Early microdose drug studies in human volunteers can minimise animal testing: Proceedings of a workshop organised by Volunteers in Research and Testing.

Testing the safety and efficacy of a successful human medicine involves many laboratory animals, which can sometimes be subjected to considerable suffering and distress. Also, it is necessary to extrapolate from the test species to humans. UK and European legislation requires that Replacement, Reduction and Refinement of animal procedures (the Three Rs) are implemented wherever possible. Over the last decade, there has been substantial progress with applying in vitro and in silico methods to both drug efficacy and safety testing. This paper is a report of the discussions and recommendations arising from a workshop on the role that might be played by human volunteer studies in the very early stages of drug development. The workshop was organised in November, 2001 by Volunteers in Research and Testing, a group of individuals in the UK which launched an initiative in 1994 to identify where and how human volunteers can participate safely in biomedical studies to replace laboratory animals. It was considered that conducting pre-Phase I very low dose human studies (sub-toxic and below the dose threshold for measurable pharmacological or clinical activity) could enable drug candidates to be assessed earlier for in vivo human pharmacokinetics and metabolism. Moreover, accelerator mass spectrometry (AMS), nuclear magnetic resonance (NMR) spectroscopy and positron emission tomography (PET) are potentially useful spectrometric and imaging methods that can be used in conjunction with such human studies. Some, limited animal tests would still be required before pre-Phase I microdose studies, to take account of the potential risk posed by completely novel chemicals. The workshop recommended that very early volunteer studies using microdoses should be introduced into the drug development process in a way that does not compromise volunteer safety or the scientific quality of the resulting safety data. This should improve the selection of drug candidates and also reduce the likelihood of later candidate failure, by providing in vivo human ADME data, especially for pharmacokinetics and metabolism, at an earlier stage in drug development than is currently the case.

Pharmacokinetics of propionyl-L-carnitine in humans: evidence for saturable tubular reabsorption.

AIMS: Propionyl-L-carnitine (PLC) is an endogenous compound which, along with L-carnitine (LC) and acetyl-L-carnitine (ALC), forms a component of the endogenous carnitine pool in humans and most, if not all, animal species. PLC is currently under investigation for the treatment of peripheral artery disease, and the present study was conducted to assess the pharmacokinetics of intravenous propionyl-L-carnitine hydrochloride. METHODS: This was a placebo-controlled, double-blind, parallel group, dose-escalating study in which 24 healthy males were divided into four groups of six. Four subjects from each group received propionyl-L-carnitine hydrochloride and two received placebo. The doses (1 g, 2 g, 4 g and 8 g) were administered as a constant rate infusion over 2 h and blood and urine were collected for 24 h from the start of the infusion. PLC, ALC and LC in plasma and urine were quantified by h.p. l.c. RESULTS: All 24 subjects successfully completed the study and the infusions were well tolerated. In addition to the expected increase in PLC levels, the plasma concentrations and urinary excretion of LC and ALC also increased above baseline values following intravenous propionyl-L-carnitine hydrochloride administration. At a dose of 1 g, PLC was found to have a mean (+/- s.d.) half-life of 1.09 +/- 0.15 h, a clearance of 11.6 +/- 0.24 l h-1 and a volume of distribution of 18.3 +/- 2.4 l. None of these parameters changed with dose. In placebo-treated subjects, endogenous PLC, LC and ALC underwent extensive renal tubular reabsorption, as indicated by renal excretory clearance to GFR ratios of less than 0.1. The renal-excretory clearance of PLC, which was 0.33 +/- 0.38 l h-1 under baseline condition, increased (P < 0. 001) from 1.98 +/- 0.59 l h-1 at a dose of 1 g to 5.55 +/- 1.50 l h-1 at a dose of 8 g (95% confidence interval for the difference was 2.18,4.97). As a consequence, the percent of the dose excreted unchanged in urine increased (P < 0.001) from 18.1 +/- 5.5% (1 g) to 50.3 +/- 13.3% (8 g). The renal-excretory clearance of LC and ALC also increased substantially after PLC administration and there was evidence for renal metabolism of PLC to LC and ALC. CONCLUSIONS: Intravenous administration of propionyl-L-carnitine hydrochloride caused significant increases in the renal excretory clearances of PLC, LC and ALC, due to saturation of the renal tubular reabsorption process - as a consequence there was a substantial increase with dose in the fraction excreted unchanged in urine. Despite the marked increase in the renal clearance of PLC, total clearance remained unchanged, suggesting a compensatory reduction in the clearance of the compound by non excretory routes.

The effect of cimetidine on the steady-state pharmacokinetics and pharmacodynamics of warfarin in humans.

OBJECTIVE: The interaction of multiple oral doses of cimetidine on the steady-state pharmacokinetics and pharmacodynamics of warfarin was investigated in six healthy male volunteers. METHODS: The subjects were given individually adjusted doses of warfarin to achieve therapeutic levels of prothrombin activity. The established daily maintenance oral dose of warfarin was kept stable throughout the trial and, on study days 8-14, each volunteer received a 800-mg daily dose of cimetidine. The degree of anticoagulant response produced by warfarin was quantified by the determination of both the prothrombin time and factor-VII clotting activity. RESULTS: Cimetidine co-administration had no significant effect on the pharmacokinetics of the more potent S-warfarin but significantly increased by 28% (P < 0.05) mean R-warfarin trough plasma concentrations and decreased by 23% (P < 0.05) mean R-warfarin apparent clearance. Both prothrombin time and factor-VII clotting activity displayed considerable inter-subject variability and were not significantly affected by concurrent cimetidine treatment. The reduction of apparent clearance of R-warfarin by cimetidine was found to be the effect of inhibition of the formation of warfarin metabolites as determined by apparent formation clearance values (+/-SD) of R-6-hydroxywarfarin (31.1+/-7.4 ml/h baseline; 18.5+/-4.5 ml/h at end of cimetidine treatment; P < 0.01), and R-7-hydroxywarfarin (6.9+/-1.3 ml/h baseline; 4.3+/-1.1 ml/h at end of cimetidine treatment; P < 0.01). CONCLUSION: Cimetidine stereoselectively affects the steady-state pharmacokinetics of warfarin by inhibiting the disposition of the less potent R-warfarin in humans. However, this interaction is likely to be of minimal clinical significance in most patients.

Enhanced cofermentation of glucose and xylose by recombinant Saccharomyces yeast strains in batch and continuous operating modes.

Agricultural residues, such as grain by-products, are rich in the hydrolyzable carbohydrate polymers hemicellulose and cellulose; hence, they represent a readily available source of the fermentable sugars xylose and glucose. The biomass-to-ethanol technology is now a step closer to commercialization because a stable recombinant yeast strain has been developed that can efficiently ferment glucose and xylose simultaneously (coferment) to ethanol. This strain, LNH-ST, is a derivative of Saccharomyces yeast strain 1400 that carries the xylose-catabolism encoding genes of Pichia stipitis in its chromosome. Continuous pure sugar cofermentation studies with this organism resulted in promising steady-state ethanol yields (70.4% of theoretical based on available sugars) at a residence time of 48 h. Further studies with corn biomass pretreated at the pilot scale confirmed the performance characteristics of the organism in a simultaneous saccharification and cofermentation (SSCF) process: LNH-ST converted 78.4% of the available glucose and 56.1% of the available xylose within 4 d, despite the presence of high levels of metabolic inhibitors. These SSCF data were reproducible at the bench scale and verified in a 9000-L pilot scale bioreactor.

Atovaquone has no effect on the pharmacokinetics of phenytoin in healthy male volunteers.

The potential pharmacokinetic interaction between atovaquone and phenytoin was investigated in 12 healthy male volunteers. Each volunteer received a single 600 mg oral dose of phenytoin in the two treatment periods. On one occasion phenytoin was taken alone and on the other pre-treatment with 2000 mg atovaquone taken as two doses of 1000 mg as a microfluidized suspension. The mean (+/- s.d.) peak plasma concentrations (Cmax), apparent total clearance (CL/F) and terminal half-life (t1/2) for phenytoin when administered alone were 10.6(1.8) mg 1(-1), 24.3 (7.7) ml min-1 and 25(8) h, respectively. When administered together with atovaquone, phenytoin Cmax, CL/F and t1/2,z were 10.9 (2.0) mg 1(-1), 23.8 ml min-1 and 24(6) h, respectively. There were no statistically significant differences in any of these plasma pharmacokinetic parameters. There were also no statistically significant differences in the fraction of circulating drug not bound to plasma protein or urinary excretion of 5-hydroxyphenyl-phenyl-hydantoin. In conclusion, there was no effect of atovaquone on the pharmacokinetics of phenytoin or its major metabolite after a single dose.

The effect of food on cabergoline pharmacokinetics and tolerability in healthy volunteers.

The effect of food on the pharmacokinetics and tolerability of cabergoline in man was investigated. For this purpose an open, randomized, single-dose study was conducted in 12 healthy male volunteers who received 1 mg cabergoline as tablets both under fasting conditions and after a breakfast containing a substantial amount of carbohydrates, fat, and proteins, in a crossover fashion. The two treatments were separated by a 4 week washout period. Plasma and urine were collected up to 336 and 168 h respectively after administration and cabergoline concentration was measured in both fluids using a validated radioimmunoassay. Tolerability assessment included haematology, blood chemistry, and urinalysis, blood pressure and heart rate measurements, and ECG. Under both fasting and fed conditions low but persistent cabergoline plasma levels were observed in the present study up to 2 weeks after drug intake, in agreement with the long-lasting prolactin-lowering activity of the drug. In subjects receiving cabergoline under fed or fasting conditions, Cmax values averaged 44 and 54 pg mL(-1), AUC(0-336 h) averaged 6392 and 5331 pg h mL(-1), Ae(0-168 h) averaged 12.7 and 11.9 micrograms, and t1/2 averaged 109.7 and 101.3 h, respectively. No statistically significant difference was found when Cmax, AUC(0-336 h), t1/2, and Ae(0-168 h) from subjects treated under fasting and fed conditions were compared. Median tmax values in subjects treated under fasting or fed conditions were identical (2.5 h). The statistical analysis applied to the parameters chosen to evaluate the variations in the blood pressure profiles observed either supine or standing did not show any significant difference between the fed and fasting conditions. Heart rate values were not significantly modified after cabergoline under either fed or fasting conditions. Laboratory evaluation showed some minor deviations from normal, which were not clinically relevant (only one subject showed an occasional and transient elevation in alkaline phosphatase which disappeared in the subsequent laboratory evaluations) and were considered for the most part not to be drug related. Eleven subjects reported adverse events (one after both treatments, five only after drug intake under fasting conditions, and five only after drug intake with food.

The relevance of pharmacokinetics in the development of biotechnology products.

Biotechnology derived medicines will have an increasing impact not only upon medical practice but also upon the working lives of many pharmaceutical scientists. Whilst such medicines may be viewed as highly sophisticated to the clinician and scientist, the computer will still rightly demand that they are both efficacious and safe. Impacting as it does upon all phases of drug development and facilitating quantitative relationship between administered dose and systemic drug concentration, pharmacokinetics has an important role to play in the development of all medicines. Bioanalysis is an essential prelude to any pharmacokinetic investigation. For many biotechnology products the immunoassay and bioassay methodologies employed are often relatively non-specific and imprecise and yield assay dependent pharmacokinetic parameters. Other factors may also confound the pharmacokinetic evaluation of biotechnological products. In vivo binding proteins (including antibodies) may not only interfere with bioanalytical methodology but also have a significant effect on the pharmacokinetics and biological activity of certain macromolecules. Antibody formation is a particular problem in the preclinical evaluation of human proteins. Unlike most conventional pharmaceuticals, the rate and time of delivery into the systemic circulation is a fundamental component of the biological activity of many biological molecules.

Comprehensive pharmacokinetics of urinary human follicle stimulating hormone in healthy female volunteers.

PURPOSE: The study determined the pharmacokinetics of urinary human follicle stimulating hormone (u-hFSH) in 12 down-regulated healthy female volunteers. METHODS: Following pituitary desensitization, baseline FSH serum levels were measured over a 24-hour period. Then each subject received, in random order, single doses of u-hFSH (Metrodin), 75 IU, 150 IU and 300 IU iv, and 150 IU im on four occasions separated by washout periods of one week. Blood and urine samples were collected at preset times. FSH levels were measured by a immuno-radiometric assay and an in vitro rat granulosa cells aromatase bioassay. RESULTS: All doses of u-hFSH were well tolerated. After an iv bolus, the pharmacokinetics of FSH were well described by a two-compartment open model. Immunoassay data showed that the total exposure to FSH was proportional to the administered dose. Mean total clearance of FSH was approximately 0.5 L.h-1 and renal clearance was 0.14 L.h-1. The volume of distribution at steady-state was around 8 liters. The distribution half-life was 2 h and the terminal half-life nearly one day. After im injection, almost two thirds of the administered dose was available systemically. The in vitro bioassay confirmed this pharmacokinetic analysis. CONCLUSIONS: The estimation of the elimination half-life of around one day indicates that the maximal effect of a given dose of u-hFSH administered daily cannot be observed until 3 to 4 days of repeated administration. This indicates that, on a pure pharmacokinetic basis, physicians should wait at least 4 days to assess the efficacy of a given dose of u-hFSH and that they should not modify dosage too frequently.

Absence of effect of ranitidine on blood alcohol concentrations when taken morning, midday, or evening with or without food.

OBJECTIVE: To investigate the effects of multiple dosing with ranitidine (300 mg four times a day) on the absorption of a moderate dose of alcohol (0.5 gm.kg-1), consumed postprandially or on an empty stomach at different times of day, and to investigate if coadministration of ranitidine affects psychomotor function. METHODS: Two double-blind, randomized, two-way crossover, and placebo-controlled studies were performed in a university research establishment. Study subjects were 36 (18 in each study) normal, healthy, nonalcoholic men aged from 25 to 48 years. Subjects received either 300 mg ranitidine four times a day or placebo for 8 days with oral alcohol (0.5 gm.kg-1) in the morning on day 4, at midday on day 6, and in the evening on day 8. Alcohol was consumed 45 minutes after standard meals and 30 minutes after ranitidine in the first study; it was consumed on an empty stomach 30 minutes after ranitidine in the second study. RESULTS: Maximum blood alcohol concentrations, area under the blood alcohol concentration--time curve, and time to maximum concentration were not significantly different during ranitidine coadministration compared with coadministration of placebo. This result held true for each time of day and for fed and fasting states. Similarly, ranitidine had no detectable effect on any of the results from tests of psychomotor function. CONCLUSION: Irrespective of the time of day, ranitidine has no statistically or clinically significant effects on blood alcohol profiles.

Clinical pharmacology of recombinant human follicle-stimulating hormone (FSH). I. Comparative pharmacokinetics with urinary human FSH.

OBJECTIVE: To assess and compare the pharmacokinetics of recombinant human FSH with those of a reference preparation of urinary human FSH. DESIGN: Urinary human FSH and recombinant human FSH (Metrodin and Gonal-F; Laboratoires Serono, Aubonne, Switzerland) were administered in a balanced, random order, crossover sequence as a single i.v. dose of 150 or 300 IU separated by 1 week of washout to 12 pituitary down-regulated, healthy female volunteers. Serum FSH concentrations were measured by an immunoradiometric assay (IRMA) and by an in vitro rat granulosa cell aromatase bioassay. Urine FSH concentrations were measured by IRMA. RESULTS: The mean concentration-time profiles after 150 IU of urinary human FSH and recombinant human FSH were superimposed, and the mean profile after 300 IU of recombinant human FSH was double that of the 150 IU dose. The data for both FSH preparations were well described by a biexponential equation. Total clearance of the preparations was comparable, judging from immunoassay and bioassay data (0.5 and 0.15 L/h, respectively). Based on the immunoassay, renal clearance of urinary human FSH was 0.1 L/h, whereas for recombinant human FSH it was slightly lower at 0.07 L/h, indicating that less than one fifth of the administered dose was excreted in the urine. Immunoassay showed that the two preparations were similar in terms of initial and terminal half-lives (2 and 17 hours, respectively). The volumes of distribution at steady state (11 L) were similar. The results of the in vitro bioassay confirmed this pharmacokinetic analysis. Just after i.v. administration, an initial decrease in the serum bioassay:immunoassay ratio was observed because of dilution of urinary human FSH or of recombinant human FSH in the residual endogenous FSH pool. Then the ratio increased progressively with time, suggesting either metabolic selection or activation of both types of injected human FSH toward forms with greater in vitro bioactivity. The bioassay:immunoassay ratio returned to baseline by day 7. CONCLUSION: The results obtained in this study indicate that the following [1] the pharmacokinetic characteristics of recombinant human FSH are similar to those of urinary human FSH; [2] the terminal half-life of human FSH is approximately 1 day; [3] after a single i.v. injection of human FSH a progressive increase in FSH bioassay: immunoassay ratio is observed; and [4] clinical use of recombinant human FSH could follow protocols and treatment regimens currently applied to urinary human FSH.

Clinical pharmacology of recombinant human follicle-stimulating hormone. II. Single doses and steady state pharmacokinetics.

OBJECTIVE: To assess the single-dose pharmacokinetics of a recombinant human FSH preparation (Gonal-F; Laboratoires Serono, Aubonne, Switzerland), administered by i.v., IM, and SC routes and its pharmacokinetics at steady state after multiple dosing by the SC route. DESIGN: Twelve healthy down-regulated female volunteers received in random order three single doses of recombinant human FSH (150 IU, i.v., IM, and SC), with each administration separated by 1 week. The volunteers then received multiple recombinant human FSH doses by the SC route (150 IU one time per day) for 7 days. Follicle-stimulating hormone concentrations were measured by an immunoradiometric assay and an in vitro granulosa cell aromatase bioassay. RESULTS: After a single administration, the pharmacokinetics of recombinant human FSH were well-described by a two-compartment model after i.v. administration and by a one-compartment model with first order absorption after IM or SC administration. The mean total clearance of FSH was approximately 0.6 L/h, and renal clearance accounted for one tenth of the total elimination after i.v. administration. The distribution half-life was close to 2 hours. The terminal half-life was nearly 1 day when estimated either by modeling the i.v. data set or from analysis of the terminal phase of the steady state pharmacokinetic curve or from the time taken to reach steady state after repeated SC administrations. After single IM and SC injection, two thirds of the administered dose was available systemically. The cumulation factor for repeated SC administration was approximately 3 when steady state was reached. The in vitro bioassay data confirmed these estimations. The temporal evolution of the bioassay:immunoassay ratio suggests either metabolic selection or activation of recombinant human FSH toward forms with greater in vitro bioactivity. CONCLUSION: The estimation of the elimination half-life of approximately 1 day indicates that the maximal effect of a given dose of recombinant human FSH administered daily cannot be observed until 3 to 4 days of repeated administration. This indicates that, on a pure pharmacokinetic basis, physicians should wait at least 4 days to assess the efficacy of a given dose of recombinant human FSH and that they should not modify dosage too frequently.

The effect of renal function on the pharmacokinetics of ranitidine.

This open study evaluated the influence of renal function on the pharmacokinetics of ranitidine (50 mg i.v. infusion given over 6 min). Five groups, each of 8 subjects, 1 with normal renal function and 4 with different degrees of renal impairment were studied. Renal function was assessed in each patient by 51Cr-EDTA (glomerular filtration rate, GFR), creatinine clearance (GFR) and N-methylnicotinamide clearance (reflecting glomerular and tubular function). Sixteen blood samples (5 ml) taken up to 48 h post dose from each subject were analysed for plasma ranitidine concentrations by reversed phase HPLC. Patient groups with renal impairment had significantly increased AUC infinity and t1/2 with corresponding decreases in CLp and lambda z when compared with normal subjects. There was also a significant increase in tmax but not in Cmax. There was a high linear correlation between the degree of renal impairment and ranitidine clearance. In patients with GFR < or = 20 ml min-1, the AUC infinity mean ratio (compared with normal subjects) was up to 4.6 while for patients with GFR 20-50 ml min-1, the average AUC infinity ratio was 2.6. It is recommended that the dose of ranitidine is halved in patients with GFR < or = 20 ml min-1.

Human transbuccal absorption of diclofenac sodium from a prototype hydrogel delivery device.

The buccal delivery of the nonsteroidal antiinflammatory drug, diclofenac sodium (Voltaren), from a prototype hydrogel was studied in man in a randomized crossover design of buccal delivery and i.v. infusion. After a 30-min delay, plasma levels of diclofenac increased to near steady-state levels of 100 ng/ml by 3 hr. With each subject serving as his own control, the i.v. infusion data facilitated the calculation of a mean steady-state flux of diclofenac sodium of 2.1 +/- 0.6 mg/cm2-hr across human buccal mucosa and a time lag of 1.0 +/- 0.5 hr. The large flux of this ionized species indicates that the traditional lipoidal model of buccal permeation based on the partition coefficient is inadequate.

Further insight into the stereoselective interaction between warfarin and cimetidine in man.

The urinary profile of R- and S-warfarin, following administration of a single (25 mg) oral dose of racemic warfarin, alone or on day 4 of a 9-day chronic administration of cimetidine (800 mg once daily), was investigated in eight healthy male volunteers. Based on estimated apparent formation clearance values, cimetidine inhibited significantly only the formation of R-6-hydroxywarfarin and R-7-hydroxywarfarin.

The pharmacokinetics and pharmacodynamics of nifedipine at steady state during concomitant administration of cimetidine or high dose ranitidine.

Ranitidine may be used at doses of up to 300 mg twice daily in the healing of duodenal ulcers, and this study investigated the potential for a pharmacokinetic or pharmacodynamic interaction between nifedipine 10 mg three times daily and ranitidine 300 mg twice daily compared with cimetidine 800 mg daily and placebo in a randomised crossover study in 18 healthy male subjects. Twelve blood samples were taken on the fifth day in each treatment period and assayed for nifedipine by h.p.l.c. Pulse, blood pressure and ECG recordings were also taken. Cimetidine, but not ranitidine, produced significant changes in the pharmacokinetics of nifedipine at steady state. Mean +/- s.d. values of AUC were 105 +/- 40 micrograms l-1 for placebo treatment, 111 +/- 45 micrograms l-1 h for ranitidine and 211 +/- 64 micrograms l-1 h for cimetidine (P less than 0.001), and Cmax values were 33 +/- 14, 39 +/- 27 and 76 +/- 40 micrograms l-1 (P less than 0.001), respectively. Neither ranitidine nor cimetidine produced statistically significant changes in the pharmacological response to nifedipine.

The pharmacokinetics of the enantiomers of flurbiprofen in patients with rheumatoid arthritis.

Plasma and synovial fluid concentrations of the enantiomers of flurbiprofen were measured in 15 rheumatoid patients receiving 100 mg racemic flurbiprofen twice daily. Pharmacokinetic parameters showed considerable variability within the group of patients, although differences in S(+)/R(-) plasma concentration ratios were small. The average values (+/- s.d.) of oral plasma clearance, volume of distribution and elimination half-life for R(-)-flurbiprofen were 0.075 (+/- 0.066) l min-1, 12.47 (+/- 5.79) l and 138 (+/- 61) min, respectively. The average values (+/- s.d.) of oral plasma clearance, volume of distribution and elimination half-life for S(+)-flurbiprofen were 0.057 (+/- 0.035) l min-1, 12.81 (+/- 4.43) l and 155 (+/- 49) min, respectively. S(+)/R(-) ratios (+/- s.d.) rose from 1.06 (+/- 0.12) to 1.75 (+/- 0.61) at the end of the 12 h interval in plasma and from 1.18 (+/- 0.13) to 1.47 (+/- 0.24) over the measured time course in synovial fluid. Increases in S(+)/R(-) ratios may be clinically important as they demonstrate accumulation of the pharmacologically active species.

Investigations into the potential effects of multiple dose ketorolac on the pharmacokinetics and pharmacodynamics of racemic warfarin.

1. The potential interaction between racemic warfarin given as a 25 mg single oral dose and chronically administered ketorolac was studied in 12 young healthy male volunteers. 2. Ketorolac produced no major change in the pharmacokinetics of (R)- or (S)-warfarin. 3. Ketorolac did not alter the pharmacodynamic profile of racemic warfarin. 4. Ketorolac increased template bleeding time by a factor of 1.35 as compared with placebo. 5. The results suggest that the ketorolac-warfarin interaction is unlikely to be of major clinical significance; however, combined use of ketorolac and warfarin in patients should be undertaken with due caution and appropriate monitoring.

An assessment of the effects of impaired renal function and haemodialysis on the pharmacokinetics of fluconazole.

1. The oral pharmacokinetics of fluconazole were studied in three groups of volunteers (n = 5) with various degrees of renal function (GFR greater than 70 ml min-1; 20-70 ml min-1; less than 20 ml min-1) and in a group of patients with chronic end-stage renal failure requiring regular haemodialysis. 2. The pharmacokinetics of fluconazole were markedly affected by impaired renal function with the elimination of half-life in Group III (GFR less than 20 ml min-1) being approximately three times that observed in normal volunteers (Group I). 3. Fluconazole renal clearance was positively correlated with GFR. 4. Non-renal clearance of fluconazole decreased with decreasing renal function. 5. Approximately 38% of the 50 mg dose of fluconazole was removed by haemodialysis extending over a 3 h period.