Tigecycline activity: low resistance rates but problematic disc breakpoints revealed by a multicentre sentinel survey in the UK

Abstract

Objectives: A multicentre surveillance of tigecycline activity against relevant pathogens in the UK.

Methods: Forty-three representative UK hospitals were each asked to test 150 consecutive, clinically significant isolates from hospitalized patients, excluding those from urine or faeces. The sentinel laboratories tested all these isolates against a panel of antibiotics by the BSAC disc method and a selection of isolates were retested centrally by the BSAC agar dilution method.

Results: The isolates collected comprised Staphylococcus aureus (39.9%), Escherichia coli (13.3%), streptococci (11.9%), enterococci (6.3%), Klebsiella (6.1%), coagulase-negative staphylococci (6.1%), Enterobacter spp. (3.9%), Proteus spp. (2.2%) and other Gram-negative species (10.3%). The laboratories' disc susceptibility testing found that 4% of isolates were resistant to tigecycline, using the zone breakpoints in place at that time. However, centralized retesting by agar dilution showed that many of these 'resistant' isolates were susceptible, with resistance only confirmed in a range of Gram-negative isolates and one S. aureus. These findings reduced the estimated resistance rate to 2.4%, or to 0.8% if Proteus spp. were ignored as intrinsically resistant to tigecycline. Sixty-two percent of the isolates found resistant by the disc method but susceptible by agar dilution had zones of inhibition within experimental error (4 mm) of the published breakpoints.

Conclusions: Tigecycline resistance was rare in isolates causing clinically significant infections in the UK and was overestimated ∼2-fold by the BSAC disc method. Adjustment of the breakpoints might overcome this problem but at the risk of increasing the false susceptibility rate. The best advice is to perform dilution tests, e.g. by gradient strip test on isolates with borderline results, especially in severe infection and when tigecycline use is intended.