In-vivo phenotyping for CYP3A by a single-point determination of midazolam plasma concentration.

We investigated whether a single plasma midazolam concentration could serve as an accurate predictor of total midazolam clearance, an established in-vivo probe measure of cytochrome P450 3A (CYP3A) activity. In a retrospective analysis of data from 224 healthy volunteers, non-compartmental pharmacokinetic parameters were estimated from plasma concentration-time curves following intravenous (IV) and/or oral administration. Based on statistical moment theory, the concentration at the mean residence time (MRT) should be the best predictor of the total area under the curve (AUC). Following IV or oral midazolam administration, the average MRT was found to be approximately 3.5 h, suggesting that the optimal single sampling time to predict AUC was between 3 and 4 h. Since a 4-h data point was common to all studies incorporated into this analysis, we selected this time point for further investigation. The concentrations of midazolam measured 4 h after an IV or oral dose explained 80 and 91% of the constitutive interindividual variability in midazolam AUC, respectively. The 4-h midazolam measurement was also an excellent predictor of drug-drug interactions involving CYP3A induction and inhibition. Compared with baseline values, the direction and magnitude of change in midazolam AUC and the 4-h concentration were completely concordant for all study subjects. We conclude that a single 4-h midazolam concentration following IV or oral administration represents an accurate marker of CYP3A phenotype under constitutive and modified states. Moreover, the single-point approach offers an efficient means to phenotype and identify individuals with important genetic polymorphisms that affect CYP3A activity.

Clinical pharmacokinetics of troglitazone.

Troglitazone is a new thiazolidinedione oral antidiabetic agent approved for use to improve glycaemic control in patients with type 2 diabetes. It is rapidly absorbed with an absolute bioavailability of between 40 and 50%. Food increases the absorption by 30 to 80%. The pharmacokinetics of troglitazone are linear over the clinical dosage range of 200 to 600 mg once daily. The mean elimination half-life ranges from 7.6 to 24 hours, which facilitates a once daily administration regimen. The pharmacokinetics of troglitazone are similar between patients with type 2 diabetes and healthy individuals. In humans, troglitazone undergoes metabolism by sulfation, glucuronidation and oxidation to form a sulfate conjugate (M1), glucuronide conjugate (M2) and quinone metabolite (M3), respectively. M1 and M3 are the major metabolites in plasma, and M2 is a minor metabolite. Age, gender, type 2 diabetes, renal impairment, smoking and race do not appear to influence the pharmacokinetics of troglitazone and its 2 major metabolites. In patients with hepatic impairment the plasma concentrations of troglitazone, M1 and M3 increase by 30%, 4-fold, and 2-fold, respectively. Cholestyramine decreases the absorption of troglitazone by 70%. Troglitazone may enhance the activities of cytochrome P450 (CYP) 3A and/or transporter(s) thereby reducing the plasma concentrations of terfenadine, cyclosporin, atorvastatin and fexofenadine. It also reduces the plasma concentrations of the oral contraceptive hormones ethinylestradiol, norethindrone and levonorgestrel. Troglitazone does not alter the pharmacokinetics of digoxin, glibenclamide (glyburide) or paracetamol (acetaminophen). There is no pharmacodynamic interaction between troglitazone and warfarin or alcohol (ethanol). Pharmacodynamic modelling showed that improvement in fasting glucose and triglyceride levels increased with dose from 200 to 600 mg. Knowledge of systemic troglitazone exposure within a dose group does not improve the prediction of glucose lowering response or adverse effects beyond those based on the administered dose.

Steady-state pharmacokinetics and dose proportionality of troglitazone and its metabolites.

This study evaluated the steady-state pharmacokinetics and dose proportionality of troglitazone, metabolite 1 (sulfate conjugate), and metabolite 3 (quinone metabolite) following administration of daily oral doses of 200, 400, and 600 mg troglitazone for 7 days (per dosing period) to 21 subjects. During each dosing period, plasma samples were collected predose on days 1, 5, 6 and 7 and serially for 24 hours on day 7. Steady-state plasma concentrations for troglitazone, metabolite 1, and metabolite 3 were achieved by day 7. Troglitazone was rapidly absorbed with mean tmax values of 2.7 to 2.9 hours. Mean Cmax and AUC(0-24) values for troglitazone, metabolite 1, and metabolite 3 increased proportionally with increasing troglitazone doses over the clinical dose range of 200 mg to 600 mg administered once daily. Mean troglitazone CL/F, percent fluctuation, and AUC ratios of metabolite 1 and metabolite 3 to troglitazone were similar across dose groups. These data suggest that the pharmacokinetics and disposition of troglitazone and its metabolites are independent of dose over the dose range studied. Thus, troglitazone, metabolite 1, and metabolite 3 displayed linear pharmacokinetics at steady-state.

Effect of troglitazone on the pharmacokinetics of an oral contraceptive agent.

Fifteen healthy women participated in a study to determine the effect of multiple doses of troglitazone on the pharmacokinetics of Ortho-Novum 1/35 (35 micrograms ethinyl estradiol [EE] and 1 mg norethindrone [NE]). Participants received three cycles (21 days each of active drug followed by 7 days without medication) of Ortho-Novum. During the third cycle, participants also received troglitazone 600 mg qd for 22 days. Pharmacokinetic profiles of EE and NE were determined on day 21 of the second and third cycles. Progesterone and sex hormone binding globulin (SHBG) levels were also measured. Troglitazone decreased EE Cmax and AUC(0-24) by 32% and 29%, respectively. Likewise, troglitazone decreased NE Cmax and AUC(0-24) by 31% and 30%, respectively. Plasma SHBG concentrations increased from 113 nmol/L during cycle 2 to 220 nmol/L during cycle 3. Troglitazone reduced plasma unbound AUC for NE by 49%. Serum progesterone levels were lower than 1.5 ng/mL on all occasions. Thus, coadministration of troglitazone and Ortho-Novum decreases the systemic exposure to EE and NE. A higher dose of oral contraceptive or an alternate method of contraception should be considered for patients treated with troglitazone.

Effect of troglitazone on urinary excretion of 6beta-hydroxycortisol.

Coadministration of troglitazone reduces the plasma concentrations of terfenadine and ethinyl estradiol. Because these drugs are metabolized at least in part by cytochrome P450 3A (CYP3A), it is possible that troglitazone induces CYP3A activity, thereby reducing plasma concentrations of these agents. Known inducers of CYP3A, such as rifampin, phenytoin, carbamazepine, and phenobarbital, increase the urinary excretion of 6beta-hydroxycortisol and the ratio of 6beta-hydroxycortisol to cortisol. This evaluation examined the effect of troglitazone on urinary excretion of 6beta-hydroxycortisol and the ratio of 6beta-hydroxycortisol to cortisol as a marker for CYP3A induction. Urine samples were collected from 11 subjects who completed a study evaluating the effect of multiple-dose administration of troglitazone 400 mg once daily (days 11-20) on the steady-state pharmacokinetics of digoxin (0.25 mg daily on days 1-20). A single urine sample was collected at baseline on day 1, and 24-hour urine samples were collected on days 10 and 20. Mean +/- standard deviation 24-hour excretion rate of cortisol was unchanged, whereas that of 6beta-hydroxycortisol increased during troglitazone administration. The ratio of 24-hour urinary 6beta-hydroxycortisol to cortisol excretion increased from 7.4 +/- 2.7 on day 10 to 16.1 +/- 6.2 on day 20. The ratio observed on day 10 was similar to that obtained at baseline. These results are consistent with the hypothesis that troglitazone induces CYP3A activity.

Effect of troglitazone on steady-state pharmacokinetics of digoxin.

Twelve healthy subjects participated in a study to determine the effect of multiple doses of troglitazone on the steady-state pharmacokinetics of digoxin. Subjects received digoxin 0.25 mg orally once daily on days 1 through 20 and 400 mg of troglitazone orally once daily on days 11 through 20. Serial plasma samples and 24-hour urine samples collected before and after the doses on days 10 and 20 were analyzed for digoxin using a radioimmunoassay method. Eleven subjects completed the study. Administration of multiple oral doses of digoxin and troglitazone was well tolerated. Mean values for maximum concentration (Cmax), time to Cmax (tmax), and area under the concentration-time curve from 0 to 24 hours (AUC0-24) of digoxin on day 10 were similar to those on day 20. Mean day 10 digoxin values for minimum concentration (Cmin), apparent oral clearance (Cl/F), total urinary excretion from 0 to 24 hours (Ae0-24), and renal clearance (Clr) were also similar to corresponding values on day 20. Thus, concomitant administration of multiple-dose troglitazone does not alter the steady-state pharmacokinetics of digoxin.

Aging and drug interactions. III. Individual and combined effects of cimetidine and cimetidine and ciprofloxacin on theophylline metabolism in healthy male and female nonsmokers.

The individual and combined effects of cimetidine and ciprofloxacin on theophylline metabolism were examined in healthy young and elderly male and female nonsmokers. Single-dose studies of theophylline pharmacokinetics were performed at base line and on the fifth day of each of three treatment regimens consisting of 400 mg cimetidine every 12 hr, 500 mg ciprofloxacin every 12 hr and the combination of cimetidine and ciprofloxacin. Base-line theophylline plasma clearance and formation clearance of theophylline metabolites decreased with age in both gender groups to a similar extent (20% less in elderly men than in young men; 24% less in elderly women than in young women). Individually, cimetidine and ciprofloxacin produced proportionate declines in plasma theophylline clearance that were similar among the four groups (range, 23.4-32.7% decrease). The combined regimen yielded further impairment in theophylline elimination compared with each agent alone (range, 35.9-42.6% decrease). Cimetidine was a nonselective inhibitor of theophylline metabolic pathways in young men, but it exerted a greater inhibitory effect on N-demethylation pathways in the other groups. Ciprofloxacin inhibited N-demethylations of theophylline to a greater extent than the hydroxylation pathway. Coadministration of these two inhibitors further reduced the formation of theophylline metabolites. The proportionate reduction in formation clearance of theophylline metabolites was similar among the four groups. Thus, the response to inhibition of theophylline metabolism by cimetidine and ciprofloxacin is not influenced by age or gender.

Acetylator phenotype in relation to age and gender in the Baltimore Longitudinal Study of Aging.

This study investigated in healthy Caucasians the possible occurrence of age and gender-associated differences in NAT2 acetylator phenotype. Acetylator phenotype was determined after a single oral dose of 100 mg dapsone during testing of oral glucose tolerance in 510 Caucasian volunteers aged from 19 to 93 years, 339 men and 171 women, from the Baltimore Longitudinal Study of Aging. Participants were classified as slow or rapid acetylators according to the ratio of monoacetyldapsone to dapsone concentration in plasma. The ratio dividing the two groups, 0.30, was chosen after inspection of a probit plot and histogram of the monoacetyldapsone/dapsone ratios. Fifty-one percent of the participants were slow acetylators and 49% were rapid acetylators. Because there was no significant difference between the sexes in the monoacetyl-dapsone/dapsone ratios, all 510 participants were pooled into a single group for further analysis. In the combined analysis, there was a small decline in the prevalence of the slow acetylator phenotype with age, but this age effect accounted for less than 1% of the total variance in the monoacetyldapsone/dapsone ratio (r2 = 0.009). Also, it was shown in a group of 20 participants that administration of glucose with dapsone does not influence the determination of acetylator phenotype. In a large healthy Caucasian. American population, there was no biologically important effect of age or sex on the distribution of NAT2 acetylator phenotype.

Meta-analysis of steady-state pharmacokinetics of troglitazone and its metabolites.

The object of this study is to evaluate the effects of age, gender, age-by-gender interaction, Type II diabetes, body weight, race, smoking, and formulation on steady-state pharmacokinetics of troglitazone, Metabolite 1 (sulfate conjugate), and Metabolite 3 (quinone metabolite) following multiple-dose oral administration of troglitazone. Pharmacokinetic parameter estimates [Cl/F (apparent oral clearance), AUC0-24 (area under plasma concentration-time curve), and ratio of AUC for troglitazone to Metabolite 1 and to Metabolite 3] obtained from 84 healthy volunteers and 171 patients with Type II diabetes in 8 studies were analyzed using a graphical method (for race and smoking) or a weighted ANCOVA model incorporating gender, health status (healthy vs Type II diabetes), and formulation as main effects; age, age-by-gender interaction, and body weight as continuous covariates. Ratio of AUC for troglitazone to metabolites was also examined by inspection of log-probit plots. Age, gender, age-by-gender, Type II diabetes, and formulation had negligible effects on troglitazone Cl/F, AUC0-24 (all analytes), and AUC ratio of troglitazone to metabolites. Race and smoking did not appear to influence steady-state pharmacokinetics of troglitazone and its metabolites. Although body weight was a significant covariate for AUC0-24 and Cl/F, the explanatory power of the overall model was weak (R2 < 0.2). Log-probit plots did not reveal a polymorphic distribution in AUC ratio of troglitazone to Metabolite 1 or Metabolite 3. Based on pharmacokinetics, dose adjustment for troglitazone in relation to the demographic factors examined is not required due to their poor predictive ability on steady-state pharmacokinetics of troglitazone and its metabolites.

Lack of effect of type II diabetes on the pharmacokinetics of troglitazone in a multiple-dose study.

Twelve patients with type II diabetes and 12 age-, weight-, and gender-matched healthy subjects participated in a study comparing the pharmacokinetics of troglitazone, metabolite 1 (sulfate conjugate), and metabolite 3 (quinone) after oral administration of 400 mg of troglitazone every morning for 15 days. Serial plasma samples collected after the dose on days 1 and 15 were analyzed for troglitazone, metabolite 1, and metabolite 3 using a validated HPLC method. Steady state plasma concentrations of troglitazone and its metabolites were achieved by the fifth day of troglitazone administration in both groups. Mean day 15 Cmax, tmax, AUC0-24, and Cl/F values of troglitazone were 1.54 micrograms/mL, 3.25 hours, 15.6 micrograms.hr/mL, and 461 mL/min, respectively, in patients with type II diabetes. Corresponding parameter values were 1.42 micrograms/mL, 2.63 hours, 12.5 micrograms.hr/mL, and 558 mL/min, respectively, in healthy subjects. Elimination t1/2 was approximately 24 hours in both groups. Mean day 15 pharmacokinetic parameter values for metabolite 1 and metabolite 3 were similar in the two groups. Ratio of AUC of metabolite 1 to troglitazone was 6.2 and 6.7, respectively, in patients and in healthy subjects. Ratio of AUC of metabolite 3 to troglitazone was 1.1 in both groups. Thus, steady-state pharmacokinetics and disposition of troglitazone and its metabolites in patients with type II diabetes were similar to those in healthy subjects.

Relative potency of mexiletine, lidocaine, and tocainide as inhibitors of rat liver CYP1A1 activity.

Mexiletine and tocainide are lidocaine congeners that share similar chemical structures. Clinical studies suggest that the in vivo inhibitory effect of mexiletine on the CYP1A family of isoforms is substantially greater than that of tocainide. We investigated the inhibitory property of mexiletine, lidocaine, and tocainide on the in vitro activity of the cytochrome P4501A1 (CYP1A1) isozyme in the rat. Hepatic microsomes were prepared from rat livers induced with 3-methylcholanthrene. The rate of ethoxyresorufin-O-dealkylation (EROD) was used as an index of CYP1A1 activity. Vmax and KM of the reactions were determined from Lineweaver-Burk plots. The Ki values for the inhibitors were derived from Dixon plots. Results showed that mexiletine is a competitive inhibitor, lidocaine is a mixed inhibitor, and tocainide is a noncompetitive inhibitor of EROD. The Ki values for mexiletine and tocainide were 0.30 +/- 0.02 mM and 12.4 +/- 0.7 mM, respectively. Two Ki values for lidocaine were determined. They were 0.65 +/- 0.07 mM and 4.1 +/- 1.3 mM, respectively. The relative inhibitory potency of these agents on rat CYP1A1 activity is mexiletine > lidocaine > tocainide. This difference in potency, which is most likely attributable to the change in the chemical composition in the aliphatic chain among the compounds, suggests that these compounds may be useful probes for studying the mechanism of the interaction with the active site of CYP1A1.

Characterization of the independent and combined effects of two inhibitors on oxidative drug metabolism in rat liver microsomes.

To evaluate how two inhibitors influence oxidative drug metabolism, this study investigated the inhibitory effects of mexiletine with cimetidine and mexiletine with lidocaine, both individually and in combination, on the oxidative metabolism of two probe substrates, aminopyrine and aniline in rat liver microsomes. Mexiletine was a competitive inhibitor of aminopyrine N-demethylation, whereas cimetidine was a mixed type of inhibitor (Ki = 2.00 +/- 0.04 and 0.20 +/- 0.02 mM, respectively). For aniline hydroxylation, mexiletine exhibited a mixed type of inhibition, whereas lidocaine was a noncompetitive inhibitor (Ki = 0.60 +/- 0.07 and 8.50 +/- 0.12 mM, respectively). The combined inhibition of either mexiletine with cimetidine or mexiletine with lidocaine on aminopyrine and aniline metabolism was close to the fully additive effects of the individual compounds when their individual concentrations were below a 2-fold Ki concentration, regardless of the apparent kinetic inhibition type. The combined inhibition was less than fully additive when the individual concentrations were twice the Ki or above. These results demonstrate that, when two inhibitors of oxidative drug metabolism are combined, both the Ki values and the concentrations of inhibitors play important roles in determining the extent of additive inhibition of enzyme activity.

Theophylline metabolism in healthy nonsmokers and in patients with insulin-dependent diabetes mellitus.

This study examined the pharmacokinetics of theophylline and formation of its metabolites in patients with insulin-dependent diabetes mellitus (IDDM) and in sex-, age-, and weight-matched healthy nonsmokers (n = 8 per group). Subjects received a single dose of 5 mg/kg theophylline intravenously. The pharmacokinetic parameter values of theophylline (plasma clearance, elimination half-life, and volume of distribution) and the formation clearance of theophylline metabolites (3-methylxanthine, 1-methyluric acid, and 1,3-dimethyluric acid) were similar between the two groups. Plasma free fraction of theophylline was higher in the subjects with diabetes than in the healthy subjects (0.61 +/- 0.04 versus 0.56 +/- 0.02; p < 0.001). In the group of subjects with diabetes, there was a positive correlation between hemoglobin A1c values and plasma theophylline clearance (r = 0.76; p < 0.05), formation clearance of 1,3-dimethyluric acid (r = 0.78; p < 0.05), and formation clearance of 1-methyluric acid (r = 0.71; p < 0.05). These results suggest that patients with IDDM and poor glycemic control are more likely to have an increased rate of theophylline metabolism.

Individual and combined effects of cimetidine and ciprofloxacin on theophylline metabolism in male nonsmokers.

1. The individual and combined effects of cimetidine and ciprofloxacin on theophylline metabolism were examined in six young male nonsmokers. 2. Treatment sequence consisted of 7 days each of cimetidine 400 mg p.o. every 12 h. ciprofloxacin 500 mg p.o. every 12 h, and the combination of cimetidine and ciprofloxacin. 3. Studies of theophylline pharmacokinetics were performed at baseline and on the fifth day of each regimen. 4. Individually, cimetidine and ciprofloxacin decreased the clearance of theophylline by 25% and 32%, respectively. Therapy with the combined regimen resulted in a 41% reduction in theophylline clearance, which was greater than that achieved with each drug alone (P < 0.01). 5. Ciprofloxacin, in contrast to cimetidine, inhibited N-demethylations of theophylline to a significantly greater extent than the hydroxylation pathway. Combined treatment produced a further decline in formation of 1,3-dimethyluric acid than each drug alone. 6. These data suggest that coadministration of cimetidine and ciprofloxacin exerts a greater impairment of theophylline biotransformation than each inhibitor alone. The enhanced inhibitory effect from the two inhibitors will occur only when sub-maximal doses of each individual agent are used.

Aging and the response to inhibition and induction of theophylline metabolism.

The twofold to threefold higher incidence of adverse drug reactions in elderly as opposed to younger patients is due mainly to more severe disease and the requirement for more complex drug treatment regimens. The incidence of adverse drug reactions increases with the number of prescribed drugs. Because of multiple drug use by the elderly, the potential for drug interactions is greater in this patient group. Surprisingly, the effect of age on the clinical pharmacology of drug interactions has not been thoroughly investigated. Our studies have shown that cimetidine inhibits and phenytoin induces the metabolism of theophylline to a similar extent in healthy male nonsmokers and smokers. Preliminary analysis of the results of a study to investigate the inhibition of theophylline metabolism by cimetidine and ciprofloxacin administered in combination to healthy male and female nonsmokers also does not show an age difference in response. Additional careful studies are needed to evaluate further the pharmacology and clinical importance of pharmacokinetic and pharmacodynamic drug interactions in the elderly.

The effect of tocainide on theophylline metabolism.

The effect of 5 days of oral tocainide (400 mg every 8 h) on the kinetics of theophylline given as a single 5 mg kg-1 i.v. infusion over 30 min was investigated in eight healthy male nonsmokers. Treatment with tocainide decreased the plasma clearance of theophylline from 37.5 +/- 6.9 (mean +/- s.d.) to 33.7 +/- 5.0 ml kg-1 h-1 (difference -3.8, 95% CI, -1.7 to -5.9; P = 0.004) and increased its terminal elimination half-life from 9.7 +/- 2.5 to 10.4 +/- 2.1 h (difference 0.7, 95% CI, 0.2 to 1.2; P = 0.011). Tocainide decreased the formation clearances of 3-methylxanthine and 1-methyluric acid, but the formation clearance of 1,3-dimethyluric acid was unaltered. These data indicate that tocainide exerts a modest inhibitory effect on theophylline metabolism. The magnitude of this change is substantially smaller than that reported to be produced by mexiletine.

Inhibition of theophylline metabolism by mexiletine in young male and female nonsmokers.

The influence of mexiletine (200 mg every 8 hours) on theophylline metabolism was studied in young male (n = 7) and female (n = 8) nonsmokers. A single-dose study of theophylline kinetics was performed at baseline and after 5 days of mexiletine treatment. With mexiletine the plasma clearance of theophylline decreased from 33.5 +/- 2.6 (mean +/- SEM) to 17.9 +/- 1.0 ml/kg per hour in the female group (p less than 0.001) and from 32.3 +/- 2.6 to 19.3 +/- 1.3 ml/kg per hour in the male group (p less than 0.001). The elimination half-life was prolonged by 74% and 103% in the male and female groups, respectively. Mexiletine decreased the formation of all theophylline metabolites in both groups. Within each group, the demethylation pathways were affected more than the hydroxylation pathway. These data indicate that mexiletine is a potent inhibitor of theophylline metabolism. This effect is not influenced by gender. Concurrent administration of these drugs may require a dose reduction of theophylline by as much as 50% to minimize the risk of toxicity.

Hepatic drug metabolism and aging.

Although there is considerable variation in the effect of age on drug biotransformation, the metabolism of many drugs is impaired in the elderly. Age-related physiological changes, such as a reduction in liver mass, hepatic metabolising enzyme activity, liver blood flow and alterations in plasma drug binding may account for the decreased elimination of some metabolised drugs in the elderly. It is difficult, however, to separate an effect of aging from a background of marked variation in the rate of metabolism due to factors such as individual metabolic phenotype, environmental influences, concomitant disease states and drug intake. The prevailing data suggest that initial doses of metabolised drugs should be reduced in older patients and then modified according to the clinical response. In most studies the elderly appear as responsive as young individuals to the effects of compounds which induce or inhibit the activity of cytochrome P450 isozymes. Concurrent use of other agents, which induce or inhibit drug metabolism, mandates dose adjustment as in younger patients. Many questions remain unanswered. For instance, limitations of in vitro studies prevent any firm conclusion about changes in hepatic drug metabolising enzyme activity in the elderly. With aging, some pathways of drug metabolism may be selectively affected, but this has not been adequately scrutinised. The possibility that metabolism of stereoisomers may be altered in the elderly has not been adequately tested. The effect of aging on the distribution of polymorphic drug metabolism phenotypes is still not established, despite potential implications for disease susceptibility and survival advantage.

Dose-dependent inhibition of theophylline metabolism by disulfiram in recovering alcoholics.

The influence of disulfiram on theophylline metabolism was studied in 20 recovering alcoholics. Ten of the patients, who were selected at random, received 250 mg of disulfiram daily. The other 10 patients received 500 mg of disulfiram daily. Two single-dose studies of theophylline kinetics were performed--one as a baseline control and the other after 1 week of treatment with disulfiram. With disulfiram pretreatment, the plasma clearance of theophylline was decreased from 105.7 +/- 10.2 (mean +/- SEM) to 83.1 +/- 8.1 ml/kg per hour (p less than 0.001) in the 250 mg group and from 94.3 +/- 13.3 to 65.4 +/- 10.7 ml/mg per hour (p less than 0.001) in the 500 mg group. The elimination half-life was prolonged significantly in both groups. The percent reduction in theophylline clearance was greater in the 500 mg group (32.5 +/- 3.1; range, 21.6 to 49.6) than it was in the 250 mg group (21.2 +/- 1.7; range, 14.6 to 29.6; p less than 0.01). Disulfiram decreased the formation of all theophylline metabolites in smokers in both treatment groups. In each group, the hydroxylation pathway was affected more than the demethylation pathway. These data indicate that at therapeutic doses disulfiram exerts a dose-dependent inhibitory effect on theophylline metabolism. Depending on the dose of disulfiram, a dose reduction of theophylline by as much as 50% may be necessary to minimize the risk of toxicity.