Commentary: Intermittent Hypoxia Severity in Animal Models of Sleep Apnea

Obstructive sleep apnea (OSA) is a common disorder that leads to problems including intermittent hypoxia (IH) and arousals from sleep. To simulate consequences of OSA, some studies expose rodents to IH with the intention of simulating the oxygen profile experienced by OSA patients. Some IH experiments induce hemoglobin oxygen saturation (SaO₂) to fall to 50–70% during the nadir phase. In a recent review, we stated that this degree of hypoxemia is more severe than that experienced by typical OSA patients (Chopra et al., 2015). Farre et al. challenged this statement (Farré et al., 2018), pointing out that small mammals such as rodents have a right-shifted oxyhemoglobin dissociation curve (ODC) compared to humans (Schmidt-Nielsen, 1970). They argue that the higher arterial partial oxygen pressure (PaO₂) for a given SaO₂ confers mice with “better oxygen reserve.” To achieve PaO₂ nadir values similar to those experienced by severe OSA patients they contend, “SaO₂ in mice should be much lower than the SaO₂ observed in patients.” We disagree with this statement, which relies on PaO₂ as an indicator of tissue oxygenation. Arterial O₂ content (CaO₂) is determined by classic formula:

$$\begin{array}{l}\text{Ca}{\text{O}}_2=(1.34\text{ x Hb x Sa}{\text{O}}_2)+(0.003\text{ xPa}{\text{O}}_2)\end{array}$$

Systemic O₂ delivery (DO₂) is the product of blood flow (Q) and CaO₂. From these equations, it is apparent that PaO₂ as it determines the amount of dissolved O₂ gas, itself has a negligible contribution to CaO₂ or DO₂. As oxygenated blood reaches target tissues, O₂ dissociates from hemoglobin to maintain a favorable capillary-to-tissue PO₂ driving gradient. When O₂ diffuses into tissues, systemic capillary PaO₂ falls, and is replenished by upstream oxygenated hemoglobin. Aerobic metabolic processes then consume cell O₂. The balance between O₂ supply and demand determines the cellular PO₂. Therefore, cellular PO₂–the parameter that truly determines oxygen adequacy - depends upon DO₂, capillary density, and rates of O₂ utilization. A right-shifted ODC means that O₂ is unloaded from hemoglobin more rapidly, but this does not increase the total amount of O₂ delivered.

To propose that mice should have their SaO₂ lowered in order to achieve a PaO₂ equal to that of humans is tantamount to suggesting that one should load a truck to only half its capacity, because it unloads boxes twice as quickly. If we invoke this logic, and use the figure provided by Farre et al., mouse SaO₂ would have to be decreased to ~75% to match a normoxic human PaO₂ at ~65 mm Hg. Mice exposed to hypoxemia of this magnitude exhibit robust erythrocytosis (Fagan, 2001) and activate anaerobic metabolic pathways (Jun et al., 2012, 2014) indicating that mice are effectively hypoxic at a higher PaO₂ than humans. Their higher PaO₂ may be necessary to maintain adequate tissue PO₂ (as opposed to having “better oxygen reserve”), as mice have dramatically higher mass-specific metabolic rates than humans. Therefore, we should be targeting equivalent drops in SaO₂ between humans and rodents if the goal is to reduce distal O₂ delivery to the same extent.

Implicit in the argument by Farre et al. is that a right-shifted ODC is advantageous in mitigating effects of low SaO₂. There is no evidence that we could find to support this argument. A right-shifted ODC is advantageous in states such as shock or hemorrhage ensuring O₂ is maximally transferred to ischemic tissues (Agostoni et al., 1975; da Luz et al., 1975; Cornum et al., 1998; Morgan, 1999). Conversely, transfusion of blood depleted of 2,3 diphosphoglycerate to left-shift the ODC (Riggs et al., 1973) decreases tissue oxygen supply. These examples pertain to conditions when hemoglobin is fully O₂ saturated (i.e., SaO₂ is constant). What is the effect of shifting the ODC during hypoxemia (e.g., high altitude, OSA)? Here, effects of hemoglobin O₂ affinity are not straightforward. Left shifting the ODC increases O₂ uptake in the pulmonary capillaries, but compromises peripheral tissue O₂ unloading. At high altitude, this trade-off is advantageous for survival; the “tipping point” occurs when O₂ uptake becomes diffusion limited (Storz and Moriyama, 2008). Indeed, rats exposed to severe hypoxia survived longer with a left-shifted ODC (Eaton et al., 1974). The rightward ODC curve of rodents may actually be counter-productive in the setting of ambient hypoxia.

In conclusion, we should not “titrate” SaO₂ in different species to match their PaO₂, based on different hemoglobin O₂ affinities. We stand by our statement that IH experiments that lower the SaO₂ of mice to nadirs of 50–70% are severe. Our intent was not to dismiss the importance or validity of these IH models. We merely object to the claim that SaO₂ needs to be lowered more in mice than humans to simulate consequences of OSA.

Author Contributions

JJ and ES collaboratively wrote the manuscript.

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.