547358 research-article2014
DVR0010.1177/1479164114547358Diabetes & Vascular Disease ResearchNandy et al.
Original Article
The effect of liraglutide on endothelial function in patients with type 2 diabetes
Diabetes & Vascular Disease Research 2014, Vol. 11(6) 419–430 © The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1479164114547358 dvr.sagepub.com
Debashis Nandy1, Christopher Johnson1, Rita Basu1, Michael Joyner1, Jason Brett2, Claus Bo Svendsen3 and Ananda Basu1
Abstract This single-centre, 12-week, double-blind, placebo-controlled trial assessed how the human glucagon-like-peptide 1 analogue liraglutide impacted endothelial function in adult patients (n = 49) with type 2 diabetes and no overt cardiovascular disease. Patients were randomized to liraglutide, placebo or glimepiride. At baseline and Week 12, venous occlusion plethysmography was used to measure forearm blood flow (FBF) in response to acetylcholine (ACh) and sodium nitroprusside (SNP) before and after l-NG-monomethyl arginine (L-NMMA) infusion. At Week 12, AChmediated FBF increased with liraglutide and decreased with placebo; however, the between-treatment difference was not significant (p = 0.055). Inhibition of ACh-mediated FBF after L-NMMA infusion increased with liraglutide and decreased with placebo; this between-treatment difference was also not significant (p = 0.149). No change in FBF was observed with SNP. Liraglutide did not significantly impact endothelium-dependent vasodilation after 12 weeks; however, additional investigations looking at the effect of liraglutide on endothelial function in alternative vasculature and during the postprandial period are warranted. Keywords Glucagon-like peptide-1, liraglutide, endothelial function
Introduction The endothelium plays a crucial role in the regulation of vascular tone and integrity, preventing atherogenesis through a balanced release of relaxing and contracting factors regulating fibrinolytic and prothrombotic activity and inhibition of leukocyte and platelet adhesion.1 Damage to the vascular endothelium can affect its normal function and predispose to atherogenesis through inflammation and impairment of endothelium-dependent vasodilation. Thus, endothelial dysfunction is often considered as an early manifestation in the development of atherosclerosis and cardiovascular (CV) disease.2 Insulin resistance and hyperglycaemia, and therefore type 2 diabetes, are associated with endothelial dysfunction,3 which may partly explain the increased risk of CV disease observed in this population compared with normoglycaemic individuals.4 Other comorbidities frequently associated with type 2 diabetes, such as dyslipidaemia, obesity and hypertension, are also linked with endothelial dysfunction, complicating the atherosclerotic process.3 The vasodilatory response of the peripheral vascular endothelium is often used as an indirect measure of endothelial function.5 Endothelial dysfunction across the forearm vascular bed is considered a surrogate endpoint
of endothelial function in the coronary circulation,6 is associated with various CV risk factors and is an independent predictor of CV events in both healthy and at-risk individuals.7–10 Therefore, it seems plausible that improvements in endothelial function may reduce CV risk. Glucagon-like peptide-1 (GLP-1) is an incretin hormone involved in maintaining glucose homeostasis after food ingestion by stimulating insulin and inhibiting glucagon secretion through glucose-dependent mechanisms.11 Interestingly, native human GLP-1 has also been shown to improve endothelium-dependent forearm blood flow (FBF) in healthy individuals12 and endothelial dysfunction in individuals with type 2 diabetes and stable coronary artery disease.13 To exploit the beneficial properties of native GLP-1, which has limited therapeutic potential due 1Mayo
Clinic, Rochester, MN, USA Nordisk, Inc., Plainsboro, NJ, USA 3Novo Nordisk A/S, Søborg, Denmark 2Novo
Corresponding author: Ananda Basu, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA. Email: [email protected]
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
420
Diabetes & Vascular Disease Research 11(6)
to its short half-life in circulation, GLP-1 receptor agonists (RAs) with longer half-lives have been developed.14,15 Several of these, including exenatide and liraglutide, are currently available for treating type 2 diabetes.16,17,18 Exenatide and liraglutide have both been shown to improve markers of endothelial function in vitro and in animal models.19–21 Additional studies with liraglutide confirmed that this effect is dependent on the GLP-1 receptor.22 Although an exploratory study in patients with type 2 diabetes found a tendency for liraglutide to increase retinal endothelial function,23 more definitive studies looking at the effect of liraglutide on other human vascular beds are lacking. The aim of this study was to investigate the effect of 12 weeks of liraglutide treatment on endothelial function, as assessed by FBF with venous occlusion plethysmography, in individuals with type 2 diabetes.
Materials and methods Participants The study was conducted at a single site in the United States. Participants were eligible if they were aged 40– 70 years, had a body mass index of 40 kg/m2 or lower, were diagnosed with type 2 diabetes and treated with lifestyle changes alone or metformin monotherapy for at least 3 months before screening and had a glycated haemoglobin (HbA1c) of 6.5%–9.0%. Main exclusion criteria were previous treatment with any antihyperglycaemic drugs other than metformin within 3 months of screening, recurrent major hypoglycaemia or hypoglycaemic unawareness, present use of any drug – except metformin – that could affect glucose and/or interfere with the outcome of the trial, hypersensitivity to trial products, high blood pressure (defined as >140/90 mmHg), initiation or change in concomitant blood pressure–lowering or lipid-lowering medication within 4 weeks prior to screening, overt CV disease, impaired renal or hepatic function, or cancer.
Trial design and conduct This was a 12-week, single-centre, double-blind, placebocontrolled, parallel group trial, with an open-label active comparator (glimepiride) arm, designed to evaluate the effect of liraglutide on endothelial function in participants with type 2 diabetes (Figure 1(a)). Participants were randomized to liraglutide 1.8 mg once daily, liraglutide placebo once daily or glimepiride 4 mg/day (open-label) in a 1:1:1 ratio using a computer-generated code. Randomization was stratified based on prior therapy with diet and lifestyle changes alone or metformin monotherapy. The goal of the open-label glimepiride arm was to achieve similar glycaemic control to the liraglutide arm, in an effort to determine whether any effect of liraglutide on endothelial function was independent of glycaemic control. To minimize gastrointestinal discomfort, the dose of liraglutide was escalated in weekly
increments of 0.6 mg to the target dose of 1.8 mg. The same dose-escalation schedule was used with placebo to ensure blinding. The dose of glimepiride was started at 1 mg/day, increased to 2 mg/day at Week 2 and to the final maintenance dose of 4 mg/day at Week 3. Dose reductions of glimepiride were allowed for unacceptable hypoglycaemia. The trial was conducted in accordance with the Declaration of Helsinki and adhered to Good Clinical Practice guidelines issued by the International Conference of Harmonization. The Mayo Institutional Review Board reviewed and approved the protocol, amendments and participant informed consent forms before trial initiation.
Study endpoints The primary endpoint was change from baseline (Week 0) to Week 12 in acetylcholine (ACh)-mediated FBF at euglycaemia for liraglutide versus placebo. Secondary endpoints included ACh-mediated FBF at euglycaemia for liraglutide versus glimepiride; amount of ACh-mediated FBF suppression induced by l-NG-monomethyl arginine (L-NMMA) at euglycaemia and change from baseline to Week 12 in sodium nitroprusside (SNP)–mediated FBF at euglycaemia, HbA1c, fasting plasma glucose (FPG) and body weight. Other endpoints included change from baseline to Week 12 in fasting lipid profile (total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, very-low-density lipoprotein cholesterol, triglycerides, free fatty acids and apolipoprotein B) and biomarkers of CV risk (high-sensitivity C-reactive protein, brain natriuretic peptide, vascular endothelial growth factor, adiponectin, interleukin-6 and tumour necrosis factor-alpha). Additionally, changes from baseline to Week 12 in ACh-mediated and SNP-mediated FBF at ambient blood glucose were exploratory endpoints and will be reported elsewhere. Safety and tolerability assessment. Safety and tolerability assessments included incidences of treatment-emergent adverse events (TEAEs) and treatment-emergent serious adverse events (TESAEs) (coded using Medical Dictionary for Regulatory Activities (MedDRA) version 13.0). In this study, TEAEs were defined as adverse events (AEs) that occurred, or events that increased in severity, on or after the first dose date and not later than 7 days after the last dose date. The severity of TEAEs was defined as follows: mild (transient symptoms, no interference with the subject’s daily activities), moderate (marked symptoms, moderate interference with the subject’s daily activities) or severe (TESAE) (considerable interference with the subject’s daily activities). Hypoglycaemia events are defined in section ‘Hypoglycaemia definitions’ of Appendix 1.
Assessment of endothelial function Endothelial function was assessed at baseline and Week 12 by measuring FBF using venous occlusion plethysmography
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
421
Nandy et al.
(a) Active liraglutide titrated weekly: 0.6 mg → 1.2 mg → 1.8 mg (n=16) • Type 2 diabetes • Diet/exercise or metformin monotherapy • Age 40–70 years (inclusive) • HbA1c 6.5–9.0% (inclusive)
Liraglutide placebo titrated weekly: 100 µL → 200 µL → 300 µL (n=16) Open-label glimepiride titrated weekly: 1 mg → 2 mg → 4 mg (n=17) Subjects on metformin at baseline remained on a stable pre-study dose
Week
–2
4 0 Baseline: VOP-1 and VOP-2
8
12 13 End of trial: Follow-up VOP-1 and VOP-2
(b) AV-line insertion: start FBF measurements Admit to CRU
ACh
Saline
06:00
06:30
07:30
Bolus insulin infusion
VOP 1
SNP
Saline
ACh
Saline
08:00
08:15
SNP
Saline
08:30
ACh
Saline
Saline
09:00
L-NMMA
AV-line removal
VOP 2
SNP
Saline
ACh
Saline
14:00 (Latest time to start VOP at euglycemia)
Ambient glucose
SNP
Saline
15:30
L-NMMA
Euglycemia
Figure 1. Schematic showing the (a) overall trial design and (b) details of VOP-1 and VOP-2. Patients were fasted overnight, and FBF was measured four times/min throughout. From 07:30 h, fasting insulin, C-peptide and glucagon levels were measured every 30 min. ACh was infused to the brachial artery at 2, 4 and 8 µg/100 mL forearm volume/min and SNP at 0.5, 1 and 2 µg/100 mL forearm volume/min; each dose lasted for 2 min. L-NMMA was infused at 5 mg/min for the first 10 min and then 1 mg/min thereafter. Bolus insulin was infused at 0.1–0.15 U/kg and then adjusted to achieve and maintain euglycaemia (90–100 mg/dL; blood glucose was monitored every 10–15 min). VOP: venous occlusion plethysmography; FBF: forearm blood flow; ACh: acetylcholine; SNP: sodium nitroprusside; L-NMMA: l-NG-monomethyl arginine; CRU: clinical research unit; AV: arterial and venous; HbA1c: glycated haemoglobin.
as described previously.12 For this study, the timing and conditions of venous occlusion plethysmography used are shown in Figure 1(b). In brief, participants were admitted to the Mayo Clinical Research Unit at 06:00 h, having fasted overnight. An arterial catheter was placed in the non-dominant arm, also as described previously,12 to allow FBF measurements with venous occlusion plethysmography. A three-port connector was placed in series with the catheter and a pressure transducer to allow drug infusion and constant measurement of arterial blood pressure. Cannulae were also placed in a vein in the dominant forearm to obtain blood for laboratory analyses and insulin infusion and in an arterialized hand vein for blood sampling of fasting insulin, C-peptide and glucagon every 30 min until the end of the study. Two sets of blood flow measurements were performed consecutively: one at ambient blood glucose and the other at euglycaemia (plasma glucose concentration 90–100 mg/dL). From 07:30 to 08:00 h, FBF responses to increasing doses of ACh (IOLAB Pharmaceuticals, Claremont, CA, USA) and SNP (Elkin Simms) were assessed as described previously.12 At 08:15 h, L-NMMA (Calbiochem, St Louis, MO, USA), at a rate of 5 mg/min for 10 min followed by 1 mg/min for the rest of the procedure, was infused into the brachial artery to inhibit endothelial nitric oxide synthase (eNOS) production,
and FBF responses to ACh and SNP were assessed again as above. At 09:00 h, all arterial infusions were stopped, and a bolus dose of regular human insulin (0.1–0.15 U/kg) was given. Intravenous regular human insulin infusion through the forearm venous cannula was also started at a variable rate to maintain the desired plasma glucose concentration, which was monitored every 10–15 min in blood drawn from the arterialized hand vein. When target glucose levels had been maintained for 2 h or at 14:00 h (whichever came first), FBF responses to ACh, in the absence and presence of L-NMMA, and SNP were assessed again as above. Heart rate was continuously monitored throughout the procedure through a five-lead electrocardiogram.
Analyses Procedures followed for the analyses of blood flow responses and glucose and hormone concentrations have been described previously.12 Blood flow analyses were done at maximum doses of ACh (8 µg/100 mL/min) and SNP (2 µg/100 mL/min) and were saline-corrected. For C-peptide, glucagon and insulin, the mean of the first four values is summarized as ambient and the mean of the last four values as euglycaemia.
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
422
Diabetes & Vascular Disease Research 11(6)
Statistical analyses Statistical analyses of all blood flow and efficacy endpoints were performed using an analysis of covariance (ANCOVA) model, with treatment as fixed effect and baseline value as a covariate on the full analysis set (i.e. all randomized participants). Two-sided 95% confidence intervals were calculated, and the significance level was set at 5%. The primary efficacy analysis compared the effect of liraglutide versus placebo on ACh-mediated FBF euglycaemia; comparisons between liraglutide versus glimepiride and glimepiride versus placebo were secondary. Post hoc analyses included a summary of the percentage of participants with maximal FBF response at maximal dose of ACh and SNP, while changes from baseline to Week 12 in the integrated area under the curve (AUC) of ACh-mediated FBF at euglycaemia were assessed using the same ANCOVA model used for the primary and secondary endpoints. Safety analyses were performed on data from the safety population, which included all participants exposed to at least one dose of the drug or who underwent at least one venous occlusion plethysmography procedure. Hypoglycaemic events were analysed using Fisher’s exact test. AEs, laboratory parameters, vital signs and physical examinations were compared between groups with descriptive statistics.
Power calculation The power calculation was based on the two-sample t-test with equal variances, as described previously.12 With 14 subjects in each treatment group, there was an 80% power to detect a difference of 1.1 standard deviation unit (twosided α = 0.05).
Results Participant disposition, demographics and baseline characteristics A total of 49 participants were enrolled and randomized to receive liraglutide (n = 16), placebo (n = 16) or glimepiride (n = 17). Of these, 46 participants (94%) completed the study and 3 participants withdrew: 2 in the placebo group and 1 in the glimepiride group (Figure 2). Demographic and baseline characteristics were generally similar between groups, with the exception of age and duration of diabetes, which were slightly greater in the placebo arm compared with the liraglutide or glimepiride arm (Table 1).
FBF responses to ACh at euglycaemia After 12 weeks, ACh-mediated FBF at euglycaemia increased from baseline in the liraglutide and glimepiride arms and decreased in the placebo arm (Figure 3(a)). After 12 weeks, least square (LS) mean changes (LS
mean ± standard error (SE)) were 4.2 ± 2.6, −3.2 ± −2.8 and 2.2 ± 2.6 mL/100 mL/min in the liraglutide, placebo and glimepiride arms, respectively. The between-treatment differences were not statistically significant for liraglutide versus placebo (p = 0.055), liraglutide versus glimepiride (p = 0.568) and glimepiride versus placebo (p = 0.167) (Figure 3(a)). Review of a scatterplot of the individual data was also not suggestive of an effect (Figure 3(b)). There was no significant correlation (r = −0.211, p = 0.1602) between these changes in ACh-mediated FBF at euglycaemia and changes in HbA1c from baseline to Week 12 (data not shown). At baseline, ACh-mediated FBF was similar before and after L-NMMA infusion in all three treatment groups. At Week 12, ACh-mediated FBF after L-NMMA infusion was suppressed in the liraglutide and glimepiride arms but not in the placebo arm compared with the values observed before L-NMMA infusion. Thus, compared with baseline, L-NMMA inhibition of ACh-mediated FBF at Week 12 increased with liraglutide and glimepiride treatment and decreased with placebo. However, between-treatment differences were not significant (Figure 4).
FBF responses to SNP at euglycaemia After 12 weeks of treatment, mean changes from baseline (LS mean ± SE) were 3.5 ± 2.7, −1.0 ± 2.9 and 2.7 ± 2.7 mL/100 mL/min with liraglutide, placebo and glimepiride, respectively. Between-treatment differences were not statistically significant (liraglutide vs placebo: p = 0.265; liraglutide vs glimepiride: p = 0.852 and glimepiride vs placebo: p = 0.344; Figure 5).
Glucose, insulin, C-peptide and glucagon measurements during the venous occlusion plethysmography procedures FPG, C-peptide, glucagon and insulin levels at baseline and Week 12 at euglycaemia are summarized in Table 2. Target blood glucose values were achieved with the insulin infusion at both Week 0 and Week 12. Mean blood glucose values during the euglycaemic clamp did not differ between treatment groups at Week 0 (95.4 vs 95.8 vs 97.2 mg/dL, liraglutide vs placebo vs glimepiride) and Week 12 (94.3 vs 96.1 vs 94.1 mg/dL, liraglutide vs placebo vs glimepiride).
Glycaemic control Change in HbA1c from baseline to Week 12 was −0.63 ± 0.44%, −0.09 ± 0.45% and −0.55 ± 0.45% with liraglutide, placebo and glimepiride treatment, respectively. Between-treatment differences were significant for liraglutide versus placebo (p = 0.002) and glimepiride versus placebo (p = 0.010). The goal of similar glycaemic control between the liraglutide and glimepiride arms was achieved
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
423
Nandy et al.
Enrollment
Assessed for eligibility (n=108)
Randomized (n=49)
Allocation Allocate to liraglutide (n=16) • Received liraglutide (n=16)
Allocate to placebo (n=16) • Received placebo (n=16)
Allocate to glimepiride (n=17) • Received glimepiride (n=17)
Follow-up
Lost to follow-up (n=0) Discontinued intervention (n=0)
Lost to follow-up (n=0) Discontinued intervention (n=2) due to: • AE (back pain, n=1) • Other (unknown, n=1)
Lost to follow-up (n=0) Discontinued intervention (n=1) due to inability to perform VOP at EOS
Analysis
Analyzed (n=16) • Excluded from analysis (n=0)
Analyzed (n=14) • Excluded from analysis (n=2) due to lack of results for primary endpoint
Analyzed (n=16) • Excluded from analysis (n=1) due to lack of results for primary endpoint
Figure 2. CONSORT diagram showing flow of patients during the study. VOP: venous occlusion plethysmography; AE: adverse event; EOS: end of study.
(p = 0.6207). Compared with placebo, significant improvements in FPG were also observed in the liraglutide and glimepiride arms (Appendix Table 3).
Body weight
pressure (LS mean ± SE) from baseline to Week 12 was 3.0 ± 2.0, −2.0 ± 2.1 and 0.8 ± 1.9 mmHg in the liraglutide, placebo and glimepiride arms, respectively. Betweentreatment differences for systolic and diastolic blood pressure were not statistically significant.
After 12 weeks of treatment, body weight decreased from baseline in the liraglutide and placebo groups and increased in the glimepiride group. Between-treatment differences were significant for liraglutide versus placebo, liraglutide versus glimepiride and glimepiride versus placebo (Appendix Table 4).
Lipid profile and biomarkers of CV risk
Blood pressure
Safety and tolerability
Change in systolic blood pressure (LS mean ± SE) from baseline to Week 12 was −4.6 ± 2.4, −4.8 ± 2.6 and −0.2 ± 2.4 mmHg in the liraglutide, placebo and glimepiride treatment arms, respectively. Change in diastolic blood
AEs. There was a single TESAE (prostate cancer), which occurred in the placebo group. The percentage of patients who reported TEAEs was greater with liraglutide (68.8%) and placebo (56.3%) than with glimepiride (29.4%).
Change from baseline to Week 12 in lipid profiles and biomarkers of CV risk is summarized in the appendix (Appendix Tables 5 and 6, respectively). Betweentreatment differences were not significant.
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
424
Diabetes & Vascular Disease Research 11(6)
Table 1. Summary of demographic and selected baseline characteristics.
Male sex (%) Age (years) Race (% White) Weight (kg) BMI (kg/m2) Duration of diabetes (years) Previous treatment Diet and exercise Metformin ACh-mediated FBF at ambient glucose (mL/100 mL/min) SNP-mediated FBF at ambient glucose (mL/100 mL/min) ACh-mediated FBF at euglycaemia (mL/100 mL/min) SNP-mediated FBF at euglycaemia (mL/100 mL/min) HbA1c (%) FPG (mmol/L)
Liraglutide (n = 16)
Placebo (n = 16)
Glimepiride (n = 17)
10 (62.5) 57.7 (9.0) 16 (100.0) 95.1 (13.1) 32.7 (4.5) 5.3 (4.1)
10 (62.5) 60.3 (7.3) 16 (100.0) 90.6 (13.5) 31.6 (4.2) 8.4 (4.6)
2 (12.5) 14 (87.5) 7.9 (5.7) 11.0 (4.5) 8.5 (6.7) 10.2 (4.4) 7.2 (0.5) 8.9 (1.7)
1 (6.3) 15 (93.8) 8.7 (5.8) 11 (5.0) 10.0 (6.6) 12.0 (6.8) 7.0 (0.5) 8.1 (1.1)
11 (64.7) 57.7 (5.3) 17 (100.0) 92.0 (14.0) 31.1 (4.9) 6.8 (8.1) 2 (11.8) 15 (88.2) 5.9 (4.4) 11.1 (5.2) 7.9 (7.6) 10.7 (5.1) 7.3 (0.5) 9.1 (1.9)
ACh: acetylcholine; BMI: body mass index; FBF: forearm blood flow; FPG: fasting plasma glucose; SD: standard deviation; SNP: sodium nitroprusside; HbA1c: glycated haemoglobin. Sex, race and previous treatment are number of patients (%); all other parameters are mean (SD).
(a)
P=0.568
Change in ACh-mediated FBF at euglycemia (mL/100 mL/min)
20.0
P=0.167
P=0.055a
15.0 10.0 5.0 4.2
0.0
2.2 –3.2
–5.0
–10.0 –15.0
Liraglutide 1.8 mg/day Placebo Glimepiride 4 mg/day
Change in ACh-mediated FBF at euglycemia (mL/100 mL/min)
(b) 60 50
49.0
40 30 20
29.7 19.2
16.4
10 0 –10 –20 –30
–17.7
Liraglutide 1.8 mg/day
–19.3
Placebo
–18.4
Glimepiride 4 mg/day
Figure 3. (a) Change in ACh-mediated FBF (LS mean ± SE) at euglycaemia in the liraglutide, placebo and glimepiride arms. (b) Scatterplot of individual ACh-mediated FBF data. ACh: acetylcholine; FBF: forearm blood flow; LS: least square; SE: standard error.
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
425
Difference in change from baseline to Week 12 in ACh-mediated FBF (mL/100 mL/min) in the absence and presence of L-NMMA
Nandy et al. P=0.634 P=0.149
15.0
P=0.320
10.0
5.0 3.7
2.3
–0.9
0.0
–5.0
–10.0
–15.0 Liraglutide 1.8 mg/day
Placebo
Glimepiride 4 mg/day
Figure 4. Difference in change from baseline to Week 12 in ACh-mediated FBF inhibition before and after L-NMMA infusion (LS mean ± SE). ACh: acetylcholine; FBF: forearm blood flow; L-NMMA: l-NG-monomethyl arginine; LS: least square; SE: standard error.
P=0.852
20.0 P=0.265
P=0.344
Change in SNP-mediated FBF at euglycemia (mL/100 mL/min)
15.0 10.0 5.0 3.5 0.0
–1.0
2.7
–5.0 –10.0 –15.0 Liraglutide 1.8 mg/day
Placebo
Glimepiride 4 mg/day
Figure 5. Change in SNP-mediated FBF (LS mean ± SE) at euglycaemia in the liraglutide, placebo and glimepiride arms. SNP: sodium nitroprusside; FBF: forearm blood flow; LS: least square; SE: standard error.
TEAEs were either mild (95%) or moderate (5%) in severity, and most (80%) were deemed by the investigator to be unlikely related to treatment. Only eight events, all of which occurred in the liraglutide group, were considered possibly related to treatment. Of these, most were of gastrointestinal nature (two events of dyspepsia, two nausea, two vomiting), one was a viral gastroenteritis and the remaining one was decreased appetite. No change in laboratory parameters, vital signs and physical examinations was observed (section ‘Laboratory parameters, vital signs
and physical examinations’ of Appendix 1). No withdrawals occurred due to AEs in patients treated with liraglutide. Hypoglycaemia. There were no episodes of major hypoglycaemia. The rate of minor hypoglycaemia was lower in the liraglutide group than in the glimepiride group (0.26 vs 2.46 events/patient/year). No events of minor hypoglycaemia were reported in the placebo group. Comparisons of the number of subjects experiencing hypoglycaemia between
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
426
Diabetes & Vascular Disease Research 11(6)
Table 2. Plasma glucose, C-peptide, glucagon and insulin levels at baseline and Week 12 at euglycaemia during the clamp and venous occlusion plethysmography measurements.
Glucose (mg/dL) Baseline Week 12 C-peptide (nmol/L) Baseline Week 12 Glucagon (pg/mL) Baseline Week 12 Insulin (µU/mL) Baseline Week 12
Liraglutide
Placebo
95.4 (8.7) 94.3 (4.3)
95.8 (4.3) 96.1 (6.1)
0.5 (0.2) 0.8 (0.3)
0.5 (0.2) 0.5 (0.1)
89.8 (29.1) 80.6 (26.6)
92.6 (25.7) 96.2 (26.0)
19.8 (17.0) 12.8 (6.9)
14.8 (7.5) 17.3 (13.9)
Glimepiride 97.2 (5.2) 94.1 (5.7) 0.4 (0.2) 0.5 (0.2) 86.9 (35.0) 75.4 (20.2) 19.5 (16.5) 14.0 (8.8)
SD: standard deviation. Data are mean (SD).
the liraglutide and placebo groups (p = 0.484) and between the liraglutide and glimepiride groups (p = 0.398) were not statistically significant.
Post hoc analyses Proportion of subjects with peak FBF response at maximal dose of ACh and SNP at euglycaemia. Peak FBF response at maximal dose of ACh was observed in 55% of patients at baseline and in 61% of patients at Week 12. In contrast, peak FBF response was seen at the maximal dose of SNP in 76% of patients at baseline and in 80% of patients at Week 12. AUC of ACh-mediated FBF at euglycaemia. After 12 weeks of treatment, change from baseline in AUC of ACh-mediated FBF (LS mean ± SE) at euglycaemia was 17.5 ± 12.3, −10.6 ± 13.3 and 0.8 ± 12.3 mL/100 mL with liraglutide, placebo and glimepiride, respectively. Between-treatment differences were not statistically significant.
Discussion In this study, no significant effect of liraglutide on endothelium-dependent vasodilation was seen. Although the p-value approached statistical significance (p = 0.055), further analysis of the distribution of the individual data did not suggest a treatment effect. There was also no effect of liraglutide on endothelium-independent vasodilation. Given that hyperglycaemia has a detrimental effect on endothelium-dependent vasodilation,3,24 our study was designed to control for acute differences in glycaemia. Euglycaemia was achieved with the insulin infusion, and the goal of achieving similar acute glycaemic control was fulfilled, with the mean blood glucose values being similar between treatment arms during the FBF assessments at both baseline and end of study.
The study was also designed to determine whether the presumed effect of liraglutide was independent of its effect on glucose. The goal of attaining a similar reduction in HbA1c with liraglutide and glimepiride was also achieved. Although the change in ACh-mediated vasodilation was numerically greater with liraglutide than glimepiride, the between-treatment difference was not statistically significant. Importantly, however, the study was not powered to detect differences in blood flow between liraglutide and glimepiride. In our study, the changes in ACh-mediated FBF and HbA1c were not correlated. These findings suggest, but do not prove, that the observed effect of liraglutide, although not statistically significantly different from placebo, is independent of its effect on glycaemic control. We chose to analyse FBF at maximum doses of ACh and SNP because this usually resulted in the maximum response in previous studies in healthy, non-diabetic individuals (A Basu, unpublished observations).3 Intriguingly, the proportion of subjects who had their peak response at the maximum dose of ACh was somewhat lower in this study than what has been observed in prior studies in people without diabetes, suggesting that the nitric oxide pathway may be impaired in patients with type 2 diabetes. Because of this finding and given that the p-value approached significance, a post hoc AUC analysis, taking into account all ACh doses, was also done. Between-treatment differences were not significant in this post hoc analysis. Our findings are consistent with some prior studies but conflict with others. Improved vascular function has been reported following native GLP-1 infusion in preclinical studies25,26 and also in clinical trials in both healthy subjects12 and patients with type 2 diabetes.13 However, it is not yet clear whether native GLP-1 exerts this beneficial effect through direct actions by the GLP-1 receptor27 or through indirect actions of its dipeptidyl peptidase (DPP)-4– mediated cleavage product, GLP-1(9–36), known to mediate important mitochondrial and other, as yet unidentified, GLP-1 receptor–independent signalling pathways.28,29 For incretin-based therapies, including GLP-1 RAs and DPP-4 inhibitors, the overall effect on vascular function is less clear. For GLP-1 RAs, a recent mechanistic study in mice that investigated the role of GLP-1 receptor activation in blood pressure control found that liraglutide had no direct effect on phosphorylated eNOS.30 Moreover, 3 months of exenatide therapy was not found to have an effect on endothelial function in patients with prediabetes and abdominal obesity, as measured by digital reactive hyperaemia, C-reactive protein, circulating oxidized lowdensity lipoprotein and vascular cell adhesion molecule-1.31 Another study examining the effect of 6 months of GLP-1 RA therapy in obese patients with type 2 diabetes found no effect on endothelial function, as measured by brachial flow–mediated dilation and carotid intimal– medial thickness.32 The low vascular impact of GLP-1 RAs described in the above studies could suggest that improved vascular
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
427
Nandy et al. function is more likely to be mediated by GLP-1 receptor– independent mechanisms, such as GLP-1(9–36) signalling. However, other studies provide evidence that GLP-1 RAs can directly improve endothelial function. An exploratory study in patients with type 2 diabetes who were in good metabolic control on metformin monotherapy found that liraglutide improved the retinal hyperaemic response to flicker light after 6 and 12 weeks of therapy.23 Another study with exenatide supports the notion that, as shown for GLP-1,13 GLP-1 RAs may also have a positive effect on endothelial function. In this observational study in patients with type 2 diabetes, endothelial function was significantly improved after 16 weeks of treatment with exenatide compared with glimepiride. Exenatide treatment and HbA1c were both independently and significantly associated with peak flow–mediated dilation. In addition, the reactivity of peripheral resistance vessels, which was indirectly evaluated through measurements of shear rate increase after ischaemia, also improved significantly with exenatide treatment. However, the study was limited by both a small number of subjects (n = 20) who were not matched in the two treatment groups and the lack of a randomized doubleblind design.33 For DPP-4 inhibitors, conflicting data have also been published regarding their effects on the vascular system. One recent study reports that 6 weeks of sitagliptin or alogliptin treatment significantly attenuated endothelial function in patients with type 2 diabetes,34 which is in support of the hypothesis that normal degradation products of native GLP-1 such as GLP-1(9–36) may be mediating vascular improvements through GLP-1 receptor–independent mechanisms. However, it has also been shown that both 12 weeks of sitagliptin treatment and 4 weeks of vildagliptin treatment resulted in significant improvement in endothelial function in patients with type 2 diabetes,35,36 possibly due to an increase in circulating levels of GLP-1, which may increase GLP-1 receptor–dependent responses, not only on vascular endothelial cells but also on myocardial and smooth muscle cells, where the GLP-1 receptor is also expressed. In addition, it has been reported that the GLP-1 RA, exenatide, when given as a single dose prior to a high-fat meal, improved postprandial endothelial function in patients with impaired glucose tolerance or recent-onset type 2 diabetes. The observed response in these patients suggests that the effect of GLP-1 RAs on endothelial function may occur in the postprandial period, possibly due to a reduction in postprandial triglycerides,37 which is supported by another study in which liraglutide was also shown to reduce postprandial triglycerides.38 This may also be the case for DPP-4 inhibitors, as shown by a recent study in patients without diabetes, in which 1 week of alogliptin treatment ameliorated postprandial endothelial dysfunction.39 In contrast to the above studies, in our study, we assessed endothelial function and lipids in the fasting state.
In the majority of the above studies, blood flow was assessed by responses in the brachial artery utilizing flow-mediated dilation, in contrast to this study, and our previously published study,12 which assessed blood flow in the regional microcirculation using venous occlusion plethysmography. Given that mechanisms of vasodilation in conduit vessels are often different from those in resistance vessels, where blood flow is regulated, it is worth noting that flow-mediated dilation assesses vasodilation in large conduit vessels, while venous occlusion plethysmography measures vasodilation in small resistance vessels and is considered a more robust method for measuring vascular activity.5 This methodological difference makes direct comparisons difficult between our findings and those of others. There were several limitations to our study. First, we did not use a somatostatin clamp to inhibit endogenous secretion of insulin and C-peptide (so as to match insulin and C-peptide levels between groups during the venous occlusion plethysmography studies), both of which can affect endothelial function.40 Second, individuals in the glimepiride treatment arm were not blinded to treatment. The fact that the study was not sufficiently powered to detect a statistically significant difference between liraglutide and glimepiride groups was also a limitation, as was the large variability present in our measurements, most likely due to the inherently increased metabolic variability of subjects with endocrine dysfunctions such as type 2 diabetes. Finally, since the FBF endpoints at euglycaemia were determined after the exploratory FBF endpoints at ambient blood glucose, sequence bias cannot be ruled out. In summary, 12 weeks of liraglutide therapy did not have a significant effect on endothelium-dependent vasodilation in this study. Further investigation with liraglutide, including assessment of endothelial function in other vascular beds and during the postprandial period, is warranted.
Key messages •• This short trial assessed how the human glucagonlike peptide-1 (GLP-1) analogue liraglutide impacted endothelial function in adult patients with type 2 diabetes and no overt cardiovascular disease. •• Liraglutide did not significantly impact endothelium-dependent vasodilation after 12 weeks. •• Additional investigations into liraglutide and endothelial function in alternative vasculature and during the postprandial period are needed. Acknowledgements The authors are grateful to Vatsala Karwe (employed at Novo Nordisk Inc. at the time of article initiation) and Anjun Cao (currently employed by Novo Nordisk) for statistical advice and input. The authors also thank Dr Rosalyn Ferguson, PhD, Watermeadow Medical (Witney, UK), supported by Novo
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
428
Diabetes & Vascular Disease Research 11(6)
Nordisk, for medical writing and editorial assistance. Clinical trial registration number: NCT00620282.
Declaration of conflicting interests D.N., C.J., R.B. and M.J. have nothing to disclosure. J.B. and C.S. are full-time employees and shareholders in Novo Nordisk.
Funding Novo Nordisk funded the trial and supported the writing assistance for the preparation of this article. A.B. was principal investigator of this trial, funded by Novo Nordisk. This article’s preparation was funded by Novo Nordisk.
References 1. Libby P. Inflammation in atherosclerosis. Nature 2002; 420: 868–874. 2. Widlansky ME, Gokce N, Keaney JF Jr, et al. The clinical implications of endothelial dysfunction. J Am Coll Cardiol 2003; 42: 1149–1160. 3. Hadi HAR and Al Suwaidi J. Endothelial dysfunction in diabetes mellitus. Vasc Health Risk Manag 2007; 3: 853–876. 4. Haffner SM, Lehto S, Rönnemaa T, et al. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med 1998; 339: 229–234. 5. Eskurza I, Seals DR, DeSouza CA, et al. Pharmacologic versus flow-mediated assessments of peripheral vascular endothelial vasodilatory function in humans. Am J Cardiol 2001; 88: 1067–1069. 6. Tagawa T, Mohri M, Tagawa H, et al. Role of nitric oxide in substance P-induced vasodilation differs between the coronary and forearm circulation in humans. J Cardiovasc Pharmacol 1997; 29: 546–553. 7. Hadi HAR, Carr CS and Al Suwaidi J. Endothelial dysfunction: cardiovascular risk factors, therapy, and outcome. Vasc Health Risk Manag 2005; 1: 183–198. 8. Perticone F, Ceravolo R, Pujia A, et al. Prognostic significance of endothelial dysfunction in hypertensive patients. Circulation 2001; 104: 191–196. 9. Heitzer T, Schlinzig T, Krohn K, et al. Endothelial dysfunction, oxidative stress, and risk of cardiovascular events in patients with coronary artery disease. Circulation 2001; 104: 2673–2678. 10. Green DJ, Jones H, Thijssen D, et al. Flow-mediated dilation and cardiovascular event prediction: does nitric oxide matter? Hypertension 2011; 57: 363–369. 11. Drucker DJ. The biology of incretin hormones. Cell Metab 2006; 3: 153–165. 12. Basu A, Charkoudian N, Schrage W, et al. Beneficial effects of GLP-1 on endothelial function in humans: dampening by glyburide but not by glimepiride. Am J Physiol Endocrinol Metab 2007; 293: E1289–E1295. 13. Nyström T, Gutniak MK, Zhang Q, et al. Effects of glucagon-like peptide-1 on endothelial function in type 2 diabetes patients with stable coronary artery disease. Am J Physiol Endocrinol Metab 2004; 287: E1209–E1215.
14. Vilsbøll T, Agersø H, Krarup T, et al. Similar elimination rates of glucagon-like peptide-1 in obese type 2 diabetic patients and healthy subjects. J Clin Endocrinol Metab 2003; 88: 220–224. 15. Cho YM, Merchant CE and Kieffer TJ. Targeting the glucagon receptor family for diabetes and obesity therapy. Pharmacol Ther 2012; 135: 247–278. 16. Montanya E. A comparison of currently available GLP-1 receptor agonists for the treatment of type 2 diabetes. Expert Opin Pharmacother 2012; 13: 1451–1467. 17. Elkinson S and Keating GM. Lixisenatide: first global approval. Drugs 2013; 73: 383–391. 18. Poole RM and Nowlan ML. Albiglutide: first global approval. Drugs 2014; 74: 929–938. 19. Arakawa M, Mita T, Azuma K, et al. Inhibition of monocyte adhesion to endothelial cells and attenuation of atherosclerotic lesion by a glucagon-like peptide-1 receptor agonist, exendin-4. Diabetes 2010; 59: 1030–1037. 20. Liu H, Dear AE, Knudsen LB, et al. A long-acting glucagonlike peptide-1 analogue attenuates induction of plasminogen activator inhibitor type-1 and vascular adhesion molecules. J Endocrinol 2009; 201: 59–66. 21. Hattori Y, Jojima T, Tomizawa A, et al. A glucagon-like peptide-1 (GLP-1) analogue, liraglutide, upregulates nitric oxide production and exerts anti-inflammatory action in endothelial cells. Diabetologia 2010; 53: 2256–2263. 22. Gaspari T, Liu H, Welungoda I, et al. A GLP-1 receptor agonist liraglutide inhibits endothelial cell dysfunction and vascular adhesion molecule expression in an ApoE-/- mouse model. Diab Vasc Dis Res 2011; 8: 117–124. 23. Forst T, Michelson G, Ratter F, et al. Addition of liraglutide in patients with Type 2 diabetes well controlled on metformin monotherapy improves several markers of vascular function. Diabet Med 2012; 29: 1115–1118. 24. Ding Y, Vaziri ND, Coulson R, et al. Effects of simulated hyperglycemia, insulin, and glucagon on endothelial nitric oxide synthase expression. Am J Physiol Endocrinol Metab 2000; 279: E11–E17. 25. Chai W, Dong Z, Wang N, et al. Glucagon-like peptide 1 recruits microvasculature and increases glucose use in muscle via a nitric oxide-dependent mechanism. Diabetes 2012; 61: 888–896. 26. Nikolaidis LA, Elahi D, Hentosz T, et al. Recombinant glucagon-like peptide-1 increases myocardial glucose uptake and improves left ventricular performance in conscious dogs with pacing-induced dilated cardiomyopathy. Circulation 2004; 110: 955–961. 27. Cabou C, Vachoux C, Campistron G, et al. Brain GLP-1 signaling regulates femoral artery blood flow and insulin sensitivity through hypothalamic PKC-δ. Diabetes 2011; 60: 2245–2256. 28. Ussher JR and Drucker DJ. Cardiovascular actions of incretin-based therapies. Circ Res 2014; 114: 1788–1803. 29. Ban K, Kim KH, Cho CK, et al. Glucagon-like peptide (GLP)-1(9-36) amide-mediated cytoprotection is blocked by exendin(9-39) yet does not require the known GLP-1 receptor. Endocrinology 2010; 151: 1520–1531. 30. Kim M, Platt MJ, Shibasaki T, et al. GLP-1 receptor activation and Epac2 link atrial natriuretic peptide secretion to control of blood pressure. Nat Med 2013; 19: 567–575.
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
429
Nandy et al. 31. Kelly AS, Bergenstal RM, Gonzalez-Campoy JM, et al. Effects of exenatide vs. metformin on endothelial function in obese patients with pre-diabetes: a randomized trial. Cardiovasc Diabetol 2012; 11: 64. 32. Hopkins ND, Cuthbertson DJ, Kemp GJ, et al. Effects of 6 months glucagon-like peptide-1 receptor agonist treatment on endothelial function in type 2 diabetes mellitus patients. Diabetes Obes Metab 2013; 15: 770–773. 33. Irace C, De Luca S, Shehaj E, et al. Exenatide improves endothelial function assessed by flow mediated dilation technique in subjects with type 2 diabetes: results from an observational research. Diab Vasc Dis Res 2013; 10: 72–77. 34. Ayaori M, Iwakami N, Uto-Kondo H, et al. Dipeptidyl peptidase-4 inhibitors attenuate endothelial function as evaluated by flow-mediated vasodilatation in type 2 diabetic patients. J Am Heart Assoc 2013; 2: e003277. 35. Kubota Y, Miyamoto M, Takagi G, et al. The dipeptidyl peptidase-4 inhibitor sitagliptin improves vascular endothelial function in type 2 diabetes. J Korean Med Sci 2012; 27: 1364–1370. 36. Van Poppel PC, Netea MG, Smits P, et al. Vildagliptin improves endothelium-dependent vasodilatation in type 2 diabetes. Diabetes Care 2011; 34: 2072–2077. 37. Koska J, Schwartz EA, Mullin MP, et al. Improvement of postprandial endothelial function after a single dose of exenatide in individuals with impaired glucose tolerance and recent-onset type 2 diabetes. Diabetes Care 2010; 33: 1028–1030. 38. Hermansen K, Baekdal TA, Düring M, et al. Liraglutide suppresses postprandial triglyceride and apolipoprotein B48 elevations after a fat-rich meal in patients with type 2 diabetes: a randomized, double-blind, placebo-controlled, cross-over trial. Diabetes Obes Metab 2013; 15: 1040–1048. 39. Noda Y, Miyoshi T, Oe H, et al. Alogliptin ameliorates postprandial lipemia and postprandial endothelial dysfunction in non-diabetic subjects: a preliminary report. Cardiovasc Diabetol 2013; 12: 8.
40. Ardigo D, Franzini L, Valtuena S, et al. Relation of plasma insulin levels to forearm flow-mediated dilatation in healthy volunteers. Am J Cardiol 2006; 97: 1250–1254.
Appendix 1 Hypoglycaemia definitions Hypoglycaemia in this trial was defined as either major hypoglycaemia – defined as an episode with severe central nervous system symptoms consistent with hypoglycaemia, in which the subject was unable to treat himself/ herself and had one of the following characteristics: (1) reversal of symptoms after either food intake or glucagon/intravenous glucose administration or (2) plasma glucose
DVR0010.1177/1479164114547358Diabetes & Vascular Disease ResearchNandy et al.
Original Article
The effect of liraglutide on endothelial function in patients with type 2 diabetes
Diabetes & Vascular Disease Research 2014, Vol. 11(6) 419–430 © The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1479164114547358 dvr.sagepub.com
Debashis Nandy1, Christopher Johnson1, Rita Basu1, Michael Joyner1, Jason Brett2, Claus Bo Svendsen3 and Ananda Basu1
Abstract This single-centre, 12-week, double-blind, placebo-controlled trial assessed how the human glucagon-like-peptide 1 analogue liraglutide impacted endothelial function in adult patients (n = 49) with type 2 diabetes and no overt cardiovascular disease. Patients were randomized to liraglutide, placebo or glimepiride. At baseline and Week 12, venous occlusion plethysmography was used to measure forearm blood flow (FBF) in response to acetylcholine (ACh) and sodium nitroprusside (SNP) before and after l-NG-monomethyl arginine (L-NMMA) infusion. At Week 12, AChmediated FBF increased with liraglutide and decreased with placebo; however, the between-treatment difference was not significant (p = 0.055). Inhibition of ACh-mediated FBF after L-NMMA infusion increased with liraglutide and decreased with placebo; this between-treatment difference was also not significant (p = 0.149). No change in FBF was observed with SNP. Liraglutide did not significantly impact endothelium-dependent vasodilation after 12 weeks; however, additional investigations looking at the effect of liraglutide on endothelial function in alternative vasculature and during the postprandial period are warranted. Keywords Glucagon-like peptide-1, liraglutide, endothelial function
Introduction The endothelium plays a crucial role in the regulation of vascular tone and integrity, preventing atherogenesis through a balanced release of relaxing and contracting factors regulating fibrinolytic and prothrombotic activity and inhibition of leukocyte and platelet adhesion.1 Damage to the vascular endothelium can affect its normal function and predispose to atherogenesis through inflammation and impairment of endothelium-dependent vasodilation. Thus, endothelial dysfunction is often considered as an early manifestation in the development of atherosclerosis and cardiovascular (CV) disease.2 Insulin resistance and hyperglycaemia, and therefore type 2 diabetes, are associated with endothelial dysfunction,3 which may partly explain the increased risk of CV disease observed in this population compared with normoglycaemic individuals.4 Other comorbidities frequently associated with type 2 diabetes, such as dyslipidaemia, obesity and hypertension, are also linked with endothelial dysfunction, complicating the atherosclerotic process.3 The vasodilatory response of the peripheral vascular endothelium is often used as an indirect measure of endothelial function.5 Endothelial dysfunction across the forearm vascular bed is considered a surrogate endpoint
of endothelial function in the coronary circulation,6 is associated with various CV risk factors and is an independent predictor of CV events in both healthy and at-risk individuals.7–10 Therefore, it seems plausible that improvements in endothelial function may reduce CV risk. Glucagon-like peptide-1 (GLP-1) is an incretin hormone involved in maintaining glucose homeostasis after food ingestion by stimulating insulin and inhibiting glucagon secretion through glucose-dependent mechanisms.11 Interestingly, native human GLP-1 has also been shown to improve endothelium-dependent forearm blood flow (FBF) in healthy individuals12 and endothelial dysfunction in individuals with type 2 diabetes and stable coronary artery disease.13 To exploit the beneficial properties of native GLP-1, which has limited therapeutic potential due 1Mayo
Clinic, Rochester, MN, USA Nordisk, Inc., Plainsboro, NJ, USA 3Novo Nordisk A/S, Søborg, Denmark 2Novo
Corresponding author: Ananda Basu, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA. Email: [email protected]
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
420
Diabetes & Vascular Disease Research 11(6)
to its short half-life in circulation, GLP-1 receptor agonists (RAs) with longer half-lives have been developed.14,15 Several of these, including exenatide and liraglutide, are currently available for treating type 2 diabetes.16,17,18 Exenatide and liraglutide have both been shown to improve markers of endothelial function in vitro and in animal models.19–21 Additional studies with liraglutide confirmed that this effect is dependent on the GLP-1 receptor.22 Although an exploratory study in patients with type 2 diabetes found a tendency for liraglutide to increase retinal endothelial function,23 more definitive studies looking at the effect of liraglutide on other human vascular beds are lacking. The aim of this study was to investigate the effect of 12 weeks of liraglutide treatment on endothelial function, as assessed by FBF with venous occlusion plethysmography, in individuals with type 2 diabetes.
Materials and methods Participants The study was conducted at a single site in the United States. Participants were eligible if they were aged 40– 70 years, had a body mass index of 40 kg/m2 or lower, were diagnosed with type 2 diabetes and treated with lifestyle changes alone or metformin monotherapy for at least 3 months before screening and had a glycated haemoglobin (HbA1c) of 6.5%–9.0%. Main exclusion criteria were previous treatment with any antihyperglycaemic drugs other than metformin within 3 months of screening, recurrent major hypoglycaemia or hypoglycaemic unawareness, present use of any drug – except metformin – that could affect glucose and/or interfere with the outcome of the trial, hypersensitivity to trial products, high blood pressure (defined as >140/90 mmHg), initiation or change in concomitant blood pressure–lowering or lipid-lowering medication within 4 weeks prior to screening, overt CV disease, impaired renal or hepatic function, or cancer.
Trial design and conduct This was a 12-week, single-centre, double-blind, placebocontrolled, parallel group trial, with an open-label active comparator (glimepiride) arm, designed to evaluate the effect of liraglutide on endothelial function in participants with type 2 diabetes (Figure 1(a)). Participants were randomized to liraglutide 1.8 mg once daily, liraglutide placebo once daily or glimepiride 4 mg/day (open-label) in a 1:1:1 ratio using a computer-generated code. Randomization was stratified based on prior therapy with diet and lifestyle changes alone or metformin monotherapy. The goal of the open-label glimepiride arm was to achieve similar glycaemic control to the liraglutide arm, in an effort to determine whether any effect of liraglutide on endothelial function was independent of glycaemic control. To minimize gastrointestinal discomfort, the dose of liraglutide was escalated in weekly
increments of 0.6 mg to the target dose of 1.8 mg. The same dose-escalation schedule was used with placebo to ensure blinding. The dose of glimepiride was started at 1 mg/day, increased to 2 mg/day at Week 2 and to the final maintenance dose of 4 mg/day at Week 3. Dose reductions of glimepiride were allowed for unacceptable hypoglycaemia. The trial was conducted in accordance with the Declaration of Helsinki and adhered to Good Clinical Practice guidelines issued by the International Conference of Harmonization. The Mayo Institutional Review Board reviewed and approved the protocol, amendments and participant informed consent forms before trial initiation.
Study endpoints The primary endpoint was change from baseline (Week 0) to Week 12 in acetylcholine (ACh)-mediated FBF at euglycaemia for liraglutide versus placebo. Secondary endpoints included ACh-mediated FBF at euglycaemia for liraglutide versus glimepiride; amount of ACh-mediated FBF suppression induced by l-NG-monomethyl arginine (L-NMMA) at euglycaemia and change from baseline to Week 12 in sodium nitroprusside (SNP)–mediated FBF at euglycaemia, HbA1c, fasting plasma glucose (FPG) and body weight. Other endpoints included change from baseline to Week 12 in fasting lipid profile (total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, very-low-density lipoprotein cholesterol, triglycerides, free fatty acids and apolipoprotein B) and biomarkers of CV risk (high-sensitivity C-reactive protein, brain natriuretic peptide, vascular endothelial growth factor, adiponectin, interleukin-6 and tumour necrosis factor-alpha). Additionally, changes from baseline to Week 12 in ACh-mediated and SNP-mediated FBF at ambient blood glucose were exploratory endpoints and will be reported elsewhere. Safety and tolerability assessment. Safety and tolerability assessments included incidences of treatment-emergent adverse events (TEAEs) and treatment-emergent serious adverse events (TESAEs) (coded using Medical Dictionary for Regulatory Activities (MedDRA) version 13.0). In this study, TEAEs were defined as adverse events (AEs) that occurred, or events that increased in severity, on or after the first dose date and not later than 7 days after the last dose date. The severity of TEAEs was defined as follows: mild (transient symptoms, no interference with the subject’s daily activities), moderate (marked symptoms, moderate interference with the subject’s daily activities) or severe (TESAE) (considerable interference with the subject’s daily activities). Hypoglycaemia events are defined in section ‘Hypoglycaemia definitions’ of Appendix 1.
Assessment of endothelial function Endothelial function was assessed at baseline and Week 12 by measuring FBF using venous occlusion plethysmography
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
421
Nandy et al.
(a) Active liraglutide titrated weekly: 0.6 mg → 1.2 mg → 1.8 mg (n=16) • Type 2 diabetes • Diet/exercise or metformin monotherapy • Age 40–70 years (inclusive) • HbA1c 6.5–9.0% (inclusive)
Liraglutide placebo titrated weekly: 100 µL → 200 µL → 300 µL (n=16) Open-label glimepiride titrated weekly: 1 mg → 2 mg → 4 mg (n=17) Subjects on metformin at baseline remained on a stable pre-study dose
Week
–2
4 0 Baseline: VOP-1 and VOP-2
8
12 13 End of trial: Follow-up VOP-1 and VOP-2
(b) AV-line insertion: start FBF measurements Admit to CRU
ACh
Saline
06:00
06:30
07:30
Bolus insulin infusion
VOP 1
SNP
Saline
ACh
Saline
08:00
08:15
SNP
Saline
08:30
ACh
Saline
Saline
09:00
L-NMMA
AV-line removal
VOP 2
SNP
Saline
ACh
Saline
14:00 (Latest time to start VOP at euglycemia)
Ambient glucose
SNP
Saline
15:30
L-NMMA
Euglycemia
Figure 1. Schematic showing the (a) overall trial design and (b) details of VOP-1 and VOP-2. Patients were fasted overnight, and FBF was measured four times/min throughout. From 07:30 h, fasting insulin, C-peptide and glucagon levels were measured every 30 min. ACh was infused to the brachial artery at 2, 4 and 8 µg/100 mL forearm volume/min and SNP at 0.5, 1 and 2 µg/100 mL forearm volume/min; each dose lasted for 2 min. L-NMMA was infused at 5 mg/min for the first 10 min and then 1 mg/min thereafter. Bolus insulin was infused at 0.1–0.15 U/kg and then adjusted to achieve and maintain euglycaemia (90–100 mg/dL; blood glucose was monitored every 10–15 min). VOP: venous occlusion plethysmography; FBF: forearm blood flow; ACh: acetylcholine; SNP: sodium nitroprusside; L-NMMA: l-NG-monomethyl arginine; CRU: clinical research unit; AV: arterial and venous; HbA1c: glycated haemoglobin.
as described previously.12 For this study, the timing and conditions of venous occlusion plethysmography used are shown in Figure 1(b). In brief, participants were admitted to the Mayo Clinical Research Unit at 06:00 h, having fasted overnight. An arterial catheter was placed in the non-dominant arm, also as described previously,12 to allow FBF measurements with venous occlusion plethysmography. A three-port connector was placed in series with the catheter and a pressure transducer to allow drug infusion and constant measurement of arterial blood pressure. Cannulae were also placed in a vein in the dominant forearm to obtain blood for laboratory analyses and insulin infusion and in an arterialized hand vein for blood sampling of fasting insulin, C-peptide and glucagon every 30 min until the end of the study. Two sets of blood flow measurements were performed consecutively: one at ambient blood glucose and the other at euglycaemia (plasma glucose concentration 90–100 mg/dL). From 07:30 to 08:00 h, FBF responses to increasing doses of ACh (IOLAB Pharmaceuticals, Claremont, CA, USA) and SNP (Elkin Simms) were assessed as described previously.12 At 08:15 h, L-NMMA (Calbiochem, St Louis, MO, USA), at a rate of 5 mg/min for 10 min followed by 1 mg/min for the rest of the procedure, was infused into the brachial artery to inhibit endothelial nitric oxide synthase (eNOS) production,
and FBF responses to ACh and SNP were assessed again as above. At 09:00 h, all arterial infusions were stopped, and a bolus dose of regular human insulin (0.1–0.15 U/kg) was given. Intravenous regular human insulin infusion through the forearm venous cannula was also started at a variable rate to maintain the desired plasma glucose concentration, which was monitored every 10–15 min in blood drawn from the arterialized hand vein. When target glucose levels had been maintained for 2 h or at 14:00 h (whichever came first), FBF responses to ACh, in the absence and presence of L-NMMA, and SNP were assessed again as above. Heart rate was continuously monitored throughout the procedure through a five-lead electrocardiogram.
Analyses Procedures followed for the analyses of blood flow responses and glucose and hormone concentrations have been described previously.12 Blood flow analyses were done at maximum doses of ACh (8 µg/100 mL/min) and SNP (2 µg/100 mL/min) and were saline-corrected. For C-peptide, glucagon and insulin, the mean of the first four values is summarized as ambient and the mean of the last four values as euglycaemia.
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
422
Diabetes & Vascular Disease Research 11(6)
Statistical analyses Statistical analyses of all blood flow and efficacy endpoints were performed using an analysis of covariance (ANCOVA) model, with treatment as fixed effect and baseline value as a covariate on the full analysis set (i.e. all randomized participants). Two-sided 95% confidence intervals were calculated, and the significance level was set at 5%. The primary efficacy analysis compared the effect of liraglutide versus placebo on ACh-mediated FBF euglycaemia; comparisons between liraglutide versus glimepiride and glimepiride versus placebo were secondary. Post hoc analyses included a summary of the percentage of participants with maximal FBF response at maximal dose of ACh and SNP, while changes from baseline to Week 12 in the integrated area under the curve (AUC) of ACh-mediated FBF at euglycaemia were assessed using the same ANCOVA model used for the primary and secondary endpoints. Safety analyses were performed on data from the safety population, which included all participants exposed to at least one dose of the drug or who underwent at least one venous occlusion plethysmography procedure. Hypoglycaemic events were analysed using Fisher’s exact test. AEs, laboratory parameters, vital signs and physical examinations were compared between groups with descriptive statistics.
Power calculation The power calculation was based on the two-sample t-test with equal variances, as described previously.12 With 14 subjects in each treatment group, there was an 80% power to detect a difference of 1.1 standard deviation unit (twosided α = 0.05).
Results Participant disposition, demographics and baseline characteristics A total of 49 participants were enrolled and randomized to receive liraglutide (n = 16), placebo (n = 16) or glimepiride (n = 17). Of these, 46 participants (94%) completed the study and 3 participants withdrew: 2 in the placebo group and 1 in the glimepiride group (Figure 2). Demographic and baseline characteristics were generally similar between groups, with the exception of age and duration of diabetes, which were slightly greater in the placebo arm compared with the liraglutide or glimepiride arm (Table 1).
FBF responses to ACh at euglycaemia After 12 weeks, ACh-mediated FBF at euglycaemia increased from baseline in the liraglutide and glimepiride arms and decreased in the placebo arm (Figure 3(a)). After 12 weeks, least square (LS) mean changes (LS
mean ± standard error (SE)) were 4.2 ± 2.6, −3.2 ± −2.8 and 2.2 ± 2.6 mL/100 mL/min in the liraglutide, placebo and glimepiride arms, respectively. The between-treatment differences were not statistically significant for liraglutide versus placebo (p = 0.055), liraglutide versus glimepiride (p = 0.568) and glimepiride versus placebo (p = 0.167) (Figure 3(a)). Review of a scatterplot of the individual data was also not suggestive of an effect (Figure 3(b)). There was no significant correlation (r = −0.211, p = 0.1602) between these changes in ACh-mediated FBF at euglycaemia and changes in HbA1c from baseline to Week 12 (data not shown). At baseline, ACh-mediated FBF was similar before and after L-NMMA infusion in all three treatment groups. At Week 12, ACh-mediated FBF after L-NMMA infusion was suppressed in the liraglutide and glimepiride arms but not in the placebo arm compared with the values observed before L-NMMA infusion. Thus, compared with baseline, L-NMMA inhibition of ACh-mediated FBF at Week 12 increased with liraglutide and glimepiride treatment and decreased with placebo. However, between-treatment differences were not significant (Figure 4).
FBF responses to SNP at euglycaemia After 12 weeks of treatment, mean changes from baseline (LS mean ± SE) were 3.5 ± 2.7, −1.0 ± 2.9 and 2.7 ± 2.7 mL/100 mL/min with liraglutide, placebo and glimepiride, respectively. Between-treatment differences were not statistically significant (liraglutide vs placebo: p = 0.265; liraglutide vs glimepiride: p = 0.852 and glimepiride vs placebo: p = 0.344; Figure 5).
Glucose, insulin, C-peptide and glucagon measurements during the venous occlusion plethysmography procedures FPG, C-peptide, glucagon and insulin levels at baseline and Week 12 at euglycaemia are summarized in Table 2. Target blood glucose values were achieved with the insulin infusion at both Week 0 and Week 12. Mean blood glucose values during the euglycaemic clamp did not differ between treatment groups at Week 0 (95.4 vs 95.8 vs 97.2 mg/dL, liraglutide vs placebo vs glimepiride) and Week 12 (94.3 vs 96.1 vs 94.1 mg/dL, liraglutide vs placebo vs glimepiride).
Glycaemic control Change in HbA1c from baseline to Week 12 was −0.63 ± 0.44%, −0.09 ± 0.45% and −0.55 ± 0.45% with liraglutide, placebo and glimepiride treatment, respectively. Between-treatment differences were significant for liraglutide versus placebo (p = 0.002) and glimepiride versus placebo (p = 0.010). The goal of similar glycaemic control between the liraglutide and glimepiride arms was achieved
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
423
Nandy et al.
Enrollment
Assessed for eligibility (n=108)
Randomized (n=49)
Allocation Allocate to liraglutide (n=16) • Received liraglutide (n=16)
Allocate to placebo (n=16) • Received placebo (n=16)
Allocate to glimepiride (n=17) • Received glimepiride (n=17)
Follow-up
Lost to follow-up (n=0) Discontinued intervention (n=0)
Lost to follow-up (n=0) Discontinued intervention (n=2) due to: • AE (back pain, n=1) • Other (unknown, n=1)
Lost to follow-up (n=0) Discontinued intervention (n=1) due to inability to perform VOP at EOS
Analysis
Analyzed (n=16) • Excluded from analysis (n=0)
Analyzed (n=14) • Excluded from analysis (n=2) due to lack of results for primary endpoint
Analyzed (n=16) • Excluded from analysis (n=1) due to lack of results for primary endpoint
Figure 2. CONSORT diagram showing flow of patients during the study. VOP: venous occlusion plethysmography; AE: adverse event; EOS: end of study.
(p = 0.6207). Compared with placebo, significant improvements in FPG were also observed in the liraglutide and glimepiride arms (Appendix Table 3).
Body weight
pressure (LS mean ± SE) from baseline to Week 12 was 3.0 ± 2.0, −2.0 ± 2.1 and 0.8 ± 1.9 mmHg in the liraglutide, placebo and glimepiride arms, respectively. Betweentreatment differences for systolic and diastolic blood pressure were not statistically significant.
After 12 weeks of treatment, body weight decreased from baseline in the liraglutide and placebo groups and increased in the glimepiride group. Between-treatment differences were significant for liraglutide versus placebo, liraglutide versus glimepiride and glimepiride versus placebo (Appendix Table 4).
Lipid profile and biomarkers of CV risk
Blood pressure
Safety and tolerability
Change in systolic blood pressure (LS mean ± SE) from baseline to Week 12 was −4.6 ± 2.4, −4.8 ± 2.6 and −0.2 ± 2.4 mmHg in the liraglutide, placebo and glimepiride treatment arms, respectively. Change in diastolic blood
AEs. There was a single TESAE (prostate cancer), which occurred in the placebo group. The percentage of patients who reported TEAEs was greater with liraglutide (68.8%) and placebo (56.3%) than with glimepiride (29.4%).
Change from baseline to Week 12 in lipid profiles and biomarkers of CV risk is summarized in the appendix (Appendix Tables 5 and 6, respectively). Betweentreatment differences were not significant.
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
424
Diabetes & Vascular Disease Research 11(6)
Table 1. Summary of demographic and selected baseline characteristics.
Male sex (%) Age (years) Race (% White) Weight (kg) BMI (kg/m2) Duration of diabetes (years) Previous treatment Diet and exercise Metformin ACh-mediated FBF at ambient glucose (mL/100 mL/min) SNP-mediated FBF at ambient glucose (mL/100 mL/min) ACh-mediated FBF at euglycaemia (mL/100 mL/min) SNP-mediated FBF at euglycaemia (mL/100 mL/min) HbA1c (%) FPG (mmol/L)
Liraglutide (n = 16)
Placebo (n = 16)
Glimepiride (n = 17)
10 (62.5) 57.7 (9.0) 16 (100.0) 95.1 (13.1) 32.7 (4.5) 5.3 (4.1)
10 (62.5) 60.3 (7.3) 16 (100.0) 90.6 (13.5) 31.6 (4.2) 8.4 (4.6)
2 (12.5) 14 (87.5) 7.9 (5.7) 11.0 (4.5) 8.5 (6.7) 10.2 (4.4) 7.2 (0.5) 8.9 (1.7)
1 (6.3) 15 (93.8) 8.7 (5.8) 11 (5.0) 10.0 (6.6) 12.0 (6.8) 7.0 (0.5) 8.1 (1.1)
11 (64.7) 57.7 (5.3) 17 (100.0) 92.0 (14.0) 31.1 (4.9) 6.8 (8.1) 2 (11.8) 15 (88.2) 5.9 (4.4) 11.1 (5.2) 7.9 (7.6) 10.7 (5.1) 7.3 (0.5) 9.1 (1.9)
ACh: acetylcholine; BMI: body mass index; FBF: forearm blood flow; FPG: fasting plasma glucose; SD: standard deviation; SNP: sodium nitroprusside; HbA1c: glycated haemoglobin. Sex, race and previous treatment are number of patients (%); all other parameters are mean (SD).
(a)
P=0.568
Change in ACh-mediated FBF at euglycemia (mL/100 mL/min)
20.0
P=0.167
P=0.055a
15.0 10.0 5.0 4.2
0.0
2.2 –3.2
–5.0
–10.0 –15.0
Liraglutide 1.8 mg/day Placebo Glimepiride 4 mg/day
Change in ACh-mediated FBF at euglycemia (mL/100 mL/min)
(b) 60 50
49.0
40 30 20
29.7 19.2
16.4
10 0 –10 –20 –30
–17.7
Liraglutide 1.8 mg/day
–19.3
Placebo
–18.4
Glimepiride 4 mg/day
Figure 3. (a) Change in ACh-mediated FBF (LS mean ± SE) at euglycaemia in the liraglutide, placebo and glimepiride arms. (b) Scatterplot of individual ACh-mediated FBF data. ACh: acetylcholine; FBF: forearm blood flow; LS: least square; SE: standard error.
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
425
Difference in change from baseline to Week 12 in ACh-mediated FBF (mL/100 mL/min) in the absence and presence of L-NMMA
Nandy et al. P=0.634 P=0.149
15.0
P=0.320
10.0
5.0 3.7
2.3
–0.9
0.0
–5.0
–10.0
–15.0 Liraglutide 1.8 mg/day
Placebo
Glimepiride 4 mg/day
Figure 4. Difference in change from baseline to Week 12 in ACh-mediated FBF inhibition before and after L-NMMA infusion (LS mean ± SE). ACh: acetylcholine; FBF: forearm blood flow; L-NMMA: l-NG-monomethyl arginine; LS: least square; SE: standard error.
P=0.852
20.0 P=0.265
P=0.344
Change in SNP-mediated FBF at euglycemia (mL/100 mL/min)
15.0 10.0 5.0 3.5 0.0
–1.0
2.7
–5.0 –10.0 –15.0 Liraglutide 1.8 mg/day
Placebo
Glimepiride 4 mg/day
Figure 5. Change in SNP-mediated FBF (LS mean ± SE) at euglycaemia in the liraglutide, placebo and glimepiride arms. SNP: sodium nitroprusside; FBF: forearm blood flow; LS: least square; SE: standard error.
TEAEs were either mild (95%) or moderate (5%) in severity, and most (80%) were deemed by the investigator to be unlikely related to treatment. Only eight events, all of which occurred in the liraglutide group, were considered possibly related to treatment. Of these, most were of gastrointestinal nature (two events of dyspepsia, two nausea, two vomiting), one was a viral gastroenteritis and the remaining one was decreased appetite. No change in laboratory parameters, vital signs and physical examinations was observed (section ‘Laboratory parameters, vital signs
and physical examinations’ of Appendix 1). No withdrawals occurred due to AEs in patients treated with liraglutide. Hypoglycaemia. There were no episodes of major hypoglycaemia. The rate of minor hypoglycaemia was lower in the liraglutide group than in the glimepiride group (0.26 vs 2.46 events/patient/year). No events of minor hypoglycaemia were reported in the placebo group. Comparisons of the number of subjects experiencing hypoglycaemia between
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
426
Diabetes & Vascular Disease Research 11(6)
Table 2. Plasma glucose, C-peptide, glucagon and insulin levels at baseline and Week 12 at euglycaemia during the clamp and venous occlusion plethysmography measurements.
Glucose (mg/dL) Baseline Week 12 C-peptide (nmol/L) Baseline Week 12 Glucagon (pg/mL) Baseline Week 12 Insulin (µU/mL) Baseline Week 12
Liraglutide
Placebo
95.4 (8.7) 94.3 (4.3)
95.8 (4.3) 96.1 (6.1)
0.5 (0.2) 0.8 (0.3)
0.5 (0.2) 0.5 (0.1)
89.8 (29.1) 80.6 (26.6)
92.6 (25.7) 96.2 (26.0)
19.8 (17.0) 12.8 (6.9)
14.8 (7.5) 17.3 (13.9)
Glimepiride 97.2 (5.2) 94.1 (5.7) 0.4 (0.2) 0.5 (0.2) 86.9 (35.0) 75.4 (20.2) 19.5 (16.5) 14.0 (8.8)
SD: standard deviation. Data are mean (SD).
the liraglutide and placebo groups (p = 0.484) and between the liraglutide and glimepiride groups (p = 0.398) were not statistically significant.
Post hoc analyses Proportion of subjects with peak FBF response at maximal dose of ACh and SNP at euglycaemia. Peak FBF response at maximal dose of ACh was observed in 55% of patients at baseline and in 61% of patients at Week 12. In contrast, peak FBF response was seen at the maximal dose of SNP in 76% of patients at baseline and in 80% of patients at Week 12. AUC of ACh-mediated FBF at euglycaemia. After 12 weeks of treatment, change from baseline in AUC of ACh-mediated FBF (LS mean ± SE) at euglycaemia was 17.5 ± 12.3, −10.6 ± 13.3 and 0.8 ± 12.3 mL/100 mL with liraglutide, placebo and glimepiride, respectively. Between-treatment differences were not statistically significant.
Discussion In this study, no significant effect of liraglutide on endothelium-dependent vasodilation was seen. Although the p-value approached statistical significance (p = 0.055), further analysis of the distribution of the individual data did not suggest a treatment effect. There was also no effect of liraglutide on endothelium-independent vasodilation. Given that hyperglycaemia has a detrimental effect on endothelium-dependent vasodilation,3,24 our study was designed to control for acute differences in glycaemia. Euglycaemia was achieved with the insulin infusion, and the goal of achieving similar acute glycaemic control was fulfilled, with the mean blood glucose values being similar between treatment arms during the FBF assessments at both baseline and end of study.
The study was also designed to determine whether the presumed effect of liraglutide was independent of its effect on glucose. The goal of attaining a similar reduction in HbA1c with liraglutide and glimepiride was also achieved. Although the change in ACh-mediated vasodilation was numerically greater with liraglutide than glimepiride, the between-treatment difference was not statistically significant. Importantly, however, the study was not powered to detect differences in blood flow between liraglutide and glimepiride. In our study, the changes in ACh-mediated FBF and HbA1c were not correlated. These findings suggest, but do not prove, that the observed effect of liraglutide, although not statistically significantly different from placebo, is independent of its effect on glycaemic control. We chose to analyse FBF at maximum doses of ACh and SNP because this usually resulted in the maximum response in previous studies in healthy, non-diabetic individuals (A Basu, unpublished observations).3 Intriguingly, the proportion of subjects who had their peak response at the maximum dose of ACh was somewhat lower in this study than what has been observed in prior studies in people without diabetes, suggesting that the nitric oxide pathway may be impaired in patients with type 2 diabetes. Because of this finding and given that the p-value approached significance, a post hoc AUC analysis, taking into account all ACh doses, was also done. Between-treatment differences were not significant in this post hoc analysis. Our findings are consistent with some prior studies but conflict with others. Improved vascular function has been reported following native GLP-1 infusion in preclinical studies25,26 and also in clinical trials in both healthy subjects12 and patients with type 2 diabetes.13 However, it is not yet clear whether native GLP-1 exerts this beneficial effect through direct actions by the GLP-1 receptor27 or through indirect actions of its dipeptidyl peptidase (DPP)-4– mediated cleavage product, GLP-1(9–36), known to mediate important mitochondrial and other, as yet unidentified, GLP-1 receptor–independent signalling pathways.28,29 For incretin-based therapies, including GLP-1 RAs and DPP-4 inhibitors, the overall effect on vascular function is less clear. For GLP-1 RAs, a recent mechanistic study in mice that investigated the role of GLP-1 receptor activation in blood pressure control found that liraglutide had no direct effect on phosphorylated eNOS.30 Moreover, 3 months of exenatide therapy was not found to have an effect on endothelial function in patients with prediabetes and abdominal obesity, as measured by digital reactive hyperaemia, C-reactive protein, circulating oxidized lowdensity lipoprotein and vascular cell adhesion molecule-1.31 Another study examining the effect of 6 months of GLP-1 RA therapy in obese patients with type 2 diabetes found no effect on endothelial function, as measured by brachial flow–mediated dilation and carotid intimal– medial thickness.32 The low vascular impact of GLP-1 RAs described in the above studies could suggest that improved vascular
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
427
Nandy et al. function is more likely to be mediated by GLP-1 receptor– independent mechanisms, such as GLP-1(9–36) signalling. However, other studies provide evidence that GLP-1 RAs can directly improve endothelial function. An exploratory study in patients with type 2 diabetes who were in good metabolic control on metformin monotherapy found that liraglutide improved the retinal hyperaemic response to flicker light after 6 and 12 weeks of therapy.23 Another study with exenatide supports the notion that, as shown for GLP-1,13 GLP-1 RAs may also have a positive effect on endothelial function. In this observational study in patients with type 2 diabetes, endothelial function was significantly improved after 16 weeks of treatment with exenatide compared with glimepiride. Exenatide treatment and HbA1c were both independently and significantly associated with peak flow–mediated dilation. In addition, the reactivity of peripheral resistance vessels, which was indirectly evaluated through measurements of shear rate increase after ischaemia, also improved significantly with exenatide treatment. However, the study was limited by both a small number of subjects (n = 20) who were not matched in the two treatment groups and the lack of a randomized doubleblind design.33 For DPP-4 inhibitors, conflicting data have also been published regarding their effects on the vascular system. One recent study reports that 6 weeks of sitagliptin or alogliptin treatment significantly attenuated endothelial function in patients with type 2 diabetes,34 which is in support of the hypothesis that normal degradation products of native GLP-1 such as GLP-1(9–36) may be mediating vascular improvements through GLP-1 receptor–independent mechanisms. However, it has also been shown that both 12 weeks of sitagliptin treatment and 4 weeks of vildagliptin treatment resulted in significant improvement in endothelial function in patients with type 2 diabetes,35,36 possibly due to an increase in circulating levels of GLP-1, which may increase GLP-1 receptor–dependent responses, not only on vascular endothelial cells but also on myocardial and smooth muscle cells, where the GLP-1 receptor is also expressed. In addition, it has been reported that the GLP-1 RA, exenatide, when given as a single dose prior to a high-fat meal, improved postprandial endothelial function in patients with impaired glucose tolerance or recent-onset type 2 diabetes. The observed response in these patients suggests that the effect of GLP-1 RAs on endothelial function may occur in the postprandial period, possibly due to a reduction in postprandial triglycerides,37 which is supported by another study in which liraglutide was also shown to reduce postprandial triglycerides.38 This may also be the case for DPP-4 inhibitors, as shown by a recent study in patients without diabetes, in which 1 week of alogliptin treatment ameliorated postprandial endothelial dysfunction.39 In contrast to the above studies, in our study, we assessed endothelial function and lipids in the fasting state.
In the majority of the above studies, blood flow was assessed by responses in the brachial artery utilizing flow-mediated dilation, in contrast to this study, and our previously published study,12 which assessed blood flow in the regional microcirculation using venous occlusion plethysmography. Given that mechanisms of vasodilation in conduit vessels are often different from those in resistance vessels, where blood flow is regulated, it is worth noting that flow-mediated dilation assesses vasodilation in large conduit vessels, while venous occlusion plethysmography measures vasodilation in small resistance vessels and is considered a more robust method for measuring vascular activity.5 This methodological difference makes direct comparisons difficult between our findings and those of others. There were several limitations to our study. First, we did not use a somatostatin clamp to inhibit endogenous secretion of insulin and C-peptide (so as to match insulin and C-peptide levels between groups during the venous occlusion plethysmography studies), both of which can affect endothelial function.40 Second, individuals in the glimepiride treatment arm were not blinded to treatment. The fact that the study was not sufficiently powered to detect a statistically significant difference between liraglutide and glimepiride groups was also a limitation, as was the large variability present in our measurements, most likely due to the inherently increased metabolic variability of subjects with endocrine dysfunctions such as type 2 diabetes. Finally, since the FBF endpoints at euglycaemia were determined after the exploratory FBF endpoints at ambient blood glucose, sequence bias cannot be ruled out. In summary, 12 weeks of liraglutide therapy did not have a significant effect on endothelium-dependent vasodilation in this study. Further investigation with liraglutide, including assessment of endothelial function in other vascular beds and during the postprandial period, is warranted.
Key messages •• This short trial assessed how the human glucagonlike peptide-1 (GLP-1) analogue liraglutide impacted endothelial function in adult patients with type 2 diabetes and no overt cardiovascular disease. •• Liraglutide did not significantly impact endothelium-dependent vasodilation after 12 weeks. •• Additional investigations into liraglutide and endothelial function in alternative vasculature and during the postprandial period are needed. Acknowledgements The authors are grateful to Vatsala Karwe (employed at Novo Nordisk Inc. at the time of article initiation) and Anjun Cao (currently employed by Novo Nordisk) for statistical advice and input. The authors also thank Dr Rosalyn Ferguson, PhD, Watermeadow Medical (Witney, UK), supported by Novo
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
428
Diabetes & Vascular Disease Research 11(6)
Nordisk, for medical writing and editorial assistance. Clinical trial registration number: NCT00620282.
Declaration of conflicting interests D.N., C.J., R.B. and M.J. have nothing to disclosure. J.B. and C.S. are full-time employees and shareholders in Novo Nordisk.
Funding Novo Nordisk funded the trial and supported the writing assistance for the preparation of this article. A.B. was principal investigator of this trial, funded by Novo Nordisk. This article’s preparation was funded by Novo Nordisk.
References 1. Libby P. Inflammation in atherosclerosis. Nature 2002; 420: 868–874. 2. Widlansky ME, Gokce N, Keaney JF Jr, et al. The clinical implications of endothelial dysfunction. J Am Coll Cardiol 2003; 42: 1149–1160. 3. Hadi HAR and Al Suwaidi J. Endothelial dysfunction in diabetes mellitus. Vasc Health Risk Manag 2007; 3: 853–876. 4. Haffner SM, Lehto S, Rönnemaa T, et al. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med 1998; 339: 229–234. 5. Eskurza I, Seals DR, DeSouza CA, et al. Pharmacologic versus flow-mediated assessments of peripheral vascular endothelial vasodilatory function in humans. Am J Cardiol 2001; 88: 1067–1069. 6. Tagawa T, Mohri M, Tagawa H, et al. Role of nitric oxide in substance P-induced vasodilation differs between the coronary and forearm circulation in humans. J Cardiovasc Pharmacol 1997; 29: 546–553. 7. Hadi HAR, Carr CS and Al Suwaidi J. Endothelial dysfunction: cardiovascular risk factors, therapy, and outcome. Vasc Health Risk Manag 2005; 1: 183–198. 8. Perticone F, Ceravolo R, Pujia A, et al. Prognostic significance of endothelial dysfunction in hypertensive patients. Circulation 2001; 104: 191–196. 9. Heitzer T, Schlinzig T, Krohn K, et al. Endothelial dysfunction, oxidative stress, and risk of cardiovascular events in patients with coronary artery disease. Circulation 2001; 104: 2673–2678. 10. Green DJ, Jones H, Thijssen D, et al. Flow-mediated dilation and cardiovascular event prediction: does nitric oxide matter? Hypertension 2011; 57: 363–369. 11. Drucker DJ. The biology of incretin hormones. Cell Metab 2006; 3: 153–165. 12. Basu A, Charkoudian N, Schrage W, et al. Beneficial effects of GLP-1 on endothelial function in humans: dampening by glyburide but not by glimepiride. Am J Physiol Endocrinol Metab 2007; 293: E1289–E1295. 13. Nyström T, Gutniak MK, Zhang Q, et al. Effects of glucagon-like peptide-1 on endothelial function in type 2 diabetes patients with stable coronary artery disease. Am J Physiol Endocrinol Metab 2004; 287: E1209–E1215.
14. Vilsbøll T, Agersø H, Krarup T, et al. Similar elimination rates of glucagon-like peptide-1 in obese type 2 diabetic patients and healthy subjects. J Clin Endocrinol Metab 2003; 88: 220–224. 15. Cho YM, Merchant CE and Kieffer TJ. Targeting the glucagon receptor family for diabetes and obesity therapy. Pharmacol Ther 2012; 135: 247–278. 16. Montanya E. A comparison of currently available GLP-1 receptor agonists for the treatment of type 2 diabetes. Expert Opin Pharmacother 2012; 13: 1451–1467. 17. Elkinson S and Keating GM. Lixisenatide: first global approval. Drugs 2013; 73: 383–391. 18. Poole RM and Nowlan ML. Albiglutide: first global approval. Drugs 2014; 74: 929–938. 19. Arakawa M, Mita T, Azuma K, et al. Inhibition of monocyte adhesion to endothelial cells and attenuation of atherosclerotic lesion by a glucagon-like peptide-1 receptor agonist, exendin-4. Diabetes 2010; 59: 1030–1037. 20. Liu H, Dear AE, Knudsen LB, et al. A long-acting glucagonlike peptide-1 analogue attenuates induction of plasminogen activator inhibitor type-1 and vascular adhesion molecules. J Endocrinol 2009; 201: 59–66. 21. Hattori Y, Jojima T, Tomizawa A, et al. A glucagon-like peptide-1 (GLP-1) analogue, liraglutide, upregulates nitric oxide production and exerts anti-inflammatory action in endothelial cells. Diabetologia 2010; 53: 2256–2263. 22. Gaspari T, Liu H, Welungoda I, et al. A GLP-1 receptor agonist liraglutide inhibits endothelial cell dysfunction and vascular adhesion molecule expression in an ApoE-/- mouse model. Diab Vasc Dis Res 2011; 8: 117–124. 23. Forst T, Michelson G, Ratter F, et al. Addition of liraglutide in patients with Type 2 diabetes well controlled on metformin monotherapy improves several markers of vascular function. Diabet Med 2012; 29: 1115–1118. 24. Ding Y, Vaziri ND, Coulson R, et al. Effects of simulated hyperglycemia, insulin, and glucagon on endothelial nitric oxide synthase expression. Am J Physiol Endocrinol Metab 2000; 279: E11–E17. 25. Chai W, Dong Z, Wang N, et al. Glucagon-like peptide 1 recruits microvasculature and increases glucose use in muscle via a nitric oxide-dependent mechanism. Diabetes 2012; 61: 888–896. 26. Nikolaidis LA, Elahi D, Hentosz T, et al. Recombinant glucagon-like peptide-1 increases myocardial glucose uptake and improves left ventricular performance in conscious dogs with pacing-induced dilated cardiomyopathy. Circulation 2004; 110: 955–961. 27. Cabou C, Vachoux C, Campistron G, et al. Brain GLP-1 signaling regulates femoral artery blood flow and insulin sensitivity through hypothalamic PKC-δ. Diabetes 2011; 60: 2245–2256. 28. Ussher JR and Drucker DJ. Cardiovascular actions of incretin-based therapies. Circ Res 2014; 114: 1788–1803. 29. Ban K, Kim KH, Cho CK, et al. Glucagon-like peptide (GLP)-1(9-36) amide-mediated cytoprotection is blocked by exendin(9-39) yet does not require the known GLP-1 receptor. Endocrinology 2010; 151: 1520–1531. 30. Kim M, Platt MJ, Shibasaki T, et al. GLP-1 receptor activation and Epac2 link atrial natriuretic peptide secretion to control of blood pressure. Nat Med 2013; 19: 567–575.
Downloaded from dvr.sagepub.com at Lucia Campus Library on June 20, 2015
429
Nandy et al. 31. Kelly AS, Bergenstal RM, Gonzalez-Campoy JM, et al. Effects of exenatide vs. metformin on endothelial function in obese patients with pre-diabetes: a randomized trial. Cardiovasc Diabetol 2012; 11: 64. 32. Hopkins ND, Cuthbertson DJ, Kemp GJ, et al. Effects of 6 months glucagon-like peptide-1 receptor agonist treatment on endothelial function in type 2 diabetes mellitus patients. Diabetes Obes Metab 2013; 15: 770–773. 33. Irace C, De Luca S, Shehaj E, et al. Exenatide improves endothelial function assessed by flow mediated dilation technique in subjects with type 2 diabetes: results from an observational research. Diab Vasc Dis Res 2013; 10: 72–77. 34. Ayaori M, Iwakami N, Uto-Kondo H, et al. Dipeptidyl peptidase-4 inhibitors attenuate endothelial function as evaluated by flow-mediated vasodilatation in type 2 diabetic patients. J Am Heart Assoc 2013; 2: e003277. 35. Kubota Y, Miyamoto M, Takagi G, et al. The dipeptidyl peptidase-4 inhibitor sitagliptin improves vascular endothelial function in type 2 diabetes. J Korean Med Sci 2012; 27: 1364–1370. 36. Van Poppel PC, Netea MG, Smits P, et al. Vildagliptin improves endothelium-dependent vasodilatation in type 2 diabetes. Diabetes Care 2011; 34: 2072–2077. 37. Koska J, Schwartz EA, Mullin MP, et al. Improvement of postprandial endothelial function after a single dose of exenatide in individuals with impaired glucose tolerance and recent-onset type 2 diabetes. Diabetes Care 2010; 33: 1028–1030. 38. Hermansen K, Baekdal TA, Düring M, et al. Liraglutide suppresses postprandial triglyceride and apolipoprotein B48 elevations after a fat-rich meal in patients with type 2 diabetes: a randomized, double-blind, placebo-controlled, cross-over trial. Diabetes Obes Metab 2013; 15: 1040–1048. 39. Noda Y, Miyoshi T, Oe H, et al. Alogliptin ameliorates postprandial lipemia and postprandial endothelial dysfunction in non-diabetic subjects: a preliminary report. Cardiovasc Diabetol 2013; 12: 8.
40. Ardigo D, Franzini L, Valtuena S, et al. Relation of plasma insulin levels to forearm flow-mediated dilatation in healthy volunteers. Am J Cardiol 2006; 97: 1250–1254.
Appendix 1 Hypoglycaemia definitions Hypoglycaemia in this trial was defined as either major hypoglycaemia – defined as an episode with severe central nervous system symptoms consistent with hypoglycaemia, in which the subject was unable to treat himself/ herself and had one of the following characteristics: (1) reversal of symptoms after either food intake or glucagon/intravenous glucose administration or (2) plasma glucose
