Serum protein binding of diazepam and its displacement by valproic acid in vitro.

Diazepam is frequently used in the epileptic patient, either for its anticonvulsant effect or in the treatment of coexisting anxiety or spasticity. The possibility of interaction with other antiepileptic drugs has been little studied. As diazepam is approximately 98% bound to plasma proteins (Van Der Kleijn et al., 1971; Thiessen et al., 1976; Klotz, Antonin & Bieck, 1976) the possibility of binding interactions with other drugs arises. Free fatty acids are known to displace diazepam in vitro (Tsutsumi et al., 1975; Sjodin, 1977) and in vivo (Colburn & Gibaldi, 1978; Sellers et al., 1980). The antiepileptic drug valproic acid is a two chain fatty acid, with similar displacing properties to endogeneous free fatty acids, including the ability to bind to fatty acid binding sites on albumin (Monks & Richens, 1979), and so we decided to investigate the effect of this drug on the binding of diazepam in serum. We were stimulated to do so by the observation that the apparent volume of distribution of diazepam tended to be higher in epileptic patients on sodium valproate (Dhillon & Richens, 1981). Binding was measured by the method of Lunde et al. (1970) in pooled serum collected from normal volunteers. 5-[C14] diazepam (specific activity 206 ,uCi/mg) (Roche Products Ltd) was added to serum in vitro in a concentration of0.52 ,umol/l. The binding of diazepam was measured over a 'therapeutic' range of concentrations, namely 0.53-4.04 ,umol/l. The effect of varying valproic acid concentrations (175-700 ,mol/l) was determined at a diazepam concentration of 1.58,mol/l. In order to examine the nature of the interaction between the two substances a Scatchard plot (Scatchard, 1949) was constructed over the relatively large range of diazepam concentrations of 3.5 x 10f6 mol/l to 2.11 x 10-4 mol/l. Diazepam binding to human serum albumin was found to be 98.1 + 0.08% at a serum diazepam concentration 1.58 Amol/l. This is in agreement with values reported earlier (Van der Kleijn et al., 1971; Klotz et al., 1976). (The coefficient of variation in the pool, within batch for the unbound fraction was 3.8% (n = 10)). Binding was found to be independent of the drug concentration over a range of 0.53-4.03 Amol/l but was shown to be concentration dependent at high serum concentrations, above those normally encountered with the therapeutic use of diazepam (3.51-2. 10 ,mol/l). The effect of valproic acid on the binding of diazepam (serum diazepam concentration 1.58 ,umol/l) is shown in Figure 1. In the absence of valproic acid, diazepam binding was 98.1 + 0.08% but with increasing concentrations of valproic acid the binding was 4-