Trajectories of glycaemia, insulin sensitivity, and insulin secretion before diagnosis of type 2 diabetes: an analysis from the Whitehall II study

Abstract

Background: Little is known about the timing of changes in glucose metabolism before occurrence of type 2 diabetes. We aimed to characterise trajectories of fasting and postload glucose, insulin sensitivity, and insulin secretion in individuals who develop type 2 diabetes.

Methods: We analysed data from our prospective occupational cohort study (Whitehall II study) of 6538 (71% male and 91% white) British civil servants without diabetes mellitus at baseline. During a median follow-up period of 9.7 years, 505 diabetes cases were diagnosed (49.1% on the basis of oral glucose tolerance test). We assessed retrospective trajectories of fasting and 2-h postload glucose, homoeostasis model assessment (HOMA) insulin sensitivity, and HOMA beta-cell function from up to 13 years before diabetes diagnosis (diabetic group) or at the end of follow-up (non-diabetics).

Findings: Multilevel models adjusted for age, sex, and ethnic origin confirmed that all metabolic measures followed linear trends in the group of non-diabetics (10,989 measurements), except for insulin secretion that did not change during follow-up. In the diabetic group (801 measurements), a linear increase in fasting glucose was followed by a steep quadratic increase (from 5.79 mmol/L to 7.40 mmol/L) starting 3 years before diagnosis of diabetes. 2-h postload glucose showed a rapid increase starting 3 years before diagnosis (from 7.60 mmol/L to 11.90 mmol/L), and HOMA insulin sensitivity decreased steeply during the 5 years before diagnosis (to 86.7%). HOMA beta-cell function increased between years 4 and 3 before diagnosis (from 85.0% to 92.6%) and then decreased until diagnosis (to 62.4%).

Interpretation: In this study, we show changes in glucose concentrations, insulin sensitivity, and insulin secretion as much as 3-6 years before diagnosis of diabetes. The description of biomarker trajectories leading to diabetes diagnosis could contribute to more-accurate risk prediction models that use repeated measures available for patients through regular check-ups.

Funding: Medical Research Council (UK); Economic and Social Research Council (UK); British Heart Foundation (UK); Health and Safety Executive (UK); Department of Health (UK); National Institute of Health (USA); Agency for Health Care Policy Research (USA); the John D and Catherine T MacArthur Foundation (USA); and Academy of Finland (Finland).

Figure 1

Fasting and 2-hour postload glucose trajectories (panels A and B) before the diagnosis of diabetes mellitus or the end of follow-up in 505 incident diabetes cases compared to 6033 non-diabetic controls. Multilevel longitudinal modeling using either linear growth model for non-diabetic and piecewise approach including cubic terms for time for incident diabetic subjects with OGTT fasting glucose (A) and 2-hour glucose (B) as outcomes. Adjusted for age, sex, ethnicity and study phase. Estimated for a hypothetical population of 72% male, 91% Caucasian aged 63 years at time 0 yrs. Error bars show 95% confidence intervals for the fixed effects. Tables show the number of measurements for each year at and before diabetes diagnosis / end of follow-up.

Figure 2

Homeostasis model assessment insulin sensitivity (HOMA2-%S) and HOMA β-cell function (HOMA2-%B) trajectories (panels A and B) before the diagnosis of diabetes mellitus or the end of follow-up in 505 incident diabetes cases compared to 6033 non-diabetic controls. Multilevel longitudinal modeling using either linear growth model for non-diabetic and non-piecewise or piecewise approach including linear or quadratic terms for time for incident diabetic subjects with HOMA2-%S (A) and HOMA2-%B (B) as outcomes. Adjusted for age, sex, ethnicity and study phase. Estimated for a hypothetical population of 72% male, 91% Caucasian aged 63 years at time 0. Error bars show 95% confidence intervals for the fixed effects. Tables show the number of measurements for each year at and before diabetes diagnosis / end of follow-up.