CORRELATION OF END TIDAL CARBON DIOXIDE TO ARTERIAL PARTIAL PRESSURE CARBON DIOXIDE MEASURED BY SENTRI NASAL CANNULA

Tekspenuh

(1)of. M. al. LEE JIA WEN. ay a. CORRELATION OF END TIDAL CARBON DIOXIDE TO ARTERIAL PARTIAL PRESSURE CARBON DIOXIDE MEASURED BY SENTRI NASAL CANNULA. A STUDY IN SPONTANEOUSLY BREATHING NONINTUBATED PATIENTS. RESEARCH PROJECT. rs. ity. SUBMITTED TO THE. FACULTY OF MEDICINE. ve. DEPARTMENT OF ANAESTHESIA AND INTENSIVE CARE. U. ni. UNIVERSITY OF MALAYA, KUALA LUMPUR,. IN PARTIAL FULFILEMENTS FOR THE DEGREE OF MASTER OF ANAESTHESIOLOGY. 2018. UNIVERSITY OF MALAYA. i.

(2) ORIGINAL LITERARY WORK DECLARATION. ay a. Name of Candidate: Lee Jia Wen Registration/Matric No: MGE 150020 Name of Degree: Masters of Anaesthesiology Title of Project Paper/Research Report/Dissertation/Thesis: Correlation of end tidal carbon dioxide to arterial partial pressure carbon dioxide measured by Sentri nasal cannula. A study in spontaneously breathing non-intubated patients. ve. rs. ity. of. M. al. I do solemnly and sincerely declare that: I am the sole author/writer of this Work; This Work is original; Any use of any work in which copyright exists was done by way of fair dealing and for permitted purposes and any excerpt or extract from, or reference to or reproduction of any copyright work has been disclosed expressly and sufficiently and the title of the Work and its authorship have been acknowledged in this Work; I do not have any actual knowledge, nor do I ought reasonably to know that the making of this work constitutes an infringement of any copyright work; I hereby assign all and every right in the copyright to this Work to the University of Malaya (“UM”), who henceforth shall be owner of the copyright in this Work and that any reproduction or use in any form or by any means whatsoever is prohibited without the written consent of UM having been first had and obtained; I am fully aware that if in the course of making this Work, I have infringed any copyright whether intentionally or otherwise, I may be subject to legal action or any other action as may be determined by UM. Date:. U. ni. Candidate’s Signature. Subscribed and solemnly declared before, Witness’s Signature Name:. Date:. ii.

(3) CORRELATION OF END TIDAL CARBON DIOXIDE TO ARTERIAL PARTIAL PRESSURE CARBON DIOXIDE MEASURED BY SENTRI NASAL CANNULA. A STUDY IN SPONTANEOUSLY BREATHING NONINTUBATED PATIENTS. ay a. ABSTRACT Background: Continuous end tidal carbon dioxide (ETCO2) monitoring by capnography is essential during procedural sedation to detect adverse respiratory event. The aim of the study. al. was to determine accuracy of ETCO2 as a measure of arterial carbon dioxide( PaCO2) by using. M. Sentri nasal cannula in non-intubated patients and to assess correlation between ETCO2 and PaCO2 across different flow rates of oxygen among sedated and non-sedated patients.. of. Methods: A prospective, interventional study. 36 patients who underwent elective surgery. ity. in which arterial line monitoring was indicated were recruited. ETCO2 was sampled via Sentri nasal cannula at various oxygen flow rates (baseline, 2,4,6L/min) with and without sedation.. rs. Arterial blood gas and ETCO2 measurement were performed and recorded. The difference. ve. between PaCO2 and ETCO2 was tested by paired t test and correlated with Pearson correlation. Bland Altman analysis was used to assess the ETCO2 bias (defined as PaCO2 minus ETCO2).. ni. Results: A significant correlation was found between PaCO2 and ETCO2 irrespective of. U. oxygen flow among sedated and non-sedated patients(r=0.75,p<0.005).ETCO2 bias at baseline was 0.9±3.2mmHg(p=0.084) , showing good limit of agreement with PaCO2. There is increase in ETCO2 bias with increase oxygen flow rate in non-sedated patient (1.6±3.5mmHg vs 2.2±4.4mmHg vs 3.1±5.2mmHg) , but they were still within normal range. There is significant increase in ETCO2 bias in sedated patient (7.2±6.1mmHg at 2L/min vs 9.3±6.4mmHg at 4L/min, p<0.001). iii.

(4) Conclusion: ETCO2 measured via Sentri nasal cannula correlate well with PaCO2 among sedated and non-sedated patients irrespective of oxygen flow rate. It is useful in tracing the trend. U. ni. ve. rs. ity. of. M. al. ay a. and monitoring of ventilation status in spontaneously breathing patients with or without sedation.. iv.

(5) ACKNOWLEDGEMENTS Foremost, I would like to express my very great appreciation to Prof. Dato’ Dr. Wang Chew Yin for her patient guidance, enthusiastic encouragement, immense knowledge and useful critiques for this research work. Her guidance helped me in all the time of research and writing of. ay a. this thesis.. My sincere thanks to Dr. Chaw Sok Hui , for her valuable and constructive suggestion. al. during the planning and development of this research work. Thanks for her assistance in keeping. M. my progress on schedule. I would also like to thank Dr Fitry and Dr Lim Siu Min for their motivation and support during the whole course of my thesis. Without their professional guidance. of. and valuable support, I would not able to complete my thesis .. ity. I am particularly grateful for the assistance given by all my fellow anaesthesia colleagues and nurses in PPUM for giving me the opportunity to conduct my thesis, helping to prepare all the. rs. equipment I need and their words of encouragement.. ve. Finally, I must express my very profound gratitude to all my family members especially. ni. my husband for providing me with unfailing support and continuous encouragement throughout. U. my years of study and through the process of researching and writing this thesis. This accomplishment would not have made possible without them. Thank you.. v.

(6) TABLE OF CONTENTS. Abstract………………………………………………………………………………......iii Acknowledgements….……………………………………………………………………v. ay a. Table of Contents.………………………………………….…………………………….vi. List of Figures……………………………………………………………………………vii. M. al. List of Tables…………………………………………………..………..........................viii List of Appendices.………………………………………………………….……….......ix. of. CHAPTER 1: INTRODUCTION.……………………………………………………….1. ity. CHAPTER 2: LITERATURE REVIEW ……………………………………………......4. rs. CHAPTER 3: METHODS ……………...…………………………….…………..….....13. ve. CHAPTER 4: RESULTS ………………………………………………………...……..17 CHAPTER 5: DISCUSSION ………………………………………………….……......36. U. ni. CHAPTER 6: CONCLUSION ………………………………………………...……......41 References …………………………………………………………………...…….........42 Appendix …………………………………………………………………….….............45. vi.

(7) LIST OF FIGURES Figure 3.1: SentriTM ETCO2 adult nasal cannula with separate prongs for oxygen delivery and carbon dioxide sampling………………………………………………………………………………..16 Figure 4.1. Correlation between ETCO2 and PaCO2 under baseline…………………………….22. ay a. Figure 4.2. Correlation between ETCO2 and PaCO2 under 2L/min, without sedation………….22 Figure 4.3. Correlation between ETCO2 and PaCO2 under 4L/min, without sedation………….23. al. Figure 4.4. Correlation between ETCO2 and PaCO2 under 6L/min, without sedation………….24. M. Figure 4.5. Correlation between ETCO2 and PaCo2 at 2L/min, with sedation ………………….25 Figure 4.6. Correlation between ETCO2 and PaCo2 at 4L/min, with sedation ………………….26. of. Figure 4.7. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at baseline………….28. ity. Figure 4.8. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at 2L/min without. rs. sedation…………………………………………………………………………………………………..29 Figure 4.9. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at 4L/min without. ve. sedation…………………………………………………………………………………………………..30. ni. Figure 4.10. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at 6L/min without. U. sedation…………………………………………………………………………………………………..31 Figure 4.11. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at 2L/min with sedation…………………………………………………………………………………………………..32 Figure 4.12. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at 4L/min with sedation…………………………………………………………………………………………………..33. vii.

(8) LIST OF TABLES. Table 3.1: Observer Assessment of alertness/sedation scale (OAA/S).………….............16 Table 4.1: Sociodemographic characteristic of patients recruited………………………….18 Table 4.2: Comparison of ETCO2, PaCO2, PaO2 and PaCO2 values without sedation at different. ay a. nasal oxygen flow rate……………………………………………………………….……........19 Table 4.3: Comparison of ETCO2, PaCO2, PaO2 and PaCO2 values with sedation at different. al. nasal oxygen flow rate……………………………………………………………….……........20. M. Table 4.4: Correlation between PaCO2 and ETCO2 across different oxygen flow rates..21 Table 4.5: Associated factors contribute to difference between PaCO2 and ETCO2 under 2L/min. of. with sedation using Linear Regression Analysis……………………………………………..34. ity. Table 4.6: Associated factors contribute to difference between PaCO2 and ETCO2 under 4L/min. U. ni. ve. rs. with sedation using Linear Regression Analysis……………………………………………..35. viii.

(9) LIST OF APPENDICES. Appendix A: Data Collection Form……………………………….………….…...…..….…45 Appendix B: Patient Consent Form …………………………………….……..………..….48. ay a. Appendix C: Patient Information Sheet (English)………………………….……..............52 Appendix D: Patient Information Sheet (Bahasa Melayu)………………….…............…55. U. ni. ve. rs. ity. of. M. al. Appendix E: Ethics Committee Approval Letter…………………………………....…..…58. ix.

(10) CHAPTER 1 : INTRODUCTION End tidal carbon dioxide (ETCO2) is the partial pressure of carbon dioxide in the end of expiration, which is primarily alveolar gas. Normal ETCO2 range is between 35mmHg to 45mmHg. ETCO2 monitoring can be achieved by capnography. It is a noninvasive device used to. ay a. measure and display partial pressure of carbon dioxide (CO2) in the respiratory gas. It is expressed as a graphic record of concentration of expiratory CO2 plotted against time.. First infrared carbon dioxide measuring and recording apparatus was introduced by Luft. al. in 1943. The early use of CO2 monitoring is mainly for research purpose. It was very bulky,. M. costly and unreliable. Over decades, capnography has developed from a research instrument into a monitoring device. In 1988 , Caplan et al showed that there are 14 cases of unexpected cardiac. of. arrest in heathy patient who received spinal anaesthesia with intraoperative use of sedation. The. ity. study showed that undetected respiratory compromise as one of the causes of the sudden cardiac arrest (Caplan , 1988). Thereafter , the attention is focused on the use of capnogram on the. rs. awake, non-intubated patient to enhance patient safety.. ve. In the past , adequacy of ventilation was assessed by qualitative clinical signs such as. ni. chest excursion, observation of the reservoir bag and auscultation of breath sounds. However , these methods are usually inadequate because patients are usually covered with a surgical drape. U. ( Bennett, 1997). In line with the updated guidelines published by American Society of Anaesthesiologist in 2011, continuous end tidal carbon dioxide monitoring by capnography is one of the integral parts of basic anesthesia monitoring during moderate and deep sedation where the airway cannot be directly observed. The measurement of CO2 in expired air directly indicates changes in the elimination from the lungs. Indirectly, it indicates the change in production of CO2 at the tissue level and in the delivery of CO2 to the lungs by the circulatory. 1.

(11) system (Bhavani-Shankar K, 1996). Therefore , it can be used to assess adequacy of ventilation and provide information regarding the metabolic and circulatory condition of a patient. Over the last decades, the needs of capnography monitoring have gained its popularity due to substantial increased in number of procedures performed under sedation outside the. ay a. operation theatre. The use of capnography to assess adequacy of ventilation has extended in emergency department , endoscopy unit and catherization laboratories. Undetected. hypoventilation and apnea during such procedure can result in significant morbidity and. al. mortality. Capnography has been shown to be the more reliable predictor of respiratory. M. depression than pulse oximetry (Soto, 2004). Although pulse oximeter is one of the gold standards of care during conscious sedation , it has its limitation. Hypoventilation is usually. of. undetected until there is a decrease in oxygen saturation. Hypoxia is a late sign of hypoventilation especially when supplemental oxygen is administered for sedated patient (. ity. Deitch, 2007). Capnography is shown to be effective in early detection of adverse respiratory. rs. events. ASA closed claimed study also showed that capnography and pulse oximetry together. ve. could have helped in the prevention of 93% of avoidable anesthesia mishaps. Studies showed that ETCO2 correlate well with partial pressure of carbon dioxide, which. ni. is a gold standard indicator of adequacy of ventilation in mechanically ventilated patients.. U. (Saunders et al,1994). In 1989, Bowe et al showed that the ETCO2 may provide a good approximation of PaCO2 in spontaneously breathing patients ( Bowe, 1989). Since then, manufacturers produced variety of nasal cannula with their unique designs to sample end tidal carbon dioxide. However, studies showed that there is a considerable variation in correlation of ETCO2 to PaCO2 measured by nasal cannula in spontaneously breathing patient. These variations may be due to mechanical factors such as oxygen administration, catheter occlusion or. 2.

(12) kinking and ambient air leaks. Patient factors such as mouth breathing, nasal obstruction or hypoventilation will affect the ETCO2 measurement as well (Fakuda , 1997). Sentri nasal cannula is a device designed to sample exhaled ETCO2 in non-intubated patients during the administration of supplementary oxygen. By delivering oxygen through one. ay a. prong and sampling exhaled gas from the other prong, the nasal cannula can provide end tidal values comparable to those achieved with intubated patients. It is fitted to the patients as same manner as conventional nasal cannula and it consists of curved prong to improve the anatomical. al. fit and non-slip toggle to provide a secure fit on patient. We conduct this study to assess the. U. ni. ve. rs. ity. of. oxygen among sedated and non-sedated patients.. M. correlation of ETCO2 and PaCO2 measured by Sentri nasal cannula across different flow of. 3.

(13) CHAPTER 2: REVIEW OF LITERATURE Ibarra and Lees introduced first carbon dioxide sampling nasal cannula into clinical practice. They modified the nasal cannula by obstructing one of the nasal prongs with a terminal sampling line of mass spectrometer for monitoring of ETCO2 in spontaneously breathing patient. ay a. (Ibarra , 1985). Norman et al found that device proposed by Ibarra and Lees resulted in artifactually low ETCO2 due to incomplete obstruction of the nasal prong. They proposed. alternative method by placing the sampling line from mass spectrometer into nasal airway.. al. Patient trial was conducted, and they concluded that neither method is as reliable as monitoring. M. ETCO2 via endotracheal tube. Nonetheless, the nasal airway produced a more consistently satisfactory carbon dioxide curve compared to the partially obstructed nasal prong (Norman,. of. 1987). Huntington and king also reported obtaining satisfactory estimation of ETCO2 by inserting 14-gauge intravenous catheter into an oxygen mask to a point “close to, but not. ity. irritating” the patient’s nose. However, the accuracy of these devices has been questioned due to. rs. dilution of expired gas by the oxygen flow through the nasal cannula (Huntington, 1986).. ve. The study by Bowe et al was the first evaluation on quantitative measurement of ETCO2 during oxygen administration in spontaneously breathing patient. They studied the correlation of. ni. ETCO2 to PaCO2 by a self-constructed modification of simple nasal cannula in 21 preoperative. U. patients. A syringe cap was fully advance into one of the nasal prongs to achieve complete occlusion . The sampling line was then attached directly to the syringe cap. They compare the PaCo2- ETCO2 gradient between spontaneously breathing and mechanically ventilated patients. They found that there was no difference in the PaCO2-ETCO2 gradients measured by modified nasal cannula(2.09±2.18mmHg) and mechanically ventilated patients (2.87±2.82mmHg). They concluded that ETCO2 monitoring via nasal cannula may provide a good approximation of. 4.

(14) PaCO2 in non-intubated patients. They recommended the use of this modified nasal cannula in patients undergoing regional anaesthesia or cataract surgery (Bowe, 1989). McNulty et al examined the relationship of peak ETCO2 and time-average ETCO2 to PaCO2 measured by modified nasal cannula. 20 pre-medicated patients undergoing various. ay a. surgical procedure were recruited during preoperative period. A 16-gauge intravenous catheter was inserted through one of the nasal prongs and ETCO2 was sampled at various oxygen flow rate(2,4 and 6L/min). They found that peak ETCO2 (mean=38.8mmHg) correlated more closely. al. with PaCO2 (mean=38.8mmHg ) than did the average ETCO2 irrespective of different oxygen. M. flow rates. However, McNulty et al mentioned that it is more difficult to obtain the peak ETCO2 in clinical practice, especially when partial airway obstruction is present. The study also showed. ity. difference (P<0.001) (McNulty, 1990).. of. that increase in respiratory rate more than 30/min resulted in increase in the PaCO2 -ETCO2. Liu et al demonstrated the accuracy of capnography in non-intubated surgical patients by. rs. using a nasopharyngeal sampling catheter system. 25 patients admitted to the intensive care unit. ve. for perioperative care were studied. Only patient with pulmonary artery catheter for hemodynamic monitoring were recruited. Supplement oxygen is administered to 20 patients in. ni. this study , 8 by face mask and 12 by nasal cannula. They found a good correlation between. U. ETCO2 and PaCO2 (r=0.61, p<0.001) in spontaneously breathing patient without chronic respiratory disease. Overall bias was 3.6mmHg with a precision of 6.8. They also found a close correlation between dead space and PaCO2-ETCO2 gradient (r=0.77, p<0.001). However, venous admixture widened the PaCO2-ETCO2 gradient ( Liu, 1991). Lenz et al investigated the suitability of capnometry for continuous postoperative monitoring of ventilation in non-intubated, spontaneously breathing patients. 19 patients who are. 5.

(15) ASA II or III underwent either open heart surgery, gastrectomy or esophagectomy were recruited into this study. A special nasal catheter was placed 8cm to 10cm into the nasopharynx to sample respiratory gas during postoperative period. They found that ETCO2 correlate well with PaCO2 in spontaneously breathing patients during postoperative period(P<0.01), which is similar with the study done by Liu et al (1991). However, they also noted that there were several patients had. ay a. a large PaCO2-ETCO2 gradient due to ventilation perfusion mismatch (Lenz, 1991).. Takaki et al compared ETCO2 and PaCO2 in subjects extubated after abdominal surgery.. al. They also determine whether different respiratory pattern will affect the PaCO2-ETCO2. M. gradient. In this study, they noticed that there were wide limits of agreement between PaCO2 and ETCO2 irrespective of breathing pattern. They found that the difference between PaCo2 and. of. ETCO2 was significant smallest during deep breathing with mouth closed (7.7mmHg±5.6) than. ity. mouth closed (14.8mmHg±8.2). Therefore, ETCO2 may not a good prediction of PaCO2 after abdominal surgery due to lung mechanics changes. They also showed that the respiratory rate. rs. measured by capnometry during breathing with the mouth consciously closed were significantly. ve. correlated with the respiratory rate measured by direct measurement (P<0.001) (Takaki, 2016). In contrast , a prospective descriptive study by Yazigi et al on morbidly obese patients after. ni. bariatric surgery showed good correlation coefficients and acceptable limits of agreement. U. between PaCO2 and ETCO2. They concluded that nasal capnography provides a good estimation of PaCO2 in morbidly obese patients after surgery. They also found that nasal capnography was reliable in monitoring respiration and detecting all apnea episodes among morbidly obese patients.(Yazigi, 2007) Ebert et al studied 4 different nasal cannula designs to compare the effectiveness of oxygen delivery and reliability of carbon dioxide waveforms in increasing flow rates(2,4 and. 6.

(16) 6Lpm). The four nasal cannulas are bifurcated nasal prongs with oxygen delivery and CO2 sensing in both nasal prongs (Hudson), separate oxygen/ carbon dioxide nasal prongs(Salter) and CO2 sensing in nasal prongs with cloud oxygen delivery outside the nasal prongs via multi vents (Oridion) and dual vents (Medline). Only 11 out of 45 participants had an arterial line for the measurement of PaO2 and PaCO2. Elbert et al found out that separate nasal cannula was the. ay a. most effective in oxygen delivery and provide most reliable ETCO2 waveform at higher fresh gas flow compared with other nasal cannula designs. However, there was no statistically. al. association was found between ETCO2 and oxygen flow rate (P=0.410) for all nasal cannula.. M. On the other hand, there is no correlation between ETCO2 and PaCO2 for the bifurcated nasal cannula irrespective of oxygen flow. This result was consistent with the study conducted with. of. Friesen et al ( 1996). They showed that bifurcated nasal cannula appeared to be design flaw for monitoring of ETCO2 during oxygen administration due to dilution of respiratory gas (Ebert ,. ity. 2015).. rs. A retrospective medical record review analysis was performed by Tai et al among non-. ve. intubated neonates admitted to neonatal ward. The reliability of measurement of ETCO2 in reflecting PaCo2 by using nasal sampling cannula was assessed. 34 neonates were recruited and. ni. 54 pairs of PaCO2 and ETCO2 samples were compared. From the study , there was a good. U. correlation between ETCO2 and PaCO2 among all participants ( n=54, r=0.809, P<0.001) and those without respiratory disease (n=20, r=0.770, P<0.001). On the contrary, there is statistically significant difference between ETCO2 and PaCO2 in patients with respiratory disease (PetCo2=38.8±9.8mmHg vs PaCo2=41.2±10.3mmHg,n=34, p=0.027 ) compared to those without ( ETCO2=40.5±7.0mmHg vs PaCo2=41.6±7.2mmHg, n=20, p=0.289). However, ETCO2 still correlate well with the PaCO2 among patients with respiratory disease( n=34,. 7.

(17) r=0.823, p<0.001) . Tai et all concluded that ETCO2 measurement via nasal sampling cannula provide accurate and noninvasive estimation of PaCO2 in non-intubated neonates (Tai, 2010). The use of procedural sedation outside the operation theatre has increased in frequency and scope. Capnometry appeared to be more sensitive than clinical assessment or desaturation in. ay a. detecting respiratory depression. Wright assessed the value of capnography and pulse oximetry in heavily sedated patients in emergency department. This is a prospective, nonrandomized, noncontrolled clinical trial conducted on 27 patients requiring sedation (benzodiazepine±. al. narcotics) for painful procedures. Ventilatory status was assessed with capnography by nasal. M. sampling cannula and pulse oximetry before, during and 2 hours after administration of the sedative agent. Types of sedation given was based on treating physician’s choice. No oxygen. of. was administered during the period of study. He found out that the average ETCO2 increased 6.2% and oxygen saturation dropped by 3.7% during the procedures. One patient developed. ity. clinically significant apnea lasting for 30 seconds which detected by apnea alarms on. rs. capnometer. In addition, 8 patients developed hypoxemia with the oxygen saturation dropped. ve. below 90% with simultaneous increase in ETCO2 during or shortly after the procedure. Wright did not correlate the ETCO2 measured by nasal cannula with the PaCO2 in this study. However,. ni. he emphasized the clinical significance of capnography measured by nasal cannula in detection. U. of unrecognized respiration depression during conscious sedation (Wright, 1992). Another pilot study by Soto et al also studied the accuracy of capnography in detecting. apnea during monitored anaesthesia care. Besides that, they also studied effect of different oxygen flow rate (Baseline, 2,4,6L/min) on detection of apnea by capnography. 39 patients who were scheduled for procedure with conscious sedation were enrolled. Transthoracic impendence and Oridion nasal cannula which can sample both nasal and oral carbon dioxide were used to. 8.

(18) monitor apnea. 26 % of the participants (10 out of 39 patients) experienced 20 seconds of apnea during the procedure but none was detected by anaesthesia care provider. All apneic episodes were detected by both capnography and transthoracic impendence. They also found that higher oxygen flow decreases the amplitude of the capnograph but did not affect apnea detection (Soto,. ay a. 2014). A randomized, controlled study by Beitz et al assessed the use of capnography. monitoring during propofol sedation for colonoscopy. A nasal cannula with an oral sampling port. al. was used to sample ETCO2 continuously and provide 2L/min oxygen. A total of 760 patients. M. presented for colonoscopy were recruited. They demonstrated significant reduction in hypoxaemia for patient with additional capnography monitoring compared to standard. of. monitoring( 38.9% vs 53.2% ;p<0.001). They emphasized importance of capnographic monitoring to reduce incidence of oxygen desaturation during conscious sedation for. ity. colonoscopy (Beitz, 2011). The result was different from the randomized controlled trial. rs. conducted by Kim et al,2014. They found out that there was only 7.5% reduction in hypoxaemia. ve. by adding capnography during deep sedation with propofol by non-anaesthesiologist. Kim et al concluded that adding capnography into routine monitoring during conscious sedation does not. ni. necessarily improve patient safety in daily practice.. U. Accuracy of ETCO2 sampled by nasal cannula was questionable due to dilution from the. oxygen flow administrated. Friesen et al identified several factors contributing to the inaccuracy of ETCO2 monitoring with nasal cannula among pediatric patients. Linear regression analysis showed that cyanotic heart disease, airway obstruction, mouth breathing and administration of oxygen through ipsilateral nasal cannula resulted in greater PaCO2-ETCO2 gradient. On the contrary, congenital heart disease with left to right shunt, respiratory rate and oxygen. 9.

(19) administration through loose face mask or separate nasal cannula does not affect the accuracy of ETCO2 sampling. They concluded that ETCO2 measured by nasal cannula correlated well with PaCO2 in sedated, spontaneous breathing patients when those factors were excluded. (Friesen , 1996). ay a. A prospective semi-blind study by Yosefy et al was conducted to determine accuracy of sidestream ETCO2 as a predictor of PaCO2 in non-intubated patients in respiratory distress. Variables that may affect this correlation were also determined. 73 patients with respiratory. al. distress and needed arterial blood gas were included into this study. There are 55 patients. M. presented with pulmonary edema (75.3%), 14 patients with exacerbation of chronic obstructive lung disease(19.2%) and 4 patients with exacerbation of bronchial asthma(5.5%). There was a. of. good correlation between ETCO2 and PaCo2 in these patients( r=0.792). They found that only age under 50 years old and temperature above 37.6% affect PaCO2 -ETCO2 gradient. Otherwise,. ity. blood pressure, respiratory rate and blood pH did not affect the PaCO2-ETCO2 gradient. (. rs. Yosefy , 2004). ve. Fakuda et al identified that reliability of ETCO2 measurement through nasal cannula is depended on both biological and mechanical factors. They found that tidal volume more than. ni. 500mls and lower respiratory rate (RR<20bpm) resulted in more reliable ETCO2 measurement (. U. Fakuda,1997 and From and Scamman,1998) This was consistent with the study conducted by McNulty et al (1990) and From and Scamman(1998). Badgwell et al also reported that low tidal volume resulted in low ETCO2 values (Badgwell, 1987). Fakuda et al also found that the smaller diameter prongs, the shorter and smaller diameter cannula resulted in improved measurement of ETCO2 due to less dead space and turbulence within the cannula. In addition, Barton et al. 10.

(20) showed that there was no significant difference between ETCO2 and PaCO2 in patients with hypocapnia and acidosis (Barton, 1993). In summary, capnography monitoring on spontaneously breathing patient during conscious sedation is important to detect any respiratory adverse events. This has increased the. ay a. needs for accurate measurement of ETCO2 by respiratory devices. Numerous literatures were conducted to debate on correlation of ETCO2 and PACO2 measured by different types of nasal cannula, modified oxygen mask, nasal/oral discriminate sampling system and nasopharyngeal. al. catheter. Nasal cannula with separate prongs for oxygen administration and ETCO2 monitoring. M. showed good correlation in a study conducted by Ebert et al. Air from the room and oxygen administration may dilute the sampled gas. As a result, the accuracy of ETCO2 in estimation of. of. PaCO2 will be affected with different oxygen flow rates. Only a few studies evaluated the accuracy of ETCO2 measurement across different flow rates of oxygen in spontaneously. ity. breathing patient. In addition, there is no study on comparison of difference of ETCO2 and. rs. PaCO2 across different flow of oxygen rates among sedated and non-sedated patients.. ve. Our primary objective is to determine the accuracy of ETCO2 as a measure of PaCO2 by using Sentri nasal cannula in spontaneously breathing patient. The secondary objectives are to. ni. evaluate the accuracy of ETCO2 measurement across different flow rates of oxygen among non-. U. sedated and sedated patients and to determine variables that affect the ETCO2 and PaCO2 gradient.. 11.

(21) CHAPTER 3: METHODS Approval from the University of Malaya Medical Center’s Institutional Review Board (Reference number:2017925-5591), as well trial registration with the National Medical Research Register (Reference number: NMRR-17-2474-38815) and the ClinicalTrial.gov (Identifier: CT. ay a. 03374644) was obtained before embarking on the study. This prospective, interventional study was carried out over a period of 6 months from March 2018 till December 2018 at University Malaya Medical Centre.. al. Explanation and written informed consent were obtained from the 36 subjects recruited. M. into this study. Patients who were classified as American Society of Anesthesiologist (ASA) physical status 1 to 3, age more than 18 years’ old and scheduled for elective surgery in which. of. direct arterial blood pressure monitoring was indicated were included into this study. Patients. ity. who were less than 18 years old, have chronic lung disease, nasal obstruction, nasal congestion or craniofacial abnormalities and allergy to propofol were excluded from this study.. rs. SentriTM ETCO2 adult nasal cannula (Intersurgical ® code 1144002) were used to sample. ve. ETCO2 during preoperative and intraoperative period. ETCO2 was obtained from sidestream. ni. sampling from the nasal cannula and measured by infrared spectroscopy (Datex-Ohmeda GE Healthcare S5, Compact monitor, Switzerland). . During preoperative period, a 20-gauge radial. U. artery catheter was inserted under aseptic technique. It was used to obtain arterial blood gas measurement of partial pressure oxygen and carbon dioxide (PaO2, PaCO2). Blood gas was analyzed by using ABL 800 flex blood gas analyzer (Radiometer Medical ApS, Denmark) with calibrations recommended by the manufacturer.. 12.

(22) All the patients were not premedicated. Standard monitoring with continuous electrography (ECG), non-invasive blood pressure(NIBP) monitoring and pulse oximetry were attached to the patient. Local anaesthesia was given prior to arterial line insertion. Following the placement of indwelling radial artery catheter, Sentri nasal cannula was placed into the nostril of the patients. No sedation was given during preoperative period. Subjects were asked to breath. ay a. normally throughout the procedure. Nasal breathing was confirmed by presence of an acceptable CO2 tracing. After 5 minutes of stabilization of ETCO2 graphic waveform, blood was drawn. al. from indwelling arterial catheter for blood gas determination (PaO2 and PaCO2).. M. Simultaneously, value of ETCO2 were recorded by a sidestream capnometer. A baseline (without oxygen flow) ETCO2, PaCO2,PaO2, SPO2, respiratory rate, blood pressure and heart. of. rate were recorded. Subjects were not asked to keep their mouth close when using the nasal cannula as it might alter the variability inherent in patients who mouth breathe during conscious. ity. sedation. After a baseline ETCO2 was established, oxygen was administered at 2,4, and 6 liters. rs. per minute for a period of five minutes for each flow. At each level of oxygen administration, the highest ETC02 reading was recorded as it best represents a full tidal volume. PaO2 and PaCO2. ve. were also obtained from arterial blood sampling at the same time. Heart rate, respiratory rate and. ni. blood pressure were routinely recorded for each oxygen flow.. U. Titrated dose of target-controlled infusion (TCI) propofol was given following regional. anaesthesia during intraoperative period. Observer Assessment of alertness/sedation scale (OAA/S) was used to assess the level of sedation. The sedation was given with the target of OAA/S score of 3, which is responds only after name is called loudly and/ or repeatedly with slurring of speech, marked relaxation of facial expression and marked ptosis. The procedure was. 13.

(23) then repeated with 2 and 4 Lpm oxygen during conscious sedation. ETCO2, PaCO2, PaO2, SPO2, RR, blood pressure and heart rate were recorded during each level of oxygen flow. Data obtained from this study was analyzed using statistical package for social sciences (SPSS) version 24.0 to make inference and to draw robust conclusions. Bland Altman was. ay a. plotted using Medcalc software. In brief, a descriptive statistic of the socio-demographic characteristics was initially done to evaluate the distribution, normality and homogeneity of the data. Frequency and percentage were reported for distribution of categorical variables while. al. continuous variables were reported as Mean ± Standard Deviation (SD) The Paired t Test was. M. applied to determine the significant difference of matched sample of baseline between ETCO2 and PaCO2 and Pearson Correlation test was done to assess the correlation between two. of. continuous data i.e. between ETCO2 and PaCO2. Bland Altman plots were generated to assess the average discrepancy between methods (the bias) between the differences of ETCO2 and. ity. PACO2. Simple and Multiple Linear Regression was proceeding to predict the factors that. rs. contribute to difference between ETCO2 and PACO2. Any relationship or difference were. ve. considered significant if p-value showed less than 0.05. Sample size was calculated based on our pilot study of 5 patients. The mean PaCO2 is. ni. 35.2mmHg with standard deviation of 3.71 while the mean ETCO2 is 30.4mmHg with standard. U. deviation of 5.94.By considering alpha of 0.05 and power of 0.8, the required sample size required for each group is 15. We increased the number to 36 patients to account for drop out.. 14.

(24) Table 3.1. Observer Assessment of alertness/sedation scale (OAA/S) Speech Normal. Facial expression Normal. Mild slowing and thickening. Mild relaxation. 3. Responds only after name is called loudly, and/or repeatedly. Slurring or prominent slowing. Marked relaxation (slacked jaw). 2. Responds only mild prodding or shaking. Few recognizable words. 1. Does not respond to mild prodding or shaking. al. of. 4. Eyes Clear, no ptosis Glazed or mild ptosis ( less than half the eyes) Glazed and marked ptosis (half the eyes or more). ay a. Level of responsiveness Responds readily to name spoken in normal tone Lethargy responses to name spoken in normal tone. M. Score 5. ity. Figure 3.1. SentriTM ETCO2 adult nasal cannula with separate prongs for oxygen delivery. U. ni. ve. rs. and carbon dioxide sampling. 15.

(25) CHAPTER 4: RESULTS A total of 36 patients who fulfilled the inclusion criteria were recruited, 20 men and 16 females. Their age ranges from 61 to 92 years old with the mean of age was 73.9 ± 7.6. Most of the studied population were Chinese (52.8%). Their average height was 157 ± 18.4 cm and the. ay a. average weight was 65.3 ± 20.4 kg with the average BMI of 24.4 ± 4.8. 63.9% of the patients were classified as ASA 2 while 36.1% were ASA 3 patients. It is found that only minority of the patients recruited were smokers ( 8.3%) compared to nonsmokers ( 92%). 30.6% of the patients. al. have underlying of heart disease which consists of ischemic heart disease and congestive heart. U. ni. ve. rs. ity. of. them was diagnosed with bronchial asthma.. M. failure. Only minority of these patients ( 11.1%)have underlying of lung disease, which all of. 16.

(26) Table 4.1. Sociodemographic characteristics of patients recruited (n=36) Mean ± Standard deviation 73.9 ± 7.6. Frequency (%). 20 (55.6) 16 (44.4) 8 (22.2) 19 (52.8) 9 (25.0). ay a. 65.3 ± 20.4 157.0 ± 18.4 24.4 ± 4.8. al. 0 23 (63.9) 13 (36.1). rs. ity. of. M. 33 (91.7) 3 (8.3) 20 (55.6) 16 (44.4) 7 (19.4) 29 (80.6) 31 (86.1) 5 (13.9) 25 (69.4) 11 (30.6) 32 (88.9) 4 (11.1). U. ni. ve. Variable Age Gender Male Female Ethnicity Malay Chinese Indian Weight Height BMI ASA I II III Smoking No Yes DM No Yes Hypertension No Yes CKD No Yes Heart disease No Yes Lung disease No Yes. 17.

(27) Table 4.2 showed data collected at different oxygen flow rates during preoperative period. No sedation was given during that period. The mean ETCO2 at baseline( without oxygen flow) was 35.2±4.7 mmHg with the mean PaCO2 of 36.2±4.2mmHg. The ETCO2 bias, defined as PaCO2 minus ETCO2 was only 0.9±3.2mmHg at baseline. ( p=0.082). There was a significant and consistent decrease in ETCO2 with increasing oxygen flow rate. The mean ETCO2 at. ay a. 2L/min, 4L/min and 6L/min was 34.4±4.3mmHg , 33.5±4.9mmHg and 33.5±4.9 mmHg. respectively. In comparing the mean ETCO2 bias at each level of oxygen administration, there. al. was significant increase in the difference of PaCO2 and ETCO2 . The differences between. M. PaCO2 and ETCO2 appears to be increasing to 1.6±3.5mmHg , 2.2±4.4mmHg and 3.1±5.2mmHg respectively when oxygen flow rate increase from 2L/min to 6L/min (P<0.05). In. of. addition, PaO2 value was increase as well with increasing oxygen flow rate. Mean PaO2 at baseline was 89.6±16.3mmHg, at 2L/min was 126.4±32.9mmHg, at 4L/min was 155±42.3and at. rs. ity. 6L/min was 179.7±51.3mmHg.. ve. Table 4.2. Comparison of ETCO2, PaCO2, PaO2 and PaCO2 (bias ± precision) values( Mean ± SD) without sedation at different nasal oxygen flow rates. 4L/Min 16.7±7.7 33.5±4.9 35.7±4.6 155±42.3. 6L/Min 17.0± 32.72±5.6 35.8±4.3 179.7±51.3. 0.9±3.2. 1.6±3.5. 2.2±4.4. 3.1±5.2. ni. Baseline 16.2±4.7 35.2±4.7 36.2±4.2 89.6±16.3. Flow Rate 2L/min 16.8±4.1 34.4±4.3 36.0±4.0 126.4±32.9. U. Variable Respiratory Rate ETCO2 PaCO2 PaO2 PaCO2-ETCO2 ( bias±precison). 18.

(28) When analyzing the mean ETCO2 measurement with sedation, the mean ETCO2 at 2L/min and 4L/min was found to be significantly lower compared to mean ETCO2 measured without sedation. The mean ETCO2 at 2L/min with sedation was 30.7±8.0mmHg while mean ETCO2 at 4L/min was 29.1±7.0. There is significant increase in ETCO2 bias with increase in oxygen flow rates when sedation was given. Mean ETCO2 bias at 2L/min was 7.2±6.1mmHg (. ay a. with sedation) compared to 1.6±3.5mmHg (without sedation), with the p value less than 0.001. With sedation, the difference between PaCO2 and ETCO2 increase significantly at 4L/min,. al. which is 9.3±6.4mmHg compared to 2.2±4.4mmHg ( without sedation).(P<0.001). In addition,. M. there is slight increase in PaCO2 when sedation was given, which is 37.8±4.5mmHg at 2L/min. of. and 38.4±4.5mmHg at 4L/min.. Table 4.3 . Comparison of ETCO2, PaCO2, PaO2 and PaCO2 (bias ± precision) values(. Flow Rate 2L/min 15.2±3.1 30.7±8.0 37.8±4.5 136.6±44.7. rs. 4L/Min 14.1±3.5 29.1±7.0 38.4±4.5 172.4±56.4. 7.2±6.1. 9.3±6.4. U. ni. ve. Variable Respiratory Rate ETCO2 PaCO2 PaO2 PaCO2-ETCO2 (bias ± precision). ity. Mean ± SD) with sedation at different nasal oxygen flow rates. 19.

(29) The Pearson Correlation test was done to assess the correlation between PaCO2 and ETCO2 across different flow with or without sedation . There is a positive linear relationship between ETCO2 and PaCO2 with or without sedation irrespective of the oxygen flow rates. There is a strong correlation between ETCO2 and PaCO2 under baseline ( without oxygen flow), with the correlation coefficient of 0.75(p<0.001). As flow of oxygen increase, the correlation. ay a. coefficient between PaCO2 and ETCO2 decreases but remain strong(P<0.005). The correlation coefficient at 2L/min , 4L/min and 6Lmin was 0.64, 0.57 and 0.47 respectively. With sedation,. al. the correlation between PaCO2 and ETCO2 remain positive despite different oxygen flow rates.. M. The Pearson correlation coefficient at 2L/min and 4L/min with sedation were 0.65 and 0.45 respectively (P<0.005).. rs. ity. Pearson correlation, r 0.75 0.64 0.57 0.47 0.65 0.45. P-value <0.001 <0.001 <0.001 0.004 0.004 0.005. U. ni. ve. Oxygen flow rate( L/Min) Baseline 2 4 6 2(with sedation) 4(with sedation ). of. Table 4.4. Correlation between PaCO2 and ETCO2 across different oxygen flow rates.. 20.

(30) of. M. al. ay a. Figure 4.1. Correlation between ETCO2 and PaCO2 under baseline. U. ni. ve. rs. ity. Figure 4.2. Correlation between ETCO2 and PaCO2 under 2L/min, without sedation. 21.

(31) U. ni. ve. rs. ity. of. M. al. ay a. Figure 4.3. Correlation between ETCO2 and PaCO2 under 4L/min , without sedation. 22.

(32) U. ni. ve. rs. ity. of. M. al. ay a. Figure 4.4. Correlation between ETCO2 and PaCO2 under 6L/min ,without sedation. 23.

(33) U. ni. ve. rs. ity. of. M. al. ay a. Figure 4.5. Correlation between ETCO2 and PaCo2 at 2L/min ,with sedation. 24.

(34) U. ni. ve. rs. ity. of. M. al. ay a. Figure 4.6. Correlation between ETCO2 and PaCo2 at 4L/min, with sedation. 25.

(35) The bland Altman plot of the difference between ETCO2 bias and the average of PaCO2 and ETCO2 were generated to assess the level of agreement between the ETCO2 and PaCO2 across different flow of oxygen (Figure 4.7-4.12). Limit of agreement between ETCO2 and PaCO2 at baseline are clinically acceptable for a wide range of measured values. (mean= 0.9417, 95% CI = -0.1271 to 2.0104, p=0.08) However, Bland Altman plot showed that there is. ay a. significant difference between ETCO2 and PaCO2 at 2l/min, 4L/min and 6L/min (P<0.05). In addition, there is a very wide limit of agreement between ETCO2 and PaCO2 at 2L/min with. al. sedation (mean=7.1417, 95% CI =5.06 to 9.22, upper limit=19.2, lower limit=-4.91, P<0.001).. M. The limit of agreement between ETCO2 and PaCO2 was the widest at 4L/min with sedation. (mean=9.26, 95%CI =7.10 to11.4, upper limit =21.8, lower limit=-3.25) with the p value less. U. ni. ve. rs. ity. of. than 0.001.. 26.

(36) Figure 4.7. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at baseline. 10 8. +1.96 SD. 6. 7.1. ay a. 4 2. Mean 0.9. al. 0. M. -2 -4. -8 25. 35. -5.2. 40. 45. ity. 30. of. -6. -1.96 SD. ve. rs. Mean of PB_PACO2 and PB_ETCO2. U. ni. Method A PB_ETCO2 Method B PB_PACO2 Differences Sample size Arithmetic mean 95% CI P (H0: Mean=0) upper limit 95% CI Lower limit 95% CI Coefficient of Repeatability 95% CI. 36 0.9417 -0.1271 to 2.0104 0.0823 7.1325 5.2890 to 8.9761 -5.2492 -7.0928 to -3.4056 -6.3772 -8.2837 to -5.1860. 27.

(37) Figure 4.8. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at 2L/min without sedation. 10. +1.96 SD. 8. ay a. 8.5. 6 4. al. 2. 1.6. M. 0. Mean. -2. of. -4. ity. -6 -8 25. 35. -5.3. 40. 45. rs. 30. -1.96 SD. ni. ve. Mean of P2PaCO2 and P2_ETCO2. U. Method A P2_ETCO2 Method B P2PaCO2 Differences Sample size 36 Arithmetic mean 1.6194 95% CI 0.4319 to 2.8070 P (H0: Mean=0) 0.0089 Upper limit 8.4985 95% CI 6.4500 to 10.5470 Lower limit -5.2596 95% CI -7.3081 to -3.2111. 28.

(38) Figure 4.9. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at 4L/min without sedation. 15 +1.96 SD 10. ay a. 10.9. 5. al. Mean 2.2. M. 0. of. -5. 20. ity. -10. 25. 30. 35. 40. -1.96 SD -6.5. 45. 50. ni. ve. rs. Mean of P4PaCO2 and P4ETCO2. U. Method A P4ETCO2 Method B P4PaCO2 Differences Sample size 36 Arithmetic mean 2.1944 95% CI 0.6941 to 3.6948 P (H0: Mean=0) 0.0054 Upper limit 10.8855 95% CI 8.2974 to 13.4736 Lower limit -6.4966 95% CI -9.0847 to -3.9085. 29.

(39) Figure 4.10. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at 6L/min without sedation. 15. +1.96 SD 13.3. ay a. 10. 5. al. Mean. M. 0. 3.1. ity. of. -5. -10 25. 35. -7.1 40. 45. rs. 30. -1.96 SD. ve. Mean of P6PaCO2 and P6ETCO2. U. ni. Method A P6ETCO2 Method B P6PaCO2 Differences Sample size 36 Arithmetic mean 3.1083 95% CI 1.3408 to 4.8758 P (H0: Mean=0) 0.0011 Upper limit 13.3471 95% CI 10.2981 to 16.3961 Lower limit -7.1304 95% CI -10.1794 to -4.0815. 30.

(40) Figure 4.11. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at 2L/min with sedation. 25 +1.96 SD. 20. ay a. 19.2 15 10. al. Mean 7.1. M. 5. -1.96 SD. of. 0. -10 15. 20. ity. -5. 25. 30. 35. -4.9. 40. 45. 50. ve. rs. Mean of I2PaCo2 and I2ETCO2. U. ni. Method A I2ETCO2 Method B I2PaCo2 Differences Sample size 36 Arithmetic mean 7.1417 95% CI 5.0618 to 9.2215 P (H0: Mean=0) <0.0001 Upper limit 19.1897 95% CI -15.6020 to -22.7775 lower limit -4.9064 95% CI -8.4942 to -1.3187. 31.

(41) Figure 4.12. Bland-Altman Plot for accuracy between PaCO2 and ETCO2 at 4L/min with sedation. 25 +1.96 SD 21.8. ay a. 20. 15. Mean. al. 10. M. 5. 9.3. -5 15. 20. ity. of. 0. 25. 30. 35. -1.96 SD -3.2 40. 45. 50. ve. rs. Mean of I4PaCO2 and I4ETCO2. U. ni. Method A I4ETCO2 Method B I4PaCO2 Differences Sample size 36 Arithmetic mean 9.2583 95% CI 7.0997 to 11.4169 P (H0: Mean=0) <0.0001 Upper limit 21.7626 95% CI -18.0390 to -25.4862 Lower limit -3.2459 95% CI -6.9695 to 0.4777. 32.

(42) Simple Linear Regression Analysis (univariable analysis) and Multiple Linear Regression Analysis (multivariable analysis) were performed to identify the associated factors that contribute to the difference between PACO2 and ETCO2 under 2L/min and 4L/min with sedation. The Simple Linear Regression analysis is a univariable analysis that gives a preliminary idea which variables were identified as the potential significant associated factor. From simple linear regression. ay a. analysis, respiratory rate and lung disease were found to be significant with p value less than 0.05. There is a significant linear relationship between respiratory rate and intraoperative difference. al. between PACO2 and ETCO2 under 2L/min (p=0.036). Besides that, there is also significant linear. M. relationship between lung disease and difference between PACO2 and ETCO2 under 2L/min with sedation (p=0.019). The potential variables were included in Multiple Linear Regression analysis.. of. However, none of the variables were found to be significant in Multiple Linear Regression. ity. analysis.. Simple Linear Regression. ve. Variables. rs. Table 4.5. Associated factors that contribute to the difference between PACO2 and ETCO2 under 2L/min with sedation using Linear Regression Analysis. ba. 0.036. BMI -0.09 -0.54,0.35 Lung disease No 0 Yes 7.52 1.33,13.71 Heart disease No 0 Yes -0.86 -5.43,3.72. 0.681. U. 0.69. a. p-value Adj bb 95% CI. 0.05,1.33. ni. RR. 95% CI. Multiple Linear Regression. 0.019. 0.52. p-value. -1.16,1.16. 0.105. 0 6.17 -0.10,12.42. 0.053. 0.705. regression coefficient adjusted regression coefficient. b. 33.

(43) From Simple Linear Regression analysis on associated factors that contribute to ETCO2 bias under 4l/min under sedation, BMI and lung disease were found to be potential significant factor (p<0.25). The potential variables were included in Multiple Linear Regression analysis. However, none of the variables were found to be significant in Multiple Linear Regression. ay a. analysis.. Table 4.6. Associated factors that contribute to the difference between PACO2 and ETCO2 under 4L/min with sedation using Linear Regression Analysis. a. p-value. -0.06. -0.69,0.57. 0.841. 0.43. 0.00,0.87. 0.051. 0.41. -0.02,0.84. 0.064. 0 5.19 -1.54,11.93. 0.126. 0 4.66 -1.86,11.17. 0.156. 0 -2.41,6.98,. 0.329. 2.29. rs. BMI Lung disease No Yes Heart disease No Yes. p-value Adj bb 95% CI. ity. RR. 95% CI. M. ba. Multiple Linear Regression. al. Simple Linear Regression. of. Variables. regression coefficient adjusted regression coefficient. U. ni. ve. b. 34.

(44) CHAPTER 5: DISCUSSION. The use of capnography in mechanically ventilated patient has become an integral part of ventilatory monitoring in anaesthesia. It is a noninvasive device which provide clinical useful information on metabolic, cardiovascular and respiratory systems. American Society of. ay a. Anesthesiologist (ASA) emphasized the importance of continuous ETCO2 monitoring during moderate or deep sedation to assess for adequacy of ventilation in 2011. Various carbon dioxide sampling device were manufactured to sample end tidal carbon dioxide among spontaneous. al. breathing patients. However, studies showed wide variation in reliability (Ebert, 2015). In. M. addition, the accuracy of ETCO2 monitoring maybe affected in the sedated patients as well.. of. In present study, we demonstrated that ETCO2 measured by Sentri nasal cannula do reflect the PaCO2 level in spontaneously breathing patients. In healthy individual, the difference. ity. between PaCO2 and ETCO2 is 2-5mmHg under normal physiologic condition (Wahba,. rs. 1996;Russel,1990). Bland Altman plot revealed that there is good level of agreement between PaCO2 and ETCO2 at baseline, with the mean ETCO2 bias of 0.9±3.2mmHg (p=0.082).. ve. However, there is increase in the difference in PaCO2 -ETCO2 gradient when oxygen flow. ni. increases. Although it is statistically significant, the PaCO2-ETCO2 gradient across different oxygen flow rates in non-sedated patients are still within the clinical acceptable range. The. U. ETCO2 bias was 1.6±3.5mmHg, 2.2±4.4mmHg and 3.1±5.2mmHg respectively with the oxygen flow rate increase from 2L/min to 6L/min. The mean PaCO2-ETCO2 difference measured by Sentri nasal cannula showed favorable result compare to the mean PaCO2-ETCO2 difference seen in intubated patient.(Raemer, 1983). 35.

(45) Previous studies showed that there was satisfactory correlation between ETCO2 sampled through nasal cannula with PaCO2 in spontaneously breathing adults as well as pediatric patients (Bowe, 1989; Tobias, 1994;McNulty,1990). In our study, we demonstrated that nasal cannula with separate nasal prongs for oxygen delivery and carbon dioxide sampling provide accurate estimation of PaCO2 value across different flow of oxygen. There is no dilution. ay a. of expired gas occur during oxygen administration when nasal cannula with separate nasal. prongs is used. It provides end tidal values which is comparable to those achieved with intubated. al. patients. This is consistent with the study done by Ebert et al. They examined different nasal. M. cannula design for the accuracy of ETCO2 measurement across different oxygen flow rate. They found that nasal cannula with separate nasal prongs was the most reliable device to sample. of. expired carbon dioxide. On the contrary, bifurcated nasal prong design which deliver oxygen and sample carbon dioxide on the same nasal prongs appeared to be most unreliable device due to. ity. fresh gas flow mixing with the exhaled CO2 (Ebert, 2015).. rs. It is postulated that accuracy of ETCO2 measurement will be affected when higher. ve. flow of oxygen is used. This is due to contamination of the sampled carbon dioxide by turbulence and obstruction of flow when oxygen flow is increased. Roy et al evaluated accuracy. ni. of ETCO2 measurement, sampled through the modified nasal cannula across different flow of. U. oxygen in sedated patients. They found that ETCO2 correlate well with PaCO2 with oxygen flow rate less than 4L/min, but were significant different at 6L/min. They also found that there was linear relationship between ETCO2 bias with oxygen flow rate.(Roy, 1991) In contrast, McNulty et al showed a good correlation of PaCO2 and ETCO2 irrespective of oxygen flow rate (McNulty,1990). We found that there is significant and consistent decrease in ETCO2 measurement with increase oxygen flow rate in our study. Still, ETCO2 correlate well with. 36.

(46) PaCO2 across different flow of oxygen among sedated and non-sedated patients, which is same as the study conducted by McNulty et al. Airway obstruction and respiratory depression are the most common serious adverse events associated with sedation administration (Tobias,1994; Tai ,2010). Hypoventilation is. ay a. characterized by increase in the PaCO2 and decrease in respiratory rate. Pulse oximetry is not a reliable device in detecting apnea or airway obstruction as hypoxaemia is a late sign of. respiratory depression (Bennett, 1997). Studies showed that many respiratory adverse events. al. were not detected by anaesthesia provider but were detected by the capnography instead (Soto. M. ,2004; Wright ,1992). Capnography provides earlier warning sign of impending airway compromise compare to pulse oximeter (Beitz, 2011;Soto ,2004; Hart, 1997 ). Therefore, early. of. intervention can be done to prevent further morbidity.( Hart, 1997). ity. In our study, there is statistically significant increase in difference in PaCO2-ETCO2 gradient among sedated patient. ETCO2 bias was 7.2±6.1mmHg at 2L/min and 9.3±6.4mmHg at. rs. 4L/min, which showed wide limit of agreement. Hypoventilation and airway obstruction can lead. ve. to falsely low ETCO2 readings due to entrainment of room air (Bowe, 1989). In addition, mouth breathing may also occur during sedation which contribute to increase in PaCO2-ETCO2. ni. gradient (Tobias, 1994;Bowe,1989). The reason for this finding is due to variable fractions of. U. tidal volume were exhaled through the mouth, leading to insufficient amount for sampling through nasal cannula. In our study, we noted capnogram was not detected in 2 patients who are predominantly mouth breather. The capnogram can be observed when they are asked to breath with mouths closed. Although it is difficult to predict PaCO2 value in sedated patient, our data showed there is strong positive linear relationship between PaCO2 and ETCO2 among sedated patient (r=0.65 at2L/min vs r=0.45 at 4L/min, p<0.005). As a result, sidestream capnography can. 37.

(47) serve as an effective adjunct method for tracing the trends and monitoring ventilatory status for any abnormal CO2 value among patient who received sedation. Several factors can affect the accuracy of ETCO2 measurement. One of the major determinants is the increase in dead space which leads to ventilation perfusion mismatch. Dead. ay a. space is defined as unit of lung that do not participant in gas exchange. Exhaled gas from these areas do not contain carbon dioxide and will dilute the gas from areas with normal ventilation perfusion ratio. As a result, the difference between PaCO2 and ETCO2 will increase. Liu et al. al. demonstrated that there is closer relation between dead space and PaCO2-ETCO2 gradient. M. compared with venous admixture (Liu, 1992). McSwain et al also showed that there is strong positive linear relationship between PaCO2-ETCO2 gradient and physiological dead space. of. (McSwain,2010). In our study, the increase in PaCO2-ETCO2 gradient is most likely due to increase in the dead space, which can be seen in elderly patients. Most of our study population. ity. are elderly patients, with the mean age of 73. It is hypothesized by McNulty et al. that there is. rs. increase in anatomical and alveolar dead space in elderly patients due to the breakdown of. ve. alveolar septae (McNulty,1990). In addition, the closing capacity is also more than functional residual capacity in elderly patient. Therefore, there is increase in V/Q mismatch and result in. ni. increase in the difference between PaCO2 and ETCO2.. U. Another factor could have influence the PaCO2-ETCO2 gradient is presence of. pulmonary disease. There is ventilation perfusion mismatch in pulmonary disease which can contribute to lower ETCO2 readings. We found that presence of pulmonary disease is one of the factors result in ETCO2 measurement error. This finding is consistent with the study conducted by Tai et al(2010). They demonstrated that ETCO2 bias is higher in patient with respiratory disease, but ETCO2 still correlate well with PaCO2. We also realized that increase in respiratory. 38.

(48) rate result in increased the PaCO2-ETCO2 gradient. It is postulated that proportion of dead space will increase when respiratory rate increase (Fakuda,1997;Barton,1991). Besides, that, Yazigi et al revealed that nasal capnography provides a good estimation of PaCO2 in obese patients after surgery (Yazigi,2007). The finding is different from our study, showing that there is increase in ETCO2 bias when body mass index increase (P>0.05). We hypothesized that closing capacity is. ay a. more than functional residual capacity in obese patient, leading to increase in pulmonary. shunting. We also observed negative PaCO2-ETCO2 difference bias in some of our patients,. al. which most commonly encounter in patient with acute hypercapnia and rebreathing of carbon. M. dioxide from the under-ventilated alveolar (McNulty, 1990). Presence of water vapor in the sampling line may also overestimate the ETCO2 value due to interference of water vapor in. U. ni. ve. rs. ity. of. measurement of carbon dioxide by infrared analyzer.. 39.

(49) CHAPTER 6: CONCLUSION. Our study showed that ETCO2 correlate well with PaCO2 with or without sedation irrespective of oxygen flow rate. As oxygen flow rate increase, the trend between ETCO2 and PaCO2 remain reliable in most patient. Although there is significant increase in the. ay a. difference between ETCO2 and PaCO2 for patients with sedation, the trend in mean PaCO2ETCO2 difference is still comparable to the trend in mean PaCO2-ETCO2 difference seen in intubated patient. Hypoventilation, mouth breathing, increase in respiratory rate and presence of. al. pulmonary disease may affect the accuracy of ETCO2 readings. We conclude that continuous. M. ETCO2 monitoring via Sentri nasal cannula is useful for tracing the trends and monitoring for. U. ni. ve. rs. ity. of. ventilatory status among the spontaneously breathing patients with or without sedation.. 40.

(50) REFERENCES. Barton CW, Wang ES. (1993). Correlation of end-tidal CO2 measurements to arterial PaCO2 in nonintubated patients. Annals of Emergency Medicine, 23(3),560-563. Badgwell JM, McLeod ME, Lerman J, Creighton RE. (1987). End-tidal PCO2 measurements sampled at the distal and proximal ends of endotracheal tube in infants and children. Anesth Analg,66,959-964.. ay a. Beitz A, Riphaus A, Meining A, Kroshage T, Geist C, Wegenpfeil S, Weber A, Junh A, …Delius SV.(2012). Capnographic monitoring reduces the incidence of arterial oxygen desaturation and hypoxemia during propofol sedation for colonoscopy: a randomized, controlled study ( ColoCap study ) . Am J Gastroenterol,107,1205-1212, doi:10.1038/ajg.2012.136.. al. Bennett J, Petersen T, Burleson JA.(1997).Capnography and ventilatory assessment during ambulatory dentoalveolar surgery. J Oral Maxillofac Surg,55, 921-925.. M. Bhavani-Shankar K, Moseley H, Kumar AY, Delph Y.(1996). Review article: capnometry and anesthesia. Can J Anesth ,39(6),342-347.. of. Bowe EA, Boysen PG, Broome JA, Klein EF Jr .(1989). Accurate determination of end-tidal carbon dioxide during administration of oxygen by nasal cannulae. J Clin Monit, 5,105110.. ity. Caplan RA, Ward RJ, Posner K, Cheney FW .(1988). Unexpected cardiac arrest during spinal anaesthesia: a closed claims analysis of predisposing factors. Anesthesiology,68 (1),5-11.. rs. Deitch K, Chudnofsky CR, Dominicl P. ( 2007). The utility of supplemental oxygen during emergency department procedural sedation and analgesia with midazolam and fentanyl: a randomized, controlled trial. Ann Emerg Med, 49(1)18.doi:10.1016/j.annemergmed.2006.06.013.. ni. ve. Ebert TJ, Novalija J, Uhrich TD, Barney JA.( 2015). The effectiveness of oxygen delivery and reliability of carbon dioxide waveform: a crossover comparison of 4 nasal cannulae. Anesth Analg, 120(2) , 342-348.. U. Fakuda K, Ichinohe T, Kaneko Y. (1997). Is measurement of end-tidal CO2 through a nasal cannula reliable? Anesth Prog,44,23-26. Friesen RH, Alswang M.(1996). End-tidal PCO2 monitoring via nasal cannula in pediatric patients: accuracy and sources of error. J Clin Monit, 12,155-159. Hart LS, Berns SD, Houck CS, Boenning DA.(1997). The value of end-tidal CO2 monitoring when comparing three methods of conscious sedation for children undergoing painful procedures in the emergency department. Pediatric emergency Care, 13(3),189-193. Huntington CT, King HK. A simpler design for mass spectrometer monitoring of the awake patients. Anesthesiology, 65,565-566.. 41.

(51) Ibarra E, Lee De. (1985). Mass spectrometer monitoring of patients with regional anaesthesia. Anesthesiology, 63, 572-573. Kim VL, Leyssius AT, Zaane BV, Denteneer M, Kalkman CJ. (2014).Capnography during deep sedation with propofol by nonanesthesiologist: a randomized controlled trial. Anesth Analg,119,49-55. Lenz G, Heipertz W, Epple E.(1991). Capnometry for continuous postoperative monitoring of nonintubated, spontaneously breathing patients. J Clin Monit,7,245-248.. ay a. Liu SY, Lee TS, Bongard F. (1992). Accuracy of capnography in nonintubated surgical patients. Chest,102,1512-1515. McNulty SE, Roy J, TorjmanM.(1990). Relationship between arterial carbon dioxide and end tidal carbon dioxide when a nasal sampling port is used. J Clin Monit,6,93-98.. al. McSwain SD, Hamel DS, Smith PB, Gentile MA, Srinivasan S, Meliones JN , Cheifetz IM. ( 2010). End-tidal and arterial carbon dioxide measurements correlate across all levels of physiologic dead space. Respir Care, 55(3), 288-293.. M. Normal EA, Zeig NJ, Ahmad I. (1986). Better designs for mass spectrometer monitoring of the awake patient. Anesthesiology, 64,664.. of. Raemer DB, Francis D, Philip JH, Gabel RA.(1983). Variation in PCO2 between arterial blood and peak expired gas during anaesthesia. Anesth Analg, 62, 1065-1069.. ity. Roy J, McNulty SE, Torjman MC(.1991).An Improved nasal prong apparatus for end-tidal carbon dioxide monitoring in awake, sedated patients. Clin Monit,7,249-252.. rs. Russel GB,Graybeal JM, Strout JC.(1990). Stability of arterial to end-tidal carbon dioxide gradients during postoperative cardiorespiratory support. Can J Anaesth,37, 560-566.. ve. Saunders MH, Kern NB, Constatino JP, Stiller RA, Strollo PJ, Studnicki KA, Castes JA, Richards TJ. ( 1994). Accuracy of end-tidal and transcutaneous PCO2 monitoring during sleep. Chest, 106,472-483.. ni. Schamman FL, From RP.(1998). Ventilatory frequency influences accuracy of end-tidal CO2 measurements. Anesth Analg, 23, 1140-1142.. U. Soto RG, Fu ES, Vila H Jr, Miguel RV. (2004). Capnography accurately detects apnea during monitored anaesthesia care. Anaesth Analg, 99, 379-382. Tai CC, Lu FL, Chen PC, Jeng SF, Chou HC, Chen YC, Tsao PN, Hsieh WS. ( 2010). Noninvasive capnometry for end-tidal carbon dioxide monitoring via nasal cannula in nonintubated neonates. Pediatr Neonatal,51(6),330-335. Takaki S, Mizutani K, Fukuchi M, Yoshida T, Idei M…Goto T. ( 2016). Deep Breathing improves end-tidal carbon dioxide monitoring of an oxygen nasal cannula-based capnometry device in subjects extubated after abdominal surgery. Respiratory Care. Doi: 10.4187/respcare.04634.. 42.

(52) Tobias JD, Flanagan JFK, Wheeler TJ, Garrett JS, Burney C. (1994). Noninvasive monitoring of end-tidal CO2 via nasal cannulas in spontaneously breathing children during the perioperative period. Crit Care Med,22,1805-1808. Wahba RW, Tessler MJ. ( 1996). Misleading end-tidal CO2 tensions. Can J Anaesth, 43, 862-866. Wright SW. (1992). Conscious sedation in the emergency department: the value of capnography and pulse oximetry. Ann Emerg Med, 21(5),551-555.. ay a. Yazigi A, Zeeni C, Richa F, Chalhoub V, Sleilaty G.(2007). The accuracy of non-invasive nasal capnography in morbidly obese patients after bariatric surgery. M.E.J Anesth, 19 (3) ,483492.. U. ni. ve. rs. ity. of. M al. Yosefy C, Hay E, Magen E, Reisin L ( 2004). End tidal carbon dioxide as a predictor of arterial PCO2 in the emergency department setting. Emerg Med J, 21,557-559. Doi: 10.1136/emj.2003.005819.. 43.

(53)