Problems with Pulse Oximetry in Screening for Carbon Monoxide Poisoning

By David J. Pierson, MD, Editor, Professor Emeritus, Pulmonary and Critical Care Medicine, University of Washington, Seattle, is Editor for Critical Care Alert.

Synopsis: In this real-world study comparing carboxyhemoglobin saturation as indicated by a commercial pulse oximeter to gold-standard co-oximetry measurements in blood samples, the oximeter gave false-positive and false-negative readings in 9% and 18%, respectively, of patients in an emergency department.

Source: Weaver LK, et al. False positive rate of carbon monoxide saturation by pulse oximetry of emergency department patients. Respir Care 2013;58:232-240.

Weaver and colleagues at Intermountain Medical Center in Murray, Utah, conducted a prospective study to determine the false-positive rate of carboxyhemoglobin (COHb) measurements by pulse oximetry (SpCO) in patients presenting to the emergency department at this level one trauma center. SpCO was determined using the Masimo Rad-57, a pulse oximeter that measures COHb as well as oxyhemoglobin saturation (SpO2). From a simultaneously obtained arterial blood sample in each patient, the investigators also determined COHb by co-oximetry, the gold-standard laboratory measurement. They defined false-positive and false-negative SpCO measurements as those lying 3 percentage points or more above or below the true co-oximetry value, respectively. A convenience sample of 1363 patients was studied over a period of 4.5 months.

Only four of the 1363 patients presented with clinical carbon monoxide (CO) poisoning, and 14 were in shock. Although the Rad-57 oximeter performed within the specifications given by its manufacturer (1 standard deviation = 3 percentage points), there were substantial numbers of both false-positive and false-negative readings as defined by the investigators. False-positive SpCO readings, ranging from 3 to 19 percentage points, were obtained from 122 individuals, or 9% of the subjects. In 247 subjects (18%), false-negative readings were registered by the oximeter, ranging from -13 to -3 percentage points. Female subjects and those with a lower perfusion index were statistically more likely to have false-positive SpCO readings. Weaver et al conclude that while the Rad-57 functioned in this study within its manufacturer’s specifications, clinicians using this oximeter should expect some SpCO readings to be significantly higher or lower than the true COHb values, and should not use SpCO to direct triage or patient management. Further, they caution that a negative SpCO level in a patient suspected of having CO poisoning should never rule out this diagnosis, and should always be confirmed by a COHb measurement in arterial blood.


A false-positive SpCO reading could result in a patient receiving unnecessary treatment for carbon monoxide poisoning, and thus potential morbidity and increased health care costs. Perhaps more importantly, a false-negative reading — that is, a normal or low screening SpCO value in someone who actually did have carbon monoxide poisoning — could be disastrous and potentially lead to an adverse outcome. While no screening device can be expected to perform perfectly in every instance, the findings of this study are troublesome given the widespread marketing and increasing use of SpCO screening with this device in both hospital and prehospital settings.

In an editorial accompanying the article by Weaver et al, Wilcox and Richards discuss the possible reasons why the available literature contains varying results with respect to the accuracy, precision, and limits of agreement of SpCO as measured by the Rad-57 in comparison to actual COHb levels.1 As they point out, much of this heterogeneity, and the fact that most studies are more favorable to the oximeter than this one, relates to aspects of study design, the use of small numbers of often normal subjects, and the utilization of the Rad-57 in near-ideal circumstances rather than in real-world applications as used by Weaver et al. As Wilcox and Richards point out, “While this study found that the device operated within the specifications of the manufacturer, given the discordance of results reported in the available literature, broad reliance on the Rad-57 CO-oximeter for general SpCO screening is premature. Clinicians must continue to have a high index of suspicion for CO poisoning and be aware of the limitations of CO oximetry, while all patients considered to be at risk must have confirmatory blood levels checked.”


  1. Wilcox SR, Richards JB. Noninvasive carbon monoxide detection: Insufficient evidence for broad clinical use. Respir Care 2013;58:376-379.