Prevalence and Predictors of Steady-State Hypoxemia in Sickle Cell Disease.

Individuals with sickle cell disease (SCD) may have arterial oxygen desaturation during the steady-state that is mainly due to a right shift of the oxyhemoglobin dissociation curve. This right shift has both non-specific causes (increased concentration of 2,3-DPG due to chronic anemia) and SCD-related causes (an effect of the intracellular concentration of hemoglobin (Hgb) S and an enhanced Bohr effect). Another possible cause is chronic cardiopulmonary disease that may be related to past acute chest syndrome (ACS). We aimed to describe the distribution of steady-state peripheral oxygen saturation (SpO₂) in a large population of children with SCD and to determine whether any simple laboratory or clinical features were predictive of SpO₂. We hypothesized that most of the variation in SpO₂ was not explained by steady-state Hgb alone, and that a history of ACS could explain some of this variability. Using our center’s comprehensive database we identified all subjects with sickle cell anemia (SS), sickle-hemoglobin C disease (SC), sickle-β⁺-thalassemia (Sβ⁺), or sickle-β⁰-thalassemia (Sβ⁰) who had been evaluated within the past 5 years for whom steady-state Hgb concentration, reticulocyte count (retic), and SpO₂ were available. All steady-state values are rolling averages calculated at routine well clinic visits. SpO₂ was determined by pulse oximetry in room air. Individuals receiving chronic transfusions were excluded. Lifetime rates of ACS were known for a subset of subjects with SS and Sβ⁰. A standard multiple regression analysis was performed between steady-state SpO₂ as the dependent variable, and steady-state Hgb and retic, age, and sex as independent variables. 585 subjects were analyzed (390 SS/Sβ⁰, 195 SC/ Sβ⁺; 47% female, 53% male). Mean age was 9.4 years (SD 5.6, range 0.2 – 19.7). Mean SpO₂ was 96.3% (SD 3.0) for SS/Sβ⁰ and 98.7% (SD 1.7) for SC/ Sβ⁺ subjects. The percentage of subjects with SpO₂ <96% and <90% was 33.1 and 2.8 for SS/Sβ⁰ and 3.6 and 0.5 for SC/ Sβ⁺. Bivariate analyses showed no correlation between Hgb and SpO₂ for SC/ Sβ⁺ subjects (N=195, Pearson R=0.024, P=0.74) and no correlation between ACS rate and SpO₂ in those with SS/Sβ⁰ (N=183, Pearson R=−0.043, P=0.56). Thus, only the 390 subjects with SS/Sβ⁰ were included in the multivariate analysis, and ACS rate was not included in the model as an idependent variable. All 4 independent variables (Hgb, retic, age, and sex) contributed significantly to prediction of SpO₂. Altogether, about 45% (adjusted 44%) of the variability in SpO₂ was explained by the model. Multiple correlation coefficient (R = 0.67) showed a significant linear relationship between independent variables and SpO₂ (F = 78.07, p < 0.001). The estimated model can be given as: SpO₂ = 94.24 + (0.58 * Hgb) − (0.16 * Age in years) + (0.64 * Female sex) − (0.20 * Retic in %). In summary, steady-state hypoxemia is common among individuals with SS and Sβ⁰, in whom decreased steady-state SpO₂ is related to decreased steady-state Hgb, increased steady-state retic, increased age, and male sex. This relationship was not found for individuals with SC and Sβ⁺. Only 5% of the variation in SpO₂ was explained by Hgb while controlling for other variables, and ACS rate was not associated with SpO₂. We conclude that most steady-state hypoxemia in individuals with SCD is explained by factors other than chronic anemia, and that hypoxemia appears to be unrelated to prior episodes of ACS.