Strategies for absorption screening in drug discovery and development.

This review gives an overview of the current approaches to evaluate drug absorption potential in the different phases of drug discovery and development. Methods discussed include in silico models, artificial membranes as absorption models, in vitro models such as the Ussing chamber and Caco-2 monolayers, in situ rat intestinal perfusion and in vivo absorption studies. In silico models such as iDEA can help optimizing chemical synthesis since the fraction absorbed (Fa) can be predicted based on structural characteristics only. A more accurate prediction of Fa can be obtained by feeding the iDEA model with Caco-2 permeability data and solubility data at various pH's. Permeability experiments with artificial membranes such as the filter-IAM technology are high-throughput and offer the possibility to group compounds according to a low and a high permeability. Highly permeable compounds, however, need to be further evaluated in Caco-2 cells, since artificial membranes lack active transport systems and efflux mechanisms such as P-glycoprotein (PgP). Caco-2 and other "intestinal-like" cell lines (MDCK, TC-7, HT29-MTX, 2/4/A1) permit to perform mechanistic studies and identify drug-drug interactions at the level of PgP. The everted sac and Ussing chamber techniques are more advanced models in the sense that they can provide additional information with respect to intestinal metabolism. In situ rat intestinal perfusion is a reliable technique to investigate drug absorption potential in combination with intestinal metabolism, however, it is time consuming, and therefore not suited for screening purposes. Finally, in vivo absorption in animals can be estimated from bioavailability studies (ratio of the plasma AUC after oral and i.v. administration). The role of the liver in affecting bioavailability can be evaluated by portal vein sampling experiments in dogs.

The pharmacokinetic properties of topical levocabastine. A review.

The linear and predictable pharmacokinetic properties of the histamine H1-receptor antagonist levocabastine make it particularly suitable for intranasal or ocular treatment of allergic rhinoconjunctivitis. Peak plasma concentrations (Cmax) occur within 1 to 2 hours of administration of single doses of levocabastine nasal spray and eye drops (0.2mg and 0.04mg, respectively). Drug absorption is incomplete after intranasal and ocular administration, with systemic availability ranging from 60 to 80% for levocabastine nasal spray and from 30 to 60% for the eye drops. However, as the amount of levocabastine applied intranasally and ocularly is small, the levocabastine plasma concentrations achieved are extremely low, with Cmax values in the ranges 1.4 to 2.2 micrograms/L and 0.26 to 0.29 micrograms/L for intranasal and ocular administration, respectively. Pharmacokinetic-pharmacodynamic modelling has indicated that the clinical benefits of levocabastine are predominantly mediated through local antihistaminic effects, although some systemic activity may contribute to the therapeutic efficacy of levocabastine nasal spray during long term use. Levocabastine undergoes minimal hepatic metabolism, i.e. ester glucuronidation, and is predominantly cleared by the kidneys. Approximately 70% of parent drug is recovered unchanged in the urine. Plasma protein binding is approximately 55% and the potential for drug interactions involving binding site displacement is negligible. Furthermore, the pharmacokinetics of this agent do not appear to be influenced by either age or gender. Levocabastine nasal spray and eye drops may thus be considered suitable for the treatment of allergic rhinoconjunctivitis in a wide patient population.

Identification of the cytochrome P450 enzymes involved in the metabolism of cisapride: in vitro studies of potential co-medication interactions.

Cisapride is a prokinetic drug that is widely used to facilitate gastrointestinal tract motility. Structurally, cisapride is a substituted piperidinyl benzamide that interacts with 5-hydroxytryptamine-4 receptors and which is largely without central depressant or antidopaminergic side-effects. The aims of this study were to investigate the metabolism of cisapride in human liver microsomes and to determine which cytochrome P-450 (CYP) isoenzyme(s) are involved in cisapride biotransformation. Additionally, the effects of various drugs on the metabolism of cisapride were investigated. The major in vitro metabolite of cisapride was formed by oxidative N-dealkylation at the piperidine nitrogen, leading to the production of norcisapride. By using competitive inhibition data, correlation studies and heterologous expression systems, it was demonstrated that CYP3A4 was the major CYP involved. CYP2A6 also contributed to the metabolism of cisapride, albeit to a much lesser extent. The mean apparent K(m) against cisapride was 8.6+/-3.5 microM (n = 3). The peak plasma levels of cisapride under normal clinical practice are approximately 0.17 microM; therefore it is unlikely that cisapride would inhibit the metabolism of co-administered drugs. In this in vitro study the inhibitory effects of 44 drugs were tested for any effect on cisapride biotransformation. In conclusion, 34 of the drugs are unlikely to have a clinically relevant interaction; however, the antidepressant nefazodone, the macrolide antibiotic troleandomycin, the HIV-1 protease inhibitors ritonavir and indinavir and the calcium channel blocker mibefradil inhibited the metabolism of cisapride and these interactions are likely to be of clinical relevance. Furthermore, the antimycotics ketoconazole, miconazole, hydroxy-itraconazole, itraconazole and fluconazole, when administered orally or intravenously, would inhibit cisapride metabolism.

Induction potential of antifungals containing an imidazole or triazole moiety. Miconazole and ketoconazole, but not itraconazole are able to induce hepatic drug metabolizing enzymes of male rats at high doses.

Male Wistar rats were dosed with miconazole, ketoconazole and itraconazole by gastric intubation once daily for up to 7 days. A dose- and time-dependent induction of the hepatic drug metabolizing enzyme system was observed for miconazole and ketoconazole, while itraconazole proved to be devoid of inductive properties even at the highest dose studied (160 mg/kg). No effect on drug metabolizing enzymes could be demonstrated for either drug at a dose level of 10 mg/kg, which is just above the antifungally active dose. At a dose of 40 mg/kg, miconazole, but not ketoconazole, significantly increased cytochrome P-450 content. At the highest dose of 160 mg/kg, both miconazole and ketoconazole increased the relative liver weight, the cytochrome P-450- and b5-content and NADPH-cyt c-reductase. Furthermore, miconazole, but not ketoconazole, increased specific microsomal aminopyrine and N,N-dimethylaniline N-demethylase activity, p-nitroanisole O-demethylase activity and UDP-glucuronyltransferase activity towards 4-nitrophenol while the specific aniline hydroxylase activity was unaffected. Ketoconazole at 160 mg/kg only induced O-demethylase activity and UDP-glucuronyltransferase activity, while it lowered the specific activities towards the other substrates. Miconazole was a relatively more potent inducer when compared to ketoconazole. Both drugs displayed biphasic effects on the mixed-function oxidase activities, which were lowered after acute administration (160 mg/kg, 1 hr before death) and were induced when determined after 23 hr had elapsed or after multiple dosage. Both drugs bound strongly to their respective induced cytochromes, giving rise to type II difference spectra, and inhibited the O-demethylase activity of the induced microsomes with an I50 of 5.2 microM for miconazole and 15.1 microM for ketoconazole. On the basis of a comparison of the enzymatic activities induced by both antimycotics with those induced by PB or 3-MC, it was concluded that miconazole behaved as a PB-type inducer, whereas ketoconazole did not belong to either category of inducers. A comparison of electrophoretograms of microsomes from different origins on SDS-PAGE revealed that miconazole increased the concentration of several proteins, whereas ketoconazole selectively induced a protein with Mr of 47,800. The protein pattern in the 50 kDa region of miconazole-induced microsomes resembled that of PB-microsomes qualitatively.

Uptake, subcellular distribution and biotransformation of 3H-labelled astemizole in cultured rat hepatocytes.

When incubated with 3H-astemizole, a potent antagonist of H1 receptor, cultured rat hepatocytes, which do not express specific receptors for this ligand, avidly take up 3H-label proportionally to the drug concentration. HPLC analysis indicates that at 10 ng 3H-astemizole/ml, cells almost entirely deplete the culture medium of the drug within 4 hr of incubation. At 37 degrees, astemizole is metabolized and released into the culture medium mainly under the form of glucuronoconjugated metabolites. Differential centrifugation of homogenates from hepatocytes incubated with 3H-astemizole indicates that astemizole and unconjugated metabolites are found in the particulate fraction, whereas astemizole and conjugated metabolites are present in the cytosol. Isopycnic centrifugation on sucrose gradient shows that the major part of the 3H-label in the particulate fraction distributes like phospholipids and NADPH cytochrome c reductase, suggesting an association with membranes and, in particular, with the endoplasmic reticulum. Chloroquine, a drug accumulating within lysosomes and acidic endosomes, decreases the uptake of 3H-astemizole by hepatocytes and induces, during isopycnic centrifugation of a particulate fraction, a shift of the 3H-label towards lower densities where it closely accompanies cathepsin B. This suggests that a minor part of astemizole accumulated in the hepatocytes could be trapped within lysosomes. These results could support the hypothesis that aspecific binding of astemizole to cellular membranes and, to a lesser extent, trapping in lysosomes could play a role in the pharmacokinetics of the drug.

The pharmacokinetics of risperidone in humans: a summary.

Risperidone is rapidly and completely absorbed after oral administration; less than 1% is excreted unchanged in the feces. The principal metabolite was found to be 9-hydroxyrisperidone. Hydroxylation of risperidone is subject to the same genetic polymorphism as debrisoquine and dextromethorphan. In poor metabolizers the half-life of risperidone was about 19 hours compared with about 3 hours in extensive metabolizers. However, becuase the pharmacology of 9-hydroxyrisperidone is very similar to that of risperidone, the half-life for the "active fraction" (risperidone +9-hydroxyrisperidone) was found to be approximately 20 hours in extensive and poor metabolizers. We found that risperidone exhibited linear elimination kinetics and that steady state was reached within 1 day for risperidone and within 5 days for the active fraction.

Plasma protein binding of risperidone and its distribution in blood.

The plasma protein binding of the new antipsychotic risperidone and of its active metabolite 9-hydroxy-risperidone was studied in vitro by equilibrium dialysis. Risperidone was 90.0% bound in human plasma, 88.2% in rat plasma and 91.7% in dog plasma. The protein binding of 9-hydroxy-risperidone was lower and averaged 77.4% in human plasma, 74.7% in rat plasma and 79.7% in dog plasma. In human plasma, the protein binding of risperidone was independent of the drug concentration up to 200 ng/ml. The binding of risperidone increased at higher pH values. Risperidone was bound to both albumin and alpha 1-acid glycoprotein. The plasma protein binding of risperidone and 9-hydroxy-risperidone in the elderly was not significantly different from that in young subjects. Plasma protein binding differences between patients with hepatic or renal impairment and healthy subjects were either not significant or rather small. The blood to plasma concentration ratio of risperidone averaged 0.67 in man, 0.51 in dogs and 0.78 in rats. Displacement interactions of risperidone and 9-hydroxy-risperidone with other drugs were minimal.

Regional brain distribution of risperidone and its active metabolite 9-hydroxy-risperidone in the rat.

Risperidone is a new benzisoxazole antipsychotic. 9-Hydroxy-risperidone is the major plasma metabolite of risperidone. The pharmacological properties of 9-hydroxy-risperidone were studied and appeared to be comparable to those of risperidone itself, both in respect of the profile of interactions with various neurotransmitters and its potency, activity, and onset and duration of action. The absorption, plasma levels and regional brain distribution of risperidone, metabolically formed 9-hydroxy-risperidone and total radioactivity were studied in the male Wistar rat after single subcutaneous administration of radiolabelled risperidone at 0.02 mg/kg. Concentrations were determined by HPLC separation, and off-line determination of the radioactivity with liquid scintillation counting. Risperidone was well absorbed. Maximum plasma concentrations were reached at 0.5-1 h after subcutaneous administration. Plasma concentrations of 9-hydroxy-risperidone were higher than those of risperidone from 2h after dosing. In plasma, the apparent elimination half-life of risperidone was 1.0 h, and mean residence times were 1.5 h for risperidone and 2.5 h for its 9-hydroxy metabolite. Plasma levels of the radioactivity increased dose proportionally between 0.02 and 1.3 mg/kg. Risperidone was rapidly distributed to brain tissues. The elimination of the radioactivity from the frontal cortex and striatum--brain regions with high concentrations of 5-HT2 or dopamine-D2 receptors--became more gradual with decreasing dose levels. After a subcutaneous dose of 0.02 mg/kg, the ED50 for central 5-HT2 antagonism in male rats, half-lives in frontal cortex and striatum were 3-4 h for risperidone, whereas mean residence times were 4-6 h for risperidone and about 12 h for 9-hydroxy-risperidone. These half-lives and mean residence times were 3-5 times longer than in plasma and in cerebellum, a region with very low concentrations of 5-HT2 and D2 receptors. Frontal cortex and striatum to plasma concentration ratios increased during the experiment. The distribution of 9-hydroxy-risperidone to the different brain regions, including frontal cortex and striatum, was more limited than that of risperidone itself. This indicated that 9-hydroxy-risperidone contributes to the in vivo activity of risperidone, but to a smaller extent than would be predicted from plasma levels. AUCs of both active compounds in frontal cortex and striatum were 10-18 times higher than those in cerebellum. No retention of metabolites other than 9-hydroxy-risperidone was observed in any of the brain regions investigated.

Location of the hydroxyl functions in hydroxylated metabolites of nebivolol in different animal species and human subjects as determined by on-line high-performance liquid chromatography-diode-array detection.

Nebivolol hydrochloride (R067555), is a new antihypertensive drug. Aromatic and alicyclic hydroxylation at the benzopyran ring systems of nebivolol are important metabolic pathways. Generally, NMR is used to unambiguously assign the sites of hydroxylation. Because of the low dose rates and the extensive metabolism of nebivolol in the different species, NMR identification is not always possible, and therefore another spectroscopic technique was searched for to address this problem. UV-chromophore absorption is affected by the kind and arrangement of adjacent atoms and groups (auxochromes). The effect of these auxochromes (e.g. -NH2, -NR2, -SH, -OH, -OR and halogens) can be strongly influenced by the pH. This paper proves that HPLC at high pH combined with on-line diode-array detection is an excellent technique for the location of the hydroxyl functions in hydroxylated metabolites of nebivolol. With this technique it is possible to differentiate between glucuronidation at the automatic and aliphatic or alicyclic hydroxyl functions.

On the pharmacokinetics of domperidone in animals and man. I. Plasma levels of domperidone in rats and dogs. Age related absorption and passage through the blood brain barrier in rats.

Domperidone, a novel gastrokinetic and antinauseant lacking central side-effects, was administered intravenously to male Wistar rats and orally to fasted rats of either sex and to 1- and 6-day old neonates at doses of 2.5 mg 14C-labelled drug/kg. The biphasic absorption of domperidone in fasted rats was extremely rapid suggesting a partial absorption from the stomach. The metabolism of domperidone was sex- and age-related: it was slower in the female rat and in the neonates. The elimination system for the metabolites was still immature in the 1-day old pups. The distribution of domperidone (and related metabolites) to the rat brain was limited, brain concentrations being lower than corresponding plasma levels in all cases. In 1-day old neonates, the blood-brain barrier was less obstructive to the passage of domperidone than in older rats. In Beagle dogs, domperidone pharmacokinetics were described by a two-compartment model with half-lives of distribution and elimination of 6 minutes and 2.45 hours respectively. The time-courses of the drug plasma levels were similar for single and repeated (once daily for 11 months) doses of 2.5, 10 and 40 mg/kg, indicating that chronic administration of domperidone, even at high dose levels, did not alter its pharmacokinetics. AUC-values increased proportionally with the dose pointing to linear pharmacokinetics over a wide dose range.

Mass spectral investigation of the metabolites of lorcainide in man.

[3H] Lorcainide hydrochloride was administered orally to healthy male volunteers. About 97% of the total radioactivity was excreted in the urine and faeces within four days of its administration. The metabolites were purified by adsorption chromatography, liquid-liquid extraction, thin layer chromatography or by gas chromatography-mass spectrometry after silylation of the samples. When there was a sufficient amount available, the samples were submitted to a nuclear magnetic resonance analysis. The results were confirmed by comparison of spectral data obtained from the reference compounds. The major metabolites of lorcainide were formed by aromatic hydroxylation, O-methylation and oxidative N-dealkylation. The urinary phenolic metabolites were present both as free aglycons and conjugates.

On the pharmacokinetics of domperidone in animals and man III. Comparative study on the excretion and metabolism of domperidone in rats, dogs and man.

The excretion and metabolism of the novel gastrokinetic and antinauseant drug domperidone were studied after oral administration of the 14C-labelled compound to rats, dogs and man, and after intravenous administration to rats and dogs. Excretion of the radioactivity was almost complete within four days. In the three species, the radioactivity was excreted for the greater part with the faeces. Biliary excretion of the radioactivity amounted to 65% of the dose 24 hours after intravenous administration in rats. Unchanged domperidone as determined by radioimmunoassay, accounted in urine for 0.3% in dogs, 0.4% in man, and in faeces for 9% in dogs and 7% in man. The main metabolic pathways of domperidone in the three species were the aromatic hydroxylation at the benzimidazolone moiety, resulting in hydroxy-domperidone -the main faecal metabolite-, and the oxidative N-dealkylation at the piperidine nitrogen, resulting in 2,3-dihydro-2-oxo-1H-benzamidazole-1-propanoic acid the major radioactive urinary metabolite- and 5-chloro-4-piperidinyl-1,3-dihydro-benzimidazol-2-one. In urine the two first metabolites were present partly as conjugates. A mass balance for the major metabolites in urine, faeces, bile and plasma samples was made up after radio-HPLC (reverse-phase HPLC with on-line radioactivity detection) of various extracts. Only minor species differences were detected.

Excretion and metabolism of lorcainide in rats, dogs and man.

After p.o. or i.v. administration of 3H-lorcainide, excretion of the radioactivity was almost complete within four days. In rats and dogs, about 35% of the dose was excreted in the urine and about 60% in the faeces. However, in humans, 62% was excreted in the urine and 35% in the faeces. In rats, about 70% of the orally administered radioactivity was excreted in the bile within 24 hours. Enterohepatic circulation was proven by "donor-acceptor" coupling in rats. Lorcainide was extensively metabolized. Urinary and faecal metabolites were isolated by extraction and high pressure liquid chromatography (HPLC), and characterized by chromatographic comparison with reference compounds, by mass spectrometry, and NMR. The mass balance for unchanged lorcainide and its major metabolites (determined by radio-HPLC) was very similar in the urine and faeces. Only minor quantitative differences were observed between intravenously and orally dosed animals, and between male and female rats. Major biotransformation pathways in the three species were: hydroxylation, O-methylation and glucuronidation. 4-Hydroxy-3-methoxy-lorcainide was the main metabolite. alpha-Oxidation resulting in alpha, 4-dihydroxy-3-methoxy-lorcainide, was observed in dogs only. Minor pathways were: oxidative N-dealkylation and amide hydrolysis. A remarkable 5-hydroxy-3,4-dimethoxy-metabolite was identified unambiguously in the three species.

On the pharmacokinetics of domperidone in animals and man. IV. The pharmacokinetics of intravenous domperidone and its bioavailability in man following intramuscular, oral and rectal administration.

The pharmacokinetics and bioavailability of domperidone, a novel gastrokinetic, were studied in healthy male subjects by comparing plasma concentrations and urinary excretion following intravenous, intramuscular, oral and rectal administration. Two oral dosage forms were studied: 10-mg tablets and a 10-mg/ml oral solution. The influence of a meal on the oral bioavailability and the dose-proportionality were also investigated. Plasma levels of intravenous domperidone could be described by a three-compartment model with a rapid distribution of 40% of the dose to a "shallow" peripheral compartment. The final elimination half-life was 7.5 hours. Peak plasma levels were reached within 30 minutes following intramuscular and oral administration and at 1-4 hours following rectal administration. Since domperidone showed an extensive first-pass elimination, AUC-values -a measure for the bioavailability- were considerably lower after oral than after parenteral administration. Equal oral and rectal doses gave a similar bioavailability. AUC-values increased proportionally with the dose over a 10-60 mg range. Cumulative urinary excretion of unchanged domperidone was proportional to corresponding AUC-values. The bioavailability was discussed in the light of the therapeutic results.

The absorption, metabolism, and excretion of the novel neuromodulator RWJ-333369 (1,2-ethanediol, [1-2-chlorophenyl]-, 2-carbamate, [S]-) in humans.

RWJ-333369 (1,2-ethanediol, [1-2-chlorophenyl]-, 2-carbamate, [S]-; CAS Registry Number 194085-75-1) is a novel neuromodulator in clinical development for the treatment of epilepsy. To study the disposition of RWJ-333369, eight healthy male subjects received a single oral dose of 500 mg of (14)C-RWJ-333369. Urine, feces, and plasma were collected for analysis for up to 1 week after dosing. Radioactivity was mainly excreted in urine (93.8 +/- 6.6%) and much less in feces (2.5 +/- 1.6%). RWJ-333369 was extensively metabolized in humans, since only low amounts of parent drug were excreted in urine (1.7% on average) and feces (trace amounts). The major biotransformation pathways were direct O-glucuronidation (44% of the dose), and hydrolysis of the carbamate ester followed by oxidation to 2-chloromandelic acid, which was subsequently metabolized in parallel to 2-chlorophenyl glycine and 2-chlorobenzoic acid (mean percentage of the dose for the three acids together was 36%). Other routes were chiral inversion followed by O-glucuronidation (11%), and aromatic hydroxylation in combination with sulfate conjugation (5%). In plasma, unchanged drug accounted for 76.5% of the total radioactivity, with the R-enantiomer and the O-glucuronide of the parent drug as the only measurable plasma metabolites. With the use of very sensitive liquid chromatography-tandem mass spectrometry techniques, only traces of aromatic (pre)mercapturic acid conjugates were detected in urine (each <0.3% of the dose), suggesting a low potential for reactive metabolite formation. In conclusion, the disposition of RWJ-333369 in humans is characterized by virtually complete absorption, extensive metabolism, and unchanged drug as the only significant circulating species.

The metabolism and fate of closantel (Flukiver) in sheep and cattle.

Closantel was reasonably well absorbed in sheep and cattle. After oral (10 mg/kg) or parenteral (5 mg/kg) administration, similar peak times (8-48 h) and peak plasma levels (45-55 micrograms/mL) are observed. Plasma level-time curves are superimposable for either route and increase linearly with the dose. The elimination half-life of closantel is 2 to 3 weeks. The relative bioavailability of 50% of oral closantel can partly be explained by incomplete absorption. Experiments in sheep with 14C-closantel revealed that the plasma radioactivity is almost exclusively due to the unmetabolized drug, metabolites accounting for less than 2%. At least 80% of the dose was excreted with the feces over the investigational period of 8 weeks, and less than 0.5% with the urine. Closantel was only poorly metabolized. Over 90% of the fecal radioactivity was due to the parent compound. Two monoiodoclosantel isomers were the only fecal metabolites detected with radio-HPLC. The distribution of closantel to tissues was limited by its high protein binding. Closantel bound strongly (greater than 99.9%) and almost exclusively to plasma albumin. Accordingly, tissue concentrations were many times lower than the corresponding plasma levels. Residual radioactivity in sheep in all tissues but liver was entirely due to closantel. About 30% to 40% of the liver radioactivity could be attributed to monoiodoclosantel. In both sheep and cattle, residual tissue concentrations decline parallel to the plasma concentrations. Consequently, the plasma kinetics of closantel reliably reflect its depletion from tissues. Independently of the dosing scheme and route of administration, the maximum daily intake by the consumer was always below the acceptable daily intake within 4 weeks after the last dose.

The plasma protein binding and distribution of etomidate in dog, rat and human blood.

The interactions of etomidate and its major metabolite (R 28 141) with plasma proteins were studied by equilibrium dialysis with a multiple cell system. A 4% human serum albumin solution was able to bind 78.5% of the etomidate, and 60.5% of R 25 141, whereas a 1.5% human gamma globulin solution bound etomidate for not more than 3% and did not bind R 28 141 at all. The association constants and free binding energies for the binding of etomidate and R 28 141 to human serum albumin were determined. Plasma protein binding of etomidate was 75.4% in the dog and 76.5% in man; in rat plasma 79.5% of the radioactivity was bound to the plasma proteins, however the etomidate was partly hydrolyzed, even in the presence of sodium fluoride. In the rat 29.7% was distributed to the blood cells, 55.9% bound to plasma proteins and 14.4% was present in plasma water; in the dog the distribution percentages were 42.1%, 43.7% and 14.2% respectively, and in man 37.7%, 47.6% and 14.7% respectively. The major metabolite of etomidate was distributed for 26.3% to the human blood cells, 47.4% was bound to plasma proteins and 26.2% was present in the plasma water; its plasma protein binding amounted to 64.3%. Etomidate was bound at or in the blood cells, whereas R 28 141 was not.

Plasma protein binding and distribution of fentanyl, sufentanil, alfentanil and lofentanil in blood.

The in vitro plasma protein binding and distribution in blood of fentanyl and three analogues were studied in rats, dogs and healthy volunteers. In human plasma, 84.4% of fentanyl was bound, 92.5% of sufentanil, 92.1% of alfentanil and 93.6% of lofentanil. Plasma protein binding of the four analgesics was independent of their concentration over the whole therapeutic range. Plasma protein binding of alfentanil was much less pH dependent than that of the three other analgesics. Attention was drawn to the possible contribution of the "acute phase' protein alpha 1-acid glycoprotein (alpha 1-AGP), of lipoproteins and of blood cells to the binding of fentanyl and its analogues in blood.