Ko 2006.

Methods	DURATION OF INTERVENTION: one year DURATION OF FOLLOW‐UP: one year RUN‐IN PERIOD: none LANGUAGE OF PUBLICATION: English
Participants	WHO PARTCIPATED: Chinese patients with type 2 diabetes and conventional oral antidiabetic drugs failure INCLUSION CRITERIA: OAD failure was defined as persistent hyperglycaemia with haemoglobin AIc (HbA1c) >= 8.5% for 6 mo or longer despite continuous use of maximal doses of conventional OAD; maximum recommended doses of various OADs were given as follows: glibenclamide 20 mg/d, gliclazide 320 mg/d, glipizide 20 mg/d, and metformin 3 g/d EXCLUSION CRITERIA: uncontrolled hypertension with sitting blood pressure (BP) >200/110 mm Hg and/or a history of myocardial infarction, cerebrovascular accident, or uncontrolled congestive heart failure during the previous 6 mo, or significant renal impairment (plasma creatinine concentration >= 150 mmol/L) DIAGNOSTIC CRITERIA: not stated CO‐MORBIDITIES: not stated CO‐MEDICATIONS: antihypertensive agents [no (%)]: I1: 31 (55.3) C1: 14 (25.0) lipid‐lowering agents [no (%)]: I1: 5 (8.9) C1: 2 (3.6)
Interventions	NUMBER OF STUDY CENTRES: 1 COUNTRY/ LOCATION: Hong Kong, China SETTING: Diabetic Clinic and Diabetes Center at AH Nethersole Hospital, in Tai PO, Hong Kong. INTERVENTION (DOSE/DAY): rosiglitazone max 8 mg/d CONTROL (DOSE/DAY): bedtime isophane insulin TREATMENT BEFORE STUDY: OAD ‐ original OAD and other medications remained the same throughout the study patients who fulfilled the inclusion criteria were referred to dietitians and diabetic nursing specialists for reinforcement of their dietary habits, drug compliance, and an understanding of OAD failure; those with HbA1c >=8.5% three months after reinforcement were included TITRATION PERIOD: oral rosiglitazone was started at 2 mg/d, insulin was begun at a dose of 6 units administered at night; the insulin dose was titrated 2 to 4 wk later by a diabetic nursing specialist with an increment of 2 to 4 units according to tolerability of the insulin injection and fasting plasma glucose (PG) improvement at 12, 24, 36, and 52 wk, all patients were seen for assessment of tolerability and compliance with treatment, and for measurement of lipid, glycemic, and other biochemical indices; insulin dosage was adjusted at each visit if this was deemed necessary, with the goal of achieving an HbA1c concentration <7.5%; if the drug was tolerable to patients, rosiglitazone was also increased to the maximum dose of 8 mg daily, with the goal of reducing HhA1c to <7.5% without the occurrence of significant hypoglycemia
Outcomes	PRIMARY OUTCOMES: not stated (differences in HbA1c) SECONDARY OUTCOMES: (not stated) lipids, BMI, FPG, blood pressure
Notes	AIM OF STUDY: to evaluate the efficacy and tolerability of rosiglitazone in patients with secondary oral anti‐diabetic drug failure and to directly compare rosiglitazone with bedtime insulin
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear