The Importance of Addressing Multidrug Resistance and Not Assuming Single-Drug Resistance in Case-Control Studies

Background. Case-control studies analyzing antibiotic exposure as a risk factor for antimicrobial resistance usually assume single-drug resistance in the bacteria under study, even though resistance to multiple antimicrobials may be present. Since antibiotic selection pressures differ depending on the susceptibility profile of the antimicrobial-resistant bacteria, an accurate assessment of whether exposure to an individual antimicrobial is a risk factor for the emergence of resistance should distinguish between single-drug–resistant and multidrug-resistant bacteria. Objective. To determine whether the exposures to individual antibiotics that were identified as independent risk factors in case-control studies differed depending on whether single-drug–resistant or multidrug-resistant bacteria were evaluated. Design. Two retrospective case-control studies were performed with data on patients harboringPseudomonas aeruginosastrains resistant only to ciprofloxacin (CRPA) and patients harboringP. aeruginosastrains resistant to ciprofloxacin and other antibiotics (multidrug-resistantP. aeruginosa[MDR-PA]). These 2 groups were compared with patients not harboringP. aeruginosa. Setting. Two tertiary care hospitals. Results. A total of 41 patients harboring CRPA and 151 patients harboring MDR-PA were identified and matched to 192 control subjects. By conditional logistic regression, independent risk factors associated with presence of CRPA were nonambulatory status (OR, 5.6 [95% confidence interval {CI}, 1.4-23];P= .02) and prior ciprofloxacin exposure (OR, 5.0 [95% CI, 1.2-21];P= .03). Independent risk factors for presence of MDR-PA were a Charlson score greater than 2 (OR, 3.3 [95% CI 1.8-6.0];P<.001) and exposure to quinolones (OR, 2.8 [95% CI, 1.2-5.0];P= .001), third- and fourth-generation cephalosporins (OR, 3.5 [95% CI, 1.7-7.1];P<.001), imipenem (OR, 3.8 [95% CI, 1.2-12.1];P= .02), and/or aminoglycosides (OR, 2.3 [95% CI, 1.04-5.1];P= .04). Conclusion. There were substantial differences in exposure to individual antimicrobials between patients harboring CRPA and patients harboring MDR-PA. Future case-control studies addressing risk factors for single-drug–resistant bacteria should consider the complete susceptibility profile of the bacteria under investigation.