METHODS

Design:

randomised controlled trial.

Allocation:

concealed.*

Blinding:

unblinded.*

Follow up period:

12 weeks.

Setting:

48 outpatient centres in the US and Canada.

Patients:

309 patients 35–80 years of age (mean age 57 y, 65% men) with ⩾1 year history of type 2 diabetes treated with a stable oral agent regimen (1 insulin secretagogue and 1 insulin sensitiser) and HbA_1c of 8–11%. Exclusion criteria included predisposition to severe hypoglycaemia; hospital admission or emergency department visit for poor diabetic control in previous 6 months; …