Typical pharmacological effects of opioids and NSOs, which also present activity at the μ-opioid receptor in the CNS, include miosis, sedation, bradycardia, hypothermia, constipation and dose-dependent respiratory depression, with the risk of fatal overdoses [26]. In intoxications involving opioids and NSOs, post-mortem findings are usually nonspecific. According to the literature [18, 27], all cases here reported did not display anatomopathological causes of death, but only signs of agonal aspiration of gastric content, which suggest the absence of neurological reflexes due to neurological impairment, as well as pulmonary and brain edema of various grades.

In case #1, protrusion of the cerebellar tonsils (a sign of severe cerebral swelling) was accompanied by unilateral necrosis of the globus pallidus. Lesions of the basal ganglia are usually bilateral and could be due to various causes, like e.g. carbon monoxide poisoning, heart disease, asphyxia, head trauma or drug overdose [28, 29]. Bilateral necrosis at the globus pallidus was detected in association with oxycodone, methadone, heroin and morphine [28, 30,31,32], so that it was suggested as a diagnostic clue of opioid intoxication [30]. Due to poor vascularization, the globus pallidus appears especially susceptible to cerebral hypoxia–ischemia. Recurrent episodes of hypoxia might be the primary cause for the necrosis, though the possibility of a direct neuronal toxicity has been discussed [32, 33]. To the best of our knowledge, this is the first unilateral globus pallidus injury found in association with a NSO consumption, even though other factors (like hypoglycemia or CPR) might also be responsible for this finding.

A wide range of concentrations has been found in fatalities related to U-47700, with sometimes very high levels of up to 2,100 ng/ml [15, 18]. The concentrations detected in case #1 (2,600 ng/ml) are among the highest so far reported, likely due to the recent consumption of a very large dose possibly combined with impaired clearance. The consumption of a large amount of U-47700 by intravenous injection, as demonstrated by the recent injection mark shown at the post-mortem examination and the extraordinarily high serum concentration, could be explained by tolerance developed after continued opioid misuse. Indeed, the man had a history of intravenous drug use and hepatitis C, which are consistent with post-mortem findings of multiple scars and a poor dental hygiene. Moreover, U-47700 was described by users for its short-lasting euphoric effects that could prompt the consumption of multiple increasing doses [13] and this might have occurred in the present case.

Given the availability of samples collected at the hospital, it was possible to estimate that U-47700 decreased to half of the initial concentration within 5.5 h, showing a serum to urine ratio of 2:1. An elimination half-life of 5.5 h is close to the finding in another case-report, where a half-life of 6 h was reported [17]. Nevertheless, poor hepatic and renal function, as shown by blood test and by post-mortem findings (signs of chronic kidney failure) might have influenced the elimination in our case. For what concerns the detection of U-47700 in post-mortem blood 9 days after the consumption (considering the 7 days of hospitalization + 2 days post-mortem interval), poor clearance, enterohepatic circulation and/or accumulation in tissues, e.g. fat tissue, as observed in pig models [34], with later redistribution to the vascular compartments are conceivable. However, the storage of the body in a refrigerated environment could also have prolonged the detection time of the NSO, since U-47700 appears rather stable up to 10 days under refrigeration [35].

The other drugs detected at the arrival to the hospital, particularly methadone, tilidine, nordazepam, clonazepam and their metabolites, were within or even below the therapeutical levels reported in the literature, although these tipically refer to serum/plasma [36]. Although these drugs, including THC, share CNS depressant effects, their role in the fatality can be considered minor. However, their detection suggests that NSOs tend to be consumed in combination with other opioids (methadone and tilidine) and benzodiazepines, as shown in other cases [35, 37,38,39]. Additional substances detected, namely naloxone, morphine, midazolam and sufentanil, were likely administered in the hospital setting. The extremely high level of U-47700 in the first clinical sample appears highly indicative of a fatal effect of the drug, justifying, together with the evaluation of circumstantial, clinical, and post-mortem findings, the assignment of a TSS of 3 to U-47700, in the setting of a combined intake of multiple drugs.

In this case it could be hypothesized that the intoxication caused primarily a CNS depression, which in turn led to a cardiac and multiorgan insufficiency, as shown by laboratory exams. The pulmonary findings might be related to a depression of the respiratory drive, followed by the aspiration of gastric material and hospitalization with CPR and mechanical ventilation, resulting in a mixed pattern of edema and emphysema.

A similar combination of drugs was detected in case #2, where U-47700 was accompanied by ethanol, methadone, duloxetine and benzodiazepines, particularly flubromazepam, and diazepam as well as its metabolites nordazepam, temazepam, and oxazepam.

The deceased had no relevant pathological findings at the autopsy, except for a fatty degeneration of the right ventricle, which is sometimes observed after chronic alcohol abuse, and could be a sign of arrhythmogenic right ventricular cardiomyopathy (ARVC), a cause of sudden cardiac death [40]. However, fatty infiltration with no myocardial atrophy, necrosis or inflammation (which are typical of ARVC) is commonly found in non-cardiac deaths [40] and the identification of severe brain edema suggests a relatively slow death, as described with opioid intoxication, rather than a sudden occurrence.

It is well known that therapeutic, toxic and fatal levels of opioids, and particularly methadone, tend to overlap [36] due to multiple factors including the development of tolerance [41]. In case #2, the circumstantial data of methadone treatment for opioids dependency matched the toxicological analysis and suggested a state of tolerance in the setting of a therapeutical consumption. Therefore, the role of co-consumed substances and U-47700 became of particular interest. The compound shows a lower potency compared to many other NSOs, but still appears to be about 10 times more potent in vivo than morphine [11]. In a series of 16 fatal (mono or mixed) U-47700-related intoxications, death was reported with concentration starting from 17 ng/ml with a median level of 247 ng/ml, which is similar to the level displayed by the deceased (204 ng/ml) in femoral blood [35]. Lower levels were detected in a mono-intoxication case by McIntyre et al. (190 ng/ml) [42] and in several cases of polydrug consumption, particularly when flubromazepam was detected [18]. The here-in presented combination of U-47700 with flubromazepam (480 ng/ml) further confirms the need to screen for designer benzodiazepines and other NPS groups beside NSOs [17, 18, 43]. Designer benzodiazepines are thought to have a synergistic effect on opioid-induced respiratory depression, as shown for traditional benzodiazepines and for ethanol (1.64 g/kg), which were both detected in case #2 [18, 44, 45]. The agitation and delirium manifested by the victim are not common in the toxidrome of opioids or designer benzodiazepines [45], but they have been reported as atypical symptoms of flubromazepam intake [46, 47] and could possibly be explained by the elevated ethanol level. Indeed, ethanol concentrations in blood and urine are consistent with acute toxicological effects after complete absorption and distribution at the time of death [48]. Other substances were either in their therapeutical range or considered metabolites [36]. Given the severe brain and pulmonary edema revealed at the post-mortem examination and the results of the toxicological analyses, the death was deemed a fatal polydrug intoxication caused by several CNS depressant agents, particularly U-47700 (TSS = 3), methadone, flubromazepam and ethanol.

A few cases on the distribution of U-47700 in tissues are available in the literature [18, 49, 50]. The whole content of the stomach (150 ml) was not analyzed, so it is possible to provide only a relative concentration.

In 26 fatalities associated with U-47700, the higher relative stomach concentrations compared to blood levels were considered suggestive of an oral intake [18].

However, in our case a syringe was found near the body and a fresh injection mark was noted at the post-mortem examination, as for injection. Studies on morphine suggest that, due to entero-hepatic circulation and reflux of duodenal contents back into the stomach, morphine levels in blood and stomach cannot be used to determine the site of administration [51]. A entero- hepatic circulation was also conceived for U-47700 due to bile and duodenum high levels in pigs following intravenous administration [34].

Nevertheless, due to its pKa, U-47700 is rapidly trapped in the acid medium of stomach contents and the (numerous) blood vessels of the stomach wall undergo autolysis already at an early stage of the postmortem interval.

Higher concentrations were detected in central blood (C) than in peripheral blood (P), with a C/P ratio of 2.3. This is consistent with a series of 26 fatalities, in which U-47700 showed a moderate propensity for post-mortem redistribution with a C/P ratio in the range of 1–8 [18]. However, several additional factors including dose, route of administration and post-mortem interval should be considered in order to assess a drug’s propensity for redistribution [52]. In case #2, the higher C concentration might be due to a concentration-gradient driven post-mortem diffusion of the drug from the stomach.

In case #3, the presence of multiple boxes and packages of drugs purchased online represented a strong hint towards the consumption of NSOs, given that internet shops are a primary source for obtaining NPS [53]. Moreover, no label reporting the presence of MeACF was noted, so that the accidental consumption of mislabeled powder seems possible.

Although lungs weighed less compared to similar cases of opioids intoxication [54], including cases #1 and #2, brain edema, and froth, a proxy for pulmonary edema, were detected.

Toxicological analyses allowed to identify a cocktail of substances, particularly MeACF, furanylfentanyl, 4-ANPP, alprazolam and its metabolite, as well as traces of other benzodiazepines. According to the data so far available, MeACF and furanylfentanyl should have a pharmacological potency comparable to fentanyl [55, 56]. Particularly, past in vitro and in vivo studies have considered that MeACF is around 3 to 5 times less potent than fentanyl [54]. However, median blood levels in intoxication cases were relatively low: 15.1 ng/ml in a series of 11 deceased described by Fogarty et al. [21] and 34 ng/g among other 11 fatalities, with the highest reported concentration of 140 ng/g in a case of monointoxication [54]. Thus, the authors concluded that MeACF carries a high risk for accidental intoxication [54]. The levels here reported are well above the ones described by other authors [21, 54], suggesting that these are consistent with a severe CNS depressant effect. The note left on the computer might additionally suggest a massive consumption of drugs with suicidal intention. Moreover, an additive worsening of the respiratory depression can be assumed in the case of co-consumption of other potent opioids, as in the present case for furanylfentanyl, and of benzodiazepines [56]. Comatose-fatal effects of furanylfentanyl were described in 17 cases of death with blood concentrations in the range 0.2–2.7 ng/ml [36, 56]. On the basis of this data, the concentration of furanylfentanyl of our case appears in a range where strong side effects (sedation, hypotension, respiratory depression) can be anticipated without the existence of an acquired opioid tolerance.

4-ANPP could represent a synthesis by-product or a metabolite of furanyfentanyl. In both cases, the compound is thought to have some biological activity, but much lower than that of morphine [56] and did likely not play a major role in the present fatality.

Alprazolam levels were slightly above the therapeutical range described by Schulz et al. [36], although multiple factors, including metabolization and redistribution, might affect the reliability of postmortem blood levels. The compound is commonly encountered in fatal polydrug intoxications and could amplify the respiratory depression induced by co-consumed drugs [57].

Although the subject was used to try drugs purchased online, and might have developed a tolerance to opioids, given the co-consumption of highly potent NSOs with supra-therapeutical (alprazolam) and therapeutical (diazepam) levels of prescription benzodiazepines and the signs of brain edema, the death was explained as a consequence of a mixed intoxication, with MeACF (TSS = 3), furanylfentanyl and alprazolam playing a major role.

To the best of our knowledge, this is the first time that concentrations of MeACF in multiple tissues are reported, beside blood and brain levels [58]. In a previous case series, brain concentrations of MeACF were 3 times higher than those retrieved in blood (0.074 mg/kg vs 0.022 mg/kg) [58]. This was also seen in our case, with fivefold higher brain concentrations compared to blood. The high concentration in the brain further confirms the hypothesized mechanism of fatal CNS depression. Moreover, together with the lower levels in the liver and in urine, it is suggested that the compound was not extensively metabolized at the time of death and led to a rather acute death. Very high relative concentrations were estimated in the gastric content (Table 2) and this might be due to an oral ingestion or insufflation (followed by swallowing) of the compound. As already shown for case #2, MeACF levels were higher in central than in peripheral blood, although only by a factor of 1.7, and this might be explained by the route of intake.

The cases here presented had several factors in common. The deceased were healthy young men with a history of psychiatric diseases and/or past drug consumption, including a propensity for self-administering novel substances. Another rather constant finding is represented by drug powder residues (sometimes mislabeled) and paraphernalia at the death scene, consisting of a domestic environment. As expected, no specific post-mortem findings were discovered, although the necrosis of the globus pallidus might be related to the NSO consumption.

As emerged from the present series and from past literature [17, 18, 54], NSOs appear to be often consumed in a setting of polydrug use, especially in combination with other opioids (classical or NSOs) and benzodiazepines (including designer benzodiazepines). The co-consumption of multiple compounds further complicates the interpretation of the role of an NPS in the fatality, considering the possibility of synergistic effects. In our series, NSOs were always considered as a major contributor to the death (TSS = 3, i.e., “… likely to have contributed to toxicity/death, even in presence of other drugs” [25]), given the high potency described in the literature, high concentrations detected in blood, low levels of other drugs and the consistency of all data. In each postmortem case involving novel compounds and especially when the complexity is increased by polydrug consumption, a multidisciplinary evaluation of circumstantial, postmortem, histological and toxicological data is of paramount importance, and further information can be obtained, when enough material is available for analysis, by the distribution of drugs in tissues.

A limitation of the standard addition method, and of the present analysis, is represented by the fact that, considering only those data points with abscissa values equal to and greater than zero, nonlinearity might occur below the spiking level zero. However, blank addition could be combined with standard addition to overcome this limitation [59].

Although blood concentrations determined by LC–MS/MS analysis and by standard addition methods were approximately in the same range (220 ng/ml vs. 204 ng/ml for case #2, and 340 ng/ml vs 266 ng/ml for case #3), standard addition remains the most reliable quantification method that can overcome matrix effects [23] and can therefore be encouraged.