PURPOSE:The pressures delivered by autotitrating continuous positive airways pressure (CPAP) devices not only treat obstructive sleep apnoea (OSA) effectively but also give potentially interesting physiological information about the forces impinging on the pharynx. In earlier work from this unit, we used correlations between autoCPAP pressure and both OSA severity and obesity, to construct an algorithm to estimate the fixed CPAP pressure a patient required for subsequent clinical use. We wished to discover if these relationships could be reliably extended to a much more obese group. METHODS:We performed a prospective cohort study in an obese population. Measurements of obesity were made, OSA severity was recorded, and the 95th centile autoCPAP pressure was recorded during 1 week of autoCPAP. Spearman's rank correlation was performed between measurements of obesity and autoCPAP pressure, and between OSA severity and autoCPAP pressure. RESULTS:Fifty-four obese individuals (median body mass index (BMI) 43.0 kg/m(2)), 52 % of whom had OSA (apnoea-hypopnoea index (AHI) ≥ 15), had a median 95th centile autoCPAP pressure of 11.8 cmH2O. We found no significant correlation between autoCPAP pressure and neck circumference, waist circumference or BMI. There was a moderate correlation between autoCPAP pressure and OSA severity (AHI r = 0.34, p = 0.02; oxygen desaturation index (ODI) r = 0.48, p < 0.001). CONCLUSIONS:In this population, neither BMI nor neck circumference nor waist circumference is predictive of autoCPAP pressure. Therefore, the previously derived algorithm does not adequately predict the fixed CPAP pressure for subsequent clinical use in these obese individuals. In addition, some subjects without OSA generated high autoCPAP pressures, and thus, the correlation between OSA severity and autoCPAP pressure was only moderate.