Nasal continuous positive airway pressure (nCPAP) has been shown to be a highly effective means of relieving upper airway obstruction of patients with obstructive sleep apnea (OSA). The upper airway of a patient with OSA acts as a Starling resister, with a collapsible segment in the oropharynx and variable airflow obstruction during the respiratory cycle.1 Application of nCPAP creates a pneumatic splint in the hypopharynx that reduces obstruction.2
The upper airway of a deeply sedated patient exhibits many similarities to that of a patient with OSA. Deep IV sedation causes varying degrees of upper airway obstruction related to alteration of pharyngeal muscle activity, and reduction in respiratory drive and lung volume.3 Propofol, commonly used as an intraoperative sedative/hypnotic drug due to its favorable pharmacokinetic and pharmacodynamic features, is associated with upper airway collapse through a disproportionate decrease in genioglossus tone as assessed by electromyography.4,5 These effects may be augmented by the supine position, respiratory consequences of regional anesthesia, and the effects of other sedatives and opiates administered during anesthetic care.6 Airway collapse may result in hypoventilation and hypercarbia. Undesirable physiological effects of hypercarbia include hypertension, tachycardia, increased pulmonary vascular resistance, acidosis, and cerebral vasodilation. The clinical impact of these effects is not clear but may be important in association with preexisting cardiac, pulmonary, or neurologic conditions. The benefit of postoperative CPAP treatment in OSA patients has been studied extensively.7 However, there have been no studies, to our knowledge, that have investigated the effect of intraoperative nCPAP application in deeply sedated surgical patients.
The purpose of this study was to assess the efficacy of CPAP in reducing arterial partial pressure of carbon dioxide (PaCO2) increase associated with deep IV sedation of adult patients undergoing total knee arthroplasty (TKA) under subarachnoid block (SAB). We hypothesized that nCPAP would decrease PaCO2.
After IRB approval of the study protocol, 40 patients gave informed written consent to be included in this prospective, randomized, controlled clinical trial conducted at New Hanover Regional Medical Center in Wilmington, NC. Inclusion criteria were ASA physical class I, II, or III patients scheduled to undergo TKA who were older than 21 years and younger than 70 years, with a body mass index (BMI) <40. Qualified study participants consented to an anesthetic plan consisting of preoperative femoral nerve block (FNB), SAB using intrathecal morphine with bupivacaine, and IV sedation during the operative procedure. Exclusion criteria included prior diagnosis of OSA, prior stroke with residual neurologic deficit, prior diagnosis of chronic obstructive pulmonary disease, a recent respiratory infection, or current pregnancy. Additionally, patients with chronic opioid or benzodiazepine dependency were excluded. This study was registered with www.ClinicalTrials.gov by Stephen B. Smith, MD, on June 5, 2012 (Identifier NCT01622647).
A random number generator program was used to assign sequential qualified study cases to control group (N = 20) or nCPAP group (N = 20) before collection of additional demographic information and before anesthetic and surgical care. The group assignment of each individual study patient was not known by the anesthesia team until after sedation and the start of the surgical procedure. After study group assignment, baseline data were collected, including age, gender, height, weight, and BMI. A questionnaire was completed to obtain a STOPBANG score.8 STOPBANG score and BMI, proven measures of OSA risk, were assessed to verify group similarity and identify potential risk for airway obstruction during sedation. Anesthetic care in the operating room was provided by 1 of 2 certified nurse anesthetists (CRNAs) working with an anesthesiologist. Participating CRNAs and anesthesiologists were blinded to all arterial blood gas (ABG) results until after completion of all study cases.
The anesthesia protocol was designed to mirror the anesthesia approach used for a large majority of patients undergoing TKA at the sponsoring facility, a community orthopedic specialty hospital. This included SAB and intraoperative deep IV sedation using propofol infusion. Additionally, intrathecal morphine and FNB were provided as components of a multimodal postoperative analgesia program used at the facility. FNB was performed 30 to 60 minutes before the operation using 20 mL of 0.5% ropivacaine, with procedural sedation using midazolam and/or fentanyl at the clinical discretion of the anesthesiologist. After transfer to the operating room, SAB was performed using bupivacaine 0.75% in dextrose 7.5%, dosed at the discretion of the anesthesia team, with the addition of 150 μg preservative-free morphine. Patients received additional midazolam and/or fentanyl as needed before the completion of the SAB procedure. After SAB induction, an infusion pump was used to deliver propofol. Dosage was titrated between 30 and 100 μg/kg/min to maintain sedation level at or near level 2 on the Modified Observer’s Assessment of Alertness/Sedation (MOAA/S) scale. All patients received supplemental oxygen using a facemask until the time that ABG-1 was collected. Nasal or oral airways were placed at the discretion of the anesthesia team, but they were not used in conjunction with nCPAP.
Five minutes after surgical skin incision, a baseline analysis (ABG-1) was collected from all study patients by radial artery puncture. Analysis was performed by a technician using the i-STAT 1 analyzer (Abbott Laboratories, Abbott Park, IL). Immediately after ABG-1 was collected, the CRNA opened a sealed group assignment envelope and those patients randomized to the nCPAP group were fitted with an nCPAP mask connected to a CPAP machine (RESMED ESCAPE-S8 model, ResMed Corp, San Diego, CA). Airway pressure was set to 8 cm of water, a mid-range setting on the device, for all CPAP group patients, and oxygen was added to the system via flowmeter as needed to maintain SaO2 at least 98%. Patients randomized to the control group continued to receive supplemental oxygen via mask. Thirty minutes after ABG-1, a second specimen (ABG-2) was collected from all study patients.
The primary objective of this study was to assess the efficacy of CPAP for reducing the PaCO2 increase associated with deep IV sedation among patients undergoing TKA under SAB. Secondarily, the protocol facilitated evaluation of any change of PaCO2 after an additional 30-minute period of deep sedation among 20 control subjects.
There are limited published data on the extent of hypercarbia that occurs during deep sedation of surgical patients, and we know of no published data concerning the impact of nCPAP during deep sedation. No formal sample size calculation was completed. Rather, a sample size of 40 subjects was chosen for a pilot study to provide an initial assessment of a PaCO2 increase during deep intraoperative sedation and the effect of nCPAP treatment. The primary end point for comparison of the control group and nCPAP group was change from baseline in PaCO2 measured 30 minutes after the institution of nCPAP in the test group. Additional data, including baseline PaCO2, STOPBANG questionnaire score, BMI, gender, patient age, and drug dosage, were collected. Continuous data are expressed as median, along with 10th and 90th quantiles. Ninety-five percent confidence intervals (CIs) for the median were determined based on order statistics. Continuous data were compared between groups using the Mann-Whitney U test, and categorical data were compared using Fisher exact test. The location shift parameter and 95% CI for the location shirt parameter were calculated using the Hodges-Lehmann approach. The Wilcoxon signed rank test was used to compare PaCO2 values at ABG-1 and ABG-2. A value of P < 0.05 was considered statistically significant. Analysis was performed using JMP software, a product of SAS Institute, Cary, NC.
Forty adult patients completed the study protocol (20 in each group) between July 2012 and July 2013. The baseline PaCO2 values for the full study population (N = 40) to the time of ABG-1 ranged from 38 to 76 mm Hg (median PaCO2 [10th–90th quantiles] = 55 mm Hg [45.8–67]). Table 1 presents a comparison of study group demographic information and data collected to the time of ABG-1. The groups were similar with respect to demographic characteristics, STOPBANG score, BMI, and dosing of midazolam and fentanyl, and there was a similar degree of hypercarbia during deep sedation in both groups.
PaCO2 values at ABG-1 (baseline) and ABG-2, as well as the change in PaCO2 values between these 2 time points, are presented in Table 2. After a 30-minute period of continuous sedation at MOAA/S level 2, ABG-2 revealed no difference in PaCO2 compared with baseline ABG-1 values in the control group (median change from baseline PaCO2 [10th–90th decile] = control: 0.95 [−4.75, 9.85]; P = 0.52; 95% CI median = −3.4 to 3.4). In contrast, after a 30-minute period of nCPAP treatment, there was a significant decline in PaCO2 in the nCPAP group compared to baseline ABG-1 values (median change from baseline PaCO2 [10th–90th decile] = control: 0.95 [−4.75, 9.85] vs −4.60 [−14.55, 3.85]; P = 0.015; 95% CI for the location shift = −9.5 to −1.3). The distribution of study group patients with mild, moderate, and severe hypercarbia (defined as 42–49, 50–59, and ≥60 mm Hg, respectively) at ABG-1 and ABG-2 is presented in Figure 1.
As noted, an oral or nasal airway was placed in 16 study patients before ABG-1. Table 3 presents PaCO2 values at ABG-1 by airway use. There was no difference in BMI or gender based upon airway use among study patients. The baseline PaCO2 value of patients who received an airway was similar to that of those who did not receive an airway (median [10th–90th quantile] = 56.0 [46.0–68.4] vs 54.1 [45.6–65.6], respectively; P = 0.33; 95% CI for location shift = −2.30 to 7.2). The PaCO2 at ABG-1 with and without airway use was also evaluated according to subsequent study group randomization. Among patients who were later randomized to the control group, the baseline PaCO2 value of those who received an airway was similar to the baseline PaCO2 of those who did not receive an airway (median [10th–90th quantile] = 56.4 [45.8–68.5] vs 56.1 [48.7–63.5], respectively; P = 0.68; 95% CI for the location shift = −6.2 to 9.8). Likewise, among patients who were later randomized to the nCPAP group, the baseline PaCO2 value of those who received an airway was similar to the PaCO2 among those who did not receive an airway (median [10th–90th quantile] = 56.0 [51.2–60.6] vs 53.2 [44.9–72.5], respectively; P = 0.21; 95% CI for the location shift = −2.8 to 9.3).
The principle aim of this study was to investigate the efficacy of nCPAP for reducing hypercarbia that accompanies deep sedation. During deep sedation of surgical patients, we observed a statistically significant reduction in PaCO2 among patients treated with nCPAP for 30 minutes versus control group patients, although the estimated difference ranged rather widely, from 1.4 to 12.6 mm Hg. For the purposes of this discussion, designations described in 1 published study may be used to categorize hypercarbia as mild, moderate, or severe based upon PaCO2 ranging from 42 to 49, 50 to 59, and ≥60 mm Hg, respectively.9 During MOAA/S level 2 sedation, moderate or severe hypercarbia was identified in 80% of control group patients and 75% of nCPAP patients when assessed by ABG-1. After a 30-minute period of continued sedation, ABG-2 demonstrated moderate or severe hypercarbia in 70% of control group patients, but in only 40% of patients treated with nCPAP during this interval. The hypothesis that nCPAP treatment would reduce hypercarbia associated with deep sedation was confirmed; however, further investigation is required to provide a more robust estimate of treatment effect.
There are limited published data that quantify a PaCO2 increase during deep IV sedation of adult patients. The data collected in this study may be a useful contribution in that regard. DeOliveira et al.10 reported a similar range of PaCO2 increase in a deeply sedated patient population with a lower BMI that was younger and all female. The clinical importance of moderate hypercarbia in this range is not clear. Deleterious effects of acute hypercarbia may include cerebral vasodilation with increase in intracranial pressure, stimulation of the sympathetic nervous system with tachycardia, and both peripheral and pulmonary vasoconstriction. Hypercarbia may provide a beneficial effect by shifting the oxyhemoglobin dissociation curve to the right, thereby promoting oxygen release in tissues. It has been postulated that this may improve healing and reduce wound infection.11–13 Underlying medical conditions may therefore influence any detrimental impact of hypercarbia. Our protocol excluded patients with conditions associated with greater tendency to develop hypercarbia during deep sedation, including chronic obstructive pulmonary disease, OSA, and BMI >40. Further studies may be useful to determine whether patients with these and other risk factors may develop increased hypercarbia during deep sedation and whether they may derive greater benefit from nCPAP treatment.
Comparison of ABG-1 and ABG-2 in the control group allowed us to observe whether PaCO2 changed after an additional 30-minute period of deep sedation. Rather than increasing further with ongoing deep sedation, PaCO2 was unchanged at ABG-2 as compared with ABG-1. This suggests a tendency for PaCO2 to plateau at a higher level than normal during a constant level of deep sedation.
In keeping with routine practice at the sponsoring facility, the study protocol incorporated a multimodal analgesia plan that included intrathecal morphine and FNB. A study by Abboud et al.14 found that ventilatory responses to CO2 showed no evidence of depression attributable to either 0.1 or 0.25 mg of intrathecal morphine. Our protocol also permitted the anesthesia team to place an oral or nasal airway when judged to be clinically indicated, but specific criteria for airway placement were not specified. An airway was placed in 40% of all study patients during deep sedation before ABG-1. The degree of hypercarbia associated with deep sedation at ABG-1 was similar across both groups irrespective of the use of an airway. It may be anticipated that an airway reduces upper airway obstruction in a manner similar to nCPAP and, therefore, placement would similarly reduce PaCO2. However, this was not observed. Studies have shown that nCPAP treatment not only splints the upper airway pneumatically to reduce obstruction but also reduces the work of breathing, and it may increase pulmonary functional residual capacity.15,16 We propose that these mechanisms likely contributed to the PaCO2 reduction observed with nCPAP treatment but not observed after airway placement. Six of the patients treated with an airway before ABG-1 were randomized to receive nCPAP treatment after ABG-1, and in all 6 patients, ABG-2 revealed a reduction in PaCO2 versus ABG-1. The number of such cases was insufficient to assess statistical significance. These issues may warrant further scientific investigation.
The STOPBANG questionnaire has been validated as a useful tool to screen for patients at risk for OSA and, potentially, at increased risk for airway obstruction during sedation. We used the questionnaire to assess and compare the likely tendency for upper airway obstruction between study groups and the groups had similar scores. One-quarter of our study population had a STOPBANG score of 5 or 6, equally distributed between study groups. Based upon the findings of Chung et al.,17 who compared STOPBANG score and subsequent diagnosis of OSA by portable polysomnography, it is likely that a significant proportion of the patients in our study had undiagnosed OSA that was moderately severe or severe.
There were several limitations to this study. First, propofol was titrated to achieve a sedation level at or near MOAA/S level 2 in all study patients. It is possible that drug titration using Bispectral Index monitoring may have allowed maintenance of a more standardized level of deep sedation. Second, our study protocol did not restrict airway placement for care of patients deemed to need it. This provision was a potentially confounding variable in the assessment of the effect of nCPAP on PaCO2 during deep sedation, but it provided some insight into airway effects on PaCO2 during deep sedation. The population size in this pilot study was insufficient to separately correlate variables such as BMI or STOPBANG score with PaCO2 during sedation in a statistically significant manner, and how these variables may influence the effectiveness of nCPAP. Finally, we selected a standardized, mid-range pressure setting for the nCPAP device used for this study. An automated CPAP device that optimizes airway pressure to patient needs may have provided superior clinical effects.
In summary, this pilot study demonstrated that deep sedation of TKA patients, in combination with FNB and SAB, resulted in moderate hypercarbia (mean and median PaCO2 = 55 mm Hg). The application of nCPAP in the nCPAP group reduced PaCO2 versus control group patients who received standard airway management, but the range of such reduction ranged widely, from 1.4 to 12.6 mm Hg. In the absence of nCPAP treatment, the baseline PaCO2 among deeply sedated patients who received an airway was not different from that of patients who did not receive an airway. Finally, among control group patients, the initial PaCO2 during deep sedation was similar to the PaCO2 when measured after an additional 30-minute period of deep sedation.
Name: Stephen B. Smith, MD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Stephen B. Smith has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Shawn Carr, CRNA.
Contribution: This author helped conduct the study.
Attestation: Shawn Carr approved the final manuscript.
Name: Stacey Psikula, CRNA.
Contribution: This author helped conduct the study.
Attestation: Stacey Psikula approved the final manuscript.
Name: Anita Das, PhD.
Contribution: This author helped analyze the data and write the manuscript.
Attestation: Anita Das has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Katherine Grichnik, MD, MS.
Contribution: This author helped analyze the data and write the manuscript.
Attestation: Katherine Grichnik has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
This manuscript was handled by: Terese T. Horlocker, MD.
We gratefully acknowledge Reese Clark, MD, Vice-President and Co-Director of The Center for Research, Education, and Quality, Pediatrix Medical Group, for his expertise in statistical analysis of our data and for his insightful comments during preparation of the manuscript.
1. Hillman DR, Platt PR, Eastwood PR. Anesthesia, sleep, and upper airway collapsibility. Anesthesiol Clin. 2010;28:443–55
2. Smith PL, Wise RA, Gold AR, Schwartz AR, Permutt S. Upper airway pressure-flow relationships in obstructive sleep apnea. J Appl Physiol (1985). 1988;64:789–95
3. Hillman DR, Platt PR, Eastwood PR. The upper airway during anaesthesia. Br J Anaesth. 2003;91:31–9
4. Eastwood PR. Evolution of changes in upper airway collapsibility during slow induction of anesthesia with propofol. Anesthesiology. 2009;111:63–71
5. Eastwood PR, Platt PR, Shepherd K, Maddison K, Hillman DR. Collapsibility of the upper airway at different concentrations of propofol anesthesia. Anesthesiology. 2005;103:470–7
6. Yamakage M, Kamada Y, Toriyabe M, Honma Y, Namiki A. Changes in respiratory pattern and arterial blood gases during sedation with propofol or midazolam in spinal anesthesia. J Clin Anesth. 1999;11:375–9
7. Vasu TS, Grewal R, Doghramji K. Obstructive sleep apnea syndrome and perioperative complications: a systematic review of the literature. J Clin Sleep Med. 2012;8:199–207
8. Chung F, Subramanyam R, Liao P, Sasaki E, Shapiro C, Sun Y. High STOP-Bang score indicates a high probability of obstructive sleep apnoea. Br J Anaesth. 2012;108:768–75
9. Lagow EE, Leeper BB, Jennings LW, Ramsay MA. Incidence and severity of respiratory insufficiency detected by transcutaneous carbon dioxide monitoring after cardiac surgery and intensive care unit discharge. Proc (Bayl Univ Med Cent). 2013;26:373–5
10. DeOliveira GS Jr, Ahmad S, Fitzgerald PC, McCarthy RJ. Detection of hypoventilation during deep sedation in patients undergoing ambulatory gynaecological hysteroscopy: a comparison between transcutaneous and nasal end-tidal carbon dioxide measurements. Br J Anaesth. 2010;104:774–8
11. Akça O, Doufas AG, Morioka N, Iscoe S, Fisher J, Sessler DI. Hypercapnia improves tissue oxygenation. Anesthesiology. 2002;97:801–6
12. Hager H, Reddy D, Mandadi G, Pulley D, Eagon JC, Sessler DI, Kurz A. Hypercapnia improves tissue oxygenation in morbidly obese surgical patients. Anesth Analg. 2006;103:677–81
13. Fleischmann E, Herbst F, Kugener A, Kabon B, Niedermayr M, Sessler DI, Kurz A. Mild hypercapnia increases subcutaneous and colonic oxygen tension in patients given 80% inspired oxygen during abdominal surgery. Anesthesiology. 2006;104:944–9
14. Abboud TK, Dror A, Mosaad P, Zhu J, Mantilla M, Swart F, Gangolly J, Silao P, Makar A, Moore J. Mini-dose intrathecal morphine for the relief of post-cesarean section pain: safety, efficacy, and ventilatory responses to carbon dioxide. Anesth Analg. 1988;67:137–43
15. Lindner KH, Lotz P, Ahnefeld FW. Continuous positive airway pressure effect on functional residual capacity, vital capacity and its subdivisions. Chest. 1987;92:66–70
16. Theerakittikul T, Ricaurte B, Aboussouan LS. Noninvasive positive pressure ventilation for stable outpatients: CPAP and beyond. Cleve Clin J Med. 2010;77:705–14
© 2015 International Anesthesia Research Society
17. Chung F, Subramanyam R, Liao P, Sasaki E, Shapiro C, Sun Y. High STOP-Bang score indicates a high probability of obstructive sleep apnoea. Br J Anaesth. 2012;108:768–75