A scoring system to predict resistance to ceftolozane/tazobactam in respiratory isolates of Pseudomonas aeruginosa.

Objectives: To develop a scoring system to predict resistance to ceftolozane/tazobactam in Pseudomonas aeruginosa strains isolated from respiratory specimens.

Methods: A case-control study was conducted to evaluate the risk factors associated with resistance to ceftolozane/tazobactam. Patients with P. aeruginosa were defined as cases if they had ceftolozane/tazobactam-resistant strains, whereas those with ceftolozane/tazobactam-susceptible strains were defined as test-negative controls. A predictive scoring system based on binary logistic regression coefficients was formulated to predict resistance to ceftolozane/tazobactam. The score's performance was assessed using ROC curves and AUC. The sensitivity, specificity and predictive values of the score were determined on the basis of a cut-off point, using the Youden index.

Results: Ceftolozane/tazobactam resistance was detected in 18.4% of P. aeruginosa isolates from 473 patients. In multivariate analysis, a history of bronchoscopy [OR (95% CI) = 2.1 (1.1-4.3), P = 0.035], invasive mechanical ventilation [OR (95% CI) = 2.4 (1.2-4.5), P = 0.009], colistin/polymyxin B use [OR (95% CI) = 3.2 (1.8-5.7), P < 0.001] and fluoroquinolone use [OR (95% CI) = 2.3 (1.1-4.8), P = 0.024] in the preceding month prior to P. aeruginosa isolation were significantly associated with ceftolozane/tazobactam resistance. The AUC (95% CI) of the score was 0.734 (0.675-0.794), with a sensitivity of 69%, specificity of 71.8%, positive predictive value of 35.5% and negative predictive value of 91.1% at the cut-off point of 2, out of a range of 0-5.

Conclusions: In respiratory tract infections caused by P. aeruginosa, use of the proposed scoring system may reduce inappropriate use of ceftolozane/tazobactam in empirical treatment.