Copyright © 2020. Published by Elsevier Editora Ltda.BACKGROUND AND OBJECTIVES: The aim of this prospective, multi-centered and multi-arm parallel randomized trial was to test the hypothesis that modified sitting positions including hamstring stretch position (HSP) and squatting position (SP) would reduce needle - bone contact events and increase the success rate of combined spinal - epidural anesthesia (CSEA) compared to traditional sitting position (TSP) in patients undergoing total knee or hip arthroplasty. PATIENTS AND METHODS: Three hundred and sixty American Society of Anesthesiologists (ASA) I-III patients, aged between 45-85 years were randomly allocated to one of three groups using computer-generated simple randomization: group TSP (n = 120), group HSP (n = 120), and group SP (n = 120). Primary outcome measures were the number of needle-bone contact and success rates. Secondary outcome measure was the ease of interspinous space identification. RESULTS: Seven patients in group SP and four of HSP could not tolerate their position and were excluded. Number of needle-bone contact, success rates, and grade of interspinous space identification were similar between groups (p = 1.000). Independent of positioning, the success rates were higher in patients whose interspinous space was graded as easy compared to difficult or impossible (p < 0.001). Success rates reduced, interspinous space identification became more challenging, and number of needle - bone contact increased as patient's body mass index (BMI) increased (p < 0.001). CONCLUSION: SP and HSP may be used as alternatives to the TSP. BMI and ease of interspinous space identification may be considered important determinants for CSEA success.