Meta-analysis reveals the association of common variants in the uncoupling protein (UCP) 1-3 genes with body mass index variability

Abstract

Background: The relationship between uncoupling protein (UCP) 1-3 polymorphisms and susceptibility to obesity has been investigated in several genetic studies. However, the impact of these polymorphisms on obesity is still under debate, with contradictory results being reported. Until this date, no meta-analysis evaluated the association of UCP polymorphisms with body mass index (BMI) variability. Thus, this paper describe a meta-analysis conducted to evaluate if the -3826A/G (UCP1); -866G/A, Ala55Val and Ins/Del (UCP2) and -55C/T (UCP3) polymorphisms are associated with BMI changes.

Methods: A literature search was run to identify all studies that investigated associations between UCP1-3 polymorphisms and BMI. Weighted mean differences (WMD) were calculated for different inheritance models.

Results: Fifty-six studies were eligible for inclusion in the meta-analysis. Meta-analysis results showed that UCP2 55Val/Val genotype was associated with increased BMI in Europeans [Random Effect Model (REM) WMD 0.81, 95% CI 0.20, 1.41]. Moreover, the UCP2 Ins allele and UCP3-55T/T genotype were associated with increased BMI in Asians [REM WMD 0.46, 95% CI 0.09, 0.83 and Fixed Effect Model (FEM) WMD 1.63, 95% CI 0.25, 3.01]. However, a decreased BMI mean was observed for the UCP2-866 A allele in Europeans under a dominant model of inheritance (REM WMD -0.18, 95% CI -0.35, -0.01). There was no significant association of the UCP1-3826A/G polymorphism with BMI mean differences.

Conclusions: The meta-analysis detected a significant association between the UCP2-866G/A, Ins/Del, Ala55Val and UCP3-55C/T polymorphisms and BMI mean differences.

Figure 1. Flowchart illustrating the search strategy used to identify association studies of UCP1-3 polymorphisms with body mass index changes for inclusion in the meta-analysis.

Figure 2. Forest plots showing individual and pooled weighted mean differences (95% CI) for the associations between UCP1-3826A/G and UCP3-55C/T polymorphisms and body mass index in Europeans, under a dominant inheritance model (A/A vs. G/G + A/G for the UCP1-3826A/G polymorphism, and C/C vs. T/T + C/T for the UCP3-55C/T polymorphism).

The area of the squares reflects the study-specific weight. The diamond shows the summary random-effect weighted mean differences estimated from the studies.

Figure 3. Forest plots showing individual and pooled weighted mean differences (95% CI) for the associations between UCP1-3826A/G and UCP3-55C/T polymorphisms and body mass index in Asians, under a dominant inheritance model (A/A vs. G/G + A/G for the UCP1-3826A/G polymorphism, and C/C vs. T/T + C/T for the UCP3-55C/T polymorphism).

The area of the squares reflects the study-specific weight. The diamond shows the summary random-effect weighted mean differences estimated from the studies.

Figure 4. Forest plots showing individual and pooled weighted mean differences (95% CI) for the associations between UCP2-866G/A, Ala55Val and Ins/Del polymorphisms and body mass index in Europeans, under a dominant inheritance model (G/G vs. A/A + G/A for the UCP2-866G/A polymorphism, Ala/Ala vs. Val/Val + Ala/Val for the UCP2 Ala55Val polymorphism, and Del/Del vs. Ins/Ins+Ins/Del for the UCP2 Ins/Del polymorphism).

The area of the squares reflects the study-specific weight. The diamond shows the summary random-effect weighted mean differences estimated from the studies.

Figure 5. Forest plots showing individual and pooled weighted mean differences (95% CI) for the associations between UCP2-866G/A, Ala55Val and Ins/Del polymorphisms and body mass index in Asians, under a dominant inheritance model (G/G vs. A/A + G/A for the UCP2-866G/A polymorphism, Ala/Ala vs. Val/Val + Ala/Val for the UCP2 Ala55Val polymorphism, and Del/Del vs. Ins/Ins + Ins/Del for the UCP2 Ins/Del polymorphism).

The area of the squares reflects the study-specific weight. The diamond shows the summary random-effect weighted mean differences estimated from the studies.