Detection of fentanyl, synthetic opioids, and ketamine in hair specimens from purposive samples of American and Italian populations

With the current crisis related to the diffusion of fentanyl and other novel opioids in several countries and populations, new and effective approaches are needed to better elucidate the phenomenon. In this context, hair testing offers a unique perspective in the investigation of drug consumption, producing useful information in terms of exposure to psychoactive substances. In this research, we applied targeted ultra‐high performance liquid chromatography tandem mass spectrometry (UHPLC–MS/MS) analytical methods to detect novel synthetic and prescription opioids and other common controlled psychoactive drugs in the keratin matrix. A total of 120 hair samples were analyzed from the United States (US) and Italy, segmented when longer than 6 cm, and then analyzed. In the 60 samples (83 segments in total) analyzed from a purposive sample of data collected in the US, fentanyl was detected in 14 cases (16.9%), with no detection of nitazens or brorphine. We also detected fentanyl metabolites, despropionyl‐p‐fluorofentanyl, and prescription opioids. In the 60 samples collected in Italy (91 segments in total), ketamine was the most prevalent compound detected (in 41 cases; 45.1%), with ketamine demonstrating a strong correlation with detection of amphetamines and MDMA, likely due to co‐use of these substances in recreational contexts. Several common drugs were also detected but no exposure to fentanyl or its analogs were detected. Results of this retrospective exploration of drug use add to increasing evidence that hair testing can serve as a useful adjunct to epidemiology studies that seek to determine biologically confirmed use and exposure in high‐risk populations.


| INTRODUC TI ON
In the past decade, illicit opioid use has progressed from nonmedical use of legal analgesic drugs such as hydrocodone, oxycodone, and tramadol, to the diffusion of illicitly manufactured fentanyl and its analogs, often referred to as "fentalogs" [1,2]. In addition, a new class of synthetic opioids referred to as nitazenes has been recently reported in several illegal drug markets [3]. Many of these novel synthetic opioids (NSOs) are considered particularly risky due to their high potency and because they are often introduced into the market as cutting/adulterant agents of drugs such as heroin or simply as cheaper substitutes for other drugs [4,5].
While heroin was the first drug to become frequently replaced or cut with NSO, these compounds began to appear to counterfeit pills representing common prescription drugs (e.g., oxycodone, alprazolam) and other illegal powder drugs such as cocaine [5][6][7][8][9][10], raising a major health concern for unaware users. In parallel, however, preference for fentanyl has increased among some populations, leaving doubt regarding whether fentanyl is more demand-led or supply-led [11]. However, it is important to note that currently, the synthetic opioid crisis is centered in North America. In 2022, in the United States (US), there were 71,238 deaths linked to use of synthetic opioids such as fentanyl [12], and in 2021, there were at least 10,000 fentanyl seizures in the US which weighed over 10,000 kg in total [13]. In Europe, however, among 12 countries providing seizure data to the European Union Early Warning System, in 2021, there were only 187 recorded fentanyl seizures (weighting 5.5 kg in total) [14]. Heroin has largely been replaced by fentanyls in the US but heroin is still the most common illicit opioid in Europe [14]. In Italy, an average of 0.74 tons of heroin have been seized annually between 2011 and 2021 [14].
Despite differences in the opioid and other drug landscapes between the US and Europe, new and effective approaches are needed to monitor shifting drug-related phenomena. For example, in the US, in 2021, there was an increase to 24,538 deaths linked to cocaine use and 32,856 deaths linked to other psychostimulant (mainly methamphetamine) use [12], but the vast majority of such cases involve co-use of opioids [15]. A greater understanding of couse (or co-exposure) of illicit fentanyl, opioids, and other psychoactive drugs can help adapt and improve existing interventions aimed to reduce overdose mortality, together with broad integrated public health strategies based on overdose education and prevention and to support the drugs debate [16][17][18].
In this context, hair analysis has proved to be an easy and effective tool to investigate the prevalence of use of psychoactive substances, since the keratin matrix allows for the detection of past drug exposure and for the investigation of the chronological profile of the exposure to one or multiple compounds. Furthermore, hair analysis is now based on multiclass methods for both well-known and emerging compounds, allowing for the investigation of different consumption patterns, including co-use of common drugs (including prescription opioids), as well as occasional vs. frequent NSO use or exposure [19][20][21]. While several papers have described multianalyte screening methods capable of detecting NSOs [22], few have presented results from real samples [23][24][25]. Polydrug use has generally been shown to be common based on the aforementioned studies, involving several psychoactive substances and not only heroin [26,27]. In general, fentanyl has been the most frequently detected compound among the class of fentalogs [8,19], suggesting that it is the most prevalent molecule while the less common analogs tend to be co-used with other drugs and are thus not consumed in isolation.
Other typical matters of current discussion (in order to provide a definitive interpretation of positive versus negative results) are: (i) the meaning of quantitative results, in terms of occasional and frequent use or exposure), and (ii) the identification of proper metabolites to discriminate direct exposure from potential external contamination.
In this paper, we present our ultra-high performance liquid chromatography tandem mass spectometry (UHPLC-MS/MS) analysis of a subset of hair samples collected in the US and in Italy based on purposive sampling methods. All samples were screend for fentalogs, prescription opioids, nitazens, brorphine, and other common controlled psychoactive drugs.

| Reagents and standards
All chemicals, including methanol, formic acid, and acetonitrile, were purchased from Sigma-Aldrich while ultra-pure water was obtained using a Milli-Q® UF-Plus apparatus (Millipore). The analytical standards of the target analytes and deuterated internal standards (norfentanyl-D5, fentanyl-D5 and oxycodone-D6) were purchased from LGC Promochem and Sigma-Aldrich (purity >99%, concentration between 0.1 mg/mL and 1 mg/mL), or kindly provided by the Italian National Early Warning System (provided at a concentration of 0.02 mg/mL). The list of target analytes is presented in Table 1.
All stock standard solutions were prepared in methanol at 1 mg/ mL and stored at −20°C until used. Working solution of 42 analytes • In the US, fentanyl was detected in 16.9% of hair segments, with no detection of nitazens or brorphine.
• In Italy, ketamine was the most prevalent compound detected, found in 45.1% of hair segments.
Detection of ketamine was strongly correlated with detection of amphetamines and MDMA.
• Hair analysis is effective in investigating the diffusion of new psychoactive drugs. TA B L E 1 List (ordered by RT) of the monitored transitions, their instrumental parameters and the related internal standard for the screened compounds. (identified among the most common synthetic opioids and those recently observed by the warning systems) and internal standard solution were prepared at the final concentration of 1 μg/mL by dilution with methanol.

| Sample collection and preparation
In this study, we focus on two purposive samples of adults-from All samples were analyzed up to a maximum of the proximal 12 cm, since the study aimed to explore the intake of drugs in the 12 months prior to collection (assuming a normal hair growth rate of 1 cm per month). When hair was ≤6 cm, it was analyzed in its entire length. When hair was longer than 6 cm (54 samples), two segments were prepared for analysis (with one representing roughly the past six months and the other representing roughly the previous 6-12 months). Therefore, a total of 174 separate segments was analyzed. The targeted screening for common drugs was performed using previously published and fully validated methods [30,31].
Existing procedures for novel opioids [32,33] were adjusted to expand the panel of screened molecules. A partial validation was performed, aimed to verify the method sensitivity and the quality of the calculated concentrations. The limits of detection are presented in Table S1, while data for trueness and precision at three different concentration levels are presented in Table S2.

| Statistical analysis
We used descriptive statistics to describe the number of segments testing positive for various drugs in each country, and among positive cases we also described the range of levels of molecules detected.
Within the Italian sample, we also computed Spearman correlations to determine the extent to which level of detection of each drug was correlated. Python version 3.11.3 has been used to compute the correlation matrices, involving numpy, pandas, and seaborn libraries.

| Testing for NSO
In  Table 2.
Three prescription opioids were detected in seven segments Novel and prescription synthetic opioid identification was much less common in the 60 samples collected in Italy. In particular, fentalogs were never detected, while only five segments (5.5%) were positive for at least one compound such as hydrocodone, oxycodone, or tramadol. Two subjects were positive for all three prescription opioids. In one case, two segments were obtained from the same sample, showing the same trend of consumption (hydrocodone at 19 and 22 pg/mg, respectively).

| Testing for common drugs
Samples were considered positive in accordance with international cut-offs for parent drugs and metabolites [35]. In the group of samples from the US, cocaine was the most prevalent substance found in the samples, with 19 segments (22.9%) resulting above the cut-off for either cocaine or its metabolite benzoylecgonine (BZE).
Cocaethylene was detected above 0.05 ng/mg only in five cases. Correlation matrices for the measured levels of common drugs in the Italian population of hair samples are presented in Figure 1.
High correlation coefficients suggest that subjects who were more exposed to ketamine were more exposed to amphetamine, possibly due to a co-use of the substances in certain recreational contexts.
In this study, we investigated samples from two different populations within these two countries, and results should not be directly used to indicate prevalence of drug use, as we used purposive sampling. As such, results are not generalizable to US or Italian populations, but rather present a snapshot of drug use within select populations in each country. Indeed, all cases of synthetic opioid detection were in the US but we focused on a sample in which many participants reported recent synthetic opioid use.
Most use of common party drugs such as ketamine and MDMA were detected in the Italian sample but we must keep in mind that these were individuals receiving harm reduction services associated with nightlife. People who attend nightclubs in particular tend to report higher prevalence of use of such drugs than the general population [38][39][40]. Prevalence of past-year ketamine use among young adults is estimated to be <1% in both the US and in Europe [14,41], although seizures of the drug appear to be increasing at a similar rate [14,42]. While prevalence of past-year heroin use is estimated to be <1% in both the US and Europe [14,43], in the US, synthetic opioids such as fentanyl analogs indeed are more available and have been involved in hundreds of thousands of deaths in recent years [13,44]. As such, it is important to note both where biological specimens are collected but also the populations from which they are obtained. This is because results will vary in particular across high-risk populations (e.g., nightclub attendees, people who utilize drug checking services) and the general population. Thanks to the longer detection window of hair (in comparison to much shorter detection windows provided by urine, saliva, and blood), drug exposures occurring 1-2 weeks up through a year before hair collection can provide retrospective results to inform scientists and public health practitioners about the diffusion of drugs in their countries. Hair analysis results based on real hair samples can provide information regarding both intentional and unintentional exposure to NPS/NSO, both with and without use of common controlled drugs. As such, hair testing can serve as an addition to epidemiology studies that seek to incorporate biological testing with survey research. The combination of surveys and hair testing can thus be used to monitor drug exposure in a more effective manner than using surveys or biological testing alone.

ACK N OWLED G M ENTS
The authors acknowledge the NDEWS team for the data collection methods and interviewing efforts.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors have no conflicts of interest to declare.