The greatest stimulus for arterial as opposed to venous specimens for blood gas studies is to obtain measurement of blood oxygen. The usual information reported is PO2 (concentration of O2 gas measured in mm of Hg or Torr), obtained with a direct-reading PO2 electrode. PO2 represents the dissolved oxygen content of plasma, analogous to the relationship of PCO2 to CO2. Reference ranges for PO2 are age related. In persons under age 60 (breathing room air), 80-95 mm Hg is considered normal. Over age 60, 1 mm Hg per year, but no more than 20 mm Hg, is subtracted from the lower limit of the reference range. A PO2 of 60-80 mm Hg is classified as mild hypoxemia, a PO2 of 40-60 mm Hg represents moderate hypoxemia, and a PO2 less than 40 mm Hg represents severe hypoxemia.

Blood oxygen studies are done for three reasons. First, they help indicate the current status of alveolar gas exchange with inspired air. PO2 provides this information. A normal PO2 while breathing room air indicates adequate pulmonary ventilation.

The second reason is to determine the amount of oxygen available to body cells. PO2 is less adequate for this purpose, because most of the oxygen in the blood is not dissolved oxygen but oxygen bound to RBC hemoglobin. The measurement one really needs is oxygen saturation, which is the actual amount of oxygen bound to hemoglobin compared with the theoretical amount that should be bound to the same amount of hemoglobin (or, the amount of hemoglobin bound to oxygen compared to the amount of hemoglobin available for binding). Then, the quantity of hemoglobin times percent saturation times the factor 1.34 gives the total quantity of oxygen in the blood (except for the very small amount of dissolved oxygen). When the hemoglobin level is normal and PO2 is normal, the percent saturation and total oxygen content are usually adequate. In fact, many blood gas machines provide a calculated oxygen saturation value derived from PO2, normal hemoglobin levels, and data from the normal oxgyen-hemoglobin dissociation curve. Although there is adequate correlation between calculated oxygen saturation and actual (true) oxygen saturation (measured in a special instrument such as a CO-Oximeter) at normal hemoglobin and PO2 values, calculated O2 saturation results can become significantly incorrect at subnormal PO2 values, due to the sigmoid (S) shape of the oxygen-hemoglobin dissociation curve. In the steep midportion of the S curve, a relatively small decrease in PO2 leads to a relatively large decrease in oxygen saturation. In addition, there are a considerable number of conditions that shift the curve to greater or lesser degree and affect oxygen saturation. Nevertheless, a decreased PO2 suggests the possibility of tissue hypoxia, and the degree of PO2 decrease provides a rough estimate of the probability and severity of tissue hypoxia. Certain conditions decrease blood oxygen content or tissue oxygen independently of PO2. These include anemia (producing decreased hemoglobin and therefore decreased oxygen-carrying capacity), carbon monoxide poisoning (CO replaces O2 on the hemoglobin molecule), acidosis (which increases oxygen dissociation from hemoglobin), and congestive heart failure (which slows blood flow and decreases tissue perfusion rate). Hemoglobins that do not carry oxygen (e.g., Hb F), if present in sufficient quantity, result in decreased O2 saturation values.

The third reason for PO2 measurement is to monitor effects of oxygen therapy. The usual goal is to raise PO2 above the lower limit of the reference range. However, in some patients, oxygen therapy may be adequate but unable to provide a normal PO2.

Oxygen saturation (SaO2) is another frequently used parameter of tissue oxygenation. This is a measurement of arterial blood oxygen content. As discussed previously, SaO2 can be measured directly by an instrument called a CO-oximeter or estimated by calculation from PO2 and hemoglobin quantity; it can also be measured indirectly by means of a pulse oximeter (discussed later).