Is the antimicrobial removal device a cost-effective addition to conventional blood cultures?

Two hundred and thirty-four blood cultures from 140 patients receiving antibiotics were processed using the antimicrobial removal device (ARD) in parallel with conventional blood cultures. One hundred and seventy cultures were obtained from patients suspected to have bacteraemia and 64 from patients known to have a positive conventional blood culture within the preceding three days. A total of 38 (16.2%) ARD-processed cultures were positive, compared with 21 (8.9%) conventional cultures (p less than 0.0001, Fisher's exact test). No instances of positive conventional cultures and negative ARD-processed cultures were identified. Thirty-three of 38 ARD-processed cultures became positive within 24 h, compared with 14 parallel conventional cultures (p less than 0.0001 Fisher's exact test). Although the yield and rapidity of isolation of bacteria from blood were improved by ARD processing, in only one of 140 patients did this information alter treatment. As the use of an ARD is associated with a sixfold increase in the cost of blood cultures, we conclude that, in our hands, general use of the device in patients receiving antibiotics is not cost-effective. Considerable care should be taken in selecting patients for ARD-processing of blood cultures.