Stability of synthetic cathinones in clinical and forensic toxicological analysis — Where are we now?

Understanding the stability of analyzed drugs in biological samples is a crucial part for an appropriate interpretation of the analytical findings. Synthetic cathinones, as psychoactive stimulants, belong to a major class of new psychoactive substances. As they are subject to several degradation pathways, they are known to clinical and forensic toxicologists as unstable analytes in biological samples. When interpreting analytical data of synthetic cathinones in biological samples, analysts must be aware that the concentration of analytes may not accurately reflect the levels at the time they were acquired owing to many factors. This review provides (i) an overview of the current scientific knowledge on the stability of synthetic cathinones and/or metabolites in various human biological samples with a focus on factors that may deteriorate their stability — such as storage temperature, length of storage, matrix, pH, type of preservatives, concentration of analytes, and the chemistry of the analytes — and (ii) possible solutions on how to avoid such degradation. The PubMed database as well as Google Scholar was thoroughly searched to find published studies on the stability of synthetic cathinones since 2007 by searching specific keywords. A total of 23 articles met the inclusion criteria and were included in this review. Synthetic cathinones that carry methylenedioxy or N -pyrrolidine ring showed higher degradation resistance over other substituted groups. Acidification of samples pH plays a crucial role at increasing the stability of cathinones even with analytes that were frequently considered as poorly stable. This review also provides several recommen-dations for best practice in planning the experimental design, preservation, and storage conditions in order to minimize synthetic cathinones' degradation in human biological samples.


| INTRODUCTION
In clinical and forensic toxicology investigations, instability of analytes under the influence of various conditions may occur in biological samples. 1 The loss of analytes may take place owing to chemical degradation, enzymatic metabolism, or the existence of interfering substances due to matrix degradation, involving (but not limited to) improper sample storage conditions, handling, and/or transportation. 2 Stability studies may help to determine the time frame that must be taken into account between sample collection and analysis, or to determine the ideal storage conditions for a sample to be preserved for a purpose of drug testing in order to minimize or reduce sample loss, thus not significantly compromising the analytical results. In toxicology laboratories, the stability of classical drugs of abuse (i.e. amphetamine, benzodiazepines, cannabinoids, cocaine, codeine, and morphine) in biological samples has been widely described. [2][3][4][5][6][7][8][9] One of the reasons of increased interest in investigating the stability of classical drugs is certainly that their existence in these samples may lead to sanctions for the users, or if the workplace drug testing (WDT) guidelines require certain laboratories to store all confirmed positive samples for a minimum of 1 year, for a potential retesting. 10,11 Thus, drug stability in biological samples is an important consideration for assuring accurate reports of drug concentration and/or its metabolites. 12 Over the past decade or so, new psychoactive substances (NPSs) gained a high popularity on the illicit drug markets, possibly because of the reducing purity and availability of classical stimulant drugs such as amphetamine, 3,4-methylenedioxymethamphetamine (MDMA), or cocaine. 13,14 Besides the latter reasons, additional factors may motivate the use including the expansion of online and virtual markets, cost, and legal status. 15,16 Furthermore, NPSs may not be included in conventional laboratory drug testing screening; indeed, there is often lack of a reliable immunoassays screening for NPSs owing to their structural similarity of classical drugs. 17 Moreover, as NPSs continuously appear on the drug market every year as structural isomers or analogs, drug testing laboratories often cannot keep track for all of them-having to continuously update their existing analytical methods and constantly obtain new reference materials, which is often time-consuming and expensive. The latter are sometimes commercially unavailable. The existence and availability of drugs that may not be initially controlled inevitably offer an attractive alternative to recreational drug users. One of the reasons is the desire to minimize the risk of being prosecuted for illicit drug use.
By far, the NPS class of synthetic cathinones (SCs), derived from naturally occurring cathinone found in the leaves of Khat, became an attractive alternative stimulant group of drugs to accompany the classical banned stimulant compounds. 18,19 These recreational drugs are also known as "legal highs" as they were initially legal in Europe and the UK and sold under deceptive names, for example, "bath salts," "research chemicals," or "plant food." 20,21 The first attempts for the synthesis of SCs started in the end of 1920s with methcathinone and mephedrone in 1928 and 1929, respectively. 22 Thereafter, a few other derivatives have been designed and investigated for possible therapeutic use, for instance, bupropion, which plays an important role in the treatment of depressive disorders. 23 Recreational SCs use only began in the mid-2000s when methylone emerged on the illicit drug market in the Netherlands and Japan. 24 27,28 which is considered the point when SCs were starting to emerge on the illicit drug markets. 29 Another reason is that other researchers have summarized the literature for stability in human biological samples before 2007 including SCs. 30 This review was restricted to studies concerning human samples only to accurately simulate the real conditions. Ultimately, this review aims to provide insights of current knowledge about the stability of SCs in biological samples and offers a critical assessment on the limited studies conducted in the mentioned time period to understand limitations and to provide suggestions and improvements in future work.

| MATERIALS AND METHODS
The most relevant findings of SCs in the available literature concerning their in vitro stability studies in human biological samples were reviewed. Different factors were used as inclusion or exclusion criteria of search selection to extrapolate the required scientific information. Published articles between January 1, 2007, and March 20, 2020, were conducted via searching of PubMed. Google Scholar was also used to identify any further relevant study. In addition, articles with at least one of the following criteria were excluded: (1) publications not in English; (2)  4-Methylephedrine (dihydromephedrone) 33,35,39,40 Buphedrone ephedrine (dihydrobuphedrone) 33 (Continues) human samples. The search was conducted by applying the following keywords: "synthetic cathinone" or "cathinone derivative" or "stability of novel psychoactive substance" or "stability of new psychoactive substance" or "stability of NPS" or "stability of synthetic cathinones" or "legal highs" or "bath salts" or "designer drugs." Further, combining keywords, for example, "AND," "NOT," and "OR," was also used to help with databases using Boolean operator. Dihydro-nor-mephedrone 40 Hydroxytolyl-mephedrone 40 N-Succinyl-nor-mephedrone 35 Nor-mephedrone 33,35,40 with NaF/citrate buffer. The other SCs, that is, methylenedioxy SC types, were also degraded, however at a lower level (5% loss). In contrast, at 5 C, all analytes that were preserved with NaF/KOx were relatively stable up to 3 days; but thereafter, cathinone, methcathinone, ethcathinone, mephedrone, and 4-FMC were less stable. The researchers pointed out that the degradation rate of SCs increased with increasing blood pH level but also varied with chemical structures. Notably, all ephedrine analytes were stable over the whole study under the same storage conditions, additives, and pH. This finding highlights the increased stability of the hydroxy-group-containing ephedrines compared with β-keto group-containing analytes, namely, SCs.
In published methods for the detection and quantification of  and >26%, respectively) and naphyrone that showed instability after three F/T cycles (degradation of >22%).
In the validation of a liquid chromatography-tandem mass spec- Analytes were spiked at 100 and 1,000 ng/ml and monitored over 6 months at −20 C, 4 C, 20 C, and 32 C. Moreover, the influence of pH was also evaluated by adjusting the pH of urine to either 4 or 8. In addition, analysis of variance (ANOVA) was also used to determine statistical significance (p = 0.05). Consequently, apart from the analyte concentrations, this study revealed that stability was found to be highly affected by other factors such as storage temperature, pH, and the chemistry of the analytes. Tertiary amines were the most stable under the given conditions than secondary amines, including ringsubstituted secondary amine. Among others, 3-FMC was found to be the most unstable with significant degradation observed after only few hours of storage, whereas MDPV and MDPBP were the most stable. Furthermore, it was noted that degradation rate decreased with lower storage temperature for all analytes, even for the most unstable ones. In addition, these analytes were found to be fairly stable in acidic urine (pH 4) when compared with those in alkaline urine (pH 8). Markedly, by comparing the most unstable analyte with the most stable with acidic and basic pH at −20 C (Figure 1), respectively, it can be observed that both analytes were relatively stable over 165 days and that their degradation rate are closely similar. Presumably, the stability of SCs at lower temperatures is highly dependent on the pH value rather than the type of analyte.
Their findings backed up their previously described hypothesis and underline the significance of sample pH and of the chemistry analytes, with some hypothesized SC substituents greatly affecting degradation resistance over other moieties.
A long-term stability study on 22 SCs and related metabolites in unpreserved urine samples was carried out by Alsenedi and Morrison. 39 The analytes covered in this study are cathinone, methcathinone, buphedrone, pentedrone, 4-FMC, mephedrone, dihydro-mephedrone and dihydro-nor-mephedrone were stable over the whole study under all the storage conditions. By contrast, both mephedrone and hydroxytolyl-mephedrone were found to be relatively unstable under the given conditions, whereas 4-carboxymephedrone and nor-mephedrone showed the poorest stability.
Overall, dihydro-mephedrone followed by dihydro-nor-mephedrone was consistently the most stable of the investigated analytes, and 4-carboxy-mephedrone represents the most unstable metabolite, with <60% remaining at both temperatures. Comparing −20 C with 4 C, analytes were generally more stable at −20 C. From their findings, it would appear the stability of dihydro-mephedrone and dihydro-normephedrone is associated with the presence of hydroxyl group resulting from keto reduction, which is missing in mephedrone and other metabolites. In addition, −20 C is the ideal storage temperature to prevent investigated analytes from degrading in blood samples. In blood samples, no indication of instability was appreciable for pentylone, N-ethylbuphedrone, methylone, eutylone, and dimethylone when stored at −26 C for 6 months, whereas the concentrations dropped significantly for the rest of the analytes, ranging from 12% (for α-PBP) to 89% (for 4-CMC). On the other hand, all analytes were unstable at RT. Similarly, 5 C for 6 months resulted in instability for all analytes, except for dimethylone, eutylone, and pentylone.
The authors concluded that storage temperature, pH, and type of biological sample had a great impact on the stability of compounds.
Notably, great influence on the stability was observed when samples were left at RT, whereas samples stored at −26 C showed greater stability. However, by comparing urine samples (pH = 5) with blood (pH = 7) at −26 C, analytes were more stable in urine, as possibly expected by the greater biochemical complexity of blood compared with urine matrix, but also again highlighting the critical role of the pH in maintaining sample integrity. As mentioned, clearly the acidity of pH plays a crucial role at increasing the stability of SCs even with those analytes that are frequently regarded as poorly stable (Figure 2).
A further experiment was conducted to compare the effect of repeated thawing and freezing samples on the stability of analytes.
From their findings, similar results were obtained for repeated thawing (24 F/T cycles) with samples not thawed.  Table 2. 3.3 | Stability of extracted samples on autosampler (from hour to days)

| Stability investigations for SCs in alternative samples
Investigations on processed (i.e. extracted) sample stability help to assess the integrity of an analyte, which has been extracted from a biological sample, for the period of its storage until the time of analysis or after initial injection. This is because instability of analytes may occur not only in biological samples but also in processed/extracted samples ready for analysis. It is therefore important to assess the impact of storage conditions against expected and/or unexpected situations that may arise when samples are left on an autosampler, for example, large batch size or technical faults of the instrument. Table 3 summarizes the key results of processed sample stability of SCs.

| Analytical techniques and types of samples for evaluating the stability of SCs
In the case of SCs, the vast majority of these methods has been con-

| Study design and reporting of results
The evaluation of processed sample stability is provided as a part of validation parameters in a number of official guidelines such as the US Food and Drug Administration (FDA), 52    prevent degradation if occurred. According to SWGTOX guideline, processed sample stability is often determined by comparing results of QC samples prepared from reference materials and then analyzing in triplicate before (reference samples) and after (stability samples) being exposed to different conditions to assess their stability against situations commonly encountered in the laboratory. In addition, it has been recommended to complete the stability evaluation at least at two concentration levels (low and high). At each concentration, average signal (i.e. absolute peak areas of analyte or ratios of peak areas of analyte to internal standard) of stability samples at each time interval is compared with average signal of reference samples. When comparing stability samples with reference samples, differences in the average signal should be higher or lower than the acceptance criterion of method accuracy (e.g. ±10% or ±20%) to consider the analyte as unstable at a given storage temperature. Where appropriate, it is recommended to perform the F/T stability over at least three cycles in triplicate.
In contrast, a slightly different approach to study processed sample stability (also for long-term testing) as recommended by FDA is that QCs at each testing point should be compared with QCs of freshly prepared calibrators and QCs and that analytes should be considered stable until the accuracy reach >±15%. In other words, the results of accuracy and precision of analyte of interest at each concentration were compared with those freshly prepared and analyzed to define the length of time that samples can be stable within acceptance criteria.

| Critical discussion
The reviewed studies clearly demonstrated that concentrations of many analytes are decreased under elevated storage temperatures.
Moreover, the reported findings are comparable for most of chemical structures and draw similar conclusions; however, some studies reported different level of degradation for certain analytes and matrices as an influence of other factors such as additives and pH. In order to better explain such discrepancies among different stability studies, the chemical, enzymatic and bacterial activities need to be considered as well, as some considerations may need to be taken into account when establishing an experimental design.
It worth noting that a number of reports were excluded in this review owing to limiting the stability to "stable" or "unstable" with no further information on storage conditions or concentration level of degradation. Secondary amine-containing SCs also possessing methylenedioxy group such as butylone, ethylone, methylone, and pentylone had also experienced degradation but at lower level. 12,31 N-Pyrrolidine without methylenedioxy significantly slowed degradation at different storage conditions and even slower for

| Storage temperature
The percentages of each matrix type from the reviewed literature to evaluate the stability of SCs (n = 23) N-pyrrolidine-methylenedioxy, 1 indicating that the methylenedioxy and/or the N-pyrrolidine ring may play a role at preventing putative chemical and/or enzymatic degradation/biotransformation pathways of the parent drug in the matrix. Another suggestion was drawn by Glicksberg and Kerrigan 1 that the inability of tertiary amines to undergo oxidative deamination may be attributed to its relative stability.
Most of the analytes (50 out of 58) in the reviewed articles are parent and not metabolites. All these parent SCs share a ketone functional group in their molecule. This ketone is readily reduced to hydroxyl group, that is, reduced SC metabolites. 36

| Concentration levels
The

| Formation of breakdown products
It is obvious that most of the stability investigations of SCs in the literature has typically been limited to the parent compounds. A related issue is knowledge of instability products that may need to be monitored as part of the analytical method. This is especially the case, as is sometimes for SCs only metabolites are detected in the biological sample and not the parent analyte.
However, very few studies further investigated the degradation products that formed as a result of parent analyte instability of SCs.
Of those, only Soh and Elliott 36  Thus, it would be ideal to carry out more studies of potential newly discovered breakdown products (resulting from in vitro instability), which alongside metabolites may possess longer detection window than the parent analyte. As mentioned earlier, the stability of dihydro-mephedrone and dihydro-nor-mephedrone metabolites over the parent analyte (i.e. mephedrone) is probably the most notable example. 40 3.6.5 | Additives implicitly assume that the pH was stable throughout the experiment. One stability investigation found that sample pH could change/increase during the experiment based on temperature conditions and/or matrix type. 41 For example, blood and urine pH were initially measured and remained the same after 6 months of storage at −26 C. However, the pH in urine sample was increased from 5 to 7 after storage at 5 C and became more alkaline (pH 8) after storage at RT, possibly owing to bacterial activity. 41 Such influences in pH may explain some variations in the stability of SCs in urine samples compared with other types of matrix at least for higher temperature conditions. Thus, the addition of antimicrobial agent (e.g. sodium azide) may tackle this problem and substantially reduce the sample alkalization. 63 This step, however, has not yet been investigated for its potential effect on the stability of SCs in urine samples.

| Selection of collection container
Generally, the selection of sample container is based on the intended purpose in a way to assure that it does not compromise the analytical investigations. For example, the choice of container size should allow the sample to be as close to full as possible to reduce uncertainty about oxidative losses due to air trapped in the top of the container or "salting out" effects, which may occur if preservatives are added to the sample. 64,65 Disposable plastic containers and/or tubes are routinely used for the collection of clinical and forensic samples, especially for urine. If plastic containers are used, it is recommended to use polypropylene and not polystyrene because of the high susceptibility of the latter to crack when frozen. On the other hand, glass tubes are more suitable for long-term storage at low temperatures. The principal issue of glass is the possibility of breakage, but this concern can be reduced by using an appropriate storage rack. 64,65 Although the stability of SCs was not compared in these containers, the reported stability of blood concentrations of a variety of other illicit drugs collected in plastic and glass Vacutainer evacuated tubes was not deemed to be significant. 66

| Experimental design
In the literature, the experimental design for reporting the stability of SCs is highly variable. For example, whereas the majority of authors used two levels of concentration, 1,12,33,39,40,43,45,46,[48][49][50] few authors used three 51 or four. 35 Although analyte stability is recommended to be assessed at two concentration levels as described before, a number of authors used only one concentration. 31,32,34,36,37,41,43,44 In most publications, it seems that the number of F/T cycles was in accordance with the abovementioned recommendations, that is, three F/T cycles. The majority of authors performed the F/T stability over three cycles. 33,43,45 However, few others performed two cycles 35 or eight cycles. 48 The vast majority of authors reported their stability data using a predefined acceptance criteria; for example, average of stability samples must be within ±10%, 40 ±15%, 41,[43][44][45][46]49,51 or ±20% of that of reference samples. 12,33,35,38,39,48,51 Alternatively, Busardò et al. 37 used a statistical approach first by assessing the mean concentration of samples over certain time intervals under various conditions.
Subsequently, polynominal regression analysis was performed to check for statistically significant differences between groups when p value was lower than 0.05 (p < 0.05). Amaratunga et al. 50 used regression analysis approach to assess the stability of processed samples by plotting the absolute peak areas of analytes against certain time intervals. Negative slope significantly different from zero (p ≤ 0.05) were regarded unstable. In this test, it is necessary to plot the absolute peak areas instead of peak area ratios. This is because any degradation that may occur to peak area of internal standard would necessarily correct peak area ratio of analyte, causing inaccurate interpretation of degradation findings. 30 Glicksberg and Kerrigan 1,12 used an ANOVA test to determine whether instability (concentration below 20%) of SCs was significantly dependent on analytes, temperature, pH, and/or concentration. The authors looked first at variances within group prior to the comparison between groups to confirm that differences within group were not significant. Soh and Elliott 36 used another approach by calculating the percentage changes between target analytes and negative internal control that is known to be stable (amitriptyline) over respective times. In this approach, both target analyte and stable analyte were added to the samples at an equal concentration. The peak height ratios of target analyte to the peak height of amitriptyline were then plotted against certain time intervals. Subsequently, the percentage differences between ratios were monitored to identify any instability.
An important limitation of certain stability studies is that neither the acceptance limits nor statistical approaches for evaluating the stability are specified. 31,32,34 As described above, only few authors have followed the standards for their stability testing, and therefore, stability reports present significant variations in the literature. This has turned out to be particularly true for the various differences in acceptance criteria used in the literature, as it may lead to underestimation/overestimation of true changes. For example, certain analytes might be reported as unstable in one study, whereas the same analytes are found stable by another study owing to the wider acceptance criteria. Obviously, the need for fixed guidelines concerning the experimental designs and evaluation of stability results is important.
Although the procedure of SWGTOX describe processed sample stability, it was used to determine the stability for both short-and long-term under various conditions. Despite comparing the results with reference samples, it is recommended to start the short-and long-term stability with freshly prepared calibrators and QCs at each analysis day/time to confirm that any loss of analyte or variations in accuracy and precision is attributed to analyte instability and not to experimental conditions.
Regardless of the chosen approach, stability study is required after a full validation of analytical method for the previous reasons. Extracted samples appear to be less susceptible to degradation regardless of the employed conditions. Moreover, differences between stability of repeated F/T cycles and freezing samples appeared to be insignificant.