Prem K. Narang

Increased Plasma Rifabutin Levels with Concomitant Fluconazole Therapy in HIV-Infected Patients

Chemoprophylaxis for opportunistic infections associated with the human immunodeficiency virus (HIV) is increasingly common; clinical studies support the administration of drugs to prevent Pneumocystis carinii pneumonia [1-3], disseminated Mycobacterium avium complex infection [4], cytomegalovirus infection [5], and fungal infections [6]. Because these agents are often administered concurrently in patients infected with HIV, many questions have been raised about the pharmacokinetic or pharmacodynamic consequences of the drugdrug interactions that may occur. Such interactions may also confound our understanding of the outcomes seen in large clinical trials. Two drugs that are often used concurrently in patients infected with HIV are rifabutin, for the prevention of M. avium complex bacteremia [4], and fluconazole, for the prevention of fungal infections [6]. Rifabutin is an antimicrobial agent similar in structure to rifampin. Fluconazole, which is used to treat cryptococcal meningitis and oropharyngeal and esophageal candidiasis [7], has been reported to be effective for the primary prevention of deep and superficial fungal infections in HIV-infected patients whose CD4 lymphocyte counts are less than 50 cells/mm3 [6]. Fluconazole and a related azole, ketoconazole, are potent inhibitors of hepatic microsomal enzymes, especially the cytochrome p450 3A group [8]. Inhibition of these enzymes has, in turn, been shown to cause clinically significant increases in circulating levels of concomitant drugs that are metabolized by these enzymes [9-15]. Our study was designed to assess a possible mechanism for the changes observed in the toxicity and efficacy of rifabutin with concomitant fluconazole therapy. We report the results of a steady-state pharmacokinetic and safety study of rifabutin and fluconazole during concurrent zidovudine therapy in HIV-infected persons. Methods Study Design This was a phase 1, open-label pharmacokinetic and safety study of 13 persons infected with HIV who were receiving maintenance therapy with zidovudine, 100 mg five times per day. The study enrolled HIV-infected adults who had CD4 lymphocyte counts between 200 and 500 cells/mm3, had no active disease by chest radiograph, had no clinically significant hepatic or renal impairment, and were receiving no other antiretroviral therapy or concomitant medications known to substantially modulate hepatic or renal function. Persons were excluded from participation if they had a history of known hypersensitivity to the study medications, had previously received treatment with cytolytic agents or radiation therapy, had had blood transfusion within 1 week of study entry, had received treatment with rifabutin or rifampin within 3 months of study entry, or had received treatment with fluconazole or other azole drugs within 4 weeks of study entry. Pregnant or lactating women were also excluded. Whenever possible, other concomitant medications were maintained at constant doses throughout the study. The study was approved by the Georgetown University Medical Center Institutional Review Board; each participant gave written informed consent before study entry. A medical evaluation was done within 1 week of study entry. Fluconazole, 200 mg, was administered orally every 24 hours beginning on day 3. On day 16, when fluconazole had reached steady state, participants returned to the outpatient clinic, where blood was drawn just before the morning doses of fluconazole and zidovudine were given. Blood and urine samples were collected during the 24 hours after drug administration. Beginning on day 17, oral rifabutin, 300 mg/d, was added to the fluconazole-zidovudine regimen. All study medications were administered concurrently. On day 30, when rifabutin had reached steady state, study participants returned to the clinic for serial blood and urine collections. Finally, fluconazole therapy was discontinued on day 31; rifabutin and zidovudine were continued for the remaining 2 weeks of the study. On day 44, participants returned to the clinic and again had serial blood and urine collections. Study participants received medical evaluations with routine laboratory testing to evaluate the safety of these therapies on days 16, 30, and 44. Each participant also returned to the outpatient clinic on the mornings of days 1, 15, 29, and 43 to provide an additional blood sample before receiving medication to estimate within-person variation in the trough concentrations of the appropriate study drugs. Drug Supply and Analysis Rifabutin was supplied as 150-mg capsules by Pharmacia, Inc. (Columbus, Ohio); fluconazole was supplied as 200-mg tablets by Pfizer, Inc. (Groton, Connecticut); and zidovudine was supplied as 100-mg capsules by Burroughs-Wellcome Company (Research Triangle Park, North Carolina). We collected all blood samples in heparinized tubes and promptly centrifuged them to separate the plasma. Plasma specimens were frozen at 20 C until they were assayed. Urine collection bottles were kept on ice or were refrigerated during the collection periods. A 10-mL aliquot of urine was placed in cryotubes and kept frozen at 20 C until it was assayed. Plasma and urine concentrations of fluconazole, rifabutin, and the 25-desacetyl metabolite of rifabutin, LM565, were determined by Harris Laboratories (Lincoln, Nebraska) using a validated high-performance liquid chromatography method as previously described [16, 17]. The respective interday precision (expressed as a percentage of relative standard deviation) and inaccuracy estimates for the quantity of fluconazole were 8% or less and 5%, respectively; those for rifabutin and LM565 were 10% or less and 5%, respectively. Pharmacokinetic Analysis Pharmacokinetic variables were estimated by noncompartmental analyses [18]. Steady-state estimates of the area under the concentration-time curve for rifabutin and fluconazole were obtained by using linear trapezoidal integration over a dosing interval. Renal clearance was estimated by dividing the amount excreted in the urine by the area under the plasma drug concentration-time curve. Statistical Analysis Statistical analyses were done using Statistical Analysis Systems software, version 6.06 (SAS Institute, Cary, North Carolina). Estimates of pharmacokinetic variables from the 12 evaluable participants in the presence or absence of the study drugs were compared using a paired, two-tailed t-test. Values are given as mean SD. Results Thirteen persons were enrolled in the study. Data from 12 participants were considered evaluable for data analysis; studies were discontinued in 1 participant before completion because of the development of a diffuse maculopapular rash 12 days after rifabutin therapy began. Data were not stratified for sex (9 men, 4 women) or race (9 white participants, 4 black participants) because of the small sample sizes within each group. Other demographic information included age (35.8 7.2 years), body weight (89 20 kg), CD4 lymphocyte count (369.3 63.1 cells/mm3), aspartate aminotransferase level (0.56 0.16 kat/L), and serum creatinine level (106 8 mol/L). No clinically significant changes in the results of physical examinations or laboratory evaluations were seen during the study in any of the evaluable study participants. Figure 1 shows the plasma concentrations of rifabutin and LM565 as a function of time from study days 30 (rifabutin and fluconazole) and 44 (rifabutin alone). Rifabutin levels were significantly higher during concurrent fluconazole treatment; the steady-state estimate of the area under the concentration-time curve increased 82% (5442 2404 ng h/mL compared with 3025 1117 ng h/mL; P 0.05). The area under the LM565 concentration-time curve over the 24-hour dosing interval increased 216% (959 529 ng h/mL compared with 244 141 ng h/mL; P 0.05). This finding was consistent among all study participants (Figure 2). Urinary excretion of rifabutin and LM565 also increased during concurrent fluconazole treatment. The amounts of rifabutin and LM565 excreted on day 30 compared with day 44expressed as a percentage of the rifabutin dosewere 2.5% 1.5% compared with 6.2% 2.0% (P < 0.01) and 0.8% 0.4% compared with 2.3% 0.9% (P < 0.01), respectively. The renal clearance of rifabutin, however, was unchanged (0.0502 0.0199 L/h kg1 and 0.0446 0.0248 L/h kg1). Figure 1. Steady-state concentration compared with time curves of the plasma concentrations (mean SD) of rifabutin and its 25-desacetyl metabolite, LM565, over one dosing interval. Figure 2. Area under the plasma concentration-time curve for rifabutin and its 25-desacetyl metabolite, LM565, when rifabutin is administered alone and in combination with fluconazole. The steady-state fluconazole plasma concentration did not change (area under the concentration-time curve over a dosing interval without rifabutin, 201.0 36.2 g h/mL; with rifabutin, 196.8 44.7 g h/mL), and rifabutin did not affect urinary excretion of fluconazole (percentage of fluconazole dose excreted without rifabutin, 73.7% 18.6%; with rifabutin, 68.8% 15.3%). Discussion Our data indicate that concurrent fluconazole administration markedly increases the steady-state plasma concentrations of both rifabutin and its equiactive 25-desacetyl metabolite, LM565, in HIV-infected persons receiving maintenance therapy with zidovudine. This is consistent with fluconazoles inhibition of cytochrome p450 3A [7]. Renal clearance of rifabutin was unchanged with fluconazole. This further supports our hypothesis that the increased rifabutin concentrations are due to the inhibition of metabolism. Interestingly, LM565 concentrations increased 2.5-fold higher than rifabutin with concurrent fluconazole, which may represent the inhibition of further metabolism of this metabolite [19]. Although study participants were receiving other concurrent medications, these were minimized, and persons who were receiving medications known to alter drug disposition were excluded from parti

Effects of Rifabutin and Rifampicin on the Pharmacokinetics of Ethinylestradiol and Norethindrone

This open‐label, randomized, three‐way crossover study of 28 healthy premenopausal women was conducted to compare the impact of concomitant rifabutin and rifampicin on the safety, pharmacokinetics, and pharmacodynamics of the oral contraceptives ethinylestradiol and norethindrone (Ortho‐Novum 1/35; Ortho Pharmaceutical, Raritan, NJ). Each participant received oral contraceptives daily for 21 days for the first control cycle, then was randomized to one of two sequences to receive oral contraceptives with concomitant rifampicin and rifabutin at equal doses of 300 mg/day for 10 days. Ethinylestradiol, norethindrone, follicle stimulating hormone (FSH), luteinizing hormone (LH), progesterone, rifampicin, and rifabutin (and metabolite) were measured in plasma over the same time frames in all three cycles. Safety was assessed from before the beginning to the end of each cycle. Twenty‐two subjects completed all three cycles. Compared with the control cycle, rifabutin and rifampicin significantly altered the disposition of the oral contraceptive. Area under the concentration—time curve from 0 to 24 hours (AUC0–24) and maximum plasma concentration (Cmax) of ethinylestradiol decreased by 64% and 42%, respectively, after coadministration with rifampicin and by 35% and 20%, respectively, after coadministration with rifabutin. The AUC0–24 of norethindrone decreased by 60% and 20% after coadministration with rifampicin and rifabutin, respectively. Unlike progesterone levels, FSH and LH levels increased during coadministration with rifampicin and rifabutin. The incidence of spotting was significantly higher after coadministration with rifampicin (36.4%) and rifabutin (21.7%) than during the control cycle (3.7%). Although both rifampicin and rifabutin affected the pharmacokinetics of ethinylestradiol and norethindrone, the magnitude of this effect was more pronounced with rifampicin. Likewise, the fact that the highest incidence of spotting occurred with rifampicin was consistent with higher metabolic induction by rifampicin. Despite the fact that there was no change in progesterone levels, it is recommended that patients be advised to use additional contraceptive methods while receiving rifabutin or rifampicin with oral contraceptives to prevent inadvertent pregnancy.

Rifabutin absorption in humans: Relative bioavailability and food effect

The relative bioavailability of the capsule dose form (150 mg) and the effect of high‐fat food were assessed in a randomized, three‐way crossover trial of rifabutin in 12 healthy male volunteers. Each subject received a single 150 mg dose as a solution (treatment A, fasted) or a capsule with food (treatment B) and without food (treatment C), with a 2‐week washout period. Serial plasma and urine samples were obtained for 168 and 48 hours, respectively, and rifabutin and its active metabolite, 25‐O‐deacetyl‐rifabutin, quantitated by a validated HPLC procedure. The mean ± SD maximum concentration for rifabutin in plasma was 238 ± 65, 156 ± 52, and 188 ± 50 ng/ml, time to reach peak concentration was 2.5 ± 0.4, 5.4 ± 1.6, and 3.0 ± 1.1 hours, and the area under the plasma concentration‐time curve from zero to infinity [AUC(0‐∞)] was 2989 ± 726, 2640 ± 891, and 2516 ± 601 ng • hr/ml for the solution and the capsule during the fed and fasted states, respectively. Percentage of dose excreted in the urine as unchanged rifabutin was 11.0% ± 2.4%, 11.4% ± 4.9%, and 9.1% ± 2.1% for treatments A, B, and C, respectively. The corresponding AUC(0‐∞) values for the equiactive metabolite 25‐O‐deacetyl‐rifabutin, were 400 ± 184, 361 ± 187, and 298 ± 102 ng • hr/ml. ANOVA showed a significant decrease (p = 0.027) in the maximum concentration for rifabutin after administration of the capsule compared with the solution during the fasted state and a significant increase (p = 0.0001) in the time to reach maximum concentration of rifabutin when the capsule was given with food (compared with capsule given after subjects had fasted). Although the rate of rifabutin absorption appears to be slower in the presence of high‐fat food, the extent of absorption from the capsule form is not altered. Mean relative bioavailability for the capsule (fasted) was 84.8% ± 18.5%, whereas the mean relative bio‐availability (fed/fasted) was 106.0% ± 24.1%. These data suggest that the tissue exposure to this antimy‐cobacterial agent is not expected to be altered by food. Rifabutin can therefore be administered with or without food for prophylaxis of Mycobacterium avium complex.

Lack of a pharmacologic interaction between rifabutin and methadone in HIV-infected former injecting drug users

Rifampin, an agent known to decrease the half-life of methadone, and rifabutin are two rifamycins that are structurally similar and share mechanisms of action. Hence the possibility of a drug-drug interaction between rifabutin and methadone was evaluated in 24 methadone-maintained, former injecting drug users infected with the human immunodeficiency virus. The study was an open-label, drug-drug interaction and safety trial in which patients were followed for 15 days. Each patient received rifabutin 300 mg as a single dose concomitantly with their individualized methadone dosage. No significant differences in methadone peak plasma concentration, time to peak plasma concentration, area under the plasma concentration-time curve, systemic clearance or renal clearance was observed in the presence of rifabutin. Seventy-five percent of the patients reported at least one symptom of narcotic withdrawal during the study, however, these symptoms were mild. A relationship between the development of narcotic withdrawal and methadone systemic exposure could not be established. Concurrent administration of rifabutin and methadone appeared to be safe in human immunodeficiency virus-infected injecting drug users maintained on stable doses of methadone and is not expected to produce any significant changes in the pharmacokinetics of methadone in these patients.

A sensitive method for quantitation of rifabutin and its desacetyl metabolite in human biological fluids by high-performance liquid chromatography (HPLC).

Sensitive HPLC-UV methodology has been developed and validated for quantitating rifabutin, an antimycobacterial, and its 25-desacetyl metabolite, LM-565, in human plasma and urine. The HPLC separation for both plasma and urine samples was performed on an ODS, 5-µm, reverse-phase column (25 cm × 4.6-cm ID) using a mobile phase of acetonitrile/0.05 M potassium phosphate, pH 4.2, with triethylamine, (38:61.5:0.5, v/v), at a flow rate of 1.0 ml/min. The separation eluate was monitored by absorbance at 275 nm. Plasma samples (1 ml) were spiked with an internal standard (medazepam), buffered at pH 7.4 and extracted with 80:20 (v/v) hexane:ethyl acetate, and then back extracted with acidified water (0.05 M H3PO4). Linearity was established between 5.0–800 and 2.5–400 ng/ml for rifabutin and LM-565, respectively. Intraday imprecision for rifabutin and LM-565 plasma quality controls prepared at 7.3 and 3.2 ng/ml, respectively, was <15% relative standard deviation (RSD). Absolute recovery for parent drug and metabolite, from plasma, was >90% throughout the respective dynamic ranges and >70% for medazepam. Urine samples (1 ml) were acidified with 50 µl of 3.6 M H2SO4 and diluted with 0.1 M ammonium acetate. Linearity was established between 100 and 5000 ng/ml for both rifabutin and LM-565. Intraday imprecision for a urine control at 200 ng/ml was ≤ 12% RSD for either component. The method is currently being used to support Phase I kinetics program for rifabutin in prophylaxis of MAC infection of AIDS patients. Application of this method to a bioavailability assessment is presented.

A model based assessment of redistribution dependent elimination and bioavailability of rifabutin.

The autoinduction characteristic of rifabutin (RIF) following multiple oral dosing was investigated via pharmacokinetic modeling. A two-compartment model with first-order absorption was fit to plasma RIF data obtained from a study conducted in healthy normal volunteers following both a single and multiple oral doses. Parameter estimates showed an elimination rate constant (k10) of about 0.12-0.14 h-1 which was independent of the single or multiple-dosing condition. The lower-than-expected drug accumulation following multiple dosing seems to suggest that prolonged dosing perturbs the linear kinetic system. However, this analysis has shown no significant changes (p > 0.05) in the rate constants describing RIF absorption, tissue distribution/redistribution, and elimination. The mean rate of drug redistribution from the tissue compartment (k21; 0.04-0.06 h-1) was twofold to threefold lower than k10, and, with a large steady-state distribution volume (Vss/F after a single dose, 1630 L), RIF elimination appears to be dependent on drug redistribution. This hypothesis was further supported by a significant correlation (p < 0.01) between RIF tissue redistribution (k21) and terminal disposition phase rate (lambda z) constants. The redistribution dependent elimination of RIF also helps explain the stability of the terminal half-life under both single and multiple-dosing paradigms. Urinary excretion of RIF and its 25-O-deacetyl metabolite totalled less than 7% of the oral dose following single dosing, and decreased to about 4% after multiple dosing. For individual patients, the decrease in urinary recovery of the 25-O-deacetyl metabolite was directly proportional to the decrease in urinary RIF recovery. In addition, both estimates of the model intercepts (A and B) were lower following multiple dosing. Further analyses revealed a linear relationship between A and B intercepts, and also between the urinary RIF recovery and the B intercept. These relationships, in conjunction with the lack of significant increase in the rate of elimination, indicate that induction of presystemic extrahepatic metabolism and/or decrease in the extent of oral absorption may be the primary causes for the lower-than-expected systemic RIF plasma levels after multiple oral dosing.

A phase I evaluation of concomitant rifabutin and didanosine in symptomatic HIV-infected patients.

It has been suggested that didanosine (ddI) may undergo hepatic metabolism. Rifabutin is an inducer of drug metabolism. Fifteen human immunodeficiency virus-infected patients whose conditions were stabilized on twice-daily doses of ddI participated in a Phase I, open-label, pharmacokinetic and safety drug interaction study between rifabutin and ddI. Twelve patients completed the study. All patients received their regular ddI dose (167-375 mg) on day 1. On days 2-13 they received once-daily rifabutin (600 mg, three patients; 300 mg, nine patients) with their regular twice-daily ddI regimen. On days 14-16 they received rifabutin alone. Serial blood and urine samples were collected for 12 h on day 1 and for 24 h on days 13 and 16, and safety evaluations were made throughout the study. Average day 1/day 13 ddI pharmacokinetic ratios and 95% confidence interval values for Cmax, AUC0-infinity, Cls/F, and t 1/2, lambda z were 1.17 (0.96-1.38), 1.13 (0.99-1.27), 0.91 (0.81-1.01), and 0.97 (0.79-1.15), respectively (p > 0.05 for all comparisons; paired t test). A 20% difference in AUC0-infinity could be detected with 90% power. Also, there were no significant changes in laboratory values or electrocardiograms, or in rifabutin pharmacokinetic parameters when the two agents were coadministered. Based on the safety and pharmacokinetic assessments, rifabutin did not appear to interact with ddI.

Improved methodology for subnanogram quantitation of doxorubicin and its 13-hydroxy metabolite in biological fluids by liquid chromatography

Sensitive and specific methodology for quantifying doxorubicin (DOX), a potent antineoplastic drug, and its 13-hydroxy metabilite, doxorubicinol (DOX-OL), in plasma and urine has been developed and validated. The plasma method uses solid-phase extraction for analyte isolation and a narrow-bore (2.0 mm i.d.) column for liquid chromatographic separation with optimized fluorescence detection. The dynamic ranges for both drug and meatbolite in plasma are linear from 0.2 to 100 ng ml−1. Drug and metabolite are quantified in unextracted, diluted urine over a 16 to 400 ng ml−1 range. Epirubicin, an epimeric analogue of doxorubicin, is used as an internal standard. Mean extraction efficiencies for drug, metabolite and internal standard from plasma are 88, 86 and 90%, respectively. The instrumental detection limit (signal-to-noise ratio = 3) for doxorubicin or metabolite was 18 pg on column, while the lower limit of quantitation (LLOQ) was 0.3 and 0.6 ng ml−1, respectively for DOX and DOX-OL. The typical intra-dry accuracy and imprecision in plasma was < 8% bias and < relative standard deviation (R.S.D.) for DOX at or above 0.3 ng ml−1, and < 15% and < 10% for DOX-OL at or above 0.6 ng ml−1. For the urine method, the average intra-day imprecision was < 7% R.S.D. and the average bias was < 5% for both drug and metabolite over the entire dynamic range. Determinations of these two components in patient samples has verified the robustness and utility of the method.

Influence of the cardioprotective agent dexrazoxane on doxorubicin pharmacokinetics in the dog

SummaryThe influence of dexrazoxane on doxorubicin pharmacokinetics was investigated in four dogs using the two treatment sequences of saline/doxorubicin or dexrazoxane/doxorubicin. Intravenous doses of 1.5 mg/kg doxorubicin and 30 mg/kg (the 20-fold multiple) dexrazoxane were given separately, with doxorubicin being injected within 1 min of the dexrazoxane dose. Both doxorubicin and its 13-dihydro metabolite doxorubicinol were quantified in plasma and urine using a validated high-performance liquid chromatographic (HPLC) fluorescence assay. The doxorubicin plasma concentration versus time data were adequately fit by a three-compartment model. The mean half-lives calculated for the fast and slow distributive and terminal elimination phases in the saline/doxorubicin group were 3.0±0.5 and 32.2±12.8 min and 30.0±4.0 h, respectively. The model-predicted plasma concentrations were virtually identical for the saline and dexrazoxane treatment groups. Analysis of variance of the area under the plasma concentration-time curve (AUC0−∞), terminal elimination rate (λZ), systemic clearance (CLs), and renal clearance (CLr) for the parent drug showed no statistically significant difference (P<0.05) between the two treatments. Furthermore, the doxorubicinol plasma AUC0−∞ value and the doxorubicinol-to-doxorubicin AUC0−∞ ratio showed no significant difference, demonstrating that dexrazoxane had no effect on the metabolic capacity for formation of the 13-dihydro metabolite. The total urinary excretion measured as parent drug plus doxorubicinol and the metabolite-to-parent ratio in urine were also unaffected by the presence of dexrazoxane. The myelosuppressive effects of doxorubicin as determined by WBC monitoring revealed no apparent difference between the two treatments. In conclusion, these results show that drug exposure was similar for the two treatment arms. No kinetic interaction with dexrazoxane suggests that its coadministration is unlikely to modify the safety and/or efficacy of doxorubicin.

A Sensitive and Specific Procedure for Quantitation of ADR-529 in Biological Fluids by High-Performance Liquid Chromatography (HPLC) with Column Switching and Amperometric Detection

An HPLC method using electrochemical detection (ED) has been validated for the determination of ADR-529 in plasma and urine using ICRF-192 as an internal standard (IS). Prior to storage and quantitation, both plasma and urine samples require acid stabilization. Acidified plasma samples were prepared for HPLC using a two column solid-phase extraction (SPE). An aliquot of buffered plasma (i.e., pH 6-7) was first deproteinated and desalted on a C-18 SPE column. The analytes were then eluted onto a C-8 SPE column where retention and selective cleanup were achieved in the cation-exchange mode via silanol interactions. Acidified urine samples were diluted in acetonitrile prior to injection. The HPLC system for plasma and urine samples employed two narrow-bore silica columns used in the weak cation-exchange mode and separated by a switching valve. To prohibit late-eluting peaks from passivating the glassy carbon working electrode, a heart-cut containing ADR-529 and the IS was vented from the first silica column to the second using an automated switching valve. Amperometric detection at an oxidation potential of +1050 mV vs a Ag/AgNO3 reference electrode was used. Linearity was validated between 5 and 500 µg/ml in plasma and between 2 and 100 µg/ml in urine. Imprecision and percentage bias were typically <10% for both plasma and urine controls throughout their respective dynamic ranges. The absolute recoveries for ADR-529 and the IS from plasma were >95%. This method is being successfully applied to the pharmacokinetic/dynamic evaluation of ADR-529 in animals and humans.