At least 30% of patients initially treated with sulphonylureas for NIDOM will have a poor response, and in the remaining 70% the subsequent failure rate is approximately 4% to 5% per year. BRL 49653C has a different mechanism of action to the sulphonylureas, and therefore the effects on fasting plasma glucose and Hb A1c are expected to be additive. Since circulatory insulin levels should decrease, and plasma glucose should be regulated, these combinations are also anticipated to slow both the progression of diabetic complications and delay the need for exogenous insulin.

The proposed study is intended primarily to determine the effectiveness of BRL 49653C by measure of glucose homeostasis as determined by Hb A1c and fasting plasma glucose, when added to sulphonylurea therapy (sulphonylureas are limited to: glibenclamide, glipazide and gliclazide). In addition, the clinical safety of BRL 49653C will be assessed in this patient population. The starting doses have been selected based on dose response studies examining safety, tolerability and efficacy in the U.S.A.

Non-insulin dependent diabetes mellitus (NIDDM) is the most common form of the hyperglycaemic state. Although NIDDM exists in all populations, its prevalence varies greatly, depending upon age and race, and appears to be increasing. It is estimated that NIDOM occurs in 5% of 50-year-old-adults and 20% of the 80-year-old population with a linear progression between these age groups. Aberrant insulin secretion from pancreatic p-cells, and insulin resistance in peripheral tissues with reduced glucose uptake (particularly in muscle and liver) are the primary pathogenic mechanisms believed to be responsible for the metabolic abnormalities associated with NIDDM. Insulin resistance progresses to overt diabetes at the rate of approximately 5% of patients per year.

Patients with insulin resistance, but no overt diabetes mellitus, are generally overweight, usually have a sedentary lifestyle, and frequently have a family history of diabetes mellitus. Insulin-resistant phenotype is commonly associated with cardiovascular risk factors such as hypertension and dyslipidemia (decreased HDL and increased VLDL, LDL and triglycerides). Current available therapies for NIDOM include insulin and oral hypoglycemic agents (e.g. sulphonylurea compounds and biguanides). Sulphonylureas achieve blood glucose control by enhancing pancreatic insulin secretion. However, a significant failure rate does exist for this mode of treatment, and many patients who have received long term treatment with sulphonylureas do require exogenous insulin therapy when there is no longer adequate pancreatic insulin secretion (even in response to sulphonylureas). Single therapy with the biguanide, metformin, or combination therapy with an additional sulphonylurea may also be administered to NIDDM patients. However, the overall effect of metformin on glycemic control seems to be smaller than that of sulphonylureas, and in many cases, combination therapy does not stop the need for insulin after a period of time.

BRL 49653C is a highly potent and orally active anti-diabetic thiazolidinedione compound. The thiazolidinediones such as BRL 49653C represent a novel class of compounds that act as "insulin sensitizers" in target tissues. Unlike sulphonylureas, BRL 49653C improves the sensitivity of muscle and liver to insulin. An agent that enhances insulin sensitivity may be effective in regulating elevated blood glucose in patients with NIDDM and theoretically may be beneficial in reducing the frequency and severity of morbid complications of this disease. BRL 49653C does not increase pancreatic insulin secretion; therefore, treatment with BRL 49653C should not have the potential to exhaust pancreatic function, or produce the untoward effect of hypoglycemia (a potential complication of treatment with sulphonylureas).

In animal testing, repeat oral administration of BRL 49653C has been effective in reducing hyperglycemia, hyperinsulinemia, hypertriglyceridemia, and free fatty acids associated with insulin-resistant models of NIDDM2. In the genetically obese mouse, the potency of BRL49653C as an anti-hyperglycemic agent, was compared with competitor insulin sensitizers, and BRL 49653C was proven to be at least 100 times more potent than CS-045 (troglitazone) and at least 30 times more potent than pioglitazone.

Six month toxicological findings were similar to those observed in one month studies. Please refer to the Investigator Brochure for more detailed information about pre-clinical studies and the pharmacology, metabolism, and toxicology of BRL 49653C.

To date, seven completed Phase I studies have shown BRL 49653C to be safe and well tolerated. The levels of exposure to BRL 49653C have been shown to be similar in both the fed and fasted state, and pharmacokinetic analysis has shown no evidence of accumulation after 10 days once daily dosing of 1 to 5 mg in obese volunteers. A total of 140 volunteers received BRL 49653C ranging in doses of 0.2 to 20 mg (20 mg dose is approximately 4 to 5-fold higher than doses expected to be administered to NIDDM patients). Of these, 140 subjects, 91 received one or more single doses of BRL 49653C up to 20 mg. This includes 10 healthy elderly males (>= 65 years of age). Additionally, 49 obese subjects have received dosages of up to 5 mg/day for 10 consecutive days.

SmithKline Beecham Phase I clinical studies have shown BRL 49653C to be safe and well tolerated. The most frequently observed adverse experience was headache. Serious adverse experiences of transient asymptomatic accelerated unifocal ventricular rhythm of less than 10 seconds duration were observed in two volunteers. The arrhythmias were noted at 27 minutes and 2 hours and 20 minutes after administration of a single dose of BRL 49653C 2mg. Hotter monitoring was performed in two studies in order to evaluate the arrythmogenicity of BRL 49653C. Study 002, was a placebo controlled, 10 day repeat dose, two-period, period balanced, cross-over study, within each of three dose groups (1, 2 or 5 mg) of BRL 49653C. Study 016 was a single dose, double-blind, randomized (with respect to order of placebo), placebo controlled, five period cross-over, oral dose rising study in which each subject received a randomly allocated placebo and four doses of BRL 49653C (5, 10, 15, 20 mg). In both of these studies, there was no evidence of protocol- defined incidence of ventricular tachycardia, ventricular premature beats, or supraventricular tachycardia.

A recent study with another thiazolidinedione compound, troglitazone, has shown that there is no increase in cardiac mass or cardiac function impairment in patients with NIDDM . To the contrary, these patients were observed to have decreases in systemic vascular resistance, with enhanced cardiac output and stroke volume. Other studies examining thiazolidinediones have shown beneficial effects on Hemoglobin A1c with statistically significant decreases indicating beneficial effects on glucose homeostasis over periods of time.

BRL 49653C has also been examined in combination with glyburide (glibenclamide). BRL 49653C was administered 2mg bd to NIDDM patients receiving stable doses of glyburide for seven days. No changes were observed in either the 24 hour glucose or insulin profiles, suggesting that the combined treatment is not associated with any acute pharmacokinetic interaction that would likely result in an acute untoward pharmacodynamics effect.

The safety, tolerability and efficacy of BRL 49653C in patients with NIDDM has been examined in a phase II, double-blind, placebo controlled dose response study. A total of 380 patients received either placebo, 005mg bid, 0.25mg bid, 1mg bid or 2mg bid for 12 weeks. Analysis of variance showed highly significant (p<0.001) treatment effects on fasting blood glucose from week 2 onwards, although it was clear that a plateau in response had not been achieved at the highest dose tested. At the higher dose of 2mg bid, there were statistically significant differences compared to placebo, at week 12 for fasting plasma glucose (-40.1 mg/dL), fructosamine (-27.1 mmol/L) and insulin (-16%), and a significant reduction in hemoglobin Alc (0.4%). The lowering of fasting glucose in the 2mg bid group with a concomitant 16% decrease in insulin levels supports the hypothesis of insulin sensitization as a mechanism of action for BRL 49653.

The drug was found to be safe with no adverse effects on vital signs, ECG or clinical chemistry. 2D M-Mode echocardiography revealed no on left ventricular mass effects, and the small reductions in hemoglobin 14.5 to 14.3 g/100ml in the highest dose group) and hematocrit (from 42.6% to 41.8%) are not regarded as being clinically significant. Furthermore, BRL 49653C was found to be well tolerated over the dose range 0.05-2.0mg bid with an adverse event profile that was similar to placebo both qualitatively and quantitatively.

During the study a total of 18 serious adverse events (sAE) were reported, 9 during the screening/run-in phase and 9 during the 12 week randomization period. There were 3 sAE in each of the placebo and 2.0mg BRL 49653 groups and I in each of the 0.05, 0.25 and 1.0mg groups. There were 2 deaths during the study, one case of myocardial infarction (2.0mg group) in a 52 year old male was designated as possibly related, and the other case (0.25mg group), a 71 year old male with cardiopulmonary arrest, was unrelated. All other sAE were specified by the investigators as being unrelated to trial medication, apart from 1 case of exercise induced non-sustained ventricular tachycardia in a 56 year old male receiving placebo. Causal relationship was given as possibly related in this case.

At least 30% of patients initially treated with sulphonylureas for NIDDM will have a poor response, and in the remaining 70% the subsequent failure rate is approximately 4% to 5% per year. BRL 49653C has a different mechanism of action to the sulphonylureas, and therefore the effects on fasting plasma glucose and Hb A1c are expected to be additive. Since circulatory insulin levels should decrease, and plasma glucose should be regulated, these combinations are also anticipated to slow both the progression of diabetic complications and delay the need for exogenous insulin.

The proposed study is intended primarily to determine the effectiveness of BRL 49653C by measure of glucose homeostasis as determined by Hb A1c and fasting plasma glucose, when added to sulphonylurea therapy (sulphonylureas are limited to: glibenclamide. glipazide and gliclazide). In addition, the clinical safety of BRL 49653C will be assessed in this patient population. The starting doses have been selected based on dose response studies examining safety, tolerability and efficacy in the USA.

Please refer to the Investigator Brochure for a review of the pre-clinical and clinical studies of BRL 49653C.

Inclusion Criteria:

• Men or women between 30-80 years of age inclusive at time of enrolment.
• Patients who had non-independent diabetes mellitus (NIDDM) defined by the criteria of the National Diabetes Data Group.
• Patients who had sulphonylurea therapy for at least 6 months and a constant dose for at least 2 months prior to visit 1.
• Patients who had fasting plasma glucose <= 15.0 mmol/L at screening. Hemoglobin A1c >= 7.5%.
• Female patients must be (1) post-menopausal, i.e. > 6 months without menstrual period, surgically sterile, or (2) using hormonal contraceptives or intrauterine contraceptive devices. Female patients who were taking hormonal contraceptives must also use an additional barrier form or intrauterine form of birth control.
• Patients who had given their written informed consent to participate.

Exclusion Criteria:

• Female patients who were pregnant, breast feeding or planning a pregnancy during the course of the study.
• Patients who had a fasting plasma glucose > 15.0 mmol/L at screening, or severity of diabetes mellitus requiring administration of insulin, or patients with ketonuria.
• Patients who had clinically significant renal or hepatic disease (i.e., patients with serum creatinine > 160 micromol/L (1.8 mg/dL); ALT, AST, total bilirubin, gamma GT, or alkaline phosphatase more than 2.5 times the upper limit of the normal laboratory range).
• Any clinically significant abnormality identified on the screening physical examination, laboratory tests, electrocardiogram which in the judgment of the investigator would preclude safe completion of the study.
• Patients who had leukocyte count < 3000/mm3 or platelet count <120,000/mm3.
• Systolic blood pressure >180mmHg or diastolic blood pressure >114mmHg while on appropriate hypertensive therapy.
• Significant anemia (hemoglobin < 11 g/dL for males or < 10g/dL for females) or diagnosis of porphyria.
• Symptomatic diabetic neuropathy of sufficient severity to require treatment for control of symptoms (eg, painful peripheral neuropathy, symptomatic orthostatic hypotension, urinary retention, gastric stasis, pedal ulcers).

Diabetic retinopathy imminently requiring treatment for preserving or restoring vision.

• Body mass index(BMI) < 22 and >38 kg/m2 (Formula: BMI= weight, kg ÷height, m2)and variation in body weight of >=5% between screening and visit2.