Predictive value of perfusion index (PI) on the discrepancy between oxygen saturation measurement via pulse oximetry and blood gas analysis

Background: Pulse oximetry (SpO2) is a critical component of patient monitoring, particularly in intensive care units (ICUs). Despite advancements, inaccuracies in SpO2 readings compared to arterial oxygen saturation (SaO2) persist, especially in critically ill patients. The perfusion index (PI), indicating the ratio of pulsatile to non-pulsatile blood flow, has been identified as a potential factor influencing SpO2 accuracy. Limited data exists on its predictive value, with conflicting results from studies involving healthy volunteers and animal models. Given the prevalence of low PI in ICU patients, further investigation is warranted. The aim of this study was to assess whether a low PI increases the likelihood of clinically significant discrepancies between SpO2 and SaO2 in critically ill patients. Methods: A retrospective cohort study was conducted, and observational data was collected from May 2021 to December 2021. Arterial blood gas samples and corresponding SpO2 and PI readings were recorded. The study included 330 data points from 80 patients, analysed across different PI ranges. Statistical analysis involved visual inspection of data via a scatter plot, Bland-Altmann analysis, and odds ratio (OR) calculation. Results: When the complete data set is analysed looking at the SpO2 and SaO2 precision is 2.3, accuracy is 2.36 and the bias is 0.33. PI values were grouped in ranges based on the greatest change in variation. Bland-Altman analysis showed wider limits of agreement in the lowest PI range of −5.8 to 7.0 (95%). Patients with a PI of <0.6 had a significantly higher chance (13.2%) of a clinically significant discrepancy (±5%) between SpO2 and SaO2 compared to other PI ranges (3.8–5.1%). The OR for this discrepancy in the <0.6 PI range was 3.36 [95% confidence interval (CI): 1.25 to 8.98, P=0.02]. Additionally, these patients had an increased risk of being more hypoxic on SaO2 than SpO2, OR of 5.57 (95% CI: 1.34 to 23.02, P=0.02). Conclusions: The performance of the pulse oximeter in all PI ranges in terms of accuracy, precision and bias was as per manufacturer standard. However, Bland-Altman analysis shows a clear trend of decreased SpO2 accuracy at low PI values <0.6 with increasing the limits of agreement as PI decreases. The chance of a clinically significant discrepancy reached significance (P=0.02) in the PI range <0.6. There is a signal in the data that the SpO2 tends to overestimate oxygen saturation at low PI (P=0.02). However, the numbers in this analysis are low and should be interpreted with caution. In conclusion, patients with a low PI have a significantly higher chance of having a clinically significant discrepancy between SpO2 and SaO2. Care should be taken to ensure these patients are not hypoxic due to monitoring inaccuracies.