Adiposity, mortality, and disease risk: insights from bioimpedance analysis and magnetic resonance imaging

Basic anthropometric (BA) indicators of adiposity, such as body mass index, may not fully capture disease risk. Whether more advanced anthropometric measurements derived from bioimpedance analysis (BIA) or magnetic resonance imaging (MRI) enhance our understanding of the relationship between adiposity and health-related outcomes is debated.

We used data from 40,338 participants from the UK Biobank imaging sub-study with anthropometric measurements derived from BIA and abdominal MRI, in addition to BA indicators, to evaluate their discriminatory performance. We studied the relationship between these adiposity indicators and all-cause mortality, risks of cardiovascular diseases (CVDs), obesity-related cancer, overall cancer, and type 2 diabetes (T2D) using Cox models adjusted for established risk factors. For each health-related outcome, relevant anthropometric indicators were selected using a stepwise approach, and the discriminatory power of each model was evaluated with cross-validated C-indexes.

MRI-derived organ morphometry indicators moderately improved risk discrimination for T2D (cross-validated C-index increased from 0.83 [95% CI: 0.81, 0.84] to 0.85 [0.83, 0.86]; adjusted p = 4.05E - 6) and obesity-related cancers (from 0.59 [0.57, 0.62] to 0.61 [0.59, 0.64]), albeit with borderline significance (adjusted p = 0.053). Although improved discrimination was also observed for overall cancer and all-cause mortality compared to BA alone, differences were not statistically significant (all adjusted p > 0.200). Conversely, the inclusion of BIA indicators generally did not lead to improved discriminatory power.

MRI-derived organ morphometry indicators may provide information beyond BA indicators for the assessment of adiposity and its association with risk of some health-related outcomes.