Article, Toxicology

Comparison of non-invasive CPAP with mask use in carbon monoxide poisoning

a b s t r a c t

Background: Carbon monoxide (CO) is one of the major causes of poisoning worldwide. We aimed to investigate the efficacy of the continuous positive airway pressure (CPAP) use in CO poisoning.

Methods: After CO poisoning, one group of patients was treated with a non-rebreather mask (NRB) and another group using the CPAP mode of mechanical ventilation (CPAP). All patients received at least 90 minute treatment. The carboxyhemoglobin saturation (SpCO) levels of all patients were measured from the fingertips with a porta- ble CO-oximeter at 0, 30, 60 and 90 min. The rates of changes in the serially measured SpCO values were obtained using the Wilcoxon signed-rank test.

Results: A total of 45 patients (24 in NRB and 21 in CPAP group) completed the study. The median initial SpCO levels were 24% (21-33) in NRB group, 25% (21-32) in CPAP group, with no statistically significant difference (p _ 0.323). At the 30th, 60th, and 90th minutes of treatment, significantly lower values were obtained from CPAP than NRB (p b 0.001). The COHb half-life was decreased significantly by CPAP [105(70-190) vs 45(30

-120), p b 0.001]. In CPAP group, the fastest decline in the SpCO level was observed for the interval of 0-30 min [Median difference: 8(3-14), p b 0.001].

Conclusions: CPAP lowered the amount of CO in the blood faster than the mask; therefore, it may be effective in the treatment of CO poisoning.

(C) 2020


Carbon monoxide, a colorless, odorless and tasteless gas, often emerges after fires and insufficient combustion of carbon compounds in car engines and gas stoves. CO is one of the major poisoning agents worldwide, causing N50,000 emergency department (ED) visits in the United States alone. It forms a carboxyhemoglobin (COHb) compound in the human body by binding to hemoglobin with 200 times greater af- finity than oxygen. Thus, it prevents the oxygen transport of hemoglo- bin, results in hypoxia in tissues, and shows poisoning effects. In addition, CO binds to heme-containing molecules, such as myoglobin and cytochrome c oxidase, preventing phosphorylation in mitochondria and reducing ATP production [1,2]. The signs of CO poisoning are non- specific, with mild exposures, headache, dizziness, Muscle pain, and neuroPsychological effects, while severe exposures can lead to confu- sion and death [3]. In addition to acute signs of poisoning, Neurological sequelae, which can cause lifelong disability, have also been observed. CO poisoning has also been shown to increase the risk of mortality three times [4].

* Corresponding author at: Department of Emergency Medicine, Adiyaman University Faculty of Medicine, Yunus Emre District, Adiyaman, Turkey.

E-mail address: [email protected] (K. Turgut).

Since CO findings are non-specific, it is difficult to diagnose, espe- cially if there is uncertainty in the patient’s history. However, the COHb level measured in blood gas, history, and physical examination findings help diagnose CO poisoning cases [5]. COHb levels above 3% in non-smokers and 10% in smokers indicate CO exposure. The amount of CO in ambient air, exposure duration, ventilation rate, and blood vol- ume are the main factors that determine the level of COHb. However, COHb levels cannot predict acute intoxication symptoms or the out- come of patients. Inflammatory processes that occur during CO poison- ing are more deterministic [3].

In the treatment of CO poisoning, the most important intervention is the removal of this gas from hemoglobin and prevention of hypoxia. The main treatment that can be used at this stage is providing routine oxy- gen through a nasal cannula or mask [2]. Normobaric oxygen adminis- tration accelerates the removal of CO from hemoglobin, but does not decrease the rate of neurological sequelae. Nevertheless, oxygen, which is easily accessible, cheap and reliable, should be given to all pa- tients experiencing CO poisoning [3]. In addition, Hyperbaric oxygen (HBO2) can be a useful adjunct in certain patients. The 2008 ACEP clin- ical policy recommended HBO2 therapy as Level C (based on prelimi- nary studies or committee decisions), although not mandatory in CO poisoning [6]. Contrary to this report, Weaver et al. reported that

0735-6757/(C) 2020

K. Turgut, E. Yavuz / American Journal of Emergency Medicine 38 (2020) 14541457 1455

HBO2 therapy should be used in CO poisoning. Their study, comparing HBO2 to normobaric Oxygen treatment, found that the former provided a 46% reduction in the rate of neurological sequelae [7]. However, this method requires special equipment and is not available in all EDs. Therefore, continuous positive airway pressure (CPAP) with a tight mask, which is often utilized in pulmonary edema, can be used to de- liver oxygen at a higher pressure than a reservoir mask as a useful sup- plemental treatment in the management of CO poisoning [2,8]. The hypothesis of this study was that CPAP would decrease the CO level faster than normobaric oxygen application with a non-rebreather mask.


Study design and setting

This study, designed as observational prospective research, was con- ducted between January 2019 and December 2019 in the ED of a tertiary hospital. Our ED provides services to an average of 25,000 patients monthly, with four physicians working at each shift. The study was started after obtaining the approval of the Medicines and Medical De- vices Agency and Clinical Research Ethics Committee of Adiyaman Uni- versity (approval number: 2019/2).

Study population and protocol

In this study, adult patients (N18 years old) presenting to the ED due to CO poisoning were prospectively evaluated. At presentation, the medical history and physical examination findings of the patients were recorded, and the carboxyhemoglobin saturation (SpCO) levels were measured with a portable CO-oximeter (MasimoSET rainbow Rad-57 Pulse CO-oximeter, Masimo, Irvine, CA).

Conscious patients with SpCO values of N20% and b35% were in- cluded in the study. These patients were divided into two groups as NRB and CPAP groups according to the order of presentation to the ED (e.g., the first patient was assigned to NRB and the second patient to CPAP group). Then, 15 L/min oxygen was given to NRB group through a non-rebreather mask and CPAP group through a non-invasive me- chanical ventilator (LTV 1200 portable ventilator) using the CPAP mode (FiO2: 100%, PEEP: 5 cm). In CPAP group, an oronasal tight CPAP mask was used to cover the mouth and nose completely. Since it would be difficult for both the patients and healthcare professionals to obtain blood samples repeatedly, the COHb level was measured by ve- nous blood gas only at the time of presentation. Then, all the repetitive measurements (30-minute intervals) were undertaken using a portable CO-oximeter, and the SpCO levels were obtained. In CPAP group, CPAP was administered to the patients for 90 min, and oxygen was continued to be given with a non-rebreather mask. According to our clinical expe- rience, many patients cannot tolerate long-term CPAP therapy; there- fore, we intentionally limited this time to 90 min. During this time, treatment was discontinued in patients who could not tolerate CPAP, and oxygen was given through a mask. In both groups, the treatment was continued for at least 90 min.

Patients with a history of any respiratory problems, such as heart failure, chronic obstructive pulmonary disease, and obstructive Sleep apnea, those with a Glasgow Coma Scale score of b15 at presentation, individuals under 18 years of age, those that did not wish to participate in the study, and pregnant women were not included in the study. The patients who left ED without informing the responsible doctor and those with a history of smoking did not have complete records, and thus were excluded from the study. Therefore, the number of patients in CPAP group decreased by three. The patients with SpCO levels of

<=20 were also excluded from the study and treated by a non- rebreather mask. Furthermore, patients with SpCO levels of>=35 were excluded from the study, and those that did not show any signs of se- vere poisoning were intervened in emergency or intensive care units by providing oxygen with a non-rebreather mask while those with

signs of serious poisoning were transferred to another hospital for HBO2 treatment. Lastly, confused patients with GCS b15 were not in- cluded in the study, and after managing their condition using a mask or intubation if necessary, these patients were transferred to intensive care units. All of the cases included in the study were selected from poi- soning cases caused by stoves with insufficient or inappropriate ventila- tion, which is commonly seen in Turkey.


The SpCO measurement was performed from the index finger of the patients by nurses trained for this procedure, and venous blood gas was simultaneously obtained to measure the COHb levels and minimalize the gap between the sampling and treatment times. In addition to seri- ally measured SpCO levels, the data related to age, gender, presentation complaints, COHb level in blood gas at the time of presentation to ED, length of ED stay, and smoking history were also recorded. The CO half-life was calculated using serially measured SpCO levels.

Venous blood gas was collected from the patients at the time of pre- sentation to ED to obtain the COHb level. Then, a portable CO-oximeter was used to monitor the levels measured from the fingertips at 30- minute intervals considering that COHb monitoring from blood is diffi- cult and undesirable for both the patient and healthcare professionals. This device is applied by attaching it to the index finger or middle finger and gives the SpCO value in 15 s. However, different results were re- ported in previous studies concerning the use of this device. Although ACEP recommended not to use non-invasive COHb measurements to di- agnose CO toxicity because of its low sensitivity, it can be used for screening undifferentiated patients in ED or out-of-hospital settings [4,5,9]. A preliminary study stated that the SpCO device could be safely used as an alternative to COHb measured in blood gas [10]. In the cur- rent study, we used the non-invasive CO-oximeter only for screening and follow-up, and not for the diagnosis of CO toxicity. The diagnosis was performed based on venous blood gas analysis. No significant dif- ference was observed between the initial SpCO and blood gas COHb levels.

Statistical analysis

In the study, the suitability of the numerical data to normal distribu- tion was determined by the Shapiro-Wilk test. According to the results, Student’s t-test and the Mann-Whitney U test were used to analyze the data that did and did not fit normal distribution, respectively. The chi- square test was used to compare qualitative data. Numerical data with normal distribution were shown as mean +- standard deviation, and those with non-normal distribution as median (minimum-maximum) values. Categorical variables were expressed as numbers and percent- ages. In the two groups, the rates of changes in the serially measured SpCO values were assessed by the Wilcoxon signed-rank and Mann- Whitney U tests. SPSS version 17 was used for all statistical analyses. p b 0.05 was considered significant.


A total of 45 patients, 24 in NRB group and 21 in CPAP group, were included in the study. The median age of all patients was 40 years, and there were no significant difference in age between the two groups (p

_ 0.432). There were 11 males and 13 females in NRB group, and

eight males and 13 females in CPAP group, with no significant gender difference between the two groups (p _ 0.6). The smoking rate was 16.7% in NRB and 33.3% in CPAP group (p _ 0.194). The patients pre- sented to the ED with headache mostly (68.9%), followed by nausea- vomiting (15.5%), dyspnea (8.9%), and dizziness (6.7%). When the two groups were examined, there was no difference in the rates of com- plaints (p _ 0.437). The median half-life was 105(70-190) minutes in NRB, 45(30-120) minutes in CPAP group, which indicated a statistically

1456 K. Turgut, E. Yavuz / American Journal of Emergency Medicine 38 (2020) 14541457

significant difference (p b 0.001). The median duration of ED stay was 125 min in CPAP group and 185 min in NRB group (p b 0.001) (Table 1).

Table 2

Changes in the SpCO levels and the COHb values of the groups.

At the time of first presentation, the median COHb level measured from blood gas was 25% (16.6-37), the median SpCO level measured by the portable CO-oximeter was 25% (21-33), and there was no signif- icant difference between the two parameters (p _ 0.528). There was no statistically significant difference in the COHb levels and SpCO values at the first presentation in both groups (p _ 0.674 and p _ 0.323, respec- tively). However, a significant difference was found between groups for the measurements undertaken at the 30th minute [21% (15-28) vs 16% (12-27), p b 0.001]. The measurements at the 60th minute were also significantly lower in CPAP group [17% (11-26) vs 10% (7-25), p b 0.001]. When the 90th minute measurements were examined, CPAP group had significantly lower values [13% (9-25) vs 7% (2-23), p b 0.001] (Table 2).


25 (16.6-37)


25.1 (20-37)



SpCO levels

Minute 0

25 (21-33)

24 (21-33)

25 (21-32)



18 (12-28)

21 (15-28)

16 (12-27)



6 (2-14)

3.5 (2-9)

8 (3-14)


(0-30 min)

p value




Minute 30

18 (12-28)

21 (15-28)

16 (12-27)


Minute 60

15 (7-26)

17 (11-26)

10 (7-25)



(30-60 min)

4 (0-10)

3.5 (0-9)

4 (2-10)


p value




Minute 60

15 (7-26)

17 (11-26)

10 (7-25)


Minute 60

10 (2-25)

13 (9-25)

7 (2-23)



4 (1-9)

3.5 (1-9)

4 (1-8)


(60-90 min)

p value




When the CO decrease rates were examined within the groups, it was seen that CPAP group had the fastest decrease from 0 to 30 min [Median difference: 8% (3-14), p b 0.001]. The difference in SpCO be- tween minutes 0 and 30 was significantly higher in CPAP group com- pared to NRB group [3.5% (2-9) vs 8% (3-14), p b 0.001]. Similarly, the

Total Median (min-max)


Median (min-max)


Median (min-max)

p value

SpCO difference between the 30th and 60th minutes was higher in CPAP group [3.5% (0-9) vs 4% (2-10), p = 0.04]. However, the change in the SpCO values from the 60th to the 90th minutes did not signifi- cantly differ between the two groups [3.5% (1-9) vs 4% (1-8), p = 0.836]. There was a significant decrease in all measurements in both groups (p b 0.001) (Table 2). The time-dependent changes in the me- dian CO levels of the two groups are shown in Fig. 1. All the patients in- cluded in the study recovered and were discharged from ED following treatment.


In CO poisoning, which is still a serious problem worldwide, the cor- nerstone of the treatment is the administration of 100% oxygen through a non-rebreather mask [7]. In addition, although there is no general con- sensus, HBO2 therapy is often recommended in CO poisoning. However, no clinical parameter, including COHb level, has value in determining the cases in which to apply HBO2 therapy [11]. It is postulated that this therapy provides neurological improvement by significantly reduc- ing the half-life of COHb. However, it remains unclear whether HBO2 is better than normobaric oxygen therapy in reducing the rate of neuro- logical sequelae [4].While some studies show that this treatment re- duces the rate of neurological sequelae [7,12,13], others report that HBO2 therapy either does not decrease or have no effect on this rate [14,15]. In addition to these conflicting results, HBO2has certain

Table 1

Characteristics of the patients.


(n = 45)


(n = 24)


(n = 21)

p value


40 (18-73)

39 (18-72)

40 (19-73)


Male gender

19 (42.2%)

11 (45.8%)

8 (38.1%)


Half-life of CO

ED stay (minutes)
















1.8 (1-5.8)







11 (24.4%)

4 (16.7%)


7 (33.3%)



34 (75.6%)

20 (83.3%)

14 (66.7%)



31 (68.9%)

17 (70.8%)

14 (66.7%)




7 (15.5%)

5 (20.8)

2 (9.5%)


3 (6.7%)

1 (4.2%)

2 (9.5%)


4 (8.9%)

1 (4.2%)

3 (14.3%)

NRB: patients treated by non-rebreather mask; CPAP: patients treated by continuous pos- itive airway pressure ventilation; CO: carbon monoxide; ED: emergency department.

COHb: carboxyhemoglobin; SpCO: COHb measured by portable CO-oximetry Dif: difference.

disadvantages that prevent it from becoming a good treatment alterna- tive, such as the high cost of application, unavailability in many centers, and related complications, including barotrauma, sinus damage, hyperoxic seizures, and Gas embolism [12].

The disadvantages of existing treatments for CO poisoning have led many scientists to seek alternative options. Delvau et al. [16] considered that giving oxygen to patients through CPAP would remove CO from tis- sues faster than a normal mask and demonstrated this using a swine model. In addition, in several case reports in the literature, it was found that the CPAP method decreased the COHb level more rapidly compared to a normal mask [2,8]. To the best of our knowledge, this is the first prospective controlled study in the literature comparing CPAP with the normal mask use in CO poisoning. In the present study, it was demonstrated that CPAP did decrease the COHb levels faster than a normal mask. In addition, it was determined that this treatment most reduced the CO levels within the first 30 min. This relieved the pa- tients’ complaints more quickly and allowed them to be discharged without remaining in ED for a long time. The emergency waiting time was 127.6 min in the CPAP group and 201.3 min in the mask group.

The half-life of CO is about 320 min with the presence of oxygen in ambient air. Providing 100% oxygen through a normal mask shortens this time by five times [12].Studies have shown that this period, which is approximately 74 min in normobaric oxygen therapy, is reduced to 42 min by HBO2 therapy [1]. In our study, the SpCO levels were

Fig. 1. Time-dependent changes in the SpCO levels.

K. Turgut, E. Yavuz / American Journal of Emergency Medicine 38 (2020) 14541457 1457

measured at 30-minute intervals, and in 81% of CPAP cases, the half-life of COHb was found to be b60 min. In the mask group, the half-life of COHb was longer than 60 min in 91.7% of the patients. The median half-life was decreased to 45 min by CPAP treatment.


The use of a portable CO oximeter rather than blood gas analysis at 30-minute intervals was a limitation of our study. Furthermore, the emergency physicians were not blinded to the study details, which may have caused bias in relation to the patients’ length of ED stay. Another limitation was that the duration of patients’ expo- sure to CO and the oxygen therapy applied in the ambulance were not considered. In addition, the patients were not followed up for neurological sequelae after discharge. Other limitations include the low number of patients and insufficient randomization of cases. The randomization of study was done according to the ED admission order of patients.


In this study, the non-rebreather mask and CPAP method were found to be effective in decreasing the amount of CO in the blood. The highest decrease rate was observed within the first 30 min for the CPAP group. In addition, it was determined that the CPAP method decreased the CO level faster than the non-rebreather mask; therefore, it may be an effective treatment method for CO poi- soning, leading to the early relief of patients’ symptoms and faster discharge from ED.


This research did not receive any specific grant from funding agen- cies in the public, commercial, or not-for-profit sectors.

CRediT authorship contribution statement

Kasim Turgut: Conceptualization, Methodology, Investigation, Data curation, Formal analysis, Writing – review & editing. Erdal Yavuz: Data curation, Investigation, Supervision, Resources, Soft- ware, Validation.

Declaration of competing interest

The authors declare that this manuscript has not been published nor is under simultaneous consideration for publication elsewhere.


We thank Associate Professor Fatih Uckardes (Department of Biosta- tistics and Medical Informatics, Faculty of Medicine, Adiyaman Univer- sity) for his contribution to the statistical analysis of the data.


  1. Rose JJ, Wang L, Xu Q, McTiernam CF, Shiva S, Tejero J, et al. Carbon monoxide poi- soning: pathogenesis, management, and future directions of therapy. Am J Respir Crit Care Med 2017;195(5):596-606.
  2. Idil H, Unek O. Non-invasive CPAP ventilation in acute carbon monoxide poisoning. Prehosp Disaster Med 2019;34(4):454-5.
  3. Weaver LK. Carbon monoxide poisoning. Undersea Hyperb Med 2020;47(1):151-69 First Quarter.
  4. Wolf SJ, Maloney GE, Shih RD, Shy BD, Brown MD, American College of Emergency Physicians. Clinical policy: critical issues in the evaluation and management of adult patients presenting to the emergency department with acute carbon monox- ide poisoning. AnnEmerg Med 2017;69(1):98-107.e6.
  5. Sebbane M, Claret PG, Mercier G, Lefebvre S, Thery R, Dumont R, et al. Emergency department management of suspected carbon monoxide poisoning: role of pulse CO-oximetry. Respir Care 2013;58(10):1614-20.
  6. Wolf SJ, Lavonas EJ, Sloan EP, Jagoda AS, American College of Emergency Physicians. Clinical policy: critical issues in the management of adult patients presenting to the emergency department with acute carbon monoxide poisoning. JEmergNurs 2008; 34(2):e19-32.
  7. Weaver LK, Hopkins RO, Chan KJ, Churchill S, Elliott G, Clemmer TP, et al. Hyperbaric oxygen for acute carbon monoxide poisoning. N Engl J Med 2002;347(14):1057-67.
  8. Roth D, Mayer J, Schreiber W, Herkner H, Laggner AN. Acute carbon monoxide poi- soning treatment by non-invasive CPAP-ventilation, and by reservoir face mask: two simultaneous cases. Am J Emerg Med 2018;36(9):1718.e5-6.
  9. Suner S, Partridge R, Sucov A, Valente J, Chee K, Hughes A, et al. Non-invasive pulse CO-oximetry screening in the emergency department identifies occult carbon mon- oxide toxicity. J Emerg Med 2008;34(4):441-50.
  10. Coulange M, Barthelemy A, Hug F, Thierry AL, De Haro L. Reliability of new pulse CO- oximeter in victims of carbon monoxide poisoning. Undersea Hyperb Med 2008;35 (2):107-11.
  11. Wolf SJ, Lavonas EJ, Sloan EP, Decker WW, Diercks DB, et al. American College of Emergency Physicians. Clinical policy: critical issues in the management of adult pa- tients presenting to the emergency department with acute carbon monoxide poi- soning. Ann Emerg Med 2008;51:138-52.
  12. Lin CH, Su WH, Chen YC, Feng PH, Shen WC, Ong JR, et al. Treatment with normobaric or hyperbaric oxygen and its effect on neuropsychometric dysfunction after carbon monoxide poisoning: a systematic review and meta-analysis of ran- domized controlled trials. Medicine (Baltimore) 2018;97(39):e12456. https://doi. org/10.1097/MD.0000000000012456.
  13. Thom SR, Taber RL, Mendiguren II, Clark JM, Hardy KR, Fisher AB. Delayed neuropsychologic sequelae after carbon monoxide poisoning: prevention by treat- ment with hyperbaric oxygen. Ann Emerg Med 1995;25:474-80.
  14. Scheinkestel CD, Bailey M, Myles PS, Jones K, Cooper DJ, Millar IL, et al. Hyperbaric or normobaric oxygen for acute carbon monoxide poisoning: a randomized controlled clinical trial. Med J Aust 1999;170:203-10.
  15. Annane D, Chadda K, Gajdos P, Jars-Guincestre MC, Chevret S, Raphael JC. Hyperbaric oxygen therapy for acute domestic carbon monoxide poisoning: two randomized controlled trials. Intensive Care Med 2011;37:486-92.
  16. Delvau N, Penaloza A, Liistro G, Thys F, Delattre IK, Hantson P, et al. Effect of pressure support ventilation on carboxyhemoglobin toxicokinetic after acute carbon monox- ide intoxication: a swine model. J Med Toxicol 2018;14(2):128-33.