Whereas venous oxygen saturation can provide important information regarding the balance of oxygen supply and demand,1–3 there is no venous equivalent to the arterial pulse oximeter. Measurement of venous oxygen saturation is currently limited to central or mixed venous saturations, both of which require the placement of central venous catheters with their associated inconvenience, cost, and risks (including but not limited to infection, thromboembolism, arterial puncture and/or rupture, and pneumothorax). Peripheral or regional venous oxygen saturation (SxvO2) may be clinically useful, and several investigators have attempted to measure Sxvo2 using transcutaneous or transmucosal devices.4–6
The goal of this experiment was to determine whether it is possible to use a standard pulse oximeter and the respiratory-induced oscillatory components of venous blood in the antecubital vein, external jugular vein, and internal jugular vein to measure SxvO2 noninvasively.
After obtaining IRB approval and individual written informed consent from 9 male and 1 female volunteers, absorbance data were collected with a Nonin OEM III pulse oximeter (Nonin Medical, Plymouth, MN), using Nonin software on a laptop computer.
Subjects were initially positioned supine and a Nonin Model 8000AA transmittance (finger probe) pulse oximeter was placed on the index finger of the right hand to measure arterial oxygen saturation (SpO2). Opaque towels were placed over the finger probe. Absorbance data were collected for 60 seconds during normal (unrestricted) respiration. Subjects were then asked to inhale nasally and exhale through pursed lips (restricted respiration), and an additional 60 seconds of data were collected.
A Nonin Model 8000R reflectance oximeter was then placed over the right antecubital vein, in axial manner (both the emitter and detector overlying the vein). If the right antecubital vein could not be visualized or palpated, the left antecubital vein was used. If neither vein was visible or palpable, a SonoSite MicroMaxx ultrasound transducer (SonoSite, Bothell, WA) was used to locate the right antecubital vein. Sixty seconds of reflectance data were then collected during both unrestricted and restricted respiration.
The right external jugular vein was identified in similar manner and the reflectance oximeter was placed on the skin surface overlying the vein in its most prominent location. Sixty seconds of reflectance data were then collected for both unrestricted and restricted respiration.
Lastly, the right internal jugular vein was identified using ultrasonography. The reflectance oximeter was then placed on the skin surface overlying the vein and 60 seconds of reflectance data were collected for both unrestricted and restricted respiration.
For all studies, reflected red (610 nm) and infrared (IR) (960 nm) electromagnetic radiation were recorded digitally at 75 Hz with 16-bit resolution.
All data analyses were conducted using computer programs written by the author (RHT, available on request). Each 60-second epoch of red and IR absorbance data was separated into its alternating (AC) and direct (DC) components using a fast Fourier transform (FFT). The DC signal was the amplitude of the FFT at 0 Hz. This is the nonoscillatory component of the reflected signal(s). For each wavelength, the AC component attributable to variations in intrathoracic pressure occurring with respiration was the amplitude of the largest harmonic component occurring between 0 and 0.67 Hz (AC<0.67 Hz). The AC component attributable to variations in arterial blood pressure occurring with cardiac contractions was the amplitude of the largest harmonic component occurring at >0.67 Hz (AC>0.67 Hz). Implicit in this approach is the assumption that the respiratory rate will not exceed 0.67 Hz and that the heart rate will not decrease to <0.67 Hz (40 bpm).
R values were then calculated twice, using the following formulas:
Using previously published correlations between R values and arterial oxygen saturation (Walton et al.6), SxvO2 was estimated.
Eight datasets were analyzed for each subject (4 anatomic locations, with free and restricted respiration at each location), for a total of 80 datasets.
Table 1 lists the characteristics of the study population. Figure 1 shows SpO2 measured with pulse oximeter on the finger and SxvO2 measured by a reflectance oximeter placed over the antecubital, external jugular, and internal jugular veins. The majority of venous measurements were suggestive of nonarterial blood, with 93% of predicted SxvO2 values <90%, 84% of predicted SxvO2 values <85%, and 57% of predicted SxvO2 values <80%. Figure 2 shows representative data from 1 subject.
Table 2 lists the calculated oxygen saturations for each anatomic site. Table 3 shows the effect of restricted respiration on signal strength (defined as AC/DC for a particular wavelength of electromagnetic radiation) at both the respiratory rate ([AC<0.67 Hz, restricted/ DC, restricted]/[AC<0.67 Hz, unrestricted/DC, unrestricted] × 100%) and heart rate ([AC>0.67 Hz, restricted/ DC, restricted]/[AC>0.67 Hz, unrestricted/DC, unrestricted] × 100%).
Our results suggest that when a reflectance oximetry probe is placed directly over a venous structure, the amplitude of the reflected signals at both the respiratory rate and the heart rate provide a measurement consistent with expected values for venous oxygen saturation. The detection of venous blood in reflected signals oscillating at the heart rate was an unexpected finding. We postulate that oscillations of venous blood at the respiratory rate are attributable to changes in intrathoracic pressure, and that oscillations of venous blood at the heart rate may be attributable to either subtle changes in venous pressure transmitted from the right heart, or to the introduction of movement by underlying arteries located in the vicinity of the veins. The implication is that exclusive analysis of the pulsations at the respiratory rate may be misleading, especially in instances when the signal strength is higher at the heart rate.
Crabtree et al.4 used an inflatable finger cuff (operating at 8 Hz) to introduce venous pulsations in the finger. They placed a finger-probe–based transmittance pulse oximeter distal to the cuff. FFT analysis revealed 2 AC signals, 1 at the heart rate (approximately 1 Hz) and another at 8 Hz. By measuring the relative amplitudes of red and IR absorbance at 8 Hz, Crabtree et al. calculated SxvO2 for the finger. When changes in SxvO2 were compared with changes in mixed venous oxygen saturation4 and changes in global consumption of oxygen5 in patients undergoing cardiac surgery, the data generated by the technique of Crabtree et al. did not appear to be reliable enough for clinical use. Importantly, Crabtree et al. did not place their device directly over venous structures, and did not attempt to measure the peripheral venous saturations.
Recently, Walton et al.6 postulated that the measurement of oximetry signals from the esophageal mucosa could be used to assess esophageal venous oxygen concentration. Their device consisted of a 2-wavelength (655 and 880 nm) reflectance oximeter designed to fit in a 20F gastric tube and advanced into the esophagus. The major advantage of this device is its placement in the thorax, which maximizes the expected venous pulsations due to intrathoracic pressure changes. One potential shortcoming of this device is the inability to place it directly over a vein. Walton et al. could not measure SxvO2 in the mucosa and could not validate the device.
Although our data suggest that measuring transcutaneous regional venous oxygen saturation may be a possibility, we did not sample venous blood in this feasibility study, and did not validate this technique. Other limitations of our study include our relatively homogeneous patient population (most were male, all had a body mass index of ≤30 kg/m2), our inability to control (or measure) intrathoracic pressure changes associated with “restricted” respiration, and our failure to standardize subject temperatures.
Several concerns that arose from analysis of our data must be addressed. Why, for instance, do estimates of SxvO2 sometimes change with initiation of restricted respiration? Why are estimates of SxvO2 obtained over the internal jugular vein similar to those obtained over the external jugular and antecubital veins? What signal processing techniques can be used to filter out clearly erroneous data? What is the minimal acquisition time required to obtain reliable data? Does changing the strength of the radiation source (red, IR) affect the reliability of the monitor? Are results improved by mechanical ventilation (less variation in respiratory rate and more profound changes in intrathoracic pressure would presumably produce cleaner FFTs)?
Mannheimer et al.7 showed that standard photoplethysmographic wavelengths are significantly attenuated by tissue depths in excess of 10 mm, and we believe that the “window” of our proposed device was similarly constrained. It is possible that changes in intrathoracic pressure could affect the amount of venous blood in the “field of view,” with blood moving in and out of the field of view (as opposed to expanding and contracting within it) potentially complicating the determination of SxvO2.
Analysis of fetal reflectance oximetry probes suggests that increasing the wavelength of red light to the near IR range (700 nm) improves the accuracy at low saturations.8 The wavelengths used by arterial pulse oximeters were selected to achieve maximal sensitivity in the 80% to 90% range8; selecting wavelengths that optimize sensitivity in the 65% to 85% range might improve SxvO2 measurements while at the same time increasing depth of penetration.
These findings point toward mechanisms by which transcutaneous techniques might be used to measure more central venous oxygen concentrations, such as those found in large, venous structures (e.g., femoral, internal jugular veins). Incorporation of techniques used in near IR spectroscopy (such as the use of multiple detectors to remove interference from intervening tissue) and fetal oximetry (wavelengths optimized for discrimination of typical venous, rather than arterial, saturations), as well as advanced signal processing techniques, may allow for targeted, transcutaneous analysis of venous blood in veins located within 1 to 2 cm of the skin surface, such as that contained in the internal jugular vein.
It is not known whether the ability to measure regional venous saturations in anesthetized, paralyzed patients is clinically useful. A device capable of measuring SxvO2 might be used to detect changes in cardiac output, assess the adequacy of regional oxygen supply:demand matching, or predict the onset of vasodilation (for instance, in the setting of sepsis). However, until there is such a device, these potential uses are speculative at best.
By placing a transmittance oximetry probe over 3 large veins (antecubital, external jugular, and internal jugular), decomposing the acquired signals into individual frequency components, and calculating R values at both the heart rate and respiratory rate, we calculated reasonable values for SxvO2 The signal strength increased significantly when subjects exhaled through pursed lips. Future validation requires peripheral and central venous blood sampling upon which advanced signal analysis techniques will be based.
Marcel E. Durieux is Section Editor of Anesthetic Preclinical Pharmacology for the Journal. This manuscript was handled by Dwayne R. Westenskow, Section Editor of Technology, Computing, and Simulation, and Dr. Durieux was not involved in any way with the editorial process or decision.
Name: Robert H. Thiele, MD.
Attestation: Study design, conduct of study, data analysis, and manuscript preparation.
Name: Jason M. Tucker-Schwartz, MS.
Attestation: Study design and data analysis.
Name: Yao Lu, MD.
Attestation: Conduct of study.
Name: George T. Gillies, PhD.
Attestation: Data analysis.
Name: Marcel E. Durieux, MD, PhD.
Attestation: Study design and manuscript preparation.
1. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, Peterson E, Tomlanovich M. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001;345:1368–77
2. Pearse RM, Hinds CJ. Should we use central venous saturation to guide management in high-risk surgical patients? Crit Care 2006;10:181
3. Reinhart K, Bloos F. The value of venous oximetry. Curr Opin Crit Care 2005;11:259–63
4. Crabtree V, Echiadis A, Smith P, Boehm M, Oc M, Bence J, Machin D, Swanavelder J, Alexiou C, Doukas G, Pidgeon W, Hadjinikolaou L, Spyt T. Prospective venox feasibility study. Conf Proc IEEE Eng Med Biol Soc 2005;2:1968–71
5. Echiadis AS, Crabtree VP, Bence J, Hadjinikolaou L, Alexiou C, Spyt TJ, Hu S. Non-invasive measurement of peripheral venous oxygen saturation using a new venous oximetry method: evaluation during bypass in heart surgery. Physiol Meas 2007; 28:897–911
6. Walton ZD, Kyriacou PA, Silverman DG, Shelley KH. Measuring venous oxygenation using the photoplethysmograph waveform. J Clin Monit Comput 2010;24:295–303
7. Mannheimer PD, Fein ME, Casciani JR. Physio-optical considerations in the design of fetal pulse oximetry sensors. Eur J Obstet Gynecol Reprod Biol 1997;72:S9–19
© 2011 International Anesthesia Research Society
8. Mannheimer PD, Casciani JR, Fein ME, Nierlich SL. Wavelength selection for low-saturation pulse oximetry. IEEE Trans Biomed Eng 1997;44:148–58