Clinical methods

This manuscript is written in accordance with the SAMPL and CONSORT guidelines.

Trial design

The study protocol has been published in details previously [7]. In summary, this was a multicenter double blind randomized placebo-controlled trial (RCT) that investigated the effects of empagliflozin 10 mg once daily versus placebo in addition to standard anti-diabetic treatment on inflammatory status in patients with concomitant T2DM and CAD. This study was conducted in accordance with the principles of the declaration of Helsinki and all subsequent revisions. Our study was approved by the ethics committee of Zanjan university of Medical Sciences, and the protocol was prospectively registered on the Iranian Registry of Clinical Trials (www.IRCT.ir, Identifier: IRCT20190412043247N2). A written informed consent was obtained prior to the study recruitment from all patients.

Participants and eligibility assessments

The included patients aged from 45 to 75 with T2DM and had documented CAD confirmed by coronary angiography (CAG) plus they had been treated with standard anti-ischemic and anti-diabetic medications for at least 3 months prior to this study. The eligible HbA1C prior to recruitment for inclusion was between 6.5 to 9.5%. The key exclusion criteria were: history of SGLT2-i medications as well as any allergic reaction to SGLT2-i drugs and the use of anti-coagulants or anti-platelet drugs (except aspirin 80 mg/daily), pioglitazone, antioxidant supplements, alcohol and anti-inflammatory drugs for at least 3 months prior to enrollment; heart failure (New York Heart Association class III-IV) or left ventricular ejection fraction (LVEF) < 40%; body mass index (BMI) ≥ 40 kg/m2; history of transient ischemic attack (TIA), acute coronary syndrome (ACS), cerebrovascular accident (CVA), coronary artery bypass graft (CABG) surgery or percutaneous coronary intervention (PCI) during past 3 month; eGFR ≤ 45; history of recent infection (1 month prior to blood sampling); presence of pregnancy or lactation; history of diabetic ketoacidosis (DKA) or malignancy; anemia (hemoglobin < 10 g/dl) or thrombocytopenia (Platelet count < 100,000/µl).

The study recruitment phase began on June 29, 2020 and ended on March 13, 2021. Patients were recruited at 4 clinical centers (Mousavi hospital [coronary angiography registry], Vali-e-Asr hospital, and two cardiology clinics of Zanjan University of Medical Sciences) in Zanjan, Iran. Before recruitment, the registered records of patients were screened to assess their eligibility based on the inclusion/exclusion criteria. Further, they were interviewed by the investigators to be recruited to the study.

Intervention, randomization and follow up

At the day of recruitment (baseline), patients went through a series of evaluations that were comprised of blood samplings, directed interviews by the filling checklists that incorporated demographic information, food frequency questionnaire (FFQ), physical activity questionnaire using international physical activity questionnaire (IPAQ), 2-dimentional transthoracic echocardiography (2D-TTE), electrocardiography (ECG), and physical examination. Further patients were randomized into 2 groups: 10 mg of empagliflozin and placebo. The randomization process was employed according to the stratified randomization method. Stratification of patients was based on age, gender and HbA1c. With the help of Winpepi software (V 11.6), the randomization sequence was created and by the sealed envelope method, the allocation sequence was concealed until the day of recruitment. Patients were followed up every 4 weeks through phone calls. At the end of weeks 2, 12, and 26, patients attended clinic visits for safety assessments, physical examination, and thorough checkups. At each visit, patients were advised to maintain a fixed physical activity throughout their participation in the trial. All patients’ drug containers were collected at the end of week 26.

Outcomes

The primary outcome was changes in levels of plasma IL-6 after 26 weeks between the two groups. Secondary outcomes were: (1) changes in levels of other inflammatory biomarkers [High-sensitivity C-reactive Protein (Hs-CRP) and Interleukin 1-beta (IL-1β)], (2) changes oxidative stress markers in plasma [superoxide dismutase (SOD) enzyme activity, reduced glutathione (GSHr), catalase (CAT) enzyme activity, ROS, total antioxidant capacity (TAC), malondialdehyde (MDA), and protein carbonyl groups (PCG)], (3) platelet function alternations (P-selectin Ag expression intensity), and (4) glycemic status [fasting blood sugar (FBS), HbA1c, and homeostatic model assessment for insulin resistance (HOMA-IR)].

All the secondary outcomes were assessed by the changes from baseline to week 26 of treatment between each study arms.

Blinding and study monitoring

Empagliflozin and placebo tablets were given to participants with the same color and shape. Following enrollment, a code was assigned to each individual patient. Patients, researchers and healthcare providers who collected information, assessed the outcomes and analyzed the final data were all blinded to the assigned treatment (Empagliflozin or placebo), classified into groups of A and B. Data source verification in addition to study monitoring were instituted by the Research Council of Zanjan University of Medical Sciences, Iran, independent from the sponsors or any conflicting interests. Adverse event, regardless of their relevance, were all documented in the standard adverse events form and were either reported by the patients or were identified by our team of investigators throughout the study follow-up (until week 28).

Experimental methodsMaterials and instruments

Elisa kits used for measuring cytokines included Human IL-6 Elisa Kit (Diaclone, France) and Human IL-1β Elisa Kit (Zellbio, Germany). Laboratory kits used for measuring the oxidative stress biomarkers (except ROS) included SOD Activity Kit (KSOD96), CAT Activity Kit (KCAT96), Thiol Content Assay Kit (KTHI96), Total Antioxidant Assay Kit (KTAC96), Lipid Peroxidation Kit (KMDA96), and Protein Carbonyl Assay Kit (KPCA96) were purchased from KIAZIST Company, Iran. 2'-7'dichlorofluorescin diacetate (H2DCFDA) (D399, Invitrogen, USA) was used to assess Cellular ROS. Platelet agonist used to activate platelet was thrombin receptor activator peptide (TRAP-6 (SFLLRN), 4,031,274.0005; Bachem, Germany) and antibodies used for flow cytometric measurement of platelet function were FITC Mouse Anti-Human CD62P (Clone AK-4) (555,523, BD Pharmingen, USA) and FITC Mouse IgG1, κ Isotype Control (Clone MOPC-21) (555,748, BD Pharmingen, USA). The instruments used for the laboratory measurements were as follow: 3-30 K Centrifuge (Sigma, Germany), FACSCalibure Flow Cytometer (BD Pharmingen, USA), UV2100 spectrophotometer (UNICO, China), Infinite M200 plate reader (Tecan, Austria), and ELISA reader apparatus (BD Biosciences, USA).

Experimental design and setting

Blood sampling was performed on the day of recruitment, on week 26 and whenever there was a suspicion for any adverse event. Sampling was performed at 2 centers; for routine laboratory tests, whole blood samples were submitted directly to the Mousavi Hospital core clinical laboratory. The eligible patients were also referred to the target hospital for sample collection. For cytokine and oxidative stress (except ROS) measurements, whole blood was centrifuged at 3500 rpm for 15 min and separated plasma was allocated in 3 microtubes and stored in an ultra-low temperature freezer tagged with the patients’ unique codes (Biobank). After all measurements the remaining samples were anonymized, and the biobank was discontinued. The process of blood sampling to plasma freezing was less than 60 min for all samples. At the end of each phase (Week 0 and 26) the collected samples were defrosted, and the mentioned biomarkers were measured according to the Kits’ standard protocol.

Measurement of intra-cellular lymphocytic ROS

Peripheral blood mononuclear cells (PBMC) isolation procedure was performed as following: (1) whole blood sample was collected using 21-gauge needle in 5 ml lithium heparin vacuum tubes. (2) Fresh heparinized blood was diluted 1:1 ratio with Phosphate-buffered saline (PBS) and mixed well at room temperature. (3) Ficoll–Hypaque density gradient centrifugation was used for PBMCs isolation [8]. (4) 10 ml diluted blood was separated into two conical tubes containing 2 mL ficoll and then centrifuged in a horizontal rotor centrifuge (40 min at 800×g at 18 °C). (5) The buffy coat layer was removed and washed with PBS (12 min at 420×g at 18 °C). Trypan blue staining was performed to determine viability and counting isolated PBMC. The overall cell viability was more than 90% in all subjects.

Intracellular lymphocytic ROS (including hydrogen peroxide, peroxynitrite, and hydroxyl radical) level was measured using H2DCFDA prob. Given the cell viability assessments the concomitant PI staining did not perform. The Isolated PBMC was divided into 2 tubes of unstained and test with 8 × 105 cells per tube. After titration, 1 µl H2DCFDA in addition to pre-heated PBS were added to the to the isolated PBMC of test tube (final volume = 1 mL) and for the unstained tube, only pre-heated PBS was added (Final volume = 1 mL). Both tubes were incubated at 37 °C in a dark and tightly sealed environment for 40 min. After the incubation period, both tubes were washed twice with PBS (× 1500 rpm for 10 min). Flow cytometry was used to discriminate between lymphocytes and monocytes using the forward and sideward scatter. After lymphocyte zone gating, the mean fluorescent intensity (MFI) of tubes were read on FL1-H Channel [9]. The data obtained by flow-cytometry was analyzed using FlowJo V7.6.5.

Performing platelet activation test

Strategies employed to reduce unwanted platelet activation included: (1) blood sampling was performed using 21-gauge needle size, (2) the platelet activation test was performed on whole blood to minimize the time interval between blood collection to platelet activation and repeated physical stress caused by ultra-centrifuging, and (3) the time interval between sample collection to data acquisition was planned to be less than 90 min.

The platelet activation test was performed in biotechnology department in school of pharmacy of Zanjan University of Medical Sciences. The protocol used in this study was obtained by the previously published optimized protocol performed by Huskens et al. [10]. Fresh blood collected into sodium citrate (3.2%) tubes were incubated at 37 °C for 10 min. After that, to minimize the formation of platelet aggregation, the blood was diluted 1:4 in pre-heated N-2-Hydroxyethylpiperazine-N'-2-Ethanesulfonic Acid (HEPES)-buffered saline. From the diluted blood, 20 µl was added to 4 tubes. Based on the label of each tube the following reagents were added: (1) Tube1 (Unstained): 80 µl of preheated HEPES-buffered saline (2) Tube 2 (Negative control): 20 µl of FITC-conjugated anti-P-selectin Ab + 60 µl of preheated HEPES-buffered saline (3) Tube 3 (Isotype control): 20 µl of FITC-conjugated Mouse IgG κ Ab + 60 µl of preheated HEPES-buffered saline (4) Tube 4 (Test): 20 µl of FITC-conjugated anti-P-selectin Ab + 60 µl mixture of preheated HEPES-buffered saline and TRAP (30 µmol/L). The tubes were incubated for exactly 20 min at 37 °C. Flow cytometry used to discriminate platelet from other cells using the forward and sideward scatter both on Log mode using the previous literature [11]. After platelet zone gating, the MFI of tubes were read on FL1-H Channel. The data obtained by flowcytometry was analyzed using FlowJo V7.6.5.

Statistical considerationSample size calculation

According to a previous study in which, the SGLT2-i arms (after intervention) had a mean plasma of IL-6 5.8 (pg/ml), and standard deviation (SD) of 8.9. The minimal important difference (MID) was calculated 4.45 using the formula 0.5× SD [12]. The distributional method was applied for calculation of sample size using the following formula (13):

$$n=\frac_}+_)}^\times (_^+_^)}^}$$

With the study power of 80% and a 2-sided alpha level of 0.05, the sample size was calculated 41 in each arm. Finally, our study intended to randomize 100 patients preparing the required space for the possible dropouts during study. The calculation of our sample size was administered by G-power software (version 3.1.9.2).

Statistical analysis

Results from categorical variables are reported as frequencies (percentage) and continuous variables are reported as means (SD). Analysis was done through the intent-to-treat approach which consists of including of all randomized participants in the same group that they were originally allocated. With reference to the central limit theorem, the means (SD) were used for the analysis of the non-normally disturbed continuous data [14]. Ahead of sample size calculation, the primary and secondary outcomes were considered confirmatory and exploratory, respectively. Between-group comparisons of changes from baseline to 26-week of treatment for the primary and secondary outcomes were based on ANCOVA model. Chi-square test was used to analyze categorical variables. Probability value of < 0.05 was considered statistically significant. All statistical tests were carried out in the in Rv.4 environment. The primary data obtained by the FFQ, and flow cytometry was processed by the Nutritionist software version 4 (N4) and FlowJo software V7.6.5, respectively.