Prediction of Acute Mountain Sickness by Pulse Oximetry: What Are the Right Questions?

To the Editor

Acute mountain sickness (AMS) is unpleasant. However, AMS is generally self-limited, easily treated, and largely preventable by gradual ascent. Portable pulse oximeters are simple to use and widely available, so it is tempting to try using them to predict who will get AMS. This might allay the anxieties of first-time travelers to high altitude and perhaps give those who are at increased risk more time to acclimatize and avoid getting AMS. As a recent editorial in Wilderness & Environmental Medicine ¹ points out, prediction of susceptibility to AMS using pulse oximetry has been an elusive goal.

The author concedes that 2 early studies using resting oxygen saturation by pulse oximetry (Spo₂) showed high sensitivity but very low specificity to predict who will get AMS. There was no mention of the methodological flaws that call into question the reported sensitivity of 100% in each of the studies. The editorial referenced 2 new studies, also using resting Spo₂, published in the same issue as the editorial, that were clearly negative.^2,3

Karinen et al ⁴ added measurements of Spo₂ immediately after exercise to resting measurements and found statistically significant differences in the AMS and non-AMS groups. The actual results (their Table 3) show high sensitivity and low specificity similar to previous studies and their own use of resting Spo₂. Attempts to use changes in Spo₂ in individual subjects have also failed to yield a useful method of predicting who will get AMS.^2,4

The editorial partly blames inaccuracies of pulse oximeters. Pulse oximeters used in the emergency department are generally accurate, but many factors such as carboxyhemoglobin and anemia can cause erroneous readings. ⁵ Some of the “[c]ommon sources of error” cited in the editorial, such as nail polish and acrylic decorations, are uncommon on treks and expeditions. Pulse oximeters may also be inherently inaccurate for technical reasons. ⁶ It is well known that pulse oximeters become highly nonlinear and underestimate arterial oxygen saturations (Sao₂) below 80%. ⁷

The real problem with using pulse oximetry to predict AMS is the assumption that small differences in Spo₂ correlate with the later development of AMS. The physiological basis for this is questionable. The Figure in the editorial shows large second-to-second fluctuations in Spo₂ at high altitude that are larger than the supposed errors in pulse oximetry. I agree with Luks and Swenson, ⁷ who concluded in their recent review of pulse oximetry at high altitude that “… clinical decisions should not be based on small differences in saturation over time or among individuals.”

Two issues of the editorial should be noted and corrected. The author refers to the result of pulse oximetry as Sao₂, which is arterial oxygen saturation measured using arterial blood gas analysis. The correct term is Spo₂, which is arterial oxygen saturation measured by pulse oximetry. The source of the Figure in the article, showing Spo₂ measurements in a resting subject at sea level and at high altitude, should be referenced and the Figure should be republished with the scales indicating the actual values on the x and y axes.

One of the joys of practicing medicine in the wilderness is the ability to rely on clinical judgment without excessive testing. Pulse oximetry is an extremely useful clinical tool in managing patients with high altitude illness, but is unlikely to have value in predicting who will get AMS.