# Thermal intervention improves pulse oximetry accuracy in critically ill patients with low perfusion: a quasi-experimental study

## TL;DR

Warming the wrist and forearm improves pulse oximetry accuracy in ICU patients with poor blood flow, reducing risks of incorrect oxygen monitoring.

## Contribution

A simple thermal intervention is shown to significantly improve pulse oximetry accuracy in critically ill patients with low perfusion.

## Key findings

- Thermal intervention increased perfusion index from 0.56 to 3.59 in ICU patients.
- Pulse oximetry accuracy improved to within ±2% of arterial oxygen saturation after warming.

## Abstract

Pulse oximetry is essential for continuous oxygen monitoring in intensive care, yet its accuracy declines in patients with low peripheral perfusion, risking both unrecognised hypoxaemia and inappropriate oxygen therapy. The perfusion index (PFI) reflects peripheral blood flow and is often reduced in critically ill patients with impaired microcirculation. Simple bedside strategies to restore PFI and improve SpO₂ accuracy remain underexplored.

In this prospective quasi-experimental study, 46 adult ICU patients with arterial catheters and baseline PFI < 1.0 underwent localised peripheral warming using a wrist–forearm heating pad for 15 min. The warming pad maintained a surface temperature of 54 °C, with directly measured skin-interface temperatures of 41–42 °C. SpO₂, PFI, and arterial oxygen saturation (SaO₂) were measured immediately before and after the intervention. The primary outcome was the change in PFI; the secondary outcome was the improvement in SpO₂ accuracy (SpO₂–SaO₂ bias).

Thermal intervention increased PFI from a median (IQR) of 0.56 (0.34–0.78) to 3.59 (2.45–4.77) (p < 0.001; Hedges’ g = 2.43). The pre-intervention SpO₂–SaO₂ bias was 4.09% (95% limits of agreement 0.61–7.56%), which decreased to 0.00% (− 1.22–1.23%) after warming. Improvements were consistent across subgroups and unrelated to cardiac index, vasoactive use, or skin pigmentation.

A brief, localised thermal intervention markedly improves peripheral perfusion and restores pulse-oximetry accuracy to within the clinically acceptable ± 2% range in critically ill patients with low PFI. However, ceiling effects at SpO₂ values near 100% and the pre–post design limit the strength of causal inference. This simple, non-invasive technique can be readily integrated into ICU practice to enhance the reliability of oxygen monitoring and reduce the risk of undetected hypoxaemia or hyperoxaemia.

The online version contains supplementary material available at 10.1186/s40560-026-00847-w.