SFE2001 Poster Presentations Diabetes and metabolism (9 abstracts)
Department of Endocrinology and Metabolic Medicine, Imperial College at St Mary's Hospital.
Use of surrogate indices of insulin resistance depends on comparability of the relationship between glucose and insulin concentrations between different groups. It is unknown whether such surrogate estimates are equally applicable in different ethnic groups with marked differences in insulin sensitivity. We compared markers of insulin sensitivity and secretion in 78 (43 European, 19 Afro-Caribbean and 16 South Asian) normoglycaemic women of (median (lower-upper quartile)) 34.7(31.8-38.8) years of age. Insulin sensitivity (Si) derived using the frequent-sampling intravenous glucose tolerance test (FSIVGTT) was compared to measures derived from baseline data: f. insulin, f. glucose/f. insulin, and HOMA-%S. Acute insulin response to glucose (AIRg), an FSIVGTT-derived index of b-cell function, was compared to HOMA-%B.
Si and AIRg differed between ethnic groups (p=0.002, p=0.006, respectively; one-way ANOVA). European women had higher Si (1.49(1.17-1.98)10-4/min/pmol/l) than both Afro-Caribbean (0.67 (0.40-1.29)10-4/min/pmol/l, p=0.003) and South Asian women (0.74(0.18-1.54)10-4/min/pmol/l, p=0.004) and lower AIRg (852(384-1295)pmol/l h) than Afro-Caribbean (2101(569-3531)pmol/l h, p=0.005) women. Si and AIRg correlated with other markers of insulin resistance and b-cell function only among European subjects (Si vs f.insulin: r=-0.45, p=0.002 ; vs HOMA-%S: r= 0.45, p=0.002, vs gluc/ins: r=0.44, p=0.003 and AIRg vs HOMA-%B: r=0.35, p=0.02; Spearman's correlation), whereas no such correlation was observed among South Asian (Si vs f. insulin: r=-0.28; vs HOMA-%S: r= 0.28, vs gluc/ins: r=0.29 and AIRg vs HOMA-%B: r=0.27, all p values >0.05) or Afro-Caribbean women (Si vs f. insulin: r=-0.11; vs HOMA-%S: r= 0.08, vs gluc/ins : r=0.14 and AIRg vs HOMA-%B: r=0.21, all p values >0.05) .
South Asian and Afro-Caribbean are more insulin resistant that European women and failed to display a similar degree of correlation between reference and surrogate measures of insulin action and secretion. Possibly, where no correlation is found, insulin action and secretion are affected differently in the fasting compared to the glucose-loaded state.