Nocturnal Hypoxemia (not sleep apnea) May Drive Reticulocytosis in Sickle Cell Disease

Introduction: Obstructive sleep apnea is a ventilatory defect that affects up to 10% of children with sickle cell disease (SCD). Sleep apnea has been implicated in cardiovascular disease in non-SCD adults and in cerebrovascular disease in children with SCD. In addition, prolonged hemoglobin oxygen desaturation from sleep disorderedbreathing would be predicted to increase sickle polymer formation and, possibly, activity of disease. We evaluated symptomatic adults from our SCD clinic for sleep apnea, in an effort to identify treatable disease modifying conditions.

Methods: 46 adults with SCD were evaluated by polysomnography (PSG) between 2012 and 2014. Subjects were referred for symptoms or signs suggestive of sleep disordered breathing, including nighttime pain, snoring, early morning headache, narrow hypo-pharynx, and/or excessive reticulocytosis. Clinical characteristics of homozygous (HbSS, n=34) and compound heterozygous (HbSC and HbSBeta+Thalassemia, n=12) subjects are shown in Table 1.

Results: Sleep apnea was diagnosed if the apnea/hypopnea index (AHI) was elevated: 5-15 (mild), 15-30 (moderate), or >30 (severe).

Over half of our patients had sleep apnea by these criteria (16/34 HbSS and 7/12 compound heterozygotes). An elevated 3% oxygen desaturation index , i.e. >15 episodes per hour during which oxygen saturation dropped by 3% or more, was seen in one-third (14/43) of all evaluable subjects, and was 17 (median, range 1.2- 192.4) in subjects with, compared with 4.6 (median, range 1-23.7) in subjects without, sleep apnea (P<.001). Sleep apnea was moderate or severe in 5/34 (14.7%) and 3/12 (25%) of screened HbSS and compound heterozygote patients, respectively. However, the presence of sleep apnea, compared with the absence of sleep apnea, was not associated with differences in measured clinical variables including WBC, Hgb, absolute reticulocyte count, LDH, urinary albumin-to-creatinine ratio, or tricuspid regurgitant jet velocity.

We then examined 8 subjects (7 HbSS, 1 SBeta+ thal) whose PSG showed >10% of sleep time afflicted by nocturnal hypoxemia. A median of 33.75% (range 16-98.6%) of sleep time was spent at <90% oxygen saturation in these subjects. Only 3 of 8 subjects with nocturnal hypoxemia had sleep apnea, based on an AHI >5, which was not different than the group as a whole (P=0.7). These 8 subjects had strikingly increased reticulocyte counts, compared with 38 subjects without nocturnal hypoxemia (median 460, range 100-923 (x10³)/μL compared with median 303, range 76-733 (x10³)/μL respectively, p=0.04), as well as a trend toward a higher LDH (median 488, range 140-752, compared with median 318, range 139-823, P=.08). Of note, hemoglobin levels were not different between subjects with and without nocturnal hypoxemia (mean 8.4±1.5 g/dL compared to mean 9.5±2.2 g/dL, respectively, p=0.17).

Conclusions & Discussion: These studies suggest that, while sleep apnea is common in adults with SCD, greater physiological relevance may arise instead from prolonged nocturnal hypoxemia and associated reticulocytosis. We speculate that significant nighttime hypoxemia and hemoglobin desaturation may stimulate both endogenous erythropoietin (and therefore reticulocyte) production and HbS polymer formation and, possibly, hemolysis. This condition, while identifiable by conventional PSG, is not coincident with sleep apnea, and may be modulated by nighttime oxygen supplementation. These findings deserve confirmation in a larger cohort of patients.