Background: Pulse oximetry fails when pulsations are weak or absent, common in patients with continuous flow left ventricular assist devices (LVADs). We developed a method to measure arterial oxygenation (Sao2) noninvasively in pulseless patients with LVADs.
Methods: The technique involves 5- to 10-s occlusions of radial and ulnar arteries on one hand. A fingertip is transilluminated alternately with light-emitting diodes emitting 660 nm (red) and 905 nm (infrared). During the approximately 1 s after release of occlusion, changing attenuance of each wavelength is measured and their red/infrared arterial blood attenuance ratio (R/IR) calculated. We studied five normal subjects breathing hyperoxic, normoxic, or hypoxic gas mixtures to establish a calibration curve, using standard pulse oximetry as the gold standard. We also studied seven pulseless patients with LVADs (two studied twice) at clinically determined oxygenation.
Results: Normal subject data showed close correlation of oxygen saturation by pulse oximetry (Spo2) with R/IR, (Spo2 = 111 - [26.7 × R/IR]; R2 = 0.975). For patients with LVADs, predicted Sao2 (from the calibration curve) tended to underestimate measured Sao2 (from arterial blood) by a clinically insignificant 1.1 ± 1.6 percentage points (mean ± SD), maximum 3.4 percentage points.
Conclusions: Preliminary results in a small number of patients demonstrate that pulseless oximetry can be used to estimate arterial saturation with acceptable accuracy. A noninvasive oximeter that does not rely on pulsatile flow would be a valuable advance in assessing oxygenation in patients with LVADs, for whom the only current option is arterial puncture, which is painful, risks arterial injury, and only provides a snapshot evaluation of oxygenation.