Over the last few years, new synthetic opioids (NSOs), primarily fentanyl derivatives are increasingly implicated in overdose deaths, and have caused a global health concern. Among the 70,237 drug-related deaths estimated in 2017 in the US, the sharpest increase occurred among those related to fentanyl analogs with more than 28,400 overdose deaths which represents 45.2% increase from 2016 to 2017 [1
]. In Europe, their use appears to be geographically localized, but the number of synthetic opioids has grown rapidly since the first substance was reported in 2009. In fact, around 50 NSOs have been reported to the EU Early Warning System with 11 being reported for the first time in 2018 [3
]. From 2015 to early 2018, the European Monitoring Center for Drugs and Drug Addiction (EMCDDA) has conducted eight joint investigations with Europol on fentanyls that have caused serious concern at European level. The two agencies investigated acetylfentanyl, acryloylfentanyl and furanylfentanyl, 4-fluoroisobutyrylfentanyl (4F-iBF), tetrahydrofuranylfentanyl (THF-F), carfentanil, methoxyacetylfentanyl and cyclopropylfentanyl. These substances have been involved in more than 250 deaths, many of which were attributed directly to these substances [4
]. Reported overdose cases include people who unknowingly buy adulterated heroin, other illicit drugs or pain relievers [3
The risk of life-threatening poisoning or deaths by NSOs is being enhanced by the following reasons: most NSOs are particularly potent on µ-opioid receptors (10 up to 10,000 more potent than morphine) and can be toxic at very low dose due to severe respiratory depression [6
]; high bioavailability using routes of administration that allow substances to rapidly reach the central nervous system such as injecting, snorting, and inhalation; ease of synthesis, purchase at low cost, and their adulteration especially with heroin facilitate the poisonings; they, in many cases, are undetectable by the standard screening methods [4
]. A recent EMCDDA study has shown that not all laboratories have the capacity to detect the more uncommon substances [3
]. Thus, the prevalence of these substances in opioid-related poisonings and deaths is most likely underreported.
In the actual context of opioids crisis, more reliable and sensitive methods dedicated to the identification and quantification of NSOs in biological and non-biological matrices are needed. Recently, these analytical methods often based on chromatography (liquid or gas) coupled with mass spectrometry (MS) detection were extensively reviewed [7
]. Even if targeted screening is highly sensitive and specific, they remain limited by the unavailability of reference standards [7
]. A new approach based on untargeted screening by liquid chromatography–high-resolution mass spectrometry (LC–HRMS) using a shared MS spectral database and without a need for reference standards has emerged in recent years [13
To study the prevalence of fentanyl derivatives among other drugs of abuse, the EMCDDA has proposed five strategies as targeted data sources: wastewater or syringe residue analysis, emergency department or drug checking service data, and web surveys [3
]. Self-reported use of drugs often includes reporting biases related to users’ ignorance about the products for their consumption which often are different from the ones that they bought, especially in case of NSOs due to adulteration. Wastewater and syringe residue analysis strategies provide more informative data regarding drug use, but remain indirectly linked to the population of the consumers. Even if these approaches are not representative of the general population and present some limitations, they provide useful, timely and complementary data that offer valuable insights into drug use in Europe [5
On the other hand, the analysis of biological specimens of drug users could be the most accurate and directly linked to their drug consumption. While blood and urine testing can provide useful information of recent exposure to drugs (few hours up to days after consumption), hair testing allows an extensive exposure pattern of past use, as it offers a wide detection window, up to months prior to sampling and is easy to collect. Given that hair is also a rot-proof material that does not decompose, it can be kept for a long period with no special storage conditions (ambient temperature) and retrospectively analyzed, as most substances are stable in the hair matrix. Lastly, unlike conventional matrices like urine, hair sample cannot be adulterated by dilution, for example, to produce negative results [18
Recently, hair testing has gained an interesting role in harm reduction, which is often used as a therapeutic strategy to disclose how individuals who use drugs learning about exposure may affect their drug taking and their adherence to treatment [10
However, hair analysis can be tricky, and some common pitfalls should be known to avoid wrong results and interpretation. The sample collection is the first step that should be considered, especially in case of segmental analysis. Given that hair growth differs in each region of the body, the collection of head hair, for example, has to be done in the vertex area behind the head where the growth rate is steady (~ 1 cm/month). Hair strand should also be oriented to distinguish between the proximal (recent) and distal side which can affect the interpretation of consumption trend (increase or decrease) [18
]. From the analytical point of view, the analysis of hair samples may be more difficult than conventional matrices like urine, because it requires more preparation steps like decontamination or digestion where losses or degradation of drugs may affect the detection and increase variability of quantitative results. The choices of the right solvent, the extraction time and the analytical instrument are some of the important factors to consider for achieving unambiguous identification and accurate quantification [19
The most serious pitfalls of hair analysis are in the interpretation of the results. While the incorporation of drugs in hair is dose related, it is well known that there is no good correlation between the dose of a drug taken and hair concentration because of interindividual variation. This can be due to differences in metabolism, frequency of use, purity of drugs, hair color, or the use of cosmetic products (dying, bleaching, etc.) among many factors that affect the incorporation of drugs into the hair matrix. It is, therefore, difficult to speculate on how much or how often an individual used drug. However, intraindividual correlation when analyzing multiple hair segments or different time frame could be used to monitor drug use pattern and detect changes in the frequency (decrease or increase) of use in the same subject [20
Finally, while a positive result from the analysis of hair suggests that a person has used or been exposed to a drug (including external contamination), a negative hair result, however, does not categorically mean that the person did not use drugs if they have been exposed to a drug infrequently or in low doses. Consequently, it is important that the limitations of the test should be highlighted when reporting the results [18
Even if this approach remains scarce and requires some particular skills, it deserves to be considered as a direct and long-term indicator of drug exposure. Some prevalence and analytical data regarding new psychoactive substance (NPS) study by hair testing in high-risk populations are now available [21
In this article, we report a case of a 43-year-old male hospitalized in intensive care unit (ICU) for 3-fluorofentanyl unintentional overdose from a mislabeled product and demonstrate the usefulness of hair to determine his exposure profile to other fentanyl analogs over 1 year.
Materials and methods
Chemicals and reagents
The reference standards of alfentanil, carfentanil, furanylfentanyl, furanylnorfentanyl, butyrylnorfentanyl, acetylnorfentanyl, norcarfetanil, norfentanyl, methylfentanyl, furanylnorfentanyl, methoxyacetylfentanyl, fentanyl-D5, acetylfentanyl-D5 and norfentanyl-D5 were purchased from LGC Standards (Molsheim, France); sufentanyl, fentanyl, and butyrylfentanyl from Lipomed AG (Arlesheim, Switzerland); 3(meta)-fluorofentanyl, ocfentanil and acetylfentanyl from Cayman Chemical Company (Ann Arbor, MI, USA); U-47700 from Chiron AS (Trondheim, Norway); acetonitrile, dichloromethane, hexane, ethyl acetate, formic acid and methanol from Sigma-Aldrich (Paris, France) in MS or high-performance liquid chromatography (HPLC) grade; sodium carbonate and hydrogen carbonate from Prolabo (Paris, France). Ultra-pure water (18 MΩ) was obtained by ultrafiltration with a Q-Pod (Millipore Corp., Molsheim, France). Formate buffer containing 2 mM ammonium formate in 0.1% formic acid was prepared in ultra-pure water and stored after each analysis at +4°C away from light for a maximum of 1 week.
Calibration standards and quality controls
Eight calibration standards containing a mixture of screened fentanyls at 1, 2.5, 5, 10, 50, 100, 500, and 1000 pg/mg of hair were prepared by spiking stock solutions of reference standards with appropriate volume into a 20-mg drug-free hair powder. Quality controls (QCs) were prepared at concentrations of 5, 250 and 750 pg/mg by the same way. Drug-free human hair was obtained from healthy subjects who are not medically treated.
Hair sample preparation
Hair locks were obtained from the posterior vertex. Hair was washed once with warm water and decontaminated twice using dichloromethane (immersion for 2 min in each step). It was then ground into a fine and homogeneous powder using a ball mill (MM200; Fisher Scientific, Illkrich, France). Each 20 mg was incubated in 1 mL of phosphate buffer at pH 5.0 at 95 °C for 10 min, in the presence of an appropriate amount of deuterated internal standards (ISs: fentanyl-D5, acetylfentanyl-D5 and norfentanyl-D5). After spiking the corresponding volume of calibration standard or QC working solution when necessary, liquid-liquid extraction was performed by 4 mL of a mixture of hexane/ethyl acetate (v/v, 1:1). After agitation and centrifugation for 20 min, the organic phase was recovered and evaporated to dryness. The residue was reconstituted in 80 µL of mobile phase and 10 µL was injected into the chromatographic system.
Liquid chromatography–tandem mass spectrometry system and conditions
A liquid chromatography–tandem mass spectrometry (LC–MS/MS) system derived from a previously published method [22
] and enriched by the addition of 17 NSOs was employed in the present study. Chromatography was performed on a Dionex Ultimate 3000 pump (ThermoFisher, Les Ulis, France) using a Hypersyl Gold PFP column (100 × 2.1 mm i.d., particle size 1.9 µm) preheated at 30 °C. The mobile phase was a gradient of acetonitrile (A) and 2 mM sodium formate buffer aqueous solution with 0.1% formic acid (B) starting from 20% (A) to 90% (A) in 10 min at a flow rate of 300 μL/min. The total run time was 12 min. Compounds were detected by a TSQ Endura triple-quadrupole mass spectrometer (ThermoFisher) equipped with an electrospray ionization source set in a positive mode with ion spray potential at + 3.5kV. Capillary temperature was set at 350°C. Nitrogen (Nitrox UHPLCMS 18, nitrogen generator; Domnick Hunter, Villefranche sur Saône, France) was employed as sheath gas at 35 arbitrary pressure unit. The argon gas collision-induced dissociation was used with a pressure of 1.5 mTorr. Data were collected in selected reaction monitoring (SRM) mode, with two m/z
transitions per analyte. Data acquisition was performed using Xcalibur and LC-Quan softwares (both ThermoFisher).
Retention times, SRM transitions with corresponding collision energies for the 17 screened NSOs and their ISs are shown in Table 1
Retention times, selected reaction monitoring transitions and collision energies of the screened new synthetic opioids (NSOs) and their internal standards (ISs)
Method validation procedure
Selectivity and carry-over
Drug-free blank hair was analyzed to determine whether endogenous hair constituents interfere or not at the retention times and on the ion channels of NSOs. A replicate of blank sample was also analyzed immediately after the highest calibration standard to determine the carry-over.
Six calibration curves were prepared with eight calibration standards ranging from 1 to 1000 pg/mg. Quantification was achieved by plotting the peak area ratios of NSOs to their respective ISs. Back-calculated concentrations of the calibration standards had to be within 85–115% of the nominal concentrations.
Limits of detection and quantification
The limit of detection (LOD) is the lowest concentration for each NSO that can be detected with a signal-to-noise ratio greater than 3. The limit of quantification (LOQ) was the lowest concentration of each NSO achievable with an accuracy of ± 20%.
Accuracy and precision
The accuracy (bias) and precision (coefficient of variation: CV) of the assay were determined at three QC levels (low 5; mid 250; high 750 pg/mg). For the intraday assay, six replicates of each QC were processed on the same day. For the interday assay, six replicates of each QC level were processed at three different days. The concentrations obtained were analyzed using analysis of variance (ANOVA), which separated the intraday and interday standard deviation and the corresponding CVs. An accuracy within the range 85–115% of the nominal values and a precision with a CV of ± 15% were required.
Matrix effect and overall method recovery
Matrix effect was defined as the ratio of the mean peak area obtained by analyzing six different blank hair matrices spiked after extraction with ISs and NSOs at two concentrations (100 and 500 pg/mg) to the mean peak area obtained in an aqueous solution at the same concentrations. Overall method recovery was defined as the ratio of the mean peak area obtained by analyzing six different blank matrices spiked before extraction with ISs and NSOs to the mean peak area obtained in an aqueous solution at the same concentrations. The determination of method recovery remains optional for mass spectrometry techniques.
As shown in a recent study in Canada, 75% of individuals reporting nonmedical opioid use and denying fentanyl use had their urine test positive for fentanyl [43
]. Another study showed that, among patients denying known exposure to fentanyl and seeking opioid withdrawal management, two-thirds tested positive for fentanyl in urine [44
]. Segmental hair analysis is especially useful for identification and monitoring the use of designer drugs, which often are different from ones that they bought; it enables to follow for months to nearly a year as we did. However, the number of cases for NSO identification and/or quantification using hair samples is still not many [45
More data are needed to know to what extent the hair analysis can be effective in harm reduction. A large-scale adoption of this approach would allow a better understanding and response to NSO crisis at both health and regulatory levels. To our knowledge, this is the first report to describe the quantification of 3-fluorofentanyl and methoxyacetylfentanyl in hair samples collected from an authentic abuser.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.