Changes in the oxyhemoglobin dissociation curve between the arterial and cerebral capillary level during normo- and hyperventilation in healthy human volunteers

We have developed a new method to model the configuration of the oxyhemoglobin dissociation curve at the cerebral capillary level, which may be used to calculate the capillary, PcapO2, and mitochondrial, PmitO2, oxygen tension in humans. In the present study, we sought to examine whether the estimated mean capillary P50 and mean Hill slope differ from measured arterial values during normo- and hyperventilation in healthy volunteers, and whether the use of the respective values results in physiologically relevant differences in PcapO2 and PmitO2. We obtained paired arterial-jugular venous blood samples during resting breathing and following 15 minutes of voluntary hyperventilation in 8 healthy volunteers. A Hill plot was used to estimate the mean capillary P50 and Hill slope. Global cerebral blood flow and the cerebral metabolic rate of oxygen, J O2, were measured by the Kety-Schmidt-technique using 133Xe as the tracer. PmitO2 was found using a standard value for the cerebral oxygen diffusion capacity, LO2, of 33.0 μmol/100g/min/kPa: J O2 = LO2 (PcapO2 - PmitO2 ). During normoventilation, neither P50 (arterial: 3.7 (0.3) kPa; capillary: 3.7 (0.1) kPa; arterial vs. capillary: p = 0.14) nor the Hill slope (arterial: 2.5 (0.1); capillary: 2.4 (0.1); arterial vs. capillary: p = 0.14) changed significantly from artery to capillary. In contrast, both P50 (arterial: 2.9 (0.1) kPa; capillary: 3.4 (0.2) kPa; arterial vs. capillary: p < 0.001) and the Hill slope (arterial: 2.0 (0.1); capillary: 2.2 (0.1); arterial vs. capillary: p < 0.001) increased significantly from artery to capillary during hyperventilation. Neither PcapO2 nor PmitO2 differed when the measured arterial rather than capillary P50 and Hill slope were used during resting breathing (bias of -0.17 kPa and limits of agreement of -0.71 kPa to 0.37 kPa) (Figure 1). By contrast, the use of the capillary P50 and Hill slope resulted in systematically higher PcapO2 and PmitO2 than when using the measured arterial values, with a bias of 0.45 kPa and limits of agreement of 0.35 kPa to 0.54 kPa (Figure 1). In conclusion, measured arterial P50 and Hill slope values may be used to provide reasonable estimates of PcapO2 and PmitO2 during resting breathing, but are probably not reliable for this purpose when the configuration of the arterial oxyhemoglobin curve is physiologically modulated, as exemplified by voluntary hyperventilation.