The authors studied consecutive severely obese patients (BMI ⩾35 kg/m²) admitted to the medicine services of three teaching hospitals. Arterial blood gas tensions were measured to assess the prevalence, predictors, and outcomes of patients with obesity associated hypoventilation (OAH), defined as Paco₂ ⩾43 mm Hg and pH ⩽7.42, with no other obvious cause of hypoventilation. Of 4332 consecutive admissions, 277 (6%) had a BMI of ⩾35 kg/m² and were screened. 127 were excluded for various reasons: 85 were unwilling to have an arterial blood gas measurement or could not provide consent, 32 used opiates, and 10 had a prior lung resection or a reduced FEV₁/FVC ratio (<50%).

Of the remaining 150 patients, OAH was found in 47 (31%). Compared with severely obese patients without OAH, patients with OAH were heavier, sleepier, more likely to have erythrocytosis, and had a lower FVC. They were more likely to require invasive mechanical ventilation (6% v 0%) and to require discharge to a long term care facility (19% v 2%). They also had a greater 18 month mortality rate (23% v 9%) which persisted after controlling for a variety of potential confounders. Surprisingly, only 23% of patients (n = 11) were given a discharge diagnosis of OAH, and only six of these were discharged with a recommendation to receive long term treatment (non-invasive nocturnal ventilation or tracheostomy).

In severely obese patients admitted to a medical service, OAH is common, associated with adverse outcomes, and under-appreciated. Clinicians should recognise the possibility of OAH in severely obese patients and consider performing arterial blood gas analysis.