A new model for the oxyhaemoglobin dissociation curve

This paper describes a new model for the oxygen-haemoglobin dissociation curve in humans. The model is based on the known structural alterations that occur in the quaternary haemoglobin molecule during oxygenation and deoxygenation. The two alternative structures, tense and relaxed, are described using hyperbolic tangent curves and linked with a probability function to obtain the completed mathematical description of the oxygen-haemoglobin dissociation curve. Model accuracy is assessed by a bias/precision analysis of calculated logit (S) and P50 against gold standard data. A mechanism for the transition between the two structures involving the chloride ion as a major allosteric effector is proposed. Results were analysed against the Siggaard-Andersen model for bias, precision and calculated P50 in four saturation ranges--0.00 < SO2 < 1.00, 0.20 < SO2 < 0.80, 0.90 < SO2 < 1.00 and 0.97 < SO2 < 1.00. In each range except for 0.20 < SO2 < 0.80, bias, precision and calculated P50 for the new model are significantly better (P < 0.05). Analysis of calculated P50 across the entire saturation range revealed significant drift out of the acceptable range in the Siggaard-Andersen model for SO2 > 0.92. The new model remained within tolerance across the saturation range 0.00 < SO2 < 1.00. The new model is significantly more accurate than the popular Siggaard-Andersen model, particularly in the range SO2 > 0.90.