The prevalence of obesity hypoventilation syndrome (OHS) in the unselected obese is unknown. Our objectives were: (i) to determine the prevalence of OHS in ambulatory obese patients not previously referred to a pulmonologist for suspicion of sleep breathing disorders and (ii) to assess whether venous bicarbonate concentration [HCO3-v ] can be used to detect OHS. In this prospective multicentric study, we measured [HCO3-v ] in consenting obese patients attending pathology analysis laboratories. Patients with [HCO3-v ] ≥ 27 mmol/L were referred to a pulmonologist for comprehensive sleep and respiratory evaluations. Those with [HCO3-v ] < 27 mmol/L were randomized to either referral to a pulmonologist or ended the study. For the 1004 screened patients, the [HCO3-v ] was ≥27 mmol/L in 24.6% and <27 mmol/L in 45.9%. A total of 29.5% who had previously consulted a pulmonologist were excluded. A population of 241 obese patients underwent sleep and respiratory assessments. The prevalence of OHS in this population was 1.10 (95% CI = 0.51; 2.27). In multivariate analysis, PaCO2 , forced expiratory volume in 1 s (FEV1 ), apnoea-hypopnoea index (AHI), BMI, use of ≥3 anti-hypertensive drugs, anti-diabetics, proton pump inhibitors and/or paracetamol were related to raised [HCO3-v ]. The prevalence of OHS in our obese population was lower than previous estimations based on hospitalized patients or clinical cohorts with sleep breathing disorders. Apart from hypercapnia, increased [HCO3-v ] may also reflect multimorbidity and polypharmacy, which should be taken into account when using [HCO3-v ] to screen for OHS.