NIRS during Organ Donation in Brain-dead Donors: Reply

We thank Drs. Dworschak and Pataraia for their interest and thoughtful comments regarding our recent publication on the effect of brain death on regional cerebral oxygenation (rS_cO₂) as determined by near-infrared spectroscopy.  Their concerns relate to our statement that brain perfusion was absent during organ donation according to the characteristics of brain death. This assumption is based on the generally accepted concept that the severely injured brain swells up to an extent, at which cerebral perfusion pressure and therefore cerebral blood flow ceases. Whether cerebral blood flow is actually zero or close to zero is more of academic interest than of clinical relevance. Even assuming the theoretical presence of remaining cerebral blood flow in some of our patients, we must emphasize the central message that cerebral oxygen saturation measurement should not indicate quasinormal values given the circumstances of brain death. In any case, the resulting hypoxia causes irreversible damage not only to the neurons but also to the cells that make up the cerebral vascular and capillary bed, which ultimately become necrotic. Flowers and Patel observed some arterial cerebral blood flow in 6 of 229 (2.6%) radionuclide studies, indicating that remaining residual cerebral blood flow is rather the exception than the rule in suspected (and even more so in confirmed) brain death. They identified risk factors for residual cerebral blood flow such as the expansible skull in infants or the existence of ventriculoperitoneal shunts and (opened) extraventricular drains, which do not apply to our cohort. Only the risk factor decompressive craniectomy was present in three of our patients (table 1).

If one follows the argumentation of Dworschak and Pataraia  and speculates that cerebral blood flow was still present despite brain death, the following should be noted. First, cerebral blood flow must have been less than 16 ml per 100 g per min or 240 ml/min to show the phenomenon of isoelectric EEG or even less than 6 ml per 100 g per min or 90 ml/min to induce irreversible neuronal cell loss both resulting in a critical low cerebral oxygen supply of less than 48 or 18 ml oxygen/min, respectively. Second, cerebral oxygen consumption must have been in the range above zero (to supply at least those cells that supply cerebral vessel walls and capillaries, without which blood flow would not be possible) and far less than 50 ml oxygen/min (which is the oxygen consumption of the normal, uninjured brain). The combination of both factors would have resulted in very low rS_cO₂ values, which did not occur in our patients. Therefore, we exclude the presence of any clinically relevant cerebral blood flow in our patients, as we actually proved by CT angiography in two of them (table 1). Moreover, the time course of rS_cO₂ values of these two patients did not differ from the other 10 patients.

Dworschak and Pataraia raised concerns that “different clamping sites of the aorta (ascending vs. descending aorta) . . . can induce differing shifts in blood volume and affect compartmentalization of intracerebral blood.” While this has been shown in mongrel dogs with intact (and unaffected) brains we could not find scientific evidence that this applies to the special situation of brain death with its no (or marginally low) cerebral perfusion.

Drs. Dworschak and Pataraia regard the “comparison of absolute rS_cO₂ values derived from three different near-infrared spectroscopy monitors” as “a further drawback.” However, these comparisons are by no means unusual and have been made before.  They are legitimate as long as the absolute values of the parameters are compared in compliance with the specifications given by the manufacturers. For the ForeSight Elite, the manufacturers state that the “oximeter module is intended for use as an adjunct monitor of absolute regional hemoglobin oxygen saturation of blood under the sensors in individuals at risk for reduced-flow or no-flow ischemic states.”  The operator’s manual of the INVOS 5100 monitor explicate that it “is intended for use as an adjunct monitor of regional hemoglobin oxygen saturation of blood in the brain or in the tissue beneath the sensor.” 

Different, contradictory information is available on the use of trend data. Although Dworschak and Pataraia state that the “INVOS devices received U.S. Food and Drug Administration clearance as trend monitors only,” the INVOS 5100 operator’s manual raises the warning that “the clinical value of trend data has not been demonstrated in disease states.”  In neither of the two operating instructions did we discover any indications that acute or individual values can only be utilized with restrictions or not at all. If both monitors actually measure hemoglobin oxygen saturation, as stated in the manuals, then it is permissible and sensible to observe both acute and trend values. In contrast, a noncalibrated monitor would output values without units (e.g., 352 arbitrary units), which could only be meaningfully analyzed in relation to previously measured values. However, this is not the case for the monitors used. Irrespective of this, we have taken note of publications according to which the absolute values of the different monitors (despite the above-mentioned information from the manufacturers) can only be compared with each other to a limited extent. 

Finally, we emphasize that it has never been the aim of our study to make near-infrared spectroscopy–based “assumptions regarding lack of residual brain perfusion in brain-dead organ donors.”  Rather, these considerations are based exclusively on the pathophysiology of brain death. We are not aware of any country in which near-infrared spectroscopy is authorized as an ancillary test to confirm brain death. Based on the data we have obtained, we consider it unlikely that such authorization would ever be granted.

We once again thank Drs. Dworschak and Pataraia for their constructive comments and hope that our responses further elucidate our conclusion that both near-infrared spectroscopy monitors, although helpful in monitoring arterial oxygen supply to the head, should not be used to detect cerebral ischemia during conditions of normal extracerebral oxygenation of the head.