Risk Factors For Polymyxin B-associated Acute Kidney Injury

ABSTRACT

Objectives: This study aimed to assess the current incidence and risk factors for polymyxin B-associated acute kidney injury (AKl) in Chinese hospitals for more effective clinical use for polymyxin B.Methods: This multicenter, retrospective cohort study included patients from 14 Chinese teaching hospitals who received polymyxin B therapy. Univariate and multivariate logistic regression models were used to determine the factors associated with polymyxin B-associated incident AKl. Furthermore, a multivariate logistic regression model was used to identify the independent risk factors for AKI.

Results: A total of 251 patients were included in the analysis. The overall incidence of AKl was 33.5% A multivariate logistic regression model identified the loading dose (hazard ratio (HR), 1.84; 95% conf. dence interval (Cl), 1.01-3.38; P 0.0491) and the use of two or more nephrotoxic drugs (HR, 3.56; 95%Cl,1.55-8.18; P= 0.0029) as independent risk factors for the occurrence of AKl. Meanwhile, the esti.mated glomerular filtration rate had a protective effect (HR, 0.99; 95% C, 0.98-0.99; P= 0.0006) on the occurrence of AKl. The daily dose, cumulative dose, and treatment duration of polymyxin B did not affect the occurrence of AKl.

HOW LONG DOES IT TAKE FOR CISTANCHE TO WORK FOR KIDNEY PATIENTS?

Introduction 

Nephrotoxicity is an important factor in evaluating the clinical effectiveness of polymyxins. When polymyxins had just been used in the clinics, the incidence of adverse events was very high, especially nephrotoxicity and neurotoxicity; although the definition of nephrotoxicity is not well established, studies have reported that its incidence was as high as 10%–50% (Falagas and Kasiakou, 2006). The high incidence of nephrotoxicity and the emergence of new antibiotics with low toxicity (such as aminoglycosides and second-generation and third-generation cephalosporins) have resulted in fewer polymyxins being used in the clinic since the 1970s.

Currently, many studies are evaluating polymyxin-associated nephrotoxicity, and more than 80% of them are specific to colistin. Kvitko et al. (2011) found that 36% (16/45) of patients receiving polymyxin B for treating bacteremia caused by Pseudomonas aeruginosa infection developed nephrotoxicity compared with only 11% (10/88) of patients who developed the same condition upon receiving other antibiotics (P = 0.002). In another study, Paul et al. (2010) compared the incidence of nephrotoxicity after treatment with colistin or other antibiotics for infections with P. aeruginosa, Acinetobacter baumannii, or Enterobacteriaceae; they found that when the former was used, nephrotoxicity's incidence was 16% (26/128) compared with only 7% (17/244) in the control group (P = 0.006). However, the incidence of polymyxin B-related nephrotoxicity reported in the literature varies widely, ranging from as low as 4% (Ramasubban et al. 2008) to as high as 60% (Kubin et al. 2012). The main reasons for such inconsistent findings across these studies are the differences in the dosage of polymyxins and disparate definitions of nephrotoxicity. For example, studies using the RIELF (Risk, Injury, Failure, Loss, and End-stage kidney disease) nephrotoxicity criteria reported a 39% (122/330) incidence of nephrotoxicity (Esaian et al. 2012; Kubin et al. 2012; Akajagbor et al. 2013; Phe et al. 2014), while those using other criteria reported a 17% (16/96) incidence (Ouderkirk et al. 2003; Sobieszczyk et al. 2004; Oliveira et al. 2009). Noteworthy, the most commonly used nephrotoxicity criteria in published clinical studies are the RIELF criteria. The Kidney Disease: Improving Global Outcomes (KDIGO) criteria for diagnosis were introduced later, and some studies have shown that more patients with AKI can be identified using the KDIGO criteria (Zhou et al. 2016).

Due to the continuous increase in carbapenem-resistant organisms, polymyxins (including colistin and polymyxin B) were re-introduced into clinical practice in the 1990s. However, the nephrotoxicity of polymyxins remained a concern for the clinicians. Polymyxin B entered the Chinese market at the end of 2017, and since then, its irregular use has become very common (such as no loading dose, low maintenance dose, etc.). Therefore, this study aimed to assess the current incidence and risk factors of polymyxin B-related nephrotoxicity in Chinese hospitals to guide clinicians toward a more effective utilization of polymyxin B.

Methods

Study design

Patient selection criteria

Adult patients(>18 years)diagnosed with hospital-acquired pneumonia (HAP)due to carbapenem-resistant A. baumannii(CRAB)or carbapenem-resistant Enterobacteriaceae who received intravenous polymyxin B was included in this study. The patients were divided into two groups based on the presence and absence of AKl: the AKl group and the non-AKl group. Patients with AKl at the time of HAP diagnosis and those who died within 48 h of polymyxin B use were excluded from this study.

Study variables definitions 

The primary outcome of this study was the occurrence of AKI during hospitalization. The definition and staging of AKI were based on the KDIGO standards (Khwaja, 2012). All potential confounding variables were collected, including demographics, underlying conditions, Charlson comorbidity index (Charlson et al. 1987), mechanical ventilation, laboratory tests, acute physiology, and chronic health evaluation II score (Knaus et al. 1985), sequential organ failure assessment score (Vincent et al. 1996), microbiological data, characteristics of polymyxin B use, and concomitant nephrontoxins. HAP was defined as new pneumonia (a lower respiratory tract infection verified by the presence of a new pulmonary infiltrate on imaging) that developed more than 48 hours after admission in non-intubated patients (Modi and Kovacs, 2020). The polymyxin B loading dose was defined as the first dose exceeding the maintenance dose. The definition of ideal body weight was as previously described (Lee et al. 2015). The estimated glomerular filtration rate (eGFR) was calculated according to a previous formula (Teo et al. 2011).

Statistical analysis

included in a stepwise multivariate logistic regression model with an entry criterion of P < 0.20 and an exit criterion of P > 0.05. All P values were two-sided and P < 0.05 was considered statistically

significant.

Results

Baseline characteristics of patients 

A total of 251 patients were included in this study for analysis. There were 84 patients in the AKI group and 167 patients in the non-AKI group. The demographic and clinical information of these patients is summarized in Table 1. The overall incidence of AKI was 33.5% (84/251), and a total of 176 (70.1%) patients were male. The most common bacteria isolated were carbapenem-resistant A. baumannii (136/251; 64.9%) and carbapenem-resistant Klebsiella pneumoniae (153/251; 60.9%). Of all the patients, 67.3% (169/251) had infections from more than one pathogen, and 27 patients had a history of chronic kidney disease (CKD). There was no significant difference between the AKI and non-AKI groups in terms of baseline serum creatinine levels (66.30 ± 29.70 vs 61.65 ± 36.60; P = 0.0552). Furthermore, there were no significant differences in disease severity between the two groups, namely in terms of the SOFA (sequential organ failure assessment) (6.25 ± 3.42 vs 5.68 ± 3.18; P = 0.1946) and APACHE II (acute physiology and chronic health evaluation II score) (16.61 ± 6.07 vs 16.98 ± 7.00; P = 0.6811) scores. In addition, the sepsis biomarkers procalcitonin and C-reactive protein were also not found to differ statistically between the two groups. However, the AKI group showed lower eGFR and higher creatinine levels when HAP was diagnosed, despite being within the normal range (94.90 ± 35.19 vs 112.55 ± 39.66; P = 0.0008; 71.50 ± 42.00 vs 57.95 ± 40.80; P = 0.0008, respectively). In the AKI group, the proportion of patients whose daily exposure dose exceeded the recommended dose was higher than that in the non-AKI group (14.3% vs 4.2%). The drug use rate and number of nephrotoxic drugs in the AKI group were also significantly higher than those in the non-AKI group.

Risk factors related to the occurrence of AKI 

The risk factors associated with the occurrence of AKI are listed in Table 2. In the multivariate analysis, only three variables (eGFR, loading dose, and use of two or more nephrotoxic drugs) showed an independent correlation with the occurrence of AKI after adjusting for underlying confounders. Among them, eGFR had a protective effect (HR, 0.99; 95% CI, 0.98–0.99; P = 0.0006) on the occurrence of AKI. The loading dose (HR, 1.84; 95% CI, 1.01–3.38; P = 0.0491) and the use of two or more nephrotoxic drugs (HR, 3.56; 95% CI, 1.55–8.18; P = 0.0029) were independent risk factors for AKI. Notably, the analysis results showed that some factors of concern such as the daily dose/actual body weight, cumulative dose, and treatment duration did not affect the occurrence of AKI.

Risk factors related to the severity of AKI

The risk factors associated with the severity of AKI are listed in Table 3. We found that patients with higher baseline creatinine levels may have more severe AKI. Whether the patient had CKD in the past and the severity of the disease was not directly related to the severity of AKI.