Phenytoin binding to human albumin.

Binding of phenytoin to human plasma proteins and to human serum albumin is studied using equilibrium dialysis method at pH 7.4 and 37 degrees C. Phenytoin is mainly bound to albumin, the percentage of bound drug being constant over a wide range of total drug concentrations. Calculation of the drug binding parameters show a low affinity, k = 745 M-1, and a high number of binding sites, n = 8. Palmitic acid and some acidic drugs, warfarin and phenylbutazone added to human serum albumin, decreased phenytoin binding in a non competitive way. Basic and non-ionizable drugs, on the other hand, did not modify phenytoin binding.     

Ceftriaxone binding to human serum albumin. Indirect displacement by probenecid and diazepam

In vitro protein binding studies were conducted to examine the interaction between ceftriaxone (CEF), probenecid (PROB) and diazepam (DIAZ). The presence of PROB and DIAZ at concentrations equal to molar albumin concentration caused a decrease in CEF affinity from 3.7 x 10(4) M-1 (control) to 1.1 x 10(4) (PROB) and 2.6 x 10(4) (DIAZ) M-1, but not in binding capacity in pooled human plasma. PROB and DIAZ at five times the molar albumin concentration also caused a decrease in CEF affinity from 4.5 x 10(4) M-1 (control) to 0.45 x 10(4) (PROB) and 3.0 x 10(4) (DIAZ) M-1 in isolated human serum albumin. DIAZ and PROB displaced one another, confirming their common binding site (Site II, the benzodiazepine site) on serum albumin. By contrast, CEF was unable to displace either PROB or DIAZ from defatted albumin. In the presence of elevated free fatty acid concentrations (four times the albumin concentration), CEF decreased the binding of both drugs. CEF free fraction (fp) in isolated human serum albumin (CEF fp = 7.7%) was increased by drugs which bind to Site I: sulfisoxazole (CEF fp = 68.1%), warfarin (CEF fp = 56.0%) and furosemide (CEF fp = 55.0%). At ten times the molar concentration of albumin, CEF displaced both warfarin (warfarin fp from 0.99 to 2.20%) and phenytoin (phenytoin fp from 17.7 to 23.4%) from defatted albumin. CEF appeared to bind to Site I (the warfarin site) on human serum albumin, and was displaced by PROB and DIAZ via a mechanism which did not involve direct competition at a common binding site.     

In vitro protein binding studies with BMS-204352: lack of protein binding displacement interaction in human serum.

BMS-204352, a maxi-K channel opener, is currently under development for the treatment of stroke. Protein binding of BMS-204352 was determined in sera from several species, namely, rat, monkey, dog, and human. Data indicated that the compound was shown to be highly protein bound in serum from all species (ca. 99.6%). In order to test for the potential for drug-drug interactions and competitive displacement of BMS-204352 by diazepam, phenytoin, propranolol, and warfarin, in vitro experiments were performed using spiked human serum and ex vivo human plasma samples. Protein binding was determined using equilibrium dialysis for 4 h at maximal therapeutic concentrations for each drug alone or in appropriate combination in spiked serum samples. Ex vivo samples from a clinical BMS-204352 study (0, 1, and 24 h) were dialyzed separately after addition of diazepam, phenytoin, propranolol, or warfarin. Drug content in biological matrices was measured for radioactivity using liquid scintillation counting. Results indicated that (1) addition of diazepam, phenytoin, propranolol, or warfarin did not alter the free fraction of BMS-204352; (2) BMS-204352 did not displace diazepam, phenytoin, propranolol, or warfarin from their protein binding sites, and (3) comparison of ex vivo plasma samples after BMS-204352 dosing indicated no impact of BMS-204352 and/or its metabolites on the free fraction of diazepam, phenytoin, propranolol, or warfarin. In conclusion, the potential for a drug-drug interaction due to alterations in protein binding with BMS-204352 is unlikely.     

Serum albumin-adjusted phenytoin levels: an approach for predicting drug efficacy in patients with epilepsy, suitable for developing countries

INTRODUCTION: The antiepileptic drug phenytoin has a high degree of plasma protein binding. Therefore, total phenytoin levels in plasma are misleading indicators of clinical efficacy. This study was designed to investigate whether serum albumin-adjusted phenytoin levels in Indian patients with epilepsy predict clinical outcome better than total phenytoin levels. PATIENTS AND METHODS: Fifty patients with epilepsy were included in the study and were followed-up for a period of 6 months. Serum albumin levels were estimated spectrophotometrically using the bromocresol green dye method, and serum phenytoin levels were estimated using high pressure liquid chromatography. Values were expressed as mean +/- SEM. Corrected phenytoin levels were calculated using the Sheiner-Tozer equation. Corrected phenytoin levels = Measured total phenytoin(micromol/l) [(albumin g/1 x 0.9)+ 0.1] 40. RESULTS: At Visit 1, mean serum albumin levels were 44.1 +/- 1.1 micromol/l and mean serum phenytoin levels were 33.9 +/- 2.8 g/l. After correction of the total phenytoin levels using the Sheiner-Tozer equation, 30% of the patients shifted to a different category. The follow-up visits showed similar results. Throughout the study, the corrected phenytoin levels were better indicators of clinical outcome than the total levels. In 23% of patients there was a significant difference between total and corrected phenytoin levels. CONCLUSION: In patients with serum albumin levels in the hyper- and hypoalbuminemic range, corrected phenytoin levels were better indicators of clinical outcome. In developing countries like India, where estimation of free drug levels is expensive and suitable equipment is not available in most centers, serum albumin-adjusted levels can be used by pharmacologists to predict response and thus assist in clinical decision-making.     

Stereoselectivity and species difference in plasma protein binding of KE-298 and its metabolites.

In vitro protein binding of KE-298 and its plasma metabolites, deacetyl-KE-298 (M-1) and S-methyl-KE-298 (M-2), was high in rat (>97%), dog (>89%) and human plasma (>99%), respectively. Human serum albumin (>93%) was the main protein involved in the binding to plasma proteins, while the binding to human serum globulins was low (16-33%). The binding of KE-298 and its metabolites in all species of plasma was stereoselective. The (+)-(S)-enantiomers of these compounds bound rat, dog and human plasma proteins to a greater extent than did the (-)-(R)-enantiomers, except that the case of KE-298 was opposite in rat plasma. The stereoselective plasma levels of these compounds in rats, dogs, or humans would likely be due to stereoselective differences in binding to plasma albumin. The protein binding of M-1 in adjuvant-induced arthritis rat plasma was >97%, and the stereoselectivity was similar to the case of normal rat plasma. KE-298 and its metabolites remarkably displaced [14C]warfarin, which bound on albumin in a solution of diluted rat serum albumin. Similarly, there was a displacement of [14C]warfarin in solutions of dog and human serum albumin, and concomitantly the displacement of [14C]diazepam. [3H]Digitoxin was not displaced by any of the enantiomers in each albumin solution. No stereoselectivity was found in displacement by enantiomers of the three compounds. These results suggest that stereoselective protein binding can be attributed to quantitative differences in binding to albumin rather than to the different binding sites.     

Plasma protein binding and interaction studies with benoxaprofen

A high proportion of the acidic anti-inflammatory compound benoxaprofen is bound to the plasma proteins of humans and rats (99.8 and 99.3 per cent respectively). The binding did not vary significantly up to total levels of about 120 μg/ml in human plasma or about 40 μg/ml in rat plasma. Above these levels the proportion bound decreased. At 50 μg/ml, more than 90 per cent of benoxaprofen was associated with the albumin fraction of normal human serum. At higher levels, this binding site became saturated and non-specific binding to all proteins appeared to occur. There was no difference between normal and arthritic subjects. In normal rat serum, there was much less bound to albumin. Total binding was reduced in adjuvant arthritic rats. due to decreased binding to α and β₁ globulins. At plasma levels in the range required for therapeutic activity, benoxaprofen did not displace warfarin, salicylate, prednisolone or dexamethasone from their binding to human plasma proteins. In contrast, therapeutic levels of phenylbutazone displaced warfarin.     

The interaction of enflurane, halothane and the halothane metabolite trifluoroacetic acid with the binding of acidic drugs to human serum albumin. An in vitro study

The interaction of the volatile anaesthetics enflurane, halothane and the halothane metabolite trifluoroacetic acid with the binding of two highly bound acidic drugs (warfarin, phenytoin) to albumin has been studied in vitro by equilibrium dialysis. Trifluoroacetic acid (TFA) inhibited the binding of both drugs to human serum albumin (HSA). Halothane, on the other hand, increased the binding of warfarin to HSA, while enflurane inhibited only the binding of phenytoin. It seems that the binding of the acidic drugs warfarin and phenytoin to HSA is more sensitive to the structures of the gases than for the basic drug diazepam which was previously shown to be equally affected by both gases. Furthermore, it seems that drugs competing for the same binding site (warfarin, phenytoin) may respond differently to conformational changes of the site. It is suggested that drugs bound to the "diazepam site" are more easily affected by the volatile anaesthetics than drugs bound to the "warfarin site".     

Effect of heparin and fatty acids on the binding of quinidine and warfarin in plasma

After an intravenous injection of heparin, the plasma protein binding of warfarin was greatly increased while the binding of quinidine seemed to be less affected. The increased binding of warfarin seemed partly due to the release of plasma non-esterified fatty acids, NEFA, but the level of NEFA alone could not explain the interindividual variations of plasma warfarin binding. Addition in vitro of palmitic acid to serum demonstrated an increased binding of warfarin and an unaltered binding of quinidine up to a serum NEFA level of 3.0 meq/l. Albumin isolated from post heparin plasma revealed an increased binding affinity for warfarin at the warfarin high affinity binding sites, and a slightly depressed affinity for quinidine. The low affinity binding sites on the albumin molecule for both drugs did not seem to be influenced by NEFA.     

The blood binding of cefotiam and cyclohexanol, metabolites of the prodrug cefotiam hexetil, in-vitro.

The binding of cefotiam and cyclohexanol to human serum, isolated proteins and erythrocytes has been studied in-vitro by equilibrium dialysis. The two molecules are 50% bound to serum proteins and the free fraction for both compounds remained constant within the therapeutic concentration range. Human serum albumin (HSA) was exclusively responsible for the cefotiam binding (48%) with a saturable process characterized by one binding site (n = 1.00 +/- 0.14) with a very weak affinity (Ka = 1457 +/- 352 M-1). Like other cephalosporins, cefotiam showed no binding to alpha 1-acid glycoprotein, lipoproteins or gamma-globulins. Cyclohexanol is mainly bound to HSA with a weak affinity (Ka approximately 1,800 M-1) but lipoproteins and alpha 1-acid glycoprotein bind about 30% of bound cyclohexanol in serum. Interactions with free fatty acids (FFA) or bilirubin were studied at physiopathological concentrations. HSA-bound cefotiam was displaced by FFA (1260 microM) and bilirubin (330 microM), whereas the cyclohexanol binding was inhibited only by FFA. The cefotiam binding site seems to be close to the warfarin site (site I) whereas cyclohexanol probably shares the diazepam site (site II) on HSA. There is no mutual inhibition of binding between cefotiam and cyclohexanol at therapeutic levels. The binding of both compounds to erythrocytes is low and restricted when measured in the presence of plasma.     

Effect of glycated albumin on phenytoin binding in elderly patients with type II diabetes mellitus

We evaluated the effect of glycated albumin on phenytoin protein binding in 36 elderly (age range 63-94 yrs) patients with type II diabetes mellitus (DM) under diet management. Serum was spiked with 15 mg/L phenytoin and incubated. A serum ultrafiltrate was obtained from each sample for determining total and free phenytoin concentrations. Glycated hemoglobin was determined by boronate-affinity chromatography, and glycated albumin was separated from nonglycated fractions with boronate-agarose gel. Glycated hemoglobin in the study group ranged from 4.3-14.6% (mean 7.8 +/- SD 2.1%) and glycated albumin ranged from 3.7-12.5% (7.4 +/- SD 2.6%). We observed no correlation between glycated albumin and the percentage of free phenytoin (r2 = -0.14; p = 0.419). The concentration of nonglycated albumin ranged from 0.66-4.28 g/dl (mean 3.45 +/- 0.67 g/dl) and was calculated from measured total and glycated albumin concentrations. A correlation between the free fraction of phenytoin and nonglycated albumin was not demonstrated (r2 = 0.22, p = 0.22). In addition, a correlation was not observed between total glycated albumin and the free fraction of phenytoin (r2 = -0.095; p = 0.58). We conclude that elderly patients with type II DM under diet control do not have significant alterations in phenytoin protein binding. The use of total serum phenytoin levels therefore appears appropriate for determining phenytoin dosages in elderly patients with well controlled type II DM.     

A study of the interaction between omeprazole and phenytoin in epileptic patients

This study was performed to determine the effect of omeprazole, given in therapeutically recommended doses, on the steady-state plasma levels of phenytoin in epileptic patients. Five men and three women of median age 34 years participated in the study. Steady-state plasma levels of phenytoin were measured once a week for 2 weeks before and after, respectively, and during 3 weeks of concomitant omeprazole treatment with 20 mg daily. Urinary excretion of phenytoin and its metabolite [5-(p-hydroxyphenyl)-5-phenyl-hydantoin] were determined before and at the end of the omeprazole treatment period. The steady-state plasma phenytoin levels as well as urinary excretion of phenytoin and its main metabolite were unchanged during omeprazole treatment. The results from this study suggest that concomitant omeprazole treatment in therapeutically recommended doses (20 mg daily) will not significantly affect the steady-state plasma levels of phenytoin in epileptic patients.     

Intraindividual relationships between serum protein binding of drugs in normal human subjects, patients with impaired renal function, and rats.

The serum protein binding of phenytoin, salicylic acid, sulfisoxazole, and warfarin was determined in normal human adults, in patients with impaired renal function (kidney donor and recipient), and in adult male Sprague--Dawley rats. The free fraction values for salicylate and sulfisoxazole were significantly correlated in all three groups. The other correlations were statistically significant in only one or two of these groups. There was a statistically significant negative correlation between albumin concentration and the free fraction values of salicylic acid and sulfisoxazole (but not of phenytoin and only under special circumstances with warfarin) in normal subjects and of phenytoin, salicylic acid, and sulfisoxazole (but not warfarin) in rats. No such correlation was observed for any of the drugs in patients with impaired renal function. These observations show that no single weakly acidic drug can serve as an index for quantitatively determining the effect of disease or species differences on the serum protein binding of other weakly acidic drugs.     

The binding of drugs to major human milk whey proteins.

The binding of nine drugs of diverse physicochemical characteristics to major human milk whey proteins is reported. This group included acids, bases and neutral drugs. No drug bound to alpha-lactalbumin, which is the protein present in greatest concentrations in mature milk. Four drugs, diclofenac, phenytoin, prednisolone and warfarin, bound to albumin but to a much lesser extent than in plasma, consistent with quantitatively less albumin in milk. None of the basic drugs studied bound to albumin. Five drugs, atenolol, diclofenac, prednisolone, propranolol and warfarin, bound to lactoferrin though the extent was minimal except for diclofenac. This group included acids, bases and neutral drugs.     

Plasma protein binding of the reversible type A MAO inhibitor cimoxatone (MD 780515)

Binding of a new selective reversible type A MAO inhibitor cimoxatone (MD 780515) to plasma proteins was studied in vitro by equilibrium dialysis. Binding to 580 microM human serum albumin (HSA) and to total plasma proteins was 93-96% and independent of cimoxatone concentration (0.15-207 microM). The drug was mainly bound to HSA with two binding sites and a moderate association constant (K = 2.9 X 10(4) M-1). Free fatty acids did not modify cimoxatone binding to HSA. Cimoxatone was also moderately bound to isolated lipoprotein fractions; alpha 1-acid glycoprotein and gamma-globulins did not play an important role in the binding of cimoxatone. MD 770222, the O-demethyl metabolite, appeared to be bound to HSA at the same binding sites as cimoxatone. However, no interaction occurred between the two compounds for 580 microM HSA. L-Tryptophan, bilirubin, the benzodiazepines flunitrazepam and oxazepam, imipramine and aspirin, did not displace cimoxatone from its binding sites. On the other hand, warfarin and phenylbutazone decreased cimoxatone binding to 29 microM HSA but no interaction occurred with 580 microM HSA.     

Effect of anticonvulsant drugs on (35S)t-butylbicyclophosphorothionate binding in vitro and ex vivo

Using several concentrations of eight anticonvulsant drugs in clinical use (carbamazepine, clonazepam, phenytoin, phenobarbital, ethosuximide, primidone, sodium valproate, and D,L-gamma-vinyl GABA), we studied their abilities in vitro to displace (35S)t-butylbicyclophosphorothionate (35S-TBPS) from its binding site in a homogenate of rat brain. Thereafter ethosuximide (150 mg/kg), phenobarbital (30 mg/kg), clonazepam (0.3 mg/kg), or phenytoin (100 mg/kg) was injected intraperitoneally into rats for 16-20 days; and the effect of drug administration on 35S-TBPS binding was studied in the cortex and hippocampus ex vivo. Phenobarbital (100 microM, P less than 0.001), ethosuximide (500 microM, P less than 0.001), and phenytoin (40 microM, P less than 0.001) decreased the specific 35S-TBPS binding in vitro by 10-16%. After drug administration of phenobarbital (concentration in plasma 168 microM), the number of binding sites decreased and the binding affinity (P less than 0.05) in the cortex increased. Other anticonvulsants did not modulate 35S-TBPS binding in vitro at the concentration analogous to therapeutic plasma levels or ex vivo at the dose used. These results suggest that the use of phenobarbital may modulate the TBPS binding site, but the role of the present findings in the anticonvulsant action of phenobarbital needs to be further studied.     

Clinical evaluation of plasma free phenytoin measurement and factors influencing its protein binding

The relationship between free phenytoin concentrations and clinical responses, and the factors influencing protein binding of phenytoin were investigated. A total of 119 plasma samples from 70 patients treated orally with phenytoin were analysed. The mean free phenytoin concentration was significantly higher in the patients who received phenytoin monotherapy and were classified as having a complete response (1.25 +/- 1.09 microg/ml) than that in the partial response group (0.59 +/- 0.07 microg/ml), whereas the mean total concentrations were not significantly different between the two groups. Samples were divided into three groups based on the free fraction of phenytoin, i.e. low, <5%; medium, 5%-10%; high, > 10%. The mean age (55.3 +/- 10.9 years) was significantly higher in the high group than in the low (42.7 +/- 21.2 years) and medium (42.8 +/- 16.0 years) groups. The mean creatinine clearance (CLcr) (55.3 +/- 10.9 ml/min) and serum albumin concentration (3.30 +/- 1.25 g/dl) were significantly lower in the high group than the low (88.3 +/- 29.0 ml/min and 4.08 +/- 0.50 g/dl, respectively) and medium (95.0 +/- 32.8 ml/min and 3.95 +/- 0.92 g/dl, respectively) groups. These results suggest that the free phenytoin concentration, rather than the total concentration, is more useful for monitoring antiepileptic effects in patients receiving phenytoin monotherapy. It was also found that the free phenytoin fraction was significantly influenced by aging, CLcr and serum albumin levels.     

Effect of valproate on free plasma phenytoin concentrations.

The plasma protein binding of phenytoin was studied in nine epileptic patients before and during addition of sodium valproate to the drug therapy. The free phenytoin fraction in plasma was significantly greater during sodium valproate treatment. The mean free fraction rose from 0.135 +/- 0.019 (s.d.) to 0.182 +/- 0.030. Total plasma phenytoin concentration fell significantly from a range of 4.3-26.2 micrograms/ml to 3.4-19.8 micrograms/ml during sodium valproate treatment. Neither the free plasma concentration nor the saliva concentration of phenytoin was significantly altered by sodium valproate. No significant correlation was found between plasma valproic acid concentrations and the change in phenytoin binding. We conclude that valproic acid displaces phenytoin from plasma protein binding sites but does not inhibit its metabolism.     

Pharmacokinetic interaction in rabbits between a new anticonvulsant, DL-3-hydroxy-3-ethyl-3-phenylpropionamide, and phenytoin

The effect of phenytoin on the disposition of DL-3-hydroxy-3-ethyl-3-phenylpropionamide (HEPP) has been studied in New Zealand white rabbits. Plasma HEPP levels decreased when the drug was administered with phenytoin. The area under the plasma concentration-time curve was reduced by 56.49% (from 43.23+/-7.0 to 18.81+/-2.03 microg h mL(-1)), the elimination half-life was also significantly (P<0.01) reduced (from 2.68+/-0.35 to 1.04+/-0.07 h) and the clearance was increased (from 0.35 to 0.81 L h(-1) kg(-1)). In-vitro protein binding to bovine serum albumin (BSA) and plasma was evaluated by equilibrium dialysis. Plasma protein binding was low (between 33.69 and 37.43% at concentrations ranging from 6.25 to 100 microg mL(-1)). The compound binds preferentially to albumin with an association constant (Ka) of 3.81 x 10(3) M(-1) at 37 degrees C. The results suggest a pharmacokinetic interaction between phenytoin and HEPP, probably on the drug-metabolizing enzyme system in the liver.     

Effect of age and sex on the plasma binding of acidic and basic drugs

Summary Protein binding of chlorpromazine, propranolol, meperidine, desipramine, salicylic acid and phenytoin was determined in plasma of 64 healthy volunteers (35 males and 29 females). An attempt was made to identify factors affecting the plasma protein binding of these drugs. Whereas plasma albumin levels decreased as a function of age in both sexes, α₁-acid glycoprotein levels increased with age, but the increase was more pronounced in males. The free plasma fraction of the acidic drugs (salicylic acid, phenytoin) and despiramine (a base) showed a significant (p<0.005) negative correlation with plasma albumin levels. The free fractions of the other three basic drugs (chlorpromazine, propranolol, meperidine) in plasma showed a significant (p<0.005) negative correlation with α₁-acid glycoprotein plasma levels. Plasma protein binding of salicylic acid, phenytoin and desipramine decreased as a function of age. Plasma protein binding of chlorpromazine, propranolol and meperidine was virtually unaffected by age or was slightly increased (chlorpromazine). Only in the case of salicylic acid could a statistically significant difference be demonstrated between males and females in the free fraction-age relationship. Stepwise multiple linear regression analysis, including age and blood chemistry values such as hematocrit, bilirubin, cholesterol, triglycerides, creatinine, BUN, albumin and α₁-acid glycoprotein as independent variables, identified age as the variable explaining most of the variability in plasma binding of salicylic acid, phenytoin and desipramine. For chlorpromazine, propranolol and meperidine α₁-acid glycoprotein was the most important determinant of plasma protein binding.     

Temperature dependence of phenytoin-protein binding in serum: effects of uremia and hypoalbuminemia

Phenytoin-protein binding was determined as a function of temperature in uremic patients with normal albumin and in uremic patients with hypoalbuminemia. Free phenytoin levels were determined in ultrafiltrates (MPS-1; Amicon, Lexington, MA, U.S.A.) from serum equilibrated at either 20, 25, 30, 35, or 40 degrees C. Scatchard analyses showed significant differences in phenytoin-protein binding affinity as a function of temperature. Linear regression plots of free phenytoin versus temperature showed the slopes of the uremic and uremic with hypoalbuminemia patients (0.546 and 0.535, respectively) to be 3 times that of patients with normal renal function (0.174). Free phenytoin was analyzed in 150 patients with normal renal function in whom the total phenytoin was in the range of 38.0-80.8 mumol/L. The ultrafiltrate, prepared at 22 +/- 0.5 degrees C, resulted in free phenytoin levels in the range of 3.2-8.3 mumol/L. We have established this as the therapeutic range for the patient population seen at this medical center.