This human cadaver study investigated whether flexor tendon repairs performed with round-bodied needles had a higher risk of pull-out compared with those performed with cutting needles. Forty human cadaver tendons were repaired (20 with each type of needle), subjected to tensile traction testing and evaluated by failure load and mode of failure. The average failure load was 50 N (SD 13 N) for tendons repaired with round-bodied needles, compared with 49 N (SD 16 N) for tendons repaired with cutting needles. Round-bodied needles resulted in more suture pull-out (18 out of 20 tendons) than cutting needles (6 out of 20 tendons). We found no differences in failure load, but significant differences in the mode of failure between round-bodied and cutting needles when used for cadaveric flexor tendon repair.