Eighty-five consecutive patients were randomized to undergo mass closure of the abdomen with no. 1 polydioxanone mounted on either a blunt-tipped (n = 46) or cutting (n = 39) needle. Gloves were changed before closure and tested for perforation afterwards using standard air or water techniques. Fourteen pairs of gloves were punctured when using a cutting needle, and three pairs when a blunt-tipped needle was used. The majority of punctures were to the non-dominant glove. The surgeon was aware of the puncture in eight of the 14 instances involving a sharp needle and in one of the three involving a blunt-tipped needle. Blunt-tipped needles, while not eliminating the risk, significantly reduced the incidence of surgical glove puncture (P < 0.001, Fisher's exact test). The use of cutting needles for abdominal closure should be abandoned.