Do Large Vessels Affect Hemodynamic Monitoring by Near Infrared Spectroscopy?

Abstract

Near Infrared Spectroscopy (NIRS) is a widely used technique to assess local tissue oxygenation and blood volume. However, several confounding factors complicate its interpretation. Among these is the potential influence exerted on NIRS variables by large vessels possibly located within the sample volume. This influence is generally neglected, according to the consideration that infrared light entering a large vessel is completely absorbed and thus not backscattered to the receiving optode. On the other hand, vessel size changes vary the amount of light absorbed and thus should profoundly affect NIRS variables. Aim of this study is to ascertain whether vessel size changes, as produced by venous occlusion, can affect NIRS blood volume indices. NIRS monitoring was performed over the basilic vein (BV) in 12 subjects (9 M – 3 F; age 31±8 yr) simultaneously to ultrasound assessment of BV size at the same location, while a second NIRS probe was positioned on the biceps muscle (BM), over a virtually vessel-free area. The arm was elevated (~60°) to induce vein collapse and subjected to venous occlusion, according to two pressure profiles: slow (from 0 to 60 mmHg in 135 s; plateau, 30 s; deflation to 0 mmHg in 135 s) and fast (step inflation to 40 mmHg maintained for 30 s). The cross-sectional area of the BV was continuously recorded along with a NIRS blood volume indicator, i.e., the total hemoglobin concentration (tHb). The tHb index detected a blood volume increase 1.7 and 2.5 times larger over BV than over BM during both slow (p<0.01) and fast (p<0.01) venous occlusions, respectively. The responses to slow venous occlusion exhibited different time courses: tHbBM exhibited a rather linear rise with time, while tHbBV showed a steeper rise at the beginning of the response, resulting in a higher convexity of the curve (p=0.01), similar to the time course of BV size. Similar patterns could be observed also in response to fast occlusions (p<0.05) and also on THI tracings (p<0.05 for both slow and fast occlusions). On average, tHbBV tracings were better correlated to BV area tracings (r=0.90) than tHbBM (r=0.83, p<0.001, after Z-transform), in response to the slow occlusion. In addition, respiratory oscillations in BV size were observed in tHbBV and not in tHbBM. In conclusion, these results challenge the long standing belief that large vessels are excluded from the NIRS sample volume. Experimental evidence strongly supports the concept that size changes of large vessels may profoundly affect NIRS variables. Neglecting this possibility may potentially introduce relevant artefacts in NIRS variables in all conditions affecting venous blood pressure, such as postural changes, or the extravascular pressure, such as muscle contraction.