Two parameters statistically explain blood glucose decrease during exercise at steady state in type 1 diabetics: Pre-exercise blood glucose and insulinemia

Introduction. - Exercise hypoglycaemia in type-1 diabetics remains hard to predict, because the precise mechanisms governing blood glucose changes during exercise in these patients remain incompletely understood. We previously demonstrated that a pre-exercise blood glucose value G(0) > 250 mg/dL does not mean as commonly believed that exercise will exacerbate hyperglycemia, and that the rate of carbohydrate oxidation at exercise (glucose pulse) does not predict hypoglycemia. In this study, we aimed at defining the role of insulin levels in this process. Methods. - Twenty-three type-1 diabetics (age 22-67 years, average weight: 68.6 +/- 2.03 kg, average height: 168 +/- 1.45) treated with basal-bolus or insulin pump underwent a steady state exercise at 50% of theoretical maximal power output (30 to 120 watts) with monitoring of blood glucose, free insulin and exercise calorimetry. Summary of facts. - Pre-exercise blood glucose Go (ranging between 111 and 422 mg/dL, averaging 2.23 +/- 0.95 g/L) decreased on average by -0.95 +/- 0.26 g/L and this decrease occurred in 21 subjects (91% of cases). This change in blood glucose is negatively correlated (not positively) to GO (r = -0.525 P <0.01), so that the higher is G(0), the deeper the blood glucose falls. The oxidation of carbohydrates is strongly correlated to the total work performed (kJ) (r = 0.875 P < 0.001) but does not predict changes in blood glucose. Insulin levels at the beginning of the test ranges between 6 and 66 mu U/mL. Subjects whose pre-exercise free insulin was less than 25 U/mL lowered their blood glucose in 67% cases (mean -0.51 +/- 0.24) while if preexercise free insulin exceeds 25 mu U/mL glucose always decreases, from 0.5 to 2.8 g/L (mean -1.7 +/- 0.42). A multivariate analysis shows that preexercise blood glucose (r = -0,671 P = 0.0098) and preexercise free insulin (r = -0.524 P = 0.033) are two variables contributing independently to lower blood glucose. Conclusion. These data confirm that during prolonged steady state exercise, carbohydrate oxidation does not predict the decrease in blood glucose. This decrease is more pronounced when basal blood glucose is high, challenging the common belief that G0 > 250 mg/dL counterindicates exercise in diabetics since it will exacerbate hyperglycemia. This decrease also appears in multivariate analysis to be determined by free insulin levels at the beginning of exercise. (C) 2011 Published by Elsevier Masson SAS.