Abstract
AIMS: Obesity, systemic inflammation, and hyperinsulinemia are all features of metabolic syndrome and frequently occur together. We aimed to evaluate the association of body mass index (BMI) with incident non-communicable chronic disease (NCD) or all-cause mortality, independent of C-reactive protein (CRP) and fasting insulin. METHODS: This prospective population-representative cohort included Canadian residents aged ≥18 years from the Canadian Health Measure Survey. The exposures were BMI, CRP, and fasting insulin and the outcomes were mortality and incident NCDs. Results were externally validated using data from the United States (NHANES) and the United Kingdom (UK Biobank). We report hazards ratios (HRs) from Cox models using percentiles for the continuous exposures: 99th, 95th, 85th, 50th, 15th, 1st versus the 5th percentile. The fifth percentile for BMI was 19.6 kg/m(2), which falls within the "healthy" range. RESULTS: Of ~8280 participants: 24.4% had obesity (BMI ≥30 kg/m(2)), 56.5% had inflammation (CRP >1 mg/L), and 36.6% had hyperinsulinemia (fasting insulin >75 pmol/L). Metabolic syndrome (≥1 feature) was more common without obesity (42.8%; ~3544 participants) than with obesity (22.9%; ~1896 participants). Participants were followed for a median of 6.7 years; 3.1% died and 7.4% developed an NCD (including 1.8% with cardiovascular disease). After adjustment for age, sex and smoking, CRP was positively associated with mortality (HR vs. referent for 1st and 99th %ile: 0.72 [95% confidence interval, CI 0.61, 0.84] and 7.67 [95% CI 2.89, 20.38]), as was fasting insulin (0.88 [95% CI 0.78, 0.99] and 3.22 [95% CI 1.12, 9.29] respectively). BMI was negatively associated with mortality (HR vs. referent for 1st and 95th %ile: 1.45 [95% CI 1.11, 1.88] and 0.46 [95% CI 0.22, 0.97]), although HR for the 99th %ile was non-significant (0.51 [95% CI 0.20, 1.27]). All three exposures were mostly positively associated with incident NCD but varied depending on the type of NCD. CONCLUSIONS: After adjustment for confounding by systemic inflammation and hyperinsulinemia, obesity was associated with a lower risk of death but with a higher risk of incident NCD. As there are more people with metabolic syndrome without obesity than with obesity, future research should prioritise the study of how to best diagnose, monitor and treat inflammation and hyperinsulinemia rather than obesity.