Abstract
OBJECTIVES: To investigate the agreement between pulse oximetry (SpO(2)) and oxygen saturation (SaO(2)) measured by CO-oximetry on arterialised earlobe blood gas (EBG) in children and adolescents with sickle cell disease (SCD). DESIGN AND SETTING: We retrospectively reviewed 39 simultaneous and paired SaO(2) EBG and SpO(2) measurements from 33 ambulatory patients with SCD (32 subjects with Haemoglobin SS and one with Haemoglobin Sß(+), 52% male, mean±SD age 11.0±3.6, age range 5-18). Measurements were performed between 2012 and 2015 when participants were asymptomatic. Hypoxaemia was defined as SaO(2) ≤93%. A Bland-Altman analysis was performed to assess the accuracy of SpO(2) as compared with EBG SaO(2). RESULTS: The mean±SD SpO(2) and SaO(2) values in the same patients were, respectively, 93.6%±3.7% and 94.3%±2.9%. The bias SpO(2)-SaO(2) was -0.7% (95% limits of agreement from -5.4% to 4.1%) and precision was 2.5%. In 9/39 (23%) cases, the difference in SpO(2)-SaO(2) was greater than the expected error range ±2%, with SaO(2) more often underestimated by SpO(2) (6/9), especially at SpO(2)values ≤93%. Thirteen participants (33%) were hypoxaemic. The sensitivity of SpO(2) for hypoxaemia was 100%, specificity 85% and positive predictive value 76%. CONCLUSIONS: Pulse oximetry was inaccurate in almost a quarter of measurements in ambulatory paediatric patients with SCD, especially at SpO(2)values ≤93%. In these cases, oxygen saturation can be confirmed through EBG CO-oximetry, which is easier to perform and less painful than traditional arterial blood sampling.