Sensitivity and specificity of waveform analysis for assessing postoperative epidural function

To characterize the accuracy of epidural waveform analysis (EWA) in assessing the functionality of thoracic epidural catheters in the immediate postoperative period (primary objective), and to determine the inter-rater reliability between EWA waveform observers (secondary outcome). Single center, prospective diagnostic accuracy cohort study. Post-anesthetic care unit of a university teaching hospital. 84 adult patients undergoing elective thoracic, gynecologic, vascular, urologic, or general surgery with preoperative placement of a thoracic epidural catheter for perioperative analgesia. EWA tracings were video recorded in the immediate postoperative period through the epidural catheter in the post-anesthetic care unit. Postoperative EWA tracings were compared with clinical assessments of the sensory block to ice produced by epidural local anesthetic in the immediate postoperative period. Additionally, intra-class correlation analysis of agreement between 3 independent (and blinded) EWA waveform observers was carried out. Among 80 patients with thoracic epidurals who completed the study protocol, 73 demonstrated postoperative functional epidurals with sensory block to ice and 7 demonstrated non-functional epidurals. EWA yielded 65 true positives, 6 true negatives, 8 false negatives, and 1 false positive. Postoperative EWA sensitivity, specificity, positive predictive value and negative predictive value, along with the 95% confidence intervals (CI) were 89% (79–95%), 86% (42–100%), 98% (92–100%), and 43% (18–71%) respectively. Intra-class correlation between waveform assessors was 0.870 (95% CI 0.818–0.910, p