We examined the pharmacokinetic (PK) and pharmacodynamic (PD) rationales to develop controlled release (CR) formulations of metformin. Unrestrained diabetic rats received the drug as intravenous bolus (i.v.), oral solution (p.o.), intra-duodenal bolus, 4-h infusion, or intra-colonic bolus. In addition, we developed two CR-gastroretentive dosage forms (CR-GRDF) that released the drug over 3 or 6 h (in vitro), and retained in the rats' stomach for 8-10 h. Metformin exhibited flip-flop PK. The colonic absorption was low but sustained and was associated with highly variable glucose-lowering effects, thus providing a PK rationale to develop CR-GRDF. In addition, the glucose-lowering effect was greater following p.o. vs. i.v. administration, despite equivalent AUC, indicating a first pass PD effect, thus, adding a PD rationale to develop metformin CR-GRDF. When administered to the diabetic rats, CR-GRDFs produced bioavailability and extent of glucose-lowering effects that were similar to those of the duodenal infusion and p.o. metformin administration. These findings are attributed to the adsorption of metformin to the intestine that yields slow and prolonged absorption even following p.o. administration of drug solution. The data indicates that unless the CR formulation could significantly extend the absorption period, it is not likely to improve glucose-lowering efficacy.