Associations of Physical Activity Intensities with Markers of Insulin Sensitivity

Background Objectively measured physical activity (PA) intensity has traditionally been categorized as light, moderate, and vigorous using laboratory calibrated cut points. The relative contribution of time spent across a spectrum of accelerometer-determined intensities on health outcomes is less clear. Purpose This study aimed to assess the relationship between objectively measured PA intensity on a continuous scale and markers of insulin sensitivity (IS). Methods Participants at high risk of type 2 diabetes mellitus were recruited from primary care (Leicestershire, UK). PA was measured using an ActiGraph accelerometer. Fasting and postchallenge glucose and insulin levels were assessed using an oral glucose tolerance test. IS was calculated using the Matsuda-IS and the HOMA-IS indices. Log-linear regression modeling was used to assess the relationship between PA intensity, in increments of 500 counts per minute, with markers of IS. Models were controlled for known confounders. Results Complete data were available for 569 participants. PA intensity was favorably associated with fasting and 2 h of insulin and IS, with the association increasing in magnitude with each increment of 500 counts per minute. Differences in HOMA-IS per 10 min of PA ranged from 12.4% (95% confidence interval = 3.7%-21.8%) to 26.8% (11.0%-44.7%) within the moderate-intensity PA category (from 2000-2499 to 3500-3999 counts per minute). For Matsuda-IS, these differences were 22.0% (10.3%-34.9%) and 34.7% (13.9%-59.3%), respectively. Significant associations for fasting insulin were no longer observed after controlling for body mass index, whereas differences associated with 2-h insulin and IS were attenuated but still significant. Conclusion PA of any intensity may positively influence glucose regulation and insulin sensitivity in individuals at high risk of type 2 diabetes mellitus in a dose-response manner. Further research is required to identify the intensity thresholds at which clinically relevant benefits occur in this population.