Thermistors are typically connected to a DC amplifier or via appropriate interfaces to an AC amplifier for printout to a polygraph. The thermistors are then placed over the mouth and nostrils of the patient and function on the principle of expired air being higher temperature than inspired air.¹ Over the course of two hundred sleep apnea studies in our lab, various disadvantages in this technique became apparent. During apparently equivalent apneic episodes, as indicated by thermistors, a wide variation in oximetry values were often observed.The variation was sometimes greater than 15%. A similar discrepancy was noted during hypopnic events. The oximetry values were supported by a variety of factors including mean SaO2 before and after the episodes, duration of episodes, and duration of transient arousals. These discrepancies led us to believe that the thermistors may be reflecting only approximate airflow changes. After the purchase of End Tidal CO2 monitors for the sleep lab, we connected them to DC amplifiers and used the channel immediately above the thermistor channel for comparison. Noticeable from the start was the three second delay in response time as compared to the thermistor signal. Once the delay was compensated for, the reasons for the discrepancies noted above became apparent. What would have been scored as an apnea if the thermistor alone had been used, was clearly a hypopnea, and what would have been scored as a hypopnea, was an apnea. How could a thermistor cause a hypopnea to appear as an apnea? Why would we see a flat line from the thermistor channel indicating apnea, and at the same time the ETCO2 monitor would still be showing a signal of approximately 30% of basal value? The reason for this could only be that the temperature of the expired air was insufficient for thermistors to measure or the amount of flow was insufficient. However, the ETCO2 monitors showed that a gas exchange was indeed taking place, therefore making the event hypopneic as opposed to true apnea. How could a thermistor cause a hypopnea to be an apnea? We would have a flatline on the ETCO2 channel indicating apnea, however the thermistor channel would still indicate a flow of the same frequency as basal value, and amplitude of approximately 30%. The End Tidal channel was not occluded as the signal would return at the end of the apparent respiratory event.During these events we have noticed a phenomenon that we refer to as “fish-mouthing”. “ Fish-mouthing” is when a patient, during an obstructive apnea, would mimic oral breathing. Pushing warm air from the mouth, unable to allow air to exit from the airway, the thermistors would register respirations but the ETCO2 monitor would show that no actual gas exchange was taking place. Despite their limitations we did find reason to maintain the thermistors instead of just replacing them with ETCO2 monitors. Humidity from the patients breath could, from time to time, temporarily cause the cannula to occlude. While the ETCO2 monitor does have a feature to correct this problem, the signal could be lost for up to a few minutes at a time. The possibility of it occluding at a crucial part of the test is significant enough for us to maintain the redundant monitoring technique. We have since recorded over two hundred more studies using this technique. We have found that the most noticeable advantage is that as there is less ambiguity of signals on the chart, we are able to increase the accuracy of the diagnosis as well as decrease the amount of time required to interpret and score the record.

______________________________________________________________________________

1 Bornstein, s. Respiratory Monitoring During Sleep: Polysomnography. In: Guilleminault C, ed., Sleeping and Waking Disorders, Indications and Techniques. Menlo Park, CA: Addison-Wesley, 1982: 185-203