Photo Credit: Sopone Nawoot
The following is a summary of “Decrease of haemoconcentration reliably detects hydrostatic pulmonary oedema in dyspnoeic patients in the emergency department – a machine learning approach,” published in the September 2024 issue of Emergency Medicine by Gavelli et al.
Researchers conducted a retrospective study to assess the efficacy of hemoglobin variation (ΔHb) generated by fluid transfer through the interstitium as a diagnostic tool for hydrostatic pulmonary edema (HPO) in the emergency department (ED) setting.
They conducted an observational retrospective study of ED patients with acute dyspnoea. Hb values were recorded at ED presentation (T0) and after 4 to 8 hours (T1), ΔHb between T1 and T0(ΔHbT1-T0) was estimated as absolute and relative values. Blind Hb values defining the cause of dyspnoea as HPO or non-HPO were found by 2 investigators. The ability of ΔHbT1-T0 to detect HPO was assessed. A machine learning technique was used to develop HPO, considering ΔHb as a covariate with baseline patient characteristics.
The results showed that 706 dyspneic patients (203 HPO and 503 non-HPO) enrolled over 19 months. Hb levels varied between patients with HPO and non-HPO at T0 and T1 (P< 0.001). ΔHbT1-T0 was more prominent in patients with HPO than non-HPO, both as relative (-8.2 [-11.2 to -5.6] vs. 0.6 [-2.1 to 3.3] %) and absolute (-1.0 [-1.4 to -0.8] vs. 0.1 [-0.3 to 0.4] g/dL) values (P< 0.001). A relative ΔHbT1-T0 of -5% detected HPO with an area under the receiver operating characteristic curve (AUROC) of 0.901 [0.896–0.906]. The Gradient Boosting Machine exhibited excellent predictive ability in identifying patients with HPO and was used to create a web-based application. ΔHbT1-T0was verified as the critical covariate for HPO prediction.
They concluded ΔHbT1-T0 in patients with acute dyspnoea identifies HPO in the ED, and the machine learning predictive tool may be a valuable tool for confirming HPO.
Source: intjem.biomedcentral.com/articles/10.1186/s12245-024-00698-y