Severe sepsis reduces accuracy of pulse oximetry

Pulse oximetry provides the means for continuous non-invasive monitoring of blood oxygenation by measuring % oxygen saturation (SpO₂). This safe, easy and convenient method of monitoring oxygenation status is ubiquitous in clinical medicine and is now a routine part of the overall assessment and monitoring of all acutely and critically ill patients, including those with sepsis. 

The validity of pulse oximetry depends on the notion that SpO₂ is an accurate estimation of arterial oxygen saturation (sO₂(a)) generated during blood gas analysis. There is a sound theoretical basis for the suggestion that this might not be the case in patients presenting with severe sepsis. 

A recently published retrospective study sought to establish the accuracy of pulse oximetry in patients with severe sepsis at the time of emergency admission. The study population comprised 88 patients with severe sepsis or septic shock who had been admitted to intensive care unit via the emergency department. All study patients had simultaneous arterial blood gas and pulse oximetry results recorded during their time spent in the emergency department. 

The accuracy of each SpO₂ measurement was assessed by the difference between SpO₂ and sO₂(a) obtained simultaneously (SpO₂ – sO₂(a)). Analysis of the data revealed that pulse oximetry overestimates sO₂(a) in patients with severe sepsis at admission. The mean difference (SpO₂ – sO₂(a)) was 2.75 %. Standard deviation of the differences was 3.1 %. 

A second part of the study sought to discover if the magnitude of difference between SpO₂ and sO₂(a) (i.e. the accuracy of pulse oximetry) is affected by any one of four clinical abnormalities that may be present in patients with severe sepsis: hypoxemia (defined as sO₂(a) <90%), acidosis (pH <7.35), hyperlactatemia (plasma lactate >2.0 mmol/L) and anemia (reduced Hb). 

In addition, the authors sought to discover if the use of vasoactive drugs, frequently prescribed for those in septic shock, affects the accuracy of pulse oximetry. Subgroup analysis revealed that acidosis,  hyperlactatemia, anemia and the use of vasoactive drugs do not affect the accuracy of SpO₂ in patients with severe sepsis, but hypoxemia does. 

The mean difference (SpO_{2 –} sO₂(a)) among patients who were hypoxemic was 4.92 % compared with just 1.89 % for non-hypoxemic patients. Fifty percent (50 %) of patients found to be hypoxemic by sO₂(a) measurement would not have been considered hypoxemic if pulse oximetry only had been used to assess oxygen status. The authors conclude that pulse oximetry overestimates sO₂(a) in patients with severe sepsis and septic shock. 

They find that this overestimation is particularly marked in patients who are hypoxemic. They recommend the use of arterial blood gases if accurate assessment of sO₂(a) is required in patients with sepsis. The paper includes discussion of the theoretical explanation for the observed inaccuracy of pulse oximetry in patients with sepsis.