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


Basal Characteristics

Of the 69 patients enrolled, data from 60 patients could be used for the final evaluation (Fig. 1). Thus, 5 patients withdrew owing to personal reasons; and after the intervention started, 1 patient was excluded due to failure to adhere to the study protocol. Based on serum pioglitazone measurements (not shown), 3 patients were excluded due to detection of pioglitazone already at the baseline, while the pioglitazone levels assessed at week 24 confirmed adherence to the study protocol in the Pio and Pio& Omega-3 subgroups.

No significant differences were observed between the subgroups in basic anthropometric and biochemical characteristics measured in the fasting state during the study (Table 1). In the subgroup analysis by the effect of the 24-week-intervention (see Δ-values in Table 1), both Pio and Pio& Omega-3 increased body weight and BMI compared with Placebo or Omega-3 (except for Placebo vs. Pio in the case of body weight; p = 0.10). Triacylglycerol, NEFA and total cholesterol in serum, blood lipoproteins including HDL cholesterol and LDL cholesterol, and markers of oxidative stress in serum including SOD activity, TBARS and GSSG/GSH ratio were not significantly affected by any of the interventions (Table 1). Magnetic resonance spectroscopy did not show any significant effect of the interventions on the ectopic lipid content (see Additional file 1 Leptin levels were increased in Pio compared with the other subgroups (p = 0.02). At week 24, adiponectin levels were higher in Pio and Pio& Omega-3 than in Placebo subgroup, reflecting ~1.6-fold fold stimulatory effect in the median concentration in both Pio and Pio& Omega-3 (p < 0.0001). Adiponectin levels in the Omega-3 subgroup were higher, reflecting 1.2-fold increase in the median in this subgroup (p < 0.001) but were unchanged in the Placebo subgroup (Table 1).

Omega-3 PhL Index was similar in all subgroups at the baseline, while at week 24 it was not affected by either Placebo or Pio, but increased to a similar extent (1.6–1.7-fold; p < 0.0001 and p = 0.0001, respectively) in response to both Omega-3 and Pio& Omega-3 (Table 1). As observed before,[26] Omega-3 PhL Index differed between individuals, showing up to ~3-fold differences when all 60 patients were compared at the baseline (Fig. 2 a-d). In the EPA + DHA supplemented subgroups (Omega-3 or Pio& Omega-3) a maximum ~2.2-fold difference in the median Omega-3 PhL Index was observed between individuals at the end of the intervention (Table 1). The variable increase in the Omega-3 PhL Index in response to EPA + DHA supplementation was independent of the pre-intervention value (Fig. 2 c, d). In contrast with the Omega-3 PhL Index (i.e., the EPA and DHA in serum phospolipids), LA content in phospholipids was not affected by any intervention and no differences in LA content between the subgroups of patients were found, either before or after the intervention (see Additional file 2

Figure 2.

Omega-3 PhL Index in individual patients. Analysis was performed at baseline (white bars) and at week 24 (black bars) in Placebo (a) Pio (b) Omega-3 (c) and Pio& Omega-3 (d) subgroups. Case numbers are indicated

Glucose Metabolism

No differences between Placebo, Pio and Pio& Omega-3 subgroups were observed in the markers of acute and long-term glycemic control, i.e., fasting blood glucose and serum HbA1c level, respectively, either at baseline, or at week 24 (Table 2). In response to Omega-3, both parameters increased by ~1.2-fold (fasting blood glucose, p = 0.02; HbA1c, p = 0.01) resulting in a significant effect of Omega-3 (see Δ-values in Table 2) and indicating a marginal deterioration of glycemic control. Glucose disposal rate, assessed using hyperinsulinemic-euglycemic clamp (M value), is a measure of insulin sensitivity. While it was similar in all the subgroups at baseline, M increased (p = 0.04) in Pio& Omega-3 subgroup during the intervention and at week 24, it was higher in the Pio& Omega-3 compared with the Omega-3 subgroup. In the subgroup analysis by the effect of the intervention (see Δ-values in Table 2), the effect of Pio (p = 0.12) and Pio& Omega-3 (p = 0.12) tended to be different from that of Omega-3. Thus, regarding the effects on insulin sensitivity, Omega-3 exerted a neutral effect compared with Placebo, while the results collectively document a marginal improvement by the Pio& Omega-3 intervention.

Energy Metabolism and Metabolic Flexibility

At week 24, indirect calorimetry was performed in conjunction with the hyperinsulinemic-euglycemic clamp. None of the measured parameters, namely REE, RQ, carbohydrate oxidation and fat oxidation were significantly different between the subgroups before or during the clamp. Non-oxidative GDR, assessed during the clamp, was also similar in all subgroups (Table 3).

Next, we attempted to detect possible differences in metabolic flexibility between the subgroups at week 24. We focused on the increase in RQ during the clamp, as a common way for assessment of metabolic flexibility to carbohydrates, which is usually impaired in insulin-resistant individuals.[27] No significant differences in the increase in median RQ between the subgroups were observed (Table 3). Therefore, a robust approach based on the evaluation of percent relative cumulative frequency (PRCF) curves of RQ values was used while pooling all RQ values from all the patients within subgroups. This was done for both fasting and clamp periods (Fig. 3). Provided that the PRCF curve represents normally distributed data, the value of EC50 of PRCF (50th percentile) corresponds to a mean RQ value, while these curves also allow to identify differences that may exist at either lower or upper levels of RQ range.[28] During the clamp, PRCF curves within each subgroup shifted to the right (i.e., towards glucose oxidation), documenting various degrees of metabolic flexibility to glucose. Compared to Placebo, metabolic flexibility was improved by all interventions in the following order of effect: Pio < Omega-3 < Pio& Omega-3 (see the PRCF curve shifts in the legend to Fig. 3).

Figure 3.

Metabolic flexibility after the interventions. RQ data from indirect calorimetry at week 24 (see Table 3) were used to construct PRCF curves, each of which represents data (~400) pooled from all patients in the given subgroup either in fasting state (dashed lines) or during clamp (solid lines). RQ values corresponding to EC50 (50th percentile value) on each of the curves were obtained and the difference between this RQ value assessed during the clamp and fasting, respectively, was used as a marker of metabolic flexibility to glucose (PRCF curve shift; Placebo, 0.04; Pio, 0.06; Omega-3, 0.06; Pio& Omega-3, 0.07)

Postprandial Metabolism

A meal test was performed at baseline and at week 24, allowing for assessment of the effects on postprandial metabolism of both glucose and lipids, and of the insulin response to a carbohydrate load. Transient increases in serum glucose, C-peptide, NEFA and triacylglycerol (and insulin; not shown) levels triggered by a standard breakfast were followed during 120 min. Data were expressed as AUC (Fig. 4). At baseline, no significant differences between the subgroups were observed; at week 24, Pio& Omega-3 subgroup showed faster metabolism of triacylglycerol (lower AUC) compared with both Placebo and Omega-3 (p = 0.04; not shown). In the subgroup analysis by the effect of the intervention, several significant results were obtained. Thus, the difference between AUC at week 24 and baseline (ΔAUC) for glucose was higher in Omega-3 compared with the other subgroups (Fig. 4a; except for Omega-3 vs. Pio& Omega-3, p = 0.06), suggesting a deterioration of glucose metabolism. In the case of C-peptide (and insulin; not shown), no significant differences between the interventions were found (Fig. 4b). Postprandial metabolism of NEFA was accelerated in response to Pio& Omega-3, as documented by the lower ΔAUC in this subgroup compared with both Placebo and Omega-3 (Fig. 4c). Regarding the metabolism of triacylglycerols (Fig. 4d), ΔAUC was similar in the Placebo, Pio and Omega-3 subgroup, while it was decreased in response to Pio& Omega-3.

Figure 4.

The effects of interventions on postprandial metabolism of selected metabolites and insulin response. A meal test was performed in overnight fasted patients at baseline and at week 24. Transient postprandial increases in serum concentrations of various analytes were evaluated during 120 min following a standard breakfast. The difference between total AUC for each analyte measured at week 24 and baseline (ΔAUC) is shown for glucose (a) C-peptide (b) NEFA (c) and triacylglycerol TG; (d). Plots represent 10th, 25th, 50th (median), 75th, and 90th percentiles. Significant differences (Kruskal-Wallis test) between the subgroups are indicated

Previous studies showed an anti-inflammatory effect of EPA + DHA at the level of some plasma cytokines in the postprandial state.[29] Therefore, various anti-inflammatory (IL-1RA and IL-10) and pro-inflammatory cytokines (MCP-1, CRP, TNF-α and IL-6), as well as proteins involved in cell adhesion (sVCAM-1, sICAM-1, sE-selectin, sP-selectin, sPECAM-1) and neovascularization (sCD105) were evaluated in serum samples collected at 120 min of the test, both at baseline and week 24. Except for sP-selectin, no significant effects of the interventions were observed (see Additional file 3