Hemoglobin Oxygen Saturation Discrepancy Using Various Methods in Patients with Sickle Cell Disease.

Objective: Sickle cell disease (SCD) is characterized by chronic hemolytic anemia and vaso-occlusive painful crises. Recurrent pulmonary microinfarction as well as increased predisposition to lung infection during vaso-occlusive crisis places these patients at high risk for hypoxemia. Furthermore, because hypoxemia is a trigger for sickling and vaso-occlusion, the detection of arterial hypoxemia and its correction are very important during painful episodes. We previously assessed reliability of pulse oximetry in pateints with SCD (

Methods: Eighteen adult hypoxic patients with SCD hospitalized during a painful episode at a tertiary care institution served as the study group, whereas 12 non-SCD hypoxic African-American patients admitted with various cardio-pulmonary diagnoses served the control group. The hemoglobin oxygen saturation was determined simultaneously by pulse oximetry (SpO₂), co-oximetry (SO₂ [functional oxyhemoglobin saturation] & FO₂Hb [oxyhemoglobin fraction]) and by calculation (SaO₂) using a normal oxyhemoglobin dissociation curve in the 2 groups.Bland-Altman analysis was carried out to evaluate agreement of various methods in the two groups. Spearman Correlations were calculated for descriptive purposes. As a secondary analysis, McNemar’s Test was used to compare the proportion of ‘critically hypoxic’ patients between two measurement techniques.

Results: The two groups were different with respect to age, systolic blood pressure, body temperature, hemoglobin level, serum bilirubin level, blood pH, methemoglobin and carboxyhemoglobin levels. Mean differences between various methods for measuring hemoglobin oxygen saturation in patients with SCD were significantly larger than the control group. Mean bias between SpO₂ and SO₂, and SpO₂ and FO₂Hb in patients with SCD were −3.1 ±4.4 (95% CI −11.7 to 5.5) and 2 ±4.1 (95% CI, −6 to 10) respectively compared to −1.4 ±1.4 (95% CI, −4 to 1.4) and 1.2 ±1.5 (95% CI, −1.8 to 4.2) in control group. A bias of −4.5 ±4 (95% CI, −12.3 to 3.8) between SpO₂ and SaO₂ was noted in patients with SCD compared to −0.08 ±2.1 (95% CI, −4.2 to 4.1) in control. Spearman correlations between various methods in control group were consistently higher (0.82 to 0.99) than patients with SCD (0.59 to 0.89). McNemar’s test for comparing the proportion of "Critically Hypoxic" disposition using a cut-point of 90% oxyhemoglobin saturation showed that there was poor agreement among methods in the sickle cell group whereas acceptable agreement was noted among methods in the control group.

Conclusion: Patients with SCD during vaso-occlusive crisis have discrepancies in oxyhemoglobin saturation measurements by various methods. Abnormal pulse oximetry values in these patients should be interpreted cautiously and supplemented by PO₂ and co-oximetry.