Hypoventilation, apnea, and airway obstruction are common during procedural sedation. Because of improvements in surgical techniques and the development of newer IV sedative drugs, conscious and deep sedation are being increasingly provided in operating rooms, clinics, and offices by trained anesthesia providers and by practitioners without specialized training in pharmacology, physiologic monitoring, or resuscitation. Pulse oximetry, routinely used during monitored anesthesia care (MAC)/sedation, is a reliable estimate of oxygenation; however, detection of apnea or airway obstruction, as evidenced by a decline in arterial oxygen saturation (SpO2), can be delayed, especially when patients are breathing supplemental oxygen1 (1). Electrical impedance respiratory rate monitoring is a well established technique (2–4), but it can be technically difficult, depending on the surgical site. Also, chest wall movement can occur with airway obstruction, which is interpreted by an impedance monitor as “breathing” (5–7). Although capnography has been shown to correlate significantly with PaCO2 level (8), the reliability of nasal capnography to monitor apnea and airway obstruction has not been examined in MAC/sedated patients, nor has it been compared with plethysmography, the “gold standard.”
The purpose of this study was to determine whether capnography accurately detects apnea during sedation. Because clinicians routinely use different oxygen flow rates during procedural sedation, a secondary goal was to determine whether oxygen flow rate affected the ability of the capnograph to detect apnea.
Forty patients scheduled to undergo procedures with MAC/sedation were enrolled after signing an IRB-approved consent form. The sample size was designed to detect a 20% difference in the apnea detection rate. One patient was converted to a general anesthetic at the start of the case and was excluded from further study participation.
Patients were excluded from study participation if they were pregnant, aged <18 yr, or could not maintain a SpO2 of >88% on room air. Dropout criteria included the need to place an artificial airway to maintain ventilation or the need to institute artificial ventilation.
A cannula designed to administer nasal oxygen and sample both nasal and oral carbon dioxide (Smart CapnoLine™O2 O2/CO2 Oral Nasal Cannula; Oridion, Jerusalem, Israel) was appropriately positioned to measure Petco2 with a handheld capnometer (NPB-70; Nellcor, Pleasanton, CA; sampling rate, 50 mL/ min via both the oral and nasal ports). We chose this cannula because of availability, and a recent abstract has shown that when this oral/nasal cannula is compared with the standard nasal cannula, there is no benefit in CO2 detection (9). The capnogram was calibrated with a 5% CO2 gas container. Transthoracic impedance monitoring (970S; Respironics, Pittsburgh, PA) was used as the control for respiratory rate monitoring on all patients, and lead placement was standardized as recommended by the manufacturer.
Before the administration of IV drugs, the accuracy of the transthoracic impedance-derived respiratory rate was verified by timed coincident visual inspection of chest motion associated with breathing for 1 min. A five-lead electrocardiogram and SpO2 monitoring were displayed continuously for all patients. Noninvasive arterial blood pressure was measured every 2.5 min. Sedation and analgesia were administered at the discretion of the anesthesia care providers (anesthesia residents and nurse anesthetists supervised by faculty anesthesiologists at a large teaching institution), who were blinded to the capnography and transthoracic impedance data. The oxygen flow rate through the nasal cannula was randomized to 0, 2, 4, and 6 L/min for 3 min at each flow rate. Every 12 min, the randomized sequence of oxygen trials was repeated throughout the study.
Data were collected at baseline and at the end of each 3-min trial, unless otherwise triggered by apnea for >20 s or SpO2 <88%. Values for SpO2 and Petco2 were collected during the last minute of each 3-min trial and averaged. Apnea or airway obstruction for 20 s, detected by using transthoracic impedance monitoring, triggered notification of the anesthesia care provider if the apnea was undetected by routine monitoring. Twenty seconds was specifically chosen because of safety concerns of the IRB.
Categorical data are presented as raw values, percentages, or incidences. Continuous data are summarized as mean ± sd. Continuous data were compared by using a two-factor (time and oxygen flow rate) analysis of variance. When the F ratio was significant at the 0.05 level, Tukey’s honestly significant difference post hoc test was used to distinguish differences in means. Logistic regression analysis was used to examine simultaneously whether apnea was predicted by age, sex, weight, smoking history, alcohol ingestion, sleep apnea, chronic obstructive pulmonary disease, sedative, analgesic, procedure, and/or duration of sedation.
Thirty-nine patients who ranged in age from 18 to 81 yr (53 ± 15 yr) and weighed from 41 to 121 kg (79 ± 20 kg) were studied. There were 21 women and 18 men. Midazolam (1.5 to 2.0 mg IV) was given to 31 patients (80%). Single-drug cases included fentanyl (n = 1), ketamine hydrochloride (n = 1), midazolam (n = 4), and propofol (n = 5). Patients who received midazolam subsequently received propofol alone (n = 10), fentanyl alone (n = 9), or both (n = 9). Procedures and the associated incidences of apnea are summarized in Table 1. No subject met dropout criteria.
Ten (26%) of 39 patients experienced 20 s of apnea. None of the 10 episodes of apnea was detected by the anesthesia provider. All were detected by both capnography and transthoracic impedance. There was no difference in apnea detection between the transthoracic impedance monitor and capnograph. The time to apnea ranged from 3 to 63 min (15 ± 14 min) after the onset of sedation. Only in one case did the SpO2 decrease to less than 90%. The SpO2 of patients after 20 s of apnea ranged from 82% to 98% (92% ± 6%). The patient who desaturated to an SpO2 of 82% was undergoing the zero oxygen flow arm of the trial. Petco2 varied inversely with oxygen flow rate. There were no differences in heart rate or arterial blood pressure throughout the study.
Procedures performed with IV sedation are being done both in and out of the operating room, and sedatives are titrated by a variety of health care providers. Hundreds of thousands of procedures are performed each year with some form of hypnotic drug. Deaths from outpatient office-based procedures have occurred, specifically in patients who have experienced respiratory or circulatory arrest, presumably because of oversedation (10). This study suggests that capnography is an accurate means of detecting apnea and that it is not affected by oxygen flow rate.
The American Society of Anesthesiologists’Standards for Basic Anesthesia Monitoring state that “During regional anesthesia and monitored anesthesia care, the adequacy of ventilation shall be evaluated, at least, by continual observation of qualitative clinical signs.” Although continuous capnography is required for all patients undergoing general anesthesia, it is optional for MAC/sedation cases. The need for CO2 monitoring has been studied by other medical specialties that frequently use procedural sedation, including gastroenterology (11), dentistry (12,13), and emergency medicine (14–16), and many specialties now recommend capnography as a standard monitoring device. No studies have compared the incidence of apnea/ obstruction detection with capnography versus impedance plethysmography, nor have the effects of oxygen flow rate on capnographic detection of apnea been evaluated.
In this study, 26% of our patients undergoing MAC with sedation/analgesia experienced apnea of at least 20 seconds. None of the patients experienced hemodynamic changes. None of the episodes was detected by the anesthesia providers, and our study did not analyze how long the apnea would have continued without intervention. Reasons for the lack of detection are unclear, but one could argue that 20 seconds is a short, possibly inconsequential episode, especially in patients breathing supplemental oxygen. During apnea, the reduction in SpO2 is moderated by the reserve of oxygen in gas-exchanging alveoli. For example, Farmery and Roe (17) reported that while patients breathed room air, apnea resulted in SpO2 ~85% within 60 seconds.
The implications of this study are that episodes of apnea for at least 20 seconds are frequent during MAC and these episodes can go undetected by anesthesia care providers. Left undetected, this persistent hypoventilation could lead to progressive hypercarbia and acidosis, which may be significant by the time the SpO2 monitor alarms in a patient receiving supplemental oxygen. Apnea can be readily and easily detected by using nasal capnography, which has the advantage of detecting airway obstruction with chest wall movement, versus chest wall plethysmography. Although supplemental oxygen delivery affects the amplitude of the capnograph, detection of apnea is unchanged.
Limitations of this study, as previously mentioned, include the 20 second limit on apnea. One could argue that allowing for a longer apnea time may end in spontaneous resolution or that such a short period is not clinically relevant, as evidenced by the lack of hemodynamic instability. However, longer apnea might result in more significant arterial oxygen desaturation and hypercapnia, placing the patient at an unnecessary and unacceptable risk. The ethics of performing a study until safety is compromised are questionable, especially when it is established that apnea and airway obstruction are undesirable events during any anesthetic.
Capnography provides both a quantitative and qualitative measure of ventilation, but it provides only an approximation of changes in tidal volume, as noted by the wave form, if gas inflow is constant. As shown in Table 2, increasing oxygen flow rates decreased the amplitude of measured CO2, probably via dilution, making the quantitative value as an assessment of adequacy of ventilation less reliable. All episodes of apnea, however, were still detected accurately regardless of flow rate.
Esophageal or precordial stethoscopes provide another inexpensive measure of ventilatory competence, but their use has been declining in recent years, with one study suggesting that only 11% of patients are monitored in this fashion (18). Outcomes studies comparing patients monitored with capnography versus auscultation have not been performed; however, one study in the dental literature compared the techniques and found the same rate of incident detection, although normal quiet breathing seemed to be more difficult to detect in adults than children (19).
The American Society of Anesthesiologists guidelines for sedation by nonanesthesiologists strongly recommend monitoring of ventilation, but they are less emphatic on the use of capnography because of the lack of supporting studies (20). Although we are cautious in our extrapolation of these results to the nonanesthesia provider in an outside location, we believe that our protocol represents a “best-case scenario.” As anesthesiologists, our primary goal is to ensure patient safety and comfort. With procedural sedation being an increasingly popular option used by trained anesthesia care providers and nontrained individuals alike, it is imperative that research efforts in procedural sedation be directed toward facilitating the diagnosis of scenarios which can lead to adverse events before they occur.
1Fu E, Neymour R, Downs J. Routine supplemental oxygen is not necessary during post-anesthesia recovery [abstract]. Anesth Analg 1999;88:S39.
1. Downs J. Prevention of hypoxemia: the simple, logical, but incorrect solution. J Clin Anesth 1994;6:180–1.
2. Coates AL, Vallinis P, Mullahoo K, et al. Pulmonary impedance as an index of severity and mechanisms of neonatal lung disease. Pediatr Pulmonol 1994;17:41–9.
3. Haborne D. Measuring respiratory rate. Arch Emerg Med 1992; 9:377–8.
4. Freundlich JJ, Erickson JC. Electrical impedance pneumography for simple nonrestrictive continuous monitoring of respiratory rate, rhythm and tidal volume for surgical patients. Chest 1974; 65:181–4.
5. Javidson J, Hosie H. Limitations of pulse oximetry: respiratory insufficiency—a failure of detection. BMJ 1993;307:372–3.
6. Hutton P, Clutton-Brock T. The benefits and pitfalls of pulse oximetry. BMJ 1993;307:457–8.
7. Wiklund L, Hok B, Stahl K, Jordeby-Jonsson A. Postanesthesia monitoring revisited: frequency of true and false alarms from different monitoring devices. J Clin Anesth 1994;6:182–8.
8. Liu SY, Lee TS, Bongard F. Accuracy of capnography in nonintubated surgical patients. Chest 1992;102:1512–5.
9. Mattison R, Dingmann C, Jameson L. Comparison of Smart Capnoline O2
nasal cannula and standard nasal O2
cannula for respiratory monitoring in sedated patients [abstract]. 2003 SAMBA Abstracts. 2003; P-34.
10. Vila H, Cantor A, Mackey D, Soto R. Florida office surgery and ambulatory surgery center outcomes following implementation of office regulations [abstract]. Anesth Analg 2003;96:S4.
11. Nelson DB, Freeman ML, Silvis SE, et al. A randomized, controlled trial of transcutaneous carbon dioxide monitoring during ERCP. Gastrointest Endosc 2000;51:288–95.
12. Anderson J, Clark P, Kafer E. Use of capnography and transcutaneous oxygen monitoring during outpatient general anesthesia for oral surgery. J Oral Maxillofac Surg 1987;45:3–10.
13. Bennett J, Peterson T, Burleson J. Capnography and ventilatory assessment during ambulatory dentoalveolar surgery. J Oral Maxillofac Surg 1997;55:921–5.
14. Hart LS, Berns SD, Houck CS, Boenning DA. The value of end-tidal CO2
monitoring when comparing three methods of conscious sedation for children undergoing painful procedures in the emergency department. Pediatr Emerg Care 1997;13: 189–93.
15. Tobias JD. End-tidal carbon dioxide monitoring during sedation with a combination of midazolam and ketamine for children undergoing painful, invasive procedures. Pediatr Emerg Care 1999;15:173–5.
16. McQuillen KK, Steele DW. Capnography during sedation/ analgesia in the pediatric emergency department. Pediatr Emerg Care 2000;16:401–4.
17. Farmery AD, Roe PG. A model to describe the rate of oxyhaemoglobin desaturation during apnoea. Br J Anaesth 1996;76: 284–91.
18. Prielipp R, Kelly J, Roy R. Use of esophageal or precordial stethoscopes by anesthesia providers: are we listening to our patients? J Clin Anesth 1995;7:367–72.
19. Croswell R, Dilley D, Lucas W, Van W. A comparison of conventional versus electronic monitoring of sedated pediatric dental patients. Pediatr Dent 1995;17:332–9.
20. Practice guidelines for sedation and analgesia by non-anesthesiologists. Anesthesiology 2002;96:1004–17.