Digital photoplethysmography to detect deep venous thrombosis in the emergency department

Study objectives: We determine the sensitivity and specificity of digital photoplethysmography (DPPG) performed by emergency physicians in patients suggestive of a deep venous thrombosis. Methods: This was a prospective longitudinal cohort study conducted in the 2 emergency departments (EDs) of Summa Health System. The EDs treat a combined 97,000 patients per year. On average, 88 patients per month are evaluated for deep venous thrombosis with duplex scans; approximately 23% are positive for deep venous thrombosis. We enrolled consecutive ED patients aged 18 years or older and with lower extremity symptoms of deep venous thrombosis requiring venous duplex scanning, who gave consent to participate. Patients were excluded if they were unable to comply with DPPG testing, had fragile or broken skin in the area on which the DPPG probe was placed, had inadequate arterial flow to the lower extremities, did not complete the required testing (duplex ultrasonography and DPPG), had a deep venous thrombosis in the previous 6 months, were at least 13 weeks pregnant, or were previously enrolled. A trained emergency physician performed and interpreted the DPPG. All patients underwent noninvasive duplex scanning; the vascular surgeonsʹ interpretation was the criterion standard. Emergency physicians were blinded to the duplex result, and the vascular technologists and vascular surgeons were blinded to the results of the DPPG. The primary study outcome was the sensitivity and specificity of the DPPG. Results: Results are presented as proportions with 95% confidence intervals (CIs). The study was discontinued early because of technical problems with the software and scans, resulting in poor quality control and low sensitivity. We evaluated the data on 116 total legs in 101 total patients (15 patients had bilateral legs evaluated). Twenty-eight legs in 23 patients were excluded because of technically inadequate scans (24% of legs; 95% CI 17% to 33%), which left 88 eligible legs in 78 patients. The prevalence of deep venous thrombosis was 12.5% (95% CI 6% to 22%). Using the old software, the sensitivity of DPPG was 40% (4/10; 95% CI 12% to 47%), and the specificity was 63% (49/78; 95% CI 51% to 74%). The manufacturer used the data obtained to reprogram the DPPG software; the sensitivity of DPPG with the new software was 90% (9/10; 95% CI 55% to 100%), and the specificity was 51% (40/78; 95% CI 40% to 63%). The sensitivity for above-knee deep venous thrombosis was 100% (95% CI 66% to 100%). Conclusion: The sensitivity for above-knee deep venous thrombosis using the new software is promising but requires further study before it can be recommended for ED use.