Abstract Background Obesity is a widespread medical condition consisting in a BMI > 30 kg/m2. The link between BMI, exercise performance and cardiovascular risk is not well defined on large set of apparently healthy subjects. Aim In this population-based trial of apparently healthy individuals we aimed at defining the exercise gas exchange phenotype by cardiopulmonary exercise testing (CPET) according to BMI strata. Methods 932 subjects (54.7% females) presenting with none to five major cardiovascular risk factors (MCVRFs; hypertension, diabetes, tabagism, dyslipidaemia, BMI> 25), underwent CPET and were divided in 3 categories (normal weight; NW, overweight; OW and obese; Ob). Cohort analysis was carried out by dividing and analysing patients into groups based on BMI, sex, and number of MCVRFs. Peak VO2 was considered normal for a predicted value > 80%. Composite endpoint was defined as the combination of the following events: cardiovasular mortality, cardiovascular hospitalization, coronary artery disease (CAD), heart failure (HF), atrial fibrillation (AF), percutaneous coronary intervention (PCI), TAVI (Transcatheter aortic valve implantation), cardiac surgery, acute pulmonary oedema or stroke/transient ischemic attack (TIA). Results the number of MCVRFs significantly increased (CI 95%, p-value 0,009) through BMI strata being in NW, OW and Ob 1.6, 2.7, 2.9 respectively. Peak VO2 significantly (CI 95%, p-value <0,001) decreased with rise in BMI (21.8ml/min/kg, 18.7 ml/min/kg, 15.5 ml/min/kg for NW, OW and Ob, respectively; p-value <0,001). A significant correlation between the increase in number of MCRVFs and the decrease in peak VO2 was documented for BMI<30 (-0.44 for NW and -0.30 for OW; p-value <0,001), whereas Ob exhibited the lowest mean peak VO2 which was unaffected by the increasing number of MCVRFs. (figure 1). HRR (Heart rate reserve) was significantly lower in Ob subjects compared to OW and NW (respectively 0.59, 0.68, 0.72 p-value <0,001). Regression models showed that neither risk factors nor CPET variables were predictive for composite endpoint in NW. Interestingly, peak VO2 (OR 0,93 [CI 0,86-0,99] p-value 0,039) and VE/VCO2 slope (OR 1,06 [CI 1,00-1,11] p-value 0,047) were predictive of composite endpoint just in OW with Ob subjects showing similar trend not reaching statistical power; Diabetes was predictive of composite endpoint in Ob cohort. KM curves reported higher event rates in OW subjects with higher VE/VCO2 slope and lower peak VO2. (figure 2) Conclusions As far as obesity is concerned, regardless of the number of MCVRFs, BMI is the single strong determinant of exercise performance and gas exchange phenotype. In OW and NW cohorts, the cumulative number of MCVRFs impacts on worse exercise performance and gas exchange inefficiency. Lower peak VO2 and higher VE/VCO2 slope are predictive of CV composite endpoint in OW and a similar trend was observed in Ob.BMI and peak VO2KM BMI 25-30