Characterization of tissue oxygen saturation and the vascular occlusion test: influence of measurement sites, probe sizes and deflation thresholds

Non-invasive measurements of tissue oxygen saturation (StO₂) using near-infrared spectroscopy have been studied to assess tissue hypoperfusion in different populations. The absolute StO₂ has been proven instrumental in predicting outcome in trauma patients [1, 2]. These same studies, however, have shown limited discriminating capacity for the absolute StO₂ value (sensitivity, 78%; specificity, 39%) and have even shown it to be at least as good as systolic blood pressure (sensitivity, 74%; specificity, 32%) [1]. The addition of a provocative test (ischaemic challenge) to the measurement of the StO₂ would potentially enhance this discriminatory power between different cardiovascular states by creating emerging parameters to follow, such as the deoxygenation and reoxygenation slopes. Furthermore, the use of such provocative tests [3], specifically the vascular occlusion test (VOT), has been shown to improve and expand the predictive ability of StO₂ to scenarios such as trauma, severe sepsis and septic shock [4, 5].

The VOT is a provocative test in which StO₂ is measured on a distal site (such as the thenar eminence (TH)) whilst a transient rapid vascular occlusion is performed, using a sphygmomanometer, for either a defined time interval (for example, 3 minutes) or until the StO₂ decreases to a defined minimal threshold. Once this threshold is reached, the vascular occlusion is released and the StO₂ is allowed to recover. Several emergent parameters arise from this technique, including the rate of deoxygenation (DeO₂), thought to reflect the local metabolic rate, the rate of reoxygenation (ReO₂), which reflects the time required to wash out stagnant blood and is thought to be determined by local cardiovascular reserve and microcirculatory flow, and the postobstructive hyperemic response.

The VOT, however, has not been standardized. In fact, multiple studies report results using different anatomical StO₂ measuring sites (TH, forearm (F), deltoid) [4, 6‐8], different near-infrared spectroscopy probe sizes (15 mm vs. 25 mm) [4, 8‐12], and different types and levels of deflation thresholds (StO₂ of 40%, 50%) versus time thresholds (3 minutes, 4 minutes, 5 minutes) [4, 6, 11, 13]. Differences between anatomical sites may potentially affect near-infrared spectroscopy-derived parameters given the fact that fat and the muscle thickness distribution along the body is variable. Furthermore, given that increasing the sensing and illuminating probe distance augments the depth that near-infrared light travels, the use of different probes may screen different capillary beds at any given time. Finally, since the primary reason to perform the VOT is to produce a transient ischaemic challenge to note the subsequent ReO₂, the level of ischaemic challenge might vary for the same ischaemic time or threshold StO₂. As a consequence, caution must be taken when comparing results from the available literature on the VOT, as the test may be inadequate [14, 15].

We hypothesized that absolute StO₂ values and VOT-derived StO₂ values obtained by monitoring different anatomical sites or using dissimilar depth may differ, and that different minimal values during the VOT (deflation thresholds) would StO₂ result in dissimilar ReO₂ rates.

The present study was approved by the Institutional Review Board of the University of Pittsburgh. Informed consent was obtained from every volunteer enrolled in the study prior to performing any procedure.

Eighteen healthy volunteers were enrolled in the present study after obtaining informed consent. Demographic characteristics of the studied population are summarized in Table 1. The mean baseline StO₂, DeO₂, ReO₂ and their coefficients of variability for these measures for the various comparisons are summarized in Table 2.

The absolute StO₂, DeO₂ and ReO₂ values for a deflation threshold of 40% were different between the TH and F for both the 15 mm and 25 mm probes (Figures 2 and 3), and were also different at the TH site for the 15 mm and 25 mm probes (Figure 4). Not surprisingly, the TH DeO₂ was not altered by varying the deflation StO₂ thresholds from 30% to 50% when measured with the 15 mm probe. This result was expected given that the DeO₂ is dependent on the metabolic conditions of the tissue, independent of the intensity of the ischemic insult and given the specific calculation methodology [1]. If the metabolic rate of the sampled tissue was not varied, therefore, the DeO₂ should not be significantly modified. The ReO₂ was higher using the 30% threshold than either the 40% or 50% thresholds, which themselves were similar (Figure 5).

The coefficients of variability for the absolute StO₂, DeO₂ and ReO₂ values were also different when comparing both sites and probes, being less with the TH than F measures and with the 15 mm than 25 mm probe size. We also observed a lower coefficient of variability with 40% and 50% thresholds as compared with the 30% threshold (Table 3).

The mean ± standard deviation hyperemic areas under the curve (Figure 1) as measured by TH_{15 mm}, TH_{25 mm}, F_{15 mm} and F_{25 mm} were 11.2 ± 8.8%O₂/minute, 7.7 ± 5.1%O₂/minute, 6.4 ± 5.7%O₂/minute and 13.2 ± 9.6%O₂/minute, respectively, and were all significantly different when compared among each other (Figure 6). Their coefficients of variability were 78.7%, 66.2%, 90.3% and 72.5%, respectively. The hyperemia durations (mean ± standard deviation) for the same measurements stated above were 2.3 ± 0.8 minutes, 1.8 ± 0.6 minutes, 1.6 ± 0.5 minutes and 2.8 ± 1.2 minutes, respectively. These were also statistically different from each other, as shown in Figure 6. Their coefficients of variability were 32.9%, 33.5%, 34.1% and 33.5%, respectively.

Baseline StO₂ values and their emergent parameters of the DeO₂ and the ReO₂, assessed by the VOT, are dependent on the site of measurement, the probe size and, to a certain extent, the ischaemic challenge. Varying any one of these components will significantly vary StO₂-related measures independent of the baseline cardiovascular status. When making inferences about cardiovascular function using these measures, therefore, care must be taken in comparing or combining results from different studies that use different measurement techniques and protocols. Presently, the VOT ischaemic challenge has not been standardized. Several investigators report the DeO₂ and the ReO₂ from a fixed ischaemic time (for example, 3 minutes) rather than to a defined threshold StO₂ value. Since the basal metabolic rate can vary amongst subjects, varying the DeO₂ proportionally [3] using a fixed ischaemic time will result in differing threshold StO₂ values within and among subjects, making changes in the ReO₂ difficult to interpret among subjects if their metabolic rates are also different because they will reach different StO₂ nadirs for a constant ischaemic time. Our data therefore support using a defined ischaemic StO₂ threshold in future studies to minimize ReO₂ variability.

We found that the baseline StO₂, DeO₂ and ReO₂ values were consistently higher in the TH than F sites. Furthermore, the magnitude of the difference between the TH and F was greater for DeO₂ and ReO₂ values than for baseline StO₂ values. There was also a trend toward a greater variability for baseline StO₂ at F than at TH sites. This variability was increased for the DeO₂ and ReO₂. Although our study does not allow us to identify why this difference occurs, most probably it reflects the greater degree of tissue and the associated vascular variability present in the F over the TH. The F has a variable layer of subcutaneous fat, which has a lower metabolic rate and minimal microcirculatory density when compared with muscle. The TH has almost no fat and is highly consistent amongst subjects. Our data therefore support using the TH site rather then the F site when measuring StO₂ and its VOT-related parameters.

We also found consistently lower baseline StO₂ values with the 15 mm probe than with the 25 mm probe, and this difference increased with VOT-related parameters. Although baseline StO₂ values displayed a similar variability (5.6% vs. 4.6%), VOT-derived measurements performed with the 15 mm probe showed less variability than those performed with the 25 mm probe. Presumably, by sampling a smaller amount of tissue volume, the 15 mm probe is less subject to measurement error. Our data support the routine use of the 15 mm probe over the 25 mm probe in future studies.

The hyperemic response, as quantified by the area under the curve and the duration, varied according to the anatomical site of measurement and the probe size used. Furthermore, great variability was found among healthy volunteers with regards to both parameters, as previously pointed out by our group [9]. These results suggest that the hyperemic response is not a reproducible parameter, and has been proven to be highly dependent on individual variability. Thus, it may not be a reliable parameter for clinical use.

The present study has two primary limitations.

First, we studied only two anatomical sites, whereas StO₂ can be measured anywhere on the body. Extrapolation of these findings to other anatomical sites, such as the abdominal wall, thigh and masseter muscles, should be done with caution. Since the VOT can only be carried out realistically on a peripheral site for which transient ischaemia will have no lasting deleterious effects, however, these studies reflect the present clinical practice of near-infrared spectroscopy VOT analysis.

Second, we did not study critically ill patients but healthy volunteers. Although this was the goal of the study, its application to critically ill patients might be limited. Previous studies by ourselves and others document that cardiovascular insufficiency can markedly increase intersubject variability [4, 9]. Our data demonstrated a consistently lower coefficient of variation whenever the combination of TH site and 15 mm probe was used, which suggests that this site may also yield more consistent data in clinical trials. Similarly, using a defined ischaemic threshold for the VOT, method-specific variability of the ReO₂ should also be minimized.