LO22: Does point-of-care ultrasonography improve diagnostic accuracy in emergency department patients with undifferentiated hypotension? An international randomized controlled trial from the SHoC-ED investigators

Abstract

Introduction: Point of care ultrasound has been reported to improve diagnosis in non-traumatic hypotensive ED patients. We compared diagnostic performance of physicians with and without PoCUS in undifferentiated hypotensive patients as part of an international prospective randomized controlled study. The primary outcome was diagnostic performance of PoCUS for cardiogenic vs. non-cardiogenic shock. Methods: SHoC-ED recruited hypotensive patients (SBP < 100 mmHg or shock index > 1) in 6 centres in Canada and South Africa. We describe previously unreported secondary outcomes relating to diagnostic accuracy. Patients were randomized to standard clinical assessment (No PoCUS) or PoCUS groups. PoCUS-trained physicians performed scans after initial assessment. Demographics, clinical details and findings were collected prospectively. Initial and secondary diagnoses including shock category were recorded at 0 and 60 minutes. Final diagnosis was determined by independent blinded chart review. Standard statistical tests were employed. Sample size was powered at 0.80 (α:0.05) for a moderate difference. Results: 273 patients were enrolled with follow-up for primary outcome completed for 270. Baseline demographics and perceived category of shock were similar between groups. 11% of patients were determined to have cardiogenic shock. PoCUS had a sensitivity of 80.0% (95% CI 54.8 to 93.0%), specificity 95.5% (90.0 to 98.1%), LR+ve 17.9 (7.34 to 43.8), LR-ve 0.21 (0.08 to 0.58), Diagnostic OR 85.6 (18.2 to 403.6) and accuracy 93.7% (88.0 to 97.2%) for cardiogenic shock. Standard assessment without PoCUS had a sensitivity of 91.7% (64.6 to 98.5%), specificity 93.8% (87.8 to 97.0%), LR+ve 14.8 (7.1 to 30.9), LR- of 0.09 (0.01 to 0.58), Diagnostic OR 166.6 (18.7 to 1481) and accuracy of 93.6% (87.8 to 97.2%). There was no significant difference in sensitivity (-11.7% (-37.8 to 18.3%)) or specificity (1.73% (-4.67 to 8.29%)). Diagnostic performance was also similar between other shock subcategories. Conclusion: As reported in other studies, PoCUS based assessment performed well diagnostically in undifferentiated hypotensive patients, especially as a rule-in test. However performance was similar to standard (non-PoCUS) assessment, which was excellent in this study.