Table 1 Summary of study characteristics

Johansen et al. 2012 [58] (Search 3B)

Pre-specified, prospective, and adjudicated meta-analysis of a phase 3 programme

To determine the cardiovascular safety of linagliptin.

RCTs (phase 3): 8

P: 5239

P ≥ 65 years: 1478

P using linagliptin: 929

P using comparators: 549

P using linagliptin: 175 (1, 617) ^a days

P using comparators: 179 (1, 619) ^a days for total comparators (169 [1, 367] ^adays for placebo and 409 [3, 619] ^a days for active comparators)

Composite of CV death (including fatal stroke and fatal MI), non-fatal stroke, non-fatal MI, and hospitalisation for UAP.

Composites of: (i) CV death, non-fatal stroke, and non-fatal MI; (ii) all adjudicated CV events which included CV death, non-fatal stroke, non-fatal MI, UAP with or without hospitalisation, SAP, and TIA; and (iii) FDA-defined custom MACE derived from 34 unadjudicated MedDRA preferred terms for stroke and MI.

Individual adjudicated components (as listed above) and total mortality.

Banerji et al. 2010 [29] (Search 3B)

Retrospective analysis of the GALIANT study which is a multicentre, randomised, open-label study

To assess the safety profile of vildagliptin compared to TZD as an add-on to metformin in patients with T2DM with mild renal impairment and with normal renal function.

P: 2613

P ≥ 65 years: 519

P normal renal function: 248

P using vildagliptin 100 mg + metformin: 157

P using TZD + metformin: 91

P mild impaired renal function: 271

P using vildagliptin 100 mg + metformin: 184

P using TZD + metformin: 87

12 weeks

Adverse events

Barnett et al. 2013 [31] (Search 3B)

Randomised, double-blind, placebo-controlled trial

To assess the effectiveness of linagliptin in elderly patients with type 2 diabetes.

P: 241

P ≥ 65 years: 241

P using linagliptin: 162

P using placebo: 79

24 weeks

Incidence and intensity of AEs, withdrawals because of AEs, hypoglycaemia, cardiovascular events, and changes in vital signs, laboratory variables, and background treatment.

Barzilai et al. 2011 [30] (Search 3A)

Randomised, double-blind, placebo-controlled trial

To assess efficacy and safety, and tolerability of sitagliptin monotherapy in elderly patients.

P: 206

P ≥ 65 years: 206

P using sitagliptin: 102

P using placebo: 104

24 weeks

AEs, SAEs, and hypoglycaemia.

Chien et al. 2011 [32] (Search 3B)

Randomised, open-labelled, parallel-group study

To examine the effectiveness and tolerability of add-on sitagliptin in elderly T2DM patients with inadequate control to existing OAD combination regimen.

P: 97

P ≥ 65 years: 97

P using sitagliptin + OAD combinations (sulfonylurea, metformin, and alpha glucosidase inhibitors): 49

P using OAD combinations (sulfonylurea, metformin, and alpha glucosidase inhibitors): 48

24 weeks

AEs.

Ferrannini et al. 2009 [54] (Search 3B)

Multicentre, randomised, double-blind, active-controlled study

To evaluate the positioning of DPP-4 inhibitors as add-on to metformin when metformin alone is not sufficient to achieve glycaemic control, the long-term efficacy and safety of vildagliptin vs. SU was examined.

P: 2789

P ≥ 65 years: 712

P using vildagliptin: 351

P using glimepiride: 361

52 weeks

AEs.

Fonseca et al. 2008 [56] (Search 3B)

Multi-centre, double-blind, parallel-group, randomised study

To report of longer term data from a clinical trial, undertaken to assess the efficacy and safety of vildagliptin therapy over one year in patients with long-standing T2DM that was inadequately controlled by insulin therapy.

P: 200

P ≥ 65 years: 60

P using vildagliptin 100 mg/d + insulin: 32

P using placebo + insulin / vildagliptin 50 mg/d + insulin: 28

52 weeks

All AEs, SAEs, and hypoglycaemic events.

Green et al. 2015 [13] (Snowballing)

Randomised, double-blind, placebo-controlled study (Trial Evaluating Cardiovascular Outcomes with Sitagliptin [TECOS])

To assess the long-term cardiovascular safety of adding sitagliptin to usual care, as compared with usual care alone, in patients with type 2 diabetes and established cardiovascular disease.

P: 14,671

P ≥ 65 years: 7735

P using sitagliptin 100 mg/d (or 50 mg/d based on eGFR^b): 3813

P using matching placebo: 3816

Median follow-up 3 years

Composite cardiovascular outcome defined as first confirmed event of cardiovascular death, nonfatal myocardial infarction, nonfatal stroke, or hospitalization for unstable angina.

Hartley et al. 2015 [37] (Search 3B)

Randomised, parallel-group, multinational, non-inferiority clinical trial with an active controlled, double-blind treatment period

To evaluate the efficacy and tolerability of sitagliptin compared with glimepiride in elderly patients with type 2 diabetes mellitus and inadequate glycemic control with diet and exercise alone.

P: 480

P ≥ 65 years: 480

P using sitagliptin (dose based on eGFR^c): 241

P using matching placebo or glimepiride (1 mg once daily up to 6 mg/day): 239

30 weeks

Primary safety endpoint: incidence of AEs of symptomatic hypoglycaemia, defined as an episode with clinical symptoms attributed to hypoglycaemia, without regard to glucose level. Asymptomatic hypoglycaemia, defined as episodes without symptoms of hypoglycaemia, but with a glucose level ≤ 70 mg/dL, was also reported.

Kadowaki et al. 2014 [38] (Search 3B)

Randomised, double-blind, placebo-controlled study

To confirm the efficacy and safety, including the risk of hypoglycaemia, of teneligliptin added to glimepiride in Japanese patients with T2DM inadequately controlled with glimepiride monotherapy.

P: 194

P ≥ 65 years: 61

P using teneligliptin + glimepiride: 27

P using placebo + glimepiride: 34

12-week randomised double-blind period with teneligliptin 20 mg or placebo.

40-week open-label period with teneligliptin 20 or 40 mg.

2-week period without any study drug.

AEs (included hypoglycaemia events) and ADRs.

Matthews et al. 2010 [55] (Search 3B)

Multicentre, randomised, double-blind, double-dummy, active-controlled study

To show that vildagliptin added to metformin is non-inferior to glimepiride in reducing HbA1c levels from baseline over 2 years.

P: 3118

P ≥ 65 years: 789

P using vildagliptin: 392

P using glimepiride: 397

2 years

AEs, SAEs, and hypoglycaemic events.

Rosenstock et al. 2013 [59] (Search 3B)

Multicentre, randomised, double-blind, active controlled study

To prospectively evaluate the efficacy and safety of alogliptin versus glipizide in elderly patients with T2DM over 1 year of treatment.

P: 441

P ≥ 65 years: 441

P using alogliptin: 222

P using glipizide: 219

54 weeks

AEs, hypoglycaemia and major cardiac events.

Schernthaner et al. 2015 [39] (Search 3B)

Multinational, randomised, double-blind, phase IIIb/IV study (GENERATION study)

To assess efficacy and safety of adjunctive saxagliptin vs glimepiride in elderly patients with type 2 diabetes and inadequate glycaemic control.

P: 720

P ≥ 75 years: 287

P using saxagliptin + metformin: 143

P using glimepiride + metformin: 144

52 weeks

AEs, proportion of patients with ≥1 confirmed/severe hypoglycaemic event.

Schweizer et al. 2009 [40] (Search 3A)

Double-blind, randomised, multicentre, active-controlled, parallel-group study

To evaluate the efficacy and tolerability of DPP-4 inhibitor, vildagliptin and metformin in drug naïve elderly patients with type 2 diabetes.

P: 335

P ≥ 65 years: 335

P using vildagliptin: 169

P using metformin: 166

24 weeks

AEs, SAEs, hypoglycaemia and cardiovascular / cerebrovascular events.

Schweizer et al. 2013 [53] (Search 3B)

Post hoc sub-analysis of a multi-centre, randomised, double-blind, parallel-group

To assess the efficacy and tolerability of vildagliptin in elderly T2DM patients with renal impairment.

P: 105

P ≥ 75 years: 105

P using vildagliptin: 50

P using placebo: 55

24 weeks

AEs, SAEs and hypoglycaemia.

Scirica et al. 2013 [12]

Scirica et al. 2014 [41]

Leiter et al. 2015 [42]

Mosenzon et al. 2015 [43]

(Search 3B)

Multicentre, randomised, double-blind, placebo-controlled trial (Saxagliptin Assessment of Vascular Outcomes Recorded in Patients with Diabetes Mellitus [SAVOR] - Thrombolysis in Myocardial Infarction [TIMI] 53 study)

Scirica et al. 2013 and Scirica et al. 2014

To evaluate the safety and efficacy of saxagliptin with respect to CV outcomes in patients with diabetes mellitus who are at risk for CV events.

Leiter et al. 2015

To examine the safety and CV effects of saxagliptin in the predefined elderly (≥ 65 years) and very elderly (≥ 75 years) subpopulations.

Mosenzon et al. 2015

To compare the incidence of fractures between patients with saxagliptin and patients with placebo.

P: 16,492

P ≥ 65 years: 8561

P using saxagliptin: 4290

P using placebo: 4271

P ≥ 75 years: 2330

P using saxagliptin: 1169

P using placebo: 1161

2 years

Scirica et al. 2013

Composite of CV death, MI, or ischemic stroke.

Scirica et al. 2014

Hospitalisation for heart failure.

Leiter et al. 2015

Primary outcome: composite of CV mortality, nonfatal MI, or nonfatal ischemic stroke.

Secondary outcomes: primary composite outcome plus hospitalisation for HF, coronary revascularization, or unstable angina and all components of primary secondary outcomes.

Mosenzon et al. 2015

Bone fractures.

Strain et al. 2013 [52] (Search 3B)

Multicentre, randomised, double-blind, placebo-controlled study

To assess the feasibility of setting and achieving investigator-defined individualised treatment targets for a period of 24 weeks in elderly patients with type 2 diabetes (drug-naive or inadequately controlled on oral agents), with the addition of a single oral agent: vildagliptin.

P: 278

P ≥ 65 years: 278

P using vildagliptin: 139

P using placebo: 139

24 weeks

AEs, SAEs, and hypoglycaemia.

White et al. 2013 [44] (Search 3B)

Multicenter, randomised, double-blind placebo-controlled trial

To determine whether alogliptin is noninferior to placebo with respect to major cardiovascular events in patients with type 2 diabetes who are at very high cardiovascular risk — those with recent acute coronary syndromes.

P: 5380

P ≥ 65 years: 1907

P using alogliptin: 973

P using placebo: 934

Median follow-up 18 months

Composite of death from cardiovascular causes, nonfatal myocardial infarction (MI), or nonfatal stroke. Principal secondary safety end point: primary composite end point with the addition of urgent revascularization due to unstable angina within 24 h after hospital admission

Chang et al. 2015 [33] (Search 3B)

Nationwide retrospective cohort study

To compare CV risks associated with second-line oral antidiabetic agents added to initial metformin therapy.

P: 36,118

Subgroup analysis by age (<65 vs ≥65 years) without number of P in each group

P using DPP-4 inhibitors + metformin: 2242

P using SU + metformin: 29,101

P using glinides + metformin: 1553

P using pioglitazone + metformin: 1283

P using α-glucosidase inhibitor + metformin: 1939

Median follow-up ranged from 215 days for the α-glucosidase inhibitor plus metformin group to 305 days for the SU plus metformin group

First hospitalization for acute MI, HF, ischemic stroke after initiation of one of the regimens studied.

Chen et al. 2015 [34] (Search 3B)

Nationwide population-based cohort study

To evaluate efficacy and safety of sitagliptin with respect to cardiovascular outcomes in patients with T2DM and recent ischaemic stroke.

P: 5145

P ≥ 75 years: 1435

P using sitagliptin: 486

P using comparison^d: 949

Mean follow-up 1.17 years (0.75)^e

Primary outcome: composite event of ischemic stroke, MI, or CV death.

Secondary outcomes: haemorrhagic stroke, nonfatal ischemic stroke, nonfatal acute MI, deaths of any cause, and hospitalisation for HF.

Safety outcomes: acute or chronic pancreatitis, hypoglycaemia, hyperosmolar hyperglycaemic state, and diabetic ketoacidosis.

Driessen et al. 2014 [45] (Reference list)

Retrospective population based cohort study

To investigate the association between the use of DPP4-I and the risk of fracture.

P: 433,632

P ≥ 70 years: 141,322

P using NIAD: 68,801

P without prescription of NIAD: 68,015

P using DPP-4 inhibitor: 4506

Median follow-up:

P using NIAD: 3.7 years (1.61–5.22)^f

P without prescription of NIAD: 3.95 years (1.79–5.22)^f

P using DPP-4 inhibitor: 5.0 years (2.95–5.16)^f

Any fracture.

Giorda et al. 2015 [48] (Search 3B)

Population-based nested case-control study

To compare the occurrence of HF in relation to DPP-4 inhibitor use versus any antidiabetic treatment.

Any admission for HF

Cases: 14,613

Controls: 146,130

P ≥ 65 years:

Cases: 13,736

Controls: 137,362

Incident HF

Cases: 7212

Controls: 72,120

P ≥ 65 years:

Cases: 6779

Controls: 67,793

Re-admission for HF

Cases: 1712

Controls: 17,222

P ≥ 65 years:

Cases: 1609

Controls: 16,189

All-cause mortality

Cases: 38,248

Controls: 382,313

P ≥ 65 years:

Cases: 36,335

Controls: 363,197

Not reported

Any admission for HF, incident HF, re-admission for HF, all-cause mortality.

Mistry et al. 2011 [57] (Search 3B)

Retrospective observational survey

To obtain efficacy and safety data on HbA1C levels and incidence of hypoglycaemia in elderly patients who were receiving vildagliptin.

P: 72

P ≥ 65 years: 72

P using vildagliptin + metformin: 52

P using vildagliptin +2 OADs (metformin, SUs and/or TZDs): 20

Median follow up:

Dual therapy: 7 months

Triple therapy: 12 months

Incidence of hypoglycaemic events before and after initiation of vildagliptin.

Ou et al. 2015 [35] (Search 3B)

Nationwide population-based observational cohort study

To compare clinical outcomes of adding DPP-4 inhibitors versus sulfonylureas to metformin therapy in patients with T2DM.

P using DPP-4 inhibitor: 10,089 (propensity score matching)

P using SU: 10,089 (propensity score matching)

P ≥ 65 years:

P using DPP-4 inhibitor: 2825

P using SU: 2825

Mean follow-up 3.3 years

All-cause mortality, MACEs (including ischemic stroke and MI), hospitalisation for HF, and hospitalisation for hypoglycaemia.

Penfornis et al. 2012 [49] (Search 3B)

Prospective cohort study

To compare DPP-4 inhibitors with COAD in the real-life treatment of elderly patients with T2DM uncontrolled on metformin alone. The primary objective was to assess the incidence of hypoglycaemic episodes in relationship with glycaemic control assessed by HbA1c level.

P: 1188

P ≥ 65 years: 1188

P using DPP-4 inhibitors: 931

P using COAD: 257

6 months

Hypoglycaemic events.

Shih et al. 2015 [36] (Search 3B)

Nested case-control study from a cohort of patients with T2DM treated with OADs

To investigate whether susceptibility to sepsis differed among patients with T2DM taking different classes of OAD.

Cases: 43,015

Controls: 43,015

P ≥ 65 years:

Cases: 41,725

Controls: 41,725

Not reported

First hospitalisation for sepsis.

Sicras-Mainar and Navarro-Artieda 2014 [50] (Search 3B)

Multicenter, retrospective, observational study

To describe the clinical (treatment adherence, metabolic control, hypoglycemia, and macrovascular complications) and economic (resource use and costs) consequences of using a combination of metformin + vildagliptin to treat type 2 diabetes in elderly patients.

P: 987

P ≥ 65 years: 987

P using metformin + vildagliptin: 270

P using metformin + SU: 717

24 months

Hypoglycaemia.

Macrovascular complications and cardiovascular events (heart disease, cardiac ischemia, acute myocardial infarction, and heart failure), cerebrovascular disease (stroke [ischemic or haemorrhagic], and transient ischemic attack), all types of peripheral arterial disease and renal disease.

Tziomalos et al. 2015 [51] (Search 3B)

Observational study

To evaluate whether prior antidiabetic treatment affects acute ischaemic stroke severity and in-hospital outcome and whether there are differences between antidiabetic agents regarding these effects.

P: 100

P ≥ 65 years: 98

P using DPP-4 inhibitors: 26

P using other antidiabetic agents: 72

Not reported

Acute ischemic stroke severity measured with the modified Rankin Scale score at discharge and with in-hospital mortality.

Viljoen et al. 2013 [46] (Search 3B)

Observational study

To study the efficacy and tolerability of DPP-4 inhibitors in older patients with type 2 diabetes whilst focusing on particular pertinent aspects relevant to care of older persons.

P: 431

P ≥ 65 years: 431

P using DPP-4 inhibitors: 129

P never treated with DPP-4 inhibitors: 302

Not reported

Hypoglycaemia.

Yu et al. 2015 [47] (Search 3B)

Cohort study with a nested case-control analysis

To determine whether the use of incretin-based drugs, including DPP-4 inhibitors and GLP-1 analogs, is associated with an increased risk of CHF among patients with T2DM.

P: 57,737

Incident cases of hospitalised CHF: 1118

Matched controls: 17,626

P ≥ 65 years:

Cases: 861

Controls: 13,572

Mean duration of treated T2DM 2.4 (3.5)^e years

Hospitalisation for a first CHF.