Multidisciplinary Approach to Deciphering Etoposide Infusion Reactions and Potential Role of Polyethersulfone Filter Membranes

Adverse drug reactions (ADRs) are an unintended reaction to a drug given at a normal dose.¹  Cancer chemotherapy is often associated with ADRs due to narrow therapeutic windows.²  Etoposide is a common chemotherapeutic agent used in a wide variety of both adult as well as pediatric cancers such as leukemia, lymphoma, neuroblastoma, and osteosarcomas.³  Common adverse effects of etoposide include hematologic toxicities such as leukopenia and thrombocytopenia, as well as gastrointestinal toxicities such as nausea, vomiting, and diarrhea. While these adverse events are expected and occur in many patients, etoposide is known to cause anaphylactic type reactions during infusion administration and has been reported to occur in approximately 0.7% to 2% of patients according to the package insert.^2–5  Symptoms of these infusion-related reactions can range from mild to life-threating and include dyspnea, flushing, erythema, chest pain, hypotension, chills, fever, and more.^2–5 

There have also been studies questioning whether the use of an in-line filter during etoposide administration may be a possible component behind the increased rates of infusion reactions in the pediatric population.^6–8  In a recent study conducted as a collaboration between Riley Hospital for Children (RH) in Indianapolis, IN, and Children’s Mercy Hospital (CMH) in Kansas City, MO, we identified an increased rate of etoposide infusion reactions in pediatric cancer patients that correlated with a 2017 implementation of in-line filters at CMH.⁹  Filtration of compounded pharmaceutical solutions serves multiple roles including removing unwanted particles, air, and precipitates, as well as sterilization of the final solution. In-line filtration is a common practice in pediatric drug delivery.¹⁰  Interactions have been reported between the pharmaceutical solution being filtered and the type of filter.¹¹  These types of interactions include sorption of the drug to the filter causing a decrease in available drug to the patient, leaching of a chemical substance from the walls of the filter into the IV solution, swelling of the membrane causing increased resistance to flow, and chemical interactions between the solution and components of the filter.¹¹  The aim of this study was to assess for any measurable change in the chemical structure of etoposide and potential degradation products that may be associated with an in-line filter equivalent to those used at CMH from 2017–2020.

To investigate the possible drug-filter interaction observed in pediatric patients,⁹  clinicians aware of the increased infusion related reactions occurring more frequently with in-line filters sought collaboration with the Clinical Pharmacology department at Indiana University School of Medicine to utilize expertise and equipment that could further investigate the interaction in a small analytical study.

Etoposide 100 mg/5 mL (Accord, Durham, NC) was added to a 250-mL bag of 0.9% sodium chloride (Baxter, Deerfield, IL) to achieve a final concentration of 0.4 mg/mL, which is the recommended maximum concentration to minimize risk of precipitation per the package insert.³  The various solvents and additional components contained within the etoposide formulation are listed in the Table, along with their relative concentrations and the maximum recommended FDA doses for each component.^3,12  The infusion was compounded under standard aseptic technique by the Hematology-Oncology Outpatient Infusion Pharmacy at RH. After preparation, the infusion bag was spiked with non-polyvinyl chloride (PVC) tubing (Becton, Dickinson and Company, Franklin Lakes, NJ) and the in-line Supor IV-5 filter (low protein binding polyethersulfone filter) (ICU Medical, San Clemente, CA), was placed. The infusion set, that is, filter and tubing, were then primed with the etoposide solution to fully saturate the in-line filter and expose it to the solution, as well as ensure solution was available at all sampling sites prior to sample collection. No solution was wasted from the infusion set prior to sample collection. Next, 5 mL samples were collected in triplicate from the following three locations: 1) the injection port of the compounded IV bag, 2) a needle-less access port on the tubing located before the in-line filter, and 3) needle-less access port after the in-line filter. Analysis was performed immediately following sample collection.

Etoposide (95.0%–105.0% purity) was purchased from Sigma Aldrich (St. Louis, MO). Methanol, acetonitrile, water, and ammonium acetate were purchased from Fisher Scientific (Fairlawn, NJ). All solvents were LC/MS grade.

A stock solution of etoposide (1 mg/mL) was prepared in methanol. Prior to analysis the etoposide stock solution was diluted to 100 ng/µL in mobile phase A (acetonitrile: 5 mM ammonium acetate, 20:80; v/v). Then a 10-µL aliquot was injected into the HPLC-UV-MS/MS system.

From each sample, 20 µL was transferred to a polypropylene tube and 180 µL of mobile phase was added. The samples were vortex mixed for 30 seconds then transferred to a high-performance liquid chromatography (HPLC) vial with a polypropylene insert. An aliquot (10 µL) of each ­sample was injected to the high-performance liquid chromatography diode array detection mass spectrometry (HPLC-DAD-MS/MS) system as described below. All samples were run in triplicate. The chromatograms of the DAD and mass spectrometry (MS) were analyzed for any possible degradation, metabolite formation, or excipients as described below.

An Accela HPLC pump coupled with a Leap PAL HTC autosampler performed the chromatographic analysis of etoposide using reverse phase chromatography at ambient temperature on an Inertsil ODS-3 150 × 4.6 mm 5-micron column. The mobile phase was delivered via gradient at a constant flow rate of 800 µL/min. Mobile phase A was acetonitrile: 5 mM ammonium acetate in water (20:80; v/v) and mobile phase B was acetonitrile: 5 mM ammonium acetate in water (80:20; v/v). The gradient began after 5 minutes at 100% of mobile phase A and then increased to 100% mobile phase B by 10 minutes in a linear fashion and held at 100% mobile phase B until 27 minutes. At 27.1 minutes the mobile phase was stepped to 100% mobile phase A and held until 30 minutes. A Thermo Dionex Ultimate 3000 PDA detector at wavelengths 254 and 280 nM monitored the column effluent. The effluent from the PDA detector then went directly into a Thermo TSQ Quantum Ultra triple-quadrupole mass spectrometer equipped with an electrospray probe. Analysis was performed using the selected reaction monitoring (SRM) scan mode of the etoposide ammonium adduct m/z, 606.2/228.8. XCalibur (version 4.0) controlled the HPLC-UV-MS/MS system. All analysis was done in triplicate to assure accuracy.

Except etoposide, no additional peaks were detected in the chromatograms of any of the samples collected during the duration of the 30-minute run at SRM 606.2/228.8 or wavelengths 210, 220, 254, or 280 nm compared to etoposide standard (Figure 1). The etoposide standard produced an absorbance peak at a range of 9.99 to 10.03 minutes into the run. A peak of similar intensity can be seen at 9.91 minutes of the IV bag port sample, 9.94 for the tubing pre-filter sample, and 9.93 for the tubing post-filter sample. Similar peaks were seen at all measured wavelengths for each sample. While additional peaks were not observed in the SRM scan of etoposide standard or at various wavelengths; they were consistent across the different sample locations with no samples producing a peak unique to a specific wavelength or time between the different sample sites.

Mass spectrometry data showed an average peak area of 151563474 (SD 25679732; CV 16.9%) for the IV bag sample, 109870416 (SD 18738338; CV 17.1%) for the pre-filter tubing sample, and 154166029 (SD 21523693; CV 14.0%) for the post-filter sample. Full results from HPLC-DAD-MS/MS analysis are shown in Figure 1. To clearly appreciate the similarity of the HPLC-MS/MS Chromatographs, Figure 2 shows HPLC-UV Λ = 254 nm, and Figure 3 shows HPLC-UV Λ = 280 nm for all experimental conditions.

We report the importance of interdisciplinary collaboration to investigate potential causes of ADRs that are observed in a clinical setting. Specifically, our results highlight two key findings: 1) HPLC-DAD-MS/MS did not detect a measurable difference between pre- and post-filter samples of etoposide infused through a hydrophilic polyethersulfone in-line filter; and 2) the absence of a chemical change of etoposide does not rule out interactions between other components of the medication solution and the filter.

To our knowledge, this is the first study to evaluate chemical composition of etoposide pre- and post-filter. There were no chemical changes in etoposide as shown by no significant difference in peak absorbance intensity in any of the three sampled conditions and compared to the etoposide standard. We did not fully rule out all the potential existence of an interaction between etoposide IV solution and the studied polyethersulfone membrane filter, but no chemical change appeared on the chromatograms.

Beyond a chemical change to etoposide mediated by a polyethersulfone membrane, other possibilities exist that could contribute to the observed increases of infusion reactions when filtering etoposide. The IV solution used to prepare etoposide for infusion contains benzyl alcohol which may potentially interact with the polyethersulfone membrane used in the IV-5 filter.^11–14  Another potential cause of infusion reactions is the method utilized for filter sterilization. A possible link has been proposed between the use of ethyl oxide gas in the sterilization of biomedical equipment that cannot tolerate heat sterilization and hypersensitivity reactions in pediatric patients.¹⁵  The Supor IV-5 filter used to conduct this analysis and match those used at the institution in which we identified the possible link between filtration and increased rates of infusion reactions in pediatric patients previously, is sterilized using ethyl oxide gas.^14–16 

Based on these results, we deduce that it is unlikely that a chemical change of etoposide mediated by an in-line filter is the cause of the increased incidence of pediatric infusion reactions observed in recent observational clinical studies.^6,9  While other possible explanations exist such as ethyl oxide gas used for IV filter sterilization, an increased risk of infusion reactions has not yet been widely reported despite the common use of these filters in medical practice.

We believe this study not only contributes to the limited literature discussing the potential for drug-administration interactions, but it also illustrates the importance of a multidisciplinary team needed to investigate a clinical hypothesis. In this case, pharmacists and a physician identified this increase in etoposide infusion reactions that coincided with a protocol change to use inline filters for etoposide infusion.⁹  They hypothesized that the filer could have a potential role in this uptick in reactions but did not have the necessary skill-sets to fully investigate. The clinical team sought out experts in HPLC-DAD-MS/MS, which is a very specialized skill and requires sophisticated equipment to successfully complete the analysis. Pharmacists working in healthcare systems, observe many ADRs, but rarely have the time, skill sets, and resources necessary to investigate the potential etiology or causality of the ADR. Drug-drug interactions or drug-nutrient interactions are often considered by pharmacist when evaluating ADRs, but interactions with tubing, filters, and components of drug sterilization may not be on the radar of pharmacists, therefore we believe it is important to share our experience of multi-disciplinary collaboration to investigate serious ADRs.

ADR

Adverse drug reactions;

CMH

Children’s Mercy Hospital;

DAD

diode array detection;

HPLC

high-performance liquid chromatography;

HPLC-DAD-MS/MS

high-performance liquid chromatography diode array detection mass spectrometry;

MS

mass spectrometry;

PVC

polyvinyl chloride;

RH

Riley Hospital for Children;

SRM

selected reaction monitoring