Insulin resistance without coexisting hyperglycaemia is not independently associated with cardiorespiratory fitness or its components in women

Abstract

Abstract Background Insulin resistance (IR) is a globally prevalent condition in which tissues’ response to insulin is diminished. Cardiorespiratory fitness (peak O2 uptake (VO2peak)) is determined by Fick’s equation (VO2 = cardiac output (CO) × arteriovenous O2 difference (Ca-vO2) = heart rate × stroke volume (SV) × Ca-vO2) and diffusion law (VO2 = diffusive O2 conductance (DO2) × constant k × venous O2 pressure (PvO2)). Several studies suggest an association between IR and decreased VO2peak, but the association has been questioned by analyses taking account of body size and composition, and potential mechanisms behind the association are unknown. Purpose To test a hypothesis that IR is not independently associated with VO2peak or its components. Methods Fifty of 144 women, who had participated in baseline experiments of three intervention studies, were included in this retrospective cross-sectional study; see Table 1 for subject characteristics. The inclusion criteria were age 18-40 years, body mass index 18.5-40 kg/m2, Caucasian origin, and available data on key measures. The exclusion criteria were diseases, medications, and other factors possibly affecting VO2peak. We measured haemoglobin, fasting glucose, and fasting insulin concentrations (blood), assessed IR (homeostasis model assessment (HOMA-IR)), and measured body size and composition (bioimpedance). The subjects performed an incremental cycling test to volitional fatigue; we measured ventilatory gas exchange (spiroergometry), arterial O2 saturation (oximetry), and SV and CO (impedance cardiography), and calculated Ca-vO2, DO2, and PvO2. We analysed VO2peak and its components as absolute (e.g. L/min) and ratiometrically scaled values (e.g. mL/min/kg), and when appropriate, also as allometrically scaled values (e.g. mL/min/kg^x, where 0 < x < 1) after identifying optimal variable-specific exponents. Results VO2peak scaled to body mass (BM) as well as SVpeak and COpeak scaled to body surface area (BSA) were inversely associated with the body size variables they were scaled to (all P < 0.013). Contrariwise, this was not the case when we optimally scaled VO2peak to fat-free mass (FFM), SVpeak to FFM^0.55, COpeak to FFM^0.50, and DO2peak to FFM (all 0.066 < P < 0.977). HOMA-IR was inversely associated with VO2peak scaled to BM but not in a multivariate analysis, whereas HOMA-IR was not associated with VO2peak scaled to FFM (Table 2); we made alike observations with SVpeak and COpeak (data not shown). Further, VO2peak scaled to BM as well as SVpeak and COpeak scaled to BSA were or tended to be up to 15 % higher in a group with HOMA-IR<3.0 than in a group with HOMA-IR≥3.0, while other peak exercise data did not differ between the groups (Table 1). Conclusion Body size and composition significantly affect conclusions drawn regarding VO2peak and its components. Taking this into account, IR without coexisting hyperglycaemia is not independently associated with VO2peak or its components in women.Table 1Table 2