The Reproducibility of the Point of Care Microcirculation (POEM) Score When Used to Assess Critically Ill Patients: A Multicenter Prospective Observational Study.

Abstract

The current standard of analyzing microcirculatory video microscopy is time-consuming and occurs away from the patient, limiting its clinical utility. Point-of-care assessment with incident dark field (IDF) microscopy, however, may offer greater clinical applicability. We aimed to determine the reproducibility of the Point of Care Microcirculation (POEM) tool when used at the bedside in critically ill patients. A multinational, multicenter, prospective observational study of adult intubated patients was undertaken during a 9-month period in Germany, the United Kingdom, and the United States. A user recorded a batch of four standardized video clips from each patient, calculated a POEM score and recorded the time for image acquisition. A second user blinded to the first repeated this process. Patients with video clips of poor quality were excluded. At a later date, the two users again blinded themselves to reassess both their own clips and those of the other user. Basic demographic information was recorded. Intrauser reliability (an individual user rescoring the same batch of videos after blinding), interuser reliability (a second user rescoring the other user's video batch after blinding), and test-retest reliability (two users individually capturing videos and recording POEM scores) were assessed using a linearly weighted kappa statistic for ordinal data. Sixty-five patients were included in the final analysis. Observer agreement was substantial for all tests. Intrauser agreement was 0.73 (0.95 CI 0.64-0.81), interuser agreement 0.71 (0.95 CI 0.63-0.79), and test-retest agreement 0.75 (0.95 CI 0.65-0.86). Average time to record videos and assess POEM scores 7:34 ± 3:37 minutes. Point-of-care assessment of the microcirculation using IDF video microscopy and POEM scoring appears to be both a feasible and reproducible approach to microcirculatory assessment. Testing of the score in critically ill patients showed substantial agreement within and between investigators, but further studies should validate its utility as a tool to guide shock resuscitation.