Summary Basis of Decision for Forxiga

 

Clinical Pharmacology

Type 2 diabetes mellitus (T2DM) is a metabolic disorder characterised by high levels of blood glucose due to insulin resistance and/or insulin deficiency. Dapagliflozin, the active ingredient in Forxiga, is a reversible inhibitor of sodium-glucose co-transporter 2 (SGLT2) that improves glycemic control in patients with T2DM by reducing renal glucose reabsorption leading to urinary excretion of excess glucose.

The clinical pharmacology included reports on the human pharmacodynamic and pharmacokinetic studies. The clinical pharmacological data support the use of Forxiga for the specified indication.

Oxidative metabolism accounted for the elimination of <10% of the dose in humans, so metabolism mediated by cytochrome P450 enzymes is unlikely to be an important factor in the disposition of dapagliflozin. Dapagliflozin is a weak in vitro substrate for P-glycoprotein (P-gp) and is not an inhibitor/inducer of P-gp transporter. Thus, it does not have a strong potential for pharmacological interactions and no meaningful interactions were observed in vivo.

Steady-state systemic exposures of dapagliflozin in patients with T2DM and mild, moderate or severe renal impairment were correspondingly higher than patients with normal renal function, and renal glucose clearance was meaningfully decreased in patients with moderate and severe renal impairment. Dapagliflozin is contraindicated in patients with moderate to severe renal impairment as specified in the labelling.

In healthy subjects and in patients with T2DM, an increase in the amount of glucose excreted in the urine was observed following the administration of dapagliflozin at doses ≥0.3 mg. In patients with T2DM, dapagliflozin treatment at 10 mg/day for 12 weeks was associated with a urinary glucose excretion of approximately 70 g/day. Maximal increase in 24 hour urinary glucose excretion was observed at doses ≥20 mg/day. Dapagliflozin results in osmotic diuresis and increases in urinary volume. In patients with T2DM, urinary volume increases were sustained for 12 weeks (approximately 375 mL/day). Increases in urinary sodium excretion were small, transient (2-3 days), and not associated with changes in serum sodium concentrations.

The initial New Drug Submission (Control #141791) included comparative bioavailability data to demonstrate that the change in dapagliflozin crystalline form induced by heat does not affect the in vivo performance. In addition, the effect of food on the bioavailability of dapagliflozin was evaluated. A high-fat meal decreased maximal dapagliflozin concentrations by 31% but did not affect the extent of absorption, and was not considered to be clinically significant. While the comparative bioavailability data provided were considered acceptable, revision to the Pharmacokinetics section of the Product Monograph was requested in the Notice of Non-Compliance (NON), which was revised.

For further details, please refer to the Forxiga Product Monograph, approved by Health Canada and available through the Drug Product Database.

Clinical Efficacy

The clinical efficacy of Forxiga for the approved indications was demonstrated across five Phase III pivotal studies, both as a monotherapy (MB102013) and in combination with other anti-diabetic medications, including metformin (MB102014 and D1690C00004), a sulfonylurea (glimepiride) (D1690C00005), and insulin (D1690C00006). The pivotal studies were each randomized, parallel group and multicentre, and all but one included a short-term double-blind treatment period of 24 weeks. The active comparator study D1690C00004 had a short-term period of 52 weeks. In each of the pivotal studies, the short-term treatment period was followed by one or more long-term extension periods. The primary efficacy outcome for all pivotal studies was the change in hemoglobin A1c (HbA1c, glycated hemoglobin) from baseline to the end of the pre-specified short term treatment period, with the last observation carried forward (LOCF) for patients that required rescue treatment. Secondary outcome measures varied by study and included fasting plasma glucose (FPG), body weight, and 2-hour post-prandial glucose (PPG).

Supporting studies were conducted in T2DM patients with moderate renal impairment (MB102029) and cardiovascular disease (D1690C00018 and D1690C00019). Body composition and bone mineral density in T2DM patients treated with Forxiga were also studied (D1690C00012).

Monotherapy

Study MB102013 was a multicentre, randomized, double-blind placebo-controlled study designed to evaluate the efficacy and safety of Forxiga 2.5 mg, 5 mg, and 10 mg once daily as monotherapy in adults with T2DM. Following a 2-week diet and exercise placebo lead-in period, 485 patients with HbA1c ≥7% and ≤10% were randomized to Forxiga 2.5 mg, 5 mg, or 10 mg once daily in either the morning (QAM, main cohort) or evening (QPM, exploratory cohort), or placebo in the morning only. This study was designed with a 24-week primary efficacy endpoint (AM dosing group) with an extension of 78 weeks (to 102 weeks total) for additional safety and exploratory efficacy data. The primary efficacy outcome was change in HbA1c with LOCF from baseline to Week 24 excluding data after rescue, with secondary outcomes of change in FPG and body weight in the same time period.

In the primary analysis for the QAM dosing group, there were statistically and clinically significant improvements in HbA1c from baseline to Week 24 for both the Forxiga 5 mg and 10 mg groups versus (vs.) placebo (p-value <0.001). Overall, the PM administration of Forxiga had a comparable efficacy profile to Forxiga administered in the AM. In the QAM group, the placebo-subtracted changes in HbA1c were -0.54% [95% confidence interval (95% CI): -0.84%, -0.24%] and -0.66% (95% CI: -0.96%, -0.36%) in the Forxiga 5 mg and 10 mg groups, respectively. Change in FPG from baseline to Week 24 was also statistically and clinically significant for both the Forxiga 5 mg and 10 mg QAM dosing groups versus placebo, with placebo-corrected reductions of -1.1 mmol/L (95% CI: -1.7 mmol/L, -0.5 mmol/L) and -1.4 mmol/L (95% CI: -2.0 mmol/L, -0.8 mmol/L) in the Forxiga 5 mg and 10 mg groups, respectively.

The placebo-subtracted mean change from baseline in body weight at Week 24 in the Forxiga 5 mg and 10 mg QAM groups were not statistically or clinically significant: -0.65 kg (95% CI: -1.90 kg, 0.61 kg) and -0.97 kg (95% CI: -2.20 kg, 0.25 kg), respectively. Proportions of patients in Forxiga treatment groups vs. placebo who achieved HbA1c <7% when adjusted for baseline HbA1c were not significant based on the sequential testing procedure: 12.6% and 19.2% of patients in the 5 mg and 10 mg QAM groups, respectively.

The long-term efficacy objectives were exploratory in this study, and no hypothesis testing was performed. As well, interpretation of the combined short-term and long-term exploratory efficacy results are complicated by addition of standard doses of metformin after Week 24 in all placebo patients not previously rescued in the short-term period. Nevertheless, these supplementary data do provide some support for maintenance of efficacy for the 5 mg and 10 mg doses. Mean change from baseline in HbA1c was -0.70% at Week 102 in the QAM group, for patients treated with Forxiga 5 mg, -0.61% for patients treated with Forxiga 10 mg, and -0.17% for patients treated with placebo based on the longitudinal repeated measures analysis excluding data after rescue.

Add-on Combination Therapy with Metformin

Study MB102014 was a multicentre, randomized, double-blind, placebo-controlled study designed to evaluate the efficacy and safety of Forxiga 2.5 mg, 5 mg, and 10 mg once daily as an add-on to metformin in adults with T2DM. Following a 2-week single-blind placebo lead-in period, 546 patients with HbA1c ≥7% and ≤10% and who were taking metformin at a dose of at least 1,500 mg per day were randomized to Forxiga 2.5 mg, 5 mg, or 10 mg or placebo in addition to their current dose of metformin. This study was designed with a 24 week primary efficacy endpoint with a 78 week extension. The primary efficacy outcome was change in HbA1c with LOCF from baseline to Week 24, excluding data after rescue, with secondary outcomes of change in FPG and body weight in the same time period.

Efficacy was supported by statistically (p<0.0001) and clinically significant results for the primary endpoint. The mean placebo-subtracted changes from baseline in HbA1c at Week 24 were -0.41% (95% CI: -0.61%, -0.21%) and -0.54% (95% CI: -0.74%, -0.34%) in the Forxiga 5 mg and 10 mg groups, respectively. The mean placebo-subtracted changes in FPG from baseline to Week 24 were -0.9 mmol/L (95% CI: -1.3 mmol/L, -0.5 mmol/L, p<0.0001) and -1.0 mmol/L (95% CI: -1.4 mmol/L, -0.6 mmol/L, p<0.0001) in the Forxiga 5 mg and 10 mg groups, respectively. The placebo-subtracted mean proportion of patients achieving a target HbA1c of <7% at Week 24 were 11.6% (p<0.05) and 14.7% (p<0.05) for the Forxiga 5 mg and 10 mg groups, respectively. Statistically significant (p<0.0001) differences were observed in mean placebo-subtracted change from baseline in body weight at Week 24 for all Forxiga dosing groups: -2.16 kg (95% CI: -2.81 kg, -1.50 kg) and -1.97 kg (95% CI: -2.63 kg, -1.31 kg) in the Forxiga 5 mg and 10 mg groups, respectively.

The long-term efficacy objectives were exploratory, but did provide some support for maintenance of efficacy. At Week 102, mean change from baseline in HbA1c was -0.58% for patients treated with Forxiga 5 mg plus metformin, -0.78% for patients treated with Forxiga 10 mg plus metformin and 0.02% for patients treated with placebo plus metformin based on the longitudinal repeated measures analysis excluding data after rescue. However, interpretation of these data is complicated by the high number of rescues over the course of the study and especially in the placebo group (60% placebo; 43.9% Forxiga 5 mg; and 44% Forxiga 10 mg) resulting in a sample of patients less representative of the original study population.

Study D1690C00004 was a multicentre, randomized, double-blind, active-controlled study designed to evaluate the efficacy and safety of Forxiga 10 mg as add-on therapy to metformin compared with glipizide as add-on therapy to metformin in patients with T2DM who have inadequate glycemic control on metformin therapy alone. Following a 2-week placebo lead-in period, 816 patients with HbA1c >6.5% and ≤10% and who were taking metformin at a dose of at least 1,500 mg per day were randomized to either 5 mg glipizide or 2.5 mg Forxiga and were up-titrated over 18 weeks to optimal glycemic effect (FPG <110 mg/dL, <6.1 mmol/L) or to the highest dose level (up to 20 mg glipizide or 10 mg Forxiga). At the end of the titration period, 87% of patients treated with Forxiga had been titrated to the maximum study dose (10 mg) vs. 73% treated with glipizide (20 mg). This study was designed with a 52 week primary efficacy endpoint with an extension of 52 weeks for additional safety and exploratory efficacy data. The primary efficacy outcome was change in HbA1c with LOCF from baseline to Week 52, with secondary outcomes of effect on body weight, hypoglycemic events and total body weight reduction in the same time period. The primary analysis demonstrated that as add-on to metformin therapy, the mean change in HbA1c from baseline to Week 52 for the Forxiga group (-0.52%) was non-inferior to the glipizide group (-0.52%) using the pre-defined non-inferiority margin of 0.35%, as the upper limit of the 95% CI of the difference in mean change from baseline between treatment groups was 0.11%. While the primary analysis was not conducted with the per protocol analysis set, the results of the primary analysis using the full analysis set agreed with those of the per protocol sensitivity analysis which showed a change from baseline to Week 52 of -0.55% for Forxiga and -0.56% for glipizide (95% CI: -0.12%, 0.12%).

Key secondary efficacy variables included effect on body weight, hypoglycemic events and total body weight reduction. All three of these variables favoured the Forxiga group over glipizide, as expected given that sulfonylureas are associated with weight gain and hypoglycemia. The Forxiga group was found to have a mean change in body weight from baseline to Week 52 of -3.22 kg compared to a mean change of +1.44 kg in the glipizide group. Over the 52 week period, 39.7% of glipizide patients experienced at least one episode of hypoglycemia compared to 3.4% of Forxiga patients. The proportion of patients with body weight reduction of at least 5% at Week 52 was 33.3% (95% CI: 28.7%, 37.9%) for Forxiga patients compared to 2.5% (95% CI: 1.0%, 4.0%) for glipizide patients.

The long-term efficacy objectives were exploratory, but did provide some support for maintenance of efficacy. At Week 104, mean change from baseline in HbA1c was -0.32% for patients treated with Forxiga and - 0.14% for patients treated with glipizide based on the longitudinal repeated measures analysis.

Add-on Combination Therapy with a Sulfonylurea (Glimepiride)

Study D1690C00005 was a multicentre, randomized, double-blind, placebo-controlled study designed to evaluate the efficacy and safety of Forxiga 2.5 mg, 5 mg, and 10 mg once daily in combination with glimepiride (a sulfonylurea) in adults with T2DM who have inadequate glycemic control. A total of 597 patients with HbA1c ≥7% and ≤10% and who were taking glimepiride at a dose of at least 4 mg per day were randomized to Forxiga 2.5 mg, 5 mg, or 10 mg or placebo in addition to 4 mg glimepiride. This study was designed with a 24 week primary efficacy endpoint with a 24 week extension. The primary efficacy outcome was change in HbA1c with LOCF from baseline to Week 24 excluding data after rescue, with secondary outcomes of change in FPG, 2-hour PPG, and change in body weight in the same time period.

Efficacy was supported by statistically and clinically significant results for the primary endpoint for both the Forxiga 5 mg and 10 mg dose groups; placebo-subtracted mean changes from baseline in HbA1c at Week 24 were -0.49% (p<0.0001, 95% CI: -0.67%, -0.32%) and -0.68% (p<0.0001, 95% CI: -0.86%, -0.51%) in the Forxiga 5 mg and 10 mg groups, respectively.

Mean placebo-subtracted changes in FPG from baseline to Week 24 were -1.1 mmol/L (p<0.0001, 95% CI: -1.5 mmol/L, -0.7 mmol/L) and -1.5 mmol/L (p<0.0001, 95% CI: -1.9 mmol/L, -1.1 mmol/L) for Forxiga 5 mg and 10 mg, respectively. The placebo-subtracted proportions of patients achieving a target HbA1c of <7% at Week 24 were 17.3% and 18.7% for Forxiga 5 mg and 10 mg, respectively. Statistically significant differences from placebo were observed in mean change from baseline in body weight at Week 24 for both groups. The placebo-subtracted differences in body weight were -0.84 kg (p = 0.0091, 95% CI: -1.47 kg, -0.21 kg) and -1.54 kg (95% CI: -2.17 kg, -0.92 kg) in the Forxiga 5 mg and 10 mg groups, respectively.

The long-term efficacy objectives were exploratory, but did provide some support for maintenance of efficacy. At Week 48, mean change from baseline in HbA1c was -0.56% for patients treated with Forxiga 5 mg plus glimepiride, -0.73% for patients treated with Forxiga 10 mg plus glimepiride and -0.04% for patients treated with placebo plus glimepiride based on the longitudinal repeated measures analysis excluding data after rescue.

Add-on Combination Therapy with Insulin

Study D1690C00006 was a randomized, double-blind, placebo-controlled study designed to evaluate the efficacy and safety of Forxiga 2.5 mg, 5 mg, and 10 mg once daily as an add on to insulin, alone or in combination with up to two oral anti-diabetic medications (OADs) in adults with T2DM. Following a 2-week enrollment period, 808 patients with HbA1c ≥7.5% and ≤10.5% and who were on a stable insulin regimen with up to 2 OADs were randomized to Forxiga 2.5 mg, 5 mg, or 10 mg or placebo in addition to their current dose of insulin and other OADs, if applicable. A stable insulin regimen was considered to be a mean dose of at least 30 IU of injectable insulin a day for a period of at least 8 weeks prior to the study. This study was designed with a 24 week primary efficacy endpoint with an 80-week extension period. The primary efficacy outcome was change in HbA1c with LOCF from baseline to Week 24 excluding data after rescue, with secondary outcomes of change in FPG and body weight in the same time period. A variety of human insulin products were used by patients in this study. Patients taking combinations other than insulin alone or in combination with metformin were included in this study, but their numbers were insufficient to support inclusion in the authorized indication.

The efficacy of Forxiga 5 mg and 10 mg doses was supported by statistically and clinically significant results for the primary endpoint. The placebo-subtracted mean changes from baseline in HbA1c at Week 24 were -0.52% (p<0.0001, 95% CI: -0.66%, -0.38%) and -0.60% (p<0.0001, 95% CI: -0.74%, -0.45%) in the Forxiga 5 mg and 10 mg groups, respectively. The subgroup analysis indicated a treatment-by-subgroup interaction for baseline HbA1c. The efficacy of Forxiga 10 mg was nearly two-fold greater in patients that had a baseline HbA1c >9% than in those with a baseline HbAc <9%. A significant interaction was not observed in subgroups based on use of OAD for the placebo-subtracted mean change from baseline in HbA1c at Week 24.

The placebo-subtracted changes in FPG from baseline at Week 24 (LOCF, excluding data after insulin up-titration) were -1.2 mmol/L (not significant, 95% CI: -1.7 mmol/L, -0.7 mmol/L) and -1.4 mmol/L (p<0.0001, 95% CI: -1.9 mmol/L, -0.9 mmol/L), for Forxiga 5 mg and 10 mg, respectively. The placebo-subtracted proportion of patients achieving a glycemic response (defined as HbA1c <7% at 24 weeks, LOCF, excluding data after insulin up-titration) was 11.1% (p<0.05) and 12.8% (p<0.05) for Forxiga 5 mg and 10 mg, respectively. The placebo-subtracted change in body weight from baseline at Week 24 (LOCF, excluding data after insulin up-titration) was -1.00 kg (p<0.0001, 95% CI: -1.50 kg, -0.50 kg) and -1.68 kg (p<0.0001, 95% CI: -2.19 kg, -1.18 kg) for Forxiga 5 mg and 10 mg, respectively. The results for change in body weight were statistically significant at all doses.

Use in Patients with Type 2 Diabetes and Renal Impairment

The efficacy of Forxiga is dependent on the amount of glucose filtered through the glomeruli in the kidney, which is in turn dependent on the glomerular filtration rate (GFR). Many patients with T2DM are expected to develop some degree of renal impairment, which may affect the efficacy and safety of Forxiga in this subpopulation.

The efficacy of Forxiga in patients with mild renal impairment [estimated glomerular filtration rate (eGFR) ≥60 to < 90 mL/min/1.73m²] was assessed in a pooled analysis across nine clinical studies consisting of 2,226 patients. The mean change from baseline in HbA1c and the placebo-corrected mean HbA1c reduction at 24 weeks was -1.03% and -0.54%, respectively for Forxiga 5 mg (n = 545) and 10 mg (n = 562), respectively.

Forxiga was studied in 252 T2DM patients with moderate renal impairment (eGFR ≥30 to <60 mL/min/1.73m²) in study MB102029, and in 366 patients with moderate renal impairment within a pooled analysis of nine clinical studies . In patients with moderate renal impairment, the efficacy data did not demonstrate a statistically significant improvement in HbA1c compared with placebo at Week 24, in line with the mechanism of action of Forxiga, which is dependent on renal function.

Patients with severe renal impairment (eGFR <30 mL/min/1.73m²) were not included in the clinical development program because Forxiga was not expected to be effective in this population based on the renal mechanism of action.

Forxiga is contraindicated in patients with moderate to severe renal impairment (defined as eGFR <60 mL/min/1.73m²).

Use in Patients with Type 2 Diabetes and Cardiovascular Disease

Studies D1690C00018 and D1690C00019 were multicentre, randomized, placebo-controlled clinical studies designed to evaluate the safety and efficacy of Forxiga 10 mg once daily in T2DM patients with established cardiovascular disease (CVD) and inadequate glycemic control despite stable treatment with OADs and/or insulin. A total of 1,876 patients with HbA1c ≥7.0% and ≤10.0% and CVD were randomized and treated with Forxiga 10 mg or placebo in addition to existing treatment for T2DM. Across both studies, participating patients had hypertension (96.2%), coronary heart disease (75.9%), stroke or transient ischemic attack (TIA) (20.1%), and congestive heart failure (14.3%). Usual care for T2DM in study participants included insulin and OADs, alone or in combination. The studies were designed with a double-blind 24 week primary efficacy period and an 80 week single-blinded extension period. The primary efficacy outcome in both studies was change in HbA1c with LOCF from baseline to Week 24, with a secondary outcome of change in body weight in the same time period. A secondary outcome measure was the 3-Item Endpoint of Clinical Benefit, which classified patients as responders if they achieved an absolute drop of 0.5% or more from baseline HbA1c, a relative drop of 3% or more from baseline weight, and an absolute drop of 3 mmHg or more from baseline in seated systolic blood pressure.

Superiority to placebo was demonstrated for Forxiga 10 mg in both studies. The placebo-subtracted differences in HbA1c at Week 24 were -0.46% (p<0.0001, 95% CI: -0.56%, -0.37%) and -0.40% (p<0.0001, 95% CI: -0.50%, -0.30%) in studies D1690C00018 and D1690C00019, respectively. For both studies, reductions in HbA1c were generally maintained at Week 52 and Week 104. The difference in the proportion of responders on the 3-Item Endpoint of Clinical Benefit between Forxiga 10 mg and placebo was statistically significant (P<0.0001) at 24 weeks in both studies, with a similar placebo-subtracted response rate in both studies (9.9% in study D1690C00018 and 7.0% in study D1690C00019). The proportion of patients meeting responder criteria for the 3-Item Endpoint of Clinical Benefit during long-term treatment was small, but higher with Forxiga than placebo (placebo-subtracted response rate: 5.7% in study D1690C00018, 7.4% in study D1690C00019 at Week 52). The 95% confidence intervals for comparisons of the 3-Item Endpoint of Clinical Benefit with placebo excluded zero at Week 52 in both studies and at Week 104 in study D1690C00018, but not study D1690C00019.

Body composition and bone mineral density in Type 2 Diabetic Patients

Body composition and bone mineral density in T2DM patients treated with Forxiga were assessed in study D1690C00012. This 24-week study in 182 patients found a greater reduction in total body weight from baseline to Week 24 in patients taking Forxiga 10 mg plus metformin (-2.96 kg), versus placebo plus metformin (-0.88 kg), with a significant interaction for gender [greater weight loss for males (-2.76 kg) than females (-1.22 kg)]. The reduction in total body fat mass from baseline to Week 24 was -2.22 kg for the Forxiga group and -0.74 kg for the placebo group with a reduction in percentage total body fat mass from baseline to Week 24 in the Forxiga group of 1%, whereas there was little change in the placebo group, as evaluated by dual energy x-ray absorptiometry (DXA). In an extension of this study to Week 102 there was no change in bone mineral density for the lumbar spine, femoral neck, or total hip seen in either treatment group (mean decrease from baseline for all anatomical regions <0.5%).

Overall Analysis of Efficacy

Overall, statistically and clinically significant improvements in HbA1c from baseline to Week 24 were demonstrated for all proposed indications for both the Forxiga 5 mg and 10 mg doses vs. placebo (placebo-subtracted values ranging from -0.40% to -0.68%). In patients with moderate renal impairment, the efficacy data did not demonstrate statistically significant improvement in HbA1c compared with placebo at Week 24, consistent with its renal mechanism of action. Statistically significant differences from placebo in favour of Forxiga 10 mg were also observed at 24 weeks in HbA1c, and maintained at 52 weeks in both large, randomized, placebo-controlled clinical studies in patients with cardiovascular disease who had inadequate glycemic control.

Following review of the submission, Health Canada revised the indication proposed by the sponsor for Forxiga in the original New Drug Submission (NDS, Control #141791). The indication for monotherapy was revised to reflect standard Product Monograph wording, which states that metformin should be considered first line therapy, and that Forxiga is an effective second-line treatment after metformin in T2DM patients. The proposed indication for use in combination with pioglitazone was removed as a safety precaution due to evidence suggesting an increased risk of bladder cancer with pioglitazone (see discussion below in the Clinical Safety section). The indication for use in combination with insulin was revised to reflect the combinations that were adequately supported by the clinical data provided.

For more information, refer to the Forxiga Product Monograph, approved by Health Canada and available through the Drug Product Database.

Clinical Safety

The clinical safety of Forxiga was assessed in pooled safety analyses drawn from 21 Phase IIb and III studies. The safety data submitted with the original Forxiga NDS (Control #141791) was supplemented with the sponsor's 30 Month Safety Update (30-MSU) and the cardiovascular studies D1690C00018 and D1690C00019. In addition to the pivotal and supporting studies discussed in the Clinical Efficacy section above, the safety database includes data from 12 additional studies that are not directly relevant to the authorised indications. The additional studies were conducted with Forxiga as a monotherapy and in various combinations including with pioglitazone or sitagliptin.

The safety database includes 5,936 patients treated with Forxiga for a mean exposure (excluding data after rescue) of 330.1 days and a cumulative exposure of 6,247.2 patient-years and 3,403 control patients with a mean exposure (excluding data rescue) of 296.5 days, and a cumulative exposure of 3,637.6 patient-years (PY). The placebo-controlled pool including both short-term and long-term data includes 2,026 patients treated with Forxiga 10 mg for a mean exposure (excluding data after rescue) of 374.8 days and cumulative exposure of 2,437.9 PY and 1,956 patients treated with placebo for a mean exposure (excluding data after rescue) of 289.5 days and a cumulative exposure of 2,243.6 PY.

The overall incidence of AEs in the 12-study, short-term, placebo-controlled pool in patients treated with Forxiga 5 mg and 10 mg was 61.9% and 61.5%, respectively compared to 56.9% for the placebo group.

The most commonly reported adverse events during treatment with Forxiga 5 mg or 10 mg (≥5%) were female genital mycotic infections, nasopharyngitis and urinary tract infections. Discontinuation of therapy due to adverse events in patients who received Forxiga 5 mg and 10 mg was 2.8% and 3.2%, respectively compared to 2.5% for the placebo group. The most commonly reported events leading to discontinuation and reported in at least 2 Forxiga 10 mg-treated patients were renal impairment (0.8%), decrease in creatinine clearance (0.6%), increased blood creatinine (0.3%), urinary tract infections (0.2%), and vulvovaginal mycotic infection (0.1%).

A total of 10 serious adverse events, assessed as related by the investigator, were reported in 9 patients in the short-term, placebo-controlled pool: 2 reports from patients taking Forxiga 5 mg daily (change of bowel habit, hypoglycemia); 2 reports from patients taking Forxiga 10 mg daily (constipation, rotator cuff syndrome); and 6 reports from patients in the placebo group (thrombocytopenia, acute myocardial infarction, cystitis, pyelopnephritis, overdose and loss of consciousness).

Malignancies

In the safety data submitted with the original Forxiga NDS (Control #141791), cases of thyroid neoplasm, breast cancer and bladder cancer were reported more frequently in Forxiga-treated patients than in control patients, although none of the imbalances achieved statistical significance. Based on the updated data in the 30-MSU, an imbalance in thyroid neoplasms is no longer evident. Numerical imbalances in breast and bladder cancer cases remain, however.

Breast cancer

In the 30-MSU safety dataset, breast cancer cases were reported in 12 of 2,693 Forxiga-treated women (0.45%) and 3 of 1,439 control-treated women (0.21%). Based on these data, the incidence rate (IR) for breast cancer in Forxiga-treated women is 400 per 100,000 PY (95% CI: 210, 700), compared with 190 per 100,000 PY (95% CI: 40, 560) in the control group, with an incidence rate ratio (IRR) for Forxiga versus control of 2.47 (95% CI: 0.64, 14.1). With the updated data provided with the increased drug exposure in the 30-MSU dataset, the numerical imbalance in breast cancer cases in the Forxiga-treated patients diminished significantly from the original submission, where 9 cases of breast cancer had been reported in dapagliflozin treated women and none in women receiving placebo. A total of 10 of the 12 breast cancer cases in Forxiga-treated women were reported within a year of beginning treatment, including 2 cases reported within 2 months; this suggests a direct causal relationship with Forxiga is unlikely. Considering all of the evidence, the numerical imbalance observed in breast cancer cases is regarded as a signal which can be monitored through surveillance in ongoing post-market studies.

Bladder cancer

In the 30-MSU safety dataset, a total of 10 bladder cancer cases in 6,045 Forxiga-treated patients (0.17%) followed for 6,744 PY were reported, versus 1 case in 3,512 control patients (0.03%) followed for 3,956 PY. This corresponds to incidence rates of 148 bladder cancer events per 100,000 PY for Forxiga-treated patients and 25 events per 100,000 PY for control patients, and an IRR for bladder cancer in Forxiga versus control patients of 6.11 (95% CI: 0.83, 272). Thus, with the additional exposure in the 30-MSU update, the imbalance in bladder cancer cases has not decreased. Of the 10 bladder cancer cases in Forxiga-treated patients, 5 were reported within the first 6 months of treatment, 1 patient was reported in the subsequent 6 month period, 2 more were reported after 12 to 18 months, and 2 after 18 to 24 months of treatment. This represents a short time frame for development of new malignancies, particularly considering the long latency period generally believed to be associated with bladder cancer. The patients generally displayed risk factors for bladder cancer, including smoking history, older age, male sex, and, notably, evidence of hematuria at screening or shortly thereafter, the latter consistent with the possibility of pre-existing tumours. Although not statistically significant, considering the magnitude of the imbalance in the number of bladder cancer cases in Forxiga patients compared to control patients (IRR >5-6), the stability of the imbalance over time, and the fact that the result is based on almost 10,000 patients and more than 10,000 PY of exposure in a clinical trial setting (with bladder cancer risk factors balanced between groups), the imbalance cannot be dismissed. Accordingly, the increased incidence of bladder cancer has been clearly described under the Warnings and Precautions section in the Forxiga Product Monograph and it will be recommended that Forxiga not be used in patients with a history of bladder cancer. Also, as a precautionary measure, concomitant use of Forxiga and pioglitazone will not be indicated, and a warning against such use has been included in the Forxiga Product Monograph, in light of the existing non-clinical, clinical, and epidemiologic evidence suggesting an increased risk of bladder cancer associated with pioglitazone. The sponsor has committed to further assessment of bladder cancer risk with Forxiga as a post-market commitment, through independent monitoring of bladder cancer events in a large, ongoing cardiovascular outcomes study.

Volume depletion events

Adverse events due to intravascular volume depletion (hypotension, orthostatic hypotension, syncope, dehydration) were reported more commonly in patients receiving Forxiga than in those receiving placebo, and this was evident within the first 1-2 weeks of initiating the drug. The increased risk was more marked in patients ≥65 years of age, those taking loop diuretics, and those with an eGFR <60 mL/min/1.73m². This is described in the Forxiga Product Monograph, including text in the Warnings and Precautions as well as Dosage and Administration sections. It is recommended that Forxiga not be administered to patients with evidence of volume depletion, that volume depletion be corrected prior to administration, and also that caution be exercised in case of intercurrent illnesses which may cause volume depletion. Caution is also advised with use in elderly patients, those on antihypertensives (particularly loop diuretics), those with a history of low blood pressure, and those with a history of cardiovascular disease.

Renal Impairment

Forxiga is associated with dose-dependent decreases in eGFR and increases in creatinine. Renal adverse events consistent with impairment of renal function due to Forxiga occurred more frequently in patients with moderate renal impairment (eGFR <60 mL/min/1.73m²) than in patients with mild or no renal impairment. In addition, since no clinically relevant or significant beneficial effect of Forxiga on HbA1c could be demonstrated in patients with moderate renal impairment, Forxiga is contraindicated in this population.

The Product Monograph contains text in the Warnings and Precautions and Dosage and Administration sections describing the renal effects of Forxiga, and advising of the necessity of assessment of renal function prior to initiation, as well as regular renal function monitoring and discontinuation if a patient's eGFR drops below 60 mL/min/1.73m².

Genital infections and UTIs

Genital infections were seen much more commonly in Forxiga-treated patients than in placebo-treated patients, with increases in frequency compared to placebo of 6-7 fold in women and 16-17 fold in men, and rates of recurrences were also roughly doubled. The most commonly reported events were vulvovaginal mycotic infections in women and balanitis in men. This information is captured in the Warnings and Precautions and Adverse Reactions sections of the Forxiga Product Monograph.

Urinary tract infections were also reported more commonly in patients receiving Forxiga than those receiving placebo, and were also more frequent in women than men; the incidence of infection was increased by roughly 1.4-fold in both sexes. Recurrent infections were also more common in Forxiga-treated patients than placebo patients. There was no clear evidence of an increased risk of complicated or upper tract infections, i.e. pyelonephritis.

Hypoglycemia

When Forxiga was administered as monotherapy or as an add-on to agents not traditionally associated with significant hypoglycemia, such as metformin, rates of hypoglycemic episodes were comparable between Forxiga and comparators. However, as an add-on to insulin or insulin secretagogues [for example (e.g.) sulfonylureas], Forxiga was associated with an increased risk of episodes of hypoglycemia. This information is described under the Warnings and Precautions and Adverse Reactions sections of the Forxiga Product Monograph.

Hemoconcentration

Hemoglobin and hematocrit levels increase with Forxiga administration, consistent with reductions in intravascular volume. No clear evidence of an increase in thromboembolic events was seen with Forxiga, however in the Forxiga Product Monograph caution is advised regarding use of Forxiga in patients with a pre-existing elevated hematocrit or hemoglobin level.

Hepatic adverse events

Based on the updated 30-MSU database, no clear imbalances were evident in either adverse events or abnormalities of liver tests (transaminase & bilirubin elevations) between Forxiga-treated and control patients. An imbalance in serious adverse events was noted, with 7/5,936 patients (0.1%) in the total Forxiga group versus 1/3,403 (<0.1%) in the control group reporting SAEs, however the reported events were not consistent with drug-induced liver injury.

Bone fractures

In the overall Forxiga clinical program, no increase in fracture risk was noted. However, in Study MB102029, in patients with moderate renal impairment (eGFR <60 mL/min/1.73m²), an apparent dose-dependent increase in fracture risk was evident in patients taking Forxiga 5 mg and 10 mg compared to placebo, although the numbers of events were too small to achieve statistical significance. In the general patient population, no clear effects of Forxiga were seen on bone mineral density at three sites, and, with respect to biochemical bone markers, only serum phosphate and parathyroid hormone showed consistent changes (increases) with Forxiga treatment. Forxiga is contraindicated in patients with moderate renal impairment.

Cardiovascular safety

In a double-blind, randomised, placebo- and positive-controlled crossover study, single oral doses of dapagliflozin 20 mg and 150 mg were not associated with clinically or statistically significant effects on the QTc interval, the QRS duration, the PR interval, or heart rate in healthy patients (n = 36).

Updated MACE analyses for the pooled Phase II/III studies comparing Forxiga 5 mg and 10 mg versus placebo, based on the 30-MSU and the pooled D1690C00018 and D1690C00019 studies in diabetic patients with cardiovascular disease were provided. A 95% CI upper limit of 1.8 was excluded for the composite endpoints of major adverse cardiovascular events (MACE), consisting of cardiovascular death, myocardial infarction, and stroke, and MACE plus hospitalisation for unstable angina for both doses combined as well as for each of dose separately, with favourable hazard ratios <1 for both composite endpoints in both dose groups. Results for the MACE components revealed favourable or neutral hazard ratios for cardiovascular death and myocardial infarction.

The sponsor has undertaken a post-marketing commitment to submit the results of an on-going cardiovascular outcomes study. The sponsor has also committed to further assessment of bladder cancer risk with Forxiga as a post-market commitment in the same ongoing trial.

Forxiga is associated with increases in mean low density lipoprotein C (LDL-C) and total cholesterol levels, along with increases in high density lipoprotein C (HDL-C). Rates of dyslipidemia adverse events were also increased amongst Forxiga-treated patients, compared to placebo-treated patients.

The Warnings and Precautions section of the Forxiga Product Monograph includes text describing the LDL elevations observed with Forxiga, and recommends appropriate monitoring.

Starting dose of Forxiga

Following consideration of the 30-MSU and the extensions of Studies D1690C00018 and D1690C00019, Health Canada determined that a starting dose of 5 mg should be recommended for all patients initiating treatment with Forxiga.

The clinical study data establish that the 5 mg dose of Forxiga has a clinically relevant glycemic benefit. While the 10 mg Forxiga dose confers greater glycemic control, it also results in greater glycosuria, osmotic diuresis, and increased intravascular volume depletion than the 5 mg dose, and this is manifest as an increased incidence of adverse events due to volume depletion, compared to the 5 mg dose.

The sponsor had previously suggested that a 5 mg starting dose might be appropriate in patients at risk of volume depletion, however Health Canada considers that reliably distinguishing patients at risk for hypovolemia is likely to be challenging, and furthermore that a majority of patients will likely fall into this category. As an example, Forxiga-associated volume depletion adverse events were seen more frequently in patients over 65, and yet, in 2008, patients over 65 constituted 46.8%, or roughly half, of Canadian diabetics. When patients on loop (and possibly other) diuretics are added, along with other patients at increased risk of hypovolemia, the impracticality of selecting patients who should begin therapy with 5 mg becomes apparent.

Forxiga therapy will therefore be initiated at a dose of 5 mg for all patients, with titration of the dose upwards if greater glycemic control is required, accompanied by careful monitoring of the patient's volume status. The sponsor was asked to resubmit the 5 mg Forxiga formulation, and to modify the Forxiga Product Monograph accordingly.

Issuance of Notice of Non-compliance, Notice of Deficiency and Responses

During review of the original New Drug Submission (NDS) for Forxiga (Control #141791), a Notice of Non-Compliance (NON) was issued to the sponsor on January 11, 2012 because an adequate risk-benefit analysis could not be conducted with the information provided. The following major safety issues were identified:

1. The sponsor informed Health Canada of its plans to submit additional safety data requested by a foreign regulator agency, including data from two large-scale studies conducted in T2DM patients with high cardiovascular risk (Studies D1690C0018 and D1690C00019). As these data provide important safety information in a high risk population relevant to T2DM, the review of this material was considered critical to adequately assess the benefit-risk profile of Forxiga.
2. In the initial dataset, an imbalance was identified in the number of bladder cancer cases reported between Forxiga-treated patients and controls. The preclinical data did not suggest any concern related to the development of bladder cancer in animals, and the data provided were insufficient to adequately assess the risk of development of bladder cancer in humans. Special vigilance was necessary because Forxiga is a first in class drug with a mechanism of action involving the genitourinary tract.
3. In the Forxiga clinical program, an imbalance was identified in the number of breast cancer cases reported between female Forxiga-treated patients and controls. A total of 9 cases of breast cancer were identified among female Forxiga-treated patients compared to zero cases in the comparator arms. An exploratory analysis did not demonstrate statistically significant differences between the number of breast cancer cases reported in Forxiga-treated women and the number that would be expected in the general population, but the analysis was considered to be limited.
4. There was one case of hepatitis identified as possibly related to drug-induced liver injury in a Forxiga-treated patient. In follow-up discussions with Health Canada, the sponsor noted that since filing the NDS there had been 8 new cases identified as meeting the criteria for hepatic adjudication, including one additional case of hepatitis in a Forxiga-treated patient considered by the investigator to be related to the study drug. Imbalances were noted in between the proportion of Forxiga-treated patients and controls with laboratory abnormalities related to liver function. There was also a slightly increased rate of the serious adverse event of cholelithiasis in Forxiga-treated patients. The nonclinical data indicate a low risk for potential intrinsic Forxiga-induced liver injury in humans, but Health Canada notes that non-clinical testing cannot predict some types of idiosyncratic hepatotoxicity in humans.
5. The MACE meta-analysis provided with the initial data included a large proportion of patients receiving sub-therapeutic doses. The data were not suggestive of an increased risk of MACE, however the components of the composite endpoint trended in different directions, with a non-significant decrease in cardiovascular death and myocardial infarction with Forxiga 10 mg vs. control, but a non-significant increase in stroke for Forxiga 10 mg vs. control.
6. Due to other trends or imbalances identified in the original review, Health Canada also requested additional data with respect to:
   1. Updated bone fracture data from the updated pool of clinical safety available at the time as well as long-term results related to bone mineral density.
   2. Updated information on thyroid neoplasms.
   3. Additional relevant information to support the efficacy and safety of the 5 mg dose in the population in which it was proposed as a starting dose.
   4. Updated renal safety data from the updated pool of clinical safety available at the time, including incidence of renal impairment/failure and increased blood creatinine.
   5. Additional analyses correlating baseline proteinuria status and efficacy of Forxiga.

The sponsor withdrew the NDS (Control #141791) as additional time was required to address the issues raised in the NON. The drug submission was re-filed on December 7, 2012, as Control #160877. During review, deficiencies were noted in the data package; clinical studies conducted in patients with diabetes and hypertension, and populations of interest in analyses of cardiovascular safety had not been provided. As a result, a Notice of Deficiency (NOD) was issued for Forxiga on September 27, 2013. Health Canada requested the following additional information:

1. Appendices to the 30-MSU, which were frequently cross-referenced in the main report.
2. Studies included in the safety pooled analyses but not already provided to Health Canada, including Asian monotherapy studies MB102054 and D1692C00006.
3. Studies MB102073 and MB102077, both conducted in patients with diabetes and hypertension treated with Forxiga. These studies were conducted to investigate the effect of Forxiga on blood pressure and glycemic control in patients with T2DM and inadequately controlled hypertension. Study MB102073 included hypertensive and diabetic patients on a pre-existing stable dose of an angiotensin converting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB). MB102077 included hypertensive and diabetic patients taking a pre-existing antihypertensive in addition to an ACE inhibitor or ARB. The sponsor specified that these were among the additional "core" studies completed since the initial summary of clinical safety. These studies are of particular interest in that hypertension is a common comorbidity with T2DM, and these patients are at particularly high risk for cardiovascular disease, and therefore of interest in analyses of cardiovascular safety data for the product. ACE inhibitors and ARBs are common medications used in the treatment of hypertension in patients with Type 2DM and ACE inhibitors are commonly used in patients with diabetes and resulting nephropathy [that is (i.e.) proteinuria].
4. Updated overall summaries of safety and efficacy.
5. A supporting argument to retain the proposed indication for Forxiga as an add-on to pioglitazone after it had been withdrawn from an application in Europe due to concerns raised regarding risk of bladder cancer with this combination.

The sponsor responded by providing the additional information requested by Health Canada, withdrawing the proposed indication for Forxiga as add-on to pioglitazone, and with revisions to the Forxiga Product Monograph. Health Canada's assessment of the information provided is discussed in the sections above.

Overall Analysis of Safety

Appropriate warnings and precautions are in place in the approved Forxiga Product Monograph to address the identified safety concerns. All safety issues, including uncertainties, were addressed in the Forxiga Product Monograph. The possible increased incidence of bladder cancer is described, with a recommendation that Forxiga not be used in patients with a history of bladder cancer, and as a precautionary measure, concomitant use of Forxiga and pioglitazone is not indicated, in view of the apparent increased risk of bladder cancer associated with pioglitazone. Risk of events related to intravascular volume depletion is well described in the Product Monograph, including a recommendation that Forxiga not be administered to patients with evidence of volume depletion, that volume depletion be corrected prior to administration, and that caution be exercised in case of intercurrent illnesses which may cause volume depletion, in patients who are elderly, on antihypertensives and with a history of low blood pressure or cardiovascular disease. In view of lack of significant beneficial effect of Forxiga on HbA1c and safety issues in patients with moderate and severe renal impairment, use is contraindicated in this population and the Forxiga Product Monograph advises to assess renal function prior to initiation, as well as regular renal function monitoring and discontinuation if a patient's eGFR drops below 60 mL/min/1.73m². Caution is also advised regarding use of Forxiga in patients with a pre-existing elevated hematocrit or hemoglobin level.

A Risk Management Plan was considered acceptable by the Marketed Health Products Directorate.

For more information, refer to the Forxiga Product Monograph, approved by Health Canada and available through the Drug Product Database.