Creatinine and NT-ProBNP levels could predict the length of hospital stay of patients with decompensated heart failure

Link to article at PubMed

Acta Cardiol. 2021 Jan 22:1-8. doi: 10.1080/00015385.2020.1871264. Online ahead of print.


BACKGROUND: Heart failure (HF) is a clinical syndrome that causes high morbidity and mortality with a high number of admissions and sometimes prolonged admissions. This study aimed at assessing whether parameters detected during the first 24 h of admission may predict a prolonged hospital stay in patients admitted to hospital for decompensated HF.

METHODS: From January 2016 to December 2019, 2359 admissions of decompensated HF were recorded. In-hospital transfers, de novo HF, deaths and scheduled admissions were discarded to homogenise the sample. Finally, 1196 patients were included. The sample was divided into two groups: (a) non-prolonged admission (n = 643, admission ≤7 days) or (b) prolonged admission (n = 553, admission >7 days). Clinical, analytical, electrocardiographic and echocardiographic variables obtained during the first 24 h of admission were analysed.

RESULTS: Univariate differences were found at admission in NT-ProBNP, creatinine, history of cardiac surgery, smoking and alcoholism, left and right ventricular ejection fraction, systolic blood pressure and heart rate. The ROC analysis showed significant areas under the curve for the NT-ProBNP (AUC: 0.63, 95% CI: 0.60-0.67; p < 0.001) and creatinine (AUC: 0.69, 95% CI: 0.66-0.72; p < 0.0001). The variables associated with prolonged hospital admission were NT-ProBNP (OR: 1, 95% CI: 1-1; p < 0.001), creatinine (OR: 2.2, 95% CI: 1.8-2.7; p < 0.0001) and previous smoking (OR: 1.5, 95% CI: 0.4-1; p < 0.02).

CONCLUSIONS: Variables such as creatinine and NT-ProBNP at hospital admission may define a subgroup of patients who will probably have a long hospital stay. Therefore, the planning of hospital care and transition to discharge may be enhanced.

PMID:33480331 | DOI:10.1080/00015385.2020.1871264

Leave a Reply

Your email address will not be published.