Point-of-care neutrophil and monocyte surface markers differentiate bacterial from viral infections at the emergency department within 30†min.

Rapid discrimination between viral and bacterial infections in a point-of-care setting will improve clinical outcome. Expression of CD64 on neutrophils (neuCD64) increases during bacterial infections, whereas expression of CD169 on classical monocytes (cmCD169) increases during viral infections. The diagnostic value of automated point-of-care neuCD64 and cmCD169 analysis was assessed for detecting bacterial and viral infections at the emergency department. Additionally, their value as input for machine learning models was studied. A prospective observational cohort study in patients suspected of infection was performed at an emergency department. A fully automated point-of-care flow cytometer measured neuCD64, cmCD169, and additional leukocyte surface markers. Flow cytometry data were gated using the FlowSOM algorithm. Bacterial and viral infections were assessed in standardized clinical care. The sole and combined diagnostic value of the markers was investigated. Clustering based on unsupervised machine learning identified unique patient clusters. Eighty-six patients were included. Thirty-five had a bacterial infection, 30 had a viral infection, and 21 had no infection. neuCD64 was increased in bacterial infections (P < 0.001), with an area under the receiver operating characteristic curve (AUROC) of 0.73. cmCD169 was higher in virally infected patients (P < 0.001; AUROC 0.79). Multivariate analyses incorporating additional markers increased the AUROC for bacterial and viral infections to 0.86 and 0.93, respectively. The additional clustering identified 4 distinctive patient clusters based on infection type and outcome. Automated neuCD64 and cmCD169 determination can discriminate between bacterial and viral infections. These markers can be determined within 30 min, allowing fast infection diagnostics in the acute clinical setting.

Cardiac output measurement by pulse dye densitometry: comparison with pulmonary artery thermodilution in post-cardiac surgery patients.

OBJECTIVE: Pulse-dye densitometry (PDD) could be a suitable, low-invasive alternative to thermodilution using a pulmonary artery catheter (PAC) for monitoring cardiac output. The aim of our study was to assess the reproducibility and validity of PDD compared to PAC-thermodilution. METHODS: In 43 post-cardiac surgery patients, the mean of triplicate readings of cardiac output was assessed using both methods. In a subgroup of 26 patients, a second set of measurements was obtained on average 2 h later. RESULTS: Reproducibility of consecutive measurements was slightly better for PAC-thermodilution than for PDD (median coefficient of variation of the triplicate measurements: 3.5% versus 5.4%, P < 0.01). Both methods correlated well (r = 0.84, p < 0.001). Using Bland and Altman analyses with PAC-thermodilution as the reference method, PDD showed a bias of -0.68 +/- 0.82 L/min, mainly due to differences in higher ranges of cardiac output (>6.5 L/min). Measured changes in cardiac output were 81% concordant (i.e. <1 L/min different) between both methods. CONCLUSION: PDD correlates well with PAC-thermodilution and thus deserves consideration as a low-invasive alternative for measurement and follow-up of cardiac output.