Year: 2018
Type: article
Abstract: In this issue of the American Journal of Hematology, Willen and colleagues systematically investigate whether hypoxia increases the risk of vaso-occlusive pain events in a carefully done prospective study including 140 children with sickle cell disease (SCD). The authors actually found that a slightly higher rate of pain episodes (IRR 1.1, P = .0004) was associated with higher rather than low nocturnal SpO2 and that pain events were not associated with obstructive sleep apnea.1 This result flies in the face of what is understood about sickle hemoglobin (Hb-S). Namely, that deoxy-Hb-S polymerizes, bringing about the transformation of a normally flexible discoid red blood cell (RBC) into very rigid sickle shaped RBCs (SRBC) that obstruct microvascular flow.2 Thus, we would expect that low SpO2 should be associated with higher risk of painful events. With the exception of one paper,3 hypoxia has not been shown to increase risk of VOC.4, 5 Why would a higher SpO2 be associated with a small but statistically significantly higher VOC rate? This is very intriguing and certainly contrary to the initial hypothesis put forth by the authors, who did not speculate on the significance of these findings. However, their results do not support the hypothesis that SpO2 less than 96% is associated with higher crisis frequency. This finding notwithstanding, the idea, based on the chemistry of Hb-S, that hypoxia might be the trigger for sickle vaso-occlusion and that treatment with oxygen might help persists. The finding of a clinical association between hypoxia and pain crisis would somehow be satisfying, perhaps bringing the "bench to the bedside." Kirkham et al.,6 and subsequently Quinn et al.7 found an association between nighttime hypoxia and CNS events in SCD, and Kirkham later reported that nighttime hypoxia is also associated with VOC.3 However, subsequent studies, including the present one, have not shown a relationship between daytime or nighttime hypoxia and the frequency of pain events.1, 5, 8 Why is there an apparent disconnect between physiology and clinical observation? The explanation is in part because the relationship between sickle events and hypoxia is far more complex than hypoxia alone. During our own research, we were intrigued by the initial data suggesting that nighttime hypoxemia was related to vaso-occlusion,9 and subsequently measured microvascular blood flow and autonomic nervous system (ANS) balance in response to controlled episodes of experimentally-induced hypoxia that were intended to mimic nighttime hypoxic episodes.9 We had hypothesized that hypoxia induced by five breaths of pure nitrogen would transiently decrease microvascular perfusion by inducing Hb-S polymerization. As with the present article, the data did not support our initial hypothesis, and in fact was a complete surprise. Transient hypoxia (Figure 1A) did not induce any change in microvascular perfusion (Figure 1B), but rather caused a very clear, substantial decrease in the parasympathetic part of the ANS balance that mirrored the change in SpO2 in SCD patients, but not in controls (Figure 1D). There were also clear neural-mediated periodic drops in microvascular perfusion (Figure 1B) triggered by stretch receptors in the thorax that correlated perfectly with sighs in SCD patients (one SCD subject: Figure 1B,C), but not in controls. Thus, SCD patients have dysautonomia and a hypersensitive global vasoconstriction response to stretch signals from the thorax, but hypoxia itself does not reduce perfusion. These findings suggest a mechanism that might connect sleep apnea, asthma, and hypoxia to VOC but were not at all what we thought we would find at the outset of these experiments. Changes in a single SCD subject (A, B, C) aligned with average change in parasympathetic activity (cardiac high frequency power: HFP) in 13 SCD subjects and 14 controls (D). There were no changes in microvascular perfusion (B) induced by hypoxia (A) induced by five breaths of pure N2 but clear drop in parasympathetic activity (D). Sighs (C) occurred 3.6 seconds before significant vasoconstriction events (B) in SCD subjects, but not controls [Color figure can be viewed at wileyonlinelibrary.com] The model proposed in the 1970s by Eaton and Hofrichter10 suggests that vaso-occlusion occurs when Hb-S RBC cannot escape the microvasculature into larger diameter vessels before polymerization of deoxygenated Hb-S transforms them into rigid sickle cells. This means that vaso-occlusion depends on the rate of deoxy-Hb-S polymer formation and on blood flow in the microvasculature. The former depends on both tissue oxygen extraction and oxygen supply; the latter on vasoconstriction-vascular tone and blood viscosity. Thus, it may not be surprising that hypoxemia and anemia are related to CNS events given the extremely high oxygen extraction in the brain.11 This balance may not be as critical in the periphery, where oxygen demands are less. If sighs cause drop in peripheral blood flow in SCD, certainly asthma, which is associated with unfavorable outcome in SCD, would affect both the hypoxemia-Hb-S polymerization and local blood flow part of the Eaton-Hofricter relation. Nevertheless, at least based on the current data, nighttime obstructive events do not seem to be sufficient to bring on crisis. It is interesting to note that the obstructive apnea hypopnea index was just outside statistical significance when all subjects were included.1 What accounts for the differences in the relation between SpO2 measured in the fingertip and painful crisis in SCD between various reports and how important is this? We are certain that Hb-S polymerization depends on the state of oxygenation; why then does not fingertip SpO2 predict clinical crisis? The short answer again is that this is far more complex than I described above. The oxygen demand to oxygen delivery ratio depends on hemoglobin level, regional flow and oxygen level, and these parameters are different in different parts of the body that may not be adequately reflected by measurements made in the fingertip or antecubital fossa. We may also be making these measurements at the wrong time with respect to when VOC occurs. The SpO2 measures and polysomnograms were taken months before the pain events. Furthermore, the measurement of SpO2 is somewhat flawed in SCD as pulse oximetry reports oxygen content based on the oxygen dissociation of normal hemoglobin A. SCD patients do not have normal hemoglobin-A, and while the assessment is reasonable most of the time, changes in 2–3-DPG and presence of dyshemoglobins (COHb and metHb) related to hemolytic rate can significantly alter these measurements in some patients and circumstances. Assessing pain in SCD also poses a huge problem for all SCD clinical research. The one paper showing a relation between hypoxemia and pain episodes used duration of hospitalization as an endpoint3 rather than pain episode rate1, 3, 5 used by others. All pain studies are difficult and subject to bias so we might not have a clear explanation for differences in results in different studies, especially if the effect size is small. Regardless, based on the studies at hand, we should not jump to the conclusion that SCD crisis should be treated with supplemental oxygen. Taken at face value, the current data1 certainly would argue against this as higher SpO2 was associated with more pain! Every time I hear on rounds that a new SCD admission is on "O2 for comfort," my skin crawls. There is probably no great harm in this practice, but I have heard spirited discussions on both sides of this point, again with no data. The issue of the role of O2 in SCD is very complex clinically, just like the basic physiology. We have to interpret all of these studies in context of an experimental design where the SpO2 measures were not made in proximity to the painful events. At this point, we have to conclude that there is no clear evidence that hypoxia is associated with painful crisis. However, the possibility of an association has not been excluded for sure either. Regardless, if there is an effect of hypoxia seen clinically on painful sickle events, it must be small. A study investigating SpO2, sleep apnea, and airway reactivity for several days just before painful episodes might be a better design for these questions. However, this would be difficult, if not impossible to do, even though the result may provide the pleasing bench-to-beside relationship we crave. As much as I would like this basic science relation between Hb-S deoxygenation and polymerization to be evident at the bedside, I think the return on investment for such a study would be small for patients. Conversely, oxygen delivery in contrast to SpO2, especially in the brain, is a very different evolving story.11 The differences between the physiological model and the clinical outcome underscore the need to better understand the issues related to the measurement of SpO2 in SCD and its relation to the pathophysiology of the disorder. The current data in conjunction with others highlight the complexity of the physiology and suggests that SpO2 alone is not the major source of variability in SCD. The author appreciates the editorial assistance of Dr. Martine Torres.
Source: American Journal of Hematology
DOI: 10.1002/ajh.25048
Author: Thomas D. Coates
Language: English
Subfield: Genetics
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2790721523