Nr

Auteur en jaartal

Redenen van exclusie

Mogelijke systematische reviews

1

Scheen 2015

Expert opinie

2

Brietzke 2015

Achtergrondartikel

3

Tella 2014

Niet verkrijgbaar

6

Russel 2013

Review, niet systematisch

9

Schernthaner 2012

Gepoolde analyse niet systematisch

10

Olokoba 2012

Narrative review

13

Cornell 2012

Review, niet systematisch

14

Germino 2011

Systematische review slechte kwaliteit

16

Schweizer 2010

Gepoolde analyse, niet systematisch.

17

Schwartz 2010

Systematische review matige kwaliteit

18

Keating 2010

Review, niet-systematisch

19

Neumiller 2009

Review, niet systematisch

21

Bode 2009

Niet full tekst beschikbaar

Mogelijke RCT’s

1

Mikhail 2015

Review, niet-systematisch.

2

Haak 2015

Review, niet systematisch.

3

Shankar 2015

Gepoolde analyse, niet systematisch

6

Toderika 2014

Review, niet systematisch

7

Seino 2014

Expert opinion

8

Schernthaner 2014

Gepoolde analyse, niet systematisch

9

Scheen 2014

Narrative review, artikel in Frans

10

Scheen 2014

Expert opinion.

15

Mikhail 2014

Review, niet systematisch, niet ouderen.

16

McKeage 2014

Review, niet systematisch, niet ouderen.

17

Keating 2014

Review, niet systematisch, niet ouderen.

19

Groop 2014

Gepoolde analyse, niet systematisch

23

Websky 2013

Review, niet systematisch.

26

Schweizer 2013

Dubbelpublicatie met geïncludeerde studie

35

Scheen 2012

Narrative review

36

Rendell 2012

Expert opinion

37

Penfornis 2012

Real life cohortonderzoek

40

He 2012

Farmacokinetiek en –dynamiek

48

Bourdel-Marchasson 2011

Review, niet systematisch

49

Bode 2011

Gepoolde analyse, niet systematisch

52

Scott 2010

Review, niet systematisch

62

Halimi 2010

Review, niet systematisch

66

Pratley 2009

Gepoolde analyse, niet systematisch

73

Abbatecola 2008

Review, niet systematisch

Diverse designs

7

Gejl 2015

Nested case control, cardiovasculaire eindpunten

9

Avogaro 2015

Review, niet systematisch.

16

Thomsen 2014

Case control study. 

18

Stein 2014

Expert opinion.

46

Marino 2014

Review, niet systematisch.

48

Kurioka 2014

Geen RCT

52

Kim 2014

Geen RCT

81

Valeron 2013

Review, niet systematisch

113

Lamos 2013

Review, niet systematisch

120

Ionica 2013

Geen RCT (cohort)

142

Zanchi 2012

Expert opinion

144

Von Eynatten 2012

Gepoolde analyse 2 RCT’s

200

Rendell 2011

Niet gevonden, waarschijnlijk abstract

205

Patel 2011

Abstract

211

Mistry 2011

Geen RCT, retrospectieve studie

230

Best 2011

Geen RCT. Retrospectieve analyse database

253

Pratley 2009

Abstract? Gepoolde analyse niet systematisch

261

Barzilai 2009

Andere uitkomstmaat

Nr

Auteur en jaartal

Redenen van exclusie

Mogelijke systematische reviews

2

Saeed 2014

Review, niet systematisch

3

Davis 2014

Review, niet systematisch

4

Schernthaner 2012

Gepoolde analyse, niet systematisch

6

Schweizer 2010

Gepoolde analyse, niet systematisch

7

Schwartz 2010

Systematische review maar search en beoordeling niet goed beschreven

Mogelijke RCT’s

1

Shankar 2015

Gepoolde analyse drie studies, niet systematisch

3

Scheen 2015

Review, niet systematisch

4

Scheen 2015

Review, niet systematisch

5

Ludemann 2015

Andere primaire uitkomst, kleine aantallen

7

Leiter 2015

Niet de juiste onderzoekspopulatie (ouderen met HVZ of meerdere risicofactoren), uitkomst cardiovasculair).

8

Inzucchi 2015

Gepoolde analyse van 2 RCT's, niet systematisch. Eén RCT is geïncludeerd (Barnett)

15

Ayvaz 2015

Observationeel cohort, zonder controlegroep.

17

Schernthaner 2014

Gepoolde analyse, niet systematisch

19

Round 2014

Gepoolde analyse, niet systematisch

23

Mikhail 2014

Review, niet systematisch

24

Mikhail 2014

Review, niet systematisch

25

McKeage 2014

Review, niet systematisch

26

Hsieh 2014

Observationeel cohort, zonder controlegroep.

27

von Websky 2013

Review, niet systematisch.

31

Capuano 2013

Review, niet systematisch.

39

Schweizer 2011

Gepoolde analyse, niet systematisch

43

Bourdel-Marchasson 2011

Review, niet systematisch

48

Halimi 2010

Review, niet systematisch

54

Pratley 2007

Gepoolde analyse, niet systematisch

Diverse designs

31

Sicras-Mainar 2014

Observationele retrospectieve studie

Diverse designs

7

Avogaro 2015

Narrative review

9

Pratley 2014

Review, niet systematisch

23

Moghissi 2013

Narrative review

Study reference

Study characteristics

Patient characteristics ²

Intervention (I)

Comparison / control (C) ³

Follow-up

Outcome measures and effect size ⁴

Comments

Rosenstock 2013

Type of study:

RCT

Setting:

Multinational

Country:

USA

Source of funding:

Takeda Pharmaceuticals North America, Inc.

Inclusion criteria:

Age ≥ 65 years

Patients on diet and exercise alone during 2 months before screening and with HbA1C 6,5-9,0%

Patients on oral anti-diabetic monotherapy with HbA1C 6,5-8,0%

Exclusion criteria:

Not reported

N total at baseline:

Intervention: 222

Control: 219

Important prognostic factors²:

Age ± SD:

I: 70,11 ± 4,42

C:69,8 ± 4,07

Sex:

I: 45,9% M

C: 43,8% M

Baseline HbA1C % ± SD:

I: 7,50 % ± 0,703

C: 7,45% ± 0,634

Duration of t2DM

Groups comparable at baseline?

Yes

Treatment:

Alogliptin 25 mg/d

Test:

Blood glucose levels, measured by patients using a blood glucose meter with automatic storage of the results.

Treatment:

Glipizide 5 – 10 mg/d

Test:

Blood glucose levels, measured by patients using a blood glucose meter with automatic storage of the results.

Length of follow-up:

52 weeks

Incomplete outcome data:

Intervention: 89 (41,1%)

Reasons

Hyperglycaemic rescue: 55 (24,8%)

Adverse event: 16 (7,2%)

Lost to follow up: 0

major protocol deviation: 4 (1,8%)

Voluntary withdrawal: 12 (5,4%)

Other: 2 (0,9%)

Control: 94 (43%)

Reasons

Hyperglycaemic rescue: 47 (21,5%)

Adverse event: 20 (9,1%)

Lost to follow up: 4 (1,8%)

major protocol deviation: 7 (3,2%)

Voluntary withdrawal: 16 (7,3%)

Other: 0

Outcome measures and effect size (include 95%CI and p-value if available):

Primary endpoint:

HbA1C mean changes from baseline (7,5%) at week 52: LS mean (s.e.)

I: - 0,14% (0,06)

C: -0,09% (0,07)

One-sided 97,5% CI: - ∞, 0.13%

Secondary endpoints:

Hypoglycaemia:

I: 31 episodes in 5,5% of patients

C: 323 episodes in 26,0% of patients

Severe hypoglycaemia:

I: 0 patients

C: 3 patients

Weight and lipids:  kg (s.d.)

I: -0,62 (0,227) kg

C: +0,60 (0,233) kg (P<0,001)

Triglycerides:

I: -8,0%

C: +1,2% (P=0,046)

No significant differences for  TC; HDL; LDL

Safety:

SAE’s:

I: 16 (7.2%)

C: 13 (5.9%)

Drug related SAE’s (per investigator)

I: 4 (1.8%)

C: 1 (0.5%)

Primary outcomes were conducted using the Per Protocol Set, which included patients  from the Full Analysis Set who had no major protocol violations.

Secondary outcomes were assessed with the FAS, defined as all patients who had at least one dose of study drugs and at least one post-baseline assessment for the particular variable being analysed.

Safety outcomes were assessed by the safety set, and included all patients who had taken one dose of the study drugs.

Hypoglycaemic events were mostly mild to moderate in both groups.

Common AE’s (>5%)

Urinary tract infection, dizziness, headache, diarrhoea, nasopharyngitis, arthralgia, hypertension, back pain, upper respiratory tract infection.  

Study reference

(first author, publication year)

Describe method of randomisation¹

Bias due to inadequate  concealment of allocation?²

(unlikely / likely / unclear)

Bias due to inadequate blinding of participants to treatment allocation?³

(unlikely / likely / unclear)

Bias due to inadequate blinding of care providers to treatment allocation?³

(unlikely / likely / unclear)

Bias due to inadequate blinding of outcome assessors to treatment allocation?³

(unlikely / likely / unclear)

Bias due to selective outcome reporting on basis of the results?⁴

(unlikely / likely / unclear)

Bias due to loss to follow-up?⁵

(unlikely / likely / unclear)

Bias due to violation of

intention to treat analysis?⁶

(unlikely / likely / unclear)

Rosenstock 2013

Randomized, exact method not reported.

Unclear

Unlikely

1. (Glucose levels were measured with an electronic device and stored automatically)

Unlikely

2. (After 52 weeks all glucose meters collected and measurements were collected by study personnel)

Unlikely

3. (After 52 weeks all glucose meters collected and measurements were collected by study personnel)

Unlikely

Unlikely

Unlikely

Study reference

Study characteristics

Patient characteristics ²

Intervention (I)

Comparison / control (C) ³

Follow-up

Outcome measures and effect size ⁴

Comments

Barnett 2013

Type of study:

RCT

Setting:

Multinational

Country:

Australia, Canada, Denmark,

the Netherlands, Sweden

Source of funding:

Sponsoring by Boeringer Ingelheim

Authors participated in

design of the study, interpretation of data, and drafting and revision of the manuscript. RJ participated in design of the study, performed the

statistical analysis, and participated in interpretation of data and drafting and revision of the manuscript were employees of Boehringer Ingelheim.

Inclusion criteria:

T2DM, Age ≥70 yr

Receiving stable dosis of metformin, SU, basal insulin for ≥ 8 weeks.

Exclusion criteria:

fasting plasma glucose  ≥ 13.3 mmol/L, impaired hepatic function, myocardial infarction, stroke or TIA within 3 months before study, previous bariatric surgery, current treatment with rapid-acting or premixed insulin, or systemic steroids.

Treatment within the previous 3 months with thiazolidinedione, αglucosidase inhibitor, meglitinide, GLP1 analogue, DPP4 inhibitor or anti-obesity drug.

N total at baseline:

I: 162

C: 79

N Full analysis set

I: 160

C: 78

Important prognostic factors:

age ± SD:

I: 74,9 (4,4)

C: 74,9 (4,2)

Sex:

I: 71,6% male

C:  62,0% male

Baseline HbA1C:

I: 7,8% (0,8)

C: 7,7% (0,7)

BMI:

I: 29,6 kg/m² (4,7)

C: 29,8 kg/m² (4,5)

Groups comparable at baseline?

Yes 

Treatment:

Linagliptin 5 mg/day in addition to existing glucose lowering treatment

Test:

Every 6 weeks blood sampling of HbA1c, fasting glucose, examination of vital signs, and documentation of any adverse events and use of concomitant therapies. 

Bodyweight and 12 lead ECG at randomisation and week 24

Urine and blood samples for laboratory safety assessments at randomisation, week 12 and week 24.

Treatment:

Placebo in addition to existing glucose lowering treatment

Test:

Every 6 weeks blood sampling of HbA1c, fasting glucose, examination of vital signs, and documentation of any adverse events and use of concomitant therapies. 

Bodyweight and 12 lead ECG at randomisation and week 24

Urine and blood samples for laboratory safety assessments at randomisation, week 12 and week 24.

Length of follow-up:

24 weeks

Loss-to-follow-up:

Intervention: n= 16/162

Reasons:

8 adverse event

7 protocol violation

1 refused to continue

Control: n= 5/79

Reasons:

1 adverse event

3 protocol violation

1 refused to continue

Outcome measures and effect size (include 95%CI and p-value if available):

Primary outcome:

Mean change in HbA1C from baseline (s.e.)

I: -0,61% (0,06)

C: 0,04% (0,07)

Difference linagliptin - placebo: -0,64% (0,08)

95% CI -0,81 to –0,48; P<0,0001)

Secondary outcome:

Proportion of patients achieving HbA1C ≤7,0%

I: 38,9%

C: 8,3%

OR: 8,32; 95% CI 3,32-20,84; P<0,0001

Others:

Proportion of patients with ≥ 0,5% change in baseline

I: 54,4%

C: 12,8%

Placebo adjusted mean reduction in FPG in linagliptin: -1,15 mmol/L (95% CI -1,68 to -0,62, P<0,0001)

Safety:

Proportion patients reporting an  AE. N (%)

I: 123  (75,9%)

C:60 (75,9%)

RR: 1,01; 95%CI 0,87 to 1,18; p=0,87

Proportion of patients reporting a SAE n (%):

I: 9 (5,9%)

C: 3 (3,8%)

Deaths:  None

Withdrawals due to AE’s: n (%)

I: 8  (4,9%)

C: 1 (1,3%)

Hypoglycaemia⁶ n (%)

I: 37, (22,8%)

C: 13, (16,5%)

cardiovascular events n (%)

I: 2 (1,2%)

C: 0

Mean change from baseline in bodyweight (sd)

I: -0,2 kg (2,3)

C:-0,6 kg (2,2)

between group difference 0,40; 95%CI -0,2 to 1,00; p=0,19

Doses and background treatment maintained in first 12 weeks, after that changes (defined as ≥ 20%) were allowed.

Placebo adjusted mean changes in HbA1c were similar between subgroups of age, duration of diabetes, renal function, and concomitant therapy.

SAE’s were deemed not related to study group.

Higher incidence (not significant) in patients using  linagliptin patients is likely due to the increased incidence of concomitant treatment with SU (30,5% in linagliptin group vs 16,3% in placebo group).

Study reference

(first author, publication year)

Describe method of randomisation¹

Bias due to inadequate  concealment of allocation?²

(unlikely/likely/unclear)

Bias due to inadequate blinding of participants to treatment allocation?³

(unlikely/likely/unclear)

Bias due to inadequate blinding of care providers to treatment allocation?³

(unlikely/likely/unclear)

Bias due to inadequate blinding of outcome assessors to treatment allocation?³

(unlikely/likely/unclear)

Bias due to selective outcome reporting on basis of the results?⁴

(unlikely/likely/unclear)

Bias due to loss to follow-up?⁵

(unlikely/likely/unclear)

Bias due to violation of

intention to treat analysis?⁶

(unlikely/likely/unclear)

Barnett 2013

Randomized 2:1 (intervention vs control)

Stratified by HbA1C level (<8,5% vs. ≥8,5%) and insulin use (yes vs. no) with a block size of four.

Computer-generated randomisation lists produced by sponsor.

Unlikely

Unlikely

4. (Linagliptin and placebo tablets were identical in appearance)

Unlikely

Unlikely

5.         Unclear

6.         

Unlikely

Unlikely

(primary efficacy analysis was done with the full analysis set

(all randomised patients who received at least one dose of

study medication, and who had a baseline and at least one on-treatment HbA1c measurement)

Study reference

Study characteristics

Patient characteristics ²

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size ⁴

Comments

Schernthaner

2015

Type of study:

RCT

Setting:

Multinational

Country:

Austria, Spain, Germany, Norway, Finland, Sweden, Turkey, USA.

Source of funding:

AstraZeneca, BristolMeyers Squibb

Inclusion criteria:

• T2DM
• aged≥65 years
• stable metformin monotherapy at any dose for ≥ 8weeks before enrolment
• HbA1c concentration of 7.0–9.0%

Exclusion criteria:

• T1DM
• Other antihyperglycaemic therapy than metformin monotherapy;
• treatment with systemic glucocorticoids (except for replacement therapy) or cytochrome P450 3A4 inducers
• history of ketoacidosis or hyperosmolar non-ketonic coma
• history of haemoglobinopathies
• renal impairment
• cognitive impairment;
• alcohol or illegal drug abuse;
• hypersensitivity or contraindication to the study drugs.
• AST or ALT levels >3 times the upper limit of normal
• total bilirubin>34 μmol/l
• creatine kinase >10 times the upper limit of normal.

N total at baseline: 720

Intervention: 360

Control: 360

Important prognostic factors²:

age ± SD:

I: 72.5 (5.7)

C: 72.7 (5.4)

Sex:

I: 60.3% M

C: 63.3 % M

Mean HbA1C % (sd)

I: 7.58 (0.67)

C: 7.62 (0.65)

Groups comparable at baseline?

Yes

saxagliptin

5mg/day and placebo

Added to metformin

2-week screening period, a 2-week

enrolment period, a 2-week single-blinded (to patients only)

placebo lead-in period, and a 52-week double-blinded treatment

period, which in turn comprised a 12-week titration

period and a 40-week maintenance period

glimepiride 1mg/day and placebo

Added to metformin

2-week screening period, a 2-week

enrolment period, a 2-week single-blinded (to patients only)

placebo lead-in period, and a 52-week double-blinded treatment

period, which in turn comprised a 12-week titration

period and a 40-week maintenance period

Length of follow-up:

52 weeks

Incomplete outcome data:

Intervention: n =71 (19.7%)

Reasons:

Voluntary discontinuation (n = 17)

Safety reasons (n = 1)

Protocol non-compliance (n = 1)

Incorrect enrolment (n = 0)

Lost to follow-up (n = 0)

Study-specific criteria (n = 33)

AE (n = 15)

Death (n = 1)

Other (n = 3)

Control: n = 75 (20.8%)

Reasons

Voluntary discontinuation (n = 19)

Safety reasons (n = 1)

Protocol non-compliance (n = 3)

Incorrect enrolment (n = 2)

Lost to follow-up (n = 1)

Study-specific criteria (n = 34)

AE (n = 7)

Death (n = 1)

Other (n = 7)

Outcome measures and effect size (include 95%CI and p-value if available):

The primary efficacy endpoint % patients achieving an HbA1c level of<7.0% at week 52 without confirmed/severe hypoglycaemia:

I: 136 (37.9%)

C: 137 (38.2%)

OR saxagliptine vs glimepiride  0.99 (0.73, 1.34); =0.9415

% of patients with ≥ 1 confirmed/severe hypoglycaemic event:

I: 4 (1.1%)

C: 55 (15.3%)

OR saxagliptine vs glimepiride : 0.06 (0.02, 0.17); p<0.0001

% patients achieving HbA1c <7.0 at week 52

I: 159 (44.7%)

C: 193 (54.7%)

OR saxagliptine vs glimepiride 0.67 (0.50, 0.90); p=0.0077

Adjusted mean change from baseline to week 52 in mean HbA1c % (95% CI)

I: −0.44 (−0.51, −0.37)

C: −0.64 (−0.71, −0.57)

OR saxagliptine vs glimepiride 0.20 (0.10, 0.30); p<0.0001

Adverse events, excluding hypoglycaemia n (%)

I: 213 (59.3)

C: 213 (59.3)

≥ 1 serious adverse events

I: 41 (11.4)

C: 32 (8.9)

RR 1.28 (0.83, 1.99) P=0.27

Mean reduction in bodyweight (week 52)

I: -0,8 kg (95% CI -1,2 to 0,4)

C: + 1,0 kg (95% CI 0,6 to 1,3)

Saxagliptin was equally effective as glimepiride regarding the primary efficicay outcome: % patients achieving an HbA1c level of<7.0% at week 52 without confirmed/severe hypoglycaemia.

There were significant less events of conferemed/severe hypoglycaemia in patients treated with saxagliptin.

More patients treated with glimepiride achieve a Hba1c < 7.0 at week 52.  

Study reference

(first author, publication year)

Describe method of randomisation¹

Bias due to inadequate  concealment of allocation?²

(unlikely/likely/unclear)

Bias due to inadequate blinding of participants to treatment allocation?³

(unlikely/likely/unclear)

Bias due to inadequate blinding of care providers to treatment allocation?³

(unlikely/likely/unclear)

Bias due to inadequate blinding of outcome assessors to treatment allocation?³

(unlikely/likely/unclear)

Bias due to selective outcome reporting on basis of the results?⁴

(unlikely/likely/unclear)

Bias due to loss to follow-up?⁵

(unlikely/likely/unclear)

Bias due to violation of

intention to treat analysis?⁶

(unlikely/likely/unclear)

Schernthaner 2015

Interactive web response system, stratified by age to include ~ 60% of patients aged < 75 yr and 40% ≥75 yr.

Unlikely

7. (Adequate randomisation using a interactive web response system)

unlikely

Unlikely

unlikely

Unlikely

Unlikely

8. (Loss to follow-up is adequately described and equal for placebo and intervention group)

Unlikely

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Chien et al. 2011

24-week, randomized, open-labeled, parallel-group study

Setting:

Country: Taiwan

Source of funding: not reported

Inclusion criteria:

- either gender

- older than 65 years;

- T2DM

-taking OAD combinations (sulfonylurea, metformin, and alphaglucosidase inhibitors) but had inadequate response as determined by HbA1c ≥7%

Exclusion criteria:

- any significant disease thet were likely to affect the outcome and compliance of this study à

- cardiovascular disease

-Active liver disease

-renal insufficiency (creatinine level of 1,5 mg/dl or higher

- use of insulin pump

- patients who had to receive any educational program for diabetes management  form elsewhere

N total at randomized: Intervention: 49

Control: 48

Characteristics randomized population: 

Intervention (SD or %)

Mean age:

 I: 73.5 (5.9) years

C: 72.65(5.2) years

Female (%)

I: 63.3

C: 52.1

BMI:

I: 26.2 (2.2) kg/m²

C: 26.0 (2.3) kg/m²

Mean HbA1c:

I: 9.5 (1.3) %

C: 10.0 (1.6)%

Groups were generally well-balanced with respect to baseline demographics and disease characteristics

Continue existing OAD combinations add-on with Sitaglipin (100 mg daily)

Continue existing OAD combinations (sulfonylurea, metformin, and alphaglucosidase inhibitors)

24 weeks of treatment

Discontinued:

Intervention:

N = 5

Reasons (n)

URI (1)

UTI (1)

Nasopharungitis (1)

Sinusitis (0)

Low back pain (1)

Hypolycemia (1)

Diarrhea (0)

Control:

N = 3

Reasons (n)

URI (1)

UTI (0)

Nasopharungitis (0)

Sinusitis (1)

Low back pain (0)

Hypolycemia (0)

Diarrhea (1)

HbA1c  LS- Mean change from baseline (%)

Baseline vs. week 12

I: -1.0 (-1.3, -0.8)

C:-0.0 (-0.2, 0.3)

Baseline vs. week 24

I: -1.3 (-1.6, -0.9)

C: 0.1 (-0.3, 0.5)

Achievement of treatment goal <7% HbA1c (%)

Baseline

I: 0 (0.0)

C: 0 (0.0)

Week 24

I: 13 (26.5)

C: 0 (0.0)

BMI (kg/m²)

Baseline (mean (95% CI))

I: 26.3 (25.7, 27.0)

C: 26.1  (25.4, 26.7)

Week 12 (mean (95% CI))

I: 26.1 (25.6, 26.6)

C: 26.1 (25.6, 26.6)

Week 24 (mean (95% CI))

I: 26.2 (25.8, 26.6)

C: 26.0 (25.6, 26.5)

NS

Clinical AE reported

Intervention: N= 5

Control: n= 3

Number serious AEs

 Intervention: N= 0

Control: n= 0

110 patients screened à 13 patients were not randomized because of failure to meet glycemic inclusion criteria at baseline

Barzilai et al. 2011

Randomized, double-blind, placebo-controlled trail

Setting: community- dwelling

Country: USA

Source of funding: sponsored by MercK & CO., Inc. USAà

Inclusion criteria:

-either gender

- community- dwelling

-at least 65 years of age

-type 2 diabetes that was inadequately controlled by diet and exercise alone (HbA1c 7.0-10.0%)

-patients on AHA monotherapy or low-dose combination AHA* therapy at time of screening

Exclusion criteria:

-previous treated with insulin or exenatide within 8 weeks prior to screening

- T1DM

-Active liver disease

-recent change in cardiovascular status

- an eCrCl <35 mL/min

N total at randomized: 206

Intervention: 102

Control: 104

Completed after discontinued:

Intervention: 70

Control: 57

Population analysed for treatment effect on HbA1c:

Intervention: 101

Control: 91

Population analysed for safety and tolerability:

Intervention:102

Control:104

Characteristics randomized population: 

Intervention (n= 102 (± SD))

Mean age : 71.6 (6.1) years

BMI: 30.8 (5.9) kg/m²

Mean HbA1c: 7.8 (0.6)%

Control (n= 104 (± SD))

Mean age : 72.1 (6.0) years

BMI: 31.1 (7.2) kg/m²

Mean HbA1c: 7.8 (0.7)%

Baseline demographic, anthropometric, and disease characteristics were similar between the treatment groups.

Sitaglipin 100mg q.d. or 50 mg q.d.

Placebo

24 weeks of treatment

Discontinued:

Intervention:

N = 32

Reasons (n)

Patients withdrawal consent (12)

Lack of Efficacy (11)

Clinical AE (5)

Protocol deviolation (2)

Other (2)

Control:

N = 47

Reasons (n)

Patients withdrawal consent (21)

Lack of Efficacy (18)

Clinical AE (4)

Protocol deviolation (3)

Other (1)

Glycaemic endpoints at week 24 HbA1c, % 

Baseline (mean ±SD)

I: 7.8±0.7

C: 7.7±0.8

Change from baseline (LS-mean (95%CI))

I: -0.5 (-0.7—0.2)

C: 0.2 (0.0-0.5)

Between group difference (95%): -0.7% (-0.9--0.5)

P<0.001

HbA1c, %, subgroup with baseline HbA1c <8.0% 

Baseline (mean ±SD)

I: 7.4±0.4

C: 7.4±0.3

Change from baseline (LS-mean (95%CI))

I: -0.4 (-0.6 - -0.1)

C: 0.2 (-0.1-0.4)

Between group difference (95%): -0.5% (-0.8—0.3)

P= 0.043

HbA1c, %, subgroup with baseline HbA1c ≥8.0% to <9.0% 

Baseline (mean ±SD)

I: 8.4±0.2

C: 8.4±0.3

Change from baseline (LS-mean (95%CI))

I: -0.6 (-1.0 - -0.1)

C: 0.4 (0.0-0.7)

Between group difference (95%): -0.9% (-1.4--0.4)

P= 0.043

HbA1c, %, subgroup with baseline HbA1c ≥9.0% 

Baseline (mean ±SD)

I: 9.4±0.3

C: 9.1±0.3

Change from baseline (LS-mean (95%CI))

I: -0.8 (-1.3 - -0.4)

C: 0.8 (0.1- 1.5)

Between group difference (95%): -1.6% (-1.4--0.4)

P= 0.043

At week 24 proportion of patients (%) <7.5% HbA1c

I: 35/110 (35)

C: 14/91 (15) p<0.001

Clinical adverse events proportion of patients (%)

I: 46.1

C: 52.9

Drug-related clinical adverse events proportion of patients (%)

I: 10.8

C: 8.7

Serious adverse events proportion of patients (%)

I: 6.9

C: 13.5

No AEs of hypogycemia

Body weight (LS-mean change from baseline)

I: -1.1 kg

C: -1.7kg p=0.299

*AHA therapy = antihyperglycemic agent

Hartley et al. 2015

Multinational Randomized, parallel-group double-blind, non-inferiority study

Setting: community- dwelling

Country: USA

Source of funding: provided by MercK & CO., Inc.

Inclusion criteria:

-≥65 and ≤85 years of age

- T2DM that was inadequately controlled with diet and exercise alone (HbA1C ≥ 7.0% and ≤ 9.0%)

Exclusion criteria:

-previous treated with a DPP-4 inhibitor, insulin od GLP-1 mimetics within 8 weeks prior to screening or

-treated with PPAR-ƴ agonists within 16 weeks prior screening

-T1DM

-Active liver disease

-recent history of cardiovascular disease

-inadequately controlled hypertension

-severe peripheral vascular disease, triglycerides >600 mg/dL,

-History of infection with HIV,

-hirtory of malignancy or clinically important hematologic disorder

-an eGFR <35 mL/min/1.73m²

N total randomized: 480

Intervention: 241

Control: 239

Completed after discontinued:

Intervention: 204

Control: 200

Characteristics per-protocol:

Intervention (n=197 (± SD))

Mean age: 70.6 (4.8) years

BMI: 29.7 (4.0) kg/m²

Mean HbA1c: 7.8 (0.7) %

Control (n=191 (± SD))

Mean age: 70.8 (4.9) years

BMI: 29.7 (5.1) kg/m²

Mean HbA1c: 7.8 (0.7) %

Groups were well-balanced for baseline demographics and efficacy variables

Sitagliptin 100mg q.d. or 50 mg q.d.

Glimepiride (starting dose 1mg, up-titrate as needed up to max. dose of 6 mg q.d. through week 18

30 weeks of treatment

Incomplete outcome data:

Intervention:

N 37 (15.4%)

Reasons (n)

Adverse events (3)

Lack of Efficacy (7)

Lost to follow-up (1)

Protocol Violation (1)

Withdrawal by Subject (15)

Non-compliance (0)

Physician Decision (5)

Hyperglycemia Discontiniation Criteria (5)

Control:

N 39 (16.3%)

Reasons (n)

Adverse events (4)

Lack of Efficacy (2)

Lost to follow-up (7)

Protocol Violation (5)

Withdrawal by Subject (11)

Non-compliance (1)

Physician Decision (4)

Hyperglycemia Discontiniation Criteria (5)

PP population

HbA1C

Least squares mean change form baseline ta week 30:

I: - 0.32%

C: - 0.51%

Between group difference (95%): 0.19% (0.03-0.34)*

At week 30; proportion of patients (%) <7.5% HbA1c

I: 61.9

C: 64.9

<7% HbA1c

I: 33.5

C: 46.6

<6.5% HbA1c

I: 9.1

C: 20.9%

Body weight

Least squares mean change form baseline ta week 30:

I: 0.4kg

C: 1.1kg

Between group difference (95%): -0.7kg (p = 0.011)

Randomized population

Intervention (n=241)

Control (n=236)

Once or more clinical AEs

I: 118 (49.0%)

C: 115 (48.7%)

Drug-related clinical AEs

I: 9 (3.7%)

C: 9 (3.8%)

Serious clinical AEs

I: 7 (2.9%)

C: 6 (2.5%)

Serious drug-related clinical AEs

I: 0 (0%)

C: 0 (0%)

Hypoglycaemia AEs

Symptomatic or asymptomatic

I: 3 (1.2%)

C: 15 (6.4%)

Symptomatic

I: 2 (0.8%)

C: 11 (4.7%)

Severe

I: 1 (0.4%)

C: 3 (1.3%)

*Non-inferiority of sitaglipin to glimepiride was to be declared if the upper boundary of the two-sided confidence interval for difference in means (intervention - control) for HbA1c was less than the non-inferiority margin, ᵷ = 0.4 for the analysis in the PP population.

Study reference

(first author, publication year)

Describe method of randomisation¹

Bias due to inadequate  concealment of allocation?²

(unlikely/likely/unclear)

Bias due to inadequate blinding of participants to treatment allocation?³

(unlikely/likely/unclear)

Bias due to inadequate blinding of care providers to treatment allocation?³

(unlikely/likely/unclear)

Bias due to inadequate blinding of outcome assessors to treatment allocation?³

(unlikely/likely/unclear)

Bias due to selective outcome reporting on basis of the results?⁴

(unlikely/likely/unclear)

Bias due to loss to follow-up?⁵

(unlikely/likely/unclear)

Bias due to violation of

intention to treat analysis?⁶

(unlikely/likely/unclear)

Chien et al. 2011

Randomization not described

Unclear

Unclear (open-labeled study, endpoints of this study were determines through laboratory tests)

Unclear (open-labeled study, endpoints of this study were determines through laboratory tests)

Unclear (open-labeled study, endpoints of this study were determines through laboratory tests)

Unlikely

Unlikely

Unlikely

Barzilai et al. 2011

Using a computer-generated schedule.

Randomization was stratified by age and  eCrCl

Unlikely

Unlikely (patients, study personnel, and Merck personnel remain blinded to individual treatment allocations until data collection was complete and the final database locked.)

Unlikely (patients, study personnel, and Merck personnel remain blinded to individual treatment allocations until data collection was complete and the final database locked.)

Unlikely (patients, study personnel, and Merck personnel remain blinded to individual treatment allocations until data collection was complete and the final database locked.)

Unlikely

Likely

Unlikely

Hartley et al. 2015

Using a computer-generated allocation schedule and an Interactive Voice Response System (IVRS)

Unlikely

Unlikely

(Single-blind placebo run-in)

Unclear

Unclear

Likely (Randomized n=480; I: n=241; C: n=239 à Characteristics of the per protocol patient population I: n=197; C: n=191)

Unlikely

Unlikely

Study reference

Study characteristics

Patient characteristics 

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Schweizer 2009

Type of study:

RCT

Setting:

Multinational

Country:

14 countries in Europe, the america’s, Asia, not further specified

Source of funding:

Novartis pharmaceutical corporation

Inclusion criteria:

T2DM

Aged ≥ 65 years

At screening

HbA1C: 7,0-9,0%

FGP < 15 mmol/l

BMI 22-40 kg/m²

No oral glucose lowering agents at least 12 weeks before screening or for more than 3 consecutive months in the past.

Exclusion criteria:

T1DM

acute metabolic diabetic complications within last 6 months

congestive heart failure requiring pharmacological treatment or myocardial infarction, unstable angina, stroke, coronary artery bypass surgery within past 6 months. Liver disease, renal disease .

N total at baseline: 332

Intervention: 167

Control: 165

Important prognostic factors²:

age ± SD:

I: 71,6 ± 5,2

C:70,2 ± 5,1

Sex:

I: 44,4 % M

C: 53,0 % M

HbA1C at baseline ± SD

I: 7,8 ± 0,6

C: 7,7 ± 0,6

BMI (kg/m²) ± SD

I: 29,8 ± 4,4

C: 29,4 ± 0,6

Groups comparable at baseline?

Yes

Vildagliptin 100 mg/ day

HbA1C, FPG and bodyweight measured at weeks 0, 4, 12, 16, 24

Physical examination, vital signs, ECG at week 0 and 24.

No dose adjustments were allowed

Metformin 1500 mg/day

1000mg morning + 500 mg evening

HbA1C, FPG and bodyweight measured at weeks 0, 4, 12, 16, 24

Physical examination, vital signs, ECG at week 0 and 24.

No dose adjustments were allowed

Length of follow-up:

24 weeks

Incomplete outcome data:

Intervention:

N = 27 (16,0%)

Reasons (describe)

AE’s (n=6, 3,6%)

Withdrew consent (n=13, 7,7%)

Death (n=1, 0,6%)

Abnormal laboratory values (n=2, 1,2%)

Lost to follow-up (N=3, 1,8%)

Unsatisfactory therapeutic effect (N=1, 0,6%)

Protocol violation (n=1, 0,6%)

Control:

N = 26 (15,7%)

Reasons (describe)

AE’s (n=13, 7,8%)

Withdrew consent (n=8, 4,8%)

Lost to follow-up (N=3, 1,8%)

Unsatisfactory therapeutic effect (N=2, 1,8%)

Outcome measures and effect size (include 95%CI and p-value if available):

Primary outcome:

Mean change in HbA1c from baseline (s.e.)

I: -0,64% (0,07)

C: -0,75% (0,07)

Between group difference: 0,11 (95% CI: -0,08 to 0,29; p=0,258)

Conclusion: Non-ineriority of vildagliptin to metformin

Other outcomes:

mean reduction in FPG:

I: -0,78 ± 0,12 mmol/L (p<0,0001 vs baseline)

C: - 1,27 ±0,12 mmol/L (p<0,0001 vs baseline).

Vildagliptin vs metformin P = 0,006

Mean change to baseline  in weight (sd)

I: -0,45 kg (0.2)

C:-1,25 kg (0,19)

P=  0,004

Safety:

Adverse events n(%)

I: 74 (44,3%)

C: 83 (50,3%)

RR 0.88 (0.70, 1.11) P=0.28

Serious adverse events n(%):

I: 5 (3,0%)

C: 6 (3,6%)

RR 0.82 (0.26, 2.65) p=0.74

Deaths:  None

Withdrawals due to AE’s: n (%)

I: 8  (4,9%)

C: 1 (1,3%)

RR 7.90 [1.00, 62.50] P=0.05

Hypoglycaemia n (%)

I: 0, (0%)

C: 2 (1,2%)

RR 0.20 (0.01, 4.09) P=0.29

ITT analysis is unclear.

Randomization procedure and allocation concealment not described.

Effect of vildagliptin not inferior to metformin, probably better tolerable.

Schweizer 2013

Type of study:

Post hoc sub- analysis of a RCT

Setting:

Multi-centre

Country:

USA, Switzerland, Finland and others

Source of funding:

Not reported

Inclusion criteria for this sub analysis:

•  Patients ≥ 75 yr

Inclusion criteria original article:

• T2DM
• moderate (eGFR ≥30 to<50 ml/min/1.73 m²) and or severe (eGFR <30 ml/min/1.73 m²) Renal Impairement
• age 18–85 years
• BMI: 18–42 kg/m2
• HbA1c:  6.5–10%  T2DM either untreated or treated with a sulfonylurea (SU), alpha-glucosidase inhibitor (AGI), thiazolidinedione (TZD), insulin, meglitinide or a combination of agents provided that doses were stable for at least 4 weeks before the first study visit.

Exclusion criteria original article:

• FPG ≥15 mmol/l,
• history of renal transplant
• significant cardiovascular history within 6 months
• active liver disease
• abnormal liver tests(ALT, AST or bilirubin 2× upper limit of normal

N total at baseline: 105

Intervention: 50

Control: 55

Important prognostic factors:

age ± SD:

I: 78.0 ± 2.5

C: 78.3 ± 2.7

Sex:

I: 52.0 % M

C: 54.5 % M

HbA1C at baseline ± SD

I: 7.8 ± 1.0

C: 7.9 ± 0.8

FPG mmol/L ± SD

I: 8.8 ± 3.2

C: 8.2 ± 3.1

BMI (kg/m2) ± SD

I: 31.0 ± 4.5

C: 30.0 ± 4.4

Groups comparable at baseline?

Yes

Describe intervention (treatment/procedure/test):

2-week single-blind, placebo

run-in period

Vildagliptin 50 mg qd

Patients also had to

continue their current diet/exercise regimen

and background therapy unchanged

throughout the duration of the study.

Describe  control (treatment/procedure/test):

2-week single-blind, placebo

run-in period

Placebo

Patients also had to

continue their current diet/exercise regimen

and background therapy unchanged

throughout the duration of the study.

Length of follow-up:

24 weeks

Loss-to-follow-up:

not described

Outcome measures and effect size (include 95%CI and p-value if available):

Change in mean HbA1c from baseline:

I: -1.0% ± 0.2%

C: -0.2% ± 0.2%

Between group difference

-0.8 ± 0.2 (P<0.001).

Mean change in FPG

I: -1.1 ± 0.6 mmol/L

C: not reported

Between group difference

-0.9 ± 0.5 mmol/L (P=0.088)

Hypoglycaemia:

I: 11 events in 5 patients

0.49 events per patient-year

C: 23 events in 4 patients

0.96 events per patient-year

P = 0.970.

Severe hypoglycaemia;

I: 1 event

C: 0 events

RR 3.29 [0.14, 79.06] P = 0.46

AE: n (%)

I: 29 (58.0%)

C: 40 (72.7%)

RR 0.80 [0.60, 1.06] P=0.12

SAE’s n (%)

I: 7 (14%)

C 9 (16.4%)

RR 0.86 [0.34, 2.13] P=0.74

Withdrawal due to AE’s, n (%):

I: 2 (4.0%)

C: 5 (9.1%)

RR 0.44 [0.09, 2.17] P = 0.2854

Cardiovascular disorders

I: 2 (4.0%)

C: 10 (18.2)

RR  0.19 [0.04, 0.90] P= 0.04

Comorbidities: hypertension (96.2%) dyslipidemia (68.6%) of study population.

Garber 2008

Type of study:

RCT

Setting:

Multinational

Country:

USA, Finland, Sweden, Finland, Argentina, Lithuania

Source of funding:

Novartis

Inclusion criteria:

• T2DM
• Inadequately controlled on SU monotherapy
• Baseline HbA1c of 7.5–11%.
• Previously treated with an SU for ≥ 3 months with a stable dose for ≥ 4 weeks before the screening visit.
• 18–80 years of age
• BMI of 22–45 kg/m2 and
• FPG < 15 mmol/l

Exclusion criteria:

• history of type 1 or secondary forms of diabetes,  myocardial infarction, unstable angina or coronary artery bypass surgery within the previous 6 months.
• Congestive heart failure (NYHA class III or IV)
• liver disease
• use of any oral antidiabetic drug other than an SU within the past 2 months
• ALAT or ASAT  > 3 times the ULN
• direct bilirubin >1.3 times the ULN
• serum creatinine levels >220 mmol/l
• TSH outside the normal range
• Fasting triglycerides (TGs) >7.9 mmol/l.

N total at baseline: 515

primary ITT: 408

N total ≥ 65 years: 114

I: V 50 mg: 41

I: V 100 mg: 35

C: 38

Groups comparable at baseline?

Unclear

Vildagliptin 1 dd 50 mg

Or

Vildagliptin 2 dd 50 mg

Added to glimepiride 1 dd  4 mg ( could be reduced by protocol to 2 mg in case of hypoglycaemia)

Placebo

Added to glimepiride 1 dd 4 mg ( could be reduced by protocol to 2 mg in case of hypoglycaemia)

Length of follow-up:

24 weeks

Loss-to-follow-up:

Not described for patients ≥ 65 years.

For total study population:

Intervention:

V 50 mg: 19 (14.4%)

Adverse event (n = 4, 3.0%)

Abnormal test result (n = 0)

Unsatisfactory therapeutic

effect (n =6, 4.5%)

Protocol violation (n = 2, 1.5%)

Withdrew consent (n = 4, 3.0%)

Lost to follow-up (n = 2, 1.5%)

Administrative (n = 1, 0.8%)

V100mg:  21 (15.9%)

Adverse event (n = 4, 3.0%

Abnormal test result (n = 1, 0.8%)

Unsatisfactory therapeutic

effect (n = 5, 3.8%)

Protocol violation (n = 0)

Withdrew consent (n = 4, 3.0%)

Lost to follow-up (n = 6, 4.5%)

Administrative (n = 1, 0.8%)

Reasons (describe)

Control: 36 (25.0%)

Adverse event (n = 2, 1.4%)

Abnormal test result (n = 0)

Unsatisfactory therapeutic effect (n=16, 11.1%)

Protocol violation (n=2, 1.4%)

Withdrew consent (n=12, 8.3%)

Lost to follow-up (n=4, 2.8%)

Administrative (n=0)

Incomplete outcome data: not reported

Outcome measures and effect size (include 95%CI and p-value if available):

Adjusted mean change in HbA1C  in patients ≥ 65 years:

I: V 50mg: -0.7 ± 0.1 %

I: V 100 mg: -0,8 ± 0.2% 

C: small increase of HbA1c level shown in figure, no data reported.

Adjusted mean change in HbA1c total study population:

I: V 50mg: -0.58 ± 0.1 %

I: V 100 mg: -0.63 ± 0.09%

C: + 0.07 ± 0.09%

Data for patients <65 years not reported.

Adverse events for patients ≥ 65 years not reported.

Adverse events for total study population

I V50 mg: 114 (67.1%)

I V100 mg: 112 (66,3%)

C: 113 (64,2%)

Hypoglycemia

I V50 mg: 2 (1.2 %)

I V100 mg: 6 (3.6%)

C: 1 (0,6%)

No sign difference in any comparison

Decrease of HbA1C in patients ≥ 65 years of unknown significance.

Strain 2013

24 week, multicentre, randomised, double-blind,

placebo-controlled study

Setting: outpatient centres in seven European countries

(Belgium, Bulgaria, Germany, Finland, Slovakia, Spain,

and UK)

Source of funding: Novartis Pharma AG.

Inclusion criteria:

-aged 70 years or older

-type 2 diabetes who were drug-naïve or inadequately controlled,

- HbA1c levels of ≥7,0% and ≤ 10,0%,

-fasting plasma glucose (FPG) <15 mmol/L

(270 mg/dL),

BMI: 19-45 kg/m2 at

screening (visit 1)

Exclusion criteria:

-use of insulin treatment (>7 consecutive days)

-incretin-based therapies in the preceding 12 weeks, - use of corticosteroids within 8 weeks,

-use of growth hormone

within 6 months of the screening visit

-Patients with acute metabolic diabetic disorders, myocardial infarction, coronary artery bypass surgery, or stroke within 6 months;

-unstable angina within 3 months;

-congestive heart failure

(NYHA III or IV);

-malignancy within 5 years;

-liver disease

-Substantial laboratory abnormalities including liver function tests, renal dysfunction

Assessed for eligibility n=434

assigned randomly n= 278

Allocated to vildagliptin n=139

Allocated to placebo: 139

Analysis populations

Intervention:

137 intention to treat

128 per protocol

139 safety

Control:

137 intention to treat

130 per protocol

139 safety

Age (years (SD))

Intervention: 75,1 (4,3) Control: 74,4 (4,0)

HbA1c (% (SD))

Intervention: 7,9 (0,8):

Control: 7,9 (0,7)

Body-mass index (kg/m2) Intervention: 29,1 (3,8)

Control: 30,5 (4,8)

Frailty status

Intervention:

Yes: 12 (8,6%)

No: 126 (90,6%)

Missing: 1 (0,7%)

Control:

Yes: 14 (10,1%)

No: 123 (88,5%)

Missing: 2 (1,4%)

Both groups were similar for all patient demographic

and background characteristics at baseline.

Vildagliptin

Patients in the drug-naive and other background OAD

Groups: vildagliptin (50 mg) twice daily

patients in the sulphonylurea monotherapy

group: vildagliptin (50 mg) once daily

Metformin and sulphony lureas were the most frequently

used OADs; of the randomly assigned patients, nearly

half were taking metformin, whereas a third were using

sulphonylureas.

Placebo

Patients in the drug-naive and other background OAD

Groups: placebo twice daily

patients in the sulphonylurea monotherapy

group: placebo (50 mg) once daily

Metformin and sulphony lureas were the most frequently

used OADs; of the randomly assigned patients, nearly

half were taking metformin, whereas a third were using

sulphonylureas.

Length of follow-up:

24 weeks

8 discontinued intervention:

5 adverse event(s)

0 unsatisfactory therapeutic effect

0 withdrew consent

1 lost to follow-up

1 administrative problems

1 death

8 discontinued control:

2 adverse event(s)

2 unsatisfactory therapeutic effect

3 withdrew consent

0 lost to follow-up

0 administrative problems

1 death

mean individualised HbA1c

targets set by the investigators: around 7,0% for both

treatment groups,

 0,9% (range –4,4 to –0,1) lower than the mean baseline HbA1c of 7,9% in each treatment group.

Intervention:

72/137 (52,6%) patients achieved their individualised targets as a result of education and interactions with the study team alone;

Placebo:

37/137 (27%) patients

The adjusted OR of achieving the individualised target was 3,16 (96,2%^ CI 1,81–5,52; p<0·0001).

Proportion of patients achieving individualised HbA1c targets, according to baseline characteristics

Age (years)

Intervention:

<75 years: 33/67 (49,3%)

≥75 years: 39/70 (55,7%)

Controle

<75 years: 22/86 (25,6%)

≥75 years: 15/51 (29,4%)

HbA1c levels

Intervention:

≤8%: 46/83 (55,4%)

>8%: 26/54 (48,1%)

≤9%: 65/125 (52,0%)

>9% :7/12 (58,3%)

Control:

≤8%: 27/86 (31,4%)

>8%: 10/51 (19,6%)

≤9% 34/128 (26,6%)

>9% 3/9 (33,3%)

Frailty status

Intervention:

Frail: 5/12 (41,7%)

Non-frail: 67/124 (54,0%)

Control: Frail: 6/14 (42,9%)

Non-frail: 30/121 (24,8%)

The risk ratios for the number of patients who reached the investigator-defined HbA1c target at study endpoint

1,92 (96,2% CI 1,29–2,86; p=0,0013) in the intention-to-treat analysis

1,91 (1,27–2,86; p=0,0017) in the per-protocol analysis.

AEs

Overall

I:66 (47,5%) C: 63 (45,3%)

Serious AEs

I: 8 (5,8%) C: 5 (3,6%)

Discontinuations due to AEs

I: 6 (4,3%) C: 3 (2,2%)

Deaths

I: 1 (0,7%) C: 1 (0,7%)

The incidence of hypoglycaemic events

I: 3/139 (2,2%) of 139 patients C: 1/139 (0,7%)

All hypoglycaemic

events occurred in patients using concomitant

sulphonylureas.

No severe hypoglycaemic^# events were reported in either group.

Mean body weight (n=138 in each group)

I: 79,9 kg at baseline à 79·8 kg at study endpoint

C: 80,7 kg at baseline and 79,7 kg at study endpoint

Study reference

(first author, publication year)

Describe method of randomisation¹

Bias due to inadequate  concealment of allocation?²

(unlikely/likely/unclear)

Bias due to inadequate blinding of participants to treatment allocation?³

(unlikely/likely/unclear)

Bias due to inadequate blinding of care providers to treatment allocation?³

(unlikely/likely/unclear)

Bias due to inadequate blinding of outcome assessors to treatment allocation?³

(unlikely/likely/unclear)

Bias due to selective outcome reporting on basis of the results?⁴

(unlikely/likely/unclear)

Bias due to loss to follow-up?⁵

(unlikely/likely/unclear)

Bias due to violation of

intention to treat analysis?⁶

(unlikely/likely/unclear)

Schweizer 2009

Not reported.

Unclear

Likely

9. (Linagliptin is taken once daily, metformin twice daily)

Unlikely

Unlikely

Unlikely

Unlikely

Unlikely

Schweizer 2013

Not reported

Unclear

Unclear

Unclear

Unlikely

10. (Outcomes are independent of the assessor)

Unclear

Unclear

Unclear

Garber 2008

Automatically by randomization system

Unlikely

Unclear

Unclear

Unclear

Unlikely

Likely

(High loss to follow up)

Unlikely

Strain 2013

The randomisation list was produced by an interactive response technology provider.

Unlikely

Unlikely

(Patients, investigators, people doing the asses-sments, and data

analysts were masked to the treatments from randomisation to database lock).

Unlikely

(Patients, investigators, people doing the asses-sments, and data

analysts were masked to the treatments from randomisation to database lock).

Unlikely

(Patients, investigators, people doing the asses-sments, and data

analysts were masked to the treatments from randomisation to database lock).

Unlikely

Unlikely

Unlikely

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Pawaskar et al. 2012

Type of study:

Retrospective observational study

Setting:

Ambulatory care setting 

Country:

USA

Source of funding:

Eli Lilly and Company and Amylin Pharmaceuticals, Inc.

Inclusion criteria:

-Age ≥ 65 years

-Received prescription orders for exenatide or glargine between November 1, 2006 and April 30, 2009

-Patients needed to have T2DM before prescription medication.

-At least one of the criteria satisfied following the algorithm in Brixner et al. 2009

Exclusion criteria:

- prescription of GLP-1, or insulin in any form in the pre-index period.

-If they had a diagnose for T1DM, gestational diabetes or were pregnant

N total at baseline:

Intervention: 1023

Control: 2238

After matching:

Intervention: 804

Control: 804

Important baseline characteristics (after matching):

age ± SD:

I: 71 ± 4,7

C:71 ± 4,8

Sex:

I: 44,9% M

C: 45,2% M

Baseline HbA1C % ± SD:

I: 7,6 % ± 1,3

C: 7,6 % ± 1,3

Weight ± SD:

I: 98,0 ± 18,9 kg

C: 98,6 ± 21,1 kg

BMI ± SD:

I: 35,2 ± 6,3 kg/m²

C: 34,9 ± 6,6 kg/m²

SBP ± SD:

I: 130,9 ± 16,3 mmHg

C: 131,3 ± 17,3 mmHg

DBP± SD:

I: 72,7 ± 9,7 mmHg

C: 72,9 ± 9,9 mmHg

Groups comparable at baseline?

Yes

Exenatide

Dose: not reported

Insulin glargine

Dose: not reported

Length of follow-up:

12 month 

Incomplete outcome data:

Because A1C was the primary outcome in this study, the study only included patients who had both baseline and follow-up A1C.

Primary endpoint:

HbA1C mean changes from baseline to follow-up

I: - 0,5% 

C: -0,2

P< 0,05

Mean HbA1C at follow-up:

I: 7,1% 

C: 7,3%

P< 0,05

Secondary endpoints:

Mean body weight changes from baseline to follow-up

I: - 2,8 kg 

C: -0,2 kg

P< 0,05

Mean weight at follow-up:

I: 95,2 kg 

C: 98,4 kg

P< 0,05

BMI changes from baseline to follow-up

I: - 1,0 kg/m²

C: -0,1 kg/m²

P< 0,05

SBP changes from baseline to follow-up

I: - 2,2 mmHg

C: - 1,0 mmHg

P< 0,05

No significant differences for  DBP

Proportion patients reached the HbA1C goal <7%

I: 53%

C: 43%

P< 0,01

Proportion patients reached the MBI goal <30 kg/m²

I: 26,7%

C: 21,1%

P= 0,02

Proportion patients reached the SBP goal <130 mmHg

I: 52,1%

C: 45,8%

Retrospective observational studies have missing values for some clinical outcomes, which leads to excluding matched patients without paired data from the study.

Cohorts were not randomized, however, propensity-score matching was used to control for selection bias among the observed variables. This does not control for unmeasured confounding bias

Study reference

Risk of Bias due to confounding

Risk of Bias in selection of participants into the study

Risk of Bias in measurement of exposure

Risk of Bias due to departures from intended exposure(s)

Risk of Bias due to missing data

Risk of Bias in measurement of outcomes

Risk of Bias in selection of the reported result

Pawaskar et al. 2012

Moderate

Cohorts were not randomized, however, propensity-score matching was used to control for selection bias among the observed variables. This does not control for unmeasured confounding bias

Low

Cohorts were not randomized, however, propensity-score matching was used to control for selection bias among the observed variables.

Moderate

Patients were included in the analysis if they received prescription orders for exenatide or glargine between November 1, 2006 and April 30, 2009

Moderate

Propensity-score matching was used to control for selection bias among the observed variables. This does not control for unmeasured confounding bias

Moderate

Because A1C was the primary outcome in this study, the study only included patients who had both baseline and follow-up A1C.

no information

serious

only positive result was reported

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Bode, 2013

Type of study:

Randomized double-blind placebo controlled phase 3 study

Setting:

Country:

17 countries

Source of funding:

Study was sponsored by Janssen Research and Development LLC.

Inclusion criteria:

Age 55-80 y

HbA1c 7.0-10.0%

BMI 20-40 kg/m2

Fasting plasma glucose <15.0 mmol/l

Fasting blood glucose 6.1 – 15.0 mmol/l

Exclusion criteria:

History of T1DM

Repeated FPG >= 15 mmol/l

History of myocardial infarction, unstable angina, revascularization procedure or CVA <3 months, history of cardiac disease, uncontrolled hypertension, eGFR <50ml/min/1.73m2.

N total at baseline:

Cana 300: 236

Cana 100: 241

Placebo: 237

Important baseline characteristics:

age ± SD:

Cana 300 63.4 ± 6.0

Cana 100: 64.3 ± 6.5

Placebo: 63.2 ± 6.2

Sex:

Cana 300: 54.7% M

Cana 100: 51.5% M

Placebo: 60.3% M

Baseline HbA1C % ± SD:

Cana 300: 7.7 % ± 0.8

Cana 100: 7.8 ± 0.8

Placebo: 7.8 % ± 0.8

Subjects on blood glucose lowering agents:

Cana 300: 233 (98.7%)

Cana 100: 231 (95.9%)

Placebo: 232 (97.9%)

Groups comparable at baseline?

Yes

2-week placebo run-in period

Canaglifozin 100 mg

Or

Canaglifozin 300 mg

add-on

2-week placebo run-in period

Placebo

Length of follow-up:

26 weeks

Incomplete outcome data:

82 subjects (11.2%) discontinued before 26 wk

Cana 300: 27

Cana 100: 15

Placebo: 40

Rescue therapy:

Cana 300: 1/236 (0.4%)

Cana 100: 5/241 (2.1%)

Placebo: 26/239 (10.9%)

Primary endpoint:

HbA1C mean changes from baseline to week 26

Cana 300: -0.73% (p<0.001)

Cana 100: -0.60% (p<0.001)

Placebo: -0.03%

(p: vs placebo)

Secondary endpoints:

FPG change from baseline

Cana 300: -1.1 mmol/l (p<0.001)

Cana 100: -1.0 mmol/l (p<0.001)

Placebo: +0.4 mmol/l

(p: vs placebo)

SBP changes from baseline

Cana 300: -6.8 ± 1.1 mmHg (p<0.001)

Cana 100: -3.5 ± 1.0 mmHg

(p<0.001)

Placebo: +1.1 ± 1.0 mmHg

(p: vs placebo)

Mean body weight changes from baseline

Cana 300: -2.8 kg (p<0.001) 

Cana 100: -2.2 kg (p<0.001)

Placebo: -0.1 kg

(p: vs placebo)

Fasting HDL-C mean % change from baseline (mmol/l)

Cana 300: 6.2 ± 1.2% (p<0.001)

Cana 100: 6.8 ± 1.2% (p<0.001)

Placebo: 1.5 ± 1.2%

(p: vs placebo)

Proportion patients reached the HbA1C goal <7%

Cana 300: 58.5% (p<0.001)

Cana 100: 47.7% (p<0.001)

Placebo: 28.0%

(p: vs placebo)

Fasting triglyceride

Mean % change from baseline (mmol/l)

Cana 300: 8.4 ± 3.4% (ns)

Cana 100: 2.8 ± 3.3% (ns)

Placebo: 7.6 ± 3.4%

(ns: vs placebo)

Additional endpoints:

DBP changes from baseline

Cana 300: -3.2 ± 0.6 mmHg

Cana 100: -1.6 ± 0.6 mmHg

Placebo: +0.1 ± 0.6 mmHg

(not tested for significance)

Hypoglycemia in pt not on BGLA associated with hypo’s:

Cana 300: 3/63 (4.8%)

Cana 100: 4/60 (6.7%)

Placebo: 2/62 (3.2%)

Hypoglycemia in pt on BGLA associated with hypo’s:

Cana 300: 82/173 (47.4%)

Cana 100: 78/181 (43.1%)

Placebo: 66/175 (37.7%)

Efficacy analyses:  Modified ITT-analysis: all randomized subjects who took >=1 dose of double blind study drug.

Missing values: last observation carried forward.

Safety analyses: according to treatment of longest duration.

Prespecified subgroup analyses:

Baseline HbA1c <8, 8-9, >=9

Age <65, >=65

Canaglifozin 100 and 300 mg provided modestly greater reductions in HbA1c-levels from baseline to week 26 in subjects aged <65 years compared with subjects aged >=65 years

Yale, 2013

Type of study:

Randomized double-blind placebo controlled phase 3 study

Setting: 89 centres

Country:

19 countries

Source of funding:

Study was sponsored by Janssen Research and Development LLC.

Inclusion criteria:

age ≥25 years

T2DM

HbA1c ≥7.0 and ≤10.5%

stage 3 chronic kidney disease (eGFR ≥30 and <50

ml/min/1.73 m2),

not on AHA therapy or on stable AHA regimen

stable renal function

Exclusion criteria

repeated fasting

plasma glucose (FPG) >15.0 mmol/l (270 mg/dl) during

pretreatment phase;

history of T1DM;

renal disease that required immunosuppressive therapy, dialysis or transplant;

nephrotic syndrome or inflammatory renal disease; cardiovascular disease; cardiovascular event <3 months prior

to screening; Hb <100 g/l (10 g/dl) at screening.

N total at baseline:

Cana 300: 89

Cana 100: 90

Placebo: 90

Important baseline characteristics:

age ± SD:

Cana 300 67.9 ± 8.2

Cana 100: 69.5 ± 8.2

Placebo: 68.2 ± 8.4

Sex:

Cana 300: 53.9% M

Cana 100: 64.4% M

Placebo: 63.3% M

Baseline HbA1C % ± SD:

Cana 300: 8.0% ± 0.8

Cana 100: 7.9% ± 0.9

Placebo: 8.0% ± 0.9

Groups comparable at baseline?

Yes

AHA adjustment period,

2-week, single blind,

placebo run-in period;

26-week, double-blind, core treatment period;  26-week, double-blind, extension period (data to be reported in a separate publication

Core treatment:

Canaglifozin 100 mg  or

Canaglifozin 300 mg

add-on

AHA adjustment period,

2-week, single blind,

placebo run-in period;

26-week, double-blind, core treatment period;  26-week, double-blind, extension period (data to be reported in a separate publication

Core treatment:

Placebo

Length of follow-up:

26 weeks

Incomplete outcome data:

35 subjects (12.9%) discontinued before 26 wk

Cana 300: 7

Cana 100: 15

Placebo: 13

Rescue therapy:

Cana 300: 3.3%

Cana 100: 4.4%

Placebo: 14.3%

Primary endpoint

HbA1C: LS mean changes from baseline to week 26 relative to placebo

Cana 300: -0.4% (p<0.001)

Cana 100: -0.3% (p<0.05)

Secondary efficacy endpoints

% of subjects reaching HbA1c <7.0% at week 26

Cana 300: 32.6%

Cana 100: 27.3%

Placebo: 17.2%

FPG: Change from baseline: differences in LS mean changes (95% CI) from placebo

Cana 300: –0.67 (–1.4, 0.1)  Cana 100:  –0.85 (–1.6, –0.1)

(mmol/l)

Other efficacy endpoints

Blood pressure

change from baseline:

differences in LS mean changes (95% CI) from placebo

Systolic

Cana 300: –6.1 (–10.0, –2.3)

Cana 100:  –5.7 (–9.5, -1.9) 

Diastolic

Cana 300: –2.1 (–4.3, 0.2)

Cana 100: –1.2 (–3.4, 1.0)

Fasting plasma lipids

LS mean percent changes

HDL-C

Cana 300: 3.0%

Cana 100: 4.0%

Placebo: 1.5%

Triglycerides

Cana 300: 11.9%

Cana 100: 6.2%

Placebo: 7.9%

LDL-C

Cana 300: -1.0%

Cana 100: 6.4%

Placebo: 6.3%

Body weight

Differences in LS mean percent changes (95% CI) relative to placebo

Cana 300: –1.8% (–2.6, –1.0)

Cana 100: –1.6% (–2.3, –0.8)

Adverse events

Any:

Cana 300: 66 (74.2%)

Cana 100: 71 (78.9%)

Placebo: 67 (74.4%)

Drug related:

Cana 300: 29 (32.6%)

Cana 100: 23 (25.6%)

Placebo: 20 (22.2%)

SAE

Cana 300: 10 (11.2%)

Cana 100: 10 (11.2%)

Placebo: 16 (17.8%)

Urinary tract infection

Cana 300: 7 (7.9%)

Cana 100: 5 (5.6%)

Placebo: 5 (5.6%)

Genital mycotic infection

Cana 300: 2 (2.2%)

Cana 100: 2 (2.2%)

Placebo: 0

Severe hypoglycaemia

Cana 300: 1 (1.2%)

Cana 100: 4 (4.7%)

Placebo: 1 (1.1%)

Study reference

Describe method of randomisation

Bias due to inadequate  concealment of allocation?

Bias due to inadequate blinding of participants to treatment allocation

Bias due to inadequate blinding of care providers to treatment allocation?

Bias due to inadequate blinding of outcome assessors to treatment allocation?

Bias due to selective outcome reporting on basis of the results?

Bias due to loss to follow-up?

Bias due to violation of

intention to treat analysis?

Bode 2013

Randomization was performed using an interactive voice response system / interactive web response system with a computer generated randomization schedule prepared by the sponsor before the study.

Unlikely

Unlikely

Subjects were to remain blinded to treatment assignment until the final database lock.

Unlikely

unlikely

Investigators were to remain blinded to treatment assignment until the final database lock.

Unlikely

Unlikely

12. (Loss to follow-up is adequately described and more in placebo than in intervention groups)

Unlikely

Yale 2013

Subjects were randomized using an Interactive Voice Response System/ Interactive Web Response System. Randomization was balanced by using

permuted blocks of six subjects per block and stratified based on (i) the presence or absence of atherosclerotic cardiovascular

disease and (ii) whether a subject required an AHA

adjustment period prior to randomization.

Unlikely

Endpoints were objective parameters: HbA1c, FPG,

Unlikely

After randomization, HbA1c and FPG

values were masked to the study centres

After randomization, HbA1c and FPG

values were masked to the study centres

Unlikely

Unlikely

(Loss to follow-up is adequately described and more in placebo than in intervention groups)

Unlikely

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Leiter 2014

Multicenter randomized, double-blind, age-stratified (<65 and ≥65), 24-week clinical trial with a 28-week extension

173 centres in 10 countries

Inclusion: T2DM, glycosylated hemoglobin (HbA1c) of 7.0% to 10.0%, and documented CVD.

Exclusion:

Type 1 diabetes mellitus; use of rosiglitazone or three or more oral antihyperglycemic drugs; symptoms of poorly controlled diabetes; cardiovascular events within 2 months of enrollment; congestive heart failure; systolic BP ≥160 mmHg and/or diastolic BP ≥100 mmHg at randomization; renal failure.

1489 pts were enrolled,

964 qualified after enrollment and lead-in-period and were randomized.

Participants were stratified according to age at enrollment (<65, ≥65), insulin use at randomization (no, yes), and time from most recent qualifying cardiovascular event (>1, ≤1 year).

Demographic and baseline characteristics were well balanced between treatments and age strata

Dapagliflozine 1 dd 10 mg in addition to background treatment

N=480

Placebo in addition to background treatment

N=482

Primary: 24 weeks, extension period to 52 weeks

Primary outcome (24 weeks):

Mean change in HbA1c from baseline

placebo-adjusted mean change:

total population:

I: -0.3 (-0.4 to -0.3) n=474

C: 0.1 (0.0 to 0.2) n=471, P<0.001

>65 years:

I: -0.3 (-0.4 to -0.2) n=225

C: 0.1 (-0.1 to 0.2) n=224, P<0.001

proportion of participants achieving a three item outcome measure of combined clinical benefit: simultaneous HbA1c decrease of 0.5% or greater, total BW reduction of 3% or greater, and systolic BP (SBP) reduction of 3 mmHg or more from baseline

total population

I: 47/468 = 1.9% (0.7 to 3.2)

C: 9/469 = 10.0% (7.3 to 12.8), P<0.001

>65 years:

I: 20/220 = 9.1% (5.3 to 12.9)

C: 5/220 = 2.3% (0.3 to 4.2), P<0.01

Secondary endpoints

Body weight

Adjusted mean change from baseline

total population

I: -2.5 (-2.9 to -2.2) n=480

C: -0.6 (-1.0 to -0.3) n=481, P<0.001

>65 years:

I: -2.6 (-3.1 to -2.1) n=227

C: -0.7 (-1.2 to -0.2) n=224, P<0.001

Post hoc analyses were conducted in participants aged 75 and older for the efficacy variables of HbA1c, SBP, and BW and for safety measures

4 authors were employees of Astra Zeneca and contributed to study conception and design, analysis and interpretation of data, writing and revision of manuscript.

Two of these contributed to study supervision, data acquisition.

“Sponsor’s Role: The sponsors had no part in the study and design.”

Study reference

Describe method of randomisation¹

Bias due to inadequate  concealment of allocation?

Bias due to inadequate blinding of participants to treatment allocation?

Bias due to inadequate blinding of care providers to treatment allocation?

Bias due to inadequate blinding of outcome assessors to treatment allocation?

Bias due to selective outcome reporting on basis of the results?

Bias due to loss to follow-up?

Bias due to violation of

intention to treat analysis?

Leiter 2014

Patients were assigned a unique enrolment number at visit1.

Patients were randomized using randomly permuted blocks with a block size of four. Patients were stratified into one of the eight age-by-insulin use-by-time from most recent qualifying cardiovascular event strata according to age group, use of insulin, and time from most recent qualifying cardiovascular event.

Unlikely

Unlikely

(Patients were given medication with the bottle number allocated by the IWRS/IVRS.)

Unclear (not described)

Unclear (not described)

Unlikely

13. Unlikely
14.  

Unlikely

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Zinman, 2015

Randomized double-blind, placebo-controlled trial

Source of funding:

Study was sponsored by Boehringer Ingelheim (steering committee) and Eli Lilly (cofounding).

Inclusion criteria

type 2 diabetes

≥18 years of age

body-mass index <45

eGFR ≥30 ml/min/1.73 m2

established cardiovascular disease

no glucose-lowering agents for ≥12 weeks and HbA1c ≥7.0% and ≤ 9.0%, or

stable glucose-lowering therapy for ≥12 weeks and HbA1c ≥7.0% and ≤ 10.0%.

7028 patients were randomized

7020 were treated and included

in the primary analysis

2-week, open-label,

placebo run-in period

empagliflozin 10 or 25 mg

(pooled analysis)

2-week, open-label,

placebo run-in period

placebo

Median duration of

treatment was 2.6 years, and the median observation

time was 3.1 years.

97.0% completed the study.

25.4% of  patients prematurely discontinued a study drug.

Final vital status was available

for 99.2% of patients.

primary outcome (up to 48 months)

composite of death from cardiovascular causes, nonfatal myocardial infarction (excluding silent myocardial infarction), or nonfatal stroke.

I: 490/4687 (10.5%)

C: 282/2333 (12.1%)

HR 0.86 (95% CI 0.74 to 0.99)

key secondary outcome (up to 48 months)

composite of the primary outcome plus hospitalization for unstable angina.

I: 599/4687 (12.8%)

C: 333/2333 (14.3%)

HR 0.89 (95% CI 0.78-1.01)

Other cardiovascular outcomes:

Empagliflozin:

lower risk of death from cardiovascular causes (HR 0.62; 95% CI 0.49 to 0.77; P<0.001),

lower risk of death from any cause

(HR 0.68; 95% CI, 0.57 to 0.82, P<0.001),

Lower risk of hospitalization for heart failure

(HR 0.65; 95% CI, 0.50 to 0.85; P = 0.002)

Myocardial infarction or stroke: no difference

Adverse events

No difference in (serious) adverse events, and adverse events leading to the discontinuation of a study drug.

Genital infection: higher % in empagliflozin group: 6.4% vs 1.8%.

Hypoglycaemia, UTI, volume depletion, acute renal failure, bone fracture, diabetic ketoacidosis, and thromboembolic events: no difference

Study reference

Describe method of randomisation¹

Bias due to inadequate  concealment of allocation?

Bias due to inadequate blinding of participants to treatment allocation?

Bias due to inadequate blinding of care providers to treatment allocation?

Bias due to inadequate blinding of outcome assessors to treatment allocation?

Bias due to selective outcome reporting on basis of the results?

Bias due to loss to follow-up?

Bias due to violation of

intention to treat analysis?

Zinman, 2015

computer-generated random-sequence and

interactive voice- and web-response system,

stratified according to HbA1c, BMI, renal function, geographic region.

Unlikely

Unlikely

Unlikely

Unlikely

Unlikely

15. Unclear.
16. 25.4% discontinued a study drug. Adverse events leading to discontinuation of study drug was higher in placebo than in empagliflozin group.

Unlikely

Modified ITT