Introduction:
Adipose tissue dysfunction contributes to insulin resistance in type 2 diabetes (T2D). Adipokines such as adiponectin, leptin, and their ratio are potential markers of this dysfunction, but their relevance may vary by ethnicity. This study aimed to investigate ethnic variation in how these adipokines relate to insulin resistance in people with T2D.

Methods:
Adults with T2D and no history or symptoms of cardiac disease were prospectively recruited at a single centre (NCT03132129) and underwent comprehensive phenotyping that included demographics, standard biochemistry, quantification of plasma adipokines (adiponectin and leptin), and insulin resistance was estimated using the homeostatic model assessment (HOMA-IR). In this study, White Europeans (WE) were compared to non-WE. Skewed variables were log-transformed. The association between the adipokines and their ratio (A/L) with HOMA-IR were examined using Pearson correlation. To determine if there was an independent relationship between HOMA-IR and each adipokine or their ratio, general linear regression modelling was applied, adjusting for age, sex and BMI, including an interaction term between ethnicity and the A/L ratio. Ethnicity-stratified models evaluated whether the A/L ratio and HOMA-IR associations differed between groups.

Results:

A total of 196 participants (109 WE and 87 non-WE) were included. All continuous variables are presented as mean ± SD unless stated otherwise. WE participants were slightly older (63.8 ± 6.8 vs. 61.9 ± 6.2 years), had higher BMI (31.6 ± 6.0 vs. 27.8 ± 4.4 kg/m²), with a similar proportion of men in both groups (61% vs. 59%). There were no statistically significant differences between WE and non-WE groups in adiponectin levels (reported as median [IQR]), 7.67 (6.05–10.17) vs. 7.14 (5.81–9.48) ng/mL, leptin levels 24.10 (13.27–43.31) vs. 22.39 (13.00–37.01) ng/mL, or the A/L ratio 0.288 (0.160–0.662) vs. 0.342 (0.168–0.652). The A/L ratio correlated with HOMA-IR in WE (r = −0.39, p < .001) but not in non-WE individuals (r = −0.03, p = 0.77). In multivariable linear regression, a significant interaction was seen between ethnicity and A/L ratio for HOMA-IR (B= -0.8, p= 0.003). Stratified analyses showed that in WE participants, A/L ratio was an independent predictor of HOMA-IR (B = −1.01, p < .001), with the model explaining 17% of the variance. In contrast, in non-WE participants, neither the A/L ratio (B = −0.02, p = 0.77) nor the overall model was significant.

Conclusion:

Our data indicate that the A/L ratio is a significant independent predictor of insulin resistance in WE, but not in non-WE individuals with T2D. The lack of a significant association between A/L ratio and HOMA-IR in non-WE individuals could suggest that there are other contributing factors (e.g., genetics) to insulin resistance pathways.