The “ACute and Chronic Effects of Saxagliptin” (ACCES) study is a randomized, placebo-controlled, double-blind, controlled phase 2, pilot study conducted at Jean Verdier University Hospital (Bondy, France). The study was approved both by the local ethics committee (Comité de Protection des Personnes, Reference Number: JLD/AP-protocole 27-2011) and the French National Agency for Drug Security (Agence Nationale de Sécurité du Médicament, Number: A110912-14) and was registered as a clinical trial (NumberEudraCT: NCT01521312).
There were two parts in this study. The first one explored acute and long-term effects of saxagliptin vs placebo on glucose tolerance (ACCES-Glucose study), the results of which will be reported elsewhere. The second part explored the effects of saxagliptin vs placebo on vascular parameters (ACCES-Vasc study) which are shown in the present article.
The inclusion criteria were the following: patients with healthcare insurance, diagnosed with IGT according to an oral glucose tolerance test (OGTT) performed less than 3 weeks before the inclusion, age between 18 years and 70 years, body mass index ranging from 30 to 40 kg/m². Exclusion criteria were the following: pregnancy or absence of contraception in women of childbearing age, breastfeeding women, known diabetes (defined as fasting plasma glucose (FPG) ≥ 7 mmol/L or 2 hours after OGTT ≥ 11.1 mmol/L) (21), uncontrolled hypertension (defined as blood pressure (BP) ≥ 160/110 mmHg), dyslipidemia (defined as serum total cholesterol > 6.5 mmol/l, or triglycerides > 2.3 mmol/l), renal failure (creatinine clearance < 60 ml/min), hepatic failure (prothrombin time < 70%), chronic respiratory disease, anemia (hemoglobin level < 10 g/dl), cardiac failure (from grade 2 of the NYHA classification), lower limb arterial disease (intermittent claudication, absence of lower limb peripheral pulses or ankle brachial pressure index < 0.8), cardiac arrhythmia, anti-hypertensive or current lipid-lowering treatment initiated after the diagnosis of IGT, hypersensitivity to saxagliptin or any excipients (lactose monohydrate, microcrystalline cellulose (E460i), croscarmellose sodium (E468), magnesium stearate), and patients who had previously experienced a severe hypersensitivity reaction to a DPP4i. Smoking was possible and defined as present if patient continued to smoke or had stopped for < 3 years ago.
The methods used were non-invasive. All patients gave us their agreement and signed the informed consent according to the European directives.
The anticipated effect size (∆/σ) for our study was estimated at 1.1, the power level was 80% and the probability level was set at 0.05. The estimated minimum sample size for a two-tailed hypothesis was calculated at 12 persons per group.
All participants were admitted for routine checkup for obesity to the research department of Endocrinology Diabetology and Nutrition. They were free of any acute disease at the time of screening for hyperglycemia. The OGTT was performed within one to three weeks before inclusion in all subjects with 75 g of carbohydrates after night’s fasting of 12 hours. IGT was defined as 2-hour post-OGTT plasma glucose value ≥ 7.8 mmol/L and < 11.1mmol/L, with a FPG value < 7 mmol/L (21). Participants received dietary advices during the pre-inclusion period.
From September 2012 to September 2014, we included 24 subjects. The investigations were non-invasive and performed twice: first at inclusion (Visit 1) and second, 12 weeks after inclusion (Visit 2). Also, women of childbearing age had a plasma blood pregnancy test the day before investigations to rule out any current pregnancy. Subjects were randomized to treatment with saxagliptin or placebo for 12 weeks, and treatment attribution was double blinded.
Smokers were asked not to smoke before and during the visits. At Visit 1, saxagliptin or placebo treatment was taken after the fasting blood sampling, 30 minutes before eating the standardized breakfast. Thus, the measurements at fasting during visit 1 were performed in both groups before taking the experimental treatment. During the rest of the study the experimental treatment was taken immediately before breakfast.
At Visit 1 and 2, the measurements were conducted at an ambient temperature of 22 °C to 24°C in a room where noise and light were kept to a minimum. Participants remained at rest in the supine position throughout the tests. All cardiovascular tests were non-invasive and were performed at fasting and 60 and 120 minutes after breakfast.
Participants were contacted for follow-up by telephone to note any side effect. Visit 2 followed the same study protocol as Visit 1. Patients were advised not to change their daily physical activity during the study period. Urinary pregnancy tests were performed at the hospital every month and at any time during the study in case of late menstruation. Visit 3 (three to five days after Visit 2) included measurements of glucose levels before and 2 hours after OGTT and HbA1c. The participation in the study stopped after that visit (Figure 1).
Procedures during Visit 1 and Visit 2
- Standardized breakfast
As previously reported (22) this standardized breakfast contained 75 g of carbohydrates. It consisted of unsweetened coffee or tea (200 ml), 200 ml of orange juice and 60 g of white bread with 30 g of jam and was offered to the patient who was asked to eat it in 10 minutes.
- Biological measurements
The biological assessments were performed at fasting and after the standardized breakfast (Figure 1). Plasma glucose was assessed by the oxidase method (colourimetry, Kone Optima, Thermolab System, Paris La Défense, France) at fasting and one, two and three hours after the breakfast, and HbA1c was measured (agarose electrophoresis; HYRYS, HYDRASYS, Sebia, Evry, France). Serum samples were centrifuged at − 4 °C and stored at − 80 °C. These samples were used for an enzyme-linked immunosorbent assay to measure nitrotyrosin (Cell Biolabs, San Diego, CA, USA), E-selectin (R&D Systems, Abingdon, UK) and vascular cell adhesion molecule-1 (VCAM-1; R&D Systems, Abingdon, UK) according to the manufacturer's instructions. Laser immunonephelometry (BN100; Dade-Behring) was used to measure the UAER from urine specimens.
- Microcirculatory cutaneous blood flow measurement
Cutaneous blood flow (CBF, expressed as perfusion units) was measured at baseline for 6 min and during 1 min of paced breathing (6 breaths/min) on the forearm, using laser Doppler flowmetry (Periflux System 5000®, PERIMED, Sweden). CBF changes during the 1 min of paced breathing (expressed as %) reflect microvascular autonomic activity, with CBF decreasing during inhalation due to sympathetic nervous system activation. The percentage decrease in the signal wavelet from the zenith to nadir was calculated, and the mean decrease during the three last breaths was used as the index of microvascular sympathetic activity as previously described (22,23).
Changes in CBF after transcutaneous administration of acetylcholine (Ach: Miochol-E®, 20mg of Ach chloride/vial; CHAUVIN®, topical application of 4 mg) through iontophoresis were also measured to explore endothelial function. Ach chloride was diluted in 1 mL of sterile water, after which 200 µL of this solution was placed into a special chamber of a sponge that was used for skin application. After Ach administration, we measured (i) maximal CBF, (ii) delay from Ach administration to maximal CBF (peak delay) and (iii) CBF incremental area under the curve 10 min after Ach administration, which is considered an index of microvascular endothelial function (22).
- Arterial endothelial function
Arterial endothelial function was assessed in one of the digital arteries by recording finger arterial pulsatile volume changes for 15 min using the Endopat2000® system (Itamar Medical®, Israel), which uses plethysmographic biosensors. Reactive hyperhemia was measured after occluding brachial artery blood flow for 5 min. The reactive hyperhemia index (RHI) represents the post-to-pre occlusion signal ratio on the occluded side normalized to the control side (24).
- Arterial stiffness and central BP
Pulse wave velocity (PWV) was assessed in the supine position (25). BP was measured just before PWV measurements with an Omron 750 IT BP device. Briefly, PWV was measured as the arterial distance divided by the transit time from the carotid to the femoral artery. The arterial distance was assessed from surface measurements subtracting the sternal notch femoral distance to the carotid-sternal notch distance. Transit time was automatically calculated from the foot of carotid and femoral pulses measured successively by applanation tonometry using R-wave of an ECG as a timing reference (Sphygmocor®, Atcor Medical, Australia). Only measurements with a beat-to-beat transit time standard deviation below 5% were considered of good quality and used in the study. Using the same device, we also measured (i) central BP to calculate pulse pressure (systolic BP (SBP) – diastolic BP); and (ii) augmentation index (AIx).
- Cardiac function and hemodynamic parameters
Using applanation tonometry on the radial artery (Sphygmocor®, Atcor Medical, Australia), myocardial perfusion was estimated by calculating the subendocardial viability ratio (Buckberg index: diastolic time index/tension–time index) (26,27).
We measured with trans-thoracic bio-impedance cardiography (Task Force Monitor®, CNSystems Medizintechnik, Austria) the following hemodynamic parameters: stroke volume, cardiac output, total peripheral resistance index, left ventricular ejection time, left ventricular work index and thoracic fluid content (28,29).
- Cardiovascular autonomic nervous system activity
Finger arterial BP and heart rate (HR) were recorded continuously in the supine position for 6 min at a paced breathing rate of 12 breaths/min using photo-plethysmographic sensors and synchronous 6-lead ECG recordings (Task force monitor®, CNSystems Medizintechnik, Austria). Cardiac autonomic nervous system activity was assessed using spectral analysis of HR and SBP variations (30,31). The opposite brachial artery was used to calibrate digital arterial BP measurements. Cardiac autonomic activity was expressed as normalized spectral HR variability: (i) normalized spectral index [i.e., the percentage of the high-frequency band (HF: 0.15–0.40 Hz/total spectrum) of HR variations (HFnu-RRI: vagal function); (ii) the ratio between low frequency (LF: 0.03–0.15 Hz/total spectrum) and HF bands of HR variations (LF/HF-RRI: sympatho–vagal balance) and (iii) LF band of the SBP spectrum (LF-SBP: 0.075–0.15 Hz; vascular sympathetic activity).
Variables were expressed as mean ± SD or as percentages. Comparisons between groups were performed for continuous variables by analysis of variance or the Mann–Whitney test as appropriate and by Chi-square test for comparisons of categorical variables. Quantitative variables were analyzed by different mixed models of ANOVA for longitudinal data including factor treatment (saxagliptin and placebo) and factor time with 2 levels when comparing changes between visits (i.e. V1 and V2) or three levels when analyzing the test meal response (0, 60 and 120 min after breakfast). Interaction between the 2 factors was used to test possible effect of treatment on time-dependent changes in studied parameters. Normalizing transformations were made if necessary. All tests were two-sided at 5% significance level. Statistical analyses were performed with SAS version 9.4.