Smoking and drug interactions
- Catherine Lucas, Jennifer Martin
- Aust Prescr 2013;36:102-4
- 3 June 2013
- DOI: 10.18773/austprescr.2013.037
When patients enter hospital they may have to stop smoking abruptly if the hospital has a ‘no smoking’ policy. Abrupt smoking cessation can affect the metabolism of drugs.
Cigarette smoking induces the activity of human cytochromes P450 (CYP) 1A2 and 2B6. These enzymes metabolise several clinically important drugs, including clozapine, olanzapine and methadone.
Decreased CYP1A2 activity after smoking cessation increases the risk of adverse drug reactions, with reports of increased toxicity from clozapine and olanzapine. Predicting the required dose reduction of drugs metabolised by CYP1A2 after smoking cessation is challenging. Therapeutic drug monitoring should be used when possible.
Nicotine replacement therapy does not influence CYP1A2 activity.
Despite anti-smoking campaigns, there are about 1.3 billion cigarette smokers worldwide and this number is still increasing.1 Both smoking prevalence and daily tobacco consumption are very high in patients with psychiatric disorders.2 Over two-thirds of Australians with psychosis smoke cigarettes, compared with one quarter of the general population.3
Smoking restriction or cessation is now commonly imposed on patients by 'no smoking' policies in Australian hospitals. Stopping smoking is particularly important in patients with mental health problems as they often have an adverse cardiovascular risk profile (from factors including obesity, dyslipidaemia, insulin resistance and hypertension). This is associated with high rates of premature mortality from cardiovascular disease.4
Even with nicotine replacement therapy as a cessation aid, patients with psychiatric illness appear to have more difficulty maintaining long-term abstinence than other smokers.5 Most resume smoking within five weeks of hospital discharge.6 These patients should be regularly asked about their smoking status, as an additional risk of resuming or abruptly ceasing cigarette smoking has the potential for drug interactions.7
The chemicals in smoke may interact with antipsychotics, antidepressants, benzodiazepines,8 oral contraceptives, inhaled corticosteroids and beta blockers via pharmacokinetic and pharmacodynamic (often nicotine-mediated) mechanisms.9
Cigarette smoking induces the activity of cytochrome P450 (CYP) 1A2 (via chemicals in cigarette smoke such as polycyclic aromatic hydrocarbons)10 and also CYP2B6.11 These enzymes metabolise several clinically important drugs (such as antidepressants and antipsychotics) (Box) and a number of procarcinogens (such as those in cigarettes).10,12
The effect of smoking on hepatic enzymes is not related to the nicotine component of tobacco. Nicotine replacement therapy does not influence CYP1A2 activity.13
Genetic polymorphisms of the CYP1A2 gene contribute to extensive inter-individual variability in drug metabolism14,15 and are associated with altered inducibility of gene expression in smokers.7,16 There are also marked ethnic differences in the distribution of CYP1A2 mutations,10,17 meaning that different ethnic groups respond differently when the patient stops smoking.
CYP1A2 activity is significantly higher in heavy smokers (more than 20 cigarettes/day) than in nonsmokers.18 This is likely to be clinically relevant for some drugs which have a narrow therapeutic index and are metabolised by CYP1A2, such as clozapine. The induction varies depending on the bioavailability of the components of cigarette smoke and the extent of inhalation.8 It is not known how the number of cigarettes smoked daily or inter-individual variation affects CYP1A2 induction,9 but heavier smokers appear to have a greater increase in the clearance of drugs.19
This enzyme induction is rapidly reversed when patients abruptly stop smoking, with a new steady state of CYP1A2 activity reached after approximately one week.20 This reduction in enzyme activity reduces clearance and increases the risk of adverse drug reactions for patients taking drugs metabolised by CYP1A2.20,21 These patients should be regularly asked about their smoking and the extent of their cigarette consumption.16
Cigarette smoking induces the metabolism of clozapine and olanzapine,22 resulting in lower plasma concentrations.7,16,21 The daily consumption of 7–12 cigarettes is probably sufficient to cause the maximum induction of clozapine and olanzapine metabolism.22 A 50% difference in the mean daily dose of clozapine needed by smokers and nonsmokers to reach a given blood concentration has been reported.23
Irrespective of smoking status, the mean oral bioavailability of clozapine is 27–47% and clozapine plasma concentrations have more than a 45-fold variability amongst individuals during chronic treatment.24 There are also large inter-patient differences in olanzapine exposure, with gender and genetic factors contributing.25
Non-smokers are at higher risk of adverse effects if treated with standard doses, suggesting that there is an interaction between smoking, olanzapine and clozapine.20,22 In one case report, a patient receiving olanzapine experienced extrapyramidal symptoms (including akathisia, akinesia and bradyphrenia) within days of significantly reducing tobacco consumption.21 Case reports on smoking discontinuation by patients taking clozapine outline effects including confusion,21 tonic–clonic seizures, stupor, coma26 or aspiration pneumonia.27
Clearance of clozapine has been shown to decrease when smoking is ceased, with a mean increase of 72% in plasma clozapine concentrations.27 It is suggested that daily dose reductions (of approximately 10% until the fourth day after smoking cessation) should be made whenever patients cease smoking during treatment with clozapine.20 Patients who resume smoking after leaving hospital may need their drugs and doses reviewed to account for this change.28 Therapeutic drug monitoring of clozapine is useful.
As fluvoxamine is metabolised by CYP1A2, smokers might require higher doses than those recommended from clinical trial data.8 Smoking is not anticipated to alter the pharmacokinetics of other selective serotonin reuptake inhibitors as they are not substrates of CYP450 isoenzymes induced by smoking.8
Smokers might require higher than normal doses of the tricyclic antidepressant imipramine. They do not appear to require dose adjustments of amitriptyline, nortriptyline or clomipramine.8
Warfarin's less active R isomer is eliminated to a minor extent by CYP1A2.29 Smoking may therefore potentially interact with warfarin by increasing its clearance and reducing its effect. A recent meta-analysis showed that smoking appeared to increase the warfarin dose requirement by 12%, resulting in an extra 2.26 mg per week compared with nonsmoking.30 Consequently, INR should be closely monitored when there is a change in patients' smoking status.
CYP isoenzymes (including CYP2C19, 3A4/5, 1A2, 2B6 and 2C9) convert clopidogrel and prasugrel into their active metabolites, which bind irreversibly to the receptors on platelets. As smoking is known to enhance CYP1A2 activity, theoretically it could increase the antiplatelet efficacy of these thienopyridine drugs.31
An enhanced response to clopidogrel has been seen in smokers who are CYP1A2 (163CA) A-allele carriers.32 Two retrospective analyses of large randomised clinical trials of clopidogrel showed that clopidogrel might be more effective in active smokers.33,34 However, a systematic review concluded that smoking is not associated with reduced platelet reactivity in patients on clopidogrel.31 Genetic polymorphisms seem not to impact on the activity of prasugrel.31
Caffeine is highly dependent on CYP1A2 for its metabolism. Smokers require up to four times as much caffeine as non-smokers to achieve the same plasma caffeine concentration. Caffeine can increase the concentration of clozapine and olanzapine.35
Pharmacodynamic drug interactions with tobacco smoke are largely due to nicotine.9
The vast majority of patients using methadone maintenance therapy also smoke tobacco.36 Methadone doses have been found to be higher in heavy smokers37 and methadone has been shown to increase both smoking rates and smoking satisfaction.38 Patients report less methadone-induced sedation when they smoke around the time of their methadone dose.39
Methadone attenuates nicotine withdrawal. Reducing methadone doses when the patient is trying to stop smoking could be detrimental.40
Substrates of some cytochrome P450 enzymes induced by smoking
amitriptyline, caffeine, clozapine, duloxetine, fluvoxamine, haloperidol, imipramine, olanzapine, ondansetron, paracetamol, propranolol, theophylline, warfarin (R-isomer)
bupropion, clopidogrel, cyclophosphamide, efavirenz, ifosfamide, methadone, nevirapine
Nicotine activates the central nervous system9 and this may explain the attenuated sedation observed in smokers compared to non-smokers taking benzodiazepines.41 Prescribers should be aware that when patients taking benzodiazepines stop smoking, there is a risk of central nervous system depression.
Smoking increases the adverse effects of the combined oral contraceptive pill (specifically thromboembolism, ischaemic stroke and myocardial infarction). The combined oral contraceptive pill is contraindicated in women aged 35 years or older who smoke 15 or more cigarettes a day.9 For smokers who use combined low-dose oral contraceptives, the attributable risk of death from cardiovascular disease is 19.4 per 100 000 women aged 35–44 years (vs 3.03 per 100 000 for non-smoking women of the same age).42 This risk is also presumed to be associated with other contraceptives containing oestrogen.9
Limited data suggest no convincing association between cardiovascular disease and progestogen-only pill use.43 If smoking cessation is unsuccessful, non-hormonal or progestogen-only contraceptives are preferred from a cardiovascular perspective.9
Proposed mechanisms of corticosteroid insensitivity include suppression of histone deacetylase expression and activity by cigarette smoking, causing inflammatory gene expression and a reduction in glucocorticoid function.45 Clearance of corticosteroids from the lungs may be altered by increased mucus secretion or airway permeability.46
Smokers may require higher doses of beta blockers. Although propranolol is a CYP1A2 substrate (Box), nicotine-mediated central nervous system activation may diminish the effect of beta blockers on blood pressure and heart rate.9
Drugs used to aid smoking cessation are not without their hazards, particularly in patients with psychiatric disorders.
Bupropion, a selective catecholamine reuptake inhibitor, is associated with a dose-related risk of seizures. Predisposing risk factors include concomitant administration of antipsychotics, antidepressants, excessive alcohol or those sedatives which lower the seizure threshold. Psychiatric symptoms, in particular psychosis or mania, have been observed, mainly in patients with a history of psychiatric illness, particularly bipolar disorder.
Bupropion is metabolised by CYP2B6 and inhibits the CYP2D6 pathway. Drugs predominantly metabolised by 2D6 (including metoprolol, many antidepressants and antipsychotics) should be started at the lower end of the dose range if bupropion is used. Co-administration of drugs known to induce metabolism (for example, carbamazepine and phenytoin) or inhibit metabolism (for example, valproate) may affect the activity of bupropion.
Nortriptyline, a tricyclic antidepressant shown to aid smoking cessation, also interacts with other drugs metabolised by CYP2D6.
Varenicline, a partial agonist at neuronal nicotinic acetylcholine receptors, has no known clinically significant drug interactions. However, using nicotine replacement therapy while taking varenicline can exacerbate adverse effects such as nausea and headache. As with bupropion, serious neuropsychiatric symptoms have been reported (although a causal association has not been established).
Cigarette smoking can affect drug metabolism via pharmacokinetic and pharmacodynamic mechanisms, and a change in smoking status can render patients at risk of serious adverse reactions.
Patients should be regularly monitored with regard to their smoking status and extent of cigarette consumption and doses of relevant medications adjusted accordingly.
No relevant potential conflict of interest declared
Acknowledgement: Dr Peter Donovan and Nicky Foxlee (Manager, Herston Health Sciences Library and The University of Queensland Dentistry Library)
Clinical pharmacology registrar, Royal Brisbane and Women’s Hospital, Brisbane
Clinical pharmacologist, Royal Brisbane and Women’s Hospital and Princess Alexandra Hospital Brisbane
Consultant physician, Internal Medicine, Princess Alexandra Hospital Brisbane
Associate Professor of Medicine, The University of Queensland