Background: Brush cytology is commonly used to evaluate potentially malignant strictures at ERCP. A new device with a second large lumen has been developed to allow the passage of accessories into the duct. We hypothesize that brushing obtained with this device would have a higher yield than a standard system because of the side pressure exerted by the catheter on the brush in the stricture. We report an alternating randomized trial comparing the accuracy of two systems. Methods: Method 1 uses the Cytomax brush system (CBS),Wilson-Cook, with an 8Fr-sheath diameter and a 2.5cm brush length. Brushings are obtained after stricture dilation with a separate 10Fr catheter. Method 2 uses the Howell Biliary system (HBS), Wilson-Cook, a 10Fr system with a separate channel to allow the passage of 5Fr sampling devices. The delivery device is used to dilate the stricture before brushings. In both systems bile was aspirated prior to brushings. The brush, slides, and fluid are sent for cytologic evaluation. Both catheter systems were flushed with sterile saline to salvage any cells within the sheath. Biliary biopsy forceps and FNA catheters were not routinely used. Results: 23 consecutive patients were enrolled, 11 CBS and 12 HBS. No significant demographic difference was present. The CBSgroup had 9 CBD and 2 IHD strictures (4 pancreatic adenocarcinoma, 2 cholangiocarcinoma, 5 benign strictures). This brushing technique alone had an accuracy of 81% with a sensitivity and specificity of 83% and 80%. The combination of brushing and aspirate had an accuracy 82% with sensitivity, specificity, PPV, and NPV of 83%, 80%, 83%, and 80%, respectively. The HBSgroup had 10 CBD, 1 IHD, and 1 pancreatic stricture (6-pancreatic adenocarcinoma, 1-cholangiocarcinoma, 1-metastatic gastric adenoca, 1- metastatic renal cell ca, and 3-benign strictures). This brushing technique alone had an accuracy of 67% with a sensitivity and specificity of 56% and 100%. The combination of brushing and aspirates had an accuracy of 75%, with sensitivity, specificity, PPV, NPV of 67%, 100%, 100%, and 50%, respectively. The biopsy forceps were available in 3/12 patients. All specimens were satisfactory for interpretation and there were 2 true positives and 1 true negative. Overall, the CBS accuracy and HBS accuracy for brush + aspiration cytology were 82% vs. 75% (NS, p>.05, Fisher s Exact). Conclusion: Our preliminary results suggest that the HBS sytem does not increase yield when compared to the CBS system unless intraductal biopsies are performed.