Remembering the classics: Pulse Oximetry or Capnography – How did we get here?

Charlie Cote, MD

Pulse oximetry or capnography?  Over the past few weeks, there have been several PAADs on this question. Today in most of the resource rich world, this really isn’t an either or question, after all we are rich enough to be able to afford and have access to both. Indeed, they are required. It may be hard for most of you to believe but when these devices first came to market in the 1980s, hospital and department administrators had to be convinced that these devices were “worth it”.   I’ve asked the author of some of the original articles on this subject, Dr. Charlie Cote, to tell us the story behind the story of how our anesthesia practice evolved.  Myron Yaster MD

Original articles

Coté CJ, Goldstein EA, Cote MA, Hoaglin DC, Ryan JF. A single-blind study of pulse oximetry in children.Anesthesiology. Feb 1988;68(2):184-8. PMID: 3277484

Coté CJ, Rolf N, Liu LM, et al. A single-blind study of combined pulse oximetry and capnography in children.Anesthesiology. Jun 1991;74(6):980-7. PMID: 1904206

When I was a junior attending at the Massachusetts General Hospital, I was always looking for issues to study.  At an ASA meeting I discovered stand-alone expired carbon dioxide monitoring devices as a potential operating room monitor and talked the company; they graciously provided a device for use in a prospective convenience sample of children.  At the time there was no wave form, simply a needle swinging from right to left and back. We studied 331 children and diagnosed 35 events including circuit disconnects, extubations, kinked endotracheal tubes, and endobronchial intubations.3  We found a high incidence of hyperventilation in infants < 1 year of age and hypercarbia in mask cases.  After this study all pediatric rooms and locations at the MGH were equipped with such monitors (but not the adult rooms!). We also diagnosed two cases of malignant hyperthermia long before elevated CO2 values were considered the sine qua non of MH.4

Shortly after these experiences, again at the ASA, I discovered pulse oximetry as a possible operating room monitoring device. I met Dr. William New, a Stanford anesthesiologist/engineer who had built the first clinically useful pulse oximeter device in his garage the previous year and formed a new company, Nellcor.  I asked if I could have one for a study in children and they gave me one along with a small grant to pay a summer student for a research project.  We then conducted the first prospective randomized, convenience sample, study of pulse oximetry.  The summer student, Andrew Goldstein, recorded all anesthetics on a continuous recording device and when he observed an event (desaturation ≤ 85% for 30 seconds or longer) he was instructed to inform the anesthesia team (the data were blinded) and then after resolution of the event asked if they recognized the event or if the oximeter made the diagnosis and also did they observe cyanosis.  35 such events were observed in 152 children (24 in the data unavailable vs 11 in the data available group).  Interestingly half the time when asked if the patient was cyanotic the team responded “yes” (when the saturation was above 85% or greater) and half the time when a patient had to be cyanotic (saturation <70%) they responded “no”; clearly, we were terrible at diagnosing cyanosis.  There also was a greater number of desaturation events in infants ≤2 years of age. 

Another interesting observation was when Andrew arrived at work on a Monday morning with his fingernails painted in five different colors.  He said: “I think nail polish can affect the oximeter reading” (I suspect he was playing with this device with a girlfriend, but I never asked).  We then conducted a blinded study and found that those colors whose absorption wavelength fell within that of the oximeter (between 660 and 940 nm) resulted in lower saturations; this now is a standard question on board examinations (answer blue and green) and why patients are asked to remove their nail polish prior to anesthesia.5 

Also during this study there was a major hypoxic event that occurred in a young adult patient from which she recovered due to the surgeon’s elbow kinking the tracheal tube and the resident thinking it was a vagal reflex and treating the bradycardia with two doses of atropine prior to arrest.  After that event and now with solid data from our pediatric capnography and oximeter studies, all anesthesia locations were equipped with both oximetry and capnography; this preceded the establishment of ASA monitoring standards.

These experiences then raised the question: Which is better: pulse oximetry or capnography in diagnosing potential life-threatening events?  Therefore, our group conducted the first randomized prospective study of four conditions: 1) just oximetry, 2) just capnography, 3) neither oximetry or capnography, and 4) both oximetry and capnography data available to the anesthesia team.  As in the prior studies we recorded all data continuously for further analysis, including continuous EKG.  At this time the ASA had published their monitoring guidelines so the IRB required the observer to be a fully trained anesthesiologist; Norbert Rolf from Germany was our observer.  As in the prior study, he notified the anesthesia team of evolving events if they already had not made an intervention.  We documented 260 events in 153 children. 59 events in 43 patients were considered “major” desaturation events (SpO2 ≤ 85% for ≥ 30 seconds). There were 15 “major” capnograph events, (esophageal intubation, disconnection, extubation, tracheal tube obstruction) in 11 children.  There were 130 “minor desaturation events {SpO2 ≤ 95% for 60 seconds or longer) and 70 “minor” capnograph events (hypercarbia or hypocarbia).  We again found a strong age effect with infants ≤ 6 months compared with those 7-24 month compared with those older than 24 months. Blinding the oximetry data resulted in nearly a threefold increase in events (31 vs. 12) but blinding the capnography data neither altered the frequency of desaturation events or major capnography events. Blinding the capnography data did result in a greater incidence of either hypo or hyperventilation (47 vs. 22).  Figure 1 illustrates how these monitors overlapped in helping to diagnose safety issues; figures 2 and 3 illustrate the value of oximetry and superiority over capnography. Once again, we found a poor relation between clinical diagnosis of desaturation and the actual saturation: 19 children had >5 grams of desaturated hemoglobin and 9 were described as cyanotic while 10 as acyanotic! The oximeter provided the first evidence of a major desaturation event in ~70% of cases, ~22% diagnosed clinically, and only ~8% diagnosed first by the capnograph. We concluded that oximetry provided the earliest warning of developing events, and that capnography may also provide a warning, but that desaturation frequently accompanied the capnography events.  The use of capnography did reduce the incidence of either hyper or hypoventilation and the number of such events was reduced when both monitors were used simultaneously.  One further wrinkle that may have affected our results was that we used in-line humidification, and that this frequently resulted in obstruction of the CO2 sampling line, thus there were many false alarms. In fact, between oximetry and capnography there were ~4,000 alarm situations with only 260 being clinically relevant.  I suspect that the new capnometers that do to clog with humidification would reduce false alarms from this cause.  I am also certain that having CO2 wave forms to follow greatly improves the value of capnography in the early recognition of airway problems but also provides incredible value is assessing the adequacy of chest compressions during a code.

One spinoff from this study was the issue of arrhythmias during anesthesia.  We found 24 children with persistent ventricular arrhythmias with 21/24 occurring in children managed by face mask.  In 8 patients this was associated with hypercarbia and in 8 others with signs of light anesthesia; we found no cases in infants < 2 years of age. There was a greater incidence in those anesthetized with halothane compared with other agents.6 A second spinoff study found a greater incidence of desaturation events in children with signs of a URI with an increased frequency of bronchospasm in those who were intubated.7  A third spinoff was our discovery that we could achieve excellent oximetry data when placing a pulse oximeter probe in the corner of the mouth or around the tongue (using the metal from a disposable paper face to give is shape).8

In summary I would say that these experiences are perfect examples of one question turning over a rock and discovering two more that then result in more rock turning and more studies.  Nothing beats objective clinical observations confirmed by data.

1.           Coté CJ, Goldstein EA, Cote MA, Hoaglin DC, Ryan JF. A single-blind study of pulse oximetry in children. Anesthesiology. Feb 1988;68(2):184-8.

2.           Coté CJ, Rolf N, Liu LM, et al. A single-blind study of combined pulse oximetry and capnography in children. Anesthesiology. Jun 1991;74(6):980-7.

3.           Coté CJ, Liu LM, Szyfelbein SK, et al. Intraoperative events diagnosed by expired carbon dioxide monitoring in children. Can Anaesth Soc J. May 1986;33(3 Pt 1):315-20.

4.           Baudendistel L, Goudsouzian N, Coté C, Strafford M. End-tidal CO2 monitoring. Its use in the diagnosis and management of malignant hyperthermia. Anaesthesia. Oct 1984;39(10):1000-3.

5.           Coté CJ, Goldstein EA, Fuchsman WH, Hoaglin DC. The effect of nail polish on pulse oximetry. Anesth Analg. Jul 1988;67(7):683-6.

6.           Rolf N, Coté CJ. Persistent cardiac arrhythmias in pediatric patients: effects of age, expired carbon dioxide values, depth of anesthesia, and airway management. Anesth Analg. Dec 1991;73(6):720-4.

7.           Rolf N, Coté CJ. Frequency and severity of desaturation events during general anesthesia in children with and without upper respiratory infections. J Clin Anesth. May-Jun 1992;4(3):200-3.

8.           Coté CJ, Daniels AL, Connolly M, Szyfelbein SK, Wickens CD. Tongue oximetry in children with extensive thermal injury: comparison with peripheral oximetry. Can J Anaesth. May 1992;39(5 Pt 1):454-7.

 Charlie Cote, MD