Recipients of hyperbaric oxygen treatment for carbon
monoxide poisoning and exposure circumstances☆
Jacquelyn H. Clower MPHa,⁎, Neil B. Hampson MDb,
Shahed Iqbal PhDc, Fuyuen Y. Yip PhDc
aCazador under contract to Air Pollution and Respiratory Health Branch, National Center for Environmental Health,
Centers for Disease Control and Prevention, Atlanta, GA, USA
bSection of Hyperbaric Medicine, Virginia Mason Medical Center, Seattle, WA, USA
cAir Pollution and Respiratory Health Branch, National Center for Environmental Health,
Centers for Disease Control and Prevention, Atlanta, GA, USA
Received 9 April 2011; revised 20 May 2011; accepted 22 May 2011
Background: Unintentional carbon monoxide poisoning is preventable. Severe cases are often referred
for hyperbaric oxygen treatment. To guide prevention efforts and treatment practices, this study
provides some of the most detailed current information about patients with carbon monoxide poisoning
who have been treated at hyperbaric facilities across the United States and the circumstances
surrounding their exposures. This study can help improve efforts to prevent carbon monoxide poisoning
and enhance treatment practices.
Methods: From August 2008 to January 2010, nonidentifiable, patient-level data were reported by 87
hyperbaric facilities in 39 states via an online reporting system. This reporting system was developed
collaboratively by the Undersea and Hyperbaric Medical Society and the Centers for Disease Control
Results: Among the 864 patients reported to receive hyperbaric oxygen treatment for unintentional,
non–fire-related, carbon monoxide poisoning, most of the patients were white men aged between 18 and
44 years. Only 10% of patients reported the presence of a carbon monoxide alarm at their exposure
location, and 75% reported being part of a group exposure. Nineteen patients (2%) reported a prior
carbon monoxide exposure. About half (55%) of the patients treated were discharged after treatment;
41% were hospitalized.
Conclusions: The findings in this report expand the knowledge about patients with carbon monoxide
poisoning. These results suggest that prevention efforts, such as educating the public about using carbon
monoxide alarms and targeting the most at-risk populations, may help reduce the number of exposures, the
number of persons with chronic cognitive sequelae, and the resulting burden on the health care system.
© 2012 Elsevier Inc. All rights reserved.
Carbon monoxide is a poisonous gas that is impercep-
tible to humans . It is produced by the incomplete
combustion of hydrocarbons during the burning of fuels
☆Disclaimer: The findings and conclusions in this article are those of
the authors and do not necessarily represent the official position of the
Centers for Disease Control and Prevention.
⁎Corresponding author. Centers for Disease Control and Prevention,
National Center for Environmental Health, Chamblee, GA 30341, USA.
E-mail address: JClower@cdc.gov (J.H. Clower).
0735-6757/$ – see front matter © 2012 Elsevier Inc. All rights reserved.
American Journal of Emergency Medicine (2012) 30, 846–851
such as gasoline, wood, charcoal, propane, natural gas, oil,
and kerosene . Major sources of carbon monoxide
include poorly maintained or ventilated home-heating
systems and cooking appliances, motor vehicle exhaust,
and fuel-powered equipment such as portable generators
and space heaters . Exposure to low levels of carbon
monoxide causes symptoms such as headache, nausea,
vomiting, dizziness, fatigue, and confusion; higher levels of
carbon monoxide can cause disorientation, loss of con-
sciousness, cardiorespiratory failure, and death [4,5]. The
nonspecificity of these symptoms often leads to misdiag-
nosis and likely underestimation of nonfatal cases [1,5,6].
Unintentional carbon monoxide poisoning is preventable
with minimal measures such as using a carbon monoxide
alarm. Despite the ease of this minimal precaution, more
than 20 000 people present to an emergency department,
and approximately 450 people die each year from
unintentional, non–fire-related poisoning [1,7,8]. In 2007,
the symptoms displayed by an estimated 0.3% of the
patients diagnosed with carbon monoxide poisoning in
emergency departments, and 21.6% of the patients who
were hospitalized were sufficiently severe to warrant
hyperbaric oxygen treatment . The Undersea and
Hyperbaric Medical Society (UHMS) recommends hyper-
baric treatment of patients with carbon monoxide poisoning
who manifest loss of consciousness (including transient),
abnormal neurologic findings, evidence of cardiac injury,
significant metabolic acidosis, pregnancy, or a carboxyhe-
moglobin level of 25% or higher . Patients who have
received hyperbaric oxygen treatment for carbon monoxide
poisoning have not been characterized extensively. The
objective of this study is to more clearly characterize
patients who have received hyperbaric treatment for carbon
monoxide poisoning and the circumstances surrounding
their exposures. This information can help to improve case
recognition by clinical professionals and enhance education
and communication to prevent more severe health out-
comes that can result from carbon monoxide poisoning.
Accordingly, the UHMS has partnered with the Centers for
Disease Control and Prevention to collect prompt,
continuous, detailed, patient-level data that are not captured
elsewhere. We are aware of no similar surveillance systems
elsewhere that can fulfill this role.
In collaboration with the Centers for Disease Control and
Prevention, since August 1, 2008, the UHMS has maintained
an online reporting system for those hyperbaric facilities in
the United States that treat patients with carbon monoxide
poisoning. Participating UHMS members voluntarily
respond to a panel of 38 questions, which includes
de-identified information regarding patient demographics,
treatment regimens, and circumstances surrounding the
poisonings on this select group of patients. A detailed
description of the development and operation of the system
is available elsewhere .
This descriptive analysis includes data from the first 18
months of system operation (August 1, 2008, through
January 31, 2010), during which time patient data were
reported by 87 hyperbaric facilities in 39 states. Because the
approaches to preventing intentional and fire-related carbon
monoxide exposures greatly differ from those to preventing
unintentional exposures, only unintentional and non–fire-
related poisonings were included in this analysis. Analysis
was conducted using SAS software (version 9.2; SAS
Institute Inc, Cary, NC). The US Census Bureau's definitions
for regions of the United States were used. Missing,
unknown, and other categories representing less than 5%
of responses were not reported. For bivariate analysis, χ2
Tests (P b .05) were used for categorical variables, and
Wilcoxon rank sum tests were used for the nonnormally
distributed continuous variable, carboxyhemoglobin level.
In the initial 18 months of system operation, 864 patients
were reported to have received hyperbaric oxygen treatment
for unintentional, non–fire-related carbon monoxide poison-
ing. Demographic information for this select group of
treatment for carbon monoxide poisoning and circumstances
surrounding the exposure, United States, August 2008–January
Age group (y)
Level of education
Less than high school graduate
High school graduate
College graduate or more
847Hyperbaric oxygen for carbon monoxide poisoning
patients revealed that the median age was 36 years (range,
birth to 89 years). Most of the patients were men (57%) and
non–Hispanic white (54%; Table 1). Most of the patients
(80%) spoke English primarily; Thirty percent of the
Hispanic white patients spoke English primarily. Reported
exposures also occurred more frequently in the West (34%)
and during the winter (41%).
Data were also collected on the signs and symptoms of
exposure; headache (66%), dizziness (51%), nausea/vomit-
ing (46%), and loss of consciousness (44%) were the most
commonly reported (Table 2). Loss of consciousness for
more than 1 hour was more common among men (72%). Of
the patients with evidence of cardiac ischemia (13%), most
(86%) were hospitalized after hyperbaric treatment.
Participating health care workers reported information on
the circumstances surrounding the carbon monoxide expo-
sures, such as the activity of the patient when they were
exposed. Five hundred forty-four patients (63%) were
exposed while performing a domestic activity such as
cooking or sleeping; 19% of the patients were conducting
occupational tasks such as emergency response (Table 3). Of
the domestic exposures, 252 (46%) were the result of a
furnace or boiler, and 123 (23%) were the result of a
generator. Men comprised the majority of those that reported
performing home or auto repair (84%) and occupational
(87%) or recreational (76%) activities when exposed. In
oxygen treatment for carbon monoxide poisoning, United
States, August 2008–January 2010 (n = 859)
Clinical end points in persons receiving hyperbaric
Most common symptomsa
Loss of consciousness
No. of symptoms per patient
Duration of loss of consciousnessb
Prior carbon monoxide poisoning
Initial normobaric oxygen treatment
Median (minimum, maximum)
aDoes not total 100% because each patient can experience more
than 1 symptom.
bn = 438.
cn = 369.
received hyperbaric oxygen treatment for carbon monoxide
Circumstances surrounding exposure in persons who
Variable n (%)
Location of exposure
Domestic (eg, sleeping)
Home or auto repair
Most common sourcesa
Furnace or boiler
Carbon monoxide alarm in use
Other persons exposed
aExposure can be the result of more than 1 source.
bn = 662. Six exposures had multiple fuels.
848J.H. Clower et al.
contrast, women more commonly reported conducting
domestic activities (55%) or traveling (55%). Seventy
percent of patients who were performing home or auto
repair, 63% who were participating in recreational activities,
and 93% who were occupationally exposed were between
the ages of 18 and 64 years. Seventy-five percent (649) of the
patients reported being exposed with other people, a
circumstance common among many exposures; of this
group, 485 patients (75%) were in a residence and 72%
were performing domestic activities. Thirty-eight patients
(6%) were poisoned by charcoal use, and 49% of those
poisonings occurred among Hispanic white patients. Finally,
among patients who were exposed in their residence, 11%
reported the presence of an alarm.
Prior carbon monoxide poisoning was reported by 19
patients (2%). At the time of recent exposure, 15 (79%) did
not have a carbon monoxide alarm, and 15 were discharged
after treatment (data not shown). Most of the patients spoke
English (n = 18), were between ages 18 and 64 years (n =
16), and were men (n = 13). Ten were performing domestic
activities when exposed; 7 were working.
With regard to patient disposition after hyperbaric
treatment, 55% of the patients were discharged after
hyperbaric treatment, 41% were hospitalized, and for 4%,
disposition was missing (Table 4). The discharged and
admitted groups differed significantly by sex, race/
ethnicity, age group, and primary language (P b .05).
Nine percent of hospitalized patients and 11% of
discharged patients reported the presence of a carbon
monoxide alarm; this difference was also significant.
Finally, patients who were hospitalized after treatment
were significantly more likely to exhibit signs of resultant
ischemia than patients who were released (27% vs 2%).
Understanding why certain groups of patients experience
higher rates of carbon monoxide poisoning is important for
case recognition and prevention efforts. For example, higher
proportions of carbon monoxide poisoning among men have
been reported in populations with fatal carbon monoxide
exposures; this pattern has been attributed to engaging more
frequently in high-risk behaviors such as using fuel-burning
tools . A similar pattern is seen in this study population,
given that men account for the large majority of those
performing home or auto repair and occupational activities
when they were poisoned. Most of the patients who lost
consciousness for more than 1 hour were men. Finally, men
comprise a larger proportion of patients who were hospital-
ized after treatment than patients who were discharged home,
suggesting that men were more severely exposed.
It has been reported in smaller studies that Hispanic white
patients are another population that has higher rates of
severe, unintentional carbon monoxide poisoning caused by
using charcoal [11,12]. In those studies, as well as this one,
Hispanic white cases accounted for the largest segment or a
disproportionate percentage of the population poisoned by
charcoal use. This pattern has been attributed, in part, to
recent immigrants' continuation of cultural practices com-
mon to their native country . For our analysis, the
relatively small percentage of Hispanic white patients who
reported speaking English suggests that bilingual prevention
messaging may be an important approach to reducing carbon
monoxide exposure in this population.
Seasonal and geographic patterns seen in these uninten-
tional carbon monoxide exposures may also be useful in
clinical diagnostic and prevention efforts. Poison center
calls, emergency department visits, and hospitalizations for
oxygen treatment for carbon monoxide poisoning and
circumstances surrounding the exposure by disposition
Characteristics of persons who received hyperbaric
(n = 353)
treatment (n = 475)
Age group (y)⁎
Carbon monoxide alarm⁎
Season of exposure⁎
⁎χ2test, P b .05.
†Wilcoxon rank sum test, P b .05.
24.6 (1.0-77.0) 21.0 (0.1-46.0)
849Hyperbaric oxygen for carbon monoxide poisoning
carbon monoxide exposure are typically higher in regions of
the United States with the coldest climates, presumably due
to the increased use of home-heating systems, use of
alternative heat sources, and motorists “warming up” their
vehicles in enclosed spaces [1,4]. This may help explain why
winter exposures account for a larger proportion of patients
who were hospitalized after treatment. Although the reported
exposures do follow the expected seasonal trends, they do
not follow the anticipated geographic pattern; this is likely a
reflection of the disproportionate numbers of participating
hyperbaric facilities in each region (Northeast, 16; South, 27;
Midwest, 23; and West 21).
Circumstances surrounding these poisonings suggest that
public health education and intervention about using alarms
would be effective in preventing carbon monoxide
poisonings. For instance, those patients who were hospital-
ized after treatment, as compared with those who were
discharged, were significantly less likely to report using a
carbon monoxide alarm in their home. Similarly, education
and intervention may increase the proportion of reported use
of carbon monoxide alarms in this study population (10%),
which is well below the 36% to 40% of American
households with alarms in 2009 (American Housing and
National Health Interview Surveys). These findings suggest
that carbon monoxide alarm initiatives could be effective in
reducing severe and repeated poisonings. Several other
findings also underscore the potential effectiveness of alarm
use for exposure prevention. Examples of these findings are
that furnaces, generators, and motor vehicles were primary
sources of carbon monoxide and that most of the poisonings
occurred in residential settings, which has also been
reported in those carbon monoxide cases presenting to an
emergency department . Of those reporting the presence
of an alarm in the residence where they were exposed
(11%), information was not available regarding the
functionality of the alarm, but education regarding routine
battery replacement may be able to reduce the number of
similar exposures. Finally, the considerable proportion of
patients who were poisoned with other people (75%)
reinforces the belief that fatal exposures for groups of
people can be the result of a single carbon monoxide source.
Thus, education and intervention about one simple
prevention measure can potentially prevent multiple poi-
sonings and deaths. This is unique when compared with
other health conditions and highlights the importance of
public health efforts that could reduce the number of severe
exposures and, thereby, the public health burden.
The clinical features captured by this reporting system
confirm the severity of carbon monoxide exposures in
patients referred for hyperbaric oxygen treatment and
provide some insight into treatment practices. For instance,
one of the most commonly reported symptoms, loss of
consciousness, is a symptom of more severe carbon
monoxide exposures . The high frequency of loss of
consciousness in this population likely relates its common
usage as an independent criterion for hyperbaric referral.
Also, higher carboxyhemoglobin levels generally cause
more severe neurologic symptoms ; reported carboxyhe-
moglobin levels were significantly higher among patients
who were hospitalized after treatment, which suggests more
severe exposures than in patients who were released.
Carboxyhemoglobin levels of 20% to 60%, however,
typically do not correlate well with patient symptoms or
prognosis ; the median levels observed in this analysis
fell within these boundaries, but the range of measurements
was rather wide. Information on acute cardiac injury, another
sign of severe exposure and indication for hyperbaric
treatment, also showed that patients who were hospitalized
after treatment were significantly more likely to exhibit signs
of resultant ischemia.
This analysis is subject to several limitations. The data
reported by participating physicians are not nationally
representative. In 2010, a survey was conducted on reporting
practices among the 44 facilities that reported cases in 2009 to
assess this limitation. Representatives of 26 (59%) of those
reporting facilities responded to the survey. The 26 facilities
total of 523 patientswith carbonmonoxide poisoning in 2009.
A query of the data found that these facilities reported data on
450 cases, or 86% of those treated. The 450 cases also
represent 64% of the total 706 cases for which data were
are likely a significant underestimation of the total number of
patients treated with hyperbaric oxygen for carbon monoxide
poisoning in the United States because not every case at
participating facilities is reported and not all facilities
participate in the reporting system. Finally, long-term follow-
up of this population of patients is not possible. Despite these
on the largest population of patients treated with hyperbaric
oxygen for carbon monoxide poisoning to date.
The unique information collected by this system, which is
not captured elsewhere, can be used by medical and public
health professionals to identify high-risk populations and
reduce the burden of exposures. The results will likely
contribute to a better understanding of carbon monoxide
poisoning epidemiology and its magnitude and, consequent-
ly, inform and enhance prevention efforts and treatment
practices. Therefore, maintaining and expanding this report-
ing system are vital.
The authors wish to acknowledge the Undersea and
Hyperbaric Medical Society members who voluntarily
participate in this surveillance system and Susan Dunn for
850J.H. Clower et al.
 Centers for Disease Control and Prevention (CDC). Nonfatal,
unintentional, non–fire-related carbon monoxide exposures—United
States, 2004-2006. MMWR Morb Mortal Wkly Rep 2008;57(33):
 Penny DG, editor. Carbon monoxide poisoning. Boca Raton (Fla):
CRC Press, Taylor & Francis Group; 2008.
 U.S. Environmental Protection Agency. Basic information: carbon
 The National Workgroup on Carbon Monoxide Surveillance. Carbon
monoxide: a model environmental public health indicator. Available
pdf. Accessed March 2011.
 Raub JA, Mathieu-Nolf M, Hampson NB, Thom SR. Carbon
monoxide poisoning—a public health perspective. Toxicology
 Wright J. Chronic and occult carbon monoxide poisoning: we don't
know what we're missing. Emerg Med J 2002;19:386-90.
 Iqbal S, Huay-Zong L, Clower JH, Yip FY, Elixhauser A. Hospital
2007.AmJEmergMed2011,doi:10.1016/j.ajem.2011.03.003, In press.
 Centers for Disease Control and Prevention (CDC). Carbon monoxide-
related deaths—United States, 1999-2004. MMWR Morb Mortal
Wkly Rep 2007;56:1309-12.
 Gesell LB, editor. Hyperbaric oxygen therapy indications. 12th ed.
Durham (NC): Undersea and Hyperbaric Medical Society; 2008.
 Hampson NB, Bell J, Clower JH, Dunn S, Weaver LK. Partnering with
a medical specialty society to perform online public health
surveillance. Undersea Hyperb Med 2011. In press.
 Ralston JD, Hampson NB. Incidence of severe unintentional carbon
monoxide poisoning differs across racial/ethnic categories. Public
Health Rep 2000;115:46-51.
 Hampson NB, Kramer CC, Dunford RG, Norkool DM. Carbon
monoxide poisoning from indoor burning of charcoal briquets. JAMA
 Kim YS. Seasonal variation in carbon monoxide poisoning in Korea.
J Epidemiol Community Health 1985;39:79-81.
 Kao LW, Nanagas KA. Carbon monoxide poisoning. Med Clin North
851 Hyperbaric oxygen for carbon monoxide poisoning