Imputation of PaO2 From SaO2

The Sequential Organ Failure Assessment (SOFA) score is validated as a measure of severity
of illness over time in intensive care unit (ICU) patients and can be used to follow the
course of organ dysfunction and response to treatment1. SOFA has become a frequently used
scoring system of patients in multi-organ failure, given its high sensitivity and
specificity as a predictor of morbidity and mortality in critically ill patients2. The
severity of respiratory dysfunction is measured in the SOFA score by PaO2/FiO2 (PF) ratio,
which is also used as a measure of severity of hypoxemia in patients with acute respiratory
distress syndrome (ARDS).

An invasive sampling of arterial blood gas (ABG) is required to measure the PaO2 for the PF.
Often, patients with less severe hypoxemia may not clinically undergo ABG testing on a
routine basis; hence the clinical and research utility of the SOFA scoring system is
reduced. Furthermore, previous studies revealed concerns about anemia following repeated
blood sampling; hence, the tendency to implement less invasive approaches have led to less
frequent ABG measurements in critically ill patients3. However, almost all ICU patients are
monitored with pulse oximeters, which measure the percent saturation of hemoglobin with
oxygen (SpO2). Whether SpO2 can be used to impute PaO2 for determining the PF ratio has not
been robustly evaluated in a prospective study of critically ill patients.

Prior work investigating the association between PaO2 and SpO2 includes a post hoc study of
ARDS Network patients4. This study excluded patients at altitude, used a linear model for a
highly non-linear relationship, and could not determine whether SpO2 and PaO2 were
simultaneously measured. A similar approach was applied to ARDS Network patients to derive
an SpO2-based respiratory subscore of the SOFA score5. Several similar, retrospective
studies have been performed in mechanically ventilated children, consistently using linear
models of correlation between SF and PF ratios, with similar limitations6-9.

The Ellis inversion10 of the Severinghaus equation11 provides a useful non-linear method for
imputing PaO2 from SaO2. This technique has been used in multiple cohorts of patients with
pneumonia12-14. This simple calculation can be improved by incorporating PaCO2 and pH
values, which could be available from venous blood gases that may be obtained in patients
without arterial catheters. Because Severinghaus/Ellis estimates SaO2, a method for
estimating SaO2 from SpO2 is necessary. There is no current validated and reliable method
for calculating SaO2 on the basis of a measured SpO2. Skin pigmentation affects accuracy of
SpO2, as do sex and oximeter type15,16.

The aims of this study are (1) to find out the relationship between PaO2 and SaO2 among
mechanically ventilated patients and (2) to create a reliable system to utilize SF ratios to
impute the PF ratios in assessing the respiratory parameter of the SOFA score.

Inclusion Criteria:

1. Mechanically ventilated patient in a participating ICU or ED.

2. Arterial blood gas ordered and obtained for clinical reasons.

Exclusion Criteria:

1. Age < 18 years

2. Pregnancy

3. Prisoners