Could statistical adjustments for age mask the insulin–blood pressure relationship? Academic Article uri icon


  • Even though conclusive findings regarding the relationship between insulin and blood pressure (BP) have been made, several papers still report on finding weak or non-existing relationships in various population groups. These relationships are often weak, depend on the characteristics of the study population and are usually strongly confounded by obesity and age. Subsequently many investigators adjust via statistical methods for age and measures of obesity (such as body mass index). In the present study four different datasets were used (Australian Aboriginal people (N=675), Torres Strait Islanders (N=369), African women (N=94) and Caucasian women (N=112)) and showed very weak correlations in all groups after statistical adjustments for age and obesity (ranging from r=-0.04 to 0.13). All subjects were then divided into different age groups (15-29 years; 30-40 years; >40 years) and partial correlations were performed within each age group whilst adjusting only for obesity. Results still showed correlations (ranging from r=-0.29 to 0.27) with a similar trend with increasing age. More positive correlations were shown for the youngest groups, and more negative correlations for the oldest groups, with the middle-group (30-40 years) showing the weakest correlations-seeming to be in a transitional phase from a positive to negative correlation. It is therefore suggested that when the relationship between fasting insulin and BP is assessed, age stratification be used and not statistical adjustments.


publication date

  • April 2006