Discrepancy in finger pulse oximetry reading related to positioning: a case report

Abstract

Pulse oximetry is one of the five cardinal vital signs used to monitor patients in the clinical setting, and has contributed significantly to patient safety. Unfortunately, extremes in patient positioning may lead to changes in peripheral perfusion pressures resulting in erroneous pulse oximetry readings. We present a case of a relatively well patient coming for robot-assisted laparoscopic radical prostatectomy who became hypoxic in the Trendelenburg position that spontaneously resolved upon transiting to supine. The reliability of the traditional method of assessing the pulse oximeter value via the plethysmograph is questioned and we discuss other modalities to assist in interpretation of the suspicious pulse oximetry reading.

Background

The development of pulse oximeters in medicine has greatly influenced medicine and has led to a significant improvement in patient care and safety. Since its invention, pulse oximetry use is associated with a 19-fold increase in detection of hypoxemia episodes. ¹ Despite its relevance in today’s practice in providing continuous oxygen saturation monitoring, it is not without its flaws. We present an interesting case of a large discrepancy in the finger pulse oxygen saturation reading in the Trendelenburg position to the actual arterial oxygen saturation and discuss the possible reasons for this. We also explore the use of ancillary tools to assist in determining the validity of the pulse oxygen saturation reading.

Written consent was taken from patient for publication.

Case

A 67-year-old gentleman (60.8 kg, 1.57 m, body mass index 24.7 kg/m²), non-smoker, with good effort tolerance, was admitted electively for a robot-assisted laparoscopic radical prostatectomy for prostate cancer. He had a history of quiescent gastric reflux disease, dyslipidaemia and impaired fasting glucose. One month prior, he underwent an uneventful prostate biopsy under general anaesthesia.

A three-lead electrocardiogram, pulse oximeter, non-invasive blood pressure and capnograph were applied. Induction was conducted using propofol, atracurium and fentanyl, and he was intubated with a size 8 endotracheal tube and thereafter maintained on an oxygen/air/desflurane mixture. Both arms were kept tucked in and he was shifted into lithotomy. An hour into surgery, after placement of the robotic arms and positioning into a steep Trendelenburg incline, patient’s finger oxygen saturation (Phillips FAST SpO₂ reusable rigid finger probe, standard adult size) started to fall from 100% to 60% on FiO₂ 0.5 despite a good waveform, BP 110–120/70–80 mmHg, HR 70–80 bpm (see Figure 1). There was no change in ventilator setting: volume controlled ventilation, tidal volume 450 ml and a respiratory rate of 12, and airway pressures remained stable at 15 cmH₂O. Procedure was halted, and he was placed on 100% O₂. The endotracheal tube was pulled out by 2 cm to 20 cm at the lips in view of suspected endobronchial intubation, chest was clear with bilateral equal air entry and the table was levelled, with resolution of the desaturation episode. However, upon transitioning to the Trendelenburg position, patient started to desaturate again. No difference was noticed when a nasal saturation probe was used (Phillips FAST SpO₂ reusable nasal probe). This time, direct bronchoscopy was performed confirming adequate endotracheal tube positioning. We suspected then that the SpO₂ readings were inaccurate as the patient looked pink, and the haemodynamics remained within normal limits despite a suspected severe hypoxia. This was confirmed with an arterial blood gas (PaO₂ of 220 mmHg). By this time, a low perfusion probe (Philips Masimo module, reusable rigid finger probe, standard adult size) was available and when applied onto the same finger showed a significant discrepancy in readings (100% vs 59%). The procedure thereafter proceeded uneventfully. Much later in the case, the perfusion index (PI) function was activated on the Phillips monitor (Philips, IntelliVue MP70 Anaesthesia, SW revision J) which demonstrated a difference in both the Masimo and FAST modules and their corresponding values when applied on different fingers on the same hand (see Figure 2). He was discharged home well on postoperative day 3.

Figure 1.

Discrepancy between Phillips FAST module finger probe monitor and Philips Masimo module.

Figure 2.

Perfusion numeric activated on the Phillips monitor (Philips, IntelliVue MP70 Anaesthesia, SW revision J). The orange circles are reflective of the Phillips FAST SpO2 reusable rigid finger probe while the green circles represent data from the Philips Masimo module, reusable rigid finger probe.

A retrospective review of his anaesthetic one month later did not reveal any episodes of desaturation for a procedure which was performed in the lithotomy position.

Discussion

Oxygen saturation monitors have revolutionized medicine particularly in areas of neonatology, anaesthesia and intensive care. Pulse oximetry provides us with a non-invasive, continuous means to monitor oxygen saturations. It is reliable and accurate with a fast response time, and also has good correlation between directly measured arterial oxygen saturation (r = 0.95) and measured arterial partial pressure of oxygen (r = 0.95). ² However, it is not without its limitations.

In patients with poor perfusion pressures, it is difficult for the traditional pulse oximeter to eliminate venous pulsations thereby resulting in inaccurate SpO₂ readings. The improved technology in low perfusion monitors is able to identify these aberrant venous signals and extract the arterial signal from them. Low perfusion pulse oximeters have since been developed, e.g. by Masimo and Nellcor, to counteract the limitations of the traditional pulse oximeter. The algorithms used are specifically targeted to reduce motion artefacts and to improve the integrity of the SpO₂ reading. ³ In particular, the Masimo pulse oximeter (Masimo Corporation, Mission Viejo, California, USA) analyses SpO₂ through the signal extraction technology (SET). SET calculates the patient’s arterial oxygen saturation by the use of five different but parallel algorithms. This enhanced technology removes movement artefacts and noise to produce a clearer arterial signal. ⁴ In contrast, the FAST technology by Philips works by applying the fast Fourier transformation to break down the red and infrared signals into individual frequencies, and uses this to distinguish artefacts from actual signals. Simulated conditions of decreased peripheral perfusion and increased motion in healthy individuals have been found to lead to a higher number of erroneous SpO₂ readings, and where the SpO₂ deviation was within ±3% less than 62% of the time. ⁵ In a comparison of finger, nasal, ear and forehead probes in poorly perfused cardiac surgery patients, it was found that nasal and forehead probes performed the most poorly, and finger probes were the best when globally assessed using five criteria (accuracy, precision, number and percentage of readings within 3% of standard and expected over-read limit in 95% of cases).^6–8

True desaturation can occur in a patient undergoing a laparoscopic procedure in a steep Trendelenburg position. The reasons can be divided into anaesthetic-, surgical- or equipment-related. For anaesthesia factors, the endotracheal tube may migrate and become endobronchial with a resultant one lung ventilation or bronchospasm due to irritation of the carina. During positioning, there may be inadequate tidal volumes generated, particularly if patients were placed on pressure controlled ventilation as the abdominal contents splints the diaphragm and causes an elevated airway pressure. There may also be inadvertent surgical port insertion to the subcutaneous tissues leading to subcutaneous emphysema which may track up to the thorax region or lead to pneumothorax. In addition, spuriously low SpO₂ readings may occur due to the inability of the pulse oximeter to read from poorly perfused areas, venous pulsations, nail polish or severe anaemia. We believe that the poor SpO₂ reading in this patient was likely due to poor peripheral perfusion from the Trendelenburg position as we noticed that his hypoxic episodes coincided with each time he was placed in the head-down position and resolved when he was levelled. This may be due to venous pooling over in the dependent regions resulting in poor blood flow to the distal and more elevated extremities. In our patient, the perfusion numeric at the ear was also below 0.4, suggesting that despite the head being a dependent site, the ear in the Trendelenburg position is still a relatively elevated area.

Our findings were confirmed by the activation of the perfusion numeric on the Philips monitor (see Figure 2). For a SpO2 reading of 80%, the perfusion numeric was only 0.33. This was in contrast to the PI on the Masimo probe which showed a value of 2.1. The perfusion numeric provides a value for the pulsatile portion of the measured signal caused by the pulsating arterial blood flow, and is indicative of the pulse strength at the sensor site, ranging from 0 to 10. ⁹ It is recommended by the manufacturers that a value less than 1 and definitely less than 0.3 should lead to repositioning of the probe. The main factor affecting this is peripheral perfusion. Hence, the index is higher during vasodilation, and vice versa. ¹⁰

In contrast, Masimo uses a similar term, the PI, which indicates the strength of the infrared signal returning to the monitoring site. This ranges from 0.02% (very weak pulse strength) to 20% (very strong pulse signal). There is no cut-off to their PI value and they attest that their technology is able to provide reliable SpO₂ readings even at low PI of 0.02%. As the algorithm used in both Philips FAST and Masimo are different, it is difficult to compare the absolute values on the perfusion numeric and PI respectively. In the critically ill, PI has been used to monitor peripheral perfusion and is reflective of the core-to-toe temperature difference. ¹¹ With greater understanding and utility of peripheral perfusion values, they have been used to track vascular tone in obstetric patients receiving spinal anaesthesia as a predictor of hypotension, ¹² in determining depth of anaesthesia in paediatric patients, ¹³ an indicator of the change in peripheral microcirculation, ¹⁴ and as an endothelial dysfunction evaluation method. ¹⁵

Another useful value which can be used in conjunction with the perfusion numeric and PI is the signal quality indicator found on the Philips FAST SpO₂ module. This gives an indication of the reliability of the displayed values, ⁹ and is represented as a filled triangle next to the SpO₂ reading. This should ideally be at least half filled, reflecting at least a medium signal quality. Unfortunately, both the signal quality indicator and the perfusion numeric are not routinely activated on the Phillips monitor MP70 and they require configuration. While the plethysmography waveform of the SpO₂ is commonly used as a reflection of the accuracy of the SpO₂ reading, our experience suggests that this is not reliable and we should assess a suspicious SpO₂ reading together with the signal quality indicator and perfusion numeric.

Nevertheless, in a hypoxic patient, the management strategy remains ABC: airway, breathing and circulation. The patient should also be placed on 100% O₂, and the ventilator and tubing checked for leaks. Meanwhile, the patient should be hand-ventilated to feel for lungs compliance. Low perfusion pressures as the cause of hypoxia should be a diagnosis of exclusion and all the above tests one and problems excluded before it is suspected. A value of <1 will suggest an inaccurate SpO₂ reading and should be confirmed with an arterial blood gas specimen and/or a low perfusion pulse oximeter.

To our knowledge, no studies have been done on the perfusion pressures of the fingers in different positions. We feel that this case report can lead to further studies to confirm our findings and to find out how this influences the pulse oximetry readings.

Conclusion

In conclusion, we present an unusual case of ‘hypoxia’ measured by traditional pulse oximetry that was caused by poor peripheral perfusion, which can be mitigated by the application of a low perfusion probe. We suggest that in addition to configuring the pulse oximetry to show the waveform, it should display the signal quality indicator and the perfusion numeric/index as a default. More studies are required to determine if there is any association between perfusion pressures in various positions and their effects on pulse oximetry.