Risk Factors and Outcomes for Isolation with Polymyxin B-Resistant Enterobacterales from 2018–2022: A Case-Control Study

Introduction

Multidrug-resistant Gram-negative bacilli have increased worldwide and pose a significant and growing threat that limits the effectiveness of therapeutic regimens.¹ Conventional drugs, such as polymyxin E (colistin) and polymyxin B, are being reconsidered as the last line of defense against multidrug-resistant Gram-negative bacilli infections.² Owing to the increased use of colistin in clinical settings, resistance to this antibiotic has increased, particularly among Klebsiella pneumoniae isolates, which should be monitored.^3–5 Polymyxin B-resistant (PR) bacterial infections have been reported in both nosocomial and community settings.⁶

The mechanisms underlying polymyxin B resistance in Gram-negative bacilli mainly involve intrinsic mechanisms, mutations, or adaptation. The selective pressure applied by the polymyxin dose regulates the dynamics of genetic variants within the population. To this end, the super minimum inhibitory concentration (super-MIC) levels (> 4 x MIC) of polymyxin B were reported to have caused the development of irreversible resistance in Acinetobacter baumannii.⁷ It has been suggested that the principal resistance to this antibiotic in Enterobacterales is associated with lipopolysaccharide modification, which is mediated by charged 4-amino-4-deoxy-L-arabinose (L-Ara4N; mediated by arnBCADTEF) and/or phosphoethanolamine (pEtN; mediated by pmrC, or mobile colistin resistance gene).^8–10 The mobile colistin resistance gene (mcr-1) and its variants, mcr-2 to mcr-10, mediate polymyxin B resistance.^6,11–13 A recent study showed that New Delhi metallo-β-lactamase (NDM)-producing MDR K. pneumoniae can rapidly attain high-level polymyxin resistance through cellular metabolism plasticity, lipid A modification, and outer membrane remodeling.¹⁴

In addition to monitoring the rate of resistance to polymyxins and underlying mechanisms, the identification of relevant risk factors in patients with colistin-resistant bacteria is important to prevent the spread of nosocomial infection outbreaks. Several studies have identified risk factors related to patients with colistin-resistant and/or carbapenem-resistant Enterobacterales infection or colonization.^4,15–17 However, studies focused on risk factors in Henan Province, China, are limited.

The relevant risk factors and outcomes of patients carrying PR Enterobacterales were retrospectively analyzed to provide a reference for the prevention and control of PR Enterobacterales infection outbreaks in hospital.

Materials and Methods

Study Participants

This case-control study was performed at a tertiary care teaching hospital with 5000 beds in Henan Province, China, between January 2018 and December 2022. The risk factors related to the isolation of PR Enterobacterales were investigated. The case group (n = 72) included patients with PR Enterobacterales isolated from various clinical specimens during the study period. Patients with polymyxin B-susceptible (PS) Enterobacterales isolated from clinical cultures were controls. The ratio of control to case samples was 2:1. Controls were matched for hospital ward, date of hospitalization, and pathogen with the case in a random manner. The exclusion criteria (n = 23) were as follows: patients with PR Enterobacterales isolation from intestinal screening samples during the study period and inadequate clinical or laboratory data, such as transfer or discharge without relevant clinical diagnoses and treatment processes, resulting in unclear patient outcomes. Patients with clinical cultures of Enterobacterales (eg, Providencia spp., Proteus spp., Morganella morganii, and Serratia spp.) intrinsically resistant to polymyxins were also excluded.

Definition of Outcomes

The outcomes of patients in this study were defined as follows: Treatment failure included patients who died for any reason or were discharged without improvement after the isolation of polymyxin B-resistant strains.

Clinical Data

Data were obtained from electronic medical records. Potential risk factors included demographic properties (age and sex), medical history (hospitalization, transfers between hospitals and departments, intensive care unit [ICU] admission, and surgery), comorbidities, inpatient time, admission diagnosis, antibiotic use, use of invasive medical devices (urinary catheter, mechanical ventilation, drainage tube, and nasogastric tube), and use of immunosuppressive drugs. Diagnosis was classified using the International Classification of Diseases-11 (ICD-11) based on the patient’s admission diagnosis.¹⁸

Microbiological Methods

Bacterial species identification and antimicrobial susceptibility testing were performed using the Phoenix100 automated system according to the manufacturer’s recommendations (Becton Dickinson Co., Sparks, MD, USA). The identification of Salmonella spp. was based on a serum agglutination test. The polymyxin B MICs of the isolates were determined by broth microdilution. The results were interpreted using the clinical breakpoints of the European Committee on Antimicrobial Susceptibility Testing.¹⁹ Isolates with an MIC higher than 2 mg/L for polymyxin B were considered PR. Escherichia coli ATCC 25922 and Escherichia coli NCTC 13846 (mcr-1-positive) were used as quality control strains.

Statistical Analyses

Categorical variables are reported as n (percentages); continuous variables with a normal distribution are expressed as means ± standard deviations (SD); and continuous variables with a non-normal distribution are described as medians (interquartile ranges: IQRs). Categorical data were analyzed using chi-squared or Fisher’s exact tests; the Student’s t-test or the Mann–Whitney U-test was used to analyze continuous variables. Odds ratios (ORs) with 95% confidence intervals (CIs) and P-values were used for univariate analyses. Variables with a P–value < 0.05 in the univariate analyses were entered into multivariate logistic regression. Survival curves (150 day) were created using the Kaplan–Meier method. Log rank tests were performed using Prism 8.0 (GraphPad). Other statistical analyses were performed using SPSS (version 25.0; IBM Corporation, Armonk, NY, USA). P < 0.05 was considered statistically significant.

Results

Characteristics of PR Isolates

As shown in Figure 1, the most commonly identified PR Enterobacterales was K. pneumoniae (72.2%; 52/72), followed by Salmonella enteritidis (8.3%; 6/72), and Escherichia coli (5.6%; 4/72). Among the PR isolates, 34.7% (25/72) were isolated from sputum and BALF, 22.2% (16/72) from urine, 20.8% (15/72) from blood, and the remainder from other samples.

Figure 1 Types (A) and sample sources (B) of polymyxin B-resistant Enterobacterales. Abbreviations: Kpn, Klebsiella pneumoniae; Sen, Salmonella enteritidis; Eco:, Escherichia coli; Eae, Klebsiella aerogenes; Koz, Klebsiella ozaenae; Sal, Salmonella spp.; Sdu, Salmonella Dublin; Ecg, Enterobacter cancerogenus; Ecl, Enterobacter cloacae; BALF, Bronchoalveolar lavage fluid.

Antimicrobial susceptibility testing revealed that the resistance rate for 52 K. pneumoniae isolates was significantly higher than that of the other 20 Enterobacterales isolates. In particular, 76.9% (40/52) of K. pneumoniae isolates were carbapenem-resistant (Figure 2).

Figure 2 Comparison of antibiotic resistance percentages between 52 strains of polymyxin B-resistant Klebsiella pneumoniae and 20 strains of other Enterobacterales. Abbreviations: AMP, ampicillin; PIP, piperacillin; SAM, ampicillin-sulbactam; AMC, amoxicillin-clavulanic acid; TZP, piperacillin-tazobactam; CAZ, ceftazidime; CTX, cefotaxime; FEP, cefepime; ATM, aztreonam; MEM, meropenem; IPM, imipenem; CIP, ciprofloxacin; LEV, levofloxacin; CHL, chloramphenicol; SXT, trimethoprim sulfamethoxazole; TE, tetracycline.

Patient Characteristics

A total of 72 patients carrying PR Enterobacterales were included during the 5-year study period, and 144 patients carrying PS Enterobacterales were included as controls. The median ages of cases and controls were 63 (51–73) and 62 (48–73) years, respectively (P = 0.391), and 66.67% of patients were male. Of the study patients, 63.89% were admitted to the ICU. Patient characteristics are shown in Table 1.

Table 1 Univariate and Multivariate Analyses of Risk Factors in Patients Infected with Polymyxin-Resistant Enterobacterales Strains

Risk Factors for Isolation with PR Strains

Table 1 compares the clinical characteristics of patients with PR and PS controls. Univariate analysis revealed that chronic heart disease, chronic kidney disease, organ transplantation, immunosuppressant use, invasive procedures (central venous catheter, chest drainage tube [chest], and head drainage tube [head]), and polymyxin B exposure were associated with the isolation of PR Enterobacterales.

Chronic heart disease (P = 0.012; odds ratio [OR] 1.15; 95% confidence interval [CI] 1.03–1.28), immunosuppressant use (P = 0.016; OR 1.04 [1.0–1.07]), drainage tube [head] (P = 0.006; OR 1.1 [1.0–1.1]), and polymyxin B exposure (P = 0.007; OR 1.03 [1.0–1.06]) were independently related to the isolation of PR Enterobacterales in a multivariate analysis (Table 1).

Treatment Failure Among Cases and Controls

Among the 72 patient cases, 33 (45.83%) experienced treatment failure; Conversely, within the cohort of 144 controls, 21 (14.58%) experienced treatment failure. The survival rate was significantly lower in patients in the case group than in the control group in the survival curve analysis (Figure 3).

Figure 3 Survival curves between patients with (case) and without (control) polymyxin B-resistant Enterobacterales isolation.

Univariate analyses revealed that a history of ICU admission, hypertension, nervous system disease, mechanical ventilation, urinary catheter, central venous catheter, and head drainage were associated with treatment failure (Table 2). In multivariate analysis, independent risk factors for treatment failure included admission to an ICU (P = 0.003; OR 7.1 [1.9–25.4]), hypertension (P = 0.035; OR 1.4 [1.02–1.83]), and drainage tube [head] (P = 0.044; OR 1.1 [1.0–1.15]) (Table 2).

Table 2 Univariate and Multivariate Analyses of Risk Factors for Treatment Failure in Patients with Polymyxin B-Resistant Enterobacterales Infection

Discussion

Increased polymyxin consumption for multidrug-resistant Gram-negative bacilli treatment has resulted in the development of polymyxin B resistance by these species in several countries.²⁰ This is an important public health issue. Monitoring the resistance mechanism and elucidating risk factors for polymyxin B resistance may be helpful for preventing the spread of resistant strains.

In this study, the most common PR Enterobacterales was K. pneumoniae (72.2%), consistent with previous results.^16,21,22 Additionally, 76.9% (40/52) of PR K. pneumoniae isolates were carbapenem resistant. The resistance rates of carbapenem-resistant K. pneumoniae to colistin in Taiwan²¹ and Oman²³ were 17% and 24.2%, respectively. We found that polymyxin B exposure is a risk factor for hospitalized patients with PR Enterobacterales, which is consistent with the results of previous studies.^24,25 However, most of the prior literature focuses on PR combined with carbapenem resistance and/or specific populations, such as critically ill patients, or specific infection types, such as bloodstream infection (BSI). For example, previous polymyxin B exposure was a predictor of the isolation of PR carbapenem-resistant and carbapenemase-producing Enterobacterales or colistin-resistant K. pneumoniae in other studies.^15,24 Colistin was reported to be related to the development of colistin-resistant K. pneumoniae BSI.²⁶ In another study, recent exposure to colistin or polymyxin B was related to increased opportunities for colistin resistance.²⁷ Except for polymyxin B exposure and carbapenem resistance, previous exposure to carbapenems was the risk factor for colonization or infection with PR E. coli or K. pneumoniae in a low-endemicity setting.¹⁷

Chronic heart failure, use of immunosuppressive agents, and head drainage were also risk factors for PR Enterobacterales in our study population. It was reported that coronary heart disease is an independent risk factor that predisposes patients to PR resistance.²⁸ Chronic renal failure has been shown as a risk factor in critically ill patients with PR and carbapenemase-producing Enterobacterales in a previous study, which also included other risk factors such as surgical procedures, indwelling devices (urinary catheters), and transfer between hospital wards.¹⁶ Other risk factors associated with colistin resistance in Gram-negative bacilli include neurological disease, residence in a nursing facility prior to admission, receipt of carbapenems or an antimicrobial with anti-methicillin-resistant Staphylococcus aureus activity, and receipt of ventilatory support.⁴ In addition, male sex, immunosuppression, and antibiotic use, particularly carbapenems and fluoroquinolones, are associated with mcr-1-positive E. coli infection.²⁹ Increased age (> 55 years) and antibiotic use in the preceding 90 days were associated with an increased risk in patients with colistin-resistant Enterobacterales.³⁰

Regarding patient outcomes, our study demonstrated that the rate of treatment failure in 72 patients carrying PR strains was 45.83%, which is similar to previous results that reported an overall mortality of between 35% and 66%.^16,17,22,31 In bloodstream infections caused by carbapenem-resistant K. pneumoniae, age and colistin resistance increased the 28-day mortality.³² In our study, multivariate regression analysis showed that ICU admission, hypertension, and head drainage were associated with treatment failure, different from a previous study that reported that carbapenem exposure was associated with mortality among adult ICU patients with PR Enterobacterales strains.¹⁶

This study has some limitations. First, the retrospective and single-centre research design could result in selection biases. Second, the association between possible risk factors identified in this study and the isolation of PR strains was only a statistical link and needs to be validated using more samples. Third, we did not evaluate resistance mechanisms in the PR strains. Thus, more surveillance and multicenter studies are required.

Conclusions

In conclusion, this study revealed that the common bacteria resistant to polymyxin B in Henan, China, are K. pneumoniae, S. enteritidis, and E. coli, and that more than half of K. pneumoniae are also resistant to carbapenem. Risk factor analysis showed that chronic heart failure, use of immunosuppressive agents, head drainage, and polymyxin B exposure are related to the acquisition of PR strains. ICU admission, hypertension, and head drainage are associated with treatment failure. Therefore, hospitals should take multiple measures to prevent further outbreaks of PR Enterobacterales.