How Oxygen Changes Blood Chemistry: New Equation Decodes Hidden Electrolyte Shifts

Every day in intensive care units around the world, clinicians order blood gas and electrolyte panels multiple times per patient, using the results to fine-tune ventilator settings, interpret acid–base disturbances, and guide life-saving decisions. Among the values derived from these tests is the plasma strong ion difference, or SID, the gap between fully dissociated cations and anions in the liquid fraction of blood. A falling SID signals metabolic acidosis; a rising one suggests alkalosis. But a new study published in Physiological Reports reveals that a substantial portion of SID variation has nothing to do with true acid–base pathology at all. Instead, it arises from a century-old physiological phenomenon, the Haldane effect, in which oxygen binding to hemoglobin physically rearranges electrolytes between red blood cells and plasma. The researchers have now, for the first time, quantified this effect precisely enough to predict it with a simple bedside equation.

The problem the team set out to solve is rooted in the architecture of blood itself. Whole blood is a bi-compartment system: red blood cells occupy a large fraction of blood volume and harbor an enormous reservoir of electrical charge, while the plasma carries the electrolytes clinicians actually measure. The erythrocyte membrane, however, is permeable to water and to selected strong ions, chiefly chloride and lactate. When pH shifts, or when hemoglobin loads or unloads oxygen, these ions migrate across the membrane, changing plasma SID without any net change in the total charge pool of the blood. Such redistribution is a purely physicochemical event, not a primary acid–base disturbance, yet failing to recognize it can lead clinicians to misdiagnose a patient’s acid–base status entirely.

Earlier work by the same group had produced a simplified equation predicting how pH-driven electrolyte shifts alter plasma SID, based on the titratable charges of albumin and hemoglobin, or on the whole-blood buffer power. That model improved bedside interpretation of base excess, but it carried a critical assumption: fully oxygenated blood. In real clinical life, oxygen saturation is anything but constant. Venous blood sampling, extracorporeal gas exchange, and acute changes in pulmonary gas transfer all produce large swings in hemoglobin oxygen saturation, and each swing triggers the Haldane effect through two distinct mechanisms. First, oxygen binding lowers the dissociation constant of certain weak amino acid residues on hemoglobin, reducing its buffering capacity. Second, oxygenated hemoglobin holds chloride less tightly, encouraging chloride and water to move out of the red cell and into the plasma.

To quantify these saturation-dependent effects, the researchers re-analyzed data from an in-vitro study in which venous blood from 20 healthy volunteers, aged 32 to 65, was equilibrated with room air. The design was elegantly simple. Nine-milliliter venous samples were divided into three aliquots: one left undiluted, and two mixed with a buffer solution at varying blood-to-buffer ratios, generating a broad range of hemoglobin and albumin concentrations. Each sample was analyzed for acid–base variables, then exposed to room air through repeated aspiration and rotation over roughly twenty minutes, a procedure that simultaneously oxygenated the blood and washed out carbon dioxide. After equilibration, all samples were re-analyzed, and plasma albumin was measured by centrifugation and immunoassay.

The results were striking. Equilibration with room air raised the oxygen saturation fraction by a median of 0.46, dropped carbon dioxide tension by 29 mmHg, and lifted pH by 0.25 units. Alongside these changes, plasma SID fell by a median of 5.3 mEq/L, driven by both a decrease in strong cations and an increase in strong anions, with sodium declining, chloride rising, and ionized calcium and lactate shifting as well. The drop was more pronounced in less diluted samples, hinting that the protein content of blood, particularly hemoglobin, governs the magnitude of the redistribution. Linear mixed-effects models confirmed that changes in oxygen saturation, changes in carbon dioxide tension, and the whole-blood buffer power were each independently associated with the change in SID.

With these determinants established, the team built and tested predictive equations. An extended model accounted first for pH-dependent charge changes on albumin and hemoglobin, then added a pK-dependent saturation component reflecting oxygen’s effect on hemoglobin’s buffering groups, and finally a chloride-dependent component capturing oxygen’s reduction of hemoglobin’s chloride affinity. Bland–Altman analysis showed that considering pH effects alone underestimated SID changes, with the bias worsening as oxygen saturation rose. Adding the pK-related effect improved matters; adding the chloride-related effect brought the mean bias to just −0.4 mEq/L, with limits of agreement from −2.0 to 1.3 mEq/L, and completely eliminated the dependence of the error on saturation change. In other words, both arms of the Haldane effect must be accounted for to predict plasma SID accurately.

Recognizing that intensive mathematics is impractical at the bedside, the researchers also distilled their findings into a remarkably simple formula. Using a rounded whole-blood buffer power of 1.5 times the sum of hemoglobin and albumin concentrations in grams per deciliter, plus a rounded saturation coefficient equal to hemoglobin concentration divided by four times the change in oxygen saturation, the expected change in SID can be computed with mental arithmetic. This simplified model slightly overestimated SID changes, with a mean bias of 0.8 mEq/L, but the error no longer grew with larger saturation shifts, and the limits of agreement, −0.9 to 2.4 mEq/L, remained within the range considered clinically acceptable. Over the observed range of SID changes approaching 10 mEq/L, the maximal discrepancy between predicted and measured values was roughly 2 mEq/L, comparable to the analytical variability of routine electrolyte measurements.

The simplified equation also carries mechanistic insight. Its rounded saturation coefficient closely matches correction factors proposed decades ago by Siggaard-Andersen and later adopted by Zander to adjust whole-blood base excess for measured oxygen saturation, lending independent support to the new calculations and informing ongoing debate about arterio-venous differences in base excess and SID. The slight overestimation of the simplified model stems from its assumption of a pH-invariant buffer power: in the alkalotic range reached during the experiment, buffer power actually declines as pH moves away from the pK of histidine residues, an effect the extended model captures but the rounded coefficient does not. Conversely, the extended model omits phosphate species, which likely explains its progressive underestimation in undiluted blood rich in hemoglobin and albumin.

The clinical implications could be immediate. Emergency medicine often relies on venous blood gas analysis, in which measured oxygen saturation differs substantially from arterial values, before any arterial catheter is placed. The authors demonstrate with a real clinical case that only when both pH- and saturation-dependent redistribution are considered does the venous sample yield a correct acid–base diagnosis, one later confirmed by arterial blood gas analysis. Neglecting these shifts risks misdiagnosis and potentially harmful management decisions. The study does carry limitations: magnesium and phosphate were not measured, the shared buffer coefficient may falter when albumin and hemoglobin concentrations are extremely imbalanced, and the saturation coefficients were derived from healthy volunteers, so extrapolation to disease states requires caution. Global desaturation in respiratory failure may also dilute SID changes across the entire extracellular space, producing smaller plasma effects than the in-vitro model predicts. Still, the core message stands: the hidden choreography of ions between red cells and plasma, long recognized qualitatively, can now be predicted with a few numbers and a moment of arithmetic, turning a source of diagnostic confusion into a quantifiable, manageable variable.

Subject of Research: Quantifying hemoglobin oxygen saturation effects on plasma strong ion difference via the Haldane effect

Article Title: Quantifying the effects of hemoglobin saturation on plasma strong ion difference during blood oxygenation and decarboxylation

Article References: Giosa, L., Krbec, M., Halamík, J., Bolnberger, A., Busana, M., Brusatori, S., Diemer, T., Kristensen, S. R., Heldeweg, M. L. A., Zadek, F., Rees, S. E., & Duška, F. (2026). Quantifying the effects of hemoglobin saturation on plasma strong ion difference during blood oxygenation and decarboxylation. Physiological Reports, 14(19), Article e71118. https://doi.org/10.14814/phy2.71118

Image Credits: AI Generated

DOI: 10.14814/phy2.71118

Keywords: strong ion difference, Haldane effect, hemoglobin oxygen saturation, acid–base physiology, electrolyte redistribution, red blood cells, chloride shift, buffer power, blood gas analysis, base excess, critical care, physiological Reports