The proposed LC–MS/MS method allowed confirmation of the identity, as well as the quantification, of a substantial number of illicit drugs and their metabolites in bile in postmortem cases of SD occurring both for overdose in adults and for intoxication in children.

Drugs that are lipophilic and with high molecular weight are more prone to undergo hepatic metabolism and excretion via biliary route (Ghibellini et al. 2006). Biliary elimination may impact on the pharmacological effect and toxicity of drugs, even because of reabsorption along the gastrointestinal tract, but so far there is limited information regarding the illicit drugs kinetics because of the difficulties to obtain bile samples from healthy subject (Ghibellini et al. 2006). Several illicit drugs, and in particular the metabolites produced after hepatic metabolism, were found in bile in higher amount: for instance, benzodiazepines were detected in their hydroxylated form, as well as natural and synthetic opioids in their conjugated form (Al-Asmari 2020; Bertagni et al. 1972; Kuwayama et al. 2008; Misra et al. 1977; Tominaga et al. 2016).

Measuring drugs concentration in bile is of forensic interest because of the need to test at least 2 matrices since sometimes the traditional ones are not available and the toxicology laboratory could meet technical difficulties due to putrefaction mechanisms in the biological materials collected (Ojanperä et al. 2012). Furthermore, there is evidence about the extended detection capability in bile: the concentration of several drugs are up to 520 fold times higher than the ones found in blood (Ferner and Aronson 2018). Other alternative matrices, such as vitreous humor and cerebrospinal fluid, could be employed but they need the addition of stabilizers (Bévalot et al. 2016a) or are characterized by low concentration of drugs (Tominaga et al. 2015).

To the authors’ knowledge, this is the first time a validated LC–MS/MS was used to screen such a large quantity of drugs of abuse: previous works were focused on the quantification of only one analyte or class of analytes using this technique (Al-Asmari 2019, 2020; Zuccarello et al. 2023). Using the same analytical method for several drugs in different matrices gains in value because it allows a more accurate comparison between the classes of drugs from a pharmacokinetic point of view and between the different specimens available for the analysis (Ferner and Aronson 2018; Vanbinst et al. 2002). However, the present analytical method did not focus on antidepressant drugs, as other studies did before (Zuccarello et al. 2023). Furthermore, other techniques, such as GC–MS, were typically employed for achieving the aim of the present paper (Tominaga et al. 2016).

In the proposed method, undiluted bile resulted unmanageable for analysis because of its influence on both chromatography and ionization; indeed the biological material presented a very low pH, probably due to putrefaction mechanisms (Donaldson and Lamont 2013). Instead, bile diluted 1:100 resulted to be the best material to work with since it allowed quantification of the analytes without having a significant matrix effect. Such a large dilution allowed to use a small amount of matrix, contrary to what was observed for other analytical methods (Tominaga et al. 2016).

Sample preparation of the present method is fast, simple and consists mainly in enzymatic deconjugation and protein precipitation through organic solvents. Since several drugs undergo conjugation through hepatic metabolism, enzymatic hydrolysis needs to take place to detect the original molecule; classic examples are opioids that are excreted for a large extent as glucuronide products (Concheiro et al. 2018). Furthermore, this sample preparation did not use solid-phase extraction, a time-consuming further step usually needed for purifying bile, as other previous works did (Bévalot et al. 2016b; Tominaga et al. 2016).

Validation of the analytical method according to the more recent ICH guidelines was also performed to assess an appropriate selectivity, specificity, sensibility, linearity, accuracy, precision, matrix effect, and recovery and applicability to autoptic samples.

Comparing the validation parameters with the ones of previously reported analytical methods, LOD and LOQ values referred to our LC–MS/MS method are greatly lower suggesting that we applied a novel and more sensitive MS/MS equipment in our procedure compared to the former ones (Launiainen and Ojanperä 2014; Vanbinst et al. 2002). The benzodiazepine clonazepam is an exception; in particular, Launiainen and colleagues developed a liquid chromatography/time-of-flight mass spectrometry (LC-TOF–MS) method with a LOQ value of 10 µg/L for clonazepam in order to investigate femoral blood concentration of drugs (Launiainen and Ojanperä 2014). Furthermore, Vanbinst et al. considered in their method a larger range of linearity, allowing an accurate analysis also at very high concentrations (Vanbinst et al. 2002).

Noteworthy, the already reported postmortem drugs concentration in bile resulted to be higher than our LOD values allowing identification and, in most cases, also the quantification of the illicit drugs and their metabolites found in this matrix (Launiainen and Ojanperä 2014; Tominaga et al. 2016).

Carry-over, defined as the contribution of the response of analytes and IS in the subsequent runs, is a big issue in quantitative analysis; therefore, several regulatory bodies of requirements for method validation, such as ICH and Food and Drug Administration (FDA), fixed specific acceptance criterion (Jogpethe et al. 2022). The present method satisfies this criterion. However, especially in case of high exposed samples, complete elimination of carry-over effect is not possible and there is the need to evaluate if it affects the accuracy and precision of the analytical method (Jogpethe et al. 2022).

Interestingly, matrix effect has never been taken into consideration before. Testing matrix effect, we found that the present method was not ideal for the opioid buprenorphine and the metabolite of cannabinoids 11-Nor-9-carboxy-Δ9-tetrahydrocannabinol because of the suppression due to the matrix.

Tominaga investigated the recovery and reported values from 60 to 70% for phenobarbital to > 95% for codeine (Tominaga et al. 2016). Instead, 7-amino nitrazepam resulted in the molecule with a lower recovery (79%) in our study.

The present work has the purpose also to test the applicability of the LC–MS/MS method on real samples derived from postmortem specimens and to compare the concentrations of analytes in bile with the ones of the traditional matrices (blood and urine). Beyond urine, peripheral blood was used for the comparison because it is less likely be subject to postmortem drug redistribution, referred as changes in drugs concentration due to organ injury and subsequent release of its content in blood (Yarema and Becker 2005). Also, it is important to notice that blood is representative of the potentially lethal concentration at the time of the death, instead bile and urine may demonstrate past drug exposure that could also be indicative (Bierly and Labay 2018). Indeed, as expected, products of hepatic metabolism, such as norcodeine and 6-MAM (Smith 2009), were detected in bile and urine, but not in blood.

Several works investigated bile:blood ratio in order to find a fixed relationship between the levels in the matrices (Ferner and Aronson 2018). On the basis of these evidences, we calculated bile:blood and bile:urine ratios, since bile contains drugs already metabolized. Even if our results did not differ greatly from the ones reported by Ferner et al., there is no fixed relationship, even taking into account the classes of illicit drugs (Ferner and Aronson 2018).

Even if to date there is no fixed relationship between the several matrices compared, the inter-matrix ratio resulted to be in line with the toxicokinetic of the illicit drugs tested and with previous works. Generally, comparing urine and bile concentrations, there is no wide difference since they, unlike blood, accumulate drugs after hepatic metabolism. The only metabolite that is greatly higher in bile than in urine is THC-COOH, probably due to the postmortem distribution of blood to bodily fluids and tissues and the accumulation in bile following multiple doses, as another paper has already reported (Zughaibi et al. 2023). Furthermore, the bile:blood ratio gave interesting information. For instance, hydroxylation is an important metabolic step for elimination of benzodiazepines, particularly for alprazolam (Ait-Daoud et al. 2018). After abuse of natural and synthetic opioids, their derivatives, such as 6-MAM and EDDP, were found only or in greatly higher amount in bile, as expected (Al-Asmari 2020; Baselt and Casarett 1972). Furthermore, only the concentrations of tramadol and cocaethylene were lower in bile than in blood. In the former case, this could be due to the metabolism of tramadol in O-DM-tramadol, instead the latter to the metabolite half-life, longer than cocaine and its other metabolites (Andrews 1997; Gong et al. 2014).

In conclusion, our results generally confirmed higher concentrations in bile than in blood suggesting a potential role of bile in helping pathologists to identify the cause of death. Based on our data, quantification of drugs of abuse in bile resulted convenient in most cases, except for tramadol and cocaethylene.