Continuous infusion of lidocaine in patients with cardiac arrhythmias. Unpredictability of plasma concentrations.

Recent studies suggest that lidocaine hydrochloride continues to accumulate during prolonged infusions. Plasma levels of lidocaine and monoethylglycinexylidide (MEGX) were measured in 26 patients with cardiac arrhythmias during lidocaine infusions of 15 to 69 hours' duration. Clearance varied, ranging from 3.2 to 14.7 mL/min/kg, and was significantly less in the ten patients with heart failure (5.8 +/- 1.7 mL/min/kg) as compared with the remaining 16 (8.4 +/- 2.6 mL/min/kg; P < .05). The MEGX levels were < 1 microgram/mL. In four patients, steady states were achieved at two different infusion rates, and changes in lidocaine plasma levels were generally proportional to changes in infusion rates. Lidocaine elimination half-lives ranged from 3.2 to 8.7 hours, and no accumulation continued beyond four half-lives. Clearance values, elimination half-lives, apparent volumes of distribution, and, consequently, steady-state levels were widely variable, which can be partly explained by the inclusion of patients with congestive heart failure. Monitoring of serum lidocaine levels may aid in individualization of therapy.

Plasma concentration and acetylator phenotype determine response to oral hydralazine.

The vasodepressor response to single and multiple oral doses of hydralazine, 1 mg/kg, was studied in hypertensive patients. The concentration of hydralazine in plasma was measured both by a newly developed specific and a nonspecific assay similar to those used in previous studies. Acetylator phenotype was determined following oral sulfamethazine. Plasma hydralazine concentration peaked at 1 hour after administration and was undetectable 2 hours later. Apparent hydralazine was present in plasma in higher concentration and for a longer duration than hydralazine. The peak decreases in blood pressure (BP) were proportional to plasma hydralazine concentration following administration of both single and multiple doses and were substantially maintained for 8 hours. In contrast there was no significant correlation between decreases in BP and apparent hydralazine concentrations. The plasma concentration of hydralazine after a standard oral dose varied by as much as 15-fold among individuals and was lower in rapid than slow acetylator phenotype patients. The BP responses were positively correlated with plasma hydralazine concentrations and inversely correlated with acetylator indices. Low plasma concentrations may account for poor responses of some patients to conventional oral doses of hydralazine. The applicability of acetylator phenotyping for individualization of hydralazine dosage regimens merits further evaluation.

N-Acetylprocainamide kinetics after single and repeated oral doses.

The kinetic behavior of N-acetylprocainamide (NAPA) was studied after single and repeated oral doses in six healthy subjects and five patients with cardiomyopathy. Renal clearance (CLR) of NAPA was lower in patients than in normal subjects after an initial 1-gm dose (1.3 +/- 0.4 [x +1- SD] and 2.7 +/- 0.4 ml . min-1 . kg-1, P less than or equal to 0.001) and a final 2 gm dose (1.6 +/- 0.4 and 2.6 +/- 0.5 ml . min-1 . kg-1, P less than or equal to 0.01) even though there was no difference between measured creatinine clearance (ClCR) 80.8 +/- 23.6 and 93.2 +/- 19.3 ml . min-1 [1.73 M2]-1). The decrease in the ratio of NAPA ClR to ClCR (R) could not be accounted for by age alone. A published regression formula overestimated the R ratio for patients (1.65 +/- 0.09 and 1.25 +/- 0.17, P less then 0.025), but accurately predicted the R ratio for healthy subjects (2.10 +/- 0.04 and 1.99 +/- 0.54). Mean steady-state concentration (normalized for daily dose per unit of body mass) after 2 gm every 8 hr for at least 3 days was higher for patients (P less than or equal to 0.05). Comparing the parameters for all subjects after the initial 1-gm dose to those after the last 2-gm dose with paired data, oral clearance was somewhat lower after the last dose (3.7 +/- 1.0 and 3.1 +/- 1.0 ml . min-1 . kg-1, P less than or equal to 0.01). In spite of this, before and 2 hr after dose steady-state NAPA serum concentration were generally proportional to dose over the concentration range studied. Net deacetylation of NAPA to procainamide in both groups was minimal.

Effect of food on blood hydralazine levels and response in hypertension.

A study with a nonspecific hydralazine assay reported that food increased hydralazine concentrations in plasma. We used a specific HPLC hydralazine assay to determine the effect of food on hydralazine blood levels and hemodynamic responses after oral hydralazine. Six subjects with uncomplicated essential hypertension were given 1 mg/kg hydralazine solution orally on two occasions at least 3 days apart. On 1 study day subjects fasted and on the other they were given a standard meal 45 min before hydralazine. Mean arterial pressure and heart rate were monitored for 2 hr before and for 4 hr after hydralazine and frequent venous blood samples were drawn for hydralazine assay. Hepatic blood flow was estimated by determination of indocyanine green clearance before food, after food, and 30 min after hydralazine. Peak blood hydralazine concentrations fell in all (46.2% +/- 11.5%; means +/- SE) and areas under the blood hydralazine concentration/time curves fell (45.7% +/- 9.5%) after food. This could not be explained by changes in liver blood flow. Food-related reductions in blood levels of hydralazine were associated with reduced vasodepressor effects (41.5% +/- 5.6%). It is possible that food increases intravascular conversion of hydralazine to hydralazine pyruvic acid hydrazone. The reduction in vasodepressor response suggests that patients with hypertension should take hydralazine at a fixed time in relation to meals.

Effect of erythromycin on carbamazepine kinetics.

Two recent reports of carbamazepine-induced intoxication during concurrent therapy with macrolide antibiotics prompted us to perform a carefully controlled two-way cross-over study in eight healthy male nonsmokers. Treatment A was 250 mg erythromycin every 6 hr for 5 days before and 3 days after 400 mg carbamazepine. Treatment B was 400 mg of carbamazepine alone. One half of the subjects received treatment A, then B, while the other half received treatment B, then A. There was a 4-wk washout period between treatments. Plasma samples obtained at various times up to 72 hr after the carbamazepine dose were assayed in duplicate by HPLC. The data were fit to a one-compartment open model with first-order absorption and elimination. Clearance of oral carbamazepine was lower in the presence of erythromycin (mean +/- SD, 0.290 +/- 0.074 and 0.360 +/- 0.072 1 X kg-1 X day-1). There were no differences in apparent volume of distribution (1.01 +/- 0.20 and 1.04 +/- 0.12 1 X kg-1), elimination rate constant (0.302 +/- 0.113 and 0.348 +/- 0.079 day-1), or absorption rate constant (14.5 +/- 8.7 and 15.5 +/- 16.6 day-1) between the two treatment groups. The decrease in clearance of oral carbamazepine secondary to erythromycin indicates that further clinical studies are warranted.

Determinants of systemic availability of oral hydralazine in heart failure.

Short-term therapy with oral hydralazine can favorably affect abnormal hemodynamics in patients with congestive heart failure, but the range of dosage is large. To investigate whether this variability in effective dose is a result of altered systemic availability, we studied 10 patients with congestive heart failure. Bioavailability (F) was calculated as the ratio of the blood AUC for a single 75 mg oral dose to the AUC of a 0.3 mg/kg iv dose. Acetylation capability was determined by sulfamethazine metabolic clearance (CLsmz). The F value in six subjects with CLsmz greater than 100 ml/min was 9.9% +/- 6.0% (means +/- SD) and was lower than the value of 26.2% +/- 13.0% (P less than 0.05) in the four patients with CLsmz less than 60 ml/min. Thus acetylation ability is an important consideration during low-dose hydralazine therapy (less than or equal to 225 mg/day). The clearance of the single intravenous dose of hydralazine averaged 29.5 +/- 8.0 ml/min/kg, which is not different than that reported in populations without heart failure. After oral dosage titration to induce maximum hemodynamic changes, the dose-normalized hydralazine AUC rose from 53.5 +/- 50.5 to 247.2 +/- 213.4 min/L X 10(3). Thus large oral doses of hydralazine result in disproportionate increases in systemic availability compatible with saturation of the first-pass effect or systemic clearance. In the doses required for maximum hemodynamic effects in our patients (225 to 3000 mg/day), this saturation phenomenon was a prominent determinant of systemic availability.

Effect of intravenous dose on hydralazine kinetics after administration.

Six male hypertensive patients, three rapid and three slow acetylators, each received four different intravenous hydralazine doses by constant infusion over 100 sec. Two to four days elapsed between doses. Plasma or whole-blood hydralazine concentrations were measured by HPLC after each dose. There was no influence of acetylator phenotype on hydralazine kinetics after intravenous dosing. There also was no consistent effect of dose size on hydralazine clearance or volume of distribution at doses up to 0.45 mg (2.3 mumol/kg). One subject, who received doses up to 0.6 mg/kg (3.05 mumol), had an apparent decrease in clearance at the higher doses. These findings are consistent with the fact that hydralazine is converted intravascularly to hydralazine pyruvic acid hydrazone and the fact that potentially saturable hepatic metabolic pathways play only a modest role in systemic clearance.

Effect of oral dose size on hydralazine kinetics and vasodepressor response.

Levels of hydralazine in blood are log-linearly related to its vasodepressor effect. We examined the effect of oral dose size on the proportion of hydralazine that reaches systemic circulation. Nine subjects with hypertension were given hydralazine in oral doses in the therapeutic range. Blood hydralazine levels, effective liver blood flow, blood pressure, and heart rate were measured. As the hydralazine dose increased, the ratios of the AUC of hydralazine to hydralazine dose and of peak blood hydralazine concentration to hydralazine dose increased, indicating an increase in the proportion of the dose in blood. Liver blood flow tended to increase (maximum 40%) as dose increased above 0.5 mg/kg. Vasodepressor response and degree of tachycardia increased disproportionately with increasing hydralazine dose. There were strong log-linear relationships between peak hydralazine levels and both vasodepressor response and tachycardia that did not change with increasing hydralazine dose. Thus blood hydralazine and vasodepressor response increase disproportionately with increasing hydralazine doses in hypertension.

Displacement of phenytoin from plasma binding sites by salicylate.

Six healthy male subjects received phenytoin sodium as 9 100-mg capsules alone or with aspirin in a randomized, crossover fashion. Aspirin, 975 mg every 6 hr, was started 22 hr before a phenytoin dose and continued for an additional 48 hr during blood sampling. Mean 4-hr plasma salicylate levels ranged from 104 to 157 micrograms/ml during the sampling period. Individual mean values for the free fraction of salicylate varied from 0.107 to 0.167. The fraction of free phenytoin in plasma rose from 0.128 +/- 0.004 to 0.163 +/- 0.009 when aspirin was given (p less than 0.001). Subjects had lower total phenytoin 48-hr area under the curve (AUC) values when on aspirin (323 +/- 36 without and 261 +/- 49 micrograms . hr . ml-1 with aspirin; p less than 0.001) but free phenytoin AUC values were unchanged (41.4 +/- 4.5 and 42.4 +/- 9.0 micrograms . hr . ml-1; p less than 0.5). Thus, more rapid clearance of total phenytoin probably compensated for salicylate displacement of phenytoin from plasma protein binding sites. Total phenytoin levels for therapeutic monitoring must be interpreted cautiously when patients also receive salicylate.

Hydralazine kinetics in hypertensive patients after intravenous administration.

Previous studies on intravenous hydralazine kinetics have been performed using nonselective analytical techniques that measure not only hydralazine but also certain hydralazine metabolites such as hydralazine pyruvic acid hydrazone (HPH). We studied the time course of hydralazine and HPH in eight hypertensive patients after 0.3 mg/kg intravenous with selective high-pressure liquid chromatographic assays. "Apparent" hydralazine concentrations were also determined using a nonselective gas-liquid chromatographic procedure. Total plasma clearance, CLT[72.9 +/- 4.9 (SEM) ml . min-1 . kg-1], apparent volume of distribution, Vd area (5.83 +/- 0.30 1 . kg-1), steady-state volume of distribution, Vd ss (1.83 +/- 0.17 . kg-1), and terminal half-life, t1/2 (53.7 min, harmonic mean) were independent of acetylator phenotype. The high ClT is compatible with rapid intravascular conversion of hydralazine to HPH and a high hepatic extraction ratio. Peak HPH concentrations occurred 10 to 60 min after dose; mean HPH t1/2 was 239 min. "Apparent" hydralazine concentrations were usually highest in the 2-min plasma sample and declined with a mean t1/2 of 296 min. Reports based on nonselective assay methods have underestimated CLT, Vd ss, and Vd area and have overestimated the t1/2 of hydralazine.

Hydralazine kinetics after single and repeated oral doses.

In reports on hydralazine kinetics plasma hydralazine levels have been measured with nonspecific assay techniques. The techniques used also include acid-labile hydralazine metabolites and therefore markedly overestimate hydralazine levels. We have developed specific, sensitive assay methods for the measurement of hydralazine and its major plasma metabolite, hydralazine pyruvic acid hydrazone (HPH). By these methods, we determined hydralazine and HPH kinetics after single and repeated oral doses of hydralazine in eight hypertensive patients. Hydralazine bioavailability in the fast acetylator group (9.5% single dose, 6.6% repeated doses) and in the slow acetylator group (31.3% single dose, 39.3% repeated doses) was phenotype dependent. Peak plasma levels were lower than those reported with nonspecific assays: 0.32 microM for the single dose and 0.14 microM for repeated doses in the fast acetylator group and 1.03 microM for the single dose and 0.96 microM repeated doses in the slow acetylator group. There was no alteration in kinetics and no cumulation in plasma on repeated administration. HPH plasma levels were proportional to those of hydralazine in both acetylator groups and were 2.5 to 4 times as high as those of hydralazine. Elimination half-lifes were phenotype independent, ranging from 4 to 6 hr. HPH cumulated in the rapid but not in the slow acetylator group after repeated doses of hydralazine.

Phenytoin cumulation kinetics.

Four male subjects were given phenytoin orally in single or twice-daily doses. Subjects were on 2 or 3 different dosing rates from 260 to 600 mg phenytoin sodium daily. Predose blood samples were obtained almost daily. The resulting serum levels, measured by gas-liquid chromatography, ranged from 1 to 18 micrograms/ml. Serum phenytoin concentration-time data were fit to a 1-compartment open model with zero-order input and Michaelis-Menten elimination. The resulting computer-generated parameter estimates (Vmax, 5.28 to 8.41 mg/kg/day; Km, 0.83 to 4.18 mg/1; Vd, 0.74 to 0.97 1/kg) are in agreement with the ranges of values in the literature. The time course of phenytoin cumulation is compatible with the presence of a major elimination pathway exhibiting Michaelis-Menten kinetic behavior.

Hemodynamic effects of N-acetylprocainamide in heart disease.

In six normal subjects and 6 patients with primary cardiomyopathy, left ventricular performance was evaluated at rest and during isometric handgrip exercise after 4 days of oral N-acetylprocainamide (NAPA) at each of the three dosage levels (3, 4, 5, and 6 gm/day). Changes in heart rate, blood pressure, and echocardiographic performance indices were noted during isometric exercise, but no effect of NAPA could be demonstrated. In five additional patients with ventricular dysrhythmias due to cardiac diseases, NAPA was given by vein until dysrhythmias were controlled and then a maintenance infusion was continued for 48 hr. Continuous ECG recordings showed excellent dysrhythmia control in four of the five patients, but no effect of NAPA on heart rate, blood pressure, mean pulmonary artery pressure, mean pulmonary artery wedge pressure, or cardiac output was demonstrated, either at the peak of initial infusion (serm NAPA 27 +/- 6.7 microgramsm/ml) or at steady state during the maintenance infusion (16 +/- 4.5 microgramm/ml). We conclude that NAPA by vein and mouth in clinically appropriate doses should be safe in patients with the reduced left ventricular performance due to cardiac disease.

Urine and plasma free phenytoin concentration correlation.

In an attempt to find a simple, noninvasive method for estimating the plasma free concentration (CPf) of phenytoin (PHT), the relationship between urine PHT concentration (Cu) and CPf was studied in 40 epileptic patients who were 6 to 50 yr old. Cu was determined by gas-liquid chromatography and CPf by equilibrium dialysis at 37 degrees. Cu was generally greater than CPf, but correlation between the two was strong (r2 = 0.876) and in the same range as previously published results for saliva and CPf. Correlation of Cu with CPf was not influenced by patient age, urine pH, or urine flow rate. In 24 patients timed urine collections were obtained and calculated renal clearances of PHT were shown to increase in proportion to urine flow rates. Although it is not known if renal impairment and albuminuria will alter the relationship between Cu and CPf it appears that Cu may be useful in estimating the concentration of pharmacologically active PHT in plasma.

Individualization of phenytoin dosage regimens.

Two methods for arriving at optimum, individual phenytoin dosage regimens have been evaluated in 12 patients. (1) Individual Michaelis-Menten pharmacokinetic parameters for phenytoin were estimated from two reliable steady-state phenytoin serum concentrations resulting from different daily doses: The observed steady-state phenytoin serum levels obtained after 3 to 8 wk of compliance with dosage regimens calculated from the individual pharmacokinetic parameters agreed well with predicted levels (r = 0.824, p less than 0.02). The average deviation between observed and predicted levels was 0.04 mug/ml (range, +/- 3.2 mug/ml). (2) A previously published nomogram for making adjustments in phenytoin dosage regimens: The serum phenytoin concentration actually expected from the dose indicated by the nomogram was calculated using individual pharmacokinetic parameters. The daily dose for one patient would have exceeded his estimated maximal rate of metabolism. The correlation between calculated and predicted phenytoin serum levels in the other 11 patients was weak but significant (r= 0.360, p less than 0.05). The average deviation was --3 mug/ml (range, 3.9 to --11.3 mug/ml). It was concluded that the use of individual pharmacokinetic parameters is practical and is also superior to the nomogram.

Rapid metabolism of phenytoin: a method of calculating proper dosage.

The optimal dosage of phenytoin can be accurately determined by a pharmacokinetic method. By plotting the rate of administration of phenytoin acid against the apparent plasma clearance rate, we estimated the maximum rate of metabolism and the serum concentration at which the rate of metabolism was one half the maximum rate for phenytoin and then applied the Michaelis-Menten equation to optimize the dosage of phenytoin in a 48-year-old man with uncontrolled idiopathic generalized seizures and increased metabolism of phenytoin. The patient became seizure free on a regimen of 650 mg of phenytoin daily and experienced no side effects of phenytoin over-dosage. The pharmacokinetic technique described is simple to use and can be applied in an outpatient clinic.

A method for the combined measurement of ethiofos and WR-1065 in plasma: application to pharmacokinetic experiments with ethiofos and its metabolites.

An analytical method for the combined measurement of ethiofos (WR-2721) and a major metabolite (WR-1065) in plasma is described. Plasma samples were subjected to conditions which quantitatively converted both ethiofos and bound WR-1065 to free WR-1065 which was subsequently separated by HPLC and detected electrochemically using established procedures. Although bound WR-1065 in plasma is thought to exist mainly in the form of mixed disulfides, the symmetrical disulfide, WR-33278, also was quantitatively converted to the free thiol form. Standard curves were linear over the range 0.10 to 25 micrograms/mL (0.75 to 186 mumol/L). Mean precision over the range was 5.4% (coefficient of variation, CV) and recoveries of various mixtures of ethiofos, WR-1065 and WR-33278 averaged 102% (CV = 6.6%). This analytical procedure and others specific for ethiofos, free WR-1065 and WR-33278 were applied to dosing experiments in which the parent drug and its major metabolites were variously administered to beagle dogs and rhesus monkeys. Following i.v. administration of ethiofos (120-150 mg per kg body weight) to monkeys, plasma concentrations of unchanged drug ranged from 477 micrograms/mL (2.23 mM) down to the minimum detectable limit of the analytical procedure (0.05 micrograms/mL, 0.23 microM) 2-3 hours postinfusion. Clearances averaged 43.5 +/- 13.4 (SD) mL min-1 kg-1 and half-lives observed in the 20-60 minute postinfusion period were 8-15 min.

Pharmacokinetic interactions of the macrolide antibiotics.

The macrolide antibiotics erythromycin and triacetyloleandomycin (troleandomycin) are prescribed for many types of infections. As such they are often added to other preexisting drug therapy. Thus, there are frequent opportunities for the interaction of these antibiotics with other drugs. Both erythromycin and triacetyloleandomycin appear to have the potential to inhibit drug metabolism in the liver and also drug metabolism by micro-organisms in the gut, either through their antibiotic effect or through complex formation and inactivation of microsomal drug oxidising enzymes. Of the two agents, triacetyloleandomycin appears to be the more potent inhibitor of microsomal drug metabolism. Published studies indicate that triacetyloleandomycin can significantly decrease the metabolism of methylprednisolone, theophylline and carbamazepine. Its ability to cause ergotism in patients receiving ergot alkaloids and cholestatic jaundice in patients on oral contraceptives may also be related to its inhibitory effect on drug metabolism. Erythromycin appears to be a much weaker inhibitor of drug metabolism. There are numerous reports describing apparent interactions of erythromycin with theophylline and a lesser number of reports dealing with carbamazepine, warfarin methylprednisolone and digoxin. There are sufficient data to suggest that erythromycin can, in some individuals, inhibit the elimination of methylprednisolone, theophylline, carbamazepine and warfarin. The mean change in drug clearance is about 20 to 25% in most cases, with some patients having a much larger change than others. Like tetracycline, erythromycin also appears to have the potential for increasing the bioavailability of digoxin in patients who excrete high amounts of reduced digoxin metabolites, apparently through destruction of the gut flora that form these compounds. Concurrent administration of triacetyloleandomycin with drugs whose metabolism is known to be affected or that could potentially be affected should be avoided unless appropriate adjustments in dosage are made. Coadministration of erythromycin with drugs believed to interact should be undertaken with caution and with appropriate patient monitoring. Among the other macrolide antibiotics, josamycin has seldom been involved in causing drug interactions, while midecamycin and the older derivative spiramycin have not so far been incriminated.

Nonlinear pharmacokinetics: clinical Implications.

Nonlinear pharmacokinetics (in other words, time or dose dependences in pharmacokinetic parameters) can arise from factors associated with absorption, first-pass metabolism, binding, excretion and biotransformation. Nonlinearities in absorption and bioavailability can cause increases in drug concentrations that are disproportionately high or low relative to the change in dose. One of the more important sources of nonlinearity is the partial saturation of presystemic metabolism exhibited by such drugs as verapamil, propranolol and hydralazine. In such cases, circulating drug concentrations are sensitive not only to dose size but also to rate of absorption: slower absorption may decrease the overall systemic availability. The binding of drugs to plasma constituents, blood cells and extravascular tissue may exhibit concentration dependence. This can cause pharmacokinetic parameters based on total blood or serum drug concentrations to be concentration-dependent. Often, in these cases, parameters based on free drug concentration appear linear. An important consideration in regard to concentration-dependent serum binding is the difficulty in relating total concentration to a usual therapeutic range if free concentration is a better indicator of drug effect. Measurement of free concentration is needed in these cases, particularly if the intersubject variability in binding is high. An example of this behaviour is valproic acid. Partial saturation of elimination pathways can result in the well known behaviour typical of Michaelis-Menten pharmacokinetics. Small changes in dosing rate can make much larger differences in steady-state concentration. The time to achieve a given fraction of steady-state becomes longer as the dosing rate approaches the maximum elimination rate. Alcohol and phenytoin are examples of drugs that exhibit such behaviour. The sensitivity of steady-state concentration and cumulation rate to changes in dosing rate are both influenced by the magnitude of parallel first-order elimination pathways: even a first-order pathway that is only 1 to 2% of maximum clearance (which occurs at very low concentration) can be an important determinant of steady-state concentration and cumulation rate when concentrations are high. Theophylline and salicylate have significant parallel first-order elimination pathways as well as saturable pathways. Autoinduction causes an increase in clearance with long term administration. In some cases, dosage adjustment must be made to compensate for the increase, and the possibility that the degree of induction may be dose- or concentration-dependent must be kept in mind. Carbamazepine is a major example of autoinduction. It is fortunate that only a few of the many hundreds of drugs in use exhibit nonlinear behaviour that leads to clinical implications.(ABSTRACT TRUNCATED AT 400 WORDS)

Pharmacokinetics of quinidine in male patients. A population analysis.

Quinidine pharmacokinetic behaviour was evaluated in 139 adult hospitalised men receiving oral quinidine therapy. A total of 391 serum quinidine concentrations were measured by enzyme immunoassay for routine clinical purposes. The NONMEM programme was used to examine the relationship between quinidine pharmacokinetics and several potential covariates. A 1-compartment open model with first-order absorption and elimination was assumed. The mean apparent volume of distribution (Vd) was about 230L. When measured, alpha 1-acid glycoprotein (AAG) levels were not included in the analysis. Oral quinidine clearance (CL) decreased with age, severe congestive heart failure and renal disease, and increased in patients with a history of alcohol abuse. The interpatient variability in CL and the intrapatient residual variability expressed as coefficients of variation (CV) were 28 and 31%, respectively. When AAG values were incorporated into the analysis, the only important covariates of CL were the AAG measurements and the presence of renal dysfunction as indicated by a calculated creatinine clearance of less than 50 ml/min (3 L/h). The interpatient variability in CL and the residual intrapatient CVs decreased to approximately 24 and 26%, respectively. Improvement of the CL model by inclusion of measured AAG strongly suggests that quinidine elimination is dependent on the free concentration of drug in plasma and supports the use of free serum quinidine concentrations when evaluating and monitoring quinidine therapy.