Oxygen and CO2 Retention in COPD

Revised and reviewed 6 March 2014


  • Excessive oxygen administration can lead to hypercapnic respiratory failure in some COPD patients.
  • COPD patients with more severe hypoxemia are at higher risk of CO2 retention from uncontrolled O2 administration


The traditional explanation is wrong

  • the traditional theory is that oxygen administration to CO2 retainers causes loss of hypoxic drive, resulting in hypoventilation and type 2 respiratory failure. This is a myth.
  • Patients suffering from COPD exacerbations, regardless of whether they have CO2 retention, generally have supra-normal respiratory drive (unless there is impending hypercapnic coma)

The real explanation involves:

  • increased V/Q mismatch (most important)
  • the Haldane effect

V/Q mismatch

  • In COPD, patients optimise their gas exchange by hypoxic vasoconstriction leading to altered alveolar ventilation-perfusion (Va/Q) ratios
  • Excessive oxygen administration overcomes this, leading to increased blood flow to poorly ventilated alveoli, and thus increased Va/Q mismatch and increased physiological deadspace
  • this increase in Va/Q mismatch occurs in both CO2 retainers and non-retainers, the difference is presumably one of degree

The Haldane effect

  • deoxygenated hemoglobin (Hb) binds CO2 with greater affinity than oxygenated hemoglobin (HbO2)
  • hence oxygen induces a rightward shift of the CO2 dissociation curve, which is called the Haldane effect
  • in patients with severe COPD who cannot increase minute ventilation, the Haldane effect accounts for about 25% of the total PaCO2 increase due to O2 administration


  • Target SaO2 88-92% in these patients
  • the targeted approach is associated with decreased mortality in COPD patients and less respiratory acidosis
  • The oxygen flow rate administered is not important, the (alveolar) PAO2 (and, indirectly, the SaO2) achieved is
  • Never withhold oxygen from a seriously ill hypoxic patient due to fear of cause hypercapnic respiratory failure

References and Links

Journal articles

  • Abdo WF, Heunks LM. Oxygen-induced hypercapnia in COPD: myths and facts. Crit Care. 2012 Oct 29;16(5):323. [Epub ahead of print] PMID: 23106947 PMCID: PMC3682248

FOAM and web resources

Print Friendly


  1. Arun says

    Hi Chris, just reading through the explanations and the 2nd dot point on V/Q mismatch talks about increased physiological dead space but I think what you might have meant was an increased shunt. This seems to make more sense due to vasodilation of vessels (secondary to high oxygen tension) with ongoing poor gas exchange due to poor lung parenchyma. The article you have linked to with Dundee chest states a similar thing as well.

    Would be grateful for your clarification.



    • says

      Hi Arun
      Shunt implies that there is blood flow to non-ventilated areas of lung (V/Q = 0).
      This is not necessarily the case.
      When hypoxic vasoconstriction is overcome blood flow is inappropriately distributed to alveoli that are poorly ventilated (but still have some ventilation, i.e. V/Q <1 or V/Q mismatch).
      This manifests as increased physiological dead space -- functionally the lungs as a whole behave as if dead space (non-ventilated portion of lungs) has increased.
      Hope that helps

      • Adam Drenzla says

        hi chris and arun. i think the intuitive difficulty is that both areas of high v/q (more deadspace like) and low v/q (more shunt like), and indeed pure shunt also, contribute to the calculation of physiological deadspace.

        it seems, physiological deadspace is a mathematical construct with no obvious anatomical/physiological compartment (such as anatomical deadspace plus deadspace like areas for instance)… thinking that only high v/q areas contribute to it whilst intuitive is incorrect.

        based on the equation, any part of the the gas/blood interface that causes a high paco2 will contribute toward it. shunt, low v/q, high v/q, dead space…

        for some physiological hoedown see this article by oli of icn fame and neil soni (editor of oh’s) from the vaults of icm ::


  2. says

    In your opinion, does this physiology also hold true for a chronic COPD patient on home oxygen therapy who is not currently in an exacerbation?

    For instance, it’s common to preoxygenate non-COPD patients to increase the patient’s oxygen reserve so that during periods of anesthetic induced decreased ventilation the patient’s oxygen saturation is maintained.

    In your opinion, will pre-oxygenating a COPD patient prior to anesthetic administration for EGD (15 min procedure) or colonoscopy (30 -- 45 min procedure) help keep the patient safe by increasing his/her oxygen reserve or will it be counterproductive towards my goal of avoiding hypoxemia due to the temporary decrease in minute ventilation? During anesthetics for these kind of procedures, patients are almost always breathing spontaneously and sometimes have brief periods of obstruction and/or apnea.

    My inclination is that the increased FiO2 more than compensates for the short-lived decreased minute ventilation and the net effect will be a safer anesthetic. Is the increased CO2 caused by giving excess oxygen to guard against hypoxemia during a short procedure clinically significant if the patient is not in an exacerbation of their COPD?

    If I skip the oxygen, the patient will desaturate and I will need to bag mask ventilate, and if I pre-oxygenate and the minute ventilation decreases to a clinically significant level I will still be at the same end point —- bag mask ventilation. With that reasoning, I could justify the extra oxygen as it could prevent a procedure interrupting bag mask ventilation in many people even though it could precipitate the procedure interrupting bag mask ventilation a few. A patient not in an exacerbation can handle a small increase in CO2 and will breath it off in the recovery room. Is my reasoning sound?

    • says

      If a COPD patient has enough physiological reserve to undergo procedural sedation, this won’t be an issue. Also, as you say, you can always bag them up.
      It is rarely a clinical issue if a patient has close monitoring -- the problem arises in side rooms or when helathcare workers don’t appreciate the phenomenon.