Near infra‐red spectroscopy

Near infra-red spectroscopy (NIRS) is well established for studying cerebral oxygenation and haemodynamics in neonates [ 1-31 using ‘transmission spectroscopy’, where light is shone across the whole head. In adults the large head and thick skull require ‘reflectance spectroscopy’, with the optodes a few centimetres apart on the same side of the head and several papers have recently been presented using this technique [MI. NIRS is deceptively easy to apply, but the path taken by near infra-red light is not clearly established and some understanding of the physical principles and.assumptions is necessary to avoid pitfalls in interpreting the results. The fundamental principle is straightforward: for a fixed distance, the absorption of NIR light is proportional to the concentration of the chromophore (e.g. haemoglobin) according to the Beer-Lambert law. As light enters the tissues it is scattered and a ‘random-walk’ or ‘Monte Carlo’ prediction can be made of the probability of different paths for the light according to the scattering and absorption coefficients of the different tissues. Based on estimated coefficients, the classical theory suggested that if the separation between ‘in’ and ‘out’ optodes was less than 3 an, the light would pass through extracranial tissues only, while with increasing separation an increasing proportion of the brain would be seen. One instrument (INVOS 3100) was released using two channels at 1 .O cm (surface) and 2.7 cm (brain) separation, designed so that by subtraction of the external signal a pure brain signal could be produced [7]; however, jugular bulb [8] and C02 response [9] studies showed little response to changes in cerebral oxygenation. Subsequent data using CO, challenge (isolated increase in intra-cerebral flow) [lo], selective carotid clamping [ 1 11 and indocyanine green injection [ 121 suggested that the extracranial contribution to the NIR signal was greater than had been predicted and was still noticeable at 7cm separation [lo]. Duncan et al. [I31 subsequently measured the scattering produced by adult frontal bone and found it to be greater than had been estimated; with the new coefficients it is suggested that separations less than 4 cm reflect predominantly extracranial tissues. The scattering of the skull may vary between patients; the amount of light collected at a fixed space is greater for women than for men [14] and it is likely that other areas of the skull will also have different scattering. The position of the optodes is critical. When optodes are placed either side of the insertion of the temporalis muscle, 30 s maximal teeth clenching produces true ischaemia in the muscle (reduction in high energy phosphates on MR spectroscopy-unpublished data) with a marked fall in oxidised haemoglobin (HbO), an increase in reduced (HbR) and a hyperaemic response on release of clenching [15] in parallel with the external carotid velocity, even though there is no change in middle cerebral arterial velocity (MCAV). The mass of muscle is small, but clearly affects the NIR signal even at 5cm optode separation. The suggestion in a recent paper [5] to use more lateral optode positions to improve correspondence with MCAV monitoring should be resisted if the optodes straddle temporalis. So the influence of the extracranial circulation on NIRS signals may be high and dependent on optode separation and position. Does this mean that studies of cerebral oxygenation using NIRS are flawed and are not really looking at the brain? The answer depends on two questions: can extra cranial oxygenation change under the conditions of the study, and if it can, what is the proportion of the NIRS signal coming from intraand extracranial compartments? If there is no possible change, then the scalp and skull will just reduce the magnitude of the intracranial signal, but will still permit measurement of changes. There are few data on the extracranial circulation. Factors affecting skull perfusion and oxygenation are not known and are difficult to investigate. NIR studies during craniotomy suggest that about 30% of the signal is absorbed by the skull and thus changes in skull perfusion would have a significant effect on measured NIRS. It is likely that the scalp and surface tissues are affected by changes in temperature, arterial and central venous pressures. As up to 30% of heat loss is through the head, the scalp circulation is likely to be significantly affected by ambient temperature to assist thermal regulation. Increase in core temperature should similarly increase external carotid flow to increase heat loss, but will also alter cerebral oxygen consumption resulting in increased cerebral blood flow (CBF). CBF and thus oxygenation is independent of mean arterial blood pressure between 60 and 160 mmHg and this autoregulation is maintained under moderate anaesthesia ( c 2 MAC) [ 161. It is thought that the external carotid does not exhibit significant autoregulation, but this has not been studied. Alterations in blood pressure could therefore change extracranial perfusion without changing intracranial perfusion. Changes in posture and central venous pressures should have equal effect on both intraand extracranial compartments unless the compliance of the venous systems is different. Water is a strong absorber of NIR and is responsible for about 20-70% of the total absorption, depending on the wavelengths being used. Coronary artery bypass surgery produces significant cerebral oedema on magnetic resonance imaging [ 171, but the effect on extracranial tissues is not known. It is likely therefore that under some circumstances extracranial perfusion can change independently from the intracranial circulation, but will these conditions occur during clinical studies? The greatest interest in the use of adult NIRS is for monitoring patients in whom cerebral oxygenation is unpredictable-coronary artery bypass surgery, major surgery, intensive care, hyperand hypotension-and these are the same situations where external carotid oxygenation may be altered. Hypothermic cardiopulmonary bypass in particular may require care, as cooling and rewarming is uneven, cerebral oedema is produced and perfusion during bypass is abnormal, even with ‘pulsatile’ perfusion.