J. Bennet(1) , LF.Lacey(2) , K. Koch(3) , A. Bye(2)
Glaxo funded PhD student at the Dept. of Mathematics,
Imperial College, London(1)
Clinical Pharmacokinetics, Glaxo Research and Developement
Ltd., Greenford,UK(2)
Clinical Pharmacology, Glaxo Research Institute, North
Carolina, U.S.A.(3)
GR122311X (ranitidine bismuth citrate) is a new chemical entity being developed for the treatment of peptic ulcer disease. Absorption of large amounts of bismuth from salts such as bismuth subnitrate and subgallate can produce toxicity in the form of encephalopathy, neuropathy and osteoarthropathy. There is very considerable inter-subject variability in the pharmacokinetics of bismuth from orally administered bismuth salts such as tri-potassium di- citrato bismuthate. However, very little work to-date has been carried out to investigate the population pharmacokinetics of bismuth.
The pharmacokinetics of bismuth from GR122311X were investigated in a 28 day repeat dose study. Eighteen subjects received 800mg GR122311X twice-daily for 28 days. Blood samples were collected during treatment to determine trough plasma bismuth concentrations and to determine full plasma profiles on days 1, 14, and 28. In addition, plasma bismuth concentrations were determined up to 168 days post-treatment. Plasma bismuth concentrations were determined using a validated inductively coupled plasma mass spectrometry method, with a lower limit of quantification of 0.2ng/ml. Two and three exponential disposition population pharmacokinetic models, with first-order rate of absorption (2exp and 3exp, respectively) were fitted to the data using the Gibbs sampler Bayesian method. This is an example of applying the population approach to a data rich situation, involving a drug with highly variable plasma bismuth concentrations. This work was carried out in order to better quantify the inter-subject variability in bismuth's pharmacokinetics and to predict the population distribution of bismuth's pharmacokinetic parameters (such as Cmax) at steady-state following dosing with GR122311X.
The 3exp model described the data better than the 2exp model. The terminal bismuth half-life was estimated to be approximately 21 days. The model predicted that steady-state plasma bismuth concentrations would be achieved after approximately 3 months of dosing and that the 95 percentile steady-sate Cmax, during dosing with GR122311X 800mg bid, would be approximately 15ng/ml, which is of no clinical concern. There was considerable inter-subject variability in the model's parameters, especially for the exponential coefficients, which had a coefficient of variation (CV) of greater than 50%. The absorption of bismuth from GR122311X is rapid (tmax approximately 30min). There was little data to describe the absorption phase and estimate the absorption rate constant (ka) for each individual subject, making it difficult to estimate individual values for ka using traditional modelling methods. The estimate obtained using the population approach (ka=8.7min, inter-subject CV=67%) should be a more reliable estimate.