Arterial hypoxemia almost always points to either a reduction of the inspired oxygen tension or to a lung problem. When troubleshooting a hypoxemic patient, it sometimes helps to investigate the problem by breaking it down in an organized and stepwise fashion:
- Hardware and access issues: these are more or less obvious plumbing problems which can be fixed with straightforward mechanical means such as endotracheal intubation, repositioning of the endotracheal tube, cricothyroidotomy, or cranking up the inspired oxygen pressure (FiO2). Hardware issues are, in a sense, nonbiological: you need to get oxygen to flow from point “A,” which is either the atmosphere or an oxygen tank, to Point “B,” which is the alveoli. Broadly speaking, this is airway management and it comes before everything else.
- Shunting: shunting occurs when deoxygenated blood from the pulmonary circulation enters the systemic circulation without first abutting against functioning alveolocapillary membranes. A chest radiograph is usually the first test done here because it is fast and cheap and reveals many of the more common causes of shunting such as pulmonary edema, pneumonia, and atelectasis. You fix shunting by increasing the PEEP and by getting rid of whatever substance or condition is causing the shunt: diuretics, antibiotics, pulmonary hygiene – whatever.
- Pulmonary embolism. If the hardware is in place and functioning well and the chest radiograph is clear, then PE is the next thing to think about in a hypoxemic patient. The test of choice here is, of course, a CT angiogram. This test has the added benefit of being able to pick up other more subtle forms of shunting which cannot be seen well on a chest radiograph, specifically interstitial lung disease, arteriovenous malformations, and milder forms of pulmonary edema. (When ordering a PE study, don’t write “rule out PE.” Rather, write “hypoxemia.” That way, you will get a lot more useful information from your friendly radiologist.)
- Arterial hypercarbia: the sum of all partial pressures of gasses in a closed system is constant, so a high PCO2 will “crowd out” oxygen and cause an automatic drop in PaO2 (Dalton’s law). If you are inclined to prove this diagnosis, the best way to do so is with an arterial blood gas. A normal venous PCO2 rules out arterial hypercarbia.
The above approach will enable you to figure out, in almost every instance, why your patient is hypoxemic.
While the PaO2 is the most accurate test and the reference standard for hypoxemia, it is a relatively invasive, painful and time-consuming test because it requires skilled puncture of an artery. Instead, the percent oxygen saturation of hemoglobin (SaO2) can serve as good surrogate test because it can be determined noninvasively with a pulse oximeter.
There are, however, situations where a patient can have a falsely low SaO2, despite a normal PaO2. These include:
- Shocky patients with poor peripheral circulation
- The presence of barriers (e.g., nail polish)
- Tricuspid regurgitation, and
- Mechanical failure of the pulse oximeter device itself.
Conversely, a falsely high SaO2 is also possible in patient’s with carbon monoxide poisoning and methemoglobinemia (see below).
The critical concept here is that when the PaO2 is well within normal limits, small drops in PaO2 will cause only tiny drops in the blood’s total oxygen content. This is due to hemoglobin’s incredible oxygen-carrying capacity and the sigmoidal shape of the oxygen-hemoglobin dissociation curve. A transfusion of packed red blood cells will add oxygen carrying capacity to an anemic patient and will improve total oxygen content, but it will not improve a poor PO2. Again, a low PO2 points squarely to a reduction of the inspired oxygen tension (FiO2) or to a lung problem. The PO2 needs to be fixed separately but concomitantly.
Poor systemic or local circulation. The PaO2 and hemoglobin may be normal, but oxygen is still not getting where it needs to go because of poor cardiac output or vascular supply.
The most important causes of toxic hypoxia are carbon monoxide, methemoglobin, and cyanide. Carbon monoxide binds to hemoglobin, while methemoglobin is a hemoglobin derivative. They both prevent normal oxygen delivery to cells and cause a falsely-normal pulse oximeter reading. Therefore, if carbon monoxide poisoning or methemoglobinemia are suspected, ditch your regular pulse oximeter and get an arterial blood gas with cooximetry or perform a pulse cooximetry. A cooximeter will measure the relative concentrations of various forms of hemoglobin, including oxyhemoglobin, carboxyhemoglobin, and methemoglobin.
Cyanide poisoning is called histotoxic hypoxia because cyanide does not bind hemoglobin to a significant degree. Rather, it poisons tissues (mitochondria) directly and thereby prevents them from utilizing oxygen. Thus, in patients with cyanide poisoning, oxygen will float right past the capillaries and enters the venous circulation, instead of getting dropped off at the capillary level for cellular respiration. Cells are then forced to undergo anaerobic respiration and produce lactic acid in the process.
As with carbon monoxide poisoning and methemoglobinemia, the pulse oximeter will often be normal in cyanide poisoning. However, unlike patients with carbon monoxide poisoning and methemoglobinemia, patients with cyanide poisoning will frequently have a normal cooximetry reading as well. The key to clinching the diagnosis is to look for red (hyperoxic) venous blood, either on venipuncture or on funduscopic examination, especially in the presence of an otherwise unexplained profound lactic acidosis.
By Mark Yoffe MD
- Fishman, Alfred, Fishman’s Pulmonary Diseases and Disorders (2008)
- Terry Des Jardins, Clinical Manifestations & Assessment of Respiratory Disease, 6e (2010)
[Updated October 9, 2013]