Quantification of total and unbound cefuroxime in plasma by ultra‐performance liquid chromatography tandem mass spectrometry in a cohort of critically ill patients with hypoalbuminemia and renal failure

Abstract Background Pharmacokinetic studies of cefuroxime by ultra‐performance liquid chromatography tandem mass spectrometry (UPLC‐MS/MS) have been limited to measurements of total concentrations. Here, we developed a robust method for quantifying total and unbound cefuroxime concentrations using UPLC‐MS/MS and ultrafiltration in critically ill patients with hypoalbuminemia and renal failure. Methods Method validation included accuracy, linearity, precision, repeatability, recovery, and limit of quantification (LOQ). Feasibility of the method was performed on samples obtained from randomly selected intensive care unit (ICU) patients. Total and unbound cefuroxime concentrations were quantified using UPLC‐MS/MS. Sampling times were categorized as trough (180‐1 min prior to administration), peak (10‐30 min after administration), mid (30‐360 min after administration), and continuous (sampling during administration). Pharmacokinetic/pharmacodynamic (PK/PD) targets were unbound cefuroxime concentrations above 4 times the minimum inhibitory concentration (32 mg/L). Results Intra‐assay and inter‐assay precision was <3%. Recovery was 99.7%‐100.3%, and LOQ was 0.1 mg/L. We included 11 patients (median age 72 years (range 54‐77). Median albumin serum concentrations and eGFR were 19 g/L (range 11‐40 g/L) and 48 mL/min/1.73 m2 (range 7‐115 mL/min/1.73 m2), respectively. Median trough and mid concentrations of total cefuroxime were 22.27 mg/L (range 5.42‐54.03 mg/L) and 71.49 mg/L (range 53.87‐73.86 mg/L), and median unbound fraction was 75.42% (range 27.36%‐99.75%). Median unbound cefuroxime concentrations were 11.94 mg/L (range 3.85‐32.39 mg/L) (trough) and 55.62 mg/L (range 10.03‐62.62 mg/L) (mid). Conclusion The method is precise and accurate according to ISO 15189 and within the clinical range of cefuroxime (0.5‐100 mg/L). The method was applied in ICU patients and is suitable for TDM on unbound cefuroxime concentrations.


| INTRODUC TI ON
Dutch guidelines on the management of sepsis in the intensive care unit (ICU) recommend cefuroxime for empiric therapy in patients with community-or nosocomial-acquired sepsis of unknown origin. 1 Cefuroxime is a second-generation cephalosporin antimicrobial drug with time-dependent killing against gram-negative and, to a lesser extent, gram-positive bacteria. The effect of hypoalbuminemia for cefuroxime dosing in critically ill patients with low levels of albumin or renal failure is likely to have significant consequences on the drug's pharmacodynamics (PD) and pharmacokinetics (PK).
Therapeutic drug monitoring (TDM)-based dose optimization of cefuroxime could overcome the drug's pharmacokinetic variability, increase its target attainment, and prevent toxicity by overdosing. 2,3 Pharmacokinetic studies of cefuroxime have been limited to measurements of total concentrations. [4][5][6] To our knowledge, only one study is published on the analysis of unbound concentrations of cefuroxime. However, analysis was performed by high-performance liquid chromatography (HPLC) ultraviolet detection. 7 In contrast, direct measurement of cefuroxime levels and unbound fractions using ultra-performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS) has not been published before. 8 Given the variability of drug protein binding in ICU patients with low albumin levels and renal failure, the main purpose of this study was to develop a reliable and sensitive method based on UPLC coupled with quadrupole-linear ion trap MS/MS. Secondly, we applied the optimized method to plasma samples of ICU patients and assessed the extent of cefuroxime plasma protein binding for TDMbased dose optimization.

| Chemicals and standards
Ammonium acetate of LC/MS quality was obtained from Sigma-Aldrich. LC/MS-grade methanol and 98%-100% formic acid were purchased from VWR International. MilliQ water was produced in our hospital. Pasteurized plasma protein solution (GPO) plasma and fresh frozen plasma (FFP) were from Sanquin. Cefuroxime was purchased from Fresenius Kabi and cefazolin as internal standard was purchased from Eurocept.

| UPLC-MS/MS conditions and procedure
An Acquity H-class UPLC system equipped with a BEH C18

| Sample preparation and processing
Before injection into the UPLC system, all samples were processed as follows: 0.1 mL of the solution to be analyzed was taken and spiked with 30 µL cefazolin 0.05 mg/mL (as internal standard) and 500 µL methanol:acetonitrile 90%:10% (v/v). Patient samples were thawed and vortexed shortly before analysis and processed in the same manner. This mixture was vortexed for 1 min and ultracentrifuged at 30 000 g for 10 min at 25°C. Then, 2 µL of this sample was injected and quantified as described in Section 2.2. To determine the analysis' specificity, a blank sample in GPO plasma was processed 10 times. Multiple reaction monitoring (MRM) transitions of the sample were compared to a standard containing 0.5 mg/L cefuroxime. To assess linearity, a calibration line was calculated using cefuroxime serial dilutions of 0.5, 5.0, 10, 25, 50, 75, and 100 mg/L in GPO plasma, and the correlation coefficient (r) was determined. Each standard was processed in three replicates. Reproducibility was tested by analyzing two control samples of the standard solution at 0.5 and 25 mg/L 10 times and further confirmed by testing the control samples ten times by two different analysts on two separate days. Recovery K E Y W O R D S cefuroxime, hypoalbuminemia, intensive care unit, therapeutic drug monitoring, unbound concentration was determined by analyzing two standards of cefuroxime (0.5 and 25 mg/L) ten times using the method described in Section 2.2. Finally, the limit of quantification (LOQ) was estimated by analyzing the lowest standard (0.5 mg/L) at a twofold and fivefold dilution of 0.25 and 0.1 mg/L, respectively. Recovery was measured by UPLC-MS/MS with ten injections for each diluted standard.

| Plasma protein binding
A method to quantify cefuroxime fractions bound and unbound to plasma proteins was set up using three in vitro cefuroxime stock solutions at 2, 40, and 80 mg/L prepared in six replicates and diluted in FFP. These solutions were quantified as described in Section 2.2. Subsequently, unbound cefuroxime was quantified by pipetting 0.5 mL of the solution to be analyzed into a Centrifree ® Ultrafiltration Device (Merck Millipore). After centrifugation at 1500 g for 25 min at 25°C, 0.1 mL of the filtrate was processed and unbound cefuroxime was quantified as described in Section 2.2.
Stability data (25°C for 25 minutes) were adopted from Hu and colleagues. 10 The unbound fraction concentration was expressed as (total measured concentration -protein-bound concentration)/ total measured concentration.

| Study design and patients
This prospective, noninterventional feasibility study was conducted as a pilot study at VieCuri Medical Center, an in-patient university-associated teaching hospital in the province of Limburg, the Netherlands. Dosages were 1500 mg TID for patients with a glomerular filtration rate (eGFR) >30 mL/min/1.73 m 2 , 1500 mg BID for patients with eGFR of 10-30 mL/min/1.73 m 2 , and 750 mg QD for patients with eGFR <10 mL/min/1.73 m 2 . Dialysis patients with intermittent hemodialysis (IHD) were treated with 750 mg BID, with the second administration following immediately after dialysis. Patients receiving continuous venovenous hemofiltration (CVVH) were given 750-1500 mg BID. 12 Leftover plasma samples were collected at room temperature (15-25°C) in serum tubes as part of routine patient care. Samples were centrifuged after collection and frozen at −20°C for a maximum of six months to analysis. 10 Samples were analyzed batchwise. Stability of plasma samples was not tested. Stability data of cefuroxime plasma samples were adopted from Hu and colleagues where corresponding storage times and temperatures (6 months at −20°C) were used. 10 Time between sampling and exact drug administration times were calculated using our electronic administration registration.

| Statistical analysis
This study focused on the feasibility of the proposed method for  Repeatability expressed as intra-assay coefficients of variation (CVs) was 6.18%-2.59%. Intermediate precision expressed as inter-assay CVs was 1.61%-3.77%. Mean recovery was 99.7%-100.3%.

| Clinical application of the UPLC-MS/ MS method
Patient characteristics (N = 11) are shown in Table 1.
We collected 18 usable leftover samples from 11 patients. Trough samples were collected from 9 patients. Three mid samples were collected from 3 patients. Two peak samples were collected from renal clearance (eGFR) was 48.5 (7-115) mL/min/1.73 m 2 . Figure 4 shows the unbound fraction of cefuroxime in our study population; the median was 75.42% (range 27.36%-99.75%). with endogenous substrates, such as urea and bilirubin, that accumulate due to reduced renal clearance. [17][18][19] An increase in volume of distribution by fluid resuscitation in septic patients may also account for higher unbound fractions of beta-lactam antibiotics. 20 A major strength of our study was the inclusion of a heterogeneous group of ICU patients, and blood samples were taken at random times. All ICU patients underwent standard care, making our results generalizable to other compatible ICU cohorts. Our study refines the current use of MS/MS in ICU patients, illustrating its potential to increase routine TDM for cefuroxime. [14][15][16][17] There are also limitations of our study. First, instead of applying a standard research protocol, we used leftover samples. Second, this study was restricted to adult ICU patients recovering from a mixture of medical-surgical procedures, making generalizability of our finding to other cohorts of ICU patients cumbersome. Third, we could only measure total protein-bound quantities but could not differentiate between specific proteins, such as immunoglobulins, proteins originated from total parenteral nutrition, or others. However, this limitation is of lesser clinical relevance, as we were mainly interested in the relative quantity of unbound cefuroxime required to reach target PK/PD levels. Finally, we used cefazolin as an internal standard and that might have prompted interference because cefazolin is widely used as prophylactic treatment for surgery. However, in our hospital, the pharmacists who interpret cefuroxime concentrations are aware of all the medications prescribed and administered to each patient during hospital stay. Medication prescriptions and administrations are digitally recorded in the patient data management system. This is the case for all hospitals in the Netherlands. Secondly, cefuroxime and cefazolin are not used simultaneously as defined in our local antibiotic treatment guideline, also cefazolin is quickly cleared (t½ = 1.5-2 hours). Therefore, using cefazolin as an internal standard will be not a problem for clinical practice.
In summary, we describe a simple and sensitive UPLC-MS/MS method for the quantification of cefuroxime in plasma obtained from ICU patients. Due to its high sensitivity and accuracy, this method allows pharmacokinetic analysis and TDM-based calculation of unbound cefuroxime plasma levels with new dosage regimens, which is especially relevant for ICU patients with hypoalbuminemia and renal failure.
By measuring unbound fractions of cefuroxime our method could improve the treatment of ICU patients, for whom achieving correct and effective treatment as fast as possible is of special importance. 2 The method was implemented in our hospital. UPLC-MS/ MS gives a quick output of results, so dosing strategies can be ad-