Psychosis was a clinical feature in 6.3 % of this European case series of 5529 presentations to an ED with acute recreational drug toxicity. The association with psychosis varied considerably between drugs. Not surprisingly, the associations were strongest for hallucinogens and some central stimulants. Overall, cannabis, amphetamine and cocaine were the drugs most commonly involved in presentations with psychosis. Amphetamine was the drug most frequently involved in single agent presentations, with 32.4 % single agent amphetamine cases presenting with psychosis. Amongst the NPS, psychosis was rare in presentations involving mephedrone and methedrone, but occurred more frequently with tryptamines, MDPV, methylphenidate and the SCRAs.

Among the patients presenting with recreational drug toxicity and psychosis, 63.2 % were agitated and/or aggressive (compared to 24.1 % of those without psychosis), 43.7 % had hallucinations (compared to 4.5 %), and 32.8 % were admitted to a psychiatric ward (compared to 3.3 %), indicating that this is a group of patients with substantial resource implications. Furthermore, acute psychosis induced by substance use is, until it has resolved, difficult to distinguish clinically from schizophrenic psychosis [31, 32]. Some authors have questioned whether they are separate entities at all [2, 33–35]. Longer duration of untreated psychosis is associated with poorer outcome, though a causal relationship has not been established [36]. However, there is evidence suggesting that early intervention, reducing the duration of untreated psychosis, could be helpful [37–39].

Cannabis, amphetamine and cocaine were the drugs most commonly involved in presentations with psychosis in our study. This is to be expected as they are commonly used in Europe [8] and carry the risk of inducing acute psychosis [2, 3, 40]. We found amphetamine more frequently associated with psychosis than cannabis and cocaine, both as a single agent and in the logistic regression analysis, confirming that psychosis is a major feature of acute amphetamine toxicity. Cannabis, but not cocaine, was significantly associated with psychosis in the logistic regression analysis.

Among the NPS, tryptamines, MDPV and SCRAs were the drugs most frequently involved in presentations with psychosis. In our study, 57.1 % of patients with tryptamine poisoning presented with psychosis, a larger proportion than the 23.6 % reported among 55 enquiries concerning AMT in a poison centre case series in the UK [27]. This substantiates previous reports noting psychosis as a prominent clinical feature of acute tryptamine toxicity [41]. In a study of 23 patients with acute MDPV toxicity seen by medical toxicologists in the US, psychosis was reported in 8.7 % [28]. This is less than the 27.3 % in our study, and our findings are more in keeping with a US poison centre study of 236 cases of MDPV toxicity, where 36 % had paranoia, 34 % confusion and 40 % hallucinations [18]. In our study, 15.4 % of those presenting with acute SCRA toxicity had psychosis – this is comparable to the 9-11 % reporting hallucinations and delusion to US poison centres in two case series of 1353 and 418 acute SCRA exposures [23, 24]. However, a German case series of 29 ED presentations with laboratory confirmed acute SCRA toxicity reported only one case of acute psychosis though 38 % of the patients had hallucinations or changes of perception [42]. For mephedrone, 5.7 % of patients in our study presented with psychosis, slightly less than the 8 % reported among 488 enquiries to UK poison centres [27] and the 9 % with paranoia and hallucinations among 57 presentations to the ED in a Scottish case series [43]. No psychotic features were reported among seven cases with laboratory confirmed mephedrone toxicity in a London ED [44].

Psychosis has previously been reported as a feature of acute methylphenidate toxicity as well as a potential side effect of methylphenidate treatment [45]. In our study, 23.1 % of presentations involving methylphenidate had psychosis.

Tramadol and fentanyl differed from the other opioids, with a larger proportion of presentations with psychosis. Though only seven cases with psychosis were seen involving these drugs, two were single agent cases with fentanyl and one with tramadol. We have only found one previous report of psychosis as a clinical feature of tramadol toxicity [46], and none involving fentanyl.

In addition to the risk of inducing acute psychosis, regular use of several recreational drugs, especially amphetamine, methamphetamine and cannabis, has been shown to be associated with later development of chronic psychosis or schizophrenia [2, 11, 34, 47]. Japanese studies have reported schizophrenia-like psychosis in methamphetamine users persisting for months to years after discontinuation of methamphetamine use [11, 33]. A large cohort study in California found an increased risk of later being diagnosed with schizophrenia in previously non-psychotic hospitalised users of methamphetamine or cannabis [34]. A meta-analysis of seven studies found cannabis use to be an independent risk factor for later development of chronic psychosis [47]. The risk was higher the more cannabis was used, and the earlier in life the start of the exposure [47]. It also seems that patients with previous substance use associated psychosis are at greater risk of subsequent psychotic episodes [2, 33].