Relationship Between RDW and Disease Burden in Obstructive Sleep Apnea

Stanley Abraham1 • Madeleine Basist1 • Sean Duenas1 • Vincent Dong2 • Sam Rusk3 • Yoav Nygate3 • Luis Quintero1 • Stella Hahn1 • Harly Greenberg1 • Arunabh Talwar1

Introduction

Obstructive sleep apnea (OSA) has significant physiologic consequences. The Apnea-Hypopnea Index (AHI) is used to define severity, but AHI only reflects the rate of sleep-disordered breathing events and does not correlate well with clinical outcomes that arise from hypoxia. Percentage of cumulative time with oxygen saturation below 90% (T90) and hypoxic burden (HBI) more accurately reflect nocturnal hypoxia and may be more suitable correlates to morbidity associated with sleep apnea. Rising Red Cell Distribution Width (RDW) reflects dysregulation in erythrocyte homeostasis seen with metabolic derangements like oxidative stress and inflammation which has been shown to correlate with mortality in cardiovascular disease. Previous work demonstrates a linear relationship between RDW and AHI. We postulate there is a positive correlation between RDW, T90 and HBI.

Methods

A retrospective analysis of 1797 participants at Northwell with Home Sleep Apnea Testing or in-lab Polysomnography data and complete blood counts (CBC) within 6 months of the sleep study was performed. Subjects with chronic hypoxemia or hematologic conditions that may affect CBC were excluded. Ordinary least squares (OLS) regression was performed using RDW as the dependent variable and AHI, T90, HBI and mean saturation of peripheral oxygen (SpO2) as independent variables.

Results

OLS analysis revealed a statistically significant positive correlation between RDW and T90 (coef=.807 (CI 0.426,1.188), p-value=0.000034), HBI (coef=0.000033 (0.000019,0.000054), p-value=0.002)) and AHI (coef= 0.0062 (0.0028, 0.0097), p-value= 0.0004). A negative correlation was noted between mean SpO2 and RDW (coef=-0.047 (-0.078,-0.162), p-value=0.003).

Conclusion

There is a positive correlation with RDW and sleep apnea markers of disease burden, including T90, HBI, and mean SpO2. A 100% change in T90 (0 – 100%) in a patient will result in an increase in RDW by 0.8 and a 100% change in HBI (0 – 500 %min/hour) will result in an increase in RDW by 1.5. AHI was noted to poorly correlate with RDW. An AHI of 100 resulted in a 0.62 increase in RDW versus an AHI of 0. For the range of mean SpO2 observed (range 80 – 100%) the expected change in RDW is 0.94. This relationship requires further exploration in future studies.