The effect of hemoglobin level on the prognostic value of peak exercise oxygen uptake differs between obese and non-obese patients with heart failure

Abstract Funding Acknowledgements Type of funding sources: None. Background Obesity and low hemoglobin content (Hgb <13 and <12 g/dL, men/women, respectively) are commonly observed in patients with heart failure and reduced ejection fraction (HFrEF). This overlap typically yields reductions in peak exercise oxygen uptake (V̇O2peak) which exceed those observed in both non-obese and obese patients exhibiting normal Hgb levels. However, when matched for body size, non-obese patients exhibiting low Hgb levels also demonstrate worse V̇O2peak impairment than normal Hgb level counterparts. To date, it is unclear whether the effect of Hgb level on V̇O2peak also affects the association between V̇O2peak and mortality risk in both non-obese and obese patients. Purpose To assess whether V̇O2peak that is predictive of mortality risk is directly impacted by Hgb level regardless of whether a patient is non-obese or obese. Methods Adults with HFrEF (n=346; LVEF ≤40%; 76% men) completed cardiopulmonary exercise testing (CPET) as part of routine clinical management. The majority of patients performed Naughton or Modified Naughton CPET protocols. Patients with a BMI ≥30 kg/m2 were classified as obese (n=135; 81% men). Patients were followed up to one-year for the endpoint of all-cause mortality. Results Age (63±7 vs 61±8 yrs), total body weight (77±13 vs 107±15 kg), BMI (25.2±3.0 vs 34.5±4.0 kg/m2), and Hgb (12.6±1.8 vs 13.1±1.8 g/dL) differed between non-obese and obese patients (all P<0.01), respectively. There were no significant differences between non-obese and obese patients for % men (73 vs 81 %), % ischemic etiology (46 vs 42 %), and LVEF (23±8 vs 24±9 % (all P>0.05), respectively. Peak exercise RER (1.13±0.12 vs 1.10±0.11), V̇O2peak (13.0±3.8 vs 11.9±3.3 mL/kg/min), and absolute V̇O2peak (1.01±0.37 vs 1.28±0.40 L/min) differed between non-obese and obese patients (all P<0.02), respectively. Over the 1-year tracking period, 25 and 13 deaths were observed and crude survival was 85% and 88% in non-obese and obese patients (Log-rank, P=0.465), respectively. In obese patients, V̇O2peak and Hgb interacted (χ2=4.855, P=0.028), resulting in up to a 22% reduction in all-cause mortality risk per 1.0 mL/kg/min rise in V̇O2peak for Hgb levels ≤13.1 g/dL. In non-obese patients, V̇O2peak and Hgb did not interact (χ2=0.001, P=0.977), and when Hgb was removed from the Cox regression, this resulted in up to a 29% reduction in all-cause mortality risk per 1.0 mL/kg/min rise in V̇O2peak (χ2=23.618, P<0.001). Conclusion Demonstrating low Hgb and V̇O2peak levels yields comparable effects on all-cause mortality risk in non-obese and obese patients. However, at normal Hgb levels, V̇O2peak loses prognostic value in the obese, suggesting for these patients other CPET variables should be used to complement V̇O2peak when CPET is built into risk stratification models.