Coffee, Cigarettes and Meds: What Are the Metabolic Effects?

Heavy smoking and caffeine intake are highly prevalent in patients with psychiatric disorders, both of which significantly impact the metabolism of a number of psychotropic medications. Hence, these factors should be routinely considered in making prescribing decisions.

Psychiatric Times May 2005 Vol. XXII Issue 6

The prevalence of smoking in severe mental disorders is higher than the normal population. In worldwide schizophrenia studies, the rate is approximately six times higher than in the general population (de Leon and Diaz, in press). Many patients with severe mental disorders are heavy smokers (Dalack et al., 1998; de Leon et al., 2002). Patients with severe mental disorders also tend to have high intake of caffeine (Hughes et al., 1998). This may, in part, be explained by the high coprevalence of heavy smoking, which increases caffeine metabolism (Gurpegui et al., 2004). Smoking and caffeine intake can affect the metabolism of some psychotropic medications with implications for clinical dosage adjustment.

Nicotine is mainly metabolized to cotinine by cytochrome P450 (CYP) 2A6 (CYP2A6) (Nakajima et al., 2002). Byproducts of tobacco smoke such as polycyclic aromatic hydrocarbons are inducers of cytochrome P450 isoenzyme 1A2 (CYP1A2) and of the less-understood UDP-glucuronosyl-transferases (UGTs) (de Leon, 2003; Zevin and Benowitz, 1999). This enzyme induction occurs with marijuana smoking as well. The effects of these inducers depend on the turnover of the hepatic enzymes and production of new enzymes. Thus, it takes several weeks for their maximum effect to occur and, similarly, for their enzyme-inducing effect to resolve after cessation of smoking.

Caffeine is highly dependent (>90%) on CYP1A2 for its metabolism (de Leon et al., 2003). It competitively inhibits CYP1A2 and increases the levels of medications metabolized by this enzyme. Thus the effects of caffeine are opposite that of smoking. Due to the metabolic inductive effects of smoking, smokers require three to four times the caffeine doses compared to non-smokers to get similar plasma caffeine levels and hence caffeine intake is usually higher in smokers (de Leon et al., 2003).

Second-Generation Antipsychotics

Second-generation antipsychotics that are not metabolized by CYP1A2 should not be affected by either smoking or caffeine. Thus, smoking or caffeine intake should not influence dosing of aripiprazole (Abilify) and risperidone (Risperdal) (the metabolism of both depends on CYP2D6 and CYP3A4) (Prior and Baker, 2003; Swainston Harrison and Perry, 2004), quetiapine (Seroquel) (mainly metabolized by CYP3A4), and ziprasidone (Geodon) (mainly metabolized by an aldehyde oxidase and CYP3A4) (Prior and Baker, 2003). On the other hand, the metabolism of clozapine (Clozaril) and olanzapine (Zyprexa) is mainly dependent on CYP1A2 and UGTs, and hence their levels are affected by smoking and caffeine. The Table lists the effects of smoking and caffeine on second-generation antipsychotics.

Clozapine. The breadth of therapeutic index for a drug determines the clinical significance of plasma level changes associated with smoking or caffeine intake. Clozapine has a much narrower therapeutic index compared to olanzapine. Several of clozapine's side effects are dose related (levels >1000 ng/mL have been associated with toxicity, including seizure risk and severe sedation) (Greenwood-Smith et al., 2003). In a patient on clozapine who smokes, smoking cessation would probably cause an average patient's blood clozapine level to increase by 1.5-fold two to four weeks later. Similarly, if a patient stabilized in a non-smoking environment starts to smoke heavily (more than one pack/day), the clinician may need to consider increasing the clozapine dose by a factor of 1.5 over two to four weeks. Checking for side effects and measuring a clozapine level may then be prudent, since the 1.5 factor is only a rough approximation.

Caffeine competitively inhibits CYP1A2, and a correction factor of 0.6 has been recommended for those who are on clozapine and begin to consume significant quantities of caffeine (de Leon, 2004). If a patient stabilized on clozapine in a caffeine-free environment begins to regularly consume high quantities of caffeine, it may be safest to decrease the clozapine dose (e.g., from 400 mg/day to 250 mg/day [400 x 0.6=240]). Low levels of caffeine intake may not have clinically significant interactions with clozapine. The Figure illustrates the amount of caffeine in selected beverages in the United States. There are no data on what level of caffeine intake is safe in patients taking clozapine. Steady caffeine doses in a patient stabilized on clozapine should not be of concern for clinicians. It is important to warn patients to avoid significant changes in caffeine intake without discussion with their psychiatrist. Although there are no published data on what a significant change in caffeine intake is, de Leon (2004) recommended caution with changes (increases or decreases) of daily caffeine intake more than three cups (or six caffeinated soda cans) in smokers, and with changes more than one cup of coffee (or two caffeinated soda cans) in non-smokers. For example, when a smoker taking clozapine increases caffeine intake by three cups of coffee (e.g., from two cups/day to five cups/day), clinicians should watch for increased side effects due to increased clozapine levels. When a non-smoker taking clozapine decreases caffeine intake by two cans of soda (e.g., from four cans/day to two cans/day), clinicians should watch for possible loss of clozapine response due to decreased clozapine levels.

Skogh et al. (1999) reported a case of a 38-year-old white male patient with schizophrenia stabilized on 725 mg of clozapine for seven years. The patient stopped smoking and two weeks later was admitted to the hospital with hypotension and generalized tonic-clonic seizure. The authors attributed this episode to smoking cessation. Subsequently, the patient was stabilized on 425 mg of clozapine while remaining a non-smoker.

Odom-White and de Leon (1996) reported a case of a 31-year-old African-American woman on 550 mg of clozapine consuming about 1200 mg of caffeine a day and showing plasma clozapine level of 1500 ng/mL and norclozapine of 630 ng/mL. One week after discontinuation of caffeine, the same dosage of 550 mg/day showed a much lower clozapine level of 630 ng/mL and norclozapine level of 330 ng/mL.

Olanzapine. The enzyme-inducing effects of smoking and inhibiting effects of caffeine on olanzapine levels are probably very similar to clozapine. Smoking may lower olanzapine levels by 40% (Botts et al., 2004). The clinical significance of this change may be limited in comparison to clozapine, as olanzapine has a much broader therapeutic index and high levels of olanzapine are not associated with the serious toxicity described with clozapine. However, higher than recommended doses of olanzapine may be required for some heavy smokers due to low olanzapine levels (Carrillo et al., 2003).

Smoking status may also influence a patient's liability to develop adverse effects. For example, it may have a protective effect against antipsychotic-induced parkinsonism (Jabs et al., 2003). On the other hand, smoking has been shown to increase Abnormal Involuntary Movement Scale (AIMS) scores in some patients (Ellingrod et al., 2002). However, some of these effects may be pharmacodynamic in nature and explained by the influence of nicotine on dopaminergic function.

Other Psychotropic Medications

Haloperidol. Haloperidol (Haldol) is metabolized by CYP2D6 and possibly by CYP3A5 and UGTs (de Leon et al., 2004). As CYP1A2 is not involved in haloperidol metabolism, changes in haloperidol dosage due to caffeine intake are generally not recommended. Some studies suggest that smoking reduces haloperidol levels; this effect may be explained by induction of UGTs, but it has not been well studied.

Phenothiazines. Fluphenazine (Permitil, Prolixin) clearance is increased significantly in smokers, and lower levels of fluphenazine decanoate have been reported (Jann et al., 1985). Chlorpromazine (Thorazine) levels are also reported lower in smokers (Desai et al., 2001). It is not clear if the effects of smoking on phenothiazine levels are due to CYP1A2 and/or UGT induction. The effects of caffeine intake on phenothiazines have not been studied, but some effect is possible since phenothiazines are probably partly metabolized by CYP1A2.

Antidepressants. Antidepressants that have increased metabolism and clearance from smoking include imipramine (Tofranil), clomipramine (Anafranil), fluvoxamine (Luvox) and trazodone (Desyrel) (Desai et al., 2001). It is likely that these inductive effects may be explained by CYP1A2 and/or UGT induction (de Leon, 2003). The effect of smoking on the plasma concentrations of amitriptyline and nortriptyline (Aventyl, Pamelor) is variable, and buproprion (Wellbutrin) is unaffected by smoking (Desai et al., 2001). Fluvoxamine is a powerful CYP1A2 inhibitor and may cause clinically relevant inhibition of caffeine metabolism (Yoshimura et al., 2002).

Benzodiazepines. Increased clearance of the benzodiazepines alprazolam (Xanax), demethyldiazepam, diazepam (Valium), lorazepam (Ativan) and oxazepam (Serax) is found in cigarette smokers, whereas chlordiazepoxide (Libritabs, Librium, Mitran) does not appear to be affected by smoking (Desai et al., 2001; Greenblatt et al., 1980). Induction of UGTs probably explains the effects of smoking on benzodiazepine metabolism.

Further Considerations and Concluding Thoughts

There are specific environments and situations that impact smoking and caffeine intake and thereby influence medication dosages. Hospital restrictions are one example that may have clinical importance. Smoking is banned inside U.S. hospitals; thus, clinicians need to carefully question patients and consider smoking's effects on some antipsychotic doses in the hospital (limited or no smoking) versus those in the community (unrestricted smoking). Many hospital environments also restrict caffeinated beverages. There are no simple correction factors for patients stabilized in a non-smoking and caffeine-free hospital environment when they go to environments allowing smoking and caffeine intake. Clinicians should be aware of the time course of changes in enzyme activity and monitor patients closely for reduced efficacy or increased risk of side effects. Also, clinicians should be aware of the pharmacokinetic implications of situations in which patients reduce or quit smoking or cut down their intake of caffeine.

Recent guidelines to promote physical health recommend that efforts toward nicotine cessation be part of regular psychiatric treatment (Goff et al., 2005). Psychiatrists should assess patients' motivation to quit on a routine basis, strengthen their motivation, offer full array of pharmaceutical treatments like buproprion (Zyban) and nicotine replacement, and psychosocial treatment like group and individual counseling. Involving family members and referring patients to quit groups can increase family and community support and the likelihood of smoking cessation.

In summary, cigarette smoking and caffeine intake significantly impact the metabolism of a number of psychotropic medications and hence these factors should be routinely considered in prescribing decisions. While patients should be routinely encouraged to quit smoking they also need to be educated about the importance of informing their treating psychiatrist of changes in their caffeine intake or smoking. Research in new drug development should include studies to determine the impact of smoking and caffeine intake on dosages and clinical response.

Dr. Pinninti is assistant professor of psychiatry at the University of Medicine and Dentistry of New Jersey School of Osteopathic Medicine and medical director of Steininger Behavioral Care Services, Inc.

Dr. Mago is assistant professor of psychiatry, director of the mood disorders program and associate director of consultation-liaison psychiatry at the Thomas Jefferson University.

Dr. de Leon is associate professor of psychiatry at the University of Kentucky College of Medicine and medical director of the University of Kentucky Mental Health Research Center at Eastern State Hospital in Lexington, Ky.


Botts S, Littrell R, de Leon J (2004), Variables associated with high olanzapine dosing in a state hospital. J Clin Psychiatry 65(8):1138-1143.

Carrillo JA, Herraiz AG, Ramos SI et al. (2003), Role of the smoking-induced cytochrome P450 (CYP)1A2 and polymorphic CYP2D6 in steady-state concentration of olanzapine. J Clin Psychopharmacol 23(2):119-127.

Dalack GW, Healy DJ, Meador-Woodruff JH (1998), Nicotine dependence in schizophrenia: clinical phenomena and laboratory findings. Am J Psychiatry 155(11):1490-1501.

de Leon J (2003), Glucuronidation enzyme genes and psychiatry. Int J Neuropsychopharmacol 6(1):57-72.

de Leon J (2004), Atypical antipsychotic dosing: the effect of smoking and caffeine. Psychiatr Serv 55(5):491-493.

de Leon J, Diaz FJ (in press), A meta-analysis of worldwide studies demonstrates an association between schizophrenia and tobacco smoking behaviors. Schizophr Res.

de Leon J, Diaz FJ, Rogers T et al. (2003), A pilot study of plasma caffeine concentrations in a US sample of smoker and nonsmoker volunteers. Prog Neuropsychopharmacol Biol Psychiatry 27(1):165-171.

de Leon J, Diaz FJ, Wedlund P et al. (2004), Haloperidol half-life after chronic dosing. J Clin Psychopharmacol 24(6):656-660.

de Leon J, Tracy J, McCann E et al. (2002), Schizophrenia and tobacco smoking: a replication study in another US psychiatric hospital. Schizophr Res 56(1-2):55-65.

Desai HD, Seabolt J, Jann MW (2001), Smoking in patients receiving psychotropic medications: a pharmacokinetic perspective. CNS Drugs 15(6):469-494.

Ellingrod VL, Schultz SK, Arndt S (2002), Abnormal movements and tardive dyskinesia in smokers and nonsmokers with schizophrenia genotyped for cytochrome P450 2D6. Pharmacotherapy 22(11):1416-1419.

Goff DC, Cather C, Evins AE et al. (2005), Medical morbidity and mortality in schizophrenia: guidelines for psychiatrists. J Clin Psychiatry 66(2):183-194.

Greenblatt DJ, Divoll M, Harmatz JS, Shader RI (1980), Oxazepam kinetics: effects of age and sex. J Pharmacol Exp Ther 215(1):86-91

Greenwood-Smith C, Lubman DI, Castle DJ (2003), Serum clozapine levels: a review of their clinical utility. J Psychopharmacol 17(2):234-238.

Gurpegui M, Aguilar MC, Martinez-Ortega JM et al. (2004), Caffeine intake in outpatients with schizophrenia. Schizophr Bull 30(4):935-945.

Hughes JR, McHugh P, Holtzman S (1998), Caffeine and schizophrenia. Psychiatr Serv 49(11):1415-1417.

Jabs BE, Bartsch AJ, Pfuhlmann B (2003), Susceptibility to neuroleptic-induced parkinsonism-age and increased substantia nigra echogenicity as putative risk factors. Eur Psychiatry 18(4):177-181.

Jann MW, Ereshefsky L, Saklad SR (1985), Clinical pharmacokinetics of the depot antipsychotics. Clin Pharmacokinet 10(4):315-333.

Nakajima M, Kuroiwa Y, Yokoi T (2002), Interindividual differences in nicotine metabolism and genetic polymorphisms of human CYP2A6. Drug Metab Rev 34(4):865-877.

Odom-White A, de Leon J (1996), Clozapine levels and caffeine. J Clin Psychiatry 57(4):175-176.

Prior TI, Baker GB (2003), Interactions between the cytochrome P450 system and the second-generation antipsychotics. J Psychiatry Neurosci 28(2):99-112.

Skogh E, Bengtsson F, Nordin C (1999), Could discontinuing smoking be hazardous for patients administered clozapine medication? A case report. Ther Drug Monit 21(5):580-582.

Swainston Harrison T, Perry CM (2004), Aripiprazole: a review of its use in schizophrenia and schizoaffective disorder. Drugs 64(15):1715-1736.

Yoshimura R, Ueda N, Nakamura J et al. (2002), Interaction between fluvoxamine and cotinine or caffeine. Neuropsychobiology 45(1):32-35.

Zevin S, Benowitz NL (1999), Drug interactions with tobacco smoking. An update. Clin Pharmacokinet 36(6):425-438.