Tradjenta Drug Information

Glycemic Control Trials in Adults with Type 2 Diabetes Mellitus

TRADJENTA has been studied as monotherapy and in combination with metformin, sulfonylurea, pioglitazone, and insulin. TRADJENTA has also been studied in patients with type 2 diabetes mellitus and severe chronic renal impairment. In patients with type 2 diabetes mellitus, treatment with TRADJENTA produced clinically significant improvements in hemoglobin A1c (A1C), fasting plasma glucose (FPG), and 2-hour post-prandial glucose (PPG) compared with placebo.

Monotherapy A total of 730 patients with type 2 diabetes mellitus participated in 2 double-blind, placebo-controlled studies, one of 18 weeks' and another of 24 weeks' duration, to evaluate the efficacy and safety of TRADJENTA monotherapy. In both monotherapy studies, patients currently on an antihyperglycemic agent discontinued the agent and underwent a diet, exercise, and drug washout period of about 6 weeks that included an open-label placebo run-in during the last 2 weeks. Patients with inadequate glycemic control (A1C 7% to 10%) after the washout period were randomized; patients not currently on antihyperglycemic agents (off therapy for at least 8 weeks) with inadequate glycemic control (A1C 7% to 10%) were randomized after completing the 2-week, open-label, placebo run-in period.

In the 18-week trial, only patients ineligible for metformin were recruited. In the 18-week trial, 76 patients were randomized to placebo and 151 to TRADJENTA 5 mg; in the 24-week trial, 167 patients were randomized to placebo and 336 to TRADJENTA 5 mg. Patients who failed to meet specific glycemic goals during the 18-week trial received rescue therapy with pioglitazone and/or insulin; metformin rescue therapy was used in the 24-week trial.

Treatment with TRADJENTA 5 mg daily provided statistically significant improvements in A1C, FPG, and 2-hour PPG compared with placebo (Table 5). In the 18-week trial, 12% of patients receiving TRADJENTA 5 mg and 18% who received placebo required rescue therapy. In the 24-week trial, 10.2% of patients receiving TRADJENTA 5 mg and 20.9% of patients receiving placebo required rescue therapy. The improvement in A1C compared with placebo was not affected by gender, age, race, prior antihyperglycemic therapy, baseline BMI, or a standard index of insulin resistance (HOMA-IR). As is typical for trials of agents to treat type 2 diabetes mellitus, the mean reduction in A1C with TRADJENTA appears to be related to the degree of A1C elevation at baseline.

In these 18- and 24-week studies, the changes from baseline in A1C were -0.4% and -0.4%, respectively, for those given TRADJENTA, and 0.1% and 0.3%, respectively, for those given placebo. Change from baseline in body weight did not differ significantly between the groups. Table 5 Glycemic Parameters in Placebo-Controlled Monotherapy Trials of TRADJENTA* 18-Week Trial 24-Week Trial TRADJENTA 5 mg Placebo TRADJENTA 5 mg Placebo *Full analysis population using last observation on trial **18-week trial: Placebo, n=68; TRADJENTA, n=136 24-week trial: Placebo, n=147; TRADJENTA, n=306 ***18-week trial.

HbA1c: ANCOVA model included treatment, reason for metformin intolerance and number of prior oral anti-diabetic medicine(s) (OADs) as class-effects, as well as baseline HbA1c as continuous covariates. FPG: ANCOVA model included treatment, reason for metformin intolerance and number of prior OADs as class-effects, as well as baseline HbA1c and baseline FPG as continuous covariates. 24-week trial. HbA1c: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c as continuous covariates.

FPG: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c and baseline FPG as continuous covariates. PPG: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c and baseline postprandial glucose after two hours as covariate. A1C (%) Number of patients n = 147 n = 73 n = 333 n = 163 Baseline (mean) 8.1 8.1 8.0

Change from baseline (adjusted mean***) -0.4 0.1 -0.4 0.3 Difference from placebo

(adjusted mean) (95% CI) -0.6 (-0.9, -0.3) -- -0.7 (-0.9, -0.5) -- Patients achieving A1C <7%** 32 8 77 17 FPG (mg/dL) Number of patients n = 138 n = 66 n = 318 n = 149 Baseline (mean) 178 176 164 166 Change from baseline (adjusted mean***) -13 7 -9 15 Difference from placebo (adjusted mean) (95% CI) -21 (-31, -10) -- -23 (-30, -16) -- 2-hour PPG (mg/dL) Number of patients Data not available Data not available n = 67 n = 24 Baseline (mean) -- -- 258 244 Change from baseline (adjusted mean***) -- -- -34 25 Difference from placebo (adjusted mean) (95% CI) -- -- -58 (-82, -34) -- Add-on Combination Therapy with Metformin A total of 701 patients with type 2 diabetes mellitus participated in a 24-week, randomized, double-blind, placebo-controlled trial designed to assess the efficacy of TRADJENTA in combination with metformin. Patients already on metformin (n = 491) at a dosage of at least 1,500 mg per day were randomized after completing a 2-week, open-label, placebo run-in period. Patients on metformin and another antihyperglycemic agent (n = 207) were randomized after a run-in period of approximately 6 weeks on metformin (at a dosage of at least 1,500 mg per day) in monotherapy.

Patients were randomized to the addition of either TRADJENTA 5 mg or placebo, administered once daily. Patients who failed to meet specific glycemic goals during the studies were treated with glimepiride rescue. In combination with metformin, TRADJENTA provided statistically significant improvements in A1C, FPG, and 2-hour PPG compared with placebo (Table 6). Rescue glycemic therapy was used in 7.8% of patients treated with TRADJENTA 5 mg and in 18.9% of patients treated with placebo.

A similar decrease in body weight was observed for both treatment groups. Table 6 Glycemic Parameters in Placebo-Controlled Trial for TRADJENTA in Combination with Metformin* TRADJENTA 5 mg + Metformin Placebo + Metformin *Full analysis population using last observation on trial **TRADJENTA 5 mg + Metformin, n=485; Placebo + Metformin, n=163 ***HbA1c: ANCOVA model included treatment and number of prior oral OADs as class-effects, as well as baseline HbA1c as continuous covariates. FPG: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c and baseline FPG as continuous covariates.

PPG: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c and baseline postprandial glucose after two hours as covariate. A1C (%) Number of patients n = 513 n = 175 Baseline (mean) 8.1

Change from baseline (adjusted mean***) -0.5 0.15 Difference from placebo + metformin

(adjusted mean) (95% CI) -0.6 (-0.8, -0.5) -- Patients achieving A1C <7%** 127 15 FPG (mg/dL) Number of patients n = 495 n = 159 Baseline (mean) 169 164 Change from baseline (adjusted mean***) -11 11 Difference from placebo + metformin (adjusted mean) (95% CI) -21 (-27, -15) -- 2-hour PPG (mg/dL) Number of patients n = 78 n = 21 Baseline (mean) 270 274 Change from baseline (adjusted mean***) -49 18 Difference from placebo + metformin (adjusted mean) (95% CI) -67 (-95, -40) -- Initial Combination Therapy with Metformin A total of 791 patients with type 2 diabetes mellitus and inadequate glycemic control on diet and exercise participated in the 24-week, randomized, double-blind, portion of this placebo-controlled factorial trial designed to assess the efficacy of TRADJENTA as initial therapy with metformin. Patients on an antihyperglycemic agent (52%) underwent a drug washout period of 4 weeks' duration. After the washout period and after completing a 2-week single-blind placebo run-in period, patients with inadequate glycemic control (A1C ≥7.0% to ≤10.5%) were randomized.

Patients with inadequate glycemic control (A1C ≥7.5% to <11.0%) not on antihyperglycemic agents at trial entry (48%) immediately entered the 2-week, single-blind, placebo run-in period and then were randomized. Randomization was stratified by baseline A1C (<8.5% vs ≥8.5%) and use of a prior oral antidiabetic drug (none vs monotherapy). Patients were randomized in a 1:2:2:2:2:2 ratio to either placebo or one of 5 active-treatment arms. Approximately equal numbers of patients were randomized to receive initial therapy with 5 mg of TRADJENTA once daily, 500 mg or 1,000 mg of metformin twice daily, or 2.5 mg of linagliptin twice daily in combination with 500 mg or 1,000 mg of metformin twice daily.

Patients who failed to meet specific glycemic goals during the trial were treated with sulfonylurea, thiazolidinedione, or insulin rescue therapy. Initial therapy with the combination of linagliptin and metformin provided significant improvements in A1C and fasting plasma glucose (FPG) compared to placebo, to metformin alone, and to linagliptin alone (Table 7). The adjusted mean treatment difference in A1C from baseline to week 24 (LOCF) was -0.5% (95% CI -0.7, -0.3; p<0.0001) for linagliptin 2.5 mg/metformin 1,000 mg twice daily compared to metformin 1,000 twice daily; -1.1% (95% CI -1.4, -0.9; p<0.0001) for linagliptin 2.5 mg/metformin 1,000 mg twice daily compared to TRADJENTA 5 mg once daily; -0.6% (95% CI -0.8, -0.4; p<0.0001) for linagliptin 2.5 mg/metformin 500 mg twice daily compared to metformin 500 mg twice daily; and -0.8% (95% CI -1.0, -0.6; p<0.0001) for linagliptin 2.5 mg/metformin 500 mg twice daily compared to TRADJENTA 5 mg once daily. Lipid effects were generally neutral.

No meaningful change in body weight was noted in any of the 6 treatment groups. Table 7 Glycemic Parameters at Final Visit (24-Week Trial) for Linagliptin and Metformin, Alone and in Combination in Randomized Patients with Type 2 Diabetes Mellitus Inadequately Controlled on Diet and Exercise** Placebo TRADJENTA 5 mg Once Daily* Metformin 500 mg Twice Daily Linagliptin 2.5 mg Twice Daily* + Metformin 500 mg Twice Daily Metformin 1,000 mg Twice Daily Linagliptin 2.5 mg Twice Daily* + Metformin 1,000 mg Twice Daily *Total daily dosage of TRADJENTA is equal to 5 mg **Full analysis population using last observation on trial ***Metformin 500 mg twice daily, n=140; Linagliptin 2.5 mg twice daily + Metformin 500 twice daily, n=136; Metformin 1,000 mg twice daily, n=137; Linagliptin 2.5 mg twice daily + Metformin 1,000 mg twice daily, n=138 ****HbA1c: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c as continuous covariates. FPG: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c and baseline FPG as continuous covariates.

A1C (%) Number of patients n = 65 n = 135 n = 141 n = 137 n = 138 n = 140 Baseline (mean) 8.7 8.7 8.7 8.7 8.5

Change from baseline (adjusted mean****) 0.1 -0.5 -0.6 -1.2 -1.1 -1.6 Difference

from placebo (adjusted mean) (95% CI) -- -0.6 (-0.9, -0.3) -0.8 (-1.0, -0.5) -1.3 (-1.6, -1.1) -1.2 (-1.5, -0.9) -1.7 (-2.0, -1.4) Patients achieving A1C <7%*** 7 14 26 41 42 74 Patients (%) receiving rescue medication 29.2 11.1 13.5 7.3 8.0

FPG (mg/dL) Number of patients n = 61 n = 134 n

= 136 n = 135 n = 132 n = 136 Baseline (mean) 203 195 191 199 191 196 Change from baseline (adjusted mean****) 10 -9 -16 -33 -32 -49 Difference from placebo (adjusted mean) (95% CI) -- -19 (-31, -6) -26 (-38, -14) -43 (-56, -31) -42 (-55, -30) -60 (-72, -47) Active-Controlled Trial vs Glimepiride in Combination with Metformin The efficacy of TRADJENTA was evaluated in a 104-week, double-blind, glimepiride-controlled, non-inferiority trial in patients with type 2 diabetes mellitus with insufficient glycemic control despite metformin therapy. Patients being treated with metformin only entered a run-in period of 2 weeks' duration, whereas patients pretreated with metformin and one additional antihyperglycemic agent entered a run-in treatment period of 6 weeks' duration with metformin monotherapy (dosage of ≥1,500 mg/day) and washout of the other agent. After an additional 2-week placebo run-in period, those with inadequate glycemic control (A1C 6.5% to 10%) were randomized 1:1 to the addition of TRADJENTA 5 mg once daily or glimepiride.

Randomization was stratified by baseline HbA1c (<8.5% vs ≥8.5%), and the previous use of antidiabetic drugs (metformin alone vs metformin plus one other OAD). Patients receiving glimepiride were given an initial dosage of 1 mg/day and then electively titrated over the next 12 weeks to a maximum dosage of 4 mg/day as needed to optimize glycemic control. Thereafter, the glimepiride dosage was to be kept constant, except for down-titration to prevent hypoglycemia. After 52 and 104 weeks, TRADJENTA and glimepiride both had reductions from baseline in A1C (52 weeks: -0.4% for TRADJENTA, -0.6% for glimepiride; 104 weeks: -0.2% for TRADJENTA, -0.4% for glimepiride) from a baseline mean of 7.7% (Table 8). The mean difference between groups in A1C change from baseline was 0.2% with 2-sided 97.5% confidence interval (0.1%, 0.3%) for the intent-to-treat population using last observation carried forward.

These results were consistent with the completers analysis. Table 8 Glycemic Parameters at 52 and 104 Weeks in Trial Comparing TRADJENTA to Glimepiride as Add-On Therapy in Patients Inadequately Controlled on Metformin** Week 52 Week 104 TRADJENTA 5 mg + Metformin Glimepiride + Metformin (mean Glimepiride dosage 3 mg) TRADJENTA 5 mg + Metformin Glimepiride + Metformin (mean Glimepiride dosage 3 mg) *p<0.0001 vs glimepiride; † p=0.0012 vs glimepiride **Full analysis population using last observation on trial ***Hypoglycemic incidence included both asymptomatic events (not accompanied by typical symptoms and plasma glucose concentration of ≤70 mg/dL) and symptomatic events with typical symptoms of hypoglycemia and plasma glucose concentration of ≤70 mg/dL. ****HbA1c: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c as continuous covariates. FPG: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c and baseline FPG as continuous covariates.

Hypoglycemia incidence (%): Cochran-Mantel-Haenszel test was performed on the patient population contained in the treated set, to compare the proportion of patients with hypoglycemic events between patients treated with linagliptin and patients treated with glimepiride. A1C (%) Number of patients n = 764 n = 755 n = 764 n = 755 Baseline (mean) 7.7 7.7 7.7

Change from baseline (adjusted mean****) -0.4 -0.6 -0.2 -0.4 Difference from glimepiride

(adjusted mean) (97.5% CI) 0.2 -- 0.2 -- FPG (mg/dL) Number of patients n = 733 n = 725 n = 733 n = 725 Baseline (mean) 164 166 164 166 Change from baseline (adjusted mean****) -8* -15 -2 † -9 Hypoglycemia incidence (%)*** Number of patients n = 776 n = 775 n = 776 n = 775 Incidence**** 5.3* 31.1 7.5*

Patients treated with linagliptin had a mean baseline body weight of 86

kg and were observed to have an adjusted mean decrease in body weight of 1.1 kg at 52 weeks and 1.4 kg at 104 weeks. Patients on glimepiride had a mean baseline body weight of 87 kg and were observed to have an adjusted mean increase from baseline in body weight of 1.4 kg at 52 weeks and 1.3 kg at 104 weeks (treatment difference p<0.0001 for both timepoints). Add-On Combination Therapy with Pioglitazone A total of 389 patients with type 2 diabetes mellitus participated in a 24-week, randomized, double-blind, placebo-controlled trial designed to assess the efficacy of TRADJENTA in combination with pioglitazone. Therapy was stopped in patients on oral antihyperglycemic therapy for a period of 6 weeks (4 weeks followed by a 2-week, open-label, placebo run-in period). Drug-naïve patients entered directly into the 2-week placebo run-in period.

After the run-in period, patients were randomized to receive either TRADJENTA 5 mg or placebo, both in addition to pioglitazone 30 mg daily. Patients who failed to meet specific glycemic goals during the studies were treated with metformin rescue. Glycemic endpoints measured were A1C and FPG. In initial combination with pioglitazone 30 mg, TRADJENTA 5 mg provided statistically significant improvements in A1C and FPG compared to placebo with pioglitazone (Table 9). Rescue therapy was used in 7.9% of patients treated with TRADJENTA 5 mg/pioglitazone 30 mg and 14.1% of patients treated with placebo/pioglitazone 30 mg.

Patient weight increased in both groups during the trial with an adjusted mean change from baseline of 2.3 kg and 1.2 kg in the TRADJENTA 5 mg/pioglitazone 30 mg and placebo/pioglitazone 30 mg groups, respectively (p = 0.0141). Table 9 Glycemic Parameters in Placebo-Controlled Trial for TRADJENTA in Combination Therapy with Pioglitazone* TRADJENTA 5 mg + Pioglitazone Placebo + Pioglitazone *Full analysis population using last observation on trial **HbA1c: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c as continuous covariates. FPG: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c and baseline FPG as continuous covariates. A1C (%) Number of patients n = 252 n = 128 Baseline (mean) 8.6

Change from baseline (adjusted mean**) -1.1 -0.6 Difference from placebo + pioglitazone

(adjusted mean) (95% CI) -0.5 (-0.7, -0.3) -- Patients achieving A1C <7% 108 39 FPG (mg/dL) Number of patients n = 243 n = 122 Baseline (mean) 188 186 Change from baseline (adjusted mean**) -33 -18 Difference from placebo + pioglitazone (adjusted mean) (95% CI) -14 (-21, -7) -- Add-On Combination with Sulfonylureas A total of 245 patients with type 2 diabetes mellitus participated in an 18-week, randomized, double-blind, placebo-controlled trial designed to assess the efficacy of TRADJENTA in combination with sulfonylurea (SU). Patients on sulfonylurea monotherapy (n = 142) were randomized after completing a 2-week, single-blind, placebo run-in period. Patients on a sulfonylurea plus one additional oral antihyperglycemic agent (n = 103) were randomized after a wash-out period of 4 weeks and a 2-week, single-blind, placebo run-in period. Patients were randomized to the addition of TRADJENTA 5 mg or to placebo, each administered once daily.

Patients who failed to meet specific glycemic goals during the studies were treated with metformin rescue. Glycemic endpoints measured included A1C and FPG. In combination with a sulfonylurea, TRADJENTA provided statistically significant improvements in A1C compared with placebo following 18 weeks' treatment; the improvements in FPG observed with TRADJENTA were not statistically significant compared with placebo (Table 10). Rescue therapy was used in 7.6% of patients treated with TRADJENTA 5 mg and 15.9% of patients treated with placebo. There was no significant difference between TRADJENTA and placebo in body weight.

Table 10 Glycemic Parameters in Placebo-Controlled Trial for TRADJENTA in Combination with Sulfonylurea* TRADJENTA 5 mg + SU Placebo + SU SU = sulfonylurea *Full analysis population using last observation on trial **TRADJENTA 5 mg + SU, n=156; Placebo + SU, n=82 ***HbA1c: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c as continuous covariates. FPG: ANCOVA model included treatment and number of prior OADs as class-effects, as well as baseline HbA1c and baseline FPG as continuous covariates A1C (%) Number of patients n = 158 n = 82 Baseline (mean) 8.6

Change from baseline (adjusted mean***) -0.5 -0.1 Difference from placebo + SU

(adjusted mean) (95% CI) -0.5 (-0.7, -0.2) -- Patients achieving A1C <7%** 23 3 FPG (mg/dL) Number of patients n = 155 n = 78 Baseline (mean) 180 171 Change from baseline (adjusted mean***) -8 -2 Difference from placebo + SU (adjusted mean) (95% CI) -6 (-17, 4) -- Add-On Combination Therapy with Metformin and a Sulfonylurea A total of 1,058 patients with type 2 diabetes mellitus participated in a 24-week, randomized, double-blind, placebo-controlled trial designed to assess the efficacy of TRADJENTA in combination with a sulfonylurea and metformin. The most common sulfonylureas used by patients in the trial were: glimepiride (31%), glibenclamide (26%), and gliclazide (26%, not available in the United States). Patients on a sulfonylurea and metformin were randomized to receive TRADJENTA 5 mg or placebo, each administered once daily. Patients who failed to meet specific glycemic goals during the trial were treated with pioglitazone rescue.

Glycemic endpoints measured included A1C and FPG. In combination with a sulfonylurea and metformin, TRADJENTA provided statistically significant improvements in A1C and FPG compared with placebo (Table 11). In the entire trial population (patients on TRADJENTA in combination with sulfonylurea and metformin), a mean reduction from baseline relative to placebo in A1C of -0.6% and in FPG of -13 mg/dL was seen. Rescue therapy was used in 5.4% of patients treated with TRADJENTA 5 mg and in 13% of patients treated with placebo. Change from baseline in body weight did not differ significantly between the groups.

Table 11 Glycemic Parameters in Placebo-Controlled Trial for TRADJENTA in Combination with Metformin and Sulfonylurea* TRADJENTA 5 mg + Metformin + SU Placebo + Metformin + SU SU = sulfonylurea *Full analysis population using last observation on trial **TRADJENTA 5 mg + Metformin + SU, n=742; Placebo + Metformin + SU, n=247 ***HbA1c: ANCOVA model included treatment as class-effects and baseline HbA1c as continuous covariates. FPG: ANCOVA model included treatment as class-effects, as well as baseline HbA1c and baseline FPG as continuous covariates. A1C (%) Number of patients n = 778 n = 262 Baseline (mean) 8.2

Change from baseline (adjusted mean***) -0.7 -0.1 Difference from placebo (adjusted mean)

(95% CI) -0.6 (-0.7, -0.5) -- Patients achieving A1C <7%** 217 20 FPG (mg/dL) Number of patients n = 739 n = 248 Baseline (mean) 159 163 Change from baseline (adjusted mean***) -5 8 Difference from placebo (adjusted mean) (95% CI) -13 (-18, -7) -- Add-On Combination Therapy with Insulin A total of 1,261 patients with type 2 diabetes mellitus inadequately controlled on basal insulin alone or basal insulin in combination with oral drugs participated in a randomized, double-blind, placebo-controlled trial designed to evaluate the efficacy of TRADJENTA as add-on therapy to basal insulin over 24 weeks. Randomization was stratified by baseline HbA1c (<8.5% vs ≥8.5%), renal function impairment status (based on baseline eGFR), and concomitant use of oral antidiabetic drugs (none, metformin only, pioglitazone only, metformin + pioglitazone). Patients with a baseline A1C of ≥7% and ≤10% were included in the trial including 709 patients with renal impairment (eGFR <90 mL/min), most of whom (n=575) were categorized as mild renal impairment (eGFR 60 to <90 mL/min). Patients entered a 2 week placebo run-in period on basal insulin (e.g., insulin glargine, insulin detemir, or NPH insulin) with or without metformin and/or pioglitazone background therapy. Following the run-in period, patients with inadequate glycemic control were randomized to the addition of either 5 mg of TRADJENTA or placebo, administered once daily.

Patients were maintained on a stable dosage of insulin prior to enrollment, during the run-in period, and during the first 24 weeks of treatment. Patients who failed to meet specific glycemic goals during the double-blind treatment period were rescued by increasing background insulin dosage. TRADJENTA used in combination with insulin (with or without metformin and/or pioglitazone), provided statistically significant improvements in A1C and FPG compared to placebo (Table 12) after 24 weeks of treatment.

The mean total daily insulin dosage at baseline was 42 units for patients treated with TRADJENTA and 40 units for patients treated with placebo. Background baseline diabetes mellitus therapy included use of: insulin alone (16.1%), insulin combined with metformin only (75.5%), insulin combined with metformin and pioglitazone (7.4%), and insulin combined with pioglitazone only (1%). The mean change from baseline to Week 24 in the daily dosage of insulin was +

IU in the placebo group and +0.6 IU in the

TRADJENTA group. The mean change in body weight from baseline to Week 24 was similar in the two treatment groups. Table 12 Glycemic Parameters in Placebo-Controlled Trial for TRADJENTA in Combination with Insulin* TRADJENTA 5 mg + Insulin Placebo + Insulin *Full analysis population using last observation carried forward (LOCF) method on trial **TRADJENTA + Insulin, n=595; Placebo + Insulin, n=593 ***HbA1c: ANCOVA model included treatment, categorical renal function impairment status and concomitant OADs as class-effects, as well as baseline HbA1c as continuous covariates.

FPG: ANCOVA model included treatment, categorical renal function impairment status and concomitant OADs as class-effects, as well as baseline HbA1c and baseline FPG as continuous covariates. A1C (%) Number of patients n = 618 n = 617 Baseline (mean) 8.3

Change from baseline (adjusted mean***) -0.6 0.1 Difference from placebo (adjusted mean)

(95% CI) -0.7 (-0.7, -0.6) -- Patients achieving A1C <7%** 116 48 FPG (mg/dL) Number of patients n = 613 n = 608 Baseline (mean) 147 151 Change from baseline (adjusted mean***) -8 3 Difference from placebo (adjusted mean) (95% CI) -11 (-16, -6) -- The difference between treatment with linagliptin and placebo in terms of adjusted mean change from baseline in HbA1c after 24 weeks was comparable for patients with no renal impairment (eGFR ≥90 mL/min, n=539), with mild renal impairment (eGFR 60 to <90 mL/min, n= 565), or with moderate renal impairment (eGFR 30 to <60 mL/min, n=124). Renal Impairment A total of 133 patients with type 2 diabetes mellitus participated in a 52 week, double-blind, randomized, placebo-controlled trial designed to evaluate the efficacy and safety of TRADJENTA in patients with both type 2 diabetes mellitus and severe chronic renal impairment. Participants with an estimated (based on the four variables modified diet in renal disease equation) GFR value of <30 mL/min were eligible to participate in the trial. Randomization was stratified by baseline HbA1c (≤8% and >8%) and background antidiabetic therapy (insulin or any combination with insulin, SU or glinides as monotherapy and pioglitazone or any other antidiabetics excluding any other DPP-4 inhibitors). For the initial 12 weeks of the trial, background antidiabetic therapy was kept stable and included insulin, sulfonylurea, glinides, and pioglitazone.

For the remainder of the trial, dosage adjustments in antidiabetic background therapy were allowed. At baseline in this trial, 62.5% of patients were receiving insulin alone as background diabetes mellitus therapy, and 12.5% were receiving sulfonylurea alone. After 12 weeks of treatment, TRADJENTA 5 mg provided statistically significant improvement in A1C compared to placebo, with an adjusted mean change of -0.6% compared to placebo (95% confidence interval -0.9, -0.3) based on the analysis using last observation carried forward (LOCF). With adjustments in antidiabetic background therapy after the initial 12 weeks, efficacy was maintained for 52 weeks, with an adjusted mean change from baseline in A1C of -0.7% compared to placebo (95% confidence interval -1.0, -0.4) based on analysis using LOCF.

Cardiovascular Safety Trials in Patients with Type 2 Diabetes Mellitus

CARMELINA The cardiovascular risk of TRADJENTA was evaluated in CARMELINA, a multi-national, multi-center, placebo-controlled, double-blind, parallel group trial comparing TRADJENTA (N=3,494) to placebo (N=3,485) in adult patients with type 2 diabetes mellitus and a history of established macrovascular and/or renal disease. The trial compared the risk of major adverse cardiovascular events (MACE) between TRADJENTA and placebo when these were added to standard of care treatments for diabetes mellitus and other cardiovascular risk factors. The trial was event driven, the median duration of follow-up was 2.2 years and vital status was obtained for 99.7% of patients.

Patients were eligible to enter the trial if they were adults with type 2 diabetes mellitus, with HbA1c of 6.5% to 10%, and had either albuminuria and previous macrovascular disease (39% of enrolled population), or evidence of impaired renal function by eGFR and Urinary Albumin Creatinine Ratio (UACR) criteria (42% of enrolled population), or both (18% of enrolled population). At baseline the mean age was 66 years and the population was 63% male, 80% White, 9% Asian, 6% Black or African American, and 36% were of Hispanic or Latino ethnicity. Mean HbA1c was 8.0% and mean duration of type 2 diabetes mellitus was 15 years. The trial population included 17% patients ≥75 years of age and 62% patients with renal impairment defined as eGFR <60 mL/min/1.73 m 2. The mean eGFR was 55 mL/min/1.73 m 2 and 27% of patients had mild renal impairment (eGFR 60 to 90 mL/min/1.73 m 2 ), 47% of patients had moderate renal impairment (eGFR 30 to <60 mL/min/1.73 m 2 ) and 15% of patients had severe renal impairment (eGFR <30 mL/min/1.73 m 2 ). Patients were taking at least one antidiabetic drug (97%), and the most common were insulin and analogues (57%), metformin (54%) and sulfonylurea (32%). Patients were also taking antihypertensives (96%), lipid lowering drugs (76%) with 72% on statin, and aspirin (62%). The primary endpoint, MACE, was the time to first occurrence of one of three composite outcomes which included cardiovascular death, non-fatal myocardial infarction or non-fatal stroke.

The trial was designed as a non-inferiority trial with a pre-specified risk margin of 1.3 for the hazard ratio of MACE. The results of CARMELINA, including the contribution of each component to the primary composite endpoint, are shown in Table 13. The estimated hazard ratio for MACE associated with TRADJENTA relative to placebo was 1.02 with a 95% confidence interval of. The upper bound of this confidence interval, 1.17, excluded the risk margin of 1.3. The Kaplan-Meier curve depicting time to first occurrence of MACE is shown in Figure 1. Table 13 Major Adverse Cardiovascular Events (MACE) by Treatment Group in the CARMELINA Trial TRADJENTA 5 mg n = 3,494 Placebo n = 3,485 Hazard Ratio Number of Subjects (%) Incidence Rate per 1,000 PY* Number of Subjects (%) Incidence Rate per 1,000 PY* (95% CI) *PY=patient years **A patient may have experienced more than one component; therefore, the sum of the components is larger than the number of patients who experienced the composite outcome. Composite of first event of CV death, non-fatal myocardial infarction (MI), or non-fatal stroke (MACE) 434 57.7 420 56.3 1.02 CV death** 255 32.6 264 34.0 0.96 Non-fatal MI** 156 20.6 135 18.0 1.15 Non-fatal stroke** 65 8.5 73 9.6 0.88 Figure 1 Kaplan-Meier: Time to First Occurrence of MACE in the CARMELINA Trial Figure 1 CAROLINA The cardiovascular risk of TRADJENTA was evaluated in CAROLINA, a multi-center, multi-national, randomized, double-blind, parallel group trial comparing TRADJENTA (N=3,023) to glimepiride (N=3,010) in adult patients with type 2 diabetes mellitus and a history of established cardiovascular disease and/or multiple cardiovascular risk factors.

The trial compared the risk of major adverse cardiovascular events (MACE) between TRADJENTA and glimepiride when these were added to standard of care treatments for diabetes mellitus and other cardiovascular risk factors. The trial was event driven, the median duration of follow-up was 6.23 years and vital status was obtained for 99.3% of patients. Patients were eligible to enter the trial if they were adults with type 2 diabetes mellitus with insufficient glycemic control (defined as HbA1c of 6.5% to 8.5% or 6.5% to 7.5% depending on whether treatment-naïve, on monotherapy or on combination therapy), and were defined to be at high cardiovascular risk with previous vascular disease, evidence of vascular related end-organ damage, age ≥70 years, and/or two cardiovascular risk factors (duration of diabetes mellitus >10 years, systolic blood pressure >140 mmHg, current smoker, LDL cholesterol ≥135 mg/dL). At baseline the mean age was 64 years and the population was 60% male, 73% White, 18% Asian, 5% Black or African American, and 17% were of Hispanic or Latino ethnicity.

The mean HbA1c was 7.15% and mean duration of type 2 diabetes mellitus was 7.6 years. The trial population included 34% patients ≥70 years of age and 19% patients with renal impairment defined as eGFR <60 mL/min/1.73 m 2. The mean eGFR was 77 mL/min/1.73m 2. Patients were taking at least one antidiabetic drug (91%) and the most common were metformin (83%) and sulfonylurea (28%). Patients were also taking antihypertensives (89%), lipid lowering drugs (70%) with 65% on statin, and aspirin (47%). The primary endpoint, MACE, was the time to first occurrence of one of three composite outcomes which included cardiovascular death, non-fatal myocardial infarction or non-fatal stroke. The trial was designed as a non-inferiority trial with a pre-specified risk margin of 1.3 for the upper bound of the 95% CI for the hazard ratio of MACE. The results of CAROLINA, including the contribution of each component to the primary composite endpoint, are shown in Table 14. The Kaplan-Meier curve depicting time to first occurrence of MACE is shown in Figure 2. Table 14 Major Adverse Cardiovascular Events (MACE) by Treatment Group in the CAROLINA Trial TRADJENTA 5 mg n=3,023 Glimepiride (1 mg to 4 mg) n=3,010 Hazard Ratio Number of Subjects (%) Incidence Rate per 1,000 PY* Number of Subjects (%) Incidence Rate per 1,000 PY* (95% CI) *PY=patient years **A patient may have experienced more than one component; therefore, the sum of the components is larger than the number of patients who experienced the composite outcome Composite of first event of CV death, non-fatal myocardial infarction (MI), or non-fatal stroke (MACE) 356 20.7 362 21.2 0.98 CV death** 169 9.2 168 9.2 1.00 Non-fatal MI** 145 8.3 142 8.2 1.01 Non-fatal stroke** 91 5.2 104 6.0 0.87 Figure 2 Time to First Occurrence of 3P-MACE in CAROLINA Figure 2