Tobacco smoking and its potential drug interactions

Tobacco smoking continues to be the leading cause of chronic illness and preventable death worldwide^[1]
. Around a fifth of all deaths in both the UK and the United States are estimated to be related to smoking, with morbidity and mortality related to respiratory disease, cardiovascular disease and various types of cancer^[1],[2],[3]
. While smoking rates in some nations are on the decline, overall they remain high worldwide. Average adult male smoking rates in the UK, United States and Australia are around 15–17% compared with 50–60% in Russia and China^[4]
. There is also a worrying increase in tobacco consumption in low human development index countries, such as those in sub-Saharan Africa and India^[4]
.

From a treatment perspective, smoking behaviour has a pharmacological impact on several commonly used drugs. This impact can be clinically significant; therefore, the pharmacokinetic and pharmacodynamic impact of smoking should be considered in medicine selection, as well as the introduction and cessation of other therapies. This article aims to outline the considerations pharmacists and other healthcare professionals should be aware of in this context, including the impact of smoking cessation on a drug’s pharmacology and what should be done in practice to avoid the associated effects.

Nicotine pharmacology

During cigarette smoking, nicotine is inhaled and rapidly absorbed into the pulmonary circulation, reaching the arterial circulation and brain within 10–20 seconds and peak concentration within a few minutes^[5]
. When nicotine diffuses into brain tissue it binds to nicotinic cholinergic receptors in the brain, which triggers the release of neurotransmitters (e.g. dopamine, adrenaline, acetylcholine, gamma aminobutyric acid and beta endorphin), and also stimulates the release of the growth hormone, prolactin, and adrenocorticotrophic hormone^[6],[7]
. Subsequently, as nicotine is then promptly distributed to other body tissues, the levels start to decline after around 20–30 minutes.

Long-term cigarette smoking leads to both tolerance and dependence, with smoking becoming a self-medicating and self-regulating behaviour to reduce withdrawal symptoms. Studies have demonstrated that smokers increase their cigarette intake with low-nicotine cigarettes and, in those individuals with urinary acidification, this results in increased nicotine clearance^[7]
.

Tobacco smoking releases carbon monoxide and other toxins, such as polycyclic aromatic hydrocarbons (PAHs), primarily responsible for induction of hepatic enzymes CYP1A2 and CYP2B6, and to a lesser extent CYP3A4 and CYP2C19^[8],[9]
. CYP1A2 is a member of the cytochrome P450 superfamily of hepatic enzymes and is involved in the metabolism of many drugs^[10]
. The degree of enzyme induction shows significant inter-individual variability and may be dose dependent, with one study demonstrating dose-dependent increases in CYP1A2 induction with increases in number of cigarettes smoked per day^[8],[9]
. The effects of smoking should be considered in individuals with a smoking rate of more than ten cigarettes per day^[8]
. Genetic polymorphisms of the CYP1A2 gene and ethnic differences also contribute to inter-individual variability in both smokers and nonsmokers. Upon smoking cessation, enzyme induction rapidly recovers within days of cessation and reaches a new steady state within one week^[11]
.

The effects of smoking and abrupt smoking cessation are clinically relevant in the management of patients who are treated with a CYP1A2 substrate.

Pharmacokinetic and pharmacodynamic interactions

Tobacco smoking is known to cause interactions with many drugs from different classes. The following sections summarise the clinically relevant pharmacokinetic and pharmacodynamic interactions. A summary of the clinically important pharmacokinetic and pharmacodynamic interactions discussed in these sections can be found in the Table.

Psychotropic drugs

Smoking rates are high among people with mental health conditions, compared with the general population^[12]
. Epidemiological studies from the UK, United States and Australia have demonstrated lifetime smoking rates up to 50–60% in those with mental health conditions, with the highest incidences occurring during periods of acute exacerbation of psychosis or depression^{[12],[13],[14]}
. Therefore, it is highly likely that pharmacy professionals will come into contact with patients who have mental health conditions and who smoke. In addition, they are also highly likely to be in receipt of psychotropic drugs.

Antipsychotics

The effect of smoking and CYP1A2 induction has been clearly demonstrated with clozapine and olanzapine^[15]
. Clozapine is metabolised primarily by CYP1A2 and the difference in enzyme activity accounts for the majority of variance in clozapine metabolism. Drug interactions between clozapine and both inducers and inhibitors of CYP3A4 and CYP2D6 also suggest clearance is mediated in some part by these enzymes^[16]
. Therapeutic drug monitoring and pharmacokinetic studies have shown significant decreases in clozapine clearance after smoking cessation^{[15],[16],[17],[18],[19]}
. Meyer demonstrated a mean increase of clozapine levels of 57% after smoking cessation^[16]
, Li et al. showed a 45% mean increase in clozapine clearance^[18]
and a meta-analysis by Tsuda et al. demonstrated that the concentration/dose ratio was significantly lower in smokers than in nonsmokers^[15]
. Case reports have shown clozapine toxicity upon cessation of smoking^[18]
.

Olanzapine is metabolised primarily by CYP1A2 (50–60% of metabolism) and also via CYP2D6, CYP3A4 and glucuronidation^[15]
. Therapeutic drug monitoring studies have shown significantly reduced concentrations of olanzapine in smokers relative to nonsmokers. Gex-Fabry et al. retrospectively reviewed therapeutic drug monitoring data and showed mean concentration reduction of ~12% in smokers compared with nonsmokers^[20]
. A prospective therapeutic drug monitoring study showed that steady-state plasma olanzapine concentration/dose ratio closely correlated with CYP1A2 metabolic activity, and described a statistically significant five-fold lower dose-corrected steady-state plasma olanzapine concentration in smokers^[21]
. The study authors demonstrated that the mean percentage difference in a standard measure of treatment effectiveness after 15 days of antipsychotic therapy was significantly higher in nonsmokers than in smokers, suggesting treatment effect correlated with drug levels. A meta-analysis by Tsuda et al. of 683 smokers and 451 nonsmokers demonstrated that olanzapine concentration/dose ratio was significantly lower in smokers^[15]
. Owing to the rapid decrease in enzyme induction after smoking cessation, patients on clozapine or olanzapine who cease smoking should undergo immediate dose reductions and therapeutic drug monitoring, where available, as a priority. Patients commencing treatment on clozapine and olanzapine should also be routinely counselled about possible drug interactions with smoking and the potential for experiencing adverse events when starting or stopping smoking, so they understand the importance of discussing changes in smoking behaviours with their prescribers.

Risperidone, metabolised by the CYP2B6 pathway, has a lower concentration/dose ratio in smokers; moderate and heavy smokers have been found to require larger doses of risperidone^[22]
. Interaction with antipsychotics that are minor substrates for CYP1A2 (e.g. haloperidol, perphenazine and chlorpromazine) may also be possible with smoking, although are likely to be less clinically significant^[23]
.

Antidepressants

Antidepressants metabolised via the CYP1A2, CYP2D6 and CYP3A4 pathway have been shown to have decreased serum concentrations in smokers — the best evidence exists for fluvoxamine, mirtazapine, trazodone and duloxetine^[24]
. A recent review of therapeutic drug monitoring data demonstrated a reduction in concentration/dose of amitriptyline, escitalopram and mirtazapine in smokers, even when age and sex variables were accounted for^[25]
. Although the clinical significance of this remains unclear, it is possible that sudden smoking cessation may increase serum drug concentrations and increase negative side effects (rather than cause toxicity).

Benzodiazepines

Smokers report less sedation on benzodiazepines compared with nonsmokers — a potential pharmacodynamic effect owing to stimulation from nicotine^[26]
. Consideration should be given to the potential risk of sedation upon sudden cessation of smoking in patients on benzodiazepines.

Methadone

Smoking results in both pharmacokinetic and pharmacodynamic interactions with methadone. Methadone is a substrate of CYP2B6, which is induced by PAH exposure in smokers. Nicotine affects the endogenous opioid system, leading to higher concentrations of beta endorphin in nicotine users. Smoking rates in methadone users are high and reduced methadone sedation and higher methadone doses have been shown in smokers^{[27],[28],[29]}
. Methadone also attenuates nicotine withdrawal^[27]
. Close monitoring should occur when a patient changes smoking status, with dose adjustments made according to individual patient response.

Anticoagulation drugs

Warfarin is metabolised primarily via CYP2C9 and, to a lesser extent by CYP1A2, CYP2C19 and CYP3A4. A systematic review and meta-analysis of interactions between smoking and warfarin demonstrated an effect of smoking on both warfarin pharmacokinetics and warfarin dose requirement^[30]
. A pharmacokinetic study in nine people showed smoking cessation led to a 13% increase in steady-state warfarin concentration, 13% decrease in warfarin clearance and 23% increase in warfarin half-life^[31]
. Pooled analysis by Nathisuwan et al. showed that smoking was associated with a 12% increase in warfarin dosage requirement compared with nonsmoking and smokers required 2.26mg more of warfarin per week compared with nonsmokers^[30]
. Lenzini et al. compared pharmacogenetic and clinical protocols for warfarin initiation in 265 orthopaedic patients and found that smoking status was a significant predictor of warfarin dose, and included smoking in their therapeutic algorithm^[32]
. Patients commencing warfarin therapy should be counselled about the potential smoking–drug interaction and more frequent monitoring should be performed in patients who either commence smoking or suddenly cease smoking.

Antiplatelet drugs

The thienopyridine P2Y12 inhibitors clopidogrel and prasugrel are converted to their active metabolites by CYP isoenzymes, including CYP1A2, CYP2B6, CYP2C19, CYP3A4/5 and CYP2C9, thus enzyme induction by smoking could theoretically increase the antiplatelet activity of these agents; however, study results in this area are varied^{[33],[34],[35],[36]}
. Park et al. examined whether genetic variations in the cytochrome and drug transporter system are associated with the effect of smoking on clopidogrel response by measuring clopidogrel on-treatment platelet reactivity in 1,431 consecutive patients (smokers n=249) who underwent coronary angiography. They reported an enhanced clopidogrel response in smokers who were carriers of the CYP1A2 (-163C.A) A-allele^[34]
.

The PARADOX study was a prospective, randomised, double-blind, double-dummy, placebo-controlled crossover study of objectively assessed nonsmokers (n=56) and smokers (n=54) with stable coronary artery disease receiving aspirin therapy. The researchers aimed to evaluate the effect of smoking on the pharmacokinetics and pharmacodynamics of clopidogrel and prasugrel therapy. The results demonstrated higher active metabolite exposure and an increase in platelet inhibition in smokers treated with clopidogrel^[35]
. No significant difference was found in prasugrel smokers compared with nonsmokers^[35]
.

Subgroup analysis of 7,062 patients aged <75 years from the primary TRILOGY ACS trial cohort randomised to prasugrel versus clopidogrel were evaluated through 30 months by baseline and time-dependent smoking status. This analysis showed platelet reactivity was lower in current smokers with no significant interaction by treatment with prasugrel versus clopidogrel^[36]
. It is difficult to make clinical recommendations in this setting; however, an awareness of these potential interactions in medication prescribing is useful.

Antihypertensive drugs

In smokers, as with nonsmokers, careful attention should be paid to management of cardiovascular risk factors. Hypertension is a major risk for atherosclerosis by deranging the normal prothrombotic/fibrinolytic balance at the coronary vascular endothelium^[37]
. Patients with hypertension should be offered lifestyle modification advice (see Box) plus pharmacotherapy. First-line agents for hypertension include thiazide diuretics, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers and calcium channel blockers. Current evidence does not support the initial use of beta-blockers for hypertension in the absence of specific cardiovascular comorbidities, such as ischaemic heart disease^[38]
.

Despite these recommendations, there is some evidence that smokers with hypertension gain additional benefits from first-line beta-blocker therapy that are not seen in the non smoking population^[37]
. Cigarette smoking inactivates nitrous oxide (a potent vasodilator) within the vasculature, thereby disrupting endothelial function^[39]
.  Arterial hypertension separately impairs the coronary vascular endothelium, leading to prothrombotic and hypercoagulable effects. Fibrinogen, plasminogen activator inhibitor-1 (PAI-1) and homocystine (Hcy) are associated with cardiovascular risk via these associated endothelial effects. Vyssoulis et al. studied the effect of vasodilating beta-blockers on endothelial function and prothrombotic state in 550 patients with uncomplicated essential hypertension. They found that smokers have elevated baseline values of fibrinogen, PAI-1 and Hcy compared with nonsmokers. When treated with celiprolol, carvedilol or nebivolol, all parameters were reduced more in smokers than nonsmokers, with nebivolol demonstrating the most favourable effects^[37]
. This suggests that smokers may benefit from vasodilating beta-blockers, although this has not yet been demonstrated in terms of cardiovascular end points and mortality data. Further studies are needed before changes to current guidelines can be recommended.

Respiratory drugs

Theophylline, which has a narrow therapeutic window, is predominantly metabolised by CYP1A2 and increased clearance of 60–100% has been demonstrated in smokers^[40],[41]
. Dosage must be individualised and guided by theophylline concentration measurements in smokers and also upon sudden cessation of smoking where theophylline toxicity may occur. Tobacco smoking may also reduce efficacy of inhaled corticosteroids, potentially leading to prolonged exacerbation of asthma^[40]
.

Antineoplastic drugs

Smoking has been shown to be an independent risk factor for both developing cancer and responding to therapy^[42]
. Both nicotine and smoke-related toxins, such as PAHs, have complex pharmacodynamic effects in cancer treatment, with interactions between apoptosis, tumour angiogenesis, host immune function and chemosensitivity^[42]
. Pharmacokinetic interactions have been demonstrated with erlotininb and irinotecan, and are postulated in cyclophosphamide and ifosfamide, thus drug interaction between smoking and chemotherapy may also play a role.

Erlotinib

A pharmacokinetic study by Hamilton et al. in 32 healthy male subjects aimed to compare the pharmacokinetic variables of erlotinib in current smokers with nonsmokers after receiving a single oral 150mg or 300mg dose of erlotinib. The authors found that erolotinib and its metabolite OSI-410 demonstrated significantly reduced exposure to erlotinib in smokers, with a 2.8-fold lower area under the curve and a peak serum concentration two-thirds that of nonsmokers^[43]
.

Irinotecan

Irinotecan clearance has been shown to be significantly faster in smokers^[44]
. Systemic exposure to its active metabolite SN-38 is also lower^[44]
. Both of these factors explain the reduction in haematologic toxicity in smokers treated with irinotecan. Irinotecan metabolism is sensitive to CYP3A induction^[44]
.

Cyclophosphamide and ifosfamide

The alkylating chemotherapy agents ifosfamide and cyclophosphamide are metabolised by CYP2B6, 2C9 and 3A4, all of which are induced by smoking^[45]
. The effect of smoking on the pharmacokinetics of these agents has not been studied. However, smoking has been identified as an independent risk factor in multiple large studies where cyclophosphamide and/or ifosfamide were involved in therapy. Lamar et al. found smoking to be an independent risk factor for progression-free and overall survival in the treatment of 136 patients with aggressive non-Hodgkin lymphoma on dose adjusted etoposide, prednisone, vincristine, cyclophosphamide and doxorubicin with rituximab (DA-EPOCH-R)^[46]
. In this study, chemotherapy (including cyclophosphamide) was dose adjusted based on haematological toxicity. A prospective study in 166 patients with soft tissue sarcomas by Gannon et al. found that smoking affected distant metastasis-free survival and progression-free survival in patients with soft tissue sarcoma treated with radiotherapy +/- ifosfamide and doxorubicin^[47]
. Further research, including pharmacokinetic studies in smokers versus nonsmokers, is warranted to clarify the impact of smoking on ifosfamide and cyclophosphamide levels, and nicotine’s impact on cancer cell biology and chemosensitivity.

Pharmacotherapies used to support smoking cessation

Smoking cessation is associated with great benefits to health and should be routinely recommended by pharmacists and other healthcare professionals. At the time of smoking cessation, pharmacists should give consideration to its pharmacokinetic effects, as outlined above, along with the potential drug interactions with pharmacotherapies used to help the person abstain from cigarettes. The CYP1A2 induction that occurs with cigarette smoking is owing to PAHs in tobacco smoke, not the nicotine itself^[8],[9]
. Therefore, this enzyme induction will not occur in patients receiving nicotine replacement therapy (NRT) and dosing adjustments should take place accordingly^[48]
.

Stopping smoking can be difficult and only around 5% of those who try to quit without professional support remain abstinent at one year. Assistance with smoking cessation includes behavioural support, NRT and other pharmacotherapies^[49]
. Recommended pharmacotherapies for smoking cessation include NRT, varenicline and bupropion. For optimal outcomes, varenicline should be used for at least 12 weeks and NRT or bupropion for a minimum of 8 weeks^[50]
.

Bupropion

Bupropion, a selective catecholamine reuptake inhibitor, reverses the negative effect associated with nicotine withdrawal^[51]
. Bupropion is associated with an infrequent but possible dose-related risk of seizures. Using other drugs that lower the seizure threshold, or concurrently reducing benzodiazepines or other antiseizure medications, increases this risk of seizures. Bupropion is metabolised by CYP2B6 to hydroxybupropion and other metabolites. Co-administration of drugs known to induce CYP2B6 metabolism (e.g. carbamazepine and phenytoin) or inhibit metabolism through a phase II process (e.g. valproate) may affect the activity of bupropion.

Bupropion acts as an inhibitor in the CYP2D6 pathway, a common metabolic pathway for many antidepressant and antipsychotic medicines including venlafaxine and tricyclic antidepressants. Metoprolol, tramadol, methadone and several other cardiovascular and analgesic medicines are also substrates of CYP2D6 and dosages may need to be reduced if bupropion is prescribed^[52]
. There have been case reports of delirium in patients where bupropion was prescribed with risperidone^[53]
or duloxetine^[54]
, possibly owing to impaired metabolism of hydroxybupropion, a CYP2D6 substrate.

Varenicline

Varenicline binds with high affinity and selectivity at the α4β2 neuronal nicotinic acetylcholine receptor, where it acts as a partial agonist and relieves nicotine withdrawal symptoms^[51]
. Its binding both alleviates symptoms of craving and withdrawal, and reduces the rewarding and reinforcing effects of smoking by preventing nicotine binding to α4β2 receptors^[55]
. To date, varenicline has no known clinically significant drug interactions^[56]
. Concern about severe psychiatric and cardiac side effects have been studied through US Food and Drug Administration post-marketing randomised control trials (RCTs).

A multinational RCT of more than 8,000 smokers found no significant increase in rates of moderate-to-severe neuropsychiatric adverse events with either varenicline or bupropion relative to nicotine patch or placebo in those with or without psychiatric disorders (although the trial excluded patients with symptomatically unstable psychiatric conditions)^[57]
. While this study did not specifically look at drug interactions, over 45% of the psychiatric cohort were on psychotropic medicines and it was reported that adverse event rates with NRT, bupropion and varenicline were comparable to placebo^[57]
. A non-treatment extension phase of this study examined rate and onset of cardiovascular events and found no significant difference in time to onset of major adverse cardiovascular events with varenicline or bupropion versus placebo^[58]
. Although these studies were not designed to investigate pharmacological interactions, the findings may help to reassure those concerned about pharmacodynamic interactions adversely affecting the psychiatric or cardiovascular health of patients.

E-cigarettes

E-cigarettes are vapourising devices that heat a liquid in a cartridge or reservoir into an aerosol for inhalation. The substance contained in the electronic cigarette may contain varying amounts of nicotine or other chemicals with the purpose of providing different flavours. Owing to this lack of standardisation, it is not possible to make categorical statements about the chemical properties or safety of e-cigarettes^[49]
. Based on the available information, e-cigarettes are currently considered safer than conventional cigarettes^[59]
.

Summary

Smoking induces a number of enzymes that affect drug metabolism. Smoking cessation leads to prompt reversal of enzyme induction, which leads to significantly altered drug concentrations. Smoking cessation should always be encouraged and supported; however, prescribers should also give careful consideration to the doses of medicines that are metabolised via CYP1A2 and CYP2B6 in smokers, particularly when a person ceases smoking. Therapeutic drug monitoring, where available, should also be considered. NRT and varenicline may be used to increase chances of successful cessation and do not have any known significant drug–drug interactions. Bupropion is another option for smoking cessation, but this drug has multiple potential interactions that should be taken into consideration at initiation and cessation of therapies.