ABSTRACT
Limited information is available on the urinary excretion of colistin in infected patients. This study aimed to investigate the pharmacokinetics of colistimethate sodium (CMS) and formed colistin in urine in patients with multidrug-resistant (MDR) Gram-negative bacterial infections. A pharmacokinetic study was conducted on 12 patients diagnosed with an infection caused by an extremely drug-resistant (XDR) P. aeruginosa strain and treated with intravenous CMS. Fresh urine samples were collected at 2-h intervals, and blood samples were collected predose (Cmin ss) and at the end of the CMS infusion (Cmax ss) for measurement of concentrations of CMS and formed colistin using high-performance liquid chromatography (HPLC). CMS urinary recovery was determined as the summed amount of CMS and formed colistin recovered in urine for each 2-h interval divided by the CMS dose. There were 12 enrolled patients, 9 of whom were male (75%). Data [median (range)] were as follows: age, 65.5 (37 to 86) years; colistimethate urinary recovery 0 to 6 h, 42.6% (2.9% to 72.8%); range of concentrations of colistin in urine, <0.1 to 95.4 mg/liter; Cmin ss and Cmax ss of colistin in plasma, 0.9 (<0.2 to 1.4) and 0.9 (<0.2 to 1.4) mg/liter, respectively. In 6/12 (50%) patients, more than 40% of the CMS dose was recovered in the urine within the first 6 h after CMS administration. This study demonstrated rapid urinary excretion of CMS in patients within the first 6 h after intravenous administration. In all but one patient, the concentrations of formed colistin in urine were above the MIC for the most predominant isolate of P. aeruginosa in our hospital. Future studies are warranted for optimizing CMS dosage regimens in urinary tract infection (UTI) patients.
INTRODUCTION
Colistin has emerged as a last-line therapy in the last decade for the treatment of nosocomial infections caused by multidrug-resistant (MDR) Gram-negative bacteria, such as Pseudomonas aeruginosa and Acinetobacter baumannii (1). Despite several recent clinical studies evaluating the pharmacokinetics of colistin and its prodrug colistimethate sodium (CMS) in plasma in different types of patients (2–13), very limited information is available on the recovery of CMS and formed colistin in urine from patients. Only three studies have evaluated urinary excretion and concentrations in urine in healthy volunteers (14–16), two in cystic fibrosis patients (3, 6) and two studies on patients undergoing renal replacement therapies (RRT) (4, 5). A better understanding of the urinary recovery of CMS and formed colistin is important for investigation of colistin-associated nephrotoxicity, the major dose-limiting factor, and optimizing intravenous CMS doses for treatment of urinary tract infections (UTI) caused by MDR Gram-negative pathogens. The present study aimed to examine the pharmacokinetics of CMS and formed colistin in urine in patients with MDR Gram-negative bacterial infections and the relationship between urinary concentrations of CMS and formed colistin and colistin-associated nephrotoxicity.
RESULTS
During the study period, 12 patients were enrolled, including 9 males (75%), with a median (range) age of 65.5 (37 to 86) years. All of them presented with an MDR Gram-negative bacterial infection caused by an MDR P. aeruginosa strain: 5 (41.7%) respiratory tract, 4 (33.3%) urinary tract, 2 (16.7%) intra-abdominal, and 1 (8.3%) skin and soft tissue infection. The median (range) CMS treatment duration and length of hospital stay were 15 (8 to 34) and 26.5 (14 to 102) days, respectively, and none of the patients died.
The characteristics of the patients and concentrations of CMS and formed colistin in plasma and urine are presented in Tables 1 and 2.
Characteristics of the included patients and plasma CMS and formed colistin concentrations
Urine recovery of CMS and formed colistin
Two of the four patients diagnosed with a UTI caused by an MDR P. aeruginosa strain achieved clinical cure (patients 3 and 9), and the other two patients could not be evaluated (one due to the presence of Enterococcus faecalis in the urine [patient 1] and the other with orchyepidemitis due to the same strain of MDR P. aeruginosa [patient 12]). Three of these four patients (75%) achieved microbiological eradication (patients 3, 9, and 12), while for the third (patient 1), no further cultures were performed.
For two of the three patients (3 and 9) with UTI who achieved microbiological cure, the CMS urinary recovery rates were 47.6% and 72.8% (Table 2). For the third (patient 12), the CMS concentrations could not be determined due to technical issues. For the patient with persistence of positive urinary cultures, the CMS urinary recovery was much lower (2.9%) (patient 1) (Table 2). This patient had severe renal impairment at the beginning of the CMS treatment.
In all the patients, the concentration of formed colistin in plasma was below the suggested therapeutic level of 2.5 mg/liter (8). In all but one patient, the concentration of colistin in urine was much higher than those achieved in plasma and also above the MIC (0.5 mg/liter) for the most predominant strain of P. aeruginosa in our hospital.
The maximum concentration of CMS achieved in urine per dose was not correlated with the initial estimated glomerular filtration rate (eGFR) (Spearman rho, 0.436; P = 0.180) or the initial serum creatinine level (Spearman rho, −0.364; P = 0.272). In 6/12 (50%) patients, more than 40% of the CMS dose was recovered in the urine within the first 6 h after CMS administration. The daily CMS dose (in milligrams) showed a significantly positive linear correlation with the maximum concentrations of CMS observed in urine during the first 6 h (rho, 0.837; P = 0.001) and also with the CMS urinary recovery within 6 h (rho, 0.608; P = 0.047).
At the end of the treatment, nephrotoxicity was present in only 3 (25%) of the patients (1 with risk stage[R] and 2 with injury stage [I] according to RIFLE categories), who also received other nephrotoxicity drugs (Table 2). The CMS urinary recovery rates were not different between patients with and without nephrotoxicity at the end of treatment (25.3% [range, 20.9% to 49.0%] and 45.1% [2.9% to 72.8%; P > 0.999], respectively). In addition, the CMS urinary recovery was not correlated with the final serum creatinine level or final eGFR (Spearman rho, −0.373, P = 0.259, and Spearman rho, 0.436, P = 0.180, respectively).
DISCUSSION
Colistin use has reemerged in recent years for the treatment of MDR Gram-negative bacterial infections. However, limited information is available on its pharmacokinetics in urine in infected patients. Our present study is the first pharmacokinetic study of CMS and formed colistin in urine in patients with MDR Gram-negative bacterial infections who are not undergoing any RRT and with urine samples collected at 2-h intervals over 6 h. The 2 more recent studies that examined the urinary excretion of CMS and formed colistin in healthy volunteers showed high urinary excretion of CMS (13, 14). In the most recent study by Couet et al., the urinary excretion of CMS was reported to be 70% after the administration of a single dose of 1 million IU (MIU) of CMS. This fact could be explained by the high estimated typical renal clearance of CMS (103 ml/min), a value close to the glomerular filtration rate. In our present study, the urinary excretion of CMS within the first 6 h after CMS administration was >40% in half of the patients. Although the majority of our patients had poor initial renal function (for 4 of them, the initial eGFR was <60 ml/min), no correlation was observed between the initial renal function (eGFR) and the urinary concentration of CMS.
CMS is mainly eliminated by the kidneys, with extensive secretion after filtration by glomeruli in the kidney, while the positively charged colistin undergoes substantial reabsorption by the renal tubular cells (1). Our results showed high excretion of the CMS dose in the first 6 h after administration (more than 40% of the administered CMS dose in 50% of the 12 patients) (Table 1). Compared to the two studies of healthy volunteers, urine samples in our study were collected fresh, immediately transported on ice, centrifuged in a cold room, and frozen at −80°C; this procedure minimized any potential postexcretion hydrolysis of CMS. Almost all of the 12 patients in this study achieved high urinary concentrations of colistin (0.5 to 95.4 mg/liter), the antibacterial entity formed from CMS, and the concentrations were equal to or above the MIC value (0.5 mg/liter) for the most predominant isolate of P. aeruginosa in our hospital. The exposure of formed colistin in plasma in our patients was much lower than in the urine within 6 h. This was due to the fact that only a very small fraction of the CMS dose that was not excreted in urine was converted to colistin in blood (1). Our study has a small sample size, and large clinical pharmacokinetic/pharmacodynamic studies are warranted to evaluate whether doses of CMS lower than the current recommendation should be administered for the treatment of UTI to reduce the high incidence of colistin-associated nephrotoxicity. A linear correlation was observed between the CMS dose and the CMS urinary concentrations, and it is very likely that lower doses of CMS could lead to lower concentrations of CMS in urine. However, no studies have evaluated the potential impact of the plasma or urine CMS concentration or the colistin levels in urine and the appearance of nephrotoxicity.
In our study, nephrotoxicity at the end of treatment was observed in 3 (25%) patients, with one at stage R (risk) and two at stage I (injury). These three patients presented CMS urinary recovery rates of 25.3% (R; patient 2), 20.9% (I; patient 6), and 49.0% (I; patient 8), and the Cmax ss and Cmin ss of formed colistin in plasma were lower than the cutoff point predictor of nephrotoxicity proposed by our group (2.42 mg/liter) (7). Consequently, this adverse event did not seem to be related to a high exposure to colistin or CMS in plasma.
The urinary formed colistin concentrations were high (4.5 to 15.8 mg/liter in patient 6 and 8.5 to 21.3 mg/liter in patient 8) in two patients but very low in the third (0.1 to 0.6 mg/liter in patient 2). In addition, all 3 patients were receiving at least other nephrotoxic drugs, such as furosemide, amikacin, or vancomycin nonsteroidal anti-inflammatory drugs (NSAIDs). Consequently, the presence of these concomitant nephrotoxic drugs could also have influenced the occurrence of nephrotoxicity. In addition, CMS urinary recovery rates were not different between patients with and without nephrotoxicity at the end of treatment, and they showed a correlation only with the CMS initial dose. For all these reasons, the real influence of CMS and colistin levels in urine on the incidence of nephrotoxicity should be evaluated in a larger clinical study.
In conclusion, our study demonstrated rapid urinary excretion of CMS within the first 6 h in patients after intravenous administration. Larger clinical studies are warranted in future for optimizing CMS dosage regimens in UTI patients in order to optimize its use in this specific patient population while minimizing any potential nephrotoxicity.
MATERIALS AND METHODS
Patients.This was a prospective pharmacokinetic study conducted with 12 patients with MDR Gram-negative bacterial infections treated with intravenous CMS in a university hospital from January to September 2014. All the patients were diagnosed with infection caused by an extremely drug-resistant (XDR) P. aeruginosa strain and treated with intravenous CMS for at least 48 h. Exclusion criteria were an age of <18 years, pregnancy or breastfeeding, and the need for renal replacement therapies. Informed consent was obtained from all participating patients or their legal representatives. The study protocol was approved by the Clinical Research Ethical Committees (CEIC), Parc de Salut Mar.
Data collection.The following clinical data were collected: demographics, type of MDR Gram-negative bacterial infection, CMS daily dose (selected by the responsible clinician), number of days of treatment, renal function at the start and end of CMS therapy (serum creatinine and eGFR by the Modification of Diet in Renal Disease [4-variable version] [MDRD-4] equation), presence of nephrotoxicity at the end of treatment evaluated by the RIFLE criteria (17), other concomitantly administered nephrotoxic drugs (e.g., NSAIDs, furosemide, contrast agent, angiotensin-converting enzyme inhibitors, aminoglycosides, and vancomycin), and crude mortality during hospital admission. CMS for injection (GES Genéricos Españoles) was dosed in MIU (range, 1 to 9 MIU/day). The prescribed dose of CMS was administered intravenously in 100 ml normal saline over 30 min with the commercially available CMS formulation for injection, each vial containing 1 MIU of CMS (equivalent to 30 mg colistin base activity [CBA] or 80 mg CMS) (19).
Pharmacokinetic study.Three-milliliter blood samples were collected immediately predose (Cmin ss) and 30 minutes after the end of infusion (Cmax ss) on day 3 or 4 of treatment (when steady state was achieved).
Fresh directly voided urine samples were collected from 0 to 2 h, 2 to 4 h, and 4 to 6 h after the start of the CMS infusion. The CMS/colistin concentration was calculated at that time and then multiplied by the volume of urine to obtain the total amount of CMS excreted renally over the time frame. Concentrations of CMS and formed colistin in plasma and urine samples were measured by validated high-performance liquid chromatography (HPLC) methods (20, 21), with minor modifications as described previously (7). The urinary recovery of CMS was determined as the summed amount of CMS and formed colistin (after conversion using different molecular weights) recovered in urine during the collection period divided by the dose of CMS (18). Nephrotoxicity during CMS treatment was defined as a 1.5-fold or more increase in serum creatinine and/or a decrease in the eGFR of >25 (17).
Statistical analysis.Quantitative variables were expressed as median and interquartile range, and a Spearman test was employed for bivariate correlations. The Statistical Package for the Social Sciences (SPSS) version 15.0 was employed. For all analyses, a two-sided P value of <0.05 was considered to be statistically significant.
FOOTNOTES
- Received 7 December 2016.
- Returned for modification 27 January 2017.
- Accepted 13 May 2017.
- Accepted manuscript posted online 30 May 2017.
- Copyright © 2017 American Society for Microbiology.