SF Marshall, HL Elliott, and PA Meredith
University Department of Medicine and Therapeutics, Western Infirmary, Glasgow, Scotland
Simvastatin is a HMG CoA reductase inhibitor which reduces total cholesterol (TC) through its principal effect on LDL cholesterol (LDLC). In a previous study dosing in the recommemded range (10-40mg) resulted in a maximum TC reduction of 31%1. The spread in achieved cholesterol concentrations was 3.3-7.7 mmol/l (95% CI) indicating marked inter-individual variation in the resultant response. It is possible that differential patient characteristics or concomitant drug treatments might contribute to this variability. The aim of this extension study was to identify and quantify possible covariate relationships within simvastatin effect on circulating lipoproteins.
Data were obtained from 41 hypertensive hypercholesterolaemic patients (10 M:31 F, 57 6 yrs) who received placebo, 10, 20 and 40 mg simvastain in random order each for 12 week treatment periods. TC and the major lipid subfraction were measured at 12 weekly intervals. Population analysis of dose response relationships was undertaken using first order estimation in NONMEM. To determine the most appropriate underlying model a hierachy of pharmacodynamic models models were fitted as previously shown 2, with nested intra and inter subject variability being estimated in each case. The most appropriate model was selected on the basis of the likelihood ratio test. When a Emax relationship was deemed superior, individual patient Emax and D50 values were estimated. If graphical relationships between individual structural parameter estimates and patients' covariates could be identified then they were formally tested by formation of full-reduced model pairs. The reductions in TC and LDLC levels were characterised as an Emax model with D50 values of 5mg and 6mg respectively. Forty milligrams simvastatin produced a fall equivalent to 90% of the maximum predicted response. Plots of parameter estimates against weight and placebo baseline cholesterol (PBC) showed some correlation and each improved the fit of the basic model. However, at most only a modest (14%) reduction in intersubject variability could be demonstrated. Where PBC was included in the Emax model, the limits of TC baseline range (5.8 & 11.4mmol/l) produce Emax estimatesof 25% and 45%. For LDLC, which had baseline limits of 3.6mmol/l & 7.9 mmol/l the corresponding estimates were 32% and 55%.
Using a population based approach it has proved possible to characterise the dose-response relationship for simvastatin and identify correlations with body weight and PBC. Although inclusion of these covariate relationships effects the clinical response, much of the intersubject variability still remains. In general, early investigation of the influence of covariates on the dose response relationship may provide useful information on the potential causes of treatment failure in routine clinical practice.
1) Farish.E, etal A double-blind twelve week palcebo controlled study to assess the efficacy of simvastatin in the treatment of hypercholesterolaemiain hyperstensive patients. J Drug Dev 1990;3(Suppl. 1):259-263.
2) Marshall.S.F, etal A population approach to dose versus response relationship for simvastatin in hypertensive hypercholesterolaemic patients. PAGE 94.