Effects of Conjugated Linoleic Acid and Metformin on Insulin Sensitivity in Obese Children

Randomized Clinical Trial

Nayely Garibay-Nieto; Gloria Queipo-García; Flor Alvarez; Mayra Bustos; Erendira Villanueva; Fernando Ramírez; Mireya León; Estibalitz Laresgoiti-Servitje; Ravindranath Duggirala; Teresa Macías; Sergio Cuevas; Abel Jalife; Miguel Fonseca-Sánchez; Fabiola Serratos; Juan Carlos López-Alvarenga


J Clin Endocrinol Metab. 2017;102(1):132-140. 

In This Article


Participants and Demographics

Enrollment occurred from August 2012 to July 2014. One hundred ninety-eight individuals were potentially eligible; 83 met inclusion criteria, signed consent and assent forms, and were randomized to receive MET (n = 24), PLB (n = 30), and CLA (n = 29). Fifty patients completed the 16-week intervention; 1 external outlier was identified and excluded during the analysis (PLB group). In 1 case, EHCT performance was technically impossible (CLA group). For this reason, we report the results of 48 executed clamps (Fig. 1). Throughout the study, 1 patient was eliminated when a preexisting lipoma was surgically removed without notifying the research team (PLB group); a second patient with psychosocial anomalies and suspected pregnancy was eliminated as well (PLB group). Twenty-nine patients were eliminated due to poor medication compliance or due to lack of interest (MET, n = 10; PLB, n = 10; and CLA, n = 9). Pubertal development (Tanner stage 1, defined as prepubertal; Tanner stages 2 to 3, defined as early puberty; and Tanner stages 4 to 5, defined as late puberty) was assessed after a clinical inspection of the mammary glands, testes volume, and pubic hair. Demographic and baseline characteristics were similar among the groups (Table 1).

Figure 1.

Flowchart representing the number of subjects at study enrollment and study termination. Fourteen clamp studies were conducted in the PLB group, 17 in the MET-treated group, and 17 in the CLA-treated group.

Anthropometric and Metabolic Results

No significant differences were observed in baseline anthropometric and metabolic parameters or insulin resistance measured by surrogate indexes of insulin resistance (fasting insulinemia, HOMA-IR, and QUICKI). Distribution of Tanner stage status did not differ among the groups (χ2 test, P = 0.415).

The overall impact of the intervention showed a positive effect on weight, height, BMI, and waist circumference, as well as on surrogated indexes of insulin resistance and physical fitness score in all of the groups (Table 2). No statistically significant differences were observed in these parameters between treatment groups. No differences were evident when comparing surrogate indices of insulin resistance during the postintervention phase among the groups.

Insulin Sensitivity Measured by EHCT

The primary outcome, insulin sensitivity, calculated as the Rd value, showed significant difference between the CLA group compared with PLB (6.53 ± 2.54 vs 5.05 ± 1.46, P = 0.035, Cohen's d effect size of 74%) (Table 3). Moreover, fasting insulinemia (Fig. 2) and HOMA-IR (Fig. 3) significantly decreased in the CLA group (P = 0.04). The adjusted analysis for controlling the influence of modifying or confounding variables such as BMI, change in BMI, age, Tanner stage, as well as dietary and physical program compliance, over the Rd value, showed that the Tanner stage had an independent effect over the Rd value (P < 0.001). When ANCOVA was executed and the aforementioned variables were controlled, no statistically significant differences were found between the three groups with regards to Rd value. Nonetheless, a clinically relevant effect size remained evident when comparing the CLA and PLB groups (Cohen's d effect size of 37%), suggesting a decrease in insulin resistance in patients receiving CLA. The effect size of MET vs PLB was 10% (5.72 ± 3.1 vs 5.38 ± 3) and that of MET vs CLA was 20% (5.72 ± 3.1 vs 6.34 ± 2.8), favoring the CLA-treated group. However, these effect sizes were not clinically relevant.

Figure 2.

Fasting insulin serum concentrations were significantly lower in the CLA-treated group at study termination, whereas no differences were found in the MET-treated patients or in the PLB group. The error bars show SEM.

Figure 3.

HOMA-IR was significantly lower in the CLA-treated group at study termination compared with initial values. No differences were found in the MET-treated and control groups. The error bars show SEM.

We analyzed the changes between initial and final serum triglycerides and high-density lipoprotein (HDL) cholesterol by ANCOVA. For these particular variables, no Tanner or change in BMI modified postintervention levels. Furthermore, HDL cholesterol and triglyceride baseline levels did show an influence over the final levels.

Lipid Profile and Adverse Effects

Patients in the CLA group had a statistically significant increase in serum triglycerides when compared with MET (169.8±69 vs 113.1±27, P = 0.027), but not significant when compared with PLB (P = 0.13). Moreover, HDL levels were lower in the CLA group when compared with MET (36.8 ± 5.4 vs 44.86 ± 8.7, P = 0.009), whereas there were no differences when compared with PLB (P = 0.26). The main differences favoring MET treatment over the lipid profile were evident only when compared with CLA.

Nonserious adverse events most commonly reported were abdominal pain, diarrhea, dizziness, headache, nausea, and gastritis. The frequency and severity of symptoms were similar in the three groups (analysis of variance, P = 0.314; χ2 test, P = 0.28). Patients exhibiting lack of compliance and/or dropout did not show a difference between groups. Additionally, a Little's missing completely at random analysis (P > 0.13) was conducted to ensure that patient elimination was actually random and homogeneous in all of the groups.

Muscle Biopsies' Analyses

The analyses of IRS1, IRS2, and IRS4 revealed that only IRS2 was modulated in the CLA group, showing a 3.56-fold increase compared with the control group (P = 0.043). The rest of the genes did not show statistically significant differences. These data support the fact that CLA has a critical effect in the molecular insulin pathway through the upregulation of ISR2. This mechanism might be related to optimal glucose uptake observed in CLA-treated patients.