No Absolutes in Neuromonitoring for Carotid Endarterectomy
To the Editor:
I have some concerns about the recently published study by Beese et al.1 While I applaud the number of patients they studied, they misled the readers regarding the reliability of somatosensory evoked potentials (SEP) for carotid endarterectomies. I question the validity of their data because of their description of sensor location, inadequate data presentation, and conclusions that are not warranted by their results. I also disagree with some of the points raised in their discussion.
To their credit, the authors addressed some of these concerns, but inadequately, I believe. My 5-year experience with near-infrared spectroscopy (NIRS), the NIRO 500, and the INVOS 3100A and 4100 convinces me that the INVOS accurately measures brain hemoglobin oxygenation even in pathological states. Understanding what is being measured and how best to apply it in its many potential and untested uses is the challenge before us.
First, the authors tacitly imply that SEP monitoring is 100% sensitive and specific by stating that the avalidity of SEP has been established. In fact, previous studies involving 200 to 500 patients report sensitivity of 60% to 100% and specificity of 94% to 100%.2 3 4 No neuromonitoring technique monitors the entire brain. SEP monitors the somatosensory strip over the postcentral gyrus. The middle cerebral artery feeds most of the cerebral cortex and is most likely to be affected by carotid clamping. Frontal placement of the sensor monitors part of the ACA and MCA arterial distributions.
The authors also chose the aloss of SEP, where the N20 and P25 are not discernible, to call for a shunt. Shunts are usually called with a P25 amplitude reduction of ≥50%. In our study,5 any changes in SEP, even decreases of <50%, were correlated with rSo2 decreases of ≥10 units.
Second, if comparisons are to be made with median SEP, some portion of the middle cerebral artery territory should be in the view of the sensor. The sensor should be placed as close to the hairline as possible, with the detectors and light-emitting diode (LED) parallel to and about 0.5 to 1.0 cm from the hairline and as far lateral as possible over the temporal lobe. Patients with a high hairline or who are bald allow placement of the detectors and LED parallel to the sagittal sinus and over the watershed zone. Additionally, avoidance of the frontal sinus is essential. If the inferior aspect of the sensor is placed immediately above the eyebrow, the optodes can be over the frontal sinus. The uncertainty of sensor placement renders the entire study subject to question.
It is important to know what is being “seen” by the sensor. A review of the CT scan, usually available in the operating room, may reveal the presence of an infarct or other anomaly. Infarcted brain could give normal values of oxygenation,6 likely due to the measurement of sequestered, pooled venous blood in nonmetabolizing tissue. This drives home the fact that cerebral oximetry measures oxygen availability and not cerebral blood flow.
Third, their data presentation makes it impossible for the reader to relate individual rSo2 change to the loss or persistence of SEP as presented in Figure 1 in their article. Absolute values were not presented except for the median and ranges. If the authors presented the absolute values or the mean and SD, rather than the median and SD, the 95% confidence interval could be calculated.
Fourth, Beese et al cite 13 different articles to support their statement that absolute rSo2 values are of no value. Of these articles, those reporting poor correlation with jugular venous oxygen saturation in pathological states, such as during carotid cross-clamping, are clearly misguided. Cerebral oximetry should not correlate with jugular venous saturation in pathology because jugular venous saturation is not a reliable indicator of hemispheric ischemia. If blood flow to the ipsilateral hemisphere is completely interrupted by the carotid cross-clamp, the entire outflow is from the opposite hemisphere. In several articles, the validity of cerebral oximetry was questioned because of the lack of correlation with brain blood flow, specifically when blood flow is absent in dead or brain-dead persons. It is critical to understand that the oximeter measures capillary oxyhemoglobin percent saturation and not blood flow.
The INVOS uses the concept of spatial resolution, and the rSo2 index is actually the ratio of the HbO2 signal relative to the isobestic point, thereby giving the absolute percentage of oxyhemoglobin or hemoglobin oxygen saturation. It is true that one cannot calculate the absolute oxyhemoglobin concentration without knowing the path length, but the ratio of HbO2 to total Hb is absolute and valid. In my experience, the normal value in adults ranges between 65 and 75 and in children between 55 and 65, or about 10 points lower than adults. I have not seen values much outside of this range except in patients with pathology. Regarding the variability in anormal people, it is of interest to note that oxygen microelectrodes with tip diameters of 1 μ placed in the cerebral cortex yield values ranging from 0 to arterial Po2. Could this distribution of values be somehow skewed in some anormal patients? It is also uncertain whether some patients may have AV malformations or other problems that may affect cerebral oximetry readings and still be anormal.
Understanding how best to apply and use this new technology in its many potential applications, not only on the brain but on other organs as well, will help further scientific knowledge and improve patient care and thereby reduce the cost of patient care.
- Copyright © 1999 by American Heart Association
Beese U, Langer H, Lang W, Dinkel M. Comparison of near-infrared spectroscopy and somatosensory evoked potentials for the detection of cerebral ischemia during carotid endarterectomy. Stroke. 1998;29:2032–2037.
Horsch S, Vleeschauwer P, Ktenidis K. Intraoperative assessment of cerebral ischemia during carotid surgery. J Cardiovasc Surg. 1990;31:599–602.
Nemoto EM, Yonas H, Kassam A. Cerebral oximetry: cerebral oxygen supply and demand. J Neurosurg Anesth. 1998;10:266. Abstract.