Clusters of comorbidities and in-hospital mortality in elderly patients affected by acute heart failure: a single-cohort prospective study

Background: acute heart failure (AHF) is negatively affected by the presence of comorbidities. We aimed to evaluate if different associations of comorbidities could differently affect in-hospital mortality for AHF in a cohort of elderly patients. Materials and Methods: we enrolled all the consecutive patients admitted to our Internal Medicine department in the years 2014-2017 for AHF investigating age, sex, in-hospital mortality and presence of 16 common internistic comorbidities. Association of chronic pathologies was explored with Pearson’s bivariate test, selecting only clusters of ≥2 significantly associated comorbidities. Then, we obtained ROC curves predicting in-hospital mortality for AHF from binary logistic regressions adopting each cluster as predictor. Results: Mean age was 82.56(±8, 92), females representing 53.7% of the sample. In-hospital mortality was 13%. We obtained 11 different clusters, 6 of which predicting significantly in-hospital mortality. The first (anaemia, dementia, diabetes, AF) had AUC:0.678;95%CI:0.585-0.810;p=0.032. The second (hypertension, CKD, CVD) had AUC:0.673;95%CI:0.525- 0.821;p=0.037. The third (dyslipidaemia, AF, PAD, CVD, haematologic disorders) had AUC:0.716;95%CI:0.564-0.868;p=0.009.The fourth (AF, anaemia, dyslipidaemia, diabetes, CVD, IPB) had AUC:0.764;95%CI:0.658- 0.870;p=0.002.The fifth (CVD, OSAS, dyslipidaemia, AF, PAD, rheumatologic diseases) had AUC:0.687;95%CI:0.557-0.818;p=0.024.The last (OSAS, AF, IPB) had AUC:0.695;95%CI:0.565-0.825;p=0.019. Discussion: In this small sample, some clusters of comorbidities were not associated to an increased risk of in-hospital mortality in AHF. However, we identified some patterns which are strongly associated to in-hospital death. These cluster deserve the best clinical attention and the most extensive therapy.