Dogma of the Arterial Blood Gas
Modern blood gas analysis came into widespread use during the polio epidemic of the early 1950s to assess the respiratory status of patients and the efficiency of the new breathing support equipment, i.e., the iron lung. Smaller and faster testing equipment was soon devised, and arterial blood gas (ABG) testing came into common use in the new Intensive Care Units being set up all over the world in the 1960’s. Instead of only being used to assess respiratory status, the ABG was called upon to assess a patient’s overall physiologic status. The convenience and speed of blood gas testing made rapid clinical decisions possible. This has given rise to the universal opinion that ABGs are the most valuable laboratory test in critically ill patients. [1] Values from ABGs are of such import that they are being used in complex scoring systems to assess a patient’s risk of dying. [2]
Many diseases and critical conditions will often lead to abnormal ABG results. Respiratory status is assessed by the CO2 and O2 components. The tests are also used to determine the presence of metabolic acidosis, respiratory acidosis, diabetic ketoacidosis, as well as respiratory alkalosis, the rare metabolic alkalosis and the confusing and contradictory mixed acid/base disorders. The ABG is considered a gold standard for assessing a patient’s overall physiologic status [1]. But is it really? For example, an ABG showing respiratory alkalosis may only mask a serious acidotic condition as revealed by the VBG. (See the examples below.*)
I learned to interpret ABGs almost 50 years ago from physicians and print materials. Things have not changed very much since then. The equipment is faster and easier to use, but the interpretation of the test hasn’t changed. At the same time I was learning about ABGs, the first perfusionist I ever met and the one who initially trained me OTJ, Jerry Swett, was teaching me about venous blood gases (VBGs). His premise was that they were more important than ABGs during CPB and also in critical care patients in general [3].
Accurate VBGs are often considered difficult to obtain. It is thought that unless the sample comes directly from the main pulmonary artery (usually via a Swan-Ganz catheter), it is not mixed well enough [4]. Arterial blood leaves the heart from a single artery. But the venous blood enters the heart from several different portals. So it is not uniform in its makeup until it is well mixed by the lungs and ventricles and becomes arterialized. But Jerry realized that venous blood draining into the heart-lung machine was well mixed and therefore represented the blood most closely related to the physiologic conditions of the tissues because it comes directly from the capillaries without being modified by the perfusionist. Jerry could tell if the oxygenator was working simply by watching the blue-red venous blood change to bright red. And if the oxygenator were not removing CO2 adequately, an elevated pvCO2 would soon disclose that fact.
From my experience, the “venous blood mixing” issue is not really a problem. The patient does not need to be on pump or have a pulmonary artery catheter. Samples taken from a right atrial line or even a superior cava line will give enough information to determine if the patient is in trouble long before an ABG will reflect the problem.
As the years passed, I continued to use primarily VBGs during CPB. Then ECMO came along, and I learned many new things about both ABGs and VBGs in non-CPB critically ill patients. I learned that all too often a “normal” ABG did not reliably reflect the patient’s true status. In the critically ill patient, an ABG by itself is almost useless unless accompanied by a VBG drawn at the same time. In fact, a lone ABG may be worse than useless because it may give false reassurance of a “not-so-serious” condition as opposed to the patient’s truly critical status. By the time an ABG becomes seriously abnormal, the patient may be too far gone to revive. At best, the ABG is only reliable in determining the lungs’ ability to transfer oxygen. But even that evaluation is invalid in most univentricular patients.
What about CO2? Doesn’t the ABG reliably detect if the lungs are providing adequate ventilation? No, because sometimes an elevated arterial CO2 is the result of a massive amount of CO2 in the venous blood passing through lungs that are well ventilated. (See the examples below). In this situation the lungs, even when working normally, are incapable of blowing the CO2 down to a normal, arterial level. When this occurs, the elevated arterial CO2 is called “breakthrough hypercapnia.” Rigorous clinical attempts to ventilate off the excess arterial CO2 can result in an arterial respiratory alkalosis that only masks a serious venous respiratory acidosis.
What about pH? Doesn’t the ABG accurately reflect the acid/base status of the patient? Again, no; as the examples below will illustrate.
Modern texts will say things like this about VBGs: “The values of a VBG and ABG are comparable (arterial and venous values are NOT (sic) significantly different for practical purposes) except in the cases of O2 and CO2.”[5] “VBG analysis compares well with ABG analysis for pH estimations in adults.”[6]
But these statements are only true in healthy, normal people; not in the critically ill. For example, compare these two gases drawn at the same time from a critical patient in the ICU and not on CPB (pH / pCO2 / pO2 / base):
ABG = 7.35 / 43 / 155 / -1.9
VBG = 7.19 / 74 / 20 / -1.7
The ABG provides only false reassurance that the patient’s respiratory and physiologic status are within normal limits. The pulse oximeter and end-tidal CO2 monitor both have normal values. But the VBG says that this patient has a serious respiratory acidosis that is not caused by under ventilation. In addition, there is venous hypoxemia and probably tissue hypoxia. This patient soon died of a refractory pulmonary hemorrhage.
Below are three sets of gases taken from a cardiac arrest patient during CPR. This first set is taken after 5 minutes of CPR. The ABG would imply that the Code Team is doing a good job of CPR by maintaining a slight arterial alkalosis. But the VBG shows that the team is rapidly losing ground:
ABG = 7.52 / 28 / 436 / -1.0
VBG = 7.31 / 58 / 25 / -2.0
The ABG shows alkalosis with hypocapnia while the VBG shows acidosis with hypercapnia and venous hypoxemia (SVO2 < 70%). Tissue hypoxia is most likely present regardless of the high arterial oxygen saturation. The decision to provide extracorporeal life support (ECLS) should have been immediately forthcoming with this finding.
Continuing with the same patient after 15 minutes of CPR, the blood gases appeared thus:
ABG = 7.27 / 48 / 430 / -6
VBG = 7.09 / 84 / 25 / -6
This ABG is an example of “breakthrough hypercapnea.” The arterial pCO2 is becoming unmanageable even with manual bag ventilation due to extreme venous hypercapnia.
After 35 minutes of CPR, the ABG pH finally passes the critical limit; 7.21 +/- 0.06 [7]. However, the VBG passed the critical limit more than 20 minutes earlier and now reflects the patient’s true disastrous pH which makes the hope of a successful ECLS intervention unlikely. ECLS intervention is at least 20 minutes too late.
7.11 / 38 / 322 / -19
6.82 / 96 / 20 / -20
Though it reduced the effectiveness of CPR even further by increasing intrathoracic pressure, aggressive mechanical ventilation finally succeeded in normalizing the arterial CO2. Nevertheless, this effort is totally ineffective at correcting the venous hypercapnia and hypoxemia. ECLS was not instituted, and the patient was soon declared dead.
In another example, compare these two gases drawn from another critical patient whose heart is still beating, and he is not on CPB:
ABG = 7.31 / 48 / 375 / -2
VBG = < 6.90 / 106 / 27 / ?
The ABG by itself suggests only a mild respiratory acidosis. Without the VBG to illustrate the true status of the patient, this ABG by itself is false reassurance that whatever treatment is being used is working to maintain the patient safely. However, the VBG reveals the true reality; that this patient is on the verge of dying and did die of a large hemorrhagic stroke a short time later.
Large venoarterial CO2 gradients, like those in the examples presented, are associated with increased morbidity and mortality. For the patient on ECLS, whether it is total support like CPB or only fractional support like IABP, the perfusionist should use paired VBGs and ABGs to ensure that blood flow is adequate to maintain a venoarterial CO2 gradient of 10 mmHg or less. This will help to maintain a normal intracellular pH and an SVO2 of 70% or better. [8,9,10]
The reason I have presented these extreme examples is to show just how bad the situation can become before the perfusionist is even aware of it if only reviewing ABGs. Of course, the reason to draw routine VBGs is to enable the perfusionist to intervene long before conditions become desperate either before or during ECLS; CPB included. There is a learning curve with VBG interpretation. Unfortunately, I know of no textbooks devoted to training perfusionists about VBG interpretation in ECLS patients or even non-ECLS patients. Sometimes experience is not just the best teacher but the only teacher. If you have questions about VBG interpretation, you are welcome to contact me at any time.
Many years ago I was in a large group of perfusionists being given a refresher class in blood gas interpretation by a prominent anesthesiologist. After he finished, he asked if anyone had any questions. I asked if he ever drew VBGs. He said no because the ABG provided all the pertinent information needed. He then asked how many perfusionists in the audience used VBGs. I was the only one to raise my hand. (Obviously, in that large group of perfusionists, they had all been dogmatically taught to use only ABGs when evaluating their patients.) I then asked the anesthesiologist what he would do if he got a set of blood gases like those I listed just above. He said that he had never seen anything like those results and he did not think that those results were even physiologically possible if taken as a coincident pair. So I asked him if he never drew VBGs how would he know if this was possible or not. His answer was telling; he never drew VBGs because the ABGs provided all the information he ever needed. So there was no point in drawing a VBG. This illustrates the dogma of the ABG, not just among physicians but among perfusionists as well. As to the reasons why these VBGs are so abnormal and how a perfusionist might correct them, that is a discussion for another time.
*All the blood gas examples given here are genuine and authentic in the situations described.