Bacteria were counted concurrently in the air and wounds during the fi
rst 20 min of total joint arthroplasty procedures in two theatres: a c
onventional plenum ventilated theatre with ultraviolet C (UVC) tubes i
nstalled and a filtered vertical laminar flow theatre. Four theatre en
vironments were tested: conventional theatre and clothing; conventiona
l theatre with UVC protective clothing, with UVC set to produce 100 or
300 mu W cm(-2) s(-1) irradiation; and filtered vertical laminar flow
air with staff wearing cuffed cotton/polyester clothing. When used, t
he UVC was activated 10 min after starting an operation to assess the
effect of UVC clothing alone, and of UVC radiation on bacteria already
present in the wound. Compared with conventional theatres, UVC clothi
ng reduced air counts by 38%, UVC at 100 mu W cm(-2) s(-1) by 81%, at
300 mu W cm(-2) s(-1) by 91%, and laminar flow by 92%. Wounds counts f
ell correspondingly by 66% with UVC clothing, 87% with UVC at 100 mu W
cm(-2) s(-1) and 92% both with UVC at 300 mu W cm(-2) s(-1) and lamin
ar flow. In conventional and laminar flow theatres air and wound count
s correlated closely but in UVC theatres wound counts were lower than
levels expected from prevailing air counts suggesting that UVC kills b
acteria in wounds as well as in air.