suspected to be present in a population under study (SP). In order to quantify this issue, we propose the introduction of an intuitive measure, the population comparison index (PCI), that relates the mean value of X in population SP to the mean values of X in populations Ref0 and Ref1. This measure is defined as PCI[X] = (mean[X|SP] − mean[X|Ref0])/(mean[X|Ref1] − mean[X|Ref0]) × 100[%], where mean[X|.] is the average value of X in the respective group of individuals. For interpretation, PCI[X] ≈ 0 indicates that the values of X in the population SP are similar to those in population Ref0, and hence, the impairment measured by X is not materially present in the individuals in population SP. On the other hand, PCI[X] ≈ 100 means that the individuals in SP exhibit values of X comparable to those occurring in Ref1, i.e., the constraint of interest is equally present in populations SP and Ref1. A value of 0 < PCI[X] < 100 indicates that a certain percentage of the constraint is present in SP, and it is more than in Ref0 but less than in Ref1. A value of PCI[X] > 100 means that population SP is even more affected by the constraint than population Ref1.