Combined Intervention With Pioglitazone and n-3 Fatty Acids in Metformin-treated Type 2 Diabetic Patients

Improvement of Lipid Metabolism

Jiri Veleba; Jan Kopecky Jr.; Petra Janovska; Ondrej Kuda; Olga Horakova; Hana Malinska; Ludmila Kazdova; Olena Oliyarnyk; Vojtech Skop; Jaroslava Trnovska; Milan Hajek; Antonin Skoch; Pavel Flachs; Kristina Bardova; Martin Rossmeisl; Josune Olza; Gabriela Salim de Castro; Philip C. Calder; Alzbeta Gardlo; Eva Fiserova; Jørgen Jensen; Morten Bryhn; Jan Kopecky Sr.; Terezie Pelikanova


Nutr Metab. 2015;12(52) 

In This Article

Materials and Methods

Study Design and Patients

A 24-week, parallel-group, four-arm, randomized trial (EudraCT number 2009-011106-42) was conducted in accordance with the principles of the Declaration of Helsinki (2008 revision) and with approval by the Institutional Ethical Committee. All patients provided written informed consent prior to their participation.

Inclusion criteria were 40–70 years of age, diagnosis of T2D as defined by the criteria of the American Diabetes Association and recognized by WHO, Expert Committee on the Diagnosis and Classification of Diabetes Mellitus (American Diabetes Association, 2004) diagnosed at least 3 months preceding screening visit, treatment by oral metformin as a monotherapy at a stable dose (0.5–3.0 g/day) for at least 1 month and no other antidiabetic agent, hemoglobin A1c (HbA1c) < 80 mmol/mol, fasting plasma triacylglycerols ≤ 6 mmol/l, BMI 25–45 kg/m2, ability and willingness to adhere to the protocol and signed and dated written Informed consent obtained before any trial-related activities. Exclusion criteria were type 1 diabetes, uncorrected thyroid dysfunction, significant weight gain or loss (>5 % of total body weight within the past 3 months), therapy with insulin, or warfarin or fibrates within past 3 months (statins and salicylic acid were allowed; 51 % of patients were treated with either simvastatin or atorvastatin), tachycardia (>100 beats/min; or use of stable doses of antihypertensives shorter than 3 months prior the screening and during the trial), history of cardiovascular disease (myocardial infarction in the last year, coronary revascularization including percutaneous transluminal coronary angioplasty, coronary artery bypass graft surgery in the previous year and no subsequent angina, unstable angina, congestive heart failure), pregnancy or lactation, significant renal impairment (serum creatinine >150 μmol/l), chronic or advanced hepato-biliary diseases, history of alcohol or substance abuse within the past year, allergy to any of the capsule excipients, participation in any other clinical trial during the previous 3 months, and clinically significant anemia (hemoglobin < 120 g/l for males and < 110 g/l for females) or any other abnormal hemoglobin profile.


Out of 294 patients subjected to an initial screening, 69 eligible patients (66 % men) were enrolled (Fig. 1) at the Diabetes Centre, Institute for Clinical and Experimental Medicine, Prague, Czech Republic. All measurements, procedures and sample collection were performed at week 0 and week 24 (2 visits during each week), on an outpatient basis, after overnight (8–10 h) fasting with water ad libitum. During week 0 (baseline; 3 weeks after the screening visit maximum), at the first visit, serum and muscle samples were collected (see below and Additional file 1, and a hyperinsulinemic-euglycemic clamp was performed (see below). At the second visit one week later, a standard meal test was performed (see below), followed by proton magnetic resonance spectroscopy (Magnetom Trio, Siemens, Erlangen, Germany) of liver and skeletal muscle to measure lipid content as described in Additional file 1 At the second visit, patients were randomized to (i) 5 g/day corn oil (Placebo), (ii) 15 mg/day pioglitazone (Pio; Actos, Takeda), (iii) 5 g/day EPA + DHA concentrate (Omega-3; EPAX 1050TG, EPAX AS, containing about 15 % EPA, 40 % DHA, wt/wt; i.e., ~2.8 g EPA + DHA), and (iv) the combination of pioglitazone with EPAX 1050TG (Pio& Omega-3). Randomization was performed using a computer-based algorithm arranging experimental units in blocks of four. The randomization code was kept secret and revealed after the clean-file procedure had been completed when all data had been filled in the case report forms. Placebo and Omega-3 were administered as gelatin-coated 1 g capsules. Thus, the study was double blind for EPA + DHA and open-label for pioglitazone. During week 24, patients were handled similarly as during week 0, except for also performing indirect calorimetry in conjunction with the clamp (see below).

Figure 1.

Study design

Anthropometric Measurements

Body weight (and height, data not shown) was measured using periodically calibrated scales accurate to 0.1 kg. Waist circumference was measured with a measuring tape placed at the midpoint between the lowest rib and the upper part of the iliac bone (results not shown). Body mass index (BMI) was calculated using the Quetelet formula (weight in kilograms divided by the square of the body height). Blood pressure was measured after 5 min in a seated position at rest, using a digital M6 Comfort monitor (Omron, Kyoto, Japan). Three measurements were taken 2 min apart. The first measurement was discarded, and the mean of the remaining two measurements was recorded.

Hyperinsulinemic-euglycemic Clamp

A 3 h clamp (1 mU/kg.min−1), was conducted as described previously.[21] A teflon cannula (Venflon; Viggo, Helsingborg, Sweden) was inserted into an antecubital vein for the infusion of all test substances. A second cannula was inserted into a wrist vein for blood sampling and the hand was placed in a heated (65 °C) box to achieve venous blood arterialization. A stepwise primed-continuous insulin infusion (1 mU/kg body weight.min−1 of Actrapid HM; NovoNordisk, Copenhagen, Denmark) was administrated to acutely raise and maintain the plasma concentration of insulin at ~75 μU/ml. Glycemia during the clamps was maintained at approx. 5.5 mmol/l by continuous infusion of 15 % glucose. Arterialized blood glucose concentration was determined every 5–10 min as described in Other analytical methods (see below) and the infusion rate was adjusted accordingly. Mean plasma glucose concentrations were comparable within the groups during clamps before and after 24 weeks. The coefficients of variation of glycemia during the studies were less than 5 %. Insulin sensitivity was estimated as the glucose disposal rate (M), i.e. the amount of glucose (mg/kg body weight.min−1) needed to maintain the concentration of glucose during the last 20 min of the clamp.

Indirect Calorimetry

At week 24, indirect calorimetry was conducted for 30 min (basal values in fasting state) just before and during the last 30 min of the clamp, using an open-circuit system (VMAX; SensorMedics, Anaheim, CA, USA). Oxygen consumption (VO2; ml O2/min) and carbon dioxide production (VCO2; ml CO2/min) were recorded every 1 min. To assess fuel partitioning, respiratory quotient (RQ; RQ = VCO2/VO2) was estimated and substrate utilization and resting energy expenditure (REE) were calculated. Non-oxidative glucose disposal rate (GDR) was calculated by subtracting the rate of glucose oxidation from the total rate of glucose uptake during the last 20 min of the clamp.[22]

Meal Test

A meal test was performed as before.[21] After an overnight fast, subjects received a standard breakfast (baguette Crocodille Cheese Gourmet: 180 g, energy 452.8 kcal (1895.7 kJ)) of the following composition: carbohydrates 49 g (45 % energy), proteins 18.5 g (17 % energy), lipids 18.8 g (38 % energy), of which saturated fatty acids 6.8 g, monounsaturated fatty acids 6.0 g, and polyunsaturated fatty acids 5.0 g. Serum concentrations of glucose, immunoreactive insulin, C-peptide, non-esterified fatty acids (NEFA), and triacylglycerol were measured at 0, 30, 60 and 120 min (see below in Other analytical methods). Data were expressed as area under the curve (AUC).

Content of Selected Fatty Acids in Serum Phospholipids

Serum samples from fasted patients were analyzed using shotgun lipidomics and mass spectrometry, and a sum of concentrations of phospholipids containing EPA and/or DHA divided by total concentration of all phospholipids (Omega-3 PhL Index) was used as a biomarker of EPA and DHA status. Briefly, 50 μl -aliquots of serum were transferred into disposable borosilicate glass tubes with 100 μl of methanol/butylated hydroxytoluene (1,000:1; v/v) containing internal standards for quantification of lipid species: 17:0–17:0 phosphatidic acid, phosphatidylethanolamine, phosphatidylcholine, phosphatidylglycerol, phosphatidylserine, and phosphatidylinositol, respectively (75 nM final concentration). Lipid extracts were prepared using a modified procedure of Bligh and Dyer as previously described.[23] Each lipid extract was diluted with dichloromethane/methanol/isopropanol (1:2:4, v/v/v and 5 mM ammonium acetate) prior to infusion into a mass spectrometer (MS; QTRAP 5500, Sciex, USA; equipped with Turbo V ESI) for the analysis of phospholipids. All the mass spectra and tandem mass spectra were automatically acquired using multiplexed precursor ion (PIS) and neutral loss (NL) scans in positive and negative mode.[23] Analyst 1.6.1/Lipidview 1.3 software was used to identify molecular species and to determine amounts of individual lipids based on internal standard concentrations assuming comparable ionization of standards and phospholipids. Sum formula annotation (e.g. PE 34:2) and acyl chain information coming from negative PIS (e.g. PE 34:2 – PIS 18:2) was used to calculate the additional acyl chain (e.g. PE 34:2 = 18:2 + 16:0). Only combinations of the major fatty acids (carbon:double bonds – 0:0, 12:0, 14:0, 14:1, 16:0, 16:1, 18:0, 18:1, 18:2, 18:3, 20:0, 20:1, 20:2, 20:3, 20:4, 20:5, 22:4, 22:5, 22:6) were used for further data processing. The content of linoleic acid (LA; 18:2 n-6) in serum phospholipids was also determined.

Other Analytical Methods

Serum glucose levels were analyzed using the Beckman Analyser glucose-oxidase method (Beckman Instruments, Fullerton, CA, USA), plasma immunoreactive insulin and C-peptide concentrations were determined using insulin and C-peptide IRMA kits (Immunotech, Prague, Czech Republic), HbA1c, was measured by HPLC (Tosoh, Tokyo, Japan), lipid concentrations were assessed by enzymatic methods (Roche, Basel, Switzerland) and HDL-cholesterol was measured after double precipitation with dextran and MgCl2 as described previously.[24] To assess oxidative stress, the amount of lipid peroxidation was determined as thiobarbituric acid reactive substances (TBARS) by the reaction with thiobarbituric acid, the whole blood level of reduced (GSH) and oxidized (GSSG) glutathione was determined with glutathione HPLC diagnostic kit (Chromsystems, Munich, Germany), and the activity of superoxide dismutase (SOD) was analyzed using a superoxide dismutase assay kit (Cayman Chemical, MI, USA). In fasting patients (Table 1), concentrations of serum cytokines were measured using ELISA kits from Biovendor (Czech Republic; total adiponectin, leptin), in the postabsortive state, various cytokines were analyzed by microbead Luminex® assay (Luminex Corporation, Texas, United States; see Additional file 3 Serum pioglitazone levels were estimated using mass spectrometry (see above).[25]

Study Endpoints

The primary endpoints were changes from baseline in (i) insulin sensitivity (M; hyperinsulinemic-euglycemic clamp) and (ii) fasting triacylglycerol levels at week 24. Secondary endpoints included the changes in fasting glycemia and HbA1c, and postprandial change in glucose, NEFA and triacylglycerol levels (meal test), metabolic flexibility assessed by indirect calorimetry, and inflammatory markers.

Statistical Analysis

A power calculation indicated that 16 patients were needed to detect a 7 % difference in M due to the intervention with the probability 1 at the 0.05 level of significance (assuming accuracy of measurement 5 %). All values are presented as median and interquartile range (IQR). Data from the baseline and the end of the study, and the changes (Δ) between the baseline and the end at week 24, were analyzed by the Kruskal-Wallis one-way analysis of variance (ANOVA) using SigmaStat 3.5 (SSI, San Jose, CA, USA) and the statistical software R version 3.1.0 ( The Holm–Bonferroni corrections for multiple comparisons were used. Wilcoxon signed-rank test was used to analyze the effect of intervention within each subgroup. Threshold of significance was defined at a p value of ≤ 0.05. For the analysis of the dependence of the response (Δ value) of selected variables on the corresponding change in Omega-3 PhL Index (Δ Omega-3 PhL Index), a linear regression model with dummy variables that indicate a subgroup Placebo, Pio, Omega-3, and Pio& Omega-3, respectively, was used. Models were considered with interactions since the effect of Δ Omega-3 PhL Index varies by subgroups, so Δ Omega-3 PhL Index and subgroups interact in affecting Δ value of selected variables.