Normal gait, albumin and d-dimer levels identify low risk emergency department patients: a prospective observational cohort study with 365-day 100% follow-up

L. Lyngholm, C. H. Nickel, J. Kellett, S. Chang, T. Cooksley, M. Brabrand

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

Abstrakt

BACKGROUND: If survival could be reliably predicted many patients could be safely managed outside of hospital in an ambulatory care setting. AIM: Comparison of common laboratory findings, co-morbidities, mobility and vital signs as predictors of mortality of acutely ill emergency department (ED) attendees. DESIGN: Prospective observational study. METHODS: Secondary analysis of 1334 consenting acutely ill patients attending a Danish ED. RESULTS: 67 (5%) out of 1334 patients died within 100 days. After logistic regression seven predictors of 100 days mortality remained significant: an albumin level ≤34 gm/l, D-dimer level >0.51 mg/l, an Asadollahi score (based on admission laboratory data and age) ≥12, a platelet count <159 X 1000/ml, impaired mobility on presentation, a respiratory rate ≥30 bpm and a Charlson co-morbidity index ≥3. Only 5 of the 442 without any of these variables died within 365 days. Only one of the 517 patients with a stable independent gait and normal d-dimer and albumin levels died within 100 days, none died within 30 days of assessment and 12 died within 365 days. Of the remaining 817 patients 66 (8%) died within 100 days. CONCLUSION: These findings suggest that normal gait, albumin and d-dimer levels are the most parsimonious way of identifying low risk ED patients.

OriginalsprogEngelsk
TidsskriftQJM : monthly journal of the Association of Physicians
Vol/bind113
Udgave nummer2
Sider (fra-til)86-92
ISSN1460-2725
DOI
StatusUdgivet - 1. feb. 2020

Fingeraftryk Dyk ned i forskningsemnerne om 'Normal gait, albumin and d-dimer levels identify low risk emergency department patients: a prospective observational cohort study with 365-day 100% follow-up'. Sammen danner de et unikt fingeraftryk.

Citationsformater