Olanzapine and serotonin toxicity

Hasslet and Kumar describe an interesting case of a patient developing significant adverse effects from a combination of psychotropic medications.1 Unfortunately they provided a selective review of the literature pertaining to the pathogenesis of serotonin syndrome (SS) and have attributed this case to serotonin toxicity. We do not believe this is the case and do not support their pharmacological explanation of the mechanism. The authors are quick to dismiss the possibility that the effects are a result of withdrawal from a selective serotonin reuptake inhibitor (SSRI). They suggest that the period of 4 days is insufficient time for a drug with a half-life of 33 h to cause a withdrawal syndrome. This is not supported by a large case series of withdrawal from SSRIs where the median onset time was 2.1 days in 160 reports and symptoms were noted in 86% of patients within 4 days.2 In this series of Price et al. most cases were a result of paroxetine withdrawal, which has a similar half-life to citalopram. Thus, it is quite possible, and indeed likely, that the features described by Haslett and Kumar are due to a SSRI withdrawal syndrome. These features are consistent with those described in other cases of SSRI withdrawal.2 Unfortunately the authors did not measure citalopram concentrations to confirm this. Serotonin syndrome is not a discrete syndrome, it is a spectrum of toxicity as discussed by Haslett and Kumar.1 This means that serotonin toxicity will be more common and more severe with increasing amounts of serotonergic agents and be most common with overdose of these agents. This is supported by the majority of reported cases of serotonin syndrome occurring with overdoses of SSRIs.3 Thus, if olanzapine does cause serotonin syndrome, it should be commonly reported with olanzapine overdoses. In our unit, there have been 102 olanzapine overdoses in the last 4.5 years. Of these 102 cases, there have been only five of serotonin syndrome as defined by Sternbach's criteria4 and in all five cases the patient coingested large amounts of an SSRI or venlafaxine. There were 16 overdoses where olanzapine was the only drug ingested in overdose, and there were no cases of SS in this group. This means it is highly unlikely that olanzapine causes SS. Based on the known pharmacology of olanzapine and our current understanding of the pathogenesis of SS, it is unlikely that olanzapine would cause SS. The authors briefly refer to recent animal work that supports the principle role of the 5HT2A receptor in serotonin syndrome5 and then in contrast to that evidence state that it is the 5HT1A receptor that is involved.1 There is good evidence in both animal models,5, 6 and clinical cases that support a major role of 5HT2A receptors in the some of the neuromuscular features and hyperthermia of SS. Nisijima et al. clearly demonstrated that 5HT1A antagonists did not prevent death and only modestly reduced hyperthermia in animal models, but both non-specific 5HT2 and specific 5HT2A antagonists did block these effects.6 Similarly, there are increasing numbers of cases where the 5HT2 antagonists cyproheptadine and chlorpromazine are effectively used to treat SS.7-9 Haslett and Kumar have confused the behavioral serotonin syndrome of rodents administered serotonergic agents, mediated by 5HT1A receptors10 with serotonin toxicity seen in humans. There is no evidence to support the assertion that these conditions are exactly alike in humans and rodents. The pharmacological mechanism suggested by Haslett and Kumar and others is based on the incorrect assertion that SS is mainly a result of the effect of increased levels of central nervous system 5HT on the 5HT1A receptors. If this was the case, agents that block 5HT2 and 5HT3 receptors, but are 5HT1A agonists, such as nefazodone and mirtazapine, should have the highest incidence of SS. This is not the case, with few reports of cases by these agents, and these being misattributed to mirtazapine or nefazadone.11 They further argue that 5HT2 blockade leads to increased serotonin (there is no pharmacological basis for this assertion) and that 5HT2 blockade also results in 5HT1 stimulation. However, the paper quoted to support this actually shows that 5HT2 blockade facilitates 5HT1 mediated dopamine release — a different phenomenon.12 Given their known pharmacology, atypical antipsychotics cannot be implicated in serotonin syndrome based on a few case reports;1, 13 none of which demonstrate definite causality and with no cases where an atypical antipsychotic alone has been implicated. Haslett and Kumar have suggested that their case report and one other are sufficient to refute well controlled animal studies of the effects of the atypical antipsychotics.1 They have suggested that this is due to interspecies differences and have ignored that the role of 5HT1A receptors in serotonin syndrome is based on rodent models only and that there is considerable human literature on the effectiveness of 5HT2 receptor antagonists in the treatment of serotonin syndrome.7-9 Until large case series or controlled studies are done, atypical antipsychotics should continue to be regarded as unlikely to cause SS, and in some cases may be effective in the treatment of it.5