Circulation 76,548 • Volume 20, No. 4 • Winter 2005   Issue PDF

Oxygen May Mask Hypoventilation – Patient Breathing Must Be Ensured

Leo I. Stemp, MD; Michael A. Ramsay, MD, FRCA

To the Editor

We strongly support Overdyk’s communication1 about the many patients daily at risk of respiratory depression, and would like to elaborate on 3 issues he raised in order to pinpoint some of the problems.

First, he states that “Cashman reported an incidence of respiratory depression of. . . 1.3% by bradypnea (RR <10).” In sedated patients, both respiratory rate and tidal volume, and even ETCO2, as isolated indices, fail to correlate consistently with alveolar ventilation (VA), and therefore pCO2. This of course makes sense when one considers the effect of tidal volume on dead space (VD/VT) and therefore VA, and the clinical observation that narcotic-induced respiratory slowing is frequently associated with an increase in tidal volume. The latter, in turn, reduces VD/VT and ameliorates the effect of the decrease in respiratory rate on VA. This, in fact, is the rationale for using opiates as a cornerstone of therapy in managing the early phase of respiratory failure and fatigue characterized by rapid shallow breathing in acutely ill patients. Bottom line: Bradypnea is certainly a signal that opiate therapy may be endangering the patient, but if only for better physiologic understanding, nobody should be under the impression that bradypnea correlates directly with VA and pCO2. Further, even ETCO2 monitoring is problematic since adequate tidal volumes are required to reflect the alveolar pCO2, and shallow breathing may result in a low ETCO2 because of poor mixing of expired gas. Therefore, careful trained interpretation of ETCO2 is necessary in using this as a measure of ventilation in nonintubated sedated patients.

Second, Overdyk states that, “Supplemental oxygen. . . merely postpones the patient’s insidious progress from bradypnea to apnea.” Since supplemental oxygen has no significant effect on ventilation, it has no direct causative effect on the progression from bradypnea to apnea (other than eliminating hypoxic drive). What it does do is mask that natural opiate-induced progression from bradypnea to apnea, by failing to allow the patient to become hypoxemic, which would otherwise cause a pulse oximeter alarm, thereby alerting clinicians to the respiratory danger. It is crucial to appreciate these points—that clinicians, without realizing that they are doing so, are using the pulse oximeter as a gauge of ventilation, and that oxygen masks hypoventilation as detected by pulse oximetry, by maintaining the SpO2, even to the point of apnea.2-4

Finally, Overdyk writes that using pulse oximeters is a “deceptively ineffective approach” to preventing catastrophic respiratory depression—but this is only true when supplementary oxygen is being administered. In a patient breathing room air, pulse oximetry is highly effective in signaling hypoventilation and/or airway obstruction! The problem, again, is that clinicians who use pulse oximetry to monitor ventilation may fail to appreciate the relevant physiology of the Hb-oxygen dissociation curve, and how it affects such a practice (i.e., that the application of oxygen, even 1 or 2 liters by nasal cannulae, moves the patient to the right on the Hb-oxygen dissociation curve, in which case the PaO2 no longer linearly correlates with the SpO2, and as a result, the SpO2 then no longer correlates with alveolar ventilation).

The Practice Guidelines for Sedation and Analgesia by Non-Anesthesiologists5 recommend that supplemental oxygen be administered to all patients undergoing deep sedation. Unfortunately, the Guidelines, which thousands of practitioners look to for guidance, are entirely silent on the complications of this practice. Interestingly, a recent report in the emergency medicine literature highlighted this issue.6 Apparently, these practitioners seem to have more insight into the relationship between the SpO2 and hypercapnea—and the confounding influence of supplemental oxygen—than other clinicians who deal with exactly this interface, in every patient they treat, on a daily basis. For example, Ramsay was criticized (Anesthesiology 2005;102:1066) for failing to use supplemental oxygen in patients administered a dexmedetomidine infusion for airway surgery, the communicating authors completely missing the point that pulse oximetry, expressly on room air, was a vitally—if not the only—accurate monitor of respiratory depression in the reported cases.

As Overdyk indicated, respiratory depression continues to be a frequent liability event in many settings, with PCA use alone translating into “thousands of patients with potentially catastrophic respiratory depression per day.” Similarly, during procedural IV sedation, respiratory complications continue to occur commonly, despite a veritable industry of regulations and policies designed to prevent such critical incidents. Clearly, clinicians and guidelines are missing the critical practical message that breathing is the only thing that counts. In fact, it could be argued that instead of making opiate use and IV sedation safer, pulse oximetry has introduced an unforeseen complication, because now clinicians are misled by a number, instead of being singularly attentive to the physical act of breathing and airway patency—watching the chest go up and down, and listening to the hiss of air through the nose or mouth—as they were in the old days, because back then, those were the only parameters one could monitor to keep the patient safe. Following that rationale leads to the inescapable conclusion that much, if not the majority of respiratory morbidity and mortality that has occurred in the current era (of IV sedation with monitoring via pulse oximetry) could have been avoided simply via the deliberate and purposeful exclusion of the use of supplemental oxygen (which would effectively put a brake on the level of sedation a clinician was willing to administer).

Clinical medicine does need a better mousetrap to directly measure ventilatory sufficiency in non-intubated patients, and we would agree that the APSF is the ideal organization to tackle the problem. Transcutaneous carbon dioxide monitors may provide a solution and certainly warrant further validation. In the meantime, pulse oximetry can be an effective solution, as long as clinicians understand how to use it appropriately. If PCA use on room air results in hypoxemia, then the fix is not to apply oxygen as de facto treatment, but to either decrease the opiate allowed, or if supplemental oxygen is administered,* move the patient to a more closely monitored environment.

Leo I. Stemp, MD
Springfield, MA
Michael A. Ramsay, MD, FRCA.
Dallas, TX 75246

(*The decision to use supplemental oxygen should be clinically based, as with any other drug, not reflexive just because the SpO2 drops. The benign nature of isolated hypoxemia was noted in an earlier communication.7)


  1. Overdyk, FJ. PCA presents serious risks. APSF Newsletter 2005;20(2):33.
  2. Freeman, ML, et. al. Carbon dioxide retention and oxygen desaturation during gastrointestinal endoscopy. Gastroenterology 1993;105:331-9.
  3. Ayas N, et al. Unrecognized severe postoperative hypercapnia: a case of apneic oxygenation. Mayo Clinic Proceedings 1998;73:51-4
  4. Frumin MJ, Epstein RM, et al. Apneic oxygenation in man. Anesthesiology 1959;20:789-98.
  5. American Society of Anesthesiologists Task Force on Sedation and Analgesia by Non-Anesthesiologists. Practice Guidelines for Sedation and Analgesia by Non-Anesthesiologists. Anesthesiology 2002;96:1004-17.
  6. Witting MD, Hsu S, Andres CA. The sensitivity of room-air pulse oximetry in the detection of hypercapnia. Am J Emerg Med 2005;23:497-500.
  7. Stemp L. Etiology of hypoxemia often overlooked. APSF Newsletter 2004;19(3):38-9.