Abstract
AIMS: To develop predictive formulas using short-term changes in glycaemic parameters [haemoglobin A1c (HbA1c) and fasting plasma glucose (FPG)] with sitagliptin, a highly selective dipeptidyl peptidase-4 inhibitor, to assess longer term steady-state changes in HbA1c. METHODS Results from two, 12-week, double-blind studies of sitagliptin in Japanese patients with type 2 diabetes mellitus receiving once-daily sitagliptin 100 mg were used to construct linear models to develop predictive formulas based on study 1 (S1) and to validate them using study 2 (S2). HbA1c and FPG were the primary and the key secondary end-point for both studies and were both used to develop predictive formulas. RESULTS The predictive formulas using HbA1c+/-FPG results (slope of change) from week 0 to week 4 in S1 showed high correlations between fitted and observed week 12 HbA1c: for HbA1c alone R2=0.76, for HbA1c+FPG R2=0.89. When using the sitagliptin 100 mg group of S2 data set to assess the validity of the predictive formulas, high correlations for HbA1c alone (R2=0.76) and for HbA1c+FPG (R2=0.77) were also observed. Data using a lower dose (25 mg once daily) of sitagliptin also demonstrated similar results. CONCLUSIONS The early responses (over 4 weeks) in HbA1c and FPG with sitagliptin can be used to accurately predict later responses (at week 12) in HbA1c in Japanese patients with type 2 diabetes mellitus. Additional studies applying this approach to other agents with diverse mechanisms are important.
