My watch list  

Arterial blood gas sampling

This article lacks in-line citations.

Arterial blood gas sampling is a medical technique used to check gas levels in the blood. It typically involves using a thin needle and syringe to puncture an artery, usually in the wrist, and withdraw a small amount of blood. This technique is useful for making sure that certain parts of the blood's chemistry are normal. This technique is commonly used on patients whose breathing is controlled by a mechanical respirator or who are having serious difficulties with breathing. When this procedure is performed on a small artery in the wrist, it is unlikely to cause serious complications, and the information gained may save the patient's life.

The purpose of arterial blood gas sampling is to assess patients' respiratory status as well as acid-base balance or for laboratory testing when venous blood is unavailable, and is frequently requested for seriously ill patients. An arterial blood gas (ABG) will help in the assessment of oxygenation, ventilation, and acid-base homeostasis. It can also aid in the determination of poisonings (carboxyhemaglobinemia or methemaoglobinemia) and in the measurement of lactate concentration.

It is sometimes called arterial blood gas analysis or "ABG" sampling.


Outcome Goal

Proper collection of arterial blood samples.

Arterial Blood Gas Test Definition

Blood is drawn anaerobically from a peripheral artery (Radial, Brachial, Ulnar, Femoral, Axillary, Posterior Tibial or Dorsalis pedis) via a single percutaneous needle puncture, or from an indwelling arterial cannula or catheter for multiple sample.

Either method provides a blood specimen for direct measurement of:

  • dyshemoglobins carboxyhemoglobin (COHb);

Purpose and indications

The purpose of arterial blood gas sampling is to assess patients respiratory status as well as acid base balance or for laboratory testing when venous blood is unavailable, and is frequently requested for seriously ill patients. So, an arterial blood gas (ABG) will help in the assessment of oxygenation, ventilation, and acid-base homeostasis. It can also aid in the determination of poisonings (carboxyhemaglobinemia or methemaoglobinemia) and in the measurement of lactate concentration.

Pulse oximetry will give a reasonable estimate of the adequacy of oxygenation in many circumstances but does not assess acid-base status or ventilation and should not be used alone in cases where these measurements are important.

Basic Conditions Diagnosed by ABG's:

  • Anything which prevents the body from getting rid of excess CO2, increases acid which decreases pH
  • Anything which makes to body lose CO2, decreases acid, which increases pH
  • Anything which increases HCO3 increases base which increases pH
  • Metabolic Acidosis
  • Anything which decreases HCO3 decreases base which decreases pH

Apart from helping to establish a diagnosis, blood gases may also help to as certain the severity of a particular condition (e.g. metabolic acidosis in sepsis). This information can help to establish diagnosis, monitor severity, progression, and prognosis as well as guide therapy of:

  • respiratory failure,
  • cardiac failure,
  • renal failure,
  • hepatic failure,
  • diabetic ketoacidosis,
  • poisoning
  • sepsis

Table 1. Normal Values in an ABG report at sea level:

PaO2 in kPa (mm Hg) PaCO2 in kPa (mm Hg) pH HCO3 mmols/L BD/BE Sa O2 (%)
Arterial Blood 11-13 (80-100) 4.7-5.9 (35-45) 7.36-7.44 21-28+/-2 94-100
Venous Blood 5-5.6 (37-42) 5.6-6.7 (42-50)7.34- 7.42

Table 2. Normal Values in an ABG report at sea level:

Normal Range
Anion Gap 8-16 mmols/L
Osmolar Gap 10 mmols/L
PaO2 (mmHg)/FiO2 (%) More than 3
Note: mm Hg = millimeters of mercury. At altitudes of 3,000 feet and above, the values for oxygen are lower. The arterial pO2 reduces with age. A rough guideline is that above the age of 40 years, paO2 = 105-age in years/2.

The pH of the blood is maintained within a normal range by a number of compensatory mechanisms, the most important being the body buffer mechanisms and the renal and respiratory systems. The degree of compensation varies between individuals and depends on the severity and duration of the primary problem and associated medical comorbidities. Respiratory compensation for metabolic problems is usually rapid and almost complete. The lungs respond quickly by increasing ventilation to blow off excessive carbon-dioxide (in metabolic acidosis) or decreasing ventilation to retain carbon-dioxide (in metabolic alkalosis). The latter compensation is less complete than the former for obvious reasons. The renal compensation for respiratory imbalances is slow and incomplete. The kidneys regulate extracellular fluid H+ ion concentration by secretion of H+ ions, reabsorption of filtered HCO3- ions, and the production of new HCO3- ions. Excess HCO3- is filtered into the renal tubules and eliminated in the urine. Depending on the need to excrete either an acid or a base load, the kidneys can excrete urine with a pH ranging from 4.5 to 8.0. A rough guide to the degree of compensation to primary changes in CO2 and HCO3 as a result of respiratory and metabolic imbalances respectively is shown in table 3.

Table 3.

Acid-base imbalance Carbon-dioxide (mm Hg) Plasma bicarbonate (mmol/L)
Metabolic acidosis Compensates (decreases) by 1.25 x X Decreases by X
Metabolic alkalosis Compensates (increases) by 0.75 x XIncreases by X
Acute respiratory acidosis Increases by X Compensates (increases) by 0.1 x X
Chronic respiratory acidosis Increases by X Compensates (increases) by 0.4 x X
Acute respiratory alkalosis Decreases by X Compensates (decreases) by 0.2 x X
Chronic respiratory alkalosis Decreases by X Compensates (decreases) by 0.4 x X

Abnormalities of gas exchange


The first step is to detect the presence of hypoxia (i.e., less than 60 mm Hg on room air). Patients with clinically evident respiratory problems who have been given oxygen supplementation before blood gas analysis must be assumed to be hypoxic at this stage. If a patient is being given oxygen supplementation, then the ratio of the paO2 (in mm Hg) to FiO2 (in %) is used to detect hypoxia. Usually, the oxygen saturation of the blood is also noted which correlates with the paO2 of the arterial blood and helps in establishing the diagnosis of hypoxia. The saturation as obtained by a blood gas analysis is more accurate than that obtained by a pulse oximetry, as it is not influenced by shock states and skin pigmentation.


The paCO2 level must then be noted, which will help in differentiating between type I and type II respiratory failure. In type I respiratory failure, the paCO2 will be normal or low (45).

Table 4. Causes of respiratory failure

Type I Type II
Atelectasis CNS depression (drugs, sleep, head injury)
Pulmonary edema (cardiogenic and non cardiogenic) High spinal cord lesions
Pneumonia Phrenic nerve lesions
Pleural effusion Neuromuscular disorders
Haemo/pneumothorax Severe kyphoscoliosis

Type I causes in an advanced state

Causes of type I respiratory failure include conditions with an impaired gas exchange and causes of type II respiratory failure include the causes of type I in an advanced state and conditions with impaired ventilation as shown in table 4. Differentiation between types I and II failure is essential to determine the etiology and institute further treatment. It may also rarely be used to restrict O2 supplementation in patients with type II disease because such patients are dependent on hypoxia for the respiratory drive and abolishing hypoxia might further suppress the CNS stimulation for respiration. Patients with persistent hypoxia, rising CO2 levels and respiratory acidosis require mechanical ventilation and are usually seen by the anesthetist at this stage. In patients on a ventilator, hypoxia might indicate one of several things, including:

  • Accidental disconnection of the breathing circuit ( which should be evident by the alarms, fall in O2, fall in saturation on pulse oximetry, clinical evidence of respiratory distress, etc.).
  • Development of pneumothorax which could be detected clinically and might need confirmation by X-ray before intercostal tube drainage.
  • Development or worsening of pre-existing chest problems (bronchopneumonia, ARDS, pulmonary contusion) might require changes in settings of the ventilator like increasing FiO2 levels, ventilatory rate or tidal volume, adding or increasing PEEP, or changing to other modes of ventilation.

A high CO2 level is always associated with hypoxia unless the patient is on oxygen supplementation. However, hypercarbia associated with a normal oxygen level should also be approached with the same urgency as the patient might deteriorate rapidly. The possibility of a venous sample as a cause of unexpected hypercarbia and hypoxia should be kept in mind. In patients on a ventilator, moderate rise in CO2 levels are currently considered acceptable and interventions to correct these might be associated with significant side effects including barotraumas and hypotension (permissive hypercarbia). Similarly, patients with COPD have adapted to higher levels of carbon-dioxide and might not require correction to normal levels. Acute changes in paCO2 result in predictable changes in pH. For every increase in paCO2 of 20 mm Hg (2.6 kPa) above normal, the pH falls approximately by 0.1. For every decrease of paCO2 of 10 mm Hg (1.3 kPa) below normal, the pH rises by 0.1. Any change in pH outside these parameters is therefore metabolic in origin. The kidneys take time to compensate for the change in pH the amount of renal compensation indicates the chronicity of the problem and the need for urgent correction. Correction will usually involve a combination of treatment of the cause, initiation of mechanical ventilation or modification of the settings and reduction of CO2 production.

Alveolar-arterial oxygen gradient (A-a) PO2: This is the difference in the oxygen partial pressures between the alveolar and arterial sides. In patients with type II respiratory failure, it may help to determine whether the patient has associated lung disease or just reduced respiratory effort.

The A-a gradient increases a little with age, but should be less than 2.6kPa (20mmHg). A normal gradient would imply conditions like CNS depression and neuromuscular disorders as the cause and a high gradient would imply some lung disease.

Abnormalities of acid base balance

pH of the blood: The pH is usually maintained within a narrow range by a number of buffer systems in the body. A normal pH value may still be due to a well-compensated imbalance or a mixed acid base disorder and an abnormal value is definitely due to a poorly compensated acid base problem or due to both metabolic and respiratory derangements causing an imbalance in the same direction.

Serum bicarbonate:

The actual bicarbonate is the value calculated from the blood gas sample. The standard/corrected bicarbonate is the value obtained after correction of CO2 levels to 40mm Hg and at room temperature. It gives a better estimate of the metabolic problem causing acid base imbalance. The base deficit/excess is the amount of deviation of the standard bicarbonate from the normal. The metabolic problem could either be a low (base deficit or metabolic acidosis) or high (base excess or metabolic alkalosis) standard bicarbonate.


A primary metabolic derangement will be accompanied by some degree of respiratory compensation. The ability to detect the primary abnormality and the amount of compensation is hindered by other co-existing conditions causing respiratory acidosis and/or alkalosis. Coexisting medical problems can cause both metabolic acidosis and alkalosis.

Metabolic acidosis:

Metabolic acidosis can be due to a variety of conditions. Treatment of metabolic acidosis is treatment of the cause. Direct administration of alkali (sodium bicarbonate) is reserved for severe cases. A number of conditions can result in metabolic acidosis, the most important among them being the under perfusion of tissues resulting in accumulation of lactic acid. Differentiation of the causes of metabolic acidosis requires the estimate of an entity called the 'anion gap'.

Anion gap:

Body fluids including blood may contain a variable number of ions, but the total number of anions (negative ions) and cations (positive ions) are roughly the same. The ions that are usually measured in blood are cations like sodium and potassium and anions including chloride and bicarbonate. There are unmeasured ions in both groups (cations and anions), which also contribute to the ionic constitution of blood. The measured cations are usually greater than the measured anions by about 8-16mmol/L. This is because the unmeasured anions constitute a significant proportion of the total number of anions in blood. Proteins make this up predominantly, but also included are sulphates, phosphates, lactate and ketones. Causes of a decreased anion gap include hypoalbuminaemia and severe haemodilution. Rarer causes include increase in minor cation concentrations like calcium and magnesium. Causes of a raised anion gap include dehydration and any cause of raised unmeasurable anions, like lactate, ketones and renal acids, along with treatment with drugs given as organic acids such as penicillin, salicylates and poisoning with methanol, ethanol and paraldehyde. Rarely it may be due to decreased minor cation concentrations such as calcium or magnesium.

Raised anion gap metabolic acidosis:

Accumulation of a number of acids can result in raised anion gap metabolic acidosis. In such cases, the reduction in serum HCO3- matches the anion gap. If not, a second acid base disorder should be kept in mind. When metabolic acidosis and alkalosis coexist, as in vomiting and ketoacidosis, the plasma HCO3- may be normal, and a raised anion gap may be the initial evidence of an underlying acid-base disturbance. To differentiate between the many causes of 'increased anion gap metabolic acidosis, we measure the osmolar gap that is the difference between the measured osmolarity and the calculated osmolarity.

Normal anion gap (hyperchloraemic) metabolic acidosis:

This usually results from conditions wherein there is a loss of alkali (i.e.HCO3-) or metabolic equivalent (eg, excretion of salts of organic anions in proportionate excess of chloride) or an accumulation of HCl or metabolic equivalent (eg, NH4Cl and chloride salts of amino acids). Loss of HCO3- can occur either due to GI causes or due to renal causes (renal excretion or insufficient generation). In many surgical conditions, the cause is usually obvious. Examples of extrarenal causes are excessive diarrhoea or drainage of gastrointestinal secretions, NH4Cl administration, parenteral nutrition, rapid saline infusion and congestive cardiac failure. Generation of large amounts of organic anions can sometimes produce this type of metabolic acidosis (and not one with a raised anion gap), if the kidneys can prevent their accumulation by rapid excretion.

Examples of renal causes include the various types of renal tubular acidosis (type I, type II and type IV).

Examples of different types of renal tubular acidosis (RTA)

Table 5. Examples of different types of renal tubular acidosis (RTA)

Type I (distal) Type II (proximal) Type IV (def. NH4+ production)
Urinary tract obstruction Fanconi's syndrome Cortisol deficiency
Interstitial nephritis Myeloma light chain nephropathy Urinary tract obstruction
K+ sparing diuretics Nephrotoxins
Genetic diseases

Metabolic Alkalosis:

Metabolic alkalosis can result from the loss of acid, addition of alkali or both in the kidneys or elsewhere. Extrarenal sites include stomach (loss of acid), redistribution of alkali from the intracellular stores to the ECF (as in potassium or chloride depletion), oral administration (antacids, ion-exchange resins, milk alkali syndrome, oral HCO3-) and parenteral administration of alkali (citrate in blood transfusions, bicarbonate in severe metabolic acidosis). Renal causes of alkali excess include mineralocorticoid excess, response to long-standing hypercapnia (persists even after correction of respiratory acidosis), hypokalemia (promotes H+ secretion in the distal nephron) and ECF volume depletion (impaired HCO3- excretion). Certain conditions can cause metabolic alkalosis by a number of mechanisms (e.g. diuretic use causes both ECF depletion and hypokalemia).

Respiratory Alkalosis:

The principal cause of respiratory alkalosis (hypocapnia) is hypoxia and its causes (type I respiratory failure), further treatment of which has been detailed before. Other causes of acute respiratory alkalosis include anxiety, fever, pain, sepsis, hepatic failure, CNS disorders (stroke, infections), pulmonary disorders without hypoxia (infections and interstitial lung disease), delirium tremens and drugs (salicylate intoxication). Chronic causes include high altitude hypoxia, chronic hepatic failure, chronic pulmonary disease, CNS trauma, anaemia, hyperthyroidism, beriberi and pregnancy. Treatment should be directed towards the cause.

Contraindications/Concerns for Arterial Puncture


Contraindications may be absolute unless specified otherwise. Contraindications considered as a relative in terms of the risks to the patient under the importance of obtaining the sample!

  • Cellulites or other infections over the radial artery.
  • Absence of palpable radial artery pulse.
  • Negative results of an Allen test (collateral circulation test), indicating that only one artery supplies the hand and suggest to select another extremity as the site for arterial puncture.
  • Coagulopathies or medium-to-high-dose anticoagulation therapy (eg, heparin or coumadin, streptokinase, and tissue plasminogen activator but not necessarily aspirin) may be a relative contraindication for arterial puncture.
  • History of arterial spasms following previous punctures.
  • Severe peripheral vascular disease.
  • Abnormal or infectious skin processes at or near the puncture sites.
  • Arterial grafts.
  • Arterial puncture should not be pertormed through a lesion or through or distal to a surgical shunt (eg, as in a dialysis patient). If there is evidence of infection or peripheral vascular disease involving the selected limb, an alternate site should be selected.


Sampling may be performed by trained health care personnel in a variety of settings including (but not limited to) hospitals, clinics, physician offices, extended care facilities, and the home. However, because of the need for monitoring the femoral puncture site for an extended period, femoral punctures should not be performed outside the hospital.


Arterial blood sampling should be performed under the direction of a physician specifically trained in laboratory medicine, pulmonary medicine, anesthesia, or critical care. A recognized credential MD, DO, CRTT, RRT, RN, RPFT, CPFT, MT, MLT, RCVT, CPT I, CPT II, or equivalent is strongly recommended.

Note: Femoral, axillary, and brachial arterial punctures are performed by MD ONLY!


The frequency with which sampling is repeated should depend on the clinical status of the patient and the indication for performing the procedure.

Note: Repeated puncture of a single site increases the likelihood of hematoma, scarring, or laceration of the artery. Care should be exercised to use alternate sites for patients requiring multiple punctures. An indwelling catheter may be indicated when multiple sampling is anticipated.

Puncture sites

Approved puncture sites include radial, dorsalis pedis, and brachial arteries. The brachial artery will not be used on patients in Children’s Hospital. In the Emergency Department, femoral artery is an approved puncture site. Brachial and femoral arteries should be reserved as a last option. The radial artery on non dominant hand is the ideal site for an arterial puncture for the following reasons:

  • It is small, but superficial and easily accessible, and stabilized.
  • It is easily compressible with better control of bleeding
  • There is no nerve near by to worry about.
  • The collateral arch with ulnar artery minimizes the risk of occlusion.

The Ulnar Artery may be used in the pediatric population. The ulnar artery in the adult is less accessible but may be used as a secondary site. Allen’s Test is checked prior to a Radial or Ulnar puncture.

Anatomical Review

  The radial artery runs along the lateral aspect of the volar forearm deep to the superficial fascia. The artery runs between the styloid process of the radius and the flexor carpi radialis tendon. The point of maximum pulsation of the radial artery can usually be palpated just proximal to the wrist.

Local Anesthesia


The concern that the pain induced hyperventilation or apnea could alter the results of blood gases. This issue was specifically studied and the results indicate that an unanesthetized arterial puncture does provide an accurate measurement of resting pH and Pco2. Hence, the only reason to use local anesthetic is to avoid pain to the patient. If you are proficient, the first stick can be tried without the anesthetic. I strongly recommend the use of local anesthetic for beginners.

The use of local anesthetic for arterial puncture is not universal. The proposed reason for the use of local anesthetic is: To avoid pain

Heparinized Syringe


Heparin will prevent the blood from clotting and allow the plunger to move with less resistance. Excessive amounts of heparin will alter ABG values

The syringe has to be heparinized to prevent clotting. It is important to have the right amount of heparin in the syringe. “Too much” or “too little heparin can alter the results.”

Further reading

  • Wikibooks: wikibooks:How to take an arterial blood gas sample


1. Armstrong PW, Parker JO: The complications of brachial arteriotomy. J Thorac Cardiovasc Surg 1971; 61:424

2. Browning JA, Kaiser DL, Durbin CG. The effect of guidelines on the appropriate use of arterial blood gas analysis in the intensive care unit. Respir Care 1989; 34:269-276.

3. Bruck E, Eichhorn JH, Ray-Meredith S. Shanahan JK, Slockbower JM. Percutaneous collection of arterial blood for laboratory analysis. National Committee for Clinical Laboratory Standards 1985;H11A;5(3):39-59.

4. Bucher, L. (2001). Arterial Puncture. In D.J. Lynn-McHale & K.K. Carlson (Eds.), AACN procedure manual for critical care (pp.496 – 502). Philadelphia, PA: W.B. Saunders Company.

5. Centers for Disease Control. Update: Universal Precautions for prevention of transmission of human immunodeficiency virus, hepatitis B virus, and other blood-borne pathogens in health care settings. MMWR 1988; 37:377-388.

6. Colley DP Vertebral arteriovenous fistula: an unusual complication of Swan-Ganz catheter insertion. Am J Neuroradiol 1985 Jan-Feb;6(1):103-4

7. Department of Labor, Occupational Safety and Health Administration. Occupational exposure to bloodborne pathogens. 29 CFRR Part 1910.1030. Federal Register, Friday December 06, 1991.

8. Depierraz B; Essinger A; Morin D; Goy JJ; Buchser E Isolated phrenic nerve injury after apparently atraumatic puncture of the internal jugular vein. Intensive Care Med, 15:132-4, 1989

9. DuBose Jr TD. Acidosis and alkalosis. Chapter 50. Section 7. Alteration in urinary function and electrolytes. In Fauci et al. Harrison's Principles of Internal Medicine 14th edition. 1998; 277-86.

10. Fleisher M, et al. Two whole-blood multi-analyte analyzers evaluated. Clin Chem 1989;35: 1532- 1535.

11. Frye M, DiBenedetto R. Lain D, Morgan K. Single arterial puncture vs arterial cannula for arterial gas analysis after exercise. Chest 1986;93:294-298.

12. Gardner RM, Clausen JL, Epler G. Hankinson JL, Permutt S. Plummer AL. Pulmonary function laboratory personnel qualifications. American Thoracic Society Position Paper, ATS News, November 1982.

13. Gluck SL. Acid-base. Electrolyte quintet. The Lancet 1998; 352: 474-9.

14. Hansen JE, Simmons DH. A systematic error in the determination of blood PCO2- Am Rev Respir Dis 1977; 115:1061-1063.

15. Harsten A, Berg B. Inerot S. Muth L. Importance of correct handling of samples for the results of blood gas analysis. Acta Anaesthesiol Scand 1988;32:365-368.

16. Hess D. Detection and monitoring of hypoxemia and oxygen therapy. Resp Care 2000;45:64-83.

17. Hess D, Good C, Didyoung R. Agarwal NN, Rexrode WO. The validity of assessing arterial blood gases 10 minutes after an Flo2 change in mechanically ventilated patients without chronic pulmonary disease. Respir Care 1985;30:1037-1041.

18. Luce EL, Futrell JW, Wilgis EFS, et al: Compression neuropathy following brachial artery puncture in anticoagulated patients. J Trauma 1976; 16:717-721

19. Macon WL, Futrell JW: Median-nerve neuropathy after percutaneous puncture of the brachial artery in patients receiving anticoagulants. N Engl J Med 1973: 288:139

20. McCready RA, Hyde GL, Bivins BA, Hagihara PF: Brachial arterial puncture: A definite risk to the hand. South Med J 1984; 77:6

21. Mitaishvili R, ABG Sampling. RM Global Health

22. Mitaishvili R, Arterial Puncture. [1]

23. Mitaishvili R, Radial artery Puncture. [2]

24. Moran RF, Van Kessel A. Blood gas quality assurance. NSCPT Analyzer 1981;11(I): 18-26.

25. National Committee for Clinical Laboratory Standards. Procedures for the collection of diagnostic blood specimens by skin puncture, 3rd ed. Villanova PA: NCCLS, 1992.

26. Neviaser RJ, Adams JP, May GI: Complications of arterial puncture in anticoagulant patients. J Bone Joint Surg 1976; 58A:218-220

27. Raffin TA. Indications for arterial blood gas analysis. Ann Intern Med 1986;105: 390-398.

28. Ries, AL, Fedullo PF, Clausen JL. Rapid changes in arterial blood gas levels after exercise in pulmonary patients. Chest 1983; 83: 454-456.

29. Shapiro BA, Harrison RA, Cane RD, Templin R. Clinical application of blood gases, 4th ed. St Louis: Year Book Medical Publishers Inc. 1989.

30. Sharzer LA, Baker WH: Nonthrombotic arterial occlusion. Arch Surg 1973; 106:344

31. Thorson SH, Marini JJ, Pierson DJ, Hudson LD. Variability of arterial blood gas values in stable patients in the ICU. Chest 1983;84(1): 14-18.

32. Weibley RE, Riggs CD. Evaluation of an improved sampling method for blood gas analysis from indwelling arterial catheters. Crit Care Med 1989; 17(8):803-805.

33. Williams AJ. ABC of oxygen. Assessing and interpreting arterial blood gases and acid-base balance. B M J 1998;317:1212-6.

This article is licensed under the GNU Free Documentation License. It uses material from the Wikipedia article "Arterial_blood_gas_sampling". A list of authors is available in Wikipedia.
Your browser is not current. Microsoft Internet Explorer 6.0 does not support some functions on Chemie.DE