Abstract
Background:
Tobacco smoke is known to affect plasma levels of some drugs, including the antipsychotic clozapine. The effects of suddenly stopping smoking on patients who take clozapine can be severe, as plasma concentrations are expected to rapidly rise, potentially leading to toxicity. A ban on smoking at South London and the Maudsley NHS Foundation Trust (SLaM) was implemented in 2014, and this was expected to affect the plasma concentrations of clozapine for inpatients at the time. This study aimed to determine whether plasma concentrations of clozapine were affected, and additionally, in line with observations from other authors, whether levels of reported violence would also be affected.
Methods:
The smoking habits of all patients at SLaM who smoked and were prescribed clozapine were recorded both before and after the ban. The Glasgow Antipsychotic Side Effect Scale for Clozapine (GASS-C) scale was used to evaluate side-effect burden. Clozapine doses and plasma concentrations were also collected.
Results:
In total, 31 patients were included in this study. The mean clozapine dose before the ban was 502 mg/day, and this did not change significantly after the ban. Similarly, there were no significant changes in clozapine or norclozapine plasma concentrations, or in GASS-C scores. There was no change in the amount of tobacco patients reported smoking before or after the ban. A modest but statistically significant reduction in violent incidences was observed.
Conclusions:
Our data suggest that a ban on smoking for patients taking clozapine on open wards at inpatient hospital sites had little impact on clozapine plasma concentrations, because patients continued to smoke tobacco if allowed to leave. Smoking bans may result in a reduction in violent incidences.
Introduction
Legislation banning smoking in enclosed spaces in the UK was introduced in July 2007, with a 1-year exemption from this for psychiatric units [HM Government, 2006]. From 2008, inpatients on mental health units at South London and the Maudsley NHS Foundation Trust (SLaM) could only smoke outside the unit, but were able to still smoke within the hospital grounds. A further Trust-wide ban on smoking on Trust (hospital) grounds was introduced in October 2014, resulting in patients no longer being able to smoke anywhere on the hospital site.
Polycyclic aromatic hydrocarbons in tobacco smoke induce hepatic cytochrome P450 enzymes, increasing the rate of metabolism of some drugs [Taylor et al. 2015]. The antipsychotic medication clozapine is principally metabolized by the hepatic cytochrome enzyme CYP1A2, and the induction of this enzyme by tobacco smoke results in enhanced clozapine metabolism and consequently a reduction in plasma concentration. Stopping smoking has the opposite effect and is likely to increase plasma clozapine levels.
Previous studies at other inpatient units have found dramatic rises in clozapine plasma concentrations in patients treated at constant dose following smoking bans, leading in some cases to severe side effects [Cormac et al. 2010; Meyer, 2001]. In this prospective observational study we recorded clozapine plasma concentrations and prescribing practices for inpatients who smoked tobacco during the implementation of the smoking ban at SLaM. Responding to concerns that patients may experience potentially dangerous increases in plasma concentrations on stopping smoking [Cormac et al. 2010], and that nicotine withdrawal consequent on the enforcement of the smoking ban may lead to increased levels of violence [Greeman and McClellan, 1991; Ryabik et al. 1994; Hempel et al. 2002], the aim of this study was to determine the effects of a smoking ban on clozapine plasma concentrations and/or levels of violence.
Methods
Patients were identified for inclusion by ward pharmacists. Inclusion criteria were:
admitted to any ward across the Trust (incorporating three boroughs in South East London, and a variety of acute, forensic, rehabilitation and specialist wards);
current smoker;
prescribed clozapine.
Data collected included: clozapine dose, whether nicotine replacement treatment (NRT) was prescribed, number of cigarettes or ‘roll ups’ smoked daily, clozapine and N-desmethylclozapine (norclozapine) plasma concentrations, incidences of violence (as defined as ‘violence’ by senior nursing or medical staff on the ward), leave status (on-grounds or off-grounds, with a staff member escort or unescorted), caffeine intake and adverse effect burden. Demographic data, including ethnicity, were drawn from the Trust Electronic Patient Journey System (Table 1). The number of cigarettes smoked, caffeine intake and adverse effect burden were ascertained by patient interview using the Glasgow Antipsychotic Side Effect Scale for Clozapine (GASS-C) scale. Patient interviews were conducted either by the ward pharmacist or by the primary nurse. The GASS-C is a self-report side-effect scale that ranks antipsychotic adverse effects, including those that may be considered clozapine-specific, in terms of severity and frequency. A score of 0–12 denotes absent or mild adverse effects, 13–26 moderate adverse effects and over 26 severe adverse effects. Nursing and medical staff were asked to provide information on the number of violent incidents per week.
Demographic details of the sample.
Data were collected in the fortnight before the ban was implemented and again 1 week after the ban; clozapine doses were noted at 2 and 3 weeks after the ban. Clozapine plasma concentrations were taken in the 2 weeks preceding the ban and again in the week following the ban. Analysis was completed using IBM SPSS Statistics version 22. Pre-ban and post-ban data were compared using two tailed paired sample t tests. Approval for this study was granted by the Trust Drug and Therapeutics Committee. Ethical consideration was not required as treatment was not changed, and patient-level information accessed only by healthcare professionals in the course of their usual work.
Results
In total, 10 patients (32%) reported some reduction in smoking habit after the ban. An increase in pre-dose clozapine plasma concentrations of >0.1 mg/l after the ban was seen in three patients (10%). Only one of the patients who experienced an increase in clozapine plasma levels also reported a reduction in smoking habit (from three cigarettes per day to 1–2 per day). The dose was not changed for this patient, and there was no corresponding increase in norclozapine plasma concentration (Table 2).
Pre- and post-smoking ban data.
Two-tailed paired samples test.
Pre-ban versus post-ban.
GASS-C, Glasgow Antipsychotic Side Effect Scale for Clozapine; NRT, nicotine replacement therapy.
Discussion
This study of prescribing practices of clozapine before and after a hospital smoking ban included 31 patients, all of whom were smokers before the ban was implemented and who were prescribed clozapine for the duration of the study period. The mean clozapine dose was 502 mg/day before the ban, and this did not change significantly after the smoking ban. This mean dose is in keeping with previously reported average doses for smokers [Rostami-Hodjegan et al. 2004]. Clozapine and norclozapine plasma levels also did not differ significantly before and after the ban. The amount of tobacco smoked did not change statistically significantly, although there was a nonsignificant reduction after the ban. Caffeine intake, which is also thought to affect plasma levels of clozapine [Taylor et al. 2015], did not change.
The lack of significant change in clozapine doses or plasma levels is unsurprising given the apparent lack of change in patient’s smoking habits despite the smoking ban. The results suggest that prescribers did not change clozapine doses pre-emptively around the ban date, but may instead have waited for clozapine plasma concentration data, or patient-reported smoking habits to change, before considering dose reduction. The finding by other inpatient units of rises in plasma concentrations following smoking bans is likely to reflect the effectiveness of the ban in stopping patients smoking: secure units are probably more able to enforce a total smoking ban, and less likely to allow ‘off-grounds’ leave [Cormac et al. 2010].
We found no statistically significant difference in the GASS-C score before and after the ban. Some adverse effects associated with clozapine use are plasma-concentration dependent, and so this lack of change in side-effect frequency or severity is unsurprising given the corresponding lack of change in plasma concentrations throughout the audit.
Most of the patients included in this audit were allowed ‘off-grounds’ leave, and during this leave patients continued to smoke. The increase in the proportion of patients receiving this type of leave after the ban may reflect requests from patients to be able to smoke, but this may also be confounded by the expected improvement in patient’s mental health throughout the audit period allowing for increased leave privileges to be granted.
NRT was prescribed for 55% of patients before the smoking ban, rising to 68% of patients after the ban. The increase may indicate an intention by either patients or prescribers to reduce smoking rates after the ban, but almost a third of patients were not prescribed NRT after the ban was implemented. It is our experience that encouraging patients to use NRT is difficult, as some patients lack the motivation to give up smoking: the hospital smoking ban is an imposed ‘quit date’ that is often viewed negatively by patients, who refuse to engage in any type of smoking cessation conversation. This experience has also been reported by other groups [Lawn and Pols, 2005]. It is also apparent that despite an increase in NRT prescription, the rates of tobacco consumption did not change.
Reports of violent incidences decreased in the week after the smoking ban was introduced, contrary to our hypothesis. This may be due to variation in the, fairly low, baseline rates of these events or a reduction in the opportunity for disputes between staff and patients over smoking breaks and access to cigarettes. A reduction, or no increase in violent incidences around smoking bans, has also been observed in other inpatient psychiatric units [Lawn and Pols, 2005].
Limitations
This audit included 31 patients, the majority of whom were male. The generalizability of the results may therefore be limited to similar patient groups.
Although data were gathered regarding the prescribing of NRT, we did not record whether patients used the products, or indeed how effectively they used them. In our clinical experience, NRT products are often prescribed, but not used by patients.
We relied on patients self-reporting the amount of tobacco smoked and caffeine consumed. It is likely that patients were reluctant to report any covert smoking (i.e. smoking in banned areas) that may have occurred either before or after the ban was implemented. Patients may also have underreported their regular tobacco or caffeine intake.
We followed one cohort of patients through from the beginning to the end of the audit period. It is expected that the mental wellbeing of patients will improve over the course of an inpatient admission, and so a corresponding reduction in violent incidences might also be expected. It might also be hoped that side effects would be addressed in the course of an admission, resulting in a reduction in GASS-C scores. This progression may account for the differences we found over the audit period, although it is unlikely that this made a marked difference in one week.
The range for violent incidences was wide, with the majority of patients (25; 81%) neither being involved in violence before nor after the ban. We did not gather data on the reasons for the violence, so cannot firmly attribute the incidences to smoking-related reasons or otherwise.
Conclusion
Despite widespread staff anxiety about the potential impact of a smoking ban on clozapine plasma concentrations, we found no differences in clozapine or norclozapine plasma concentrations or clozapine doses after the smoking ban was implemented. We also did not find a significant drop in the amount of tobacco smoked by patients after the ban was introduced. These data suggest that a ban on smoking for inpatients receiving clozapine on hospital sites can be managed safely, but also that the introduction of a ban has little effect on the smoking habits of patients, at least in the short term. Smoking bans on hospital sites do not affect clozapine plasma concentrations, because patients continue to smoke if allowed to leave. Smoking bans on hospital inpatient sites do not seems to be associated with increases in violence, and may indeed result in a reduction in violent incidences.
Recommendations for practice
Smoking cessation or reduction is expected to increase clozapine plasma concentrations by on average 50%, potentially causing toxicity. Conversely, starting smoking will reduce clozapine plasma concentrations and may contribute to therapeutic failure.
We recommend baseline plasma concentrations before and weekly after a change in smoking habit, with doses adjusted accordingly until a plasma concentration in the therapeutic range is achieved.
NRT should be offered to all smokers who are admitted to hospital, but it should be acknowledged that this is unlikely to lead to complete smoking cessation. Patients who wish to quit smoking must be supported to do so, but those who do not wish to quit should have their withdrawal symptoms managed effectively and in a timely manner.
Consideration should be given to predicted smoking status after discharge. Clozapine plasma concentrations should be expected to drop if smoking is resumed on discharge (resumption of smoking on discharge has been demonstrated by other authors, including Jonas and Eagle [Jonas and Eagle, 1991], for whom 80% of patients resumed smoking immediately upon discharge), and this has implications for relapse risk.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Professor Shergill was supported by a European Research Council Consolidator Award, and by the National Institute for Health Research (NIHR) Mental Health Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London. Dr MacCabe is supported by the NIHR Biomedical Research Centre for Mental Health at the South London and Maudsley NHS Foundation Trust and King’s College London.
Conflict of interest statement
Ms Gee has received consultancy fees from Sunovion, and lecturing honoraria from Sunovion, Otsuka and Janssen. Professor Taylor has received research funding from Janssen, Lundbeck and Servier; consultancy fees from Sunovion, and lecturing honoraria from Janssen, Otsuka and Servier.