Concerns also relate to three further factors

First, the use of promethazine interferes directly with employing medium-long-term psychological and behavioural strategies that, unlike promethazine, have been proven to improve sleep quality. CBTi is effective among people with co-occurring mental state disorders,^{Reference Hertenstein, Trinca, Wunderlin, Schneider, Züst and Fehér6} and has continued benefits demonstrated for as many as 10 years later.^{Reference Jernelöv, Blom, Hentati Isacsson, Bjurner, Rosén and Kraepelien7} CBTi is a short-term psychological treatment that teaches people how to regulate their sleep–wake behaviour, strengthen the association of feeling sleepy and being in bed, and provides sleep psychoeducation and techniques to manage nocturnal worrying and anxious body states, like imagery work, breathing techniques and progressive muscular relaxation. However, with a long half-life, the duration of effect of promethazine at 10–19 h causes a hangover effect and is known to impede people being able to stick to a consistent rising time (a hugely important factor in regulating one's sleep pattern). In addition, use of sedative medication unhelpfully reinforces the idea that to be asleep is to be ‘knocked out’, anaesthetised, turned off like a computer – that is to say, impervious to external stimuli. During sleep, there is some maintenance of external awareness (for evolutionary survival reasons), yet people who are not ‘confident sleepers’ monitor their body and environment vigilantly, reinforcing a cycle of worries, anxious bodily feelings and unhelpful sleep–wake behaviours.

Second, anecdotally, colleagues suggest their preference for promethazine as a sleep aid over benzodiazepine receptor agonist medications (like Z-drugs) because it is ‘less addictive’. There are current active debates in psychiatry around what it is to be ‘addictive’ (see debates on the ‘addictive’ nature of selective serotonin reuptake inhibitors, for example), but many of the features of substance dependence are true for promethazine as they are for sedative hypnotics. Promethazine tolerance develops quickly, and there are well-documented withdrawal effects from regular use. Some long-term, off-label users of promethazine report not being able to stop their medication particularly because of rebound insomnia. Long-term use of promethazine as a sleep aid is anecdotally common, but national figures are not available, and neither are deprescribing guidelines for antihistamine agents. Importantly, there is plenty of recreational use of promethazine. Promethazine has a likeability for many users, and has a ‘street value’. It is also commonly used in combination with codeine, so nicknamed ‘lean’ or ‘purple drank’, because of promethazine's known effects in potentiating an opioid high, and research also evidences people self-medicating with promethazine to cope with undertreated mental health symptoms.^{Reference Ware8} Altogether, promethazine should be seen to be at least partially dependence-forming, and in the same ways as benzodiazepine receptor agonists.

Third, the side-effect profile of promethazine is particularly concerning for people using psychiatric services. These include QT prolongation, weight gain, nightmares, delirium and perhaps most importantly, significant anticholinergic side-effects, which also increase the risk of dementia.^{Reference Gray, Anderson, Dublin, Hanlon, Hubbard and Walker9} Herein, many other common medications prescribed for patients using psychiatric services, notably antipsychotics and other sedative psychotropics, have similar side-effect profiles that are compounded by promethazine. The effect on seizure threshold has shown mixed results in animal models. Patients should be told not to drive when using promethazine, and to be abstinent of other sedative drugs like alcohol and cannabis. Self-poisoning with promethazine in combination with other psychotropics appears to substantially increase fatality.^{Reference Fusaroli, Pelletti, Giunchi, Pugliese, Bartolucci and Necibi10,Reference Oyekan, Gorton and Copeland11}

As per NICE guidelines, CBTi is the first-line treatment for people with sleeping difficulties who are using psychiatric services, this can be delivered individually, in groups or online where facilitated CBTi is not available or is preferred by the patient. Mobile phone app-based, self-directed CBTi programmes are recommended by NICE, and are available with the National Health Service in some localities, although the evidence of its benefit for people with severe psychiatric co-morbidity is less robust. There should be screening for primary sleep disorders like sleep–wake phase disorders, restless leg syndrome and respiratory sleep disorders; reconciliation of existing psychiatric medication that might contribute to sleep difficulties; treatment of psychiatric causes of insomnia (substance misuse, psychosis, etc.) and referral to local specialist sleep services where needed. If medication is indicated because of limited response to CBTi, then it should be short term (<7 days). The evidence suggests that benzodiazepine receptor agonists offer a more effective intervention, with a side-effect profile of lesser concern, and do not produce a comparable hangover effect that impedes the employment of cognitive–behavioural techniques.^{Reference De Crescenzo, D'Alò, Ostinelli, Ciabattini, Di Franco and Watanabe12} Newer orexin antagonists are emergingly prescribed: daridorexant is now recently licensed in UK for patients who have not adequately responded to CBTi. Melatonin is licenced both as a hypnotic for insomnia in adults over 55 years old and for people with an intellectual disability, as well as treatment of certain sleep–wake disorders. Patients taking sedative hypnotics in the long term should be supported to gradually reduce their medication, under the supervision of specialist sleep services in some cases.

In short, we must reflect on our apparent increasing use of this 1950s medication, and hold it to the same standards that we would any new medication. When examined, we find that there is no good evidence for its use as a sleep aid, signs of active detriment and much better alternatives. Improving provision of CBTi for people using psychiatric services ought be a response to the increasing number of prescriptions. Promethazine likely has value as a short-term sedative agent for people in acute psychiatric crisis, but this should not be conflated with improving sleep quality.