The main findings of our study were that patient history inaccurately describes drug ingestions: benzodiazepines/Z-drugs, Pregabalin, and THC were especially underreported. Furthermore, benzodiazepines/Z-drugs and Pregabalin are increasingly abused. NPS legislation in Germany effectively decreased cathinone/phenethylamine and SCRA overdoses. Agitation and aggression, anxiety, hallucinations, and psychosis as well as coma were frequent symptoms. Sedation, antidote therapy with Naloxone and/or Flumazenil, or intubation were often required. Opiates and opioids were the main cause of morbidity and mortality.

Our findings show that there is a significant difference between patient history and analytical findings. Patients reported a median of one (IQR 1–2) ingested substance classes, analytical testing revealed a median of three (IQR 2–4) substance classes. In total, 22.1% of patients tested positive for one substance class. In 2.2% of patients, no substance could be identified through toxicological analysis. In two patients, eight substance classes could be detected. There may be numerous reasons for this finding. Incomplete or inaccurate patient histories might have played a role and are a common problem in patients with drug ingestion [15,16]. Patients might only report the main substance responsible for their overdose and leave out co-ingested drugs like be benzodiazepines/Z-drugs, Pregabalin, and THC which they might deem irrelevant and therefore have a low NPV. Patients might also have been unaware of consuming additional drugs that were detected in their system. Most established drugs like amphetamines/MDMA, benzodiazepines/Z-drugs, Buprenorphine, cocaine, Methadone, opiates, Pregabalin, and THC have a high PPV. For these substance classes, we can assume that patients knew what they ingested, and this was subsequently confirmed by analytical testing. Newer substance classes like cathinones/phenethylamines and SCRA as well as Ketamine and LSD had a low PPV. For these substances, patients themselves were often unaware of what exactly they had ingested. In particular, new cathinones/phenethylamines and SCRA posed a great challenge for analytical testing because of their sometimes unknown structure and their absence in available mass spectral libraries and databases. Therefore, there is a delay between the appearance of these substances on the market and possible analytical testing [17]. NPSs are constantly evolving as alternatives to classic drugs—containing more effective substances with higher receptor affinity—leading to health care risks and posing a challenge for medical service analytical testing and legislation [5,18]. These findings are in line with Liakoni et al., who described a high analytical confirmation for heroin and cocaine and an advantage for mass spectrometry in detecting NPSs [19]. Bharat et al. describes an even higher association between self-reporting and biological testing; however, this meta-analysis was not performed in a clinical setting. It has also been shown that sensitivity estimates differ between research studies with no consequences and studies conducted in criminal justice settings [20]. Our study focused on patient reports submitted at the time of admission which is what clinicians are confronted with. Reports might have been more accurate if patients were interviewed after their acute intoxication had passed. Overall, focusing on the patients’ history alone may only represent a relatively small spectrum of reality. This is the methodology used by the Euro-DEN-Plus study group [21]. To make the study results even more robust, analytical confirmation would be desirable and could improve the validity of the results. Using our study data and the PPV and NPV of certain substance classes, we can, e.g., assume that studies like the one by Galicia et al., describing the clinical relevance of ethanol coingestion in patients with GBL/GHB intoxication, have a high validity (GBL/GHB PPV 94.1%, NPV 98.1%) [9]. This study might help with the interpretation of further Euro-DEN-Plus publications.

Urine analytics showed a clear predominance of sedatives with opiates/opioids in 57.2% of cases, followed by benzodiazepines (55.6%), Pregabalin (44.4%), THC (37.7%), and GBL/GHB (3.7%), leading to typical symptoms like coma and CNS depression. While ethanol was only co-ingested in 41.4% of cases, it was shown that ethanol co-ingestion significantly increased the risk of severe outcomes in CNS-depressing substances [10]. Sympathomimetic substances like amphetamines/MDMA (21.3%), cathinones/phenethylamines (8.0%), and cocaine (16.7%) were associated with agitation and aggression. This was also true for and SCRA (6.1%). This underlines the findings of several studies like Crulli et al., Noseda et al., Romanek et al., and Schmoll et al., showing violent thoughts and acts in one-third of users of synthetic cathinones as well as prolonged confusion as the most common complications [6,7,22,23]. Vallersnes et al. showed that psychosis was significantly associated with amphetamine, methylenedioxypyrovalerone, SCRA, tryptamine, and LSD toxicity [24]. SCRA consumption led to significantly increased rates of vomiting, seizures, and arrhythmias, also described by Waters et al. showing significantly higher rates of drowsiness, coma, agitation, seizures, and bradycardia for intoxication with lone synthetic cannabinoid receptor agonist exposures compared with cannabis use [25]. Hallucinations and psychosis associated with the use of amphetamines/MDMA and cathinones/phenethylamines as well as psychosis after cocaine ingestion were significantly reduced in polyvalent drug misuse. This was also especially true for patients who tested positive for benzodiazepines/Z-drugs, Pregabalin, and opiate/opioids—assuming an antagonistic effect of these drugs and corresponding with the findings of Heier et al. [10]. Overall, our study confirms the findings of previous studies where patient records and not urine analysis were used as a reference.

Only 22.1% of patients tested positive for one substance—predominantly amphetamines/MDMA (35 patients) and THC (43 patients)—confirming that polyvalent drug abuse predominates in Europe [1]. Polydrug use is a main risk factor for ICU requirement in recreational drug toxicity as shown by Noseda et al. [26]. In addition to trends like polyvalent substance use, the European Drug Report 2014 stresses the risks of highly potent substances, e.g., nitazene, and an overall rapidly growing market. According to the reported European drug seizures, the amount of synthetic cathinones and cannabinoids is still high, but synthetic opioids and benzodiazepines are increasing [1]. Nevertheless, cathinone/phenethylamine and SCRA decreased significantly in our center after Germany passed NPS legislation in November 2016 [27]. This law added NPS-like cathinones, phenethylamines, and SCRA, among others, to the list of controlled substances and seemed to have had an effect in our study population.

In suspected opiate/opioid overdose, application of Naloxone—the antagonist of opiates and opioids for all opioid receptors—can be considered. In benzodiazepine/z-substance toxicity, Flumazenil—the antagonist of benzodiazepines at the GABA_A-receptor—can be applied. Naloxone was applied more frequently (10.2%) compared to Flumazenil (5.4%). This is in line with current recommendations, reinforcing that Flumazenil should not be applied routinely due to severe adverse effects like triggering withdrawal syndrome or seizures [28]. The clinical identification of opiate/opioid or benzodiazepine/z-substance toxicity and the subsequent application of Naloxone or Flumazenil was correct in 87.6% and 83% of cases. Naloxone and Flumazenil together were applied in 23 cases for unknown coma. EMSs and clinicians apparently had a good understanding of drug toxicity or used Naloxone as diagnostic tool. Because of the novel highly potent synthetic opioids and benzodiazepines, it is unknown if higher doses of Naloxone or Flumazenil are needed to show a clinical effect [29]. Unfortunately, Naloxon dose was not investigated in our study. However, applying doses higher than 0.4 mg is uncommon in our department. In cases where antagonization with significantly more than 0.4 mg naloxone is required, intubation, sedation, and ventilation may be carried out at the discretion of the treating physician.

While self-discharge rates vary between countries and hospitals, the average among Euro-DEN emergency rooms was reported at 1/8th [30]. Our high rate of 68.7% might be explained by the fact that all patients were offered to stay for detoxification. By refusing detoxification and requesting discharge after the acute intoxication had been treated, discharge was considered “against medical advice”.

Although our findings demonstrate the importance of analytical drug screening in recreational drug poisonings, these methods are rarely routinely available—mainly due to high costs (e.g., equipment, solvents, specialized personnel) and long analytical run length with delayed results often arriving after primary care has been completed or the patient has even already left the hospital due to short stays (median treatment time 19 h) [31] To avoid missing substances and to maximize the number of tested substances, several analytical methods should be combined. The hospital management of intoxicated patients should not be delayed because of analytical testing.