Session: Poster Session: Diagnostic Testing 2008
Tuesday, October 28, 2008: 12:00 AM
Room: Hall C
Background: Automated microscopy can measure growth rates of small numbers of immobilized bacteria to rapidly identify MRSA. We report performance of this method using S. aureus (SA) clinical isolates, and retrospective chart review to assess potential impact on antimicrobial prescribing. Methods: Frozen SA isolates were resuspended after 24-h growth on blood agar. After 2-h, 10 μL aliquots of 1x1E6 CFU/mL were pipetted into microfluidic flowcells. 200-500 bacteria were immobilized onto poly-L-lysine coated glass. An image analyzer measured growth rates after 1-h induction by 1 μg/mL then 3-h of 6 μg/mL cefoxitin. SA was classified as MRSA or MSSA according to growth arrest within the observation time. Results were compared to the original culture. Charts were abstracted for antimicrobial days of therapy against SA for 3 days after culture obtained, and categorized as active/optimal; suboptimal/redundant; or inactive. Results: The microscopy method correctly classified 17/19 MSSA and 39/39 MRSA. The 2 incorrectly classified isolates were small colony variants from the same patient. Sensitivity and specificity in identifying MRSA was 100% [0.88-1] and 89% [0.65-0.98], respectively. For 52 charts reviewed, 208 days of therapy were prescribed vs SA in the 3 days while awaiting culture results. Based on SA susceptibility, 96/208 (46%) days of therapy were considered active/optimal; 55/208 (26%) days were suboptimal/redundant; and 57/208 (27%) days were inactive. If used directly with patient specimens, the new method could have raised the active/optimal rate from 46% to 97%. Conclusions: The new method effectively identified MRSA within 4 hours using small numbers of bacteria from clinical isolates with an inoculum size compatible with direct extraction from clinical specimens. It shows promise for reducing the high rate of inactive and suboptimal/redundant therapy observed while awaiting results from conventional cultures.