Polymyxin B MIC >1 mg/L Predicts Reduced Clinical Cure and PK/PD Target Attainment in Carbapenem-Resistant Enterobacterales Infections

A prespecified secondary analysis has identified a polymyxin B查看药品详情 (PMB) minimum inhibitory concentration (MIC) threshold of >1 mg/L as an independent predictor of reduced clinical cure and markedly lower pharmacokinetic/pharmacodynamic (PK/PD) target attainment in patients with carbapenem-resistant Enterobacterales (搜索)搜索该疾病 (CRE) infections. The single-center cohort study, published in Frontiers in Pharmacology, integrated PK/PD analyses with real-world clinical outcomes to evaluate the clinical risk-stratification value of this MIC threshold.

The study included 98 non-overlapping PMB treatment episodes contributed by 94 patients, stratified into a low-MIC group (≤1 mg/L, n = 61) and a high-MIC group (>1 mg/L, n = 37). The MIC distribution in the initial culture was 0.25 mg/L (n = 3), 0.5 mg/L (n = 29), 1 mg/L (n = 29), and 2 mg/L (n = 37). Notably, 16 patients experienced MIC drift during subsequent treatment, including 12 patients whose initial MIC of 2 mg/L shifted to ≥4 mg/L.

Clinical and Microbiological Outcomes

Clinical cure was significantly less frequent in the high-MIC group than in the low-MIC group (27.0% vs. 55.7%, P = 0.007). Microbiological eradication was also less frequent in the high-MIC group (32.4% vs. 55.7%, P = 0.036). No statistically significant differences were observed in 28-day mortality (24.3% vs. 19.7%, P = 0.618) or 90-day mortality (37.8% vs. 26.2%, P = 0.262).

After adjustment in a multivariable model, PMB MIC >1 mg/L remained independently associated with lower odds of clinical cure (adjusted OR = 0.311; 95% CI, 0.119–0.809; P = 0.017). The association was consistent in a parsimonious model (adjusted OR = 0.349; 95% CI, 0.148–0.821; P = 0.016). Baseline Sequential Organ Failure Assessment (SOFA) score was also independently associated with lower odds of clinical cure in the parsimonious model (adjusted OR = 0.871; 95% CI, 0.776–0.978; P = 0.020).

Sensitivity Analyses

The association between MIC >1 mg/L and reduced clinical cure remained significant in a generalized estimating equation model accounting for within-patient correlation (adjusted OR = 0.309; 95% CI, 0.116–0.823; P = 0.019) and when analysis was restricted to the first eligible episode per patient (n = 94; adjusted OR = 0.347; 95% CI, 0.132–0.912; P = 0.032).

In a sensitivity analysis restricted to episodes with MIC ≤2 mg/L (n = 86), which excluded the 12 episodes with MIC ≥4 mg/L that emerged during follow-up, the direction and magnitude of the association between MIC = 2 mg/L and clinical cure remained similar, although statistical precision was reduced and the confidence interval crossed unity (adjusted OR = 0.350; 95% CI, 0.118–1.037; P = 0.058). In this restricted cohort, clinical cure occurred in 34/61 episodes with MIC ≤1 mg/L and 8/25 episodes with MIC = 2 mg/L.

PK/PD Target Attainment

Although model-predicted steady-state area under the curve (AUCss, 24 h) was similar between groups (67.3 [IQR, 53.8–91.1] vs. 65.4 [53.6–82.0] mg·h/L; P = 0.823), median AUC/MIC was substantially lower in the high-MIC group than in the low-MIC group (27.4 [IQR, 12.8–43.1] vs. 98.2 [64.3–142.4]; P < 0.001). This indicates that the reduction in target attainment was primarily driven by the higher MIC denominator rather than lower systemic exposure.

Probability of target attainment (PTA) for AUC/MIC ≥50 was 90.2% (55/61) in the low-MIC group and 13.5% (5/37) in the high-MIC group (P < 0.001). The between-group difference remained significant using thresholds of 40 (96.7% vs. 27.0%) and 60 (77.0% vs. 2.7%; both P < 0.001).

Safety Profile

No significant between-group differences were observed in PMB-associated adverse events. New-onset acute kidney injury (AKI) occurred in 37.8% of high-MIC episodes and 26.2% of low-MIC episodes (P = 0.262). Neurotoxicity (21.6% vs. 9.8%, P = 0.139) and hepatotoxicity (8.1% vs. 11.5%, P = 0.738) were also comparable. A loading dose was administered in 48/98 episodes.

Clinical Implications

The current CLSI classification assigns an "intermediate" category to Enterobacterales isolates with polymyxin MIC ≤2 mg/L and does not provide a susceptible category, reflecting uncertainty regarding reliable efficacy at clinically tolerable exposures. The authors emphasize that these findings "should not be interpreted as sufficient evidence to revise a clinical breakpoint; rather, they support the potential use of MIC >1 mg/L as a clinical risk-stratification threshold and provide hypothesis-generating evidence for prospective breakpoint evaluation."

From a clinical perspective, PMB MIC >1 mg/L should prompt heightened concern regarding inadequate PK/PD target attainment. For MIC = 2 mg/L isolates, standard exposure frequently failed to achieve AUC/MIC ≥50, and dose escalation may be constrained by nephrotoxicity. If an active non-polymyxin option is available, it should generally be prioritized. When PMB remains necessary, confirmation of MIC by reference broth microdilution, early therapeutic drug monitoring where available, individualized PK/PD assessment, source control, and close nephrotoxicity monitoring are reasonable.

The study's strengths included the use of reference broth microdilution with routine external quality assurance and routine repeat testing before final reporting. However, the authors acknowledged several limitations: a single-center mixed retrospective–prospective cohort with a modest sample size and residual confounding, outcome assessors who were not blinded to microbiological or treatment information, exposure estimated using an a priori population model rather than measured concentrations, and the absence of carbapenemase genotype, polymyxin resistance mechanisms, and heteroresistance assessment. Accordingly, the observed associations should be interpreted as clinically informative but not causal, with external validation in larger, prospectively sampled cohorts required.