Exposure-response modeling and clinical trial simulation of the effect of tolterodine on QT intervals in healthy volunteers.

The objective of this analysis was to explore exposure-response modeling of data from a thorough QT (TQT) study of tolterodine in CYP2D6 extensive (EMs) and poor metabolizers (PMs). Crossover treatments of the TQT study included the recommended (2 mg twice daily) and supratherapeutic (4 mg twice daily) doses of tolterodine, moxifloxacin (400 mg once daily), and placebo. The concentration-response relationships for the QTc effects of moxifloxacin and tolterodine were described using a linear model with baseline effect, placebo effect, and a drug effect. The mixed effects modeling approach, using the first order conditional estimation method, was implemented in NONMEM. Simulated data from 250 trial replicates were used for limited predictive check and to describe the expected extreme responders in this study, under the derived model and point estimates. Modeling results for tolterodine showed linear concentrationdependent increases in QTc interval, with no difference in slopes between EMs and PMs. Modelpredicted QTc prolongations for tolterodine and moxifloxacin were consistent with their respective observed mean results. No subjects were predicted to have increases of > 60 milliseconds (ms); the predicted incidence of borderline QTc increases (> 30 and ≤ 60 ms) remained low at the supratherapeutic tolterodine dose in both PMs (9.1%) and EMs (3.9%). In conclusions, this analysis supports our clinical experience that tolterodine does not have a clinically significant effect on QT interval.

Phase I and pharmacologic study of the tyrosine kinase inhibitor SU101 in patients with advanced solid tumors.

PURPOSE: To evaluate the clinical feasibility and pharmacologic behavior of the platelet-derived growth factor (PDGF) tyrosine kinase inhibitor SU101, administered on a prolonged, intermittent dosing schedule to patients with advanced solid malignancies. PATIENTS AND METHODS: Twenty-six patients were treated with SU101 doses ranging from 15 to 443 mg/m(2) as a 24-hour continuous intravenous (IV) infusion weekly for 4 weeks, repeated every 6 weeks. Pharmacokinetic studies were performed to characterize the disposition of SU101 and its major active metabolite, SU0020. Immunohistochemical staining of PDGF-alpha and -beta receptors was performed on malignant tumor specimens obtained at diagnosis. RESULTS: Twenty-six patients were treated with 52 courses (187 infusions) of SU101. The most common toxicities were mild to moderate nausea, vomiting, and fever. Two patients experienced one episode each of grade 3 neutropenia at the 333 and 443 mg/m(2) dose levels. Dose escalation of SU101 above 443 mg/m(2)/wk was precluded by the total volume of infusate required, 2.5 to 3.0 L. Individual plasma SU101 and SU0020 concentrations were described by a one-compartment model that incorporates both first-order formation and elimination of SU0020. SU101 was rapidly converted to SU0020, which exhibited a long elimination half-life averaging 19 +/- 12 days. At the 443 mg/m(2)/wk dose level, trough plasma SU0020 concentrations during weeks 2 and 4 ranged from 54 to 522 micromol/L. Immunohistochemical studies revealed PDGF-alpha and -beta receptor staining in the majority (15 of 19) of malignant neoplasms. CONCLUSION: SU101 was well tolerated as a 24-hour continuous IV infusion at doses of up to 443 mg/m(2)/wk for 4 consecutive weeks every 6 weeks. Although further dose escalation was precluded by infusate volume constraints, this SU101 dose schedule resulted in the achievement and maintenance of substantial plasma concentrations of the major metabolite, SU0020, for the entire treatment period.

Concentration-dependent tubular secretion of acetazolamide and its inhibition by salicylic acid in the isolated perfused rat kidney.

An isolated perfused rat kidney (IPK) technique was used to study the effect of salicylic acid (SA) on the excretion of acetazolamide (AZ). Initial experiments were conducted in the absence of interactants at three nominal AZ concentrations (50, 100, and 250 micrograms/ml). Over the concentration range studied, AZ demonstrated net tubular secretion in the IPK. Significant decreases in excretion ratio (4.97 +/- 0.79-2.66 +/- 1.1) and secretory clearance (0.809 +/- 0.23-0.541 +/- 0.28) were observed with increasing AZ concentration, consistent with saturation of tubular secretion. Using a facilitated model for renal secretion, values of tubular transport parameters were obtained from a plot of excretion ratio vs. unbound AZ concentration: tmax = 118 +/- 29.4 micrograms/min, KM = 53.4 +/- 22.4 micrograms/ml, and tmax(A) = 6.31 +/- 2.82 micrograms/min. In the presence of SA (200 micrograms/ml), renal secretion of AZ was inhibited, as demonstrated by significant decreases in renal clearance (0.731 +/- 0.21-0.147 +/- 0.03) and excretion ratio (3.77 +/- 0.82-0.378 +/- 0.07). Although these findings were indicative of a reabsorption component to AZ excretion in the IPK that had not been previously proposed, the results were consistent with a previous investigation of concomitant administration of AZ and SA in humans (Br. J. Clin. Pharmacol. 27, 866, 1989), thereby endorsing utilization of the IPK as a screening tool for renal clearance mechanisms in humans.

Renal disposition and drug interaction screening of (-)-2'-deoxy-3'-thiacytidine (3TC) in the isolated perfused rat kidney.

PURPOSE: Dideoxynucleoside bases are used for the treatment of acquired immune deficiency syndrome (AIDS), acting by inhibiting reverse transcriptase and preventing human immunodeficiency virus (HIV) replication. Currently, AZT (zidovudine), ddC (zalcitibine), and ddI (didanosine) are available to the medical community to prevent the onset of AIDS in HIV-infected individuals. 3TC (-)-2'-deoxy-3'-thiacytidine, lamivudine), a new dideoxynucleoside base, is currently undergoing Phase II/III trials, and has exhibited anti-HIV replication activity, a favorable adverse event safety profile, and is eliminated via renal mechanisms. Concomitantly administered drugs could potentiate the effects of 3TC due to interaction in the kidney. METHODS: An isolated perfused rat kidney (IPK) technique was used to screen several clinically relevant drugs for potential interaction with 3TC. The following perfusions were performed: baseline 3TC; and 500 ng/mL 3TC with clinically relevant concentrations of AZT, ddC, ddI, probenecid, trimethoprim, sulfamethoxazole, ranitidine, and cimetidine. RESULTS: Renal clearance of 3TC was nonlinear between 500 and 5000 ng/mL, decreasing from 3.06 to 1.74 mL/min. Excretion ratio also decreased, from 3.67 (500 ng/mL) to 2.49 (5000 ng/mL), consistent with a decrease in 3TC secretion. AZT, ddI, and ddC elicited no or minimal effects on 3TC elimination at the concentrations studied. However, trimethoprim caused significant reductions in 3TC elimination parameters: clearance and excretion ratio decreased to 1.25 mL/min and 1.43, respectively. CONCLUSIONS: These results indicate that caution should be exercised when the combination of 3TC and trimethoprim are administered to AIDS patients.

The influence of protein binding on the elimination of acetazolamide by the isolated perfused rat kidney: evidence of albumin-mediated tubular secretion.

The impact of albumin on the renal elimination of acetazolamide, a low extraction ratio compound, was investigated in the isolated perfused rat kidney. Perfusion studies were conducted over a wide range of protein concentrations (0.25, 0.5, 0.75, 1.0, 4.0 and 6.0 g/100 ml) and an initial drug concentration of 100 micrograms/ml. Kidney viability was within normal limits among all treatment groups. Over the range of albumin levels studied, an approximate 3.4-fold increase in drug-free fraction effected a 2.8-fold increase in renal clearance. Although this finding contradicted conventional wisdom regarding extraction ratio and renal elimination, the results were consistent with a proposed ancillary role of albumin in renal tubular transport processes. An alternative clearance model was developed, analogous to earlier models of hepatic elimination. The facilitated renal clearance model utilized and validated in this investigation represents a composite of previously proposed theories, modified to account for albumin-mediated tubular secretion.

Comparison of the pharmacodynamic activity of cefotaxime plus metronidazole with cefoxitin and ampicillin plus sulbactam.

STUDY OBJECTIVE: To compare the pharmacokinetic and pharmacodynamic activity of three drug regimens: cefotaxime plus metronidazole, cefoxitin, and ampicillin-sulbactam against two organisms frequently isolated in intraabdominal infection, Escherichia coli and Bacteroides fragilis. DESIGN: Open-label, three-way crossover study. SETTING: Hartford Hospital Clinical Research Center. PARTICIPANTS: Nine healthy volunteers. INTERVENTIONS: Subjects received the following regimens: (1) a single 1-g intravenous dose of cefotaxime plus a single 500-mg oral dose of metronidazole; (2) two intravenous doses of cefoxitin, 2 g each dose given every 6 hours; and (3) two intravenous doses of ampicillin-sulbactam, 3 g each dose given every 6 hours. MEASUREMENTS AND MAIN RESULTS: Serum bactericidal titers and drug concentrations were measured over a 12-hour period. The cefotaxime-metronidazole regimen showed superior activity against E. coli compared with ampicillin-sulbactam and cefoxitin. The mean areas under the bactericidal activity curve (AUBC) for the three regimens were 550.2, 68.7, and 48.9, respectively (p = 0.0001). There was no significant difference in AUBC among the three regimens for B. fragilis. Serum concentrations of cefotaxime remained above the minimum inhibitory concentration (MIC) for E. coli significantly longer than did concentrations of ampicillin-sulbactam and cefoxitin (p = 0.0002 and p = 0.0023, respectively). Serum concentrations of metronidazole were still at 9 times the MIC for B. fragilis at the end of the 12-hour dosing interval; for ampicillin-sulbactam and cefoxitin concentrations remained above the MIC for one-half and less than one-fourth, respectively, of the dosing interval (p < 0.0001). The ratio of AUC:MIC was also favorable for metronidazole (212.2) compared with 63.4 for ampicillin-sulbactam and 9.2 for cefoxitin. CONCLUSIONS: The combination of cefotaxime-metronidazole, even at the relatively low doses used in this study, provides coverage against gram-negative and anaerobic pathogens that is at least as effective as that of cefoxitin and ampicillin-sulbactam. In addition, its cost is considerably less expensive than that of the other regimens.

Loss of bactericidal activities of quinolones during the post-antibiotic effect induced by rifampicin.

The post-antibiotic effect (PAE) is the phenomenon of persisting suppression of bacterial growth as a result of prior antimicrobial exposure. In antimicrobial therapy, multiple doses of either a single drug or a combination of drugs are common. Accordingly, the following question may arise: what impact might the PAE induced by the previous dose impose on the subsequent bactericidal action of a cycle-specific antibiotic. To answer the question, a study was conducted using pefloxacin, norfloxacin, ciprofloxacin and ofloxacin as test drugs, rifampicin as the PAE inducer and Escherichia coli ATCC 25922 as the test organism. Bacterial kill kinetics were determined for each quinolone at 4 x MIC before rifampicin treatment and during the PAE period following rifampicin challenge. The relative bactericidal activity of the quinolones during the PAE period was calculated. During the PAE period, pefloxacin, norfloxacin, ciprofloxacin and ofloxacin displayed only 6%, 8%, 7% and 33% of their normal activities, respectively. The results were compared to those obtained at low temperature (4 degrees C) and when the cells were challenged simultaneously with a quinolone and with rifampicin. The findings of this study suggest that prolonging the dosing interval by taking the PAE into account may not only lower the cost of antimicrobial therapy and the risk of toxicity, but also ensure the efficacy of subsequent doses.

Favorable amphiphilicity of nimodipine facilitates its interactions with brain membranes.

Nimodipine is a 1,4-dihydropyridine (DHP) calcium channel blocker which is used in the treatment of neurological deficits associated with subarachnoid hemorrhage. Small angle x-ray diffraction, differential scanning calorimetry, and equilibrium and kinetic binding techniques were used to study the interaction of nimodipine with bovine brain phosphatidylcholine (BBPC) membranes of varying cholesterol content. At concentrations (5 x 10(-10) M) near its Kd, the membrane partition coefficient of nimodipine was inversely related to the cholesterol to phospholipid (C:P) mole ratio in both model and native (rat synaptoneurosome) membranes. The nonspecific dissociation rate of nimodipine from BBPC was significantly slower at low C:P mole ratio (0.1:1) than at high C:P mole ratio (0.6:1). Calorimetric analysis showed that nimodipine decreased both the main phase transition temperature and cooperative unit size of melt for dimyristoyl phosphatidylcholine, dependent on membrane cholesterol content. Small angle x-ray diffraction analysis showed that nimodipine occupies a position in BBPC approx +/- 15 A from the center of the hydrocarbon core, near the hydrocarbon core/water interface. These data indicate that nimodipine is an amphiphilic molecule which rapidly washes out of and transports across membrane bilayers, facilitating its interactions with membranes and possibly its transport across the blood-brain barrier.

Characterization of methotrexate elimination and interaction with indomethacin and flurbiprofen in the isolated perfused rat kidney.

The mechanism of methotrexate (MTX) renal elimination and the interaction of MTX with the nonsteroidal anti-inflammatory drugs, indomethacin and flurbiprofen, were characterized in the isolated perfused rat kidney. Perfusion studies elucidating MTX renal disposition were performed in perfusate (4.0% bovine serum albumin) at initial MTX concentrations of 1, 5, 12 and 25 micrograms/ml and in controls without MTX. Interaction studies were performed at clinically relevant indomethacin and flurbiprofen concentrations of 2.5 and 10 micrograms/ml, respectively, and a MTX concentration of 25 micrograms/ml (representative of an oncolytic MTX dose). MTX unbound fractions were concentration and interactant independent. Kidney viability was within normal limits for this technique among all perfusions. MTX renal clearance was nonlinear, increasing from 0.310 to 0.434 ml/min over the concentration range studied. Corresponding excretion ratios increased from 0.933 to 1.52, indicative of MTX renal elimination involving the processes of filtration, secretion and reabsorption. Excretion ratio results were supported by tubular transit rate calculations. Interaction studies indicated that there was secretory inhibition of MTX as evidenced by a decrease in both excretion ratio and tubular clearance at 25 micrograms/ml of MTX. Therefore, the secretory component was significantly inhibited by indomethacin and flurbiprofen after concomitant administration of oncolytic doses of MTX.

Lomefloxacin concentrations in bone after a single oral dose.

We studied the penetration characteristics of lomefloxacin in bone in 30 patients with osteoarthritis undergoing total hip replacement. Patients were given a single oral 400 mg dose at various times from 1 to 12 hours prior to removal of bone samples. The peak plasma and bone (subchondral bone from femoral head) concentrations reached approximately 4.0 micrograms/mL at 2 hours post-dose and 3.0 micrograms/mL at 3 hours post-dose, respectively. At 12 hours post-dose both plasma and bone concentrations were still greater than 1.0 microgram/mL. Two hours after dosing the average bone-to-plasma ratio was greater than 0.6. These data indicate that a single 400 mg oral dose of lomefloxacin attains bone concentrations that are above its usual minimum inhibitory concentrations for susceptible organisms.

Quantification of in-vitro post-antibiotic effect based on the mean recovery-time. I: Theoretical perspectives and a practical procedure.

An empirical description of the post-antibiotic effect has been previously proposed on the basis of the time required for the bacterial density to increase by one log10 following antimicrobial exposure. In this study, a set of theoretical perspectives are established and the post-antibiotic effect quantitatively defined based on the mean-recovery time. The mean recovery-time can be determined by measuring the area above the growth curve of a bacterial population following antimicrobial exposure. It can be shown that the previously proposed description of the post-antibiotic effect may be regarded as a special case of the definition based on the mean-recovery. Furthermore, a practical procedure is described for the quantitation of the post-antibiotic effect from a given set of data in the form of a growth curve.

Quantification of in-vitro post-antibiotic effect based on the mean recovery-time. II: A comparison of colony counting versus photometric methods for the determination of bacterial growth.

In the preceding paper, a method for the quantification of the post-antibiotic effect (PAE) has been developed based upon the mean recovery-time of an antibiotic-treated bacterial population, showing that the mean recovery-time is determined by the area above a growth curve of the bacterial population. Such an area determination of PAE implies that this method is less likely to be influenced by the early portion of a growth curve and therefore less dependent on the ability of a technique to detect a minimum bacterial density (sensitivity limit). In this study, this implication is further demonstrated in theory and through experimentation. As a result, an approach to estimation of contribution of the early portion of a growth curve to the quantitation of PAE based on the mean recovery-time is derived. The PAE of ciprofloxacin on Escherichia coli ATCC 25922 was determined under different test conditions using colony count and optical density measurements for determining bacterial growth. Differences in durations of the PAE quantitated by the mean recovery-time from the colony count measured and optical density measured growth curves were not statistically significant (P greater than 0.05).

Teratogenicity of Ni2+ in Xenopus laevis, assayed by the FETAX procedure.

The teratogenicity of Ni2+ was tested by the FETAX (Frog Embryo Teratogenesis Assay: Xenopus) procedure in the South African frog, Xenopus laevis. In seven assays, beginning at 5 h postfertilization, groups of Xenopus embryos were incubated for 96 h in media that contained Ni2+ (added as NiCl2) at concentrations ranging from 1 x 10(-7) to 3 x 10(-3) mol/L; control groups were incubated in the same medium without added NiCl2. At 101 h postfertilization, surviving embryos were counted, fixed in formalin, and examined by microscopy to determine their developmental stages, malformations, and head-to-tail lengths. In control embryos, survival was greater than or equal to 95% and malformations were less than or equal to 7%. Malformations were found in greater than 95% of embryos exposed to Ni2+ concentrations greater than or equal to 5.6 mumol/L. The most frequent malformations in Ni(2+)-exposed embryos were ocular, skeletal, and intestinal deformities; less common malformations included facial, cardiac, and integumentary deformities. Other abnormalities, not categorized as malformations, included stunted growth, dermal hypopigmentation, and coelomic effusions or hemorrhages. The median embryolethal concentration (LC50) of Ni2+ was 365 (SE +/- 9) mumol/L; the median teratogenic concentration (EC50) was 2.5 (SE +/- 0.1) mumol/L; the Teratogenic Index (TI = LC50/EC50) was 147 (SE +/- 5), indicating that Ni2+ is a potent teratogen for Xenopus laevis. Experiments in which Ni(2+)-exposures were limited to specific 24 h periods showed that Xenopus embryos were most susceptible to Ni(2+)-induced malformations on the second and third days of life, during the most active period of organogenesis.

Influence of sucralfate on trilisate bioavailability.

Twelve healthy human volunteers were randomized to receive either choline-magnesium trisalicylate (CMT) 1.5 g orally every 12 hours or a combination of CMT plus sucralfate 1 g orally every 6 hours for 5 days before the blood sampling day. After a 3-week washout period, the subjects were crossed over to receive the alternate treatment for 5 days. The mean (+/- SD) area under the curve was 2668 (729) mg - hr/L and 2748 (716) mg - hr/L for CMT and CMT/sucralfate treatments, respectively. Mean (+/- SD) maximum concentration was 275 (69) mg/L and 283 (75) mg/L for CMT and CMT/sucralfate administrations, respectively. Mean (+/- SD) time to maximum concentration for CMT and CMT/sucralfate treatments was 1.8 (0.6) hours and 1.7 (0.7) hours, respectively. There were no significant differences detected for any parameter, therefore sucralfate does not affect rate or extent of CMT absorption.

Bioavailability and pharmacokinetics of ofloxacin in healthy volunteers.

The pharmacokinetics and bioavailability of ofloxacin in 20 healthy male volunteers were studied in an open-label, randomized, two-way crossover study. Ofloxacin (400 mg) was administered either as a 1-h infusion or as an oral tablet. The mean peak concentration after intravenous infusion was 4.30 +/- 0.69 microgram/ml, and that after oral administration was 3.14 +/- 0.53 microgram/ml, occurring 1.74 +/- 0.57 h after dosing. The bioavailability (F) of the oral dosage form of ofloxacin was virtually identical to that of the intravenous form (F = 105% +/- 7%). This complete bioavailability of ofloxacin is supportive of the use of the oral dosage form for the treatment of infections in hospitalized patients either as a replacement for intravenous ofloxacin therapy or in streamlining therapy from the intravenous to the oral route.

Platinum in blood mononuclear cells from patients after cisplatin therapy.

The feasibility of monitoring cisplatin chemotherapy by measuring platinum (Pt) concentrations in blood mononuclear cells was tested in six patients (four with squamous cell carcinomas of the head and neck, one with non-Hodgkin's lymphoma, and one with lung carcinoma). Blood samples (20 to 40 mL) were collected at intervals from 6 min to 21 days after an iv infusion of cisplatin (80 or 100 mg per m2). Blood mononuclear cells were harvested on a Ficoll-Hypaque gradient, washed repeatedly, counted, and homogenized by sonication in 0.5 mL of saline solution. Pt was analyzed in duplicate 40 microL samples of plasma (N = 26) or cell homogenates (N = 23) by electrothermal atomic absorption spectrophotometry with Zeeman background correction. Immediately after the cisplatin infusion, plasma Pt concentrations averaged 2.6 (SD +/- 0.2) mg per L and mononuclear cell Pt concentrations averaged 2.5 +/- 0.5 ng per 10(6) cells. At 24 to 26 hr post-infusion, plasma Pt concentrations averaged 1.6 +/- 0.2 mg per L and mononuclear cell Pt concentrations averaged 2.3 +/- 0.6 ng per 10(6) cells. Plasma Pt disappearance followed two-compartment kinetics in all patients; the plasma Pt disappearance half-time (T1/2, mean +/- SD) was 148 +/- 41 hours. The Pt concentrations in blood mononuclear cells diminished gradually during the period of observation; the T1/2 could not be reliably determined, but was estimated to be longer than two weeks. This study shows that Pt can be measured in mononuclear cells of patients after cisplatin treatment and that Pt disappears more slowly from blood mononuclear cells than from plasma of these patients.

Teicoplanin pharmacokinetics in healthy volunteers after administration of intravenous loading and maintenance doses.

Teicoplanin is an investigational glycopeptide antibiotic that is structurally and microbiologically similar to vancomycin. Since teicoplanin possesses a very long elimination half-life, the manufacturer suggests that the drug be administered every 12 h for the first day of therapy and once daily thereafter. We studied the multiple-dose (6 mg/kg per dose) pharmacokinetics of teicoplanin in volunteers following intravenous administration every 12 h for 5 days and then every 24 h for 9 days in an attempt to identify the optimal duration of the every-12-h loading-dose regimen. Multiple serum samples were obtained throughout the study, including intensive sampling after the first and last doses; urine was collected during the entire study. A three-exponential equation was fitted to the serum concentration data. The mean terminal-phase half-life was 157 +/- 93 h. Concentrations of teicoplanin in serum similar to those observed after the administration of the last dose (day 14) were observed following the fourth or fifth dose given every 12 h. Therefore, it is suggested that for clinical dosing regimens for teicoplanin, dosing every 12 h for approximately 48 h should be used, followed by once-daily dosing thereafter.

Absorption of ciprofloxacin administered through a nasogastric or a nasoduodenal tube in volunteers and patients receiving enteral nutrition.

The absorption of ciprofloxacin was higher when administered through a nasoduodenal tube than through a nasogastric tube, indicating that even though acidic gastric pH is needed for rapid disintegration, dissolved ciprofloxacin might not be stable in the gastric environment. If the length of exposure or the different gastric environment in each individual affects the overall absorption of ciprofloxacin, this could explain the reported variability in ciprofloxacin absorption and suggests the need for the development of an enteric-coated tablet. Further studies are needed to fully characterize the absorption of ciprofloxacin in patients with different illnesses, gastric transit times, gastrointestinal environments and of different ages.

Relative bioavailability in healthy volunteers of ciprofloxacin administered through a nasogastric tube with and without enteral feeding.

The bioavailability of ciprofloxacin after its administration through a nasogastric (NG) feeding tube was studied in six healthy volunteers. Each subject received, on separate occasions, an intact 750-mg ciprofloxacin tablet, a crushed tablet as a suspension through an NG tube, and a crushed tablet as a suspension through an NG tube while receiving enteral feeding. No statistically significant differences were observed in the area under the curve, maximum concentration in serum, and time to peak concentration among these three modes of administration. These findings suggest that ciprofloxacin is well absorbed after administration via an NG tube (compared with an orally administered intact tablet) even in the presence of enteral feeding.

Differential effects of flurbiprofen and aspirin on acetazolamide disposition in humans.

The plasma concentration-time profile of acetazolamide (AZ) following an intravenous bolus dose (5 mg kg-1) was determined during control, aspirin and flurbiprofen (FLU) treatment periods. The unbound fraction of AZ in plasma increased three-fold in the presence of salicylate (SA) while, in contrast, FLU produced consistent, but statistically insignificant, increases in binding. SA caused a two-fold decrease in both unbound AZ renal clearance and apparent volume of distribution at steady-state, while FLU produced a small, but significant, increase only in the latter. The area under the concentration-time curve for AZ in erythrocytes was increased by about 40% during SA treatment while FLU had no effect. Our results suggest that on a pharmacokinetic basis FLU may be a safer nonsteroidal anti-inflammatory drug (NSAID) to co-administer with AZ.