Colistin Pharmacokinetics in Pediatric Cancer Patients in Egypt

Colistin has been reintroduced to clinical practice after the emergence of multidrug-resistant gram-negative (MDR-GN) and the failure of other antibiotics. Pharmacokinetics and pharmacodynamic data in the pediatric population are scarce. This study aimed to highlight the pharmacokinetics of 2 colistin doses, 2.5, and 5 mg/kg/day, in febrile neutropenia pediatric cancer patients regarding patient outcomes. In a prospective, comparative study, patients suffering from MDR-GN infection were randomly recruited to receive either 2.5 or 5 mg/kg/day colistin doses. The demographic, microbiological, and treatment outcomes were collected before and after treatment. Colistin levels were determined using HPLC/MS/MS. Peak, trough, area under the concentration-time curve (AUC 24 ), and the ratio of AUC 24 to the minimum inhibitory concentration (AUC 24 /MIC) were assessed. Clinical cure was achieved in 14 (77.8%) cases in the Low-Dose (LD) group vs. 13 (81.3%) in the High-Dose (HD) group. Four (25%) patients vs. 4 (33.3%) in the LD and HD group (P= 0.69) attained an optimal plasma AUC 24 /MIC, respectively, while the therapeutic level of colistin was reached in all patients in the LD group compared to 14/16 (87.5%) in the HD group. Microbiological eradication was achieved in 93.8% and 91.6% of patients in the LD and HD groups, respectively. However, the median time to clearance was significantly lower in the LD group, 4 days vs. 7 days in the HD group (P= 0.022). In conclusion, the current study suggests that LD may be as efficacious and safe as HD in treating MDR-GN infection. However, LD colistin was associated with a shorter clearance time than HD colistin.


Introduction
Febrile neutropenia (FN) is common in hematological and solid malignancies as a consequence of cytotoxic chemotherapy and is a contributor to death in children. About one-third of children treated with these cytotoxic drugs experienced FN during the neutropenic period.
During FN, there are bloodstream infections with MDR-GNB, and the defective inflammatory and immunologic responses lead to sepsis and death [1]. Consequently, this medical emergency mandates prompt administration of antibiotic therapy to treat these patients [2].
Colistin was initially discovered in the 1940s and was used in several countries. However, its association with nephrotoxicity and neurotoxicity limited its use. Colistin, a member of the polymyxin family of antibiotics, has been reintroduced to clinical practice after the increased incidence of infections caused by MDR-GNB [3]. Colistin methanesulphonate (CMS) is a prodrug of colistin that hydrolyses in vivo, yielding the active colistin base [4]. Colistin is now used as 2 nd line agent in the treatment of carbapenem-resistant Enterobacteriaceae (CRE) and MDR-GN infections [1]. Its effectiveness against most gram-negative bacteria, including Pseudomonas aeruginosa and Acinetobacter baumannii, is well documented [2]. An excellent index to measure the efficacy of colistin is the pharmacokinetic/pharmacodynamic index (PK/PD), defined as the ratio of the area under the concentration-time curve from 0 to 24 h (AUC 24 ) to the minimum inhibitory concentration (MIC). In 2019 International Consensus Guidelines for the Optimal Use of Polymyxins recommended a target AUC 24 ≥ 50 mg.h/L to equate to a target steady-state concentration of 2 mg/L. However, lower respiratory tract infections may require a higher target AUC 24 . It is recommended to consider the dose that provides the maximum exposure to colistin with the least side effect [5].
Currently, the FDA-recommended doses for pediatric patients are 2.5-5 mg/kg /day of CBA in 2-4 divided doses [5]. However, pharmacodynamic and pharmacokinetic data in the pediatric population are still lacking.
In this study, we aimed to study the pharmacokinetic/pharmacodynamics parameters of the two currently used doses of colistin to evaluate their effectiveness and treatment outcomes in pediatric oncology patients suffering from febrile neutropenia caused by MDR-GN infection.

Study design and ethical consideration
The study was a prospective, randomized clinical study comparing two different colistin doses in pediatric cancer patients with MDR gram-negative infection. This is a sub-group analysis of a larger cohort of patients focusing only on the pharmacokinetic data of colistin.
Patients were recruited at the Pediatric Oncology Unit and the Pediatric Intermediate Care Unit at the National Cancer Institute, Cairo University. The Research Ethics Committee of the Faculty of Pharmacy, Ain Shams University, approved study protocol number (103). The study was conducted following the declaration of Helsinki and was registered at clinicaltrials.gov (ID number NCT03397914). The patient's caregiver or legal guardian was counseled about the study, and informed consent was obtained before participation.

Patients
Thirty-four pediatric cancer patients diagnosed with MDR-GN organism infection with either active or based on a previous history of GNB received intravenous colistin during hospitalization were enrolled for this study. Patients were excluded if colistin was used for less than 6 doses, i.e., 3 days, serum creatinine was more than 2 mg/dL before treatment with colistin commencement, or the patient was suffering from septic shock. In addition to colistin, all patients received carbapenem, amikacin, or tigecycline as part of the MDR-GNB treatment protocol).

Randomization
Patients were randomized using "simple computer-generated" randomization techniques to receive either a low dose (LD) (Loading dose of 2.5 mg/kg followed by a maintenance dose of 1.25 mg/Kg as a short infusion every 12 h) or a high dose (HD) (Loading dose of 5 mg/kg followed by maintenance dose 2.5 mg/Kg as a short infusion for 30min every 12 h).

Colistin administration
Colistin was supplied under the trade name Colomycin  and manufactured by Forest Laboratories, Pharma BV company, Netherlands. It is obtained as powder vials containing either 1 million (30 mg) or 2 million units (60 mg). The powder is diluted with 2.1 mL or 4.2 mL, respectively, of sterile water for injection, yielding a concentration of 15 mg/mL (0.5 million units/mL). The volume corresponding to the calculated dose is withdrawn and further diluted with a compatible solution [dextrose 5% D5w or normal saline (NS)]. The calculated dose was dissolved in 100 mL of normal saline and was administered as a 30 min intravenous infusion.

Data collection
All patients' data were collected from the patient's files, including demographic and clinical characteristics. Other clinical data included clinical cure (resolution of clinical signs and symptoms of infections), time to defervescence (fever resolution), microbiological clearance (disappearance of MDR-GN bacterial isolates on follow-up cultures), time to microbiological clearance [6], mortality, the development of nephrotoxicity (assessed and graded based on National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTC AE) v5.0) and length of hospital stay and adjusted length of stay (time from the start of colistin till the discharge of the hospital or death).

Microbiological data & antimicrobial susceptibility
Specimens from blood were obtained as clinically indicated. Identification and susceptibility testing of all gram-negative microorganisms grown on blood agar/ MacConkey agar were performed based on routine microbiological methods using Vitek automated system (Gram-negative panel).

Blood samples
Blood samples were withdrawn from the patients after reaching a steady state (after 4 days) of CMS therapy. Two blood samples (3 mL) of venous blood were drawn into a heparinized tube, the trough concentrations (C min ) were collected just before the next dose, and peak concentrations (C max ) were collected 30 min after the end of the CMS infusion. The plasma was separated by centrifugation at 2500g for 10 min within 2 h of collection. The resultant plasma was stored at -70 ºC until assayed.

Methods
One hundred microliters of plasma were mixed thoroughly with 400 L acetonitrile (Alliance Bio, USA), vortexes for 30 sec, and centrifuged at 10,000 g for 10 min at 4 C. Twenty L of the resultant clear supernatant was then injected into LC; Agilent 1260 series chromatograph; Agilent Technologies, coupled with a tandem mass spectrometry (MS/MS; AB SCIEX API 3200 LC-MS/MS system; Applied Biosystems, Q TRAP, Germany). The analytical column used was Waters X Bridge C18-5 µm (2.1x150 mm Column, Germany) at 39 °C. The mobile phase consists of a mixture of 0.1% formic acid/water and 0.1% formic acid/acetonitrile (60/40 v/v) isocratic flow, delivered at a flow rate of 0.2 mL/min. The mass spectrometer was operated in the positive ESI mode with the spray voltage set at 5.5 kV, at a temperature of 600 °C, and a curtain gas flow of 30 L/h [8]. The calculation is done by the Multiquant software program. Serial dilutions of standards were prepared at different concentrations for colistin in drug-free plasma and extracted as mentioned in sample preparation to make a calibration curve. They were detected at a retention time of 2.32 min. Quantification was performed with multiple reactions monitoring (MRM) by using curtain gas collision-induced dissociation and the following ion transitions: m/z 585.5/101.2, for Colistin A & m/z 578.5/101.2, for Colistin B with the declustering potential set at 51 V and the collision energy at 53 eV.

Pharmacokinetic calculation
The pharmacokinetics parameters colistin peak and trough labeled C max and C min , and AUC 24 were calculated using the following equation:

[9]
Where t′ is the time of infusion (h), C max is the peak concentration at the end of infusion, C min is the trough concentration at the end of the dosing interval, and K el is the elimination rate constant.
The elimination rate constant (K el ) was calculated based on the Sawchuk-Zaske method

[9]
Where t is the time difference in time between the 2 concentrations.
The number of patients reaching therapeutic concentration (≥2mg/L ), The number of patients reaching the target AUC 24 ≥ 50 mg.h/L, and the number of patients achieving a target AUC 24 / MIC= 60 mg.h/L [10] were assessed [5].

Statistical analysis
Statistical analysis was done using IBM SPSS ® Statistics version 26 (IBM ® Corp., Armonk, NY, USA). Numerical data were expressed as median and range. Qualitative data were expressed as frequency and percentage. Pearson's Chi-square test or Fisher's exact test examined the relationship between qualitative variables. For quantitative data, two groups were compared using the Mann-Whitney test (nonparametric t-test) for not normally distributed data. All tests were two-tailed. A P<0.05 was considered significant.

Results
Thirty-four pediatric oncology patients were enrolled in this study. Eighteen patients were assigned to the LD group compared to 16 patients in the HD group.
There were no significant differences in terms of age, sex, weight, diagnosis, and disease stages at presentation between groups. Their median age was 4.5 (1.8-13) years in the LD group compared to 7 (2.5-16) years in the HD group. Data are presented in Table 1.

Microbiological data
The most commonly isolated organism was Klebsiella pneumonia [8 (44.4%) vs. 4 (26.7%)], followed by E coli [4 (22.2%) vs. 4 (26.7%)] in the LD and HD groups, respectively. Details of causative organisms are presented in Table 2. Susceptibility of bacterial isolates collected from patients in the study had a median MIC of 0.75 (0.5-2) µg/mL vs. 0.88 (0.75-3) µg/mL in the LD and HD groups, respectively.  Fig. 1. On the other hand, the median levels of the peak concentration of total colistin (C max ) were insignificantly different between the LD and HD groups, [10.6 (2.22-18.52) µg/mL] vs. [7.52 (0.99-18.84) µg/mL] respectively, P= 0.528. The median duration of colistin treatment was 7.5 days in both the LD and HD groups, without significant differences between the two groups, P= 0.825.
The AUC 24 and AUC 24 / MIC were calculated using the equation described above. Nonsignificant differences were observed between the two groups. Data are described in Table 3.  Table 4 and Fig. 2 describe the clinical outcome observed in both groups. The median time to clearance was significantly lower in the LD group, showing a median of 4 days vs. 7 days in the HD group; however, microbiological clearance was attained in a comparable number of patients in both groups. The LD group had clinical cures observed in 14/18 (77.8%) patients compared to 13/16 (81.3%) patients in the HD group, P= 1.

Clinical outcomes
Only grade 1-2 nephrotoxicity was experienced in 3/18 (16.7%) and 3/16 (18.8%) patients in the LD and HD groups, respectively. (P= 1) Both groups showed a similar length of stay, adjusted length of stay, and mortality. Details are described in Table 4.
Non-significant differences existed between the numbers of patients reaching a therapeutic level in both groups. All patients in the LD group and 14/16 (87.5%) in the HD group reached therapeutic plasma concentration (≥2mg/L) [5]. The two current doses achieved a target AUC 24 / MIC>60 mg.h/L in only 4/16 (25%) and 4/12 (33.3%) in the LD and HD groups, respectively.

Discussion
Colistin may be a valuable option in treating severe pediatric nosocomial infections caused by MDR -GN organisms [1]. Yet, pediatric literature addressing colistin dosing, pharmacokinetics, and dynamics is scarce. The study aims to assess colistin pharmacokinetic parameters (C max , C min , AUC 24 , and AUC 24 /MIC) in the pediatric cancer population treated with 2.5 or 5 mg/kg/day, aiming to fill the gap of knowledge in this age group and to better elucidate the optimal dose to fight MDR-GN infections.
The current study showed that median trough levels were 0.54 (0-1.71) μg/mL for the LD group; compared to Sorlí et al. study, the observed trough concentrations (C min ) were 1.14 (0.11-5) μg/mL, for a colistin dose of 2 million units [equivalent to 2.9 mg/kg/day colistin base activity (CBA)] in 3 divided doses. Whereas for the HD group of the current study, the median trough levels were 1. 16  In the HD group, the observed C max was lower than in the LD group. This may be explained by the phenomenon of augmented renal clearance (ARC) in patients suffering from sepsis. A proposed hypothesis is that augmented renal clearance (ARC) is usually observed in patients suffering from systemic inflammatory response syndromes (SIRS), like patients suffering from burns, trauma, or sepsis. The SIRS causes an increased cardiac output resulting in enhanced renal blood flow. Thus, the kidneys clear CMS too rapidly, leading to reduced colistin bioavailability [12, 13].
Adults' pharmacokinetic studies reported that at steady-state, typical C max was estimated to be 2. The studies mentioned above reported a lower steady-state concentration than the concentration observed in the current study in both low-dose [10.6 (2.22-18.52) µg/mL] and high-dose [7.52 (0.99-18.84) µg/mL] groups. These discrepancies may be attributed to the studied population's differences in pharmacokinetics and pharmacodynamics. There is a scarcity of literature addressing colistin pharmacokinetic studies in the younger population.
However, in a study by Mesini et al., addressing the pediatric population, they reported the C max and C min concentration of colistin following the administration of a loading dose of 150 000 IU/kg (equivalent to 5 mg/kg) followed by a maintenance dose of 75000 IU/kg (equivalent to 2.5mg/kg) every 12h during 9 treatment courses to 7 children. They observed a range of C max (4.3-18.9 mg/L) and a C min (0.4-3) mg/L [17]. They calculated the AUC 24 , ranging from (33-92 mg.h/L). These findings are comparable to the results of our study.
According to previous studies, a wide range of serum colistin peak levels was observed in both groups; however, no standard levels were reported in earlier literature. The current study demonstrated that colistin has relatively safe and tolerable toxicities. The observed nephrotoxicity was only graded 1-2 and was reported in 16.7% and 18.8% of the LD and HD groups, respectively. These data were corroborated by Tamma et al., in their study, they administered 2.5 mg/kg BID, and nephrotoxicity was reported in 22% of studied children [22]. However, Bal et al. reported an incidence of 10% following the administration of 5 mg/kg in 3 divided doses, compared to the HD group of our study. The difference may be due to the administration of a loading dose, thus higher colistin exposure, and the dosing intervals (every 12 h) in the current study [23].
The limitations of this study were the small number of patients examined as well as the use of only two blood samples (trough and peak) from each patient; we need more sampling at different time intervals that will give a more accurate estimate of the AUC 24 and the AUC 24 /MIC. In addition, plasma protein binding was not measured for colistin; therefore, this study's AUC/MIC values were for total colistin.
The strength of this work is being a prospective and randomized study. We are also addressing limited data regarding colistin PK and PD studies in the pediatric population, aiming to improve their clinical outcomes.

Conclusion
In conclusion, our study demonstrates that the low dose of colistin may be as efficacious and safe as the high dose in treating MDR-GN infection. However, low-dose colistin was associated with a shorter time to clearance than high-dose colistin. The current study suggests the benefit of monitoring colistin concentrations in plasma.
No significant differences were observed between groups in terms of C max and AUC 24 ; however, the study is limited to the small number in both groups. Future research should target a larger sample of pediatric patients to validate these findings.