Polymyxin B in The Treatment of Infections Caused by Multidrug-Resistant Gram-Negative Bacteria in Children: A Retrospective Case Series and A Literature Review

Introduction

Bacterial resistance poses a significant global public health challenge. According to the World Health Organization (WHO) classification of drug-resistant bacteria based on threat severity, Multidrug-resistant Gram-negative Bacteria (MRGN) are defined as being non-susceptible to at least 1 agent in 3 or more antimicrobial categories within the antibacterial spectrum (including drug resistance and intermediaries),¹ including Carbapenem-resistant Enterobacteriaceae (CRE), Carbapenem-resistant Acinetobacter baumannii (CRAB), Carbapenem-resistant Pseudomonas aeruginosa (CRPA), represent the most critical threats deserving priority attention.² Therefore, it is imperative to actively explore treatment options for these bacteria, encompassing the continuous development of novel antibiotics, as well as research and application of existing antibiotics.

Polymyxins are important antibiotics used to combat multidrug-resistant MRGN. However, their were limited due to potential nephrotoxicity and the availability of alternative antibiotics in recent years.³ In 1990s, polymyxins experienced a resurgence in clinical practice due to the emergence of MRGNs and the scarcity of effective new antibiotics.⁴ For the past few years, new β-lactams such as ceftazidime-avibactam have emerged as the primary drugs recommended in guidelines for managing MRGN infections.^5,6 However, given the diverse distribution of carbapenemase across different regions and populations⁷ and the challenges related to accessibility and affordability of β-lactam drugs, polymyxins continue to hold significance in the treatment of MRGN infections in Chinese children.

For invasive infections, the guideline recommended systemic use of polymyxin B (PMB) over colistin. The reason for the recommendation was that PMB had better pharmacokinetic properties and lower risk of nephrotoxicity.⁸ Contradictorily, one study found that the incidence of acute kidney injury (AKI) associated with colistin was significantly lower than that associated with PMB, but the 30-day mortality rate was similar.⁹ Another study concluded that nephrotoxicity was more common in patients receiving high-dose colistin than in those receiving PMB.¹⁰ The nephrotoxicity of PMB deserves our attention. Despite having been available for over half a century, studies on the pharmacokinetics and optimal dosing of PMB remain limited, particularly in the context of paediatric use.¹¹ To assess the effectiveness and safety of PMB in treating MRGN infections in Chinese children, a retrospective study was conducted at a specialized tertiary hospital for women and children in Southwest China. Additionally, a comprehensive literature review was undertaken to provide valuable references for the judicious clinical utilization of PMB.

Methods

Retrospective Case Series

Inclusion criteria: ① The paediatric patients aged 0–18 years who received PMB were enrolled in the study who were hospitalized in West China Second University Hospital of Sichuan University, a tertiary teaching hospital special for women and children, Children’s National Regional Medical Center in Southwest China, from October 1, 2019, to November 30, 2022. ② All enrolled patients obtained approval for the use of PMB as part of the pre-authorization management within paediatric antimicrobial stewardship programs (It was identified or strongly suspected to be caused by MRGN infection, particularly Gram-negative bacteria that are resistant to carbapenems but sensitive to PMB. Its use was approved following consultation and agreement between the physicians and pharmacists). Exclusion criteria: ① The patients who did not use PMB finally were excluded even though the approval form was submitted. ② The cases received PMB treatment for less than 3 days because of the difficulty in effectiveness evaluation. This study was approved and supervised by the Ethics Committee of West China Second University Hospital of Sichuan University (No.2022–337).

Data Collection

Demographic data, medication information and laboratory test results were collected through Hospital Information System. Demographic data include age, gender, body weight, underlying medical conditions, and site of infection. Medication information included the route of administration for PMB, dosage, duration of treatment, and any concomitant therapies administered. Laboratory tests include pathogen detection from culture and metagenomic next generation sequencing reports (mNGS) results, antibiotics sensitivity reports, renal function (serum creatinine, urine volume), white blood cell count, neutrophil percentage, C-reactive protein (CRP), and procalcitonin (PCT). Furthermore, the evaluation of organ dysfunction and mortality risk prior to the initiation of PMB therapy was conducted using the Paediatric Sequential Organ Failure Assessment score (pSOFA)¹² and the Paediatric Risk of Mortality III score (PRISM III).¹³ These scores provided valuable insights into the patients’ clinical status and prognosis before the administration of PMB.

Clinical Effectiveness

Refer to the US Food and Drug Administration (FDA) guidelines for the evaluation of clinical research on anti-infective drugs, and the American Society of infectious diseases (IDSA) guidelines for the evaluation of infectious diseases,¹⁴ the therapeutic effect of PMB was divided into “effective” and “ineffective”. The favourable clinical response was defined as:

① For febrile patients, the body temperature decreased within 72 hours.

② For patients exhibiting a high inflammatory index (including neutrophil count, CRP, and PCT), a significant reduction in the inflammatory index was observed during medication, resulting in either a cure or noticeable improvement within 3 days.

③ Improvement in the patient’s symptoms, physical signs, as well as positive changes in imaging and laboratory test results were observed, leading to either recovery or discharge. Improvements in these aspects had been assessed by two intensivists and one antimicrobial pharmacist, all of whom possess extensive clinical experience. A unanimous conclusion must be reached ultimately.

The unfavourable clinical response was defined as:

① There was a lack of lack of improvement or a significant worsening of symptoms, physical signs, laboratory examination results, and imaging findings.

② It was impossible to assess treatment effectiveness if the patient either passed away, was discharged automatically, or underwent surgical intervention within 3 days of PMB administration.

③ The treatment schedule was modified (discontinue PMB and switch to alternative agent) due to the patient’s deteriorating condition after initiating PMB therapy.

Adverse Events

In the assessment of adverse drug events associated with PMB, a thorough data cleaning was conducted for relevant symptoms, physical signs, and laboratory examination results documented in the patient’s medical records. Special attention was given to the potential nephrotoxic effects of PMB, in line with guidelines outlined by the FDA. The definition and staging of AKI were based on the RIFLE standards:¹⁵

① Higher Risk: A 1.5-fold increase in serum creatinine (SCr) or a decrease >25% in glomerular filtration rate (GFR), or urine output (UO) <0.5mL/(kg·h) × 6 h.

② Kidney Injury: A 2-fold increase in SCr or a decrease >50% in GFR, or UO <0.5mL/(kg·h) × 12 h.

③ Kidney Failure: A 3-fold increase in SCr or a decrease >75% in GFR. And even if the increase in SCr is less than 3-fold so long as the new SCr value is ≥4.0 mg/dL (350 μmol/L) in the setting of an acute increase of at least 0.5 mg/dL (44 μmol/L), or UO <0.3mL/(kg·h) × 24 h or anuria × 12 h.

The incidence of AKI was calculated as the sum of these three stages and was determined based on the maximum SCr value during PMB administration. This comprehensive approach allowed for a rigorous assessment of nephrotoxicity associated with PMB use.

Statistical Analysis

Normally distributed data were represented using mean and standard deviation (SD), and comparisons were conducted using the t-test. Non-normally distributed count data were presented as median and interquartile range (IQR) and analysed using the Mann–Whitney U-test. Count data were presented as numbers and percentages, and the Chi-square test (χ²) or Fisher’s exact test was utilized for analysis. Due to the limited sample size, the number of independent variables in the logistic regression analysis was based on the “10 events per variable”.^16,17 The selection of these variables referred to the guidelines^8,18 and clinical research.¹⁹ Ultimately, there were three independent variables (whether a loading dose was used, whether it was co-administered with carbapenems, and whether it was co-administered with tigecycline) were included in the logistic regression analysis of clinical effectiveness, and two independent variables (whether a loading dose was used, whether it was co-administered with nephrotoxic drugs) were included in the logistic regression analysis of nephrotoxicity. A significance level of P ≤ 0.05 was used as the criteria for statistical significance. All statistical analyses were conducted in R 4.3.1. with following packages: openxlsx, exactRankTests, stats, plyr.

Literature Review

A literature review was conducted to collect clinical studies focusing on the use of PMB in the treatment of infections caused by MRGN in paediatric patients. Three English databases and three Chinese database, including PubMed, EMBASE, Cochrane Library, China National Knowledge Infrastructure (CNKI), Wanfang, and Chinese Scientific Journals Database (VIP) were searched. The following terms were utilized to search: “PMB”, “children”, “paediatric”, “infant”, “newborn” and “children”. The retrieval time was from the inception of the database to November 2022. The collected data were analysed as follows:

① Patient Demographics: The distributions of country, sample size, age, gender.

② Infection site and pathogen: The frequency and proportion of different infection sites, as well as the types of pathogens responsible for these infections, were tabulated and analysed. Additionally, any descriptions of drug resistance patterns among these pathogens were noted.

③ PMB usage and dose: Descriptive analysis was performed on the utilization of PMB, including the loading dose, maintenance dose, intrathecal dose, duration days.

④ Combination Therapy: Information regarding the choice of antibiotics used in combination therapy alongside PMB was collected and analysed.

⑤ Effectiveness outcomes: Mortality, effectiveness rates, and pathogen clearance rates were related to PMB therapy were recorded and analysed descriptively.

⑥ Safety outcomes: The incidence of adverse reactions, including nephrotoxicity, neurotoxicity, pigmentation, or others, were associated with PMB treatment in paediatric patients.

Results

Characteristics of Patients

There were 57 children who received PMB between October 1, 2019, to November 30, 2022, were initially identified. After screening, 3 cases were excluded from the study due to a treatment course of less than 3 days. Consequently, a final cohort of 54 children aged 0.23–192 months (1 neonate, 17 infants under 1 year old, 19 children aged 1 to 6 years old, 9 children aged 7 to 12 years old, 8 children aged 12 years and above) was included in this analysis. The demographic and clinical characteristics of these children are summarized in Table 1.

Table 1 The Demographic and Clinical Characteristics of 54 Paediatric Patients

Polymyxin B and Combination Therapy

The loading dose of PMB was administered to 35 patients (64.8%). As shown in Table 2, the dose of PMB was divided into three levels: 12 patients (22.2%) received a daily dose of less than 2.5 mg/(kg·d), 27 patients (50.0%) were given a daily dose ranging from 2.5 to 3.5 mg/(kg·d), and 15 patients (27.8%) received a daily dose between 3.5 and 4.0 mg/(kg·d). The median duration of treatment was 16 days (IQR, 14 days to 21 days), and the median cumulative dose was 489.10 mg (IQR, 331.50 mg to 715.00 mg). The execution time (the time from receiving the antibiotics sensitivity test report to the administration of PMB) of 38 patients was recorded, with an equal number of patients, 19 in each category, exceeding 48 hours and ≤48 hours. It is worth noting that there were no significant differences in the usage and dosage of PMB between favourable clinical response group and unfavourable clinical response group.

Table 2 Polymyxin B and Combination Therapy in Patients With Favourable and Unfavourable Clinical Response

Pathogens and Antibiotics Sensitivity Test

The antibiotics sensitivity tests were interpreted clinically with reference to the current clinical breakpoints of United States Committee on Antimicrobial Susceptibility Testing (S ≤ 2 mg/L, R ≥ 4 mg/L).²⁰ A total of 54 samples, bacterial cultures (n = 44) and mNGS (n = 10), were confirmed the presence of MRGN infections in 51 patients (48 patients with a single sample and 3 patients with 2 samples). Additionally, three patients with bone marrow immunosuppression and a strong suspicion of Carbapenem-Resistant Gram-Negative Bacteria (CR-GNB) infections were treated with PMB after prior antibiotic treatments with carbapenem, β-lactam, and tigecycline had proven ineffective. Based on the results of the bacterial cultures, the following MRGN isolates were identified: 15 Klebsiella pneumoniae (13 carbapenem-resistant and 2 extended spectrum β-lactamases), 13 CRAB, 11 P. aeruginosa (7 CRPA and 4 multidrug-resistant) and others (2 carbapenem-resistant Enterobacter cloacae subsp., 1 Enterobacter cloacae subsp., 1 carbapenem-resistant Aeromonas and 1 carbapenem-resistant Escherichia coli), as summarized in Table 3. Notably, 36 of these isolates had polymyxin susceptibility results, and all 36 isolates were susceptible to PMB.

Table 3 Sensitivity, Intermediate and Resistance of Bacterial Strains to Antimicrobial Agents

Clinical Effectiveness

Out of the total participants, 53.7% (29/54) exhibited favourable clinical responses, while 46.3% (25/54) had unfavourable clinical responses. The results revealed that age (P = 0.04), Cr before medication (P < 0.01), and pSOFA (P = 0.05) showed significant differences between the two groups, while the remaining baseline characteristics were balanced between the two groups (Table 1). Specifically, 95.83% (23/24) in the favourable response group and 68.75% (11/16) in the unfavourable response group achieved microorganism clearance (P = 0.03). By individually including each of the three independent variables, the results of logistic regression analysis showed that co-administration with carbapenems was associated with significantly higher effectiveness (odds rate [OR] = 3.16, 95% confidence interval [CI]: 1.02–9.86, P = 0.05). But the loading dose (OR = 0.55, 95% CI: 0.18–1.72, P = 0.31) and co-administration with tigecycline (OR = 0.83, 95% CI: 0.21–3.29, P = 0.79) did not have significantly impact on effectiveness. After simultaneously including all three independent variables in multivariate analysis, none of them, including loading dose (OR = 0.60, 95% CI: 0.18–2.02, P = 0.41), co-administration with tigecycline (OR = 0.67, 95% CI: 0.15–2.91, P = 0.59), and co-administration with carbapenems (OR = 3.07, 95% CI: 0.97–9.75, P = 0.06), showed significant differences.

Among the 54 individuals, 13.0% (7/54) died during hospital stay, of which 3.7% (2/54) died within 30 days after receiving PMB. The age range of all patients who died was 4 months to 158 months. All patients had mixed infections, involving any two combinations of MRGN, viruses, and fungi. One infant aged 4 months was even infected with all three. The diseases and medications of the seven patients who died in the hospital were as follows. Four patients were diagnosed with acute leukaemia (three of them were complicated by graft-versus-host disease). The doses of PMB were 2.5–3.5 mg/(kg·d) and durations were 3–15 days. One patient had severe aplastic anaemia and had been treated with <2.5 mg/(kg·d) of PMB for four days. One patient had T-cell lymphoblastic lymphoma and had been treated with <2.5 mg/(kg·d) of PMB for 22 days. One infant aged 4 months was diagnosed with tricho-hepato-enteric syndrome and had been treated with 3.5–4.0 mg/(kg·d) of PMB for 14 days.

Adverse Events

AKI was observed in 25.9% (14/54) patients with 11.1% (6/54) risk stage, 7.4% (4/54) injury stage and 7.4% (4/54) failure stage. The characteristics of patients with AKI group (n = 14) and without AKI group (n = 40) during treatment with PMB are shown in Supplementary Table 1. There were no significant differences in baseline characteristics between the two groups. No patients adjusted the dose of PMB due to AKI. Renal function recovery was observed in 71.42% (10/14) patients before PMB was discontinued, and one patient recovered renal function one day after PMB was discontinued. Three patients who had progressed to renal failure experienced further deterioration and were eventually discharged, with no additional records available. By individually including each of the two independent variables, the results of logistic regression analysis showed that co-administration with potent diuretic (furosemide) may be a contributing factor that increased the risk of AKI (OR = 4.91, 95% CI: 0.96–24.98, P = 0.05) and loading dose was not a contributing factor to AKI (OR = 0.97, 95% CI: 0.27–3.47, P = 0.96). When both factors were included in the model for analysis, the results showed that neither of them had a significant impact on AKI (loading dose: OR = 0.93, 95% CI: 0.25–3.47, P = 0.92; co-administration with potent diuretic: OR = 4.44, 95% CI: 0.87–22.61, P = 0.07). One patient developed erythema after using PMB, and no instances of neurological toxicity were observed.

Literature Review Results

There were 24 included literature had reported the information about PMB used in children with MRGN.^21–44 A total of 259 children’s data were collected. Among the infection sites, the respiratory system infections were the most prevalent (123 cases), followed by the circulatory system infections (80 cases) and the central nervous system infections (21 cases). The most reported pathogen in these literature was K. pneumoniae (159 cases), followed by the A. baumannii (73 cases), P. aeruginosa (37 cases). The intravenous administration dosage range for PMB was 1.5–4.0 mg/kg, while the intrathecal injection range was 0.2–0.5 mg/kg. Excluding studies with only one case, the mortality rate in the remaining studies ranged from 6.3% to 42.9%, clinical effectiveness rate ranged from 47.8% to 93.7%, and pathogen clearance rate ranged from 30.8% to 100.0%. Correspondingly, the safety outcomes of nephrotoxicity ranged from 0.0% to 82.4%, neurotoxicity ranged from 0.0% to 17.6%, pigmentation ranged from 0.0% to 7.7%. Other adverse events reported in these literature include increased eosinophil count, erythema and papule, and skin desquamation. More details are shown in Supplementary Table 2.

Discussion

In summary, our findings indicated that more than half of patients exhibited favourable clinical responses, with 53.7% (29/54) achieving positive outcomes. 95.83% (23/24) in the favourable response group and 68.75% (11/16) in the unfavourable response group achieved microorganism clearance. 13.0% (7/54) died during their hospital stay. PMB co-administration with carbapenems may be a contributing factor in clinical effectiveness and PMB co-administration with potent diuretic (furosemide) may be a contributing factor in AKI.

In recent years, the emergence of MRGN infections presents a serious clinical challenge. Common antibiotics used to treat MRGN infections include carbapenems, tigecycline, ceftazidime-avibactam, polymyxins. Bacterial resistance to polymyxins is low and they have unique advantages. They can destroy the integrity of the outer membrane of Gram-negative bacteria and have rapid bactericidal activity. They can be used in combination with other antibacterial drugs to exert synergistic anti-infection effects. The bacterial resistance rate is low. Polymyxins has been a vital tool in addressing these infections.⁴⁵ Our study confirmed the presence of 54 bacterial strains in 94.4% (51/54) patients. Antimicrobial drug use management in our institution requires explicit microbiological results and should be reserved for severe infections caused by multidrug-resistant pathogens when other antimicrobial agents are ineffective or not tolerated. Wang et al³⁹ enrolled 19 children aged 3.2–17.8 years to build population pharmacokinetics (PK) and evaluate clinical outcomes of PMB. Clinical success, defined as patient survival, improved microbiological and clinical symptoms, body temperature ≤ 37.5°C, and enhanced infection-related biochemical indicators at the end of PMB treatment, occurred in 73.7% (14/19) of patients. The 28-day mortality was 10.5% (2/19). This study showed a better clinical effectiveness, a lower mortality rate compared to our study. This could be attributed to the smaller sample size, the consistent use of loading doses and looser clinical effectiveness criteria in their study. Jia et al³⁰ conducted a retrospective study to analyse effectiveness during treatment with PMB-based regimen. This study indicated effective rate of 52.7% (29/55). It was similar to our study, including the daily dose of PMB, duration of treatment, and concomitant antibiotics (carbapenems, tigecycline). Notably, our study also suggested that the co-administration of PMB with carbapenems may have influence on clinical effectiveness. One international guideline and two Chinese guidelines recommend PMB combination therapy (based on susceptible MIC) rather than monotherapy.^8,46,47 Two studies comparing colistin monotherapy versus colistin combined with meropenem produced conflicting conclusions, indicating that combination therapy did not improve 28-day mortality and clinical failure.^48,49 Most predominant bacteria were CRAB in these two studies (77% and 78%). The main isolates identified in our study were Klebsiella pneumoniae and P. aeruginosa. Although the two polymyxins have very similar antibacterial activity in vitro, they behave differently in patients, which may affect their clinical benefit against various types of infections.⁵⁰ PMB has advantages in the treatment of severe systemic infections in critically ill patients (administered as the active ingredient), allowing rapid and reliable achievement of effective antibiotic plasma concentrations.⁵¹ The combination use of PMB with other antibiotics is a topic of ongoing debate. Due to the lack of high-quality clinical study, only PK analysis and vitro studies suggests that PMB combination therapy may be advantageous.^8,52–55 It should be noted that effectiveness estimates were affected by multiple confounding factors, including antibiotics combinations, site of infection, bacterial type, MIC differences, diseases severity, and surgical control of infection may significantly clinical effectiveness outcome.⁵⁶ Our study only included one neonate, 7 days old, who had pneumonia caused by Klebsiella pneumoniae. After PMB treatment, this neonate was unfavourable clinical responses and did not achieve microbial clearance, as well as no adverse events. Three studies were conducted to assess the effectiveness of colistin in neonates, and indicated colistin seems to be effective for treating MRGN infections.^57–59 However, no studies have yet established and reported a cohort of neonates using PMB. There were 7 deaths in our study, all of whom had mixed infections involving MRGN, viruses, or fungi. All patients developed sepsis. Additionally, their immune functions were impaired, including hematologic malignancies, severe aplastic anaemia, and genetic disorders. The increased risk of infection was significantly associated with allogeneic transplant and colonization by MRGN. Even worse, immunosuppression might be significantly associated with increased patient mortality.^60–62

PMB has been used to treat infections caused by MRGN, but it fell out of favour due to significant nephrotoxicity and neurotoxicity in paediatric patients. Specifically, there was a paucity of pharmacokinetic and pharmacodynamic information regarding the administration of polymyxins in neonates and paediatric patients, raising safety-related apprehensions.⁶³ AKI was observed in 25.9% (14/54) paediatric patients. However, the incidence of nephrotoxicity reported in the literature varies widely. The highest reached 82.4%, which received therapy with PMB or colistin.⁴¹ Most reported nephrotoxicity events were mild and recovered gradually after drug withdrawal. Measures that may reduce nephrotoxicity include avoiding exceeding recommended doses, performing therapeutic drug monitoring, and avoiding co-administration of other nephrotoxic drugs. Existing studies have reached the same conclusion as ours, that co-administration with potent diuretic (furosemide) may be a contributing factor to nephrotoxicity.⁶⁴ The mechanism by which PMB caused nephrotoxicity was related to the cumulative uptake of PMB mediated by megalin leading to cell apoptosis.⁶⁵ What’s more, furosemide promoted the accumulation of nephrotoxic antibacterial drugs (such as gentamicin) in the tubules and aggravated the damage to renal tubular epithelial cells.⁶⁶ Some studies have found some risk factors for PMB-associated AKI in adult patients, including older age, too high loading and cumulative doses, combined use of nephrotoxic drugs or vasoactive drug, septic shock, and abnormal baseline Cr.^19,67–69 However, due to the limited sample size and heterogeneous patient screening, it was a challenging to achieve consistency in identifying the risk factors associated with AKI related to PMB. Ferroptosis was one of the main mechanisms underlying PMB-induced renal injury, and decreased p53 acetylation level could inhibit ultimately attenuates AKI.⁷⁰ The results of PK model showed that an area under the curve across 24 h at steady state (AUC_ss,24h) of >100 mg·h/L was a good predictor for the probability of nephrotoxicity in adult.⁷¹ But the pharmacokinetics of PMB in paediatric patients were extremely limited. Only one study including 19 children has built a two-compartment model with first-order elimination for PMB in paediatric patients, to evaluate the appropriateness of different PMB dosages, but did not explore independent predictors of efficacy and renal injury.³⁹ Our study found that loading dose did not increase clinical effectiveness and risk of AKI, which was also consistent with Jia’s study.³⁰ Neurotoxicity was another important side-effect associated with PMB, usually manifesting as paraesthesia, dizziness, ataxia, respiratory depression, and meningeal irritation (intrathecal administration).⁷² No relevant adverse event was collected in our study, possibly due to the inability to accurately describe such symptoms in paediatric patients. The intrathecal delivery of PMB bypassed the blood-brain barrier and were sometimes used for MRGN infections in central nervous system. Notwithstanding, polymyxins could induce mitochondrial dysfunction and activation of pro-inflammatory mediators in central and peripheral nervous tissue, thereby inducing neurotoxicity.⁷³

Our study inevitably had some limitations. Firstly, due to the difficulty in collecting data on children using PMB, especially in neonate, our study was limited to a relatively small sample size of 54 patients, which made it less robust to conduct multivariate confirmatory logistic regression analysis to speculate on possible independent predictors of effectiveness and nephrotoxicity. We found that some influencing factors were significant in univariate analysis but not significant in multivariate analysis. Limited by the sample size, it was difficult to analyse the possible reasons (spurious correlation, indirect correlation or confounding) in detail. We elaborated in the discussion based on other published literature, guidelines, and clinical experience, and drew cautious conclusions. Secondly, the pharmacokinetics and pharmacodynamics of PMB in children are lacking in our study because our institution has not yet developed therapeutic drug monitoring technology in PMB. Thirdly, our study was a retrospective study. Due to the lack of a control group, the results might be affected by multiple confounding factors, and the causal relationship between multiple factors and clinical effectiveness and adverse events cannot be directly proved. Finally, the literature review only conducted a descriptive analysis of the included studies. Meta-analysis was not performed due to the heterogeneity in criteria for assessing clinical effectiveness and nephrotoxicity, as well as limited reported data in the literature.

Conclusion

Our study highlighted the importance of PMB in managing MRGN infections in paediatric patients by demonstrating favourable clinical responses and pathogen clearance rates. AKI was also a safety issue that cannot be ignored. The limited sample size might lead to the low confidence in determining factors that affect clinical effectiveness and adverse effects. More large sample studies are needed in the future to verify these conclusions.