Hemoglobin oxygen saturation discrepancy using various methods in patients with sickle cell vaso-occlusive painful crisis

Objective: To evaluate agreement among various methods for measuring oxyhemoglobin (O(2)Hb) saturation in adult hypoxic patients with sickle cell disease (SCD) during painful vaso-occlusive crisis and to compare those results with a control group. Patients and methods: The hemoglobin oxygen saturation was determined simultaneously by pulse oximetry (SpO(2)), co-oximetry [SO2 (functional oxyhemoglobin saturation) and FO(2)Hb (oxyhemoglobin fraction)] and by calculation (SaO(2)) using a normal O(2)Hb dissociation curve in 18 adult patients with SCD during vaso-occlusive crisis and 12 non-SCD patients with various cardiopulmonary diagnoses. The method proposed by Bland and Altman was used to evaluate agreement of various methods in each of the two groups. Results: Mean differences between various methods in patients with SCD were significantly larger than the control group. Limits of agreement (LOA) were also wider in the SCD group than in the control group. Mean bias between SpO(2) and SO2, and SpO(2) and FO(2)Hb in patients with SCD were -3.1 +/- 4.4 (LOA: -11.9 to 5.7) and 2 +/- 4.1 (LOA: -6.2 to 10.2) respectively, compared with -1.4 +/- 1.4 (LOA: -4.2 to 1.4) and 1.2 +/- 1.5 (LOA: -1.9 to 4.3) in the control group. A mean bias of -4.5 +/- 4 (LOA: -12.5 to 3.5) between SpO(2) and SaO(2) was noted in patients with SCD compared with -0.1 +/- 2.1 (LOA: -4.3 to 4.1) in the control group. The width of LOA for various methods in patients with SCD ranged from 9.8 to 17.6 compared with 1.3 to 8.4 in the control group. Conclusion: Patients with SCD during vaso-occlusive crisis have discrepancies in O(2)Hb saturation measurements by various methods. Abnormal pulse oximetry values in these patients should be interpreted cautiously and supplemented by arterial blood gas analysis and co-oximetry.