Ceramides, phosphatidylcholines and cognitive function in the general population – results from the STAAB cohort study

Vladimir Cejka (Würzburg)1, F. Sahiti (Würzburg)1, G. Gelbrich (Würzburg)2, C. Morbach (Würzburg)3, W. März (Mannheim)4, F. Montellano (Würzburg)2, L. Schmidbauer (Würzburg)2, P. U. Heuschmann (Würzburg)2, S. Störk (Würzburg)1

1Universitätsklinikum Würzburg Deutsches Zentrum für Herzinsuffizienz Würzburg, Deutschland; 2Universitätsklinikum Würzburg Institut für Klinische Epidemiologie und Biometrie Würzburg, Deutschland; 3Universitätsklinikum Würzburg Medizinische Klinik I, Kardiologie Würzburg, Deutschland; 4SYNLAB Holding Deutschland GmbH SYNLAB Akademie Mannheim, Deutschland



Impaired cognitive function is associated with adverse outcome in patients with heart failure. Nevertheless, the pathophysiological mechanisms are not fully understood, so far. Ceramides (CER) and phosphatidylcholines (PC) are small lipids participating in membrane transport and intracellular signaling; they have been implicated in cardiovascular risk and neurodegeneration. We investigated their association with cognitive function in the general population.

Methods and Results:

The STAAB cohort study included adult residents of the City of Würzburg, Germany, free of known heart failure. Besides a comprehensive cardiovascular profiling and fasting blood sampling, cognitive function was measured by the Montreal Cognitive Assessment (MoCA). Serum CER and PC were analyzed by liquid chromatography–mass spectrometry with a targeted LCPL platform, and concentrations (pmol/µl) were log10-transformed to approximate a normal distribution. Principal components analysis (PCA) of 4 CER and 3 PC was performed, with the eigenvalue cut-off for component extraction set to 1. Regression-derived factor scores were included as predictors into a linear regression model with the MoCA score as the dependent variable, adjusted for age and sex. Collinearity was assessed by the variance inflation factor (VIF).

At baseline, 4338 participants had completed a valid MoCA test and provided serum samples for analysis. The median MoCA score was 26 (quartiles: 24, 28; range 12–31), median age was 55 (47; 65) years, 52 % were women, 5% had diabetes mellitus, mean body mass index was 26.6 (5.1) kg/m², blood pressure 134(20)/80(16) mmHg, LDL-cholesterol 122 (35) mg/dl. PCA yielded 2 components explaining 72% of the total variance. CER had the strongest correlations with component 1, PC with component 2. In multivariable linear regression analysis, component 1 score was a significant predictor of the MoCA score, beta-coefficient -0.180, 95%CI -0.276 to -0.084, p< 0.001, while component 2 score was not a significant predictor (p=0.094). Unadjusted model R² was 14,6%. No relevant collinearity was observed (maximal VIF = 1.138).


Independent of age and sex, elevated concentrations of CER were associated with a lower cognitive function. These findings suggest a potential pathophysiological link between cardiovascular disease and cognition. In ongoing analyses, the impact of other factors including educational status and traditional cardiovascular risk factors as well as the prognostic utility of CER/PC will be evaluated.

Figure 1:  Clustering of ceramides (n=4) and phosphatidylcholines (n=3) along 2 components identified by PCA.


Table 1. Multivariable linear regression model with MoCA score as the dependent variable and age, sex, component 1 and 2 as predictors.



95% confidence interval





30.665 – 31.600

< 0.001




-0.104 – -0.087

< 0.001


Sex (for man)


-0.696 – -0.323

< 0.001


Component 1 score


-0.276 – -0.084

< 0.001


Component 2 score


-0.014 – 0.172





Diese Seite teilen