Classic Papers Revisited  |   January 2007
An Early Example of Evidence-based Medicine: Hypoxemia due to Nitrous Oxide
Author Affiliations & Notes
  • Frederick W. Cheney, M.D.
  • * Professor and Chair.
Article Information
Classic Papers Revisited / Respiratory System
Classic Papers Revisited   |   January 2007
An Early Example of Evidence-based Medicine: Hypoxemia due to Nitrous Oxide
Anesthesiology 1 2007, Vol.106, 186-188. doi:
Anesthesiology 1 2007, Vol.106, 186-188. doi:
Click on the links below to access all the ArticlePlus for this article.
Please note that ArticlePlus files may launch a viewer application outside of your web browser.
IN 1955, Dr. B. Raymond (Ray) Fink (1914–2000) was faculty in the Department of Anesthesiology, Columbia University College of Physicians and Surgeons (New York, New York) when he published one of the first articles describing a mechanism by which anesthesia could cause hypoxemia.1 At the time, cyanosis was a recognized phenomenon during recovery from general anesthesia despite apparently good ventilation. It was also recognized that this was more frequent in patients anesthetized with a nitrous oxide–oxygen mixture. There were previous reports of a decrease in arterial oxygen saturation as measured by an ear oximeter when patients began to breathe room air, especially after a nitrous oxide–oxygen anesthetic.2,3 
Dr. Fink postulated that this hypoxemia was due to the high tension at which nitrous oxide was administered for anesthesia and its high solubility in the blood relative to the nitrogen in room air. In an in vitro  experiment, he saturated water in a half-filled flask with nitrous oxide. He stoppered the flask and showed an increased pressure in the flask due to nitrous oxide diffusing out of the water. As the gas equilibrated in the flask, the oxygen concentration of the gas in the flask decreased to approximately 17%. He hypothesized that this simulated the condition in the lungs at the end of a nitrous oxide–oxygen anesthetic.
Ray followed this simple experiment with clinical observations in eight otherwise healthy patients undergoing gynecologic surgery who were anesthetized with 75% nitrous oxide–25% oxygen supplemented with sodium thiopental. Blood oxygen saturations were measured by either a Wood ear oximeter or brachial arterial blood analysis. Removal of the facemask at the end of the nitrous oxide–oxygen anesthetic resulted in a decrease in arterial oxygen saturation from 4.5% to 10.3% with an average of 7.9%, an effect that lasted approximately 10 min. Figure 2 of his article1 (fig. 1) showed the effect of air breathing after 3.5 l/1 l nitrous oxide–oxygen anesthesia in one patient. The “diffusion anoxia” peaked at 3 min and was readily reversed by inhalation of 100% oxygen.
Fig. 1. Arterial oxygen saturation during recovery from nitrous oxide–oxygen anesthesia. Patient M.R. From Fink  1; reproduced with permission. 
Fig. 1. Arterial oxygen saturation during recovery from nitrous oxide–oxygen anesthesia. Patient M.R. From Fink  1; reproduced with permission. 
Fig. 1. Arterial oxygen saturation during recovery from nitrous oxide–oxygen anesthesia. Patient M.R. From Fink  1; reproduced with permission. 
The variability of change in oxygen saturation among the patients studied was attributed to differing levels of ventilation and lung volume. Fink observed that if patients were hyperventilating, the effect would be less, and the effect would be greater in patients with low lung volumes in which to dilute the nitrous oxide. The diffusion anoxia could be reversed by increasing the inspired concentration of oxygen. This led to his recommendation that oxygen should be administered at the conclusion of nitrous oxide–oxygen anesthesia.
This article is a classic because it is an early example of the application of scientific principles to the description of a mechanism to explain the “cyanosis” often associated with emergence from anesthesia. The work is an early example of “evidence-based medicine” in that it provided scientific justification for the necessity for administration of oxygen on emergence from nitrous oxide–oxygen anesthesia.
However, the clinical significance of the effect was challenged in 1961 by Rackow et al.  ,4 also from Columbia, who studied nitrous oxide excretion in artificially ventilated patients and concluded that the decrease in arterial oxygen saturation (Sao2) observed by Fink could not be attributed to dilution of alveolar oxygen by nitrous oxide alone. The decrease in Sao2due to nitrous oxide excretion alone was 2–3% during the first 15 min of excretion, which is in contrast to the more pronounced changes seen in the Fink study. These authors also emphasized the insolubility of nitrogen and the high solubility of nitrous oxide in the blood as a major issue. The more correct term hypoxia  as opposed to anoxia  was also used in this article.
In 1969, Frumin and Edelist (Department of Anesthesiology, Albert Einstein College of Medicine, Yeshiva University, Bronx, New York) published an article titled “Diffusion Anoxia: A Critical Reappraisal.”5 They studied 18 healthy, spontaneously breathing patients during change from breathing 79% nitrous oxide–21% oxygen to air breathing and observed Sao2values above 90 and usually above 93%. However, they did observe Sao2values below 90% in 4 patients who had respiratory irregularities or airway obstruction. These low saturations lasted only as long as the period of obstruction. The authors concluded that the entity of diffusion anoxia was not a clinically significant phenomenon in healthy patients who maintained normal ventilation.
Tom F. Hornbein, M.D., an Associate Professor (and later Chair) of Anesthesiology at the University of Washington (Seattle, Washington), where Ray Fink had migrated in 1964 as Director of Anesthesia Research, questioned the validity of Frumin and Edelist’s findings in a letter to the editor of Anesthesiology.6 These investigators had measured arterial oxygen tension (Pao2) and converted it to Sao2. Hornbein pointed out that changes in Sao2are less dramatic than Pao2because of the slope of the dissociation curve at the Pao2levels studied. He opined that the patients might tolerate a “touch of” diffusion hypoxia, but when added to other possible factors threatening oxygenation during emergence from anesthesia, it does place the “diffusion” effect in the realm of clinical significance.
The Fink article stimulated subsequent studies in humans7,8 and animals8 using indwelling Pao2electrodes, which did detect a modest “diffusion effect.” The various authors emphasized its contribution to hypoxemia developing in the emergence period of anesthesia.7,8 As measurement of arterial blood gases and other modalities of lung function such as imaging, diaphragmatic function, and lung volume became clinically available, the original concept of diffusion anoxia has faded into that of a historical footnote. Alveolar collapse or atelectasis due to a variety of causes during emergence from anesthesia9 has been recognized as the major contributor to the hypoxemia or “cyanosis” observed by early practitioners of our specialty. Ray Fink did note that lung volume and ventilatory status contributed to diffusion anoxia, but he did not recognize their major contribution to emergence hypoxemia. The original “evidence” for the need for a higher inspired concentration in the patient emerging from anesthesia turns out to have been a minor player in a much larger league.
This author recalls learning the concept of diffusion anoxia during his residency (1961–1964) and teaching it to residents over the ensuing 40 yr as a faculty member at the University of Washington. Therefore, the evidence has changed, but the practice of administering oxygen during emergence from anesthesia, which was originally supported by the concept of diffusion hypoxia, is still necessary and central to the practice of modern day anesthesiology. This is why Ray Fink’s pioneering work in 1955 remains a classic in the field.
What is also remarkable about this article is that it was only Ray’s fourth peer-reviewed publication. He graduated from his residency in anesthesiology at Beth Israel Hospital in New York in 1952 and joined the faculty at Columbia as an Instructor in anesthesiology. He published three articles in 1954, two of which dealt with anesthesia equipment, and one of these described the Fink nonrebreathing valve. The third was a publication on diffusion anoxia published in Acta Anaesthesiologica Belgica  . After his classic article on diffusion anoxia in Anesthesiology, while still at Columbia, Ray turned his fertile mind to the study of laryngeal and diaphragmatic mechanics and regulation of respiration.
Ray and I joined the Department of Anesthesiology at University of Washington at about the same time in 1964. I was an Instructor and he was Professor and Research Director. Over the next 20 yr, he contributed original knowledge to the specialty in the areas of anesthesia toxicity, cell metabolism, local anesthetics, and nerve conduction. His inventive mind, sense of humor, and grasp of science were an inspiration to junior faculty such as myself.
After his retirement to Emeritus status in 1984, he remained active as a researcher and scholar until his death in October 2000. One of the unusual features of Ray’s career is that of 129 publications, he was first author of 104. This is evidence that Ray was the main source of his published creative ideas. His contributions to new knowledge in the specialty were recognized by the American Society of Anesthesiologists in 1987 when he received the Excellence in Research Award and again in 1993 when he received the Distinguished Service Award.
Therefore, this classic article on diffusion anoxia was not only an early example of evidence-based medicine, but also heralded the beginning of a long and productive research career that produced many more important scientific contributions to the scientific basis of the specialty.
Fink BR: Diffusion anoxia Anesthesiology 1955; 16:511–4Fink, BR
McClure RD, Behrmann VG, Hartman FW: The control of anoxemia during surgical anesthesia with the aid of the oxyhemograph. Ann Surg 1948; 128:685–706McClure, RD Behrmann, VG Hartman, FW
Ridley RW, Faulconer A Jr, Osborn JE: Concentrations of oxygen, nitrous oxide, nitrogen and ether and their correlation with certain physiologic variables during surgical anesthesia in man. Anesthesiology 1951; 12:276–92Ridley, RW Faulconer, A Osborn, JE
Rackow H, Salanitre E, Frumin MJ: Dilution of alveolar gases during nitrous oxide excretion in man. J Appl Physiol 1961; 16:723–8Rackow, H Salanitre, E Frumin, MJ
Frumin MJ, Edelist G: Diffusion anoxia: A critical reappraisal. Anesthesiology 1969; 31:243–9Frumin, MJ Edelist, G
Hornbein TF: More about diffusion anoxia (letter). Anesthesiology 1970; 32:284–5Hornbein, TF
Fanning GL, Colgan FJ: Diffusion hypoxia following nitrous oxide anesthesia. Anesth Analg 1971; 50:86–91Fanning, GL Colgan, FJ
Holcomb C, Erdmann W, Corssen G: The significance of diffusion hypoxemia. South Med J 1976; 69:1282–4Holcomb, C Erdmann, W Corssen, G
Duggan M, Kavanagh BP: Pulmonary atelectasis: A pathogenic perioperative entity. Anesthesiology 2005; 102:838–54Duggan, M Kavanagh, BP
Fig. 1. Arterial oxygen saturation during recovery from nitrous oxide–oxygen anesthesia. Patient M.R. From Fink  1; reproduced with permission. 
Fig. 1. Arterial oxygen saturation during recovery from nitrous oxide–oxygen anesthesia. Patient M.R. From Fink  1; reproduced with permission. 
Fig. 1. Arterial oxygen saturation during recovery from nitrous oxide–oxygen anesthesia. Patient M.R. From Fink  1; reproduced with permission.