The objectives of this study were to examine the reported accuracy of cooximetry in determination of methemoglobin levels, to report the apparent discrepant values of "measured" methemoglobin and carboxyhemoglobin, and to discuss the effect that methemoglobinemia has on pulse oximetry readings. Secondly, relative oxygen saturations measured by pulse oximetry and calculated from arterial blood gas samples were compared. A retrospective analysis was performed in an academic medical center with two patients who were diagnosed with acute methemoglobinemia after taking sulfa-containing medications. Serial pulse oximetry, arterial blood gas, and cooximeter analysis were performed for two patients. Cooximetry was performed using an IL-482 oximeter, arterial blood gas analysis was performed using a BLE 1400 oximeter, and pulse oximetry was performed using Ohmeda Biox S740, Hewlett Packard SPO2, or Marquette SPO2 models. Methemoglobin levels ranged from 6% to 48%, however, summation of methemoglobin and oxyhemoglobin were as high as 120%. As a result, negative values of carboxyhemoglobin were recorded by the cooximeter, generating a total sum of 100%. Comparison of pulse oximetry and arterial blood gas oxygen saturation (measured and calculated values, respectively) revealed significant discrepancies; methemoglobin was > 9%, whereas oxygen saturation uniformly was lower with pulse oximetry. Therefore, we recommend that when methemoglobin levels exceed 10%, cooximetry be used as a screen for methemoglobinemia and that serial cooximeter measurements be used to guide therapy and reliance on noninvasive pulse oximetry.