Noninvasive measurement of carboxyhemoglobin: how accurate is accurate enough?

Exposure to carbon monoxide (CO) is the leading cause of death by poisoning in industrialized countries. It accounts for hundreds of accidental deaths and thousands of intentional deaths annually in the United States. CO is a colorless, odorless gas that is easily absorbed through the lungs when inhaled. It competes with oxygen for binding to hemoglobin, but because hemoglobin has an affinity for CO that is 200 to 250 times greater than that for oxygen, CO poisoning leads to decreased hemoglobin oxygen-carrying capacity, impaired oxygen delivery, and tissue hypoxia. Although timely diagnosis and treatment of CO poisoning are critical, diagnosis based on clinical presentation alone is difficult because symptoms are often nonspecific and can mimic the flu or other viral illness. Definitive diagnosis therefore relies on measurement of increased carboxyhemoglobin levels, which have traditionally been performed on blood samples by using a spectrophotometer. Because of the need to screen many patients with low likelihood of clinically important CO poisoning to ensure that all truly poisoned patients are identified and, further, because of the occasional need to screen large numbers of patients in the setting of a possible mass exposure, a rapid noninvasive diagnostic method would be clinically useful. Although conventional oximeters are unable to distinguish carboxyhemoglobin from oxyhemoglobin and can mistakenly report a 100% oxygen saturation even in the presence of CO, the Rad-57 CO-Oximeter (Masimo Corporation, Irvine, CA) uses a noninvasive sensor to distinguish between oxygenated hemoglobin, deoxygenated hemoglobin, and carboxyhemoglobin. Data are obtained by passing various wavelengths of visible and infrared light through a capillary bed (typically the sensor is placed on a finger), taking advantage of the variable light absorption of the different hemoglobin species. Initial marketing clearance of the Rad-57 was granted by the Food and Drug Administration (FDA) in 2005 on the basis of company-submitted data comparing the noninvasive methodology with simultaneous invasive carboxyhemoglobin