Polymyxin B- Compared to Beta-Lactam- Based Regimens for the Treatment of Carbapenem Resistant Gram Negative Bacterial Pneumonia

Background:Infections caused by multidrug resistant (MDR) gram-negative bacteria (GNB) are an increasingly common problem in hospitalized patients. As conventional antimicrobials prove to be ineffective, this therapeutic challenge may be met by polymyxin B (PB). The use of PB for treatment of pneumonia (PNA) is controversial as previous studies have reported reduced efficacy of PB, possibly due to limited lung penetration. Nonetheless, PB is frequently used with variable success. We aim to characterize the clinical effectiveness of PB for the treatment of MDR GNB PNA.

Methods: This is a retrospective cohort study of patients from Columbia University Medical Center who had PNA with a respiratory culture positive for a carbapenem resistant GNB in 2014. Subjects met National Healthcare Safety Network criteria for PNA and received at least 48 hours of active antibiotic treatment. We compared subjects who received a PB based regimen to those who received a beta lactam (BL) based regimen (without PB). The primary outcome was 30-day all-cause mortality; additional outcomes were clinical response rate and nephrotoxicity.

Results: 50 patients met inclusion criteria; mean age was 65 years and 64% were male. Cultures were comprised of Pseudomonas aeruginosa (52%), Acinetobacter baumannii (26%), and Klebsiella pneumoniae (16%). 22 patients (44%) received a PB containing regimen. The PB group were more likely to be in an intensive care unit (ICU) (82% vs. 46%) compared to the BL group. 30-day mortality was 55% and 18% in the PB and BL groups respectively (p=0.023). Clinical response rate was 50% in PB group and 82% in the BL group (p=0.035). In a multivariable model adjusting for age, receiving an aminoglycoside, and ICU stay, the odds ratio of clinical success was 0.129 (95% confidence interval 0.02-0.73, p=0.02) for those with a PB vs. BL regimen. Of 28 patients evaluable for nephrotoxicity, 9 (75%) in the PB group developed acute renal failure on treatment compared to 4 (25%) in the BL group.

Conclusion: A PB based regimen for the treatment of MDR GNB PNA resulted in higher mortality and lower odds of clinical success compared to those who received a BL regimen, though those who received PB were more critically ill. These data highlight the need for better therapeutic options for MDR GNB PNA.