Search

Our study aimed to identify clusters of hospitalized older COVID-19 patients according to their main comorbidities and routine laboratory parameters to evaluate their association with in-hospital mortality. We performed an observational study on 485 hospitalized older COVID-19 adults (aged 80+ years). Patients were aggregated in clusters by a K-medians cluster analysis. The primary outcome was in-hospital mortality. Medical history and laboratory parameters were collected on admission. Frailty, defined by the Clinical Frailty Scale (CFS), referred to the two weeks before hospitalization and was used as a covariate. The median age was 87 (83–91) years, with a female prevalence (59.2%). Three different clusters were identified: cluster 1 (337), cluster 2 (118), and cluster 3 (30). In-hospital mortality was 28.5%, increasing from cluster 1 to cluster 3: cluster 1 = 21.1%, cluster 2 = 40.7%, and cluster 3 = 63.3% (p < 0.001). The risk for in-hospital mortality was higher in clusters 2 [HR 1.96 (95% CI: 1.28–3.01)] and 3 [HR 2.87 (95% CI: 1.62–5.07)] compared to cluster 1, even after adjusting for age, sex, and frailty. Patients in cluster 3 were older and had a higher prevalence of atrial fibrillation, higher admission NT-proBNP and C-reactive protein levels, higher prevalence of concurrent bacterial infections, and lower estimated glomerular filtration rates. The addition of CFS significantly improved the predictive ability of the clusters for in-hospital mortality. Our cluster analysis on older COVID-19 patients provides a characterization of those subjects at higher risk for in-hospital mortality, highlighting the role played by cardio-renal impairment, higher inflammation markers, and frailty, often simultaneously present in the same patient.

Background and Aims: Cardiac natriuretic peptides (NPs) exert several metabolic effects, including some on lipid metabolism. Higher NPs levels are likely to be associated with a favorable lipid profile. In in vitro studies, NPs have been found to modulate low-density lipoprotein receptor (LDLR) trafficking by preventing proprotein convertase subtilisin/kexin type 9 (PCSK9) overexpression. The aim of our study is to investigate a possible association between plasma levels of PCSK9 and N-terminal pro B-type natriuretic peptide (NT-proBNP) in vivo. Methods: We performed a cross-sectional study on 160 consecutive older male and female patients hospitalized for medical conditions. Patients taking lipid-lowering drugs and patients with an admission diagnosis of acute heart failure were excluded. Fasting blood samples were collected after clinical stabilization of the acute illness, the day before discharge. Results: The mean age was 87.8 ± 6.4 years with a female prevalence (62.5%). The median NT-proBNP was 2340 (814–5397) pg/mL. The mean plasma PCSK9 was 275.2 ± 113.2 ng/mL. We found an inverse correlation between plasma PCSK9 and NT-proBNP (r = −0.280; p = 0.001). This association was confirmed after taking into account NT-proBNP tertiles (plasma PCSK9 levels: 317.4 ± 123.6 ng/mL in the first tertile, 283.3 ± 101.8 ng/mL in the second tertile, 231.3 ± 99.0 ng/mL in the third tertile, p = 0.001) and even after an adjustment for confounding factors (beta = −0.361, p = 0.001 for ln(NT-proBNP); beta = −0.330, p = 0.001 for NT-proBNP tertiles). The strength of the correlation between plasma PCSK9 and NT-proBNP was likely greater in patients affected by type 2 diabetes mellitus (r = −0.483; p = 0.006) and in male patients (r = −0.431, p = 0.001). Conclusion: The inverse association found between PCSK9 and NT-proBNP plasma levels in our real-life clinical study supports the hypothesis that NPs may play a role in cholesterol metabolism, possibly through an inhibitory action on circulating PCSK9 concentrations, thus increasing the availability of LDLR.