Long-term safety and efficacy of linagliptin as monotherapy or in combination with other oral glucose-lowering agents in 2121 subjects with type 2 diabetes: up to 2 years exposure in 24-week phase III trials followed by a 78-week open-label extension

Summary

Aim:  The aim of this study was to evaluate the long-term safety, tolerability and efficacy of the dipeptidyl peptidase-4 inhibitor linagliptin given either alone or in combination with other oral glucose-lowering agents in persons with type 2 diabetes.

Methods:  A 78-week open-label extension study evaluated subjects who participated in one of four preceding 24-week, randomised, double-blind, placebo-controlled parent trials and who received linagliptin, linagliptin + metformin, linagliptin + metformin + a sulphonylurea or linagliptin + pioglitazone (all with linagliptin administered orally once daily). Individuals receiving one of these treatments during a previous trial continued the same treatment (n = 1532) for up to a total of 102 weeks, whereas those previously receiving placebo were switched to linagliptin (n = 589). All 2121 participants received at least one dose of the trial medication and were included in the primary safety analysis.

Results:  In subjects previously receiving active treatment, the glycosylated haemoglobin A_1c reduction achieved during the 24-week parent trials was sustained through the 78-week extension period (change from baseline to week 102: −0.8%). Drug-related adverse events were experienced by 14.3% of participants. Hypoglycaemia occurred in 13.9% of participants and was similar between those previously receiving treatment (13.6%) and those switching from placebo to linagliptin (14.6%). Hypoglycaemia occurred most frequently with the use of metformin + a sulphonylurea background therapy (11%). Overall, no clinically relevant changes in body weight were observed.

Conclusion:  Long-term treatment with linagliptin was well tolerated with no change in the safety profile observed during the extension study. Sustained long-term glycaemic control was maintained for up to 102 weeks with either linagliptin monotherapy or linagliptin in combination with other oral glucose-lowering agents.

What’s known

In 24-week, randomised, double-blind, placebo-controlled trials in subjects with type 2 diabetes, linagliptin as monotherapy or in combination with other oral glucose-lowering agents was shown to be effective in improving glycaemic control with no weight gain and without independently increasing the risk for hypoglycaemia.

What’s new

Linagliptin as monotherapy or in combination with other oral glucose-lowering agents is a safe and effective option for the long-term treatment of type 2 diabetes. The glycosylated haemoglobin A_1c reduction (change from baseline: –0.8%) in subjects previously receiving active treatment was sustained for up to 102 weeks. Overall, no relevant weight changes were observed throughout treatment during the extension period regardless of concomitant therapy.

Introduction

The management of persons with type 2 diabetes usually begins with dietary and lifestyle modification and typically progresses to pharmacological intervention. Most oral antidiabetic drugs (OADs) either address insulin resistance (e.g. metformin or thiazolidinediones) or increase the secretion of endogenous insulin (e.g. sulphonylureas, glinides or dipeptidyl peptidase-4 [DPP-4] inhibitors). As pancreatic β-cell function declines over time, the maintenance of glycaemic control with OADs becomes more challenging and treatment intensification with dual or even triple therapy frequently becomes necessary (1,2). However, common drug therapy may be associated with undesirable side effects such as excess weight gain, hypoglycaemia, gastrointestinal-related side effects and fluid retention (3). In light of these limitations, there is still a great need to explore new treatment options for type 2 diabetes.

Dipeptidyl peptidase-4 inhibitors represent a relatively new pharmacological approach to the treatment of type 2 diabetes. DPP-4 inhibitors act to lower blood glucose levels by preventing the degradation of the incretin hormones such as glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic peptide, with a resulting increase in the stimulation of insulin secretion and the inhibition of glucagon secretion in a glucose-dependent manner (4). Linagliptin is a DPP-4 inhibitor with a unique pharmacological profile and a very high affinity for the DPP-4 enzyme (5,6). It has been shown to inhibit plasma DPP-4 activity by more than 80% over 24 h, and it has a low potential for drug-drug interactions and elimination primarily via bile into the gastrointestinal tract. Because of the predominantly non-renal mode of excretion, no dose adjustment for linagliptin is required in patients with any degree of renal impairment (7). In four 24-week, randomised, double-blind, placebo-controlled phase III trials, linagliptin as monotherapy (8) or in combination with metformin (9), metformin plus a sulphonylurea (10) or a thiazolidinedione (pioglitazone) (11) was shown to be efficacious and well tolerated among treatment-naïve and/or treatment-experienced adult subjects with inadequately controlled type 2 diabetes. Trial participants who had completed one of the four 24-week, randomised, double-blind, placebo-controlled parent trials (8–11) were eligible for a 78-week open-label extension study to evaluate the long-term safety and efficacy of linagliptin.

Methods

Study design and subjects

This large extension trial was carried out at 231 sites in 32 countries (Argentina, Austria, Belgium, Canada, China, Croatia, the Czech Republic, Finland, Germany, Greece, Hungary, India, Israel, Italy, Japan, Korea, Malaysia, Mexico, the Netherlands, New Zealand, the Philippines, Poland, Romania, Russia, Slovakia, Spain, Sweden, Taiwan, Thailand, Ukraine, the United Kingdom and the United States) in persons with type 2 diabetes who had participated in one of four previous 24-week, randomised, double-blind, placebo-controlled parent trials (8–11). The open-label extension comprised a 78-week treatment period followed by a 1-week follow-up (12).

Briefly, in the four double-blind parent trials, subjects with treatment-naïve and/or oral treatment failures with inadequately controlled type 2 diabetes were randomised to receive either placebo or linagliptin 5 mg orally once daily for 24 weeks:

• • 
  as monotherapy (8);
• • 
  in addition to metformin background therapy (9);
• • 
  in addition to metformin plus a sulphonylurea background therapy (10) or;
• • 
  in addition to pioglitazone (30 mg) as initial combination treatment (11).

Participants were eligible for this open-label extension if they had completed an entire treatment period during the previous double-blind trials regardless of whether they were treated with rescue medication. Women of child-bearing potential who were not practicing an acceptable method of birth control and pregnant or nursing women were excluded from the studies as were premenopausal women (last menstruation ≤ 1 year prior to informed consent). Other exclusion criteria included alcohol or drug abuse. Those choosing to enrol in the extension study were provided with information and signed written consent obtained before participation; they continued for another 78 weeks in their originally assigned treatment groups. Thus, patients previously receiving linagliptin would have been receiving treatment for up to 102 weeks. Participants who received placebo in the previous trials were given linagliptin 5 mg in the extension trial.

Rescue therapy could be initiated at any stage of the trial if a participant had glycosylated haemoglobin A_1c (HbA_1c) > 8.5%. Depending on background medication, rescue therapy could be metformin, a sulphonylurea, pioglitazone or the additional use of insulin. During the first 6 weeks of the extension period, rescue therapy could be initiated if a participant had a confirmed fasting plasma glucose (FPG) level > 11.1 mmol/l (> 200 mg/dl) or a randomly determined glucose level > 22.2 mmol/l (> 400 mg/dl). During the last 72 weeks of the trial, rescue therapy could be initiated if a participant had a confirmed FPG level > 10.0 mmol/l (> 180 mg/dl) or a randomly determined glucose level > 19.4 mmol/l (> 350 mg/dl). If participants failed to meet these pre-specified glycaemic criteria despite rescue therapy, they were discontinued from the study. The occurrence of hypoglycaemic events that could have put the participant at continued risk was an additional withdrawal criterion during the extension.

Routine laboratory analyses and determinations of HbA_1c and plasma glucose were performed by a central laboratory (Clearstone Central Laboratories, with facilities in the United States, China, Argentina and France). The study was carried out in accordance with the Declaration of Helsinki and the International Conference on Harmonisation Good Clinical Practice principles. The protocol was approved by the independent ethics committee or institutional review board for each participating site and by health authorities if required. The 78-week open-label extension was initiated on 13 August 2008 and completed on 29 December 2010 (ClinicalTrials.gov identifier NCT00736099).

Study outcomes

The extension study was conducted primarily to assess the long-term safety and tolerability of linagliptin given either alone or in combination with other oral glucose-lowering agents. Safety and tolerability outcomes included the incidence and intensity of adverse events (AEs) during the 78-week extension, withdrawals owing to AEs, and the results of physical examinations, 12-lead electrocardiograms, vital signs and clinical laboratory assessments (haematology, clinical chemistry and urinalysis). Hypoglycaemic events and severe hypoglycaemic events also were recorded (13). In addition, an independent clinical event committee (CEC) prospectively reviewed in a masked fashion all reported treatment-emergent fatal events, suspected events of stroke, myocardial ischaemia (including myocardial infarction [MI]), hospitalisation for heart failure, stent thrombosis or revascularisation procedures. CEC members evaluated whether pre-specified criteria for adjudication endpoints (cardiovascular [CV] death, stroke, MI and hospitalisation for unstable angina) were met.

Secondary efficacy outcomes included the HbA_1c change from baseline over time, the occurrence of a treat-to-target efficacy response (HbA_1c < 7.0% at week 78) and the occurrence of a relative efficacy response (HbA_1c lowering by at least 0.5% at week 78). Other secondary outcomes included the change from baseline over time in FPG, body weight and the use of rescue therapy.

Statistical analyses

All participants who received at least one dose of trial medication (treated set) were included in the safety analysis. All participants who received at least one dose of trial medication and who had both a baseline and at least one on-treatment HbA_1c measurement (full analysis set [FAS]) were included in the efficacy analysis. All safety and efficacy variables were summarised using descriptive statistics without tests for significance. The observed cases (OC) approach was applied for the analysis of primary and secondary outcomes. For participants initiating rescue therapy, HbA_1c values were set to missing after the start of rescue medication. Participants also were analysed according to their previous exposure to linagliptin: group A comprised those participants who received linagliptin in the previous trials, whereas group B comprised those previously randomised to receive placebo. For 78-week efficacy, a coefficient of durability was calculated to compare non-inferiority of HbA_1c to 24-week results (difference between HbA_1c at 24 weeks minus HbA_1c at end of study).

Results

Subject disposition and baseline characteristics

Of 2121 subjects who received trial medication, 1880 (88.6%) completed the 78-week extension study (Figure 1). The primary reasons for discontinuation were AEs (3.7%), refusal to continue medication (2.6%) and lack of therapeutic effect (1.1%). Of the participants treated in this trial, 1532 (72.2%) received linagliptin treatment as monotherapy or in combination with other glucose-lowering drugs during the previous study and remained on the same treatment in the extension phase (group A); 589 (27.8%) received placebo during the previous studies and were switched to linagliptin monotherapy in the extension phase (group B).

Figure 1.  Study disposition. *Treated set includes all subjects who received at least one dose of treatment

Demographics and baseline characteristics of the study population are summarised in Table 1. Overall, the study cohort had a mean age of 57.5 years, 75.1% of participants were younger than 65 years and 51.8% were men. The majority of participants had a body mass index < 30 kg/m² (62.4%), had a normal or mildly impaired renal function (95.6%) and had a HbA_1c < 8% (71.2%). As expected, mean baseline HbA_1c and FPG were lower in group A than in group B (7.4% vs. 7.9% and 8.4 mmol/l vs. 9.1 mmol/l [151.6 mg/dl vs. 164.3 mg/dl], respectively). About half of the participants had a diagnosis of type 2 diabetes for more than 5 years (52.5%). Mean exposure to trial medication was 515 days (513 in group A; 520 in group B). Median exposure to trial medication was 547 days in both groups.

Table 1.   Demographics and baseline characteristics in the treated set

	Group A* (n = 1532)	Group B^† (n = 589)	Total (n = 2121)
Data are expressed as mean ± SD unless otherwise indicated. *Subjects randomised to linagliptin as monotherapy or in combination with other oral glucose-lowering agents in the four previous trials and remained on the same treatment in the extension study. ^†Patients randomised to placebo in the four previous trials and switched to linagliptin monotherapy in the extension study. ^‡eGFR ≥ 90 ml/min per 1.73 m^2. ^§eGFR: 60–90 ml/min per 1.73 m^2. ^¶eGFR: 30–60 ml/min per 1.73 m^2. **eGFR < 30 ml/min per 1.73 m^2. BMI, body mass index; eGFR, estimated glomerular filtration rate calculated by the modification of diet in renal disease equation; HbA1c, haemoglobin A1c; FPG, fasting plasma glucose; SD, standard deviation.
Age, years	57.8 ± 9.8	56.7 ± 10.0	57.5 ± 9.9
Age group, n (%)
< 65 years	1142 (74.5)	450 (76.4)	1592 (75.1)
≥ 65 years	390 (25.5)	139 (23.6)	529 (24.9)
Gender, n (%)
Male	789 (51.5)	310 (52.6)	1099 (51.8)
Female	743 (48.5)	279 (47.4)	1022 (48.2)
Race, n (%)
White	867 (56.6)	335 (56.9)	1202 (56.7)
Asian	646 (42.2)	245 (41.6)	891 (42.0)
Black/African American	8 (0.5)	3 (0.5)	11 (0.5)
Other	11 (0.7)	6 (1.0)	17 (0.8)
Body weight, kg	78.7 ± 16.9	80.1 ± 16.9	79.1 ± 16.9
BMI, kg/m^2	28.9 ± 4.8	29.2 ± 4.9	29.0 ± 4.9
BMI < 30 kg/m^2, n (%)	965 (63.0)	359 (61.0)	1324 (62.4)
BMI ≥ 30 kg/m^2, n (%)	567 (37.0)	230 (39.1)	797 (37.6)
eGFR, n (%)
Normal renal function^‡	853 (55.7)	347 (58.9)	1200 (56.6)
Mild renal impairment^§	605 (39.5)	222 (37.7)	827 (39.0)
Moderate renal impairment^¶	73 (4.8)	20 (3.4)	93 (4.4)
Severe renal impairment**	1 (0.1)	0 (0.0)	1 (0.1)
Duration of type 2 diabetes, n (%)
≤ 1 year	226 (14.8)	110 (18.7)	336 (15.8)
> 1 and ≤ 5 years	476 (31.1)	196 (33.3)	672 (31.7)
> 5 years	830 (54.2)	283 (48.0)	1113 (52.5)
HbA1c, %	7.4 ± 0.9	7.9 ± 1.0	7.5 ± 1.0
HbA1c distribution at baseline, n (%)
HbA1c < 7%	532 (34.7)	108 (18.3)	640 (30.2)
HbA1c ≥ 7 and < 8%	639 (41.7)	230 (39.1)	869 (41.0)
HbA1c ≥ 8 and < 9%	272 (17.8)	164 (27.8)	436 (20.6)
HbA1c ≥ 9%	89 (5.8)	87 (14.8)	176 (8.3)
FPG, mmol/l	8.4 ± 2.0	9.1 ± 2.1	8.6 ± 2.0
FPG, mg/dl	151.6 ± 35.3	164.3 ± 37.4	155.1 ± 36.3

Safety outcomes

Overall, 1718 subjects (81.0%) reported at least one AE during the 78-week extension phase. Generally, the incidences of AEs were similar between group A and group B (Table 2). When stratified by previous trials, the highest incidence of AEs was noted in participants receiving metformin plus a sulphonylurea background therapy (84.2%), followed by those participants receiving metformin background therapy (81.6%). Participants receiving linagliptin monotherapy had an AE incidence rate of 78.8% and those receiving linagliptin plus pioglitazone reported the lowest incidence of AEs (76.0%). Most AEs were mild or moderate and the incidence of severe AEs was low (3.8%). The incidence of AEs leading to discontinuation also was low (3.4%). Drug-related AEs were experienced by 14.3% of participants.

Table 2.   Safety during the 78-week extension period in the treated set

	Group A (n = 1532)	Group B (n = 589)	Total (n = 2121)
*Any AE which resulted in death, was immediately life-threatening, resulted in persistent or significant disability/incapacity, required or prolonged patient hospitalisation, or was a congenital anomaly/birth defect; or any other serious important medical event which may have jeopardised the patient and may have required medical or surgical intervention to prevent one of those outcomes. ^†No deaths were determined to be drug-related. ^‡Measured plasma glucose concentration ≤ 3.9 mmol/l (≤ 70 mg/dl). AE, adverse event; CV, cardiovascular; CEC, clinical event committee.
Overall incidence of AEs, n (%)
Any AE	1253 (81.8)	465 (79.0)	1718 (81.0)
Drug-related AEs	220 (14.4)	83 (14.1)	303 (14.3)
AEs leading to discontinuation	57 (3.7)	16 (2.7)	73 (3.4)
Serious AEs*	158 (10.3)	52 (8.8)	210 (9.9)
Deaths^†	5 (0.3)	3 (0.5)	8 (0.4)
Most common AE (> 5% in any group) by preferred term, n (%)
Hyperglycaemia	376 (24.5)	121 (20.5)	497 (23.4)
Hypoglycaemia^‡	209 (13.6)	86 (14.6)	295 (13.9)
Nasopharyngitis	164 (10.7)	62 (10.5)	226 (10.7)
Upper respiratory tract infection	123 (8.0)	54 (9.2)	177 (8.3)
Urinary tract infection	77 (5.0)	34 (5.8)	111 (5.2)
Back pain	72 (4.7)	31 (5.3)	103 (4.9)
Frequency of CV events confirmed by CEC, n (%)
Non-fatal stroke	10 (0.7)	2 (0.3)	12 (0.6)
Non-fatal myocardial infarction	4 (0.3)	10 (1.7)	14 (0.7)
CV death	3 (0.2)	1 (0.2)	4 (0.2)
Hospitalisation for unstable angina	5 (0.3)	3 (0.5)	8 (0.4)

Investigator-defined hypoglycaemic events occurred in 295 (13.9%) participants overall. Of those, 146 (6.9%) participants experienced hypoglycaemic events that investigators considered to be drug-related. When stratified by previous trials, the highest frequency of participants with drug-related hypoglycaemic events (Table 3) was recorded in those individuals receiving metformin plus a sulphonylurea as background therapy (11.0%). The rate was much lower for those receiving metformin background therapy (2.1%) and those receiving linagliptin monotherapy and linagliptin plus pioglitazone reported the lowest frequencies of hypoglycaemic events (0.5% and 0.2%, respectively). Among subjects who received linagliptin with a background of metformin plus a sulphonylurea, 32.2% experienced no investigator-defined hypoglycaemia. Incidence of investigator-defined hypoglycaemia was under 10% in each of the three non-sulphonylurea background subgroups.

Table 3.   Severity and frequency of hypoglycaemic events in the treated set

	Group A (n = 1532)	Group B (n = 589)	Total (n = 2121)
*Measured plasma glucose concentration ≤ 3.9 mmol/l (≤ 70 mg/dl). ^†Event requiring the assistance of another person to actively administer carbohydrate, glucagons or other resuscitative actions.
Hypoglycaemia*, n (%)	209 (13.6)	86 (14.6)	295 (13.9)
Severe hypoglycaemic events^†	10 (0.7)	3 (0.5)	13 (0.6)
Frequency of hypoglycaemic events stratified by previous trials
Linagliptin monotherapy	9 (0.6)	2 (0.3)	11 (0.5)
Linagliptin + metformin	31 (2.0)	14 (2.4)	45 (2.1)
Linagliptin + metformin + sulphonylurea	167 (10.9)	67 (11.4)	234 (11.0)
Linagliptin + pioglitazone	2 (0.1)	3 (0.5)	5 (0.2)

Severe hypoglycaemic events occurred in 0.6% of participants overall (Table 3). Hypoglycaemic events leading to trial discontinuation were reported for three (0.1%) participants, all of whom had received linagliptin in combination with metformin and sulphonylurea in the preceding 24-week study prior to entering the extension phase.

Serious AEs were reported in 9.9% of subjects, with eight deaths occurring over the 78-week extension period; none were considered by the investigator to be related to study drug. Long-term treatment with linagliptin was not associated with a clinically relevant change in body weight (−0.03 kg change in subjects previously treated with linagliptin, 0.47 kg in those switched from placebo).

Linagliptin, given either alone or in combination with other oral glucose-lowering agents, was associated with a low frequency of prospectively captured and adjudicated major CV events (Table 2). Laboratory variables and vital signs did not reveal any clinically significant findings. Four patients in the group originally treated during the parent trials (0.2% of the overall treated set) experienced pancreatitis. Of those, two cases were acute and two were chronic. Infections occurred in 37.8% of patients overall during the 102-week period. Of those, nasopharyngitis and upper respiratory infections (URTI) were most common, occurring in 10.7% and 8.3% of patients overall.

Efficacy outcomes

In participants randomised to linagliptin in the four previous trials (group A), the glucose-lowering effect of linagliptin achieved during the initial 24 weeks of treatment (mean change from baseline to week 24: −0.8%) was maintained over the 78 weeks of the extension phase (change from baseline to week 102: −0.8%). Coefficient of durability per 78 weeks, 0.14% (p < 0.0001, non-inferiority 0.3%). Mean HbA_1c change from baseline over time for the FAS (OC) is shown in Figure 2. The largest observed reduction of HbA_1c with linagliptin was noted for those receiving the initial combination of linagliptin plus pioglitazone (change from baseline to week 102: −1.5%), followed by those receiving metformin background therapy and metformin plus a sulphonylurea background therapy (change from baseline to week 102: −0.7%). With linagliptin monotherapy, the HbA_1c change from baseline to week 102 was −0.5% (Figure 3).

Figure 2.  Time course of mean (± standard error) change in haemoglobin A_1c (HbA_1c) from baseline over 102 weeks in group A (full analysis set [observed cases]). The total study period of 102 weeks comprises 24 weeks from the previous trials and 78 weeks from the extension study

Figure 3.  Time course of mean (± standard error) change in haemoglobin A_1c (HbA_1c) from baseline over 102 weeks in subjects treated with (A) linagliptin monotherapy; (B) linagliptin in combination with metformin background therapy; (C) linagliptin in combination with metformin and sulphonylurea background therapy; or (D) linagliptin plus pioglitazone as initial combination in the full analysis set (observed cases). The total study period of 102 weeks comprises 24 weeks from the previous trials and 78 weeks from the extension study

In subjects randomised to placebo in the four previous trials and switched to linagliptin monotherapy in the extension phase (group B), the change in mean HbA_1c was –0.9% after 78 weeks (Figure 4). Overall, 39.8% of participants had missing HbA_1c data at the end of the extension study, mainly because of the use of rescue therapy (25.5%; Table 4).

Figure 4.  Time course of mean (± standard error) change in haemoglobin A_1c (HbA_1c) from baseline over 78 weeks in patients switching from placebo (full analysis set [observed cases])

Table 4.   Number of subjects with non-missing and missing HbA_1c data at the end of the extension study in the full analysis set (observed cases)

	Group A (n = 1531)	Group B (n = 589)	Total (n = 2120)
*Three subjects were not counted as early discontinued owing to haemoglobin A1c (HbA1c) values measured 1 day after drug discontinuation (two subjects) and owing to intake of rescue medication before drug discontinuation (one subject).
Subjects with HbA1c data, n (%)	903 (59.0)	373 (63.3)	1276 (61.2)
Missing data owing to rescue therapy	408 (26.7)	143 (24.3)	551 (26.0)
Missing data owing to early discontinuation	180* (11.8)	58 (9.9)	238 (11.2)
Missing data owing to other reasons	40 (2.6)	15 (2.6)	55 (2.6)

Concerning the treat-to-target efficacy response, 42.3% of subjects in group A and 46.1% of those in group B reached the HbA_1c target of < 7.0% at week 78 of the extension phase. The percentage of subjects with HbA_1c lowering by ≥ 0.5% at week 78 was twice that in group B compared with group A (46.9% vs. 17.1%, respectively).

Changes in FPG showed similar trends to those seen for HbA_1c over the 78-week extension period. In group A, mean FPG values already reduced during the previous trials further decreased during the extension period. In group B, subjects had a more pronounced mean FPG reduction from baseline over time.

The proportion of participants using rescue therapy during the extension study was 30.8% (31.7% in group A; 28.4% in group B). The most common rescue therapies in group A were sulphonylureas (10.7%), pioglitazone (9.2%), metformin (8.9%) and insulin alone (2.2%). In group B, they were metformin (9.7%), sulphonylureas (9.3%), pioglitazone (5.9%) and insulin alone (2.9%). Median times to use of rescue therapy were 181.5 and 180.0 days in group A and group B, respectively.

Discussion

This long-term extension study comprising 2121 participants with type 2 diabetes provides the broadest long-term safety evidence for linagliptin to date. Individuals from four 24-week, randomised, placebo-controlled parent trials were eligible to enter a 78-week open-label extension period. This follow-up of up to 102 weeks was undertaken to explore the long-term safety and tolerability of linagliptin in monotherapy when used with other OADs. Although 81.0% of participants experienced AEs in the extension study, most events were of mild or moderate intensity and only 14.3% were designated as being drug-related.

Nearly one half of all drug-related AEs reported during the 78-week extension were hypoglycaemia (6.9%); however, the overall incidence of hypoglycaemia in subjects receiving linagliptin monotherapy was very low (0.5%). This was not surprising since incretin-stimulated release of insulin is glucose-dependent and hypoglycaemia is uncommon with DPP-4 inhibitors (14). Moreover, only 0.1% of participants discontinued treatment because of hypoglycaemia. Hypoglycaemic events occurred among 2.1% and 0.2% of participants receiving linagliptin in addition to metformin or pioglitazone, respectively. The triple regimen of linagliptin with metformin plus a sulphonylurea was associated with an 11% incidence of hypoglycaemia. These findings are in line with the results from the preceding 24-week parent trials, in which hypoglycaemic events also were more common with this triple combination therapy than with placebo (10). The increased number of participants experiencing hypoglycaemia can be attributed to the background hypoglycaemic risk associated with sulphonylurea therapy. However, linagliptin, either as monotherapy or in combination with other OADs was associated with a very low risk of severe hypoglycaemia (0.6%).

In contrast to what has been reported with other OADs such as sulphonylureas, glinides and thiazolidinediones (3), linagliptin did not induce clinically relevant weight gain. Previous 24-week clinical trials have shown that linagliptin was not associated with weight gain when given to patients receiving other oral agents or to patients who are drug-naïve (10,11), and this also was shown in this study after 102 weeks of treatment. These results concur with the findings of a recent meta-analysis of 41 randomised controlled trials of at least 12 weeks’ duration where DPP-4 inhibitors were found to have no detectable effect on body weight in placebo-controlled trials (+0.2 kg; 95% confidence interval: −0.1 to 0.6; p = 0.11) (15).

Some concerns have arisen over the potential for pancreatitis or pancreatic cancer in patients treated with incretin therapies (either DPP-4 inhibitors or GLP-1 receptor agonists) (16,17). However, observational evidence suggest that any increased risk for pancreatitis may be attributable to complications of type 2 diabetes itself (18,19). In this 78-week extension, the incidence of pancreatitis was very small (0.2%) and no cases of pancreatic cancer were reported.

A recent analysis of the World Health Organisation’s VigiBase individual case safety report dataset for adverse drug reactions showed disproportionality in the reporting of common infections in patients using DPP-4 inhibitor monotherapy compared with those using biguanides (reporting odds ratio: 2.3 [95% confidence interval: 1.9–2.7]) (20). The researchers noted that the data raise suspicions of a link between DPP-4 inhibition and infections, particularly in the upper respiratory tract, but are not conclusive. In the overall population of the present 78-week extension study, incidence rates for nasopharyngitis and URTI were 10.7% and 8.3%, respectively. Incidence for both AEs was similar for patients continuing linagliptin and those switching from placebo at entry into the extension period.

The safety findings of this 78-week extension study support those of a previous pooled analysis of data for 2523 persons with type 2 diabetes who were treated with linagliptin in placebo-controlled phase III clinical trials lasting ≤ 24 weeks (21). In that pooled population linagliptin was well tolerated, with an overall incidence of AEs and serious AEs similar to placebo (AEs, 55.8% vs. 55.0%; serious AEs, 2.8% vs. 2.7%). Participants treated with linagliptin on a background of sulphonylurea therapy experienced hypoglycaemia at a slightly increased rate. Incidences of all infections and URTIs were slightly lower with linagliptin than with placebo, but rates of nasopharyngitis were slightly higher with linagliptin.

A pooled analysis of patient data from the sitagliptin clinical trials programme (N = 10,246; 12 weeks to 2 years) found that rates of nasopharyngitis and URTI were similar for sitagliptin 100 mg and comparators (placebo and active combined) (22). Among 5429 participants receiving sitagliptin, the incidence of nasopharyngitis and URTI was 7.7% and 8.6%, respectively. In a long-term extension study of vildagliptin 50 mg bid vs. rosiglitazone 8 mg qd, the incidence of nasopharyngitis and URTI in patients receiving vildagliptin for over 2 years was 13.0% and 8.7%, respectively (23). These rates of infection with sitagliptin and vildagliptin were similar to the present linagliptin extension as well as the pooled safety analysis (21). The pooled analysis of sitagliptin patient data found lower rates of hypoglycaemia with sitagliptin than with comparators, but the difference was attributed to the use of sulphonylureas as a comparator in two of the 19 trials pooled (22).

Similar to previously reported 24-week observations, linagliptin as monotherapy or as combination therapy was well tolerated in this extension study. No relevant changes in body weight were observed.

Along with compelling safety data, the present extension study provides additional evidence for clinically meaningful and sustained improvements in glycaemic control over time with an HbA_1c reduction from baseline of −0.8% in participants with type 2 diabetes who received linagliptin during the 24-week parent trials and the 78-week extension. A coefficient of durability calculation showed that efficacy at the end of the extension was non-inferior to 24 week results.

The degree of HbA_1c reduction varied somewhat depending on background medication, with mean reductions ranging from −0.5% with linagliptin monotherapy to −1.5% with an initial combination of linagliptin plus pioglitazone. These differences follow the 24-week results of the four parent trials. The linagliptin monotherapy trial and the linagliptin plus pioglitazone trial showed mean placebo-corrected HbA_1c reductions of −0.69% and −0.51%, respectively (8,11). The linagliptin plus metformin trial and the linagliptin plus metformin and sulphonylurea trials reported mean placebo-corrected HbA_1c reductions of −0.64% and −0.62%, respectively, over 24 weeks (9,10). As shown in Figure 3, 24-week glycaemic improvements were sustained for up to 2 years with all four treatments.

Use of rescue therapy was similar between subjects continuing linagliptin therapies (group A) and those switching to linagliptin from placebo controls (group B). In both groups, the majority of participants did not use rescue therapy (68.3% and 71.6%, respectively). For those who did use a rescue therapy, the median time to rescue was approximately 6 months, which reinforces the sustained glycaemic control with linagliptin therapy.

Our findings are limited by the open-label design of the extension study. Thus, glycaemic control and safety were evaluated without blinded comparison. Inferences about the incidence of organ-specific AEs and the comparison to earlier placebo-controlled 24-week trials should be made with appropriate caution.

In conclusion, this is the largest data set of long-term clinical evidence for linagliptin to date. Findings from the 78-week open-label extension conducted among 2121 persons with type 2 diabetes demonstrate sustained glycaemic control (for up 102 weeks treatment duration) and provide safety evidence that supports the efficacy and tolerability profile established in previously reported 24-week studies. For long-term management of type 2 diabetes, linagliptin is an effective and well tolerated therapy.

Acknowledgments

The authors were fully responsible for all content and editorial decisions and were involved at all stages of manuscript development and have approved the final version. Medical writing assistance, supported financially by Boehringer Ingelheim, was provided by Audrey Koïtka-Weber and Mark Poirier (Envision Scientific Solutions) during the preparation of this manuscript. This study was sponsored by Boehringer Ingelheim.

Author contributions

RG, DRO, M-RT and SDP contributed to study concept and design, collected data, were involved in patient enrolment, contributed to endpoint adjudication and participated in the interpretation of the data. SP, AP and H-JW contributed to study concept and design and participated in the interpretation of the data. AP contributed to study concept and design, did the statistical analysis, and analysed and interpreted the data. All authors revised the manuscript for important intellectual content and have seen and approved the final version.