Arterial blood gas interpretation


Although arterial blood gas (ABG) data provide critical information to the practitioners of critical care medicine, ABG analysis is among the most frequently ordered test in the intensive care unit (ICU), is overused, and is associated with burdens to our patients (discomfort, blood loss) and healthcare systems. Therefore appropriate understanding and use of this clinical test is important for optimal care. There are no randomized trials performed on data; however, studies looking at the utility of ABG in relation to clinical outcomes have given mixed results and collectively underscore the need to use ABG information obtained in the ICU within a specific and appropriate clinical context. Since the 1950s, the development of polarographic electrodes by Clark for oxygen (O 2 ) and Severinghaus and Bradley and Stow and coworkers for carbon dioxide (CO 2 ) has permitted measurement of the partial pressures of oxygen (PaO 2 ) and carbon dioxide (PaCO 2 ) in arterial blood. Cremer, Haber, and Klemensiewicz developed the pH electrode in the early 20th century. ABG remains the definitive method to diagnose, categorize, and quantitate respiratory and metabolic failure.

Why am I obtaining the arterial blood gas?

Acute metabolic acidosis is common among critically ill and results from complex causes. Unfortunately, the clinical focus often shifts from treating the underlying causes to correcting the pH itself. Thus the primary question to ask ourselves before reflexively ordering the next ABG should be “Why am I ordering this test?”

A pH value less than 7.38 defines acidemia, whereas severe acidosis is a pH of 7.20 or lower. In acute severe metabolic acidemia, where the pH is <7.20, the incidence is 6% among the critically ill and carries an associated ICU mortality rate of almost 60%. Interestingly, studies suggest that the presence of an arterial line is the most powerful predictor for obtaining an arterial blood sample for ABG, regardless of the PaO 2 , PaCO 2 , Acute Physiology and Chronic Health Evaluation (APACHE) II score, or presence of a ventilator. , In addition, there are no published guidelines and only limited trials that provide guidance to clinicians regarding the indications for sampling ABGs. Protocolized care may be able to reduce the number of unnecessary ABGs without negative effects on patient outcomes. ,

The indications for ABG analysis have to be guided by the clinical context. Already deployed technologies, such as pulse oximetry and transcutaneous CO 2 detection, decrease the frequency of using ABGs. , Yet we still need to give clinicians a “rule of thumb” for sampling ABGs. One attempt at constructing such an “ABG indications” list might be:

  • After initiation of mechanical (invasive/noninvasive) ventilation

  • Tracking the acute respiratory distress syndrome (ARDS) clinical course

  • Presence of hypoxemic and/or hypercapnic respiratory failure

  • Presence of acute circulatory failure

  • Management of complex acid-base disorders

Arterial blood gas sampling

The ABG samples are either obtained from an arterial catheter or direct arterial puncture into a heparinized syringe. It used to be customary to flush a syringe with heparin and then use that syringe to sample ABGs; however, work in both adult and pediatric patients showed that excess heparin decreased the PaCO 2 , PaO 2 , HCO 3 , and base excess, while the pH remained unchanged. Thus excess liquid heparin tends to promote the interpretation of metabolic acidosis with respiratory compensation. ,

While the most common site for arterial puncture is the radial artery, femoral and brachial arteries are also commonly used to sample arterial blood. Risks associated with arterial punctures are hematoma formation, ischemia to the hand or lower extremity, arterial injury, pseudoaneurysms, and arteriovenous fistulae. , Once obtained, the arterial blood sample must be processed immediately and in accordance with the best laboratory practices. In addition to differences between laboratories, calibration discrepancies and contamination of electrodes with protein or other fluids may alter results. ,

Using the polarographic electrodes, PaO 2 , PaCO 2 , and pH are directly measured; oxygen saturation is calculated from standard O 2 dissociation curves and may be directly measured with a co-oximeter. , A co-oximeter is a blood gas analyzer that measures not just the partial pressure of gases but also the concentration of oxygen associated with different types of hemoglobin based on their absorption spectra (Beer-Lambert law). The use of co-oximetry is usually indicated when:

  • A toxin such as cyanide is suspected

  • Hypoxia fails to improve with the administration of oxygen

  • There is a discrepancy between the PaO 2 on a blood gas determination and the oxygen saturation on pulse oximetry (SpO 2 )

  • The clinician suspects dyshemoglobinemias such as methemoglobinemia (Met-Hb) or carboxyhemoglobinemia (CO-Hb)

Pulse oximetry, unfortunately, does not differentiate among the different types of hemoglobin. For example, in the case of Met-Hb, the SpO 2 may read 86%, but desaturation can be demonstrated with co-oximetry, recording 68% oxyhemoglobin and 32% Met-Hb. ,

The bicarbonate (HCO 3 ) concentration is then calculated using the Henderson-Hasselbalch equation:

pH = pK A + log([HCO 3 ]/[CO 2 ])

where pK A is the negative log of the dissociation constant of carbonic acid (HCO 3 ).

The base excess is the quantity of strong acid required to titrate blood to pH 7.40 with a PaCO 2 of 40 mm Hg at 37°C. In reality, acid is not titrated but calculated using a variety of normograms. Such calculations focus only on the metabolic sources for pH and [H + ] changes. Similarly, bicarbonate is a calculated value that assumes PaCO 2 to be 40 mm Hg.

The following are some of the details that contribute to erroneous readings and interpretations.

Steady state

The ABG has to be collected when a patient reaches a steady state during the clinical course, allowing for the arterial and alveolar gases to equilibrate. Equilibration may take up to 20 or 30 minutes in the case of patients with chronic obstructive pulmonary diseases (COPDs).

Anticoagulants

As mentioned in the Arterial Blood Gas Sampling section, excess heparin may affect the PaCO 2 , PaO 2 , HCO 3 , and base excess, while not interfering with the pH. Only 0.05 mL is required to anticoagulate 1 mL of blood. Knowing that the “dead space” volume of a standard 5-mL syringe is approximately 0.02 mL, it is safe to assume that just having heparin in that dead space would suffice to provide anticoagulation up to 2 mL of blood sample. New prefilled syringes (sodium or lithium heparin) overcome these problems.

Processing delay

Although the blood sample resides in the syringe until analyzed, it does consume O 2 and produce CO 2 . Red blood cell glycolysis could generate more lactic acid through the aerobic glycolysis and lower the pH. , If the sample remains unanalyzed at room temperature for more than 15 minutes, there would be significant increases in PaCO 2 and decreases in pH. When the sample is stored on ice, however, it can be processed up to 2 hours after collection without affecting the blood gas values, even on Mt. Everest.

Venous sampling

If no arterial blood sample is obtained, venous blood gas analyses would be of limited help; however, venous samples do yield estimates for PaCO 2 and lactatemia. , This is discussed later in the chapter.

At times the intended arterial puncture results in inadvertent venous sampling. One can recognize venous sampling when:

  • The practitioner fails to observe a flash of blood during syringe filling

  • The blood gas analyses clearly are not compatible with the clinical condition

  • There is unexpectedly low PaO 2 and high PaCO 2

  • SpO 2 by simultaneous pulse oximetry exceeds the SaO 2 in the measured ABG

Collection equipment and technique

If the dead space were high in the syringe, it would lower the PaCO 2 . Additionally, a needle smaller than a 25-gauge may cause hemolysis.

If an arterial line is in place, one must minimize the dead space of the system (priming the volume from sampling port to catheter tip) to prevent dilution; the safe upper limit would be two times the dead space.

Hyperventilation

Hyperventilation resulting from anxiety and/or pain may acutely alter results from baseline values.

Leukocytosis

Leukocytosis decreases the PaO 2 and pH and elevates PaCO 2 in stored samples. Such PaO 2 decreases are most noticeable at higher PaO 2 levels, are attributable to cellular oxygen consumption, and may be attenuated when samples are stored at colder temperatures.

Hypothermia

Blood gas values are temperature dependent, and as the temperature decreases, solubility of CO 2 increases and partial pressures decline. Thus if blood samples are warmed to 37°C before analysis (as is common in most laboratories), PaO 2 and PaCO 2 will be overestimated and pH underestimated in hypothermic patients. The following correction formulas can be used:

  • Subtract 5 mm Hg PO 2 per 1°C that the patient’s temperature is less than 37°C.

  • Subtract 2 mm Hg PaCO 2 per 1°C that the patient’s temperature is less than 37°C.

  • Add 0.012 pH units per 1°C that the patient’s temperature is less than 37°C.

Although there is an extensive literature on the so-called “pH-stat” and “alpha-stat” assessment of ABGs, in summary the pH-stat acid-base approach aims at maintaining the patient’s pH in a constant range via managing pH at the patient’s temperature. As in the formulas provided earlier, pH-stat is temperature corrected. On the other hand, alpha-stat aims at the ionization state of histidine, and it is maintained by managing a standardized pH (measured at 37°C). Alpha (normally about 0.55) is the ratio of protonated imidazole to total imidazole on the histidine moieties of proteins. Alpha-stat is not temperature corrected; as the patient’s temperature falls, the partial pressure of CO 2 decreases and solubility increases. Thus a hypothermic patient with a pH of 7.40 and a PCO 2 of 40 mm Hg (measured at 37°C) will actually have a lower PaCO 2 because of its lower partial pressure, and this will manifest as a relative respiratory alkalosis. pH-stat, with its goal of maintaining PaCO 2 of 40 mm Hg and pH of 7.40 at the patient’s actual temperature, results in higher PaCO 2 (respiratory acidosis) .

Hypoxemia, hypoxia, and arterial blood gas analysis

The PaO 2 is primarily used to assess oxygenation, and it is reliable within a dynamic range between 30 and 200 mm Hg; however, the reliability of the reported oxygen saturation of blood (SaO 2 ) ranges within a much narrower range: between 30 and 60 mm Hg. The noninvasive method of measuring oxygen saturation by pulse oximetry (SpO 2 ) or by ABG analysis (SaO 2 ) is a better indicator of arterial oxygen content than PaO 2 , because only ~2% of the oxygen is carried in dissolved form, and the greatest amount (98%) is carried by Hb. When using SpO 2 , one has to be cognizant about its shortcomings, including interference by indicator dyes used during procedures. ,

Hypoxemia is defined as a PaO 2 of less than 80 mm Hg in adults breathing room air; hypoxia denotes tissue- or cell-level decreases in oxygen availability. Thus tissue or end-organ hypoxia in patients with hypoxemia depends on the severity of the hypoxemia and the ability of the cardiovascular system to compensate. Hypoxia is unlikely to occur in response to mild hypoxemia itself (PaO 2 = 60–79 mm Hg) when cardiovascular reflexes and integrity remain intact. Moderate hypoxemia (PaO 2 = 45–59 mm Hg) may be associated with hypoxia in patients with anemia or cardiovascular dysfunction. Hypoxia is almost always present with severe hypoxemia at levels of PaO 2 <45 mm Hg. Although the PaO 2 might be low at 45 mm Hg, the mitochondrial oxygen partial pressure necessary to complete oxidative phosphorylation is around 0.5–3 mm Hg (i.e., several orders of magnitude lower), suggesting that this buffer may be a reason why some accommodated patients with cyanotic diseases and elite Everest climbers without supplemental oxygen might have an average PaO 2 of 26 mm Hg and yet survive without significant end-organ injury.

Acute respiratory insufficiency occurs when the lungs no longer meet the metabolic demands of the body, which by tradition is divided into two types:

  • Type I, hypoxemic respiratory insufficiency: PaO 2 ≤60 mm Hg when breathing room air at 1 atm.

  • Type II, hypercapnic respiratory insufficiency: PaCO 2 ≥50 mm Hg.

The tripartite information that can be gathered from ABG analyses are (1) oxygen saturation and content, (2) CO 2 as a marker of ventilation, and (3) acid-base status. Here we discuss all three components.

Alveolar ventilation

The arterial CO 2 partial pressure (PaCO 2 ) reflects the CO 2 content of the sample. The CO 2 content basically is the balance between the quantity of CO 2 produced and its excretion through alveolar ventilation (VA). This relationship can be expressed by the expression:

PaCO 2 ~ CO 2 /VA

The alveolar ventilation is that portion of total ventilation that participates in gas exchange with pulmonary blood. If the metabolic rate remains unchanged and CO 2 production is assumed to be in steady state, then:

1/VA ~ PaCO 2

Therefore when steady state is reached, PaCO 2 becomes a useful tool assessing alveolar ventilation. If PaCO 2 is >45 mm Hg, it is alveolar hypoventilation, and if PaCO 2 is <35 mm Hg, it is called alveolar hyperventilation (primary or compensatory for metabolic disturbance).

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here