CLIN. CHEM. 27/1, 166-168 (1981) 166 CLINICAL CHEMISTRY, Vol. 27. No. 1, 1981 The “Ultra-Free” Ultrafiltration Technique Compared with Equilibrium Dialysis for Determination of Unbound Thiopental Concentrations in Serum Donald Jung, Michael Mayersohn,1 and Donald Perrier We compare a new ultrafiltration technique, involving a unique Millipore membrane, with the classical method of equilibrium dialysis for determining the fraction of thio- pental not bound to serum proteins. This fraction, as de- termined by equilibrium dialysis at 37 #{176}C, ranged between 12 and 16% for total concentrations of 50 jig/L to 10 mg/L of serum. In contrast, ultrafiltration at 37 #{176}C yielded a 49% higher value for unbound thiopental: 26.3 (SD 2.6)%. Determined at room temperature (24 #{176}C), there was no statistically significant difference for results by the two methods: 14.2 and 15.9%, respectively. The discrepancy between results at 37 #{176}C may partly be explained by serum proteins penetrating the Ultra-Free filter. For the routine clinical measurement of unbound drug concentrations, the ultrafiltration membrane at room temperature appears to be sufficiently accurate and less time-consuming than equilibrium dialysis. AddItional Keyphrases: drug assay ultrafiltration #{149} bar- biturates . anesthetics The binding of a drug to serum proteins influences its dis- trihution, elimination, and pharmacological activity; therefore, knowledge of the extent of binding will assist in the inter- pretation of a given drug concentration in serum. Drug assays usually measure total drug concentration in serum, i.e., bound plus unbound drug. To determine the unbound concentration in plasma or serum, one must measure the concentration in a protein-free solution. The following techniques have long been used to measure unbound concentrations: equilibrium dialysis, ultrafiltration, ultracentrifugation, and gel filtration. Equilibrium dialysis is the technique most often used to study the binding of drugs and other ligands to proteins, but the long time necessary to reach equilibrium (ca. 8-24 h) is a major disadvantage. Because of simplicity, convenience, and speed, ultrafiltration techniques involving gas pressure or centrifugal force have been used to separate the protein-free phase from the protein-containing solution. Recently, a modified ultrafiltration technique involving use of a Millipore membrane (Ultra-Free#{174};Worthington Diag- nostics, Freehold, NJ 07728) and a 1-mL tuberculin syringe has been introduced to measure free anticonvulsant drug concentrations in plasma. The purpose of our study is to evaluate and compare the reliability of this new ultrafiltration approach with the “classical” and more time-consuming method of equilibrium dialysis. We chose as a model drug thiopental, an intravenous anesthetic shown to have a high capacity for plasma protein binding. l)epartment of Pharmaeeut cal Sciences and Ruth F. Golding ClinicalPharmacokinetics I.ahoratorv,College of Pharmacy, ‘I’he Universityof Arizona, ‘l’ucson, AZ 85721. Address communications to this author. Received .June 30, 1980; accepted Oct. 8, 1980. Materials and Methods Serum, Plasma, and Human Serum Albumin Plasma or serum was obtained from fresh blood samples of normal, healthy volunteers, collected in heparinized (14.3 USP units/mL) or non-heparinized test tubes. Human serum al- bumin, Fraction V, fatty-acid free (United States Biochemical Corp., Cleveland, OH 44122), was used as received and made up to a concentration of 40 g/L in water. Drugs 12-’4C]Thiopental (6.3 Ci/mol) was obtained from ICN, Isotope Nuclear Division, Irvine, CA 92715; radioactive purity was greater than 98%. Thiopental free acid was prepared from thiopental sodium (Pentothal; Abbott Lab., North Chicago, IL 60064). Final concentrations of 50 ig/L to 10 mg/L were prepared by adding a constant amount of radiolabeled drug and various amounts of unlabeled thiopental to plasma or serum. Equilibrium Dialysis Dialysis membranes (Technilab Instruments, Pequannock, NJ 07440) with an Mr cutoff of 6000 were placed between two plastic blocks of an acrylic plastic dialyzing chamber used for equilibrium dialysis. Through the access ports, 1 mL of 0.13 mol/L phosphate buffer, pH 7.4, was added to one side of the membranes, and 1 mL of plasma or serum containing labeled and unlabeled thiopental was added to the other side. The dialyzing chamber was placed in a 37 #{176}C water bath or kept at room temperature (24 #{176}C) and rotated at 60 rpm until equilibrium was reached. Then 600 tL of the buffer solution or 200 tL of serum was added to 10 mL of Instagel (Packard Instrument Co., Downers Grove, IL 60515) and liquid scin- tillation counting was performed to 2% counting error. Counts per minute (cpm) were converted to disintegrations per minute (dpm) by use of an external standard channels ratio method. After dialysis, radioactivity of the membranes was counted to determine the amount of drug bound to the membrane. The percentage of unbound thiopental was cal- culated by dividing the concentration of thiopental in the buffer solution by the concentration of thiopental in the serum after dialysis. Ultrafiltration We used the Ultra-Free anticonvulsant drug filter (lot no. A0A39020) with an Mr cutoff of 40 000. We performed ul- trafiltration by aspiration, using a 1-mL syringe plunger. One-milliliter serum samples were placed into the reservoir cup of the filter unit and allowed to stand undisturbed for at least 2 mm before reducing the pressure with the syringe. Ultrafiltration was performed at either 37 or 24 #{176}C until 100-600 pL of ultrafiltrate was collected. We analyzed the ultrafiltrate as described for equilibrium dialysis, except that we determined the fraction of thiopental that was unbound by dividing the concentration of thiopental in the ultrafiltrate