Identification of Drugs in Parenteral Pharmaceutical Preparations from a Quality Assurance and a Diversion Program by Direct Analysis in Real-Time AccuTOF^TM-Mass Spectrometry (DART-MS)

Abstract

In healthcare settings drug diversion and impairment of physicians are major concerns requiring a rapid and efficient method for surveillance and detection. A Direct Analysis in Real Time ion source coupled to a JEOL AccuTOF^TM time-of-flight mass spectrometer (DART-MS) method was developed to screen parenteral pharmaceutical formulations for potential drug diversion. Parenteral pharmaceutical formulations are also known as injectable formulations and are used with intravenous, subcutaneous, intramuscular and intra-articular administration. A library was created using the mass spectra data collected by a DART-MS operated in switching mode at 20, 60 and 90 V settings. This library contained 17 commonly encountered drugs in parenteral pharmaceutical formulations that included the surgical analgesic: fentanyl, hydromorphone and morphine; anesthetic: baclofen, bupivacaine, ketamine, midazolam, ropivacaine and succinylcholine; and a mixture of other drug classes: caffeine, clonidine, dexamethasone, ephedrine, heparin, methadone, oxytocin and phenylephrine. Randomly selected 200 de-identified parenteral pharmaceutical formulations containing one or more drugs were submitted for analysis to the FIRM Toxicology Laboratory at Virginia Commonwealth University Health and were screened using the DART-MS. The drug contents of the de-identified formulations were previously confirmed by a published high performance liquid chromatography (HPLC) method. The drugs in the formulations were rapidly and successfully identified using the generated library. The DART-MS and HPLC results were in complete agreement for all 200 parenteral pharmaceutical formulations.

Introduction

In healthcare settings drug diversion and impairment of physicians are major concerns requiring a rapid and efficient method for surveillance and detection. Harm can come to patients, the healthcare worker, fellow employees and employers if not detected and corrected. Patients may receive substandard care from their healthcare provider. The absence or dilution of an intended drug could lead to undue pain or anxiety for the patient. Adulterated or contaminated drugs may result in an allergic reaction, potential infections or death. The legal ramifications of drug diversion can include felony criminal prosecution, civil malpractice actions and actions against professional licensure. An employer could lose revenue due to diverted drugs, poor work quality or absenteeism from the healthcare provider, and be at risk for civil liabilities (1).

Research has shown that physicians, nurses, and other healthcare professionals have a higher risk of misusing drugs than other population groups, with the major factors for misuse being access and availability of the drug (2). McAuliffe et al. (3) found 19% of pharmacists reported misusing a controlled substance without a prescription. Similarly, Hughes et al. (4) reported that 17.6% of surveyed physicians reported the use of a minor opiate without a prescription in the past year (5). Anesthesiologists are up to four times more likely to be treated for drug addiction than physicians in other specialties (6). Between 1991 and 2001, 80% of United States anesthesiology residency programs reported that they had experience dealing with impaired residents (7). Opioids are the most commonly abused, with fentanyl and sufentanil being the drugs of choice for many anesthesiologists (6–8). Other commonly abused drugs were propofol, ketamine, sodium thiopental, lidocaine, nitrous oxide and potent volatile anesthetics. It is suspected that the factors relating to the high occurrence of diversion include proximity to large quantities of addictive drugs, relative ease in diverting small amounts for personal use, high-stress environment and exposure in the workplace which sensitizes the reward pathways in the brain, in turn promoting substance abuse (7, 9).

In order to ensure that healthcare workers are not diverting drugs intended for patient care, medication access points can be established. These can include auditing processes, implementation of automated dispensing hardware and software, physical controls and operational policies (2). Other methods for detection of diversion include workplace urine drug-testing programs or for-cause drug testing after an incident involving patient care.

Presented is a Direct Analysis in Real Time AccuTOF™ mass spectrometry (DART-MS) method and generated mass spectra library that allows for the rapid qualitative analysis of multiple parenteral pharmaceutical formulations as opposed to traditional chromatographic methods that could take as long to analyze a single sample or control. DART-MS is an ionization technique developed in 2005. It uses atmospheric pressure interactions of long-lived electronic excited-state atoms or vibronic excited-state molecules within the specimen being analyzed and the gases in the atmosphere (10). The ionization mechanisms depend on the polarity and reaction of the gas, the proton affinity and ionization potential of the analyte, and the presence of additives or dopants (10). It is used to analyze a wide variety of substances “including chemical warfare agents and their signatures, pharmaceutics, metabolites, peptides and oligosaccharides, synthetic organics and organometallics, drugs of abuse, explosives and toxic industrial chemicals” (11). DART-MS offers advantages over traditional mass spectrometry techniques including direct sample examination, minimal or no sample preparation and high sample throughput (12). By coupling a DART ion source to a time-of-flight mass analyzer accurate mass can be measured for compounds present on the substrate.

The generated mass spectra library contains 17 commonly encountered drugs in parenteral pharmaceutical formulations as determined by the Forensic Industrial Environmental Research and Metabolism (FIRM) Laboratory as part of Virginia Commonwealth University (VCU) Health quality assurance and diversion program. The drugs included surgical analgesic: fentanyl, hydromorphone and morphine; anesthetic: baclofen, bupivacaine, ketamine, midazolam, ropivacaine and succinylcholine; and a mixture of other drug: caffeine, clonidine, dexamethasone, ephedrine, heparin, methadone, oxytocin and phenylephrine.

Experimental

Reagents

The primary reference materials for baclofen, bupivacaine, caffeine, clonidine, dexamethasone, ephedrine, fentanyl, heparin, hydromorphone, ketamine, methadone, midazolam, morphine, oxytocin, phenylephrine, ropivacaine and succinylcholine obtained from the Department of Pharmacy at VCU Health. Polyethylene glycol (PEG 600) was purchased from ULTRA Inc. (North Kingstown, Rhode Island). Melting point tubes were obtained from Corning Incorporated (Corning, New York). High performance liquid chromatography (HPLC) grade methanol was purchased from Fisher Scientific (Hanover Park, Illinois). Medical grade nitrogen and helium were purchased from National Welders Supply Company (Richmond, Virginia).

Equipment

A DART ion source was coupled to JEOL AccuTOF^TM mass spectrometer (JEOL Inc Tokyo, Japan) and was operated in positive-ion mode. The system was controlled by Mass Center software (version 1.3.0.1000; JEOL, Peabody, MA, USA). The helium gas flow rate was 2.0 L/min and the gas temperature was 350°C. The resolution of the DART was 6000 FWHM. The instrument was operated using the following settings: ion guide peak voltage 400 V, reflectron voltage 900 V, orifice 1 was operated in function switching mode having voltages of 20, 60 and 90 V. The voltages were cycled, with each voltage being applied for 0.25 s. Orifice 2 voltage was 5 V, and the ring lens was 8 V. The mass range was 100−1,000 Da. The individual total ion chromatograms were produced as a function of response vs. time and were analyzed for each of the orifice 1 voltages. An internal mass calibration was completed by analyzing polyethylene glycol (PEG 600) with an average molecular weight of 600 Da diluted in methanol (12). The mass calibration was considered acceptable if the curve fit was less than or equal to 1E-1 over the entire mass range. A mass calibration reference spectrum was acquired with each sample and was applied to each of the three orifice 1 voltages of 20, 60 and 90 V for each data set.

Analysis

Sampling was achieved by dipping the closed end of a cleaned glass melting point tube into the sample aliquot of the primary reference or parenteral pharmaceutical formulations and “wanding” the melting point tube into the DART-MS gas stream. The primary reference materials were diluted with methanol to one fourth their lowest parenteral pharmaceutical formulations prior to DART-MS analysis. Each sample was analyzed using three aliquots, with two samplings per aliquot. The more intense signal was used for data analysis. If the sample overloaded the detector due to concentration, the sample was diluted with methanol and reanalyzed. The data was reduced by Data Reduction in TSSPro3.0 software. The spectra were produced by averaging a portion of the background and then reducing the background, calibrating, and then averaging the peaks of interest in the total ion chromatogram. The reduced spectra files were then analyzed using Mass Mountaineer and library matching. Criteria for an acceptable match were established and include the requirement that the measured mass of the [M + H]⁺ of the compound fell within ±5 mmu or mDa of the theoretical calculated [M + H]⁺ for that compound. Using typical mass spectra generated from the analysis of the primary reverence materials (Figures 1–5) and the theoretical calculated [M + H]⁺, a mass spectra library was generated for parenteral pharmaceutical formulations.

Figure 2.

DART-MS generated library spectra at 20, 60 and 90 V settings of bupivacaine, caffeine, clonidine and dexamethasone.

Figure 3.

DART-MS generated library spectra at 20, 60 and 90 V settings of ephedrine, fentanyl, heparin and hydromorphone.

Figure 4.

DART-MS generated library spectra at 20, 60 and 90 V settings of, and ketamine, methadone, midazolam and morphine.

Figure 1.

DART-MS spectra at 20, 60 and 90 V settings of baclofen identify the changes generated ion fragments.

Figure 5.

DART-MS generated library spectra at 20, 60 and 90 V settings of oxytocin, phenylephrine, ropivacaine and succinylcholine.

The 200 samples analyzed were randomly selected de-identified parenteral pharmaceutical formulations that had been collected as part of routine testing by the FIRM Toxicology Laboratory at VCU. The samples were a mixture of commercial parenteral pharmaceutical formulations and compounded parenteral pharmaceutical formulations. All of the samples had been previously analyzed and confirmed via HPLC using a previously published method (13). These samples were analyzed by the DART-MS method described herein. The drugs in these parenteral pharmaceutical formulations were identified using the generated library.

Results

The 200 samples of parenteral pharmaceutical formulations analyzed by DART-MS using the generated library resulted in the successful identification all seventeen expected drugs. A total of 171 of the formulations were determined to contain only one drug. The remaining 29 formulations contained between two and four drugs. Overall, 16 of the 17 drugs were detected as the single active ingredient in at least one formulation. Only baclofen was detected solely in combination with other drugs, Table I. The DART-MS and previously determined HPLC results were in complete agreement for all 200 parenteral pharmaceutical formulations.

Table I.

Identification of drugs in 200 parenteral pharmaceutical formulations using the DART-MS and there prepared concentration ranges

Drug	Number of formulations	Concentrations
Bupivacaine	7	0.1−0.25%
Caffeine	13	125 mg/mL
Clonidine	8	0.1−2 mg/mL
Dexamethasone	1	24 mg/mL
Ephedrine	8	0.02−0.1 mg/mL
Fentanyl	21	0.01−50 ug/mL
Heparin	2	20−100 units/mL
Hydromorphone	20	0.01−50 mg/mL
Ketamine	3	20−100 mg/mL
Methadone	15	1−10 mg/mL
Midazolam	20	1 mg/mL
Morphine	26	0.1−50 mg/mL
Oxytocin	3	20−100 units/mL
Phenylephrine	3	0.02−0.1 mg/mL
Ropivacaine	19	0.1−0.25%
Succinylcholine	2	20 mg/mL
Baclofen, bupivacaine	1	20 ug/mL, 13 mg/mL
Fentanyl, bupivacaine	3	2 or 10 ug/mL, 0.1%
Hydromorphone, bupivacaine	4	6 mg/mL, 25 mg/mL
Morphine, bupivacaine	5	10 mg/mL, 25 mg/mL
Hydromorphone, bupivacaine, clonidine	6	0.01−0.10 mg/mL, 13−28 ug/mL, 0.1−0.6 ug/mL
Morphine, bupivacaine, clonidine	8	0.01−0.10 mg/mL, 13−28 ug/mL, 0.1−0.6 ug/mL
Morphine, bupivacaine, clonidine, baclofen	2	7 ug/mL, 13 ug/mL, 0.5 ug/mL, 0.1 ug/mL

Discussion

Over the last 17 years the FIRM Laboratory has routinely analyzed approximately 100 parenteral pharmaceutical formulations/month submitted from various dispensing pharmacies throughout our own medical center as part of VCU Health quality assurance and diversion program. Cases where drug diversion has been detected have been rare as this program is widely known throughout the medical center. The majority of cases were drug diversion was detected have been “for cause” testing that originated from entities outside of our medical center. The ability to use the presented DART-MS method and generated library for the rapid screening of parenteral pharmaceutical formulations can result in quick action by hospital pharmacies and/or risk management departments in cases where drug diversion and impairment of healthcare workers is suspected.

DART-MS utilized both a soft ionization often resulting in the [M + H]⁺ ion and has the ability to fragment the ions within a sample. It has been established that the higher the orifice 1 voltage, the more fragmentation may occur (14). This fragmentation is useful in the identification of drugs. Baclofen (Figure 1) for example at an orifice 1 setting of 20 V produced a mass spectrum with the [M + H]⁺ ion at m/z 214.0629 and the dimer [M² + H]⁺ at m/z 427.1262, and only one fragment ion at m/z 135.1043 (C₉H₁₃N^•+). At 60 and 90 V settings, the m/z 135.1043 peak becomes part of the background. New ions at m/z 197.0602 (C₁₀H₁₂ClNO^•+), m/z 179.0264 (C₁₀H₈ClO^•), and m/z 151.0315 (C₉H₈Cl^•) are detected. Further, hydromorphone and morphine have the same chemical formula, and therefore the same exacted theoretical mass. For morphine at the 20 V setting the [M + H]⁺ ion at m/z 286.1438 (C₁₇H₂₀NO₃ ⁺) and the fragment that represents the loss of the alcohol group at m/z 268.1338 (C₁₇H₁₈NO^•) can be detected. At 60 and 90 V, the spectrum contains an ion at m/z 211.0759 (C₁₄H₁₁O₂ ^•). Hydromorphone at the 20 V setting contains the [M + H]⁺ ion at 286.1438 (C₁₇H₂₀NO₃ ⁺) and the dimer [M² + H]⁺. The dimer at m/z 571.2808 was not detected at 60 and 90 V settings. Instead of the m/z 268.1338 or the m/z 211.0759 peaks seen in the morphine spectra, the hydromorphone spectrum contains an ion at m/z 243.1021 (C₁₅H₁₅O₃ ^•) allowing for the ability to distinguish these isomers.

Like other mass spectroscopy techniques, DART-MS is useful in determining the presence of halogens in a molecule. The mass spectra generated for clonidine at the 20, 60 and 90 V settings contained the [M+H]⁺ ion at m/z 230.0246 (C₉H₁₀Cl₂N₃ ⁺), as well as a [M + H] due to ³⁷Cl at m/z 232.0217 at approximately ⅔ the height. The dimer [M² + H]⁺ dimer at m/z 459.0448 was detected only at the 20 V setting. Clonidine did not appear to fragment at the 60 or 90 V setting.

Oxytocin has a theoretical calculated [M + H]⁺ ions of m/z 1007.4437. This mass is outside of the detectable range of the DART-MS and therefore cannot be detected. Heparin and succinylcholine do not produce their [M + H]⁺ ions on the DART-MS even at the 20 V setting. These three drugs were identified based on their characteristic fragmentation patterns, even without detectable [M + H]⁺ ions.

Despite the lack of chromatographic separation in DART-MS, individual drugs found in a mixture were identified. A mixture of morphine, bupivacaine and clondine contained the [M + H]⁺ ions of each drug detected, as well as the most abundant fragments (Figure 6). In a mixture of bupivacaine and fentanyl (0.8 and 0.01 mg/mL, respectively) where the bupivacaine is a greater concentration, both drugs were identified using the generated DART-MS library.

Figure 6.

DART-MS spectra at 20, 60 and 90 V settings of a mix drug parenteral pharmaceutical formulations.

Conclusion

Utilizing a DART-MS in combination with the generated library, 17 different drugs commonly found in parenteral pharmaceutical formulations were successfully identified either alone or in combination. A total of 200 de-identified formulations that had been previously analyzed by HPLC were screened by DART-MS and the drugs were identified by the generated library. The drugs contained in these formulations, were identified by the protonated molecular species [M + H]⁺ and/or unique fragmentation species. All of the drugs contained in these formulations, whether as single drug or as mixed drug parenteral pharmaceutical formulations were identified, without chromatographic separation. The DART-MS results matched the previously determined HPLC results for all 200 formulations.

DART-MS could be a useful screening technique for diversion testing in a hospital setting. The sampling and data analysis on the DART-MS is rapid and accurate, with minimal or no sample preparation needed.

Funding

This project was supported in part by the National Institute on Health (NIH) Center for Drug Abuse grant P30DA033934.