Summary Basis of Decision for Byetta ™

3.3.1 Pharmacodynamics

The pharmacodynamic (PD) studies consistently demonstrated that exenatide, 0.05 μg/kg or greater, decreased fasting and post-prandial plasma-glucose concentrations in diabetes mellitus type 2 patients. Moreover, exenatide lowered plasma-glucose levels in a dose-dependent fashion. Doses in the range of 0.05 to 0.2 μg/kg were shown to be optimal for glycaemic control. Doses of 0.02 μg/kg or less did not effectively reduce plasma-glucose levels while doses of 0.3 μg/kg or more were poorly tolerated. Considering that the average body weights in these studies ranged from 74 to 97 kg, a fixed dosing regimen of 10 µg was determined to achieve the best balance between effectiveness and tolerability.

Exenatide mediates its effects on plasma-glucose levels via multiple mechanisms of action. Exenatide delivered intravenously at steady state concentrations restored first-phase and augmented second-phase insulin release in type 2 diabetic patients, relative to placebo-treated controls. In some studies, exenatide was demonstrated to enhance glucose-dependent insulin secretion.

A PD study with healthy males appeared to support exenatide's ability to increase plasma-insulin secretion under euglycemic, but not hypoglycaemic, conditions. In a PD study with diabetes mellitus type 2 patients, exenatide produced a dose-dependent decrease in plasma-glucose levels under fasting conditions along with a concomitant increase in insulin levels. A similar finding was also reported for euglycemic healthy males under fasting conditions. However, when administered before a meal, at relatively low glucose concentrations in subjects with type 2 diabetes, exenatide markedly reduced postprandial glucose without changing insulin-plasma profiles. Under these conditions, in healthy subjects, exenatide either had no effect or decreased insulin secretion despite a marked suppression of plasma-glucose levels demonstrating that exenatide's effect of insulin secretion is glucose dependant. When exenatide was administered following a meal, insulin secretion was enhanced due to the physiologic rise in blood glucose post prandially. This may help exert greater glycaemic control, but has the potential to induce hypoglycaemia as low glucose values were observed in these healthy subjects. When exenatide is administered as recommended, within an hour preceding a meal, it would appear that insulin-independent mechanisms of action such as slowing of gastric emptying and a reduction in glucagon secretion could be the predominant contributors to glucose control with a minor contribution from insulin action.

Glucagon is secreted by pancreatic alpha cells in response to low plasma-sugar concentrations. It induces the hepatic conversion of stored glycogen into glucose and raises blood glucose concentrations. Exenatide was demonstrated to suppress plasma-glucagon concentrations under hyperglycaemic conditions in individuals with type 2 diabetes. The effect was generally modest and not dose-dependent, but was observed relatively consistently across the studies.

Another mechanism of action is the delayed rate of gastric-emptying. Because acetaminophen is predominantly absorbed from the small intestine, it was used as a representative oral medication in a number of clinical PD studies. Exenatide significantly and dose-dependently decreased plasma concentrations of acetaminophen relative to placebo. In a separate study using scintigraphy, 10 μg exenatide administered subcutaneously twice daily delayed the gastric emptying of both a solid and liquid test meal. The delay in gastric emptying may account for a very mild association of exenatide with a reduction in appetite; however, appetite suppression was generally very weak, inconsistent and could as well have been due to the gastrointestinal side-effects of exenatide.

3.3.2 Pharmacokinetics

Absorption

Exenatide is rapidly absorbed after subcutaneous administration to patients with Type 2 diabetes mellitus reaching median peak plasma concentrations in approximately 2.1 hours. Exenatide exposure (AUC parameters) increased proportionally over the therapeutic dose range of 5 µg to 10 µg. The rate and extent of absorption of exenatide after subcutaneous administration in the arm or thigh was essentially the same as that after subcutaneous administration in the abdomen.

Distribution

The mean apparent volume of distribution was 28.3 L following subcutaneous administration of a single dose of exenatide. The distribution volume is reflective of a relatively large tissue distribution. Data from a study using an ex vivo human placental perfusion system show that the potential for exenatide to cross the placental barrier is considered low. This supports the conclusion that exenatide used during pregnancy should result in minimal direct exposure of the peptide to the foetus.

Metabolism

In vitro studies using kidney membrane preparations demonstrated the potential for exenatide to be degraded within the kidney. In the presence of membrane preparations from mouse, rabbit, monkey, and human kidneys, exenatide was completely degraded after approximately 40 minutes.

Excretion

The kidney was the major route of elimination. Consistent with the major contribution of renal elimination to exenatide clearance, the mean exenatide half-life increased with greater degrees of renal impairment. Clinically relevant reductions in exenatide clearance were observed at creatinine clearance values of <30 mL/min.

Special Populations

Population pharmacokinetic analyses suggest that age, gender, race, and degree of obesity do not seem to influence the pharmacokinetic properties of exenatide.

In subjects with mild to moderate renal impairment (creatinine clearance 30 to 80 mL/min), exenatide clearance was reduced compared to subjects with normal renal function (13% reduction in mild and 36% reduction in moderate renal impairment). In subjects with end-stage renal disease receiving dialysis (creatinine clearance <30 L/hr), exenatide clearance was significantly reduced and exposure to exenatide was observed to be 3.37-fold higher in this group compared to the healthy group. No dosage adjustment is required in patients with mild renal impairment (creatinine clearance 50 to 80 mL/min); however, caution should be applied when initiating or escalating doses of Byetta from 5 µg to 10 µg in patients with moderate renal impairment (creatinine clearance 30 to 50 mL/min). Byetta should not be used in patients with end-stage renal disease or severe renal impairment (creatinine clearance <30 mL/min), including patients receiving dialysis.

Pharmacokinetic studies have not been conducted in subjects with a diagnosis of acute or chronic hepatic insufficiency. Hepatic dysfunction is not expected to affect blood concentrations of exenatide as exenatide is cleared primarily by the kidney.

3.3.3 Clinical Efficacy

The use of Byetta with metformin, with a sulfonylurea, and both metformin and a sulfonylurea to improve glycaemic control in patients with type 2 diabetes mellitus was primarily supported by three pivotal 30-week, randomized, double-blind, placebo-controlled, three-arm parallel, multicentre studies: Studies 112, 113, and 115, respectively; with corresponding uncontrolled, open-label, multicentre long-term extensions.

A total of 1,447 (Intent-to-treat patients = 1,446) patients were randomized in the three pivotal studies. After a 4-week placebo lead-in period, patients were randomly assigned to receive Byetta 5 µg BID, Byetta 10 µg BID, or placebo BID before the morning and evening meals, in addition to their existing oral antihyperglycemic agent (metformin and/or a sulfonylurea). All patients assigned to Byetta began a treatment initiation period with 5 µg BID for 4 weeks. After 4 weeks, those patients either continued to receive Byetta 5 µg BID or had their dose increased to 10 µg BID for 26 weeks. Patients assigned to placebo received placebo BID throughout the study. The primary efficacy endpoint in each of the 30-week studies was the mean change in HbA1c from baseline to 30 weeks. Several secondary efficacy endpoints were also assessed including changes in HbA1c at other timepoints, fasting plasma glucose, and body weight. The extension studies evaluated the long-term effects of subcutaneous Byetta 10 µg BID on efficacy and safety over 82 weeks. However, conclusions based on the extension period are limited due to the design limitations (open-label with no placebo or active treatment arm for comparison). Several non-pivotal studies were also submitted for additional support.

Use with Metformin

Study 112 [number of patients (n)=336] demonstrated significant mean reductions in HbA1c at Week 30 in the Byetta treatment arms, with changes from baseline of 0% for placebo, -0.46% for Byetta 5 µg, and -0.86% for Byetta 10 µg.

At Week 30, completers of Study 112 (n=225 ) were recruited into the extension study (Study 112E) andreceived open-label Byetta 5 µg BID for 4 weeks, followed by Byetta 10 µg BID for up to 82 weeks. Within the extension, Cohort I consisted of 150 patients previously on active treatment, and Cohort II consisted of 75 former placebo patients. In Cohort I, 124 patients completed Week 52 and 79 patients completed Week 82. In Cohort II, 60 patients completed Week 52 and 39 patients completed Week 82. In Cohort I, mean reductions in HbA1c from the original baseline were apparent at Week 30 (-0.7%) and were sustained through to Week 52 (-1.0%) and Week 82 (-1.3%). In Cohort II, the mean reductions in HbA1c from the extension baseline [that is (i.e.) Week 30 of Study 112] to Week 52 and Week 82 were -1.2% and -1.3%, respectively. In Cohort I, mean reductions in body weight were observed from the original baseline to Week 52 (-3.9 kg), Week 82 (-5.4 kg) and at all intermediate visits. For Cohort II, mean reductions in body weight were also observed from the extension baseline to Week 52 (-3.0 kg), Week 82 (-3.3 kg), and at all intermediate visits.

Use with a Sulfonylurea

Study 113 (n=377) demonstrated significant mean reductions in HbA1c at Week 30 in the Byetta treatment arms, with changes from baseline of -0.5% for Byetta 5 μg and -0.9% for Byetta 10 μg compared with an increase of 0.1% for placebo. Approximately 75-80% of patients in each of the three treatment arms used either concomitant glibenclamide (glyburide) or glipizide as the concomitant sulfonylurea during the study. Both glyburide and glipizide are approved in Canada, but glipizide is not marketed in Canada.

The extension to Study 113 (Study 113E) included 223 patients exposed to Byetta 10 µg BID. Cohort I consisted of 159 patients that were previously on active treatment and Cohort II consisted of 64 patients previously treated with placebo. Of the 223 patients, there were 127 (79.9%) Cohort I patients and 20 (31.3%) Cohort II patients exposed to Byetta for >54 weeks, and 61 (38.4%) Cohort I patients and 0 (zero) Cohort II patients exposed to Byetta for >82 weeks (includes duration of Byetta exposure in Study 113).

In Cohort I, mean reductions in HbA1c from the original baseline were apparent at Week 30 (-0.7 %) and were sustained through to Week 52 (-1.2 %) and Week 82 (-1.5%). In Cohort II, the mean reductions in HbA1c from the extension baseline i.e., Week 30 of Study 113] to Week 52 and Week 82 were -1.2% and -1.3%, respectively. In Cohort I, mean reductions in body weight were observed from the original baseline to Week 52 (-2.9 kg), Week 82 (-4.3 kg) and at all intermediate visits. For Cohort II, mean significant reductions in body weight were also observed from the extension baseline to Week 52 (-1.7 kg), Week 82 (-3.9 kg), and at all intermediate visits.

Use with both Metformin and a Sulfonylurea

Study 115 (n = 733) demonstrated significant mean reductions in HbA1c at Week 30 in the Byetta treatment arms, with changes from baseline of -0.66% for Byetta 5 µg BID, and -0.88% for Byetta 10 µg BID, compared with an increase of 0.12% for placebo.

The open-label extension to Study 115 (Study 115E) included 526 patients exposed to Byetta 10 µg BID. Cohort I consisted of 359 patients that were previously treated with Byetta and Cohort II consisted of 167 patients previously treated with placebo. Of the 526 patients, there were 293 (81.6%) Cohort I patients and 69 (41.3%) Cohort II patients exposed to Byetta for >54 weeks and 161 (44.8%) Cohort I patients and 0 (zero) Cohort II patients exposed to Byetta for 82 weeks (includes duration of Byetta exposure in Study 115). In Cohort I, mean reductions in HbA1c from the original baseline were apparent at Week 30 (-0.8%) and were sustained through to Week 52 (-1.0 %) and Week 82 (-1.0%). In Cohort II, the mean reductions in HbA1c from the extension baseline (i.e. Week 30 of Study 115) to Week 52 and Week 82 were -1.3% and -1.2%, respectively. In Cohort I, mean reductions in body weight were observed from the original baseline to Week 52 (-2.9 kg), Week 82 (-4.2 kg) and at all intermediate visits. For Cohort II, mean reductions in body weight were also observed from the extension baseline to Week 52 (-2.1 kg), Week 82 (-3.1 kg), and at all intermediate visits.

3.3.4 Clinical Safety

The clinical safety of Byetta in patients with Type 2 diabetes was primarily assessed in the three 30-week pivotal studies along with their extension studies (described in section 3.3.3 Clinical Efficacy).

Nausea was the most common treatment-emergent adverse event in all three studies: Study 112 (placebo 23.0%, Byetta 5 µg 36.4%, and Byetta 10 µg 45.1%); Study 113 (placebo 7.3%, Byetta 5 µg 39.2%, and Byetta 10 µg 51.2%); and Study 115 (placebo 20.6%, Byetta 5 µg 39.2%, and Byetta 10 µg 48.5%). Nausea was also observed to be the principal adverse event in the extension studies: Study 112E (Cohort I 44% and Cohort II 56%), Study 113E (Cohort I 39% and Cohort II 56%), and Study 115E (Cohort I 39% and Cohort II 58%).

Hypoglycaemia was the second most frequent adverse event in the Byetta treatment arms compared to placebo in the two pivotal studies in which Byetta was used in combination with a sulfonylurea-alone, or in combination with a sulfonylurea and metformin (Studies 113 and 115, respectively). The incidence of hypoglycemia in Study 113 was: placebo 3.3%; Byetta 5 µg 14.4%; and Byetta 10 µg 35.7%. The incidence of hypoglycemia in Study 115 was: placebo 12.6%; Byetta 5 µg 19.2%; and Byetta 10 µg 27.8%). Study 112, which used Byetta in combination with metformin did not demonstrate an increased incidence of hypoglycemia compared to placebo. The Product Monograph includes a recommendation that a decrease in the dose of sulfonylurea may be considered to reduce the risk of hypoglycaemia. The majority of cases of hypoglycaemia were considered mild to moderate in intensity.

Other common adverse events included vomiting, diarrhea, feeling jittery, dizziness, headache, and dyspepsia. The adverse reactions were usually mild to moderate in intensity. In all of the extension studies, there were no apparent adverse trends with respect to cardiovascular events, pancreatitis, neoplasms, or renal failure.

At Week 30, 43.2%, 40.8%, and 48.5% of the patients treated with Byetta in Studies 112, 113, and 115, respectively, had treatment-emergent anti-exenatide antibodies. The impact of the presence of these antibodies on the efficacy of Byetta has not been established.

Cardiovascular Events

Safety concerns were raised from results of a placebo-controlled study (GWCI) that evaluated the electrophysiological effects of a single 10 µg dose of exenatide on the 12-lead electrocardiogram in 62 healthy volunteers (thorough QT study). The results showed that a single subcutaneous dose of 10 µg was associated with statistically significant increases in heart rate and PR interval from 1 to 10 hours post-dosing, and statistically significant QTc interval prolongation from 1 to 3 hours post-dosing.

The maximum increase in mean heart rate was 10.21 beats per minute (bpm) at 2-hours post-dosing. The incidence of subjects with heart rate increases >90 bpm was 4.8% for exenatide versus 1.6% for placebo. In integrated long-term placebo-controlled trials, the incidence of treatment-emergent events of tachyarrhythmia (including supraventricular and ventricular tachyarrhythmias) was 0.6% for exenatide versus 0.5% for placebo.

The maximum increase in the mean PR interval occurred at 1-hour post-dosing and was 7.13 milliseconds (ms). The incidence of subjects with PR intervals >200 ms (i.e., first degree heart block) that appeared or worsened during treatment was 19% for the exenatide group versus 11% for the placebo group. Cases of second and third degree heart block and sudden cardiac death have been reported during post-marketing use of exenatide.

The QTc interval showed a maximum mean increase of 6.34 [90% confidence interval (CI) 4.12, 8.56] ms (QTcP) at 2 hours post-doing with exenatide. QTc prolongation in this study showed a positive correlation with exenatide concentration and a negative correlation with glucose concentration. In integrated long-term efficacy and safety studies, the incidence of treatment-emergent events for the Standardized MedDRA Query QT prolongation/torsade de pointes was 0.6% for exenatide versus 0% for placebo. Post-marketing events of ventricular fibrillation, cardiac arrest, and sudden cardiac death have been noted.

The sponsor agreed to submit results of another QT study using both therapeutic and supratherapeutic doses. Study results are expected by the end of 2011. In addition, the sponsor has indicated that an ongoing post-market cardiovascular outcomes study (EXSCEL) utilizing Bydureon will adjudicate arrhythmias related to major adverse cardiovascular events and heart failure with approaches in accordance with United States Food and Drug Administration (FDA) guidelines. Bydureon, the once weekly formulation of exenatide, is under review in the United States and Europe.

Pancreatitis

A review of clinical studies, post-marketing spontaneous experience, and published scientific literature cumulatively through August 31, 2008 was submitted. Although some cases of pancreatitis were observed in the exenatide clinical development program, the data did not suggest a difference in exposure-adjusted incidence rates of this condition between cohorts receiving Byetta compared to control groups receiving placebo or insulin; however, these data are limited by exposure and not powered for safety analysis. In this review, the incidence rate (per 1,000 person years of exposure) for pancreatitis events was 2.3 for Byetta, 2.7 for placebo, and 2.5 for insulin.

The post-market data provide a much wider exposure to exenatide than clinical trial data but are limited by poor and incomplete information collection. Cases were obtained primarily from approximately three-year and one-year marketing periods in the United States and the European Union, respectively (both ending August 2008). Overall, there were 251 pancreatitis events (79 acute, 8 necrotising, 1 haemorrhagic, 4 chronic, and 2 relapsing). Six of these cases were fatal (3 of which were classified as definite pancreatitis diagnoses). In addition, there were 52 cases identified for pancreatic enzyme abnormality. Based on the post-market data cumulative to August 31, 2008, an association between exenatide and pancreatitis is suspected.

As acute pancreatitis is potentially life-threatening, the Product Monograph includes warning information regarding the possible association of pancreatitis with Byetta, common symptoms of pancreatitis, recommendations to prescribers to consider alternative antidiabetic therapies in patients with a history of pancreatitis or other risk factors for pancreatitis, and a recommendation to promptly discontinue Byetta in cases of suspected pancreatitis. The sponsor has also proposed additional post-market actions to monitor this potential safety concern, including collecting information on the incidence of pancreatitis in an ongoing, long-term, post-market cardiovascular outcomes study (EXSCEL) utilizing Bydureon, the once weekly formulation of exenatide.

Neoplasms

In female rats given exenatide subcutaneously at a systemic exposure 130 times the human exposure (based on maximum recommended dose of 20 µg/day) for 2 years, there was an apparent numerical increase in benign thyroid C-cell adenomas. This incidence was not statistically significant when adjusted for survival. There was no tumorigenic response in male rats. There was no evidence of tumours in a 104-week carcinogenicity study in mice administered exenatide subcutaneously at systemic exposures up to 95 times the human exposure.

Other GLP-1 receptor agonists [Victoza (liraglutide) and Bydureon (the once weekly formulation of exenatide)] have been shown to cause thyroid C-cell tumours (adenomas and/or carcinomas) at clinically relevant exposures in rats and mice. The relevance of these results to humans has not yet been determined. Until further long-term data in humans are available, Byetta should not be used in patients with a personal or family history of medullary thyroid carcinoma or in patients with Multiple Endocrine Neoplasia syndrome type 2 (MEN 2).

To investigate the relationship between exenatide treatment and the potential for development of neoplasms, the sponsor performed a cumulative review of clinical trial data through to September 30, 2007. As of that date, 4,365 subjects had participated in completed company-sponsored long-term controlled and uncontrolled clinical studies. A total of 2,796 subjects (mean age 55.9 years) received exenatide. Of the remaining 1,569 subjects, 911 received placebo and 658 received comparator insulin. Of the 2,796 exenatide-treated subjects, 37 subjects (17 women and 20 men) reported cases of malignant neoplasms. There were 5 exenatide-treated subjects who experienced treatment-emergent adverse events identified by the MedDRA preferred terms: thyroid neoplasm (n=3), thyroid cyst (n=1), and benign neoplasm of the thyroid gland (n=1). The sponsor reports that none of the cases suggested a causal relationship to exenatide and that the rate of neoplasms in clinical trials had declined since the initial exenatide safety database integration in 2003.

After excluding pre-existing conditions and events of skin cancer (basal cell and squamous cell cancers) for comparison to the United States general population annual incidence rates, exposure-adjusted incidence rates for the 25 remaining cases of malignant neoplasms for exenatide-treated women (n=11) and men (n=14) in clinical trials were 875.2 and 861.6 per 100,000 subject years, respectively. These clinical data incidence rates are similar to the background incidence rates for women and men of comparable age in the United States general population (819.0 and 965.1 per 100,000 subject years, respectively). The sponsor concluded that "together, neither the clinical trial nor the spontaneous reports of neoplasms appears to provide a signal for relatedness to exenatide".

As of November 30, 2009, no clinical trial events of medullary thyroid carcinoma have been reported in users of Byetta. As well, there have been no post-market reports of medullary thyroid cancer in users of Byetta since market launch in the United States in 2005, although these data must be considered in light of the known under-reporting of spontaneous post-market adverse events.

The sponsor also performed a post-market, United States claims database signal detection analysis of over 29,000 patients using Byetta compared to a similar number of matched patients using metformin or glyburide between June 1, 2005 and June 30, 2008. The results of that study did not demonstrate a differential risk of thyroid neoplasm (benign or malignant) among patients initiating Byetta. However, during the course of this review, the sponsor subsequently advised Health Canada that the results are unreliable due to methodological issues related to the assessment tool (i3 Aperio) used in the study. The sponsor has therefore proposed to conduct an epidemiologic study to examine the potential risk of thyroid neoplasms in humans but these results are not anticipated until 2013. The sponsor has also proposed additional post-market actions to monitor this potential safety concern, including collecting information on incidence of thyroid and other malignancies, including pancreatic cancer, in an ongoing, long-term, post-market cardiovascular outcomes study (EXSCEL) utilizing Bydureon (the once weekly formulation of exenatide).

Until further long-term data in humans are available, warning information has been included in the Product Monograph indicating that other GLP-1 receptor agonists have been shown to cause thyroid C-cell tumours (adenomas and/or carcinomas) at clinically relevant exposures in rats and mice but the relevance of these results to humans has not yet been determined. As such, Byetta should not be used in patients with a personal of family history of medullary thyroid carcinoma or in patients with Multiple Endocrine Neoplasia syndrome type 2.

Acute Renal Failure

On November 2, 2009, during the course of this review, the United States FDA posted information for healthcare professionals related to post-market reports of altered kidney function in patients using Byetta. The United States FDA posting indicated that some, but not all, cases occurred in patients with pre-existing kidney disease or in patients with one or more risk factors for developing altered kidney function (including volume depletion related to diarrhea and/or vomiting, concomitant medical conditions or concomitant medications).

As well, the sponsor notes that although subjects with severe renal impairment (creatinine clearance ≤30 mL/min) have not been studied, clinically relevant reductions in exenatide clearance, accompanied by an increased intensity of gastrointestinal side effects, such as nausea and vomiting, were observed for subjects with end-stage renal disease (those receiving haemodialysis). In patients with end-stage renal disease receiving dialysis, mean exenatide clearance was reduced to 0.9 L/h compared with 9.1 L/h in healthy subjects. Furthermore, of the subjects in the long-term controlled trials, a significant proportion (between 30% and 60%) developed anti-exenatide antibodies. The sponsor notes that in population analyses, antibody-positive subjects were predicted, on average, to have mean exenatide clearance half the value observed in antibody-negative subjects. The sponsor attributes this phenomenon as likely due to the fact that the larger antibody-exenatide complex cannot be eliminated by renal filtration. It is therefore of additional concern that patients with end-stage renal failure could develop anti-exenatide antibodies and even higher exposures to exenatide as related to further reduction in exenatide clearance. As a result, Byetta is contraindicated in patients with end stage renal disease or severe renal impairment, and sufficient warnings and recommendations are stated in the Product Monograph.

The sponsor has also proposed additional post-market actions to monitor this potential safety concern, including collecting information on incidence of renal failure in an ongoing, long-term, post-market cardiovascular outcomes study (EXSCEL) utilizing Bydureon, the once weekly formulation of exenatide.