Comparison of 19-gauge aspiration needles versus 20-gauge forward-bevel needles for diagnosing autoimmune pancreatitis: a single-centre randomised controlled trial

The diagnosis of autoimmune pancreatitis (AIP) is multidisciplinary. Pathological diagnosis is pivotal when distinguishing AIP from malignancies. In this study, we aim to compare the diagnostic performance and sample adequacy of 19-gauge endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) needles with 20-gauge endoscopic ultrasound-guided fine-needle biopsy (EUS-FNB) needles in patients with suspected AIP. We conducted a single-centre, parallel-group, prospective, randomized controlled trial at Peking Union Medical College Hospital (Beijing, China). Patients with suspected AIP were enrolled between October 2018 and August 2023 and were randomly assigned to undergo tissue specimen collection using either a 19-gauge EUS-FNA needle or a 20-gauge EUS-FNB needle. The primary endpoint was the diagnostic accuracy against International Consensus Diagnostic Criteria, with secondary endpoints including histologic sensitivity, technical success rate, tissue adequacy, and adverse event rate. 46 patients were enrolled and 45 underwent EUS (FNB n = 23, FNA n = 22), amongst 42 had a final diagnosis of AIP. Diagnostic accuracy, as the primary endpoint, was 52.2% (95% CI 33.0–70.8) with 20-G FNB versus 47.8% (95% CI 29.2–67.0) with 19-G FNA (P = 1.00). Sensitivity for histologic diagnosis of AIP was 47.6% (95% CI 28.3–67.6) in both groups (P = 1.00). Core tissue (≥ 500 μm) was obtained in 65.2% vs. 68.2% (P = 0.83). Adverse events were infrequent and similar (4.3%; both mild; P = 1.00). We did not observe a statistically significant difference in diagnostic efficacy or core tissue acquisition between 19G FNA and 20G forward-bevelled FNB needles for patients with AIP; however, confidence intervals were wide and smaller clinically important differences could not be excluded. The 19-G FNA needle might serve as an alternative in centres where FNB needles are unavailable. This study was retrospectively registered on 22 November 2018 at ClinicalTrials.gov (NCT03753815); first participant enrolled in October 2018.