To the Editor: We commend the authors of the recent publication in Diabetologia entitled ‘Glucose patterns during an oral glucose tolerance test and associations with future diabetes, cardiovascular disease and all-cause mortality rate’ [1] and would encourage them to repeat their investigation using insulin response patterns in people with normal glucose tolerance. Hyperinsulinaemia is becoming well-established in the aetiology of many metabolic diseases [2], yet many people will be unaware that they are at risk. Insulin levels are posited to be elevated for many years, possibly even decades, before changes in blood glucose are noted [3, 4]. Although fasting insulin is used in many studies as a disease risk marker, it has a high coefficient of variation leading to concerns about repeatability [5]. There has been little emphasis on the use of insulin response patterns to assess the risk of metabolic disease, yet the patterns observed when tracking blood insulin levels following an OGTT may provide an earlier opportunity to assess disease risk compared with changes in blood glucose levels. Hayashi and colleagues noted that cumulative incidence of type 2 diabetes in Japanese-American men over 10 years of follow-up was between 38% and 48% in those who reached a peak insulin level at 120 min following a 75 g OGTT, which was higher than those whose insulin levels peaked at 30 or 60 min (cumulative incidence, <16%) [6]. However, these response patterns were based on relative insulin serum levels in a cohort that included a high proportion of participants with pre-existing impaired fasting glucose and/or impaired glucose tolerance [5]. Between 1970 and 1990, Dr J. R. Kraft collated a large database holding data from multiple-sampled OGTTs with insulin assays. Further analysis of this database suggests that up to 75% of people with normal glucose tolerance may have hyperinsulinaemia [7], but there are no longitudinal outcomes to support risk calculations. There is a distinct paucity of data on insulin response patterns and subsequent risk of cardiovascular disease and we believe the authors would advance the field with further data analysis.