Three different levels of hyperchloremia were induced in healthy Friesian calves to study the effects of chloride on blood oxygen transport. By infusion, the calves received either 5 ml/kg of 0.9% NaCl (low-level hyperchloremia; group A), 5 ml/kg of 7.5% NaCl (moderate hyperchloremia;group B), or 7.5 ml/kg of 7.5% NaCl (high-level hyperchloremia; group C). Blood was sampled from the jugular vein and the brachial artery. Chloride concentration, hemoglobin content, arterial and venous pH, PCO2 , and PO2 were determined. At each time point (0, 15, 30, 60, and 120 min), the whole blood oxygen equilibrium curve (OEC) was measured under standard conditions. Ingroups B andC, hyperchloremia was accompanied by a sustained rightward shift of the OEC, as indicated by the significant increase in the standard PO2 at 50% hemoglobin saturation. Infusion of hypertonic saline also induced relative acidosis. The arterial and venous OEC were calculated, with body temperature, pH, and PCO2 values in arterial and venous blood taken into account. The degree of blood desaturation between the arterial and the venous compartments [O2 exchange fraction (OEF%)] and the amount of oxygen released at tissue level by 100 ml of bovine blood (OEF vol%) were calculated from the arterial and venous OEC combined with the PO2 and hemoglobin concentration. The chloride-induced rightward shift of the OEC was reinforced by the relative acidosis, but the altered PO2 values combined with the lower hemoglobin concentration explained the absence of any significant difference in OEF (% and vol%). We conclude that infusion of hypertonic saline induces hyperchloremia and acidemia, which can explain the OEC rightward shift observed in arterial and peripheral venous blood.

  • oxygen affinity
  • red blood cells
  • calves
  • hyperchloremia

in experiments on bovine hemoglobin solutions, Fronticelli et al. (8) showed that chloride decreases the oxygen affinity of hemoglobin and may act in bovines as a physiological modulator of blood oxygen transport. Using whole tonometered blood, Gustin et al. (9, 10) showed more recently that under standard conditions (pH 7.4, PCO2 40 Torr, temperature 37°C), chloride modulates binding of oxygen to the red blood cells of adult and neonate bovines, shifting the oxygen equilibrium curve (OEC) to the right. The OEC is also influenced by other factors such as pH, PCO2 , temperature, and the 2,3-diphosphoglycerate concentration. From these data it appears that chloride could be a good pharmacological tool for shifting the OEC to the right in vivo in calves. As recently reviewed by Cambier et al. (2), hypertonic saline solutions have been used in clinical conditions, such as hemorrhaghic shock, endotoxic shock, and hypochloremic hypokalamic alkalosis, to improve cardiovascular function and blood biochemical paremeters. However, the influence of hypertonic saline solutions on the different stages of oxygen transport from the lung to the tissues remains unclear. To assess the chloride effects in combination with other biochemical changes induced in vivo by saline solution administration, we recorded the standard OEC in blood sampled from healthy calves with different levels of hyperchloremia induced by the administration of saline. We also measured temperature, the acid-base balance, the hemoglobin concentration, and PO2 and PCO2 in arterial ( PaO2 and PaCO2 , respectively) and venous blood ( PvO2 and PvCO2 , respectively) to assess the in vivo effects of chloride on blood oxygen release.


Experimental procedure.

Healthy male Friesian calves, 7–44 days old, were infused with saline via the jugular vein so as to obtain three different levels of hyperchloremia. Group A (low-level hyperchloremia) received 5 ml/kg of 0.9% NaCl, group B (moderate hyperchloremia) received 5 ml/kg of 7.5% NaCl, and group C (high-level hyperchloremia) received 7.5 ml/kg of the latter solution. Because of the decrease in the percentage of fetal hemoglobin during the first months of life in calves (10), the animals were selected to obtain a similar age range in the three groups. The infusion rate was in all cases 1 ml ⋅ kg−1 ⋅ min−1. At time 0, defined as the start of infusion, and at 15, 30, 60, and 120 min, we used heparinized syringes to collect, under anaerobic conditions, 1 ml of venous and 1 ml of arterial blood. These samples were taken from the jugular vein and brachial artery, respectively. The syringes were placed on ice, and the pH, PO2 , and PCO2 were measured immediately (AVL, Biomedical Instruments, Graz, Austria). The rectal temperature was measured to correct the blood-gas values and pH. At each time point, we also collected venous blood samples from the jugular vein into 20-ml syringes containing 0.1 ml heparin (Liquémine Roche, 5,000 IU/ml). Heparin is classically used as an anticoagulant for such research purposes because of its limited effect on the biochemical composition of blood (12). The blood was immediately stored at 4°C, and the determination of OEC was performed within 1 day after the blood was drawn from the animal, a period during which we observed no modification in control animals. Plasma was also stored at 4°C after 2 ml of venous blood were centrifuged for 15 min at 4,000 rpm for chloride determination.

Blood analysis.

The OEC of oxyhemoglobin was measured by a dynamic method under standard conditions (pH 7.4, PCO2 40 Torr, temperature 37°C) (3). A 15-ml blood sample was deoxygenated in a rotary tonometer with a gas mixture composed of 5.6% CO2-94.4% N2. The blood was placed in an analyzer and equilibrated with this first gas mixture. For 15 min, PO2 was slowly increased from 0 to 320 Torr by introducing a second gas mixture composed of 5.6% CO2-94.4% O2. Oxygen saturation was measured by photometry (LED, 660 nm) as a function of PO2 , the latter being measured polarographically ( PO2 electrode, Eschweiler, Kiel, Germany). Changes in plasma pH were corrected automatically (to pH 7.4) by addition of NaOH (1 N) or HCl (1 N) in the analyzer. A temperature of 37°C and a PCO2 of 40 Torr were maintained throughout the experiment. For each curve, 100 points were measured automatically; their PO2 and oxygen saturation coordinates were digitized, stored on a floppy disk, and processed on an IBM-compatible personal computer (HP Vectra, QS/16S, Hewlett-Packard). The computer reproduced the curves on a laser printer (Laser Jet Series 2, Hewlett-Packard) and processed the data. The accuracy of our method for measuring the OEC, expressed by the standard deviation of the PO2 at 50% hemoglobin saturation (P50), is 0.1 Torr for six curves recorded from the same blood sample (analytic error of the analyzer) and 0.3 Torr for 11 samples taken from the same control subject over a 30-day period (analytic error associated with intraindividual variations). Oxygen affinity changes were evaluated by measuring P50 under standard conditions (standard P50).

The hemoglobin concentration, expressed in grams per 100 milliliters of blood, was determined with an OSM3 radiometer (Radiometer).

Hematocrit was measured by microcentrifugation (Martin Christ), and mean corpuscular hemoglobin concentration was calculated by dividing hemoglobin concentration by hematocrit.

Chloride concentrations (mM) were determined by a titrimetric method by using Hg(NO3)2, as instructed in Merckotest no. 3311.

Arterial and venous pH and PaO2 , PaCO2 , PvO2 , and PvCO2 were measured with an AVL 995 (Biomedical Instruments).

Oxygen exchange fraction (OEF) calculation.

The OEF% is the difference in saturation between PaO2 and PvO2 values measured simultaneously, with the position and shape of the OEC in both compartments taken into account. In practice, the arterial and venous OEC values were calculated from the standard OEC corrected for the effects of pH, temperature (4), and PCO2 (unpublished observations), although the influence of the latter is of minor importance compared with the other correcting factors.

The amount of oxygen released in vivo by 100 ml of bovine blood at tissue level (OEF vol%) was calculated as followsOEFvol%=[Hb]×BO2×(OEF%/100)+α(PaO2PvO2) where Bo 2 is the hemoglobin oxygen capacity (1.39 ml O2/g Hb) (10), [Hb] is the hemoglobin concentration (g/100 ml), α is the oxygen solubility coefficient for blood at the temperature of the experiment (0.003 ml ⋅ 100 ml−1 ⋅ Torr−1), and PaO2 and PvO2 are in Torr.

Statistical analyses.

The values of parameters measured at time 0 are included in Table 1and are expressed as the means of individual values ± SE. Others results are expressed as the mean of the individual differences recorded between the value measured in blood drawn at the indicated time postinfusion and the value recorded at time 0[Σ( χti χt0 )/n]. Means are given with the standard error. All data were subjected to a normality test. The influence of time and treatment was assessed by two-way variance analysis. When statistical significance was reached, the following tests were done: within the same group, results were compared by using a paired Student’st-test; among groups, results were compared by using a nonpaired Student’st-test. Effects were considered significant when the P value did not exceed 0.05.

View this table:
Table 1.

Value of each parameter measured in each group at time 0


Table 1 shows the absolute value of each parameter measured in each group at time 0. No statistical differences were observed between groups at time 0, except for the hematocrit, the hemoglobin concentration, and, consequently, the amount of oxygen released in vivo at tissue level by 100 ml of bovine blood (OEF vol%). All parameters were within their physiological ranges (seediscussion).

Figure 1 depicts the effects of NaCl infusion on plasma chloride concentration after infusion. The level of hyperchloremia was directly related to the amount of chloride administered. In group A, significant but very limited hyperchloremia was observed at 60 min postinfusion. Ingroups B andC, hyperchloremia was statistically higher than in group A and was sustained to the end of the experiment.

Fig. 1.

Effects of NaCl infusion on plasma chloride (Cl) concentration after infusion. Each value is mean ± SE of differences (Δ) recorded between values measured after infusion and at time 0; n, no. of calves. The following treatments were applied: group A: 0.9% NaCl, 5 ml/kg; group B: 7.5% NaCl, 5 ml/kg; group C: 7.5% NaCl, 7.5 ml/kg. Significantly different from value measured attime 0 in same group: P < 0.05;★★ P < 0.01;★★★ P < 0.001. Significantly different from corresponding value measured ingroup A at the same time:Δ P < 0.05;ΔΔ P < 0.01;ΔΔΔ P < 0.001. Significantly different between groups B and C at the same time, †† P < 0.01.

Figure 2 shows the effects of NaCl infusion on the standard P50.Group A displayed no significant change. Moderate and high levels of hyperchloremia induced a rightward shift of the OEC, as illustrated by the significant increase in standard P50 recorded ingroups B andC. Although the magnitude of the shift did not differ significantly between groups B and C, the effects lasted longer in the latter group.

Fig. 2.

Effect of NaCl infusion on standard Formula at 50% hemoglobin saturation (P50 std) after infusion. Each value is mean ± SE of Δ between values measured at time indicated postinfusion and those measured at time 0; n, no. of calves. For applied treatments and statistical meaning of symbols, see legend for Fig. 1.

Body temperature was not significantly influenced by infusion and remained within the range shown in Table 1. Table2 shows the effects of NaCl infusion on pH in arterial and venous blood and on PaO2 , PaCO2 , PvO2 , and PvCO2 . The effects of NaCl infusion on hemoglobin concentration, hematocrit, and mean corpuscular hemoglobin concentration are also noted. In group A, no changes were observed in any of these parameters. In group B, the pH tended to decrease in both the arterial and the venous compartment. The only significant change was in venous blood at 120 min. In group C, relative blood acidosis was significant. Mild but significant hypercapnia was sometimes observed in both compartments. Decreased hemoglobin concentration and hematocrit throughout the experiment were observed in groups B and C. For mean corpuscular hemoglobin content, no changes were observed in any groups.

View this table:
Table 2.

Evolution of blood gases, pH, [Hb], Hct, MCHC, OEF%, and OEF vol% after infusion

Figure 3 shows the effects of NaCl infusion on P50 in venous and arterial blood. In group A, no significant change was observed. In group B, an increase in P50 was observed at 15 min postinfusion in both venous and arterial blood. Changes recorded ingroup C were of similar magnitude as in group B, but were more sustained, remaining significant even 120 min after the start of infusion.

Fig. 3.

Effects of NaCl infusion on P50 in venous ( Formula ) and arterial ( Formula ) blood after infusion. Each value is mean ± SE of Δ between values measured at indicated time postinfusion and those measured attime 0; n, no. of calves. For applied treatments and statistical meaning of symbols, see legend for Fig. 1. ND, not determined.

Figure 4 shows the effects of NaCl infusion on OEF%. Table 2 gives detailed data for each time point. No changes in OEF were observed because the increase normally induced by the rightward OEC shift described above was counterbalanced by the increase in PvO2 . Indeed, at the level of hemoglobin saturation corresponding to this PvO2 , the OEC slopes markedly. Thus, although limited and even not significant, the changes in PvO2 strongly influence the level of saturation of hemoglobin.

Fig. 4.

Evolution of arterial oxygen equilibrium curve (OECa), venous oxygen equilibrium curve (OECv), and oxygen exchange fraction (OEF%) after infusion of 7.5% NaCl, 7.5 ml/kg. So 2, O2 saturation.A: mean reference values measured attime (t) 0 in 6 animals.B: mean values measured in same animals 30 min postinfusion. Formula , Formula , and venous ( Formula ) and arterial Formula ( Formula ) are also shown.


By infusion of saline, we have induced three levels of hyperchloremia in healthy calves to study the effects of chloride on blood oxygen transport. To obtain the highest hyperchloremia level, we used the same saline solution as for moderate hyperchloremia (7.5% NaCl), but more of it (7.5 ml/kg instead of 5 ml/kg), because it is well known that administration of more concentrated saline solutions induces convulsions and mortality (18, 20).

The method used here to plot the OEC is a proven technique (3), but, because the curves were recorded up to 1 day after the blood was drawn, we performed two checks to make sure that the delay had not affected the results obtained. First, OECs remained unaltered over this period, as illustrated by preliminary data obtained with blood sampled in older cattle. Indeed, when measured immediately after blood sampling, the mean standard P50 was 26.6 ± 0.4 Torr; when measured 24 h later, it was 26.5 ± 0.4 Torr (n = 22). Second, the mean chloremia value obtained was also unaltered by delayed measurement, being 100.7 ± 0.8 mmol/l immediately after sampling and 101.6 ± 0.8 mmol/l 24 h later (n = 31).

All blood parameters recorded at time 0 are within their physiological ranges (6, 10, 11, 14,19), suggesting that all calves in the study were healthy. The hemoglobin and the hematocrit values were lower ingroup C than in group A, but all values remained within the physiological range, which are 6–12 g/100 ml (19) and 24–46% (11), respectively. Because no physiological values are available in the literature for P50 in arterial and venous blood or for the parameters derived from it (OEF%, OEF vol%), no comparison is possible.

The data in Fig. 1 show that the level of chloremia can be adjusted from low to high by varying the concentration and volume of saline administered. The levels of chloremia obtained are in agreement with those measured in endotoxemic calves infused with hypertonic saline (7% NaCl, 4 ml/kg; baseline value 103 vs. 118 meq/l 15 min after infusion) (5). As shown in Fig. 2, calves with moderate to high hyperchloremia (groups B andC) display an increased standard P50. Four hypotheses might explain this modification. First, because only infusion of hypertonic saline affects the standard P50, the OEC shift might be due to hypertonicity. This hypothesis can be rejected on the basis of previous in vitro data showing that hypertonicity obtained by adding sucrose to bovine blood does not affect the OEC (9). These data are in agreement with those obtained by Murphy et al. (13) in vitro with human blood, demonstrating that the changes in P50 were not related to the increase in the mean corpuscular hemoglobin concentration. Moreover, the fact that no significant changes in mean corpuscular hemoglobin concentration were recorded during the experiment suggests that red cell shrinkage did not occur after NaCl infusion in our experiment. A second explanation might be the dilution effect due to infusion itself, but, to date, no physiological mechanism accounting for such an effect has been described. Furthermore, although groups A and B received the same volume of saline, the postinfusion evolution of the standard P50 was markedly different in the two groups. The possible changes in the intracellular pH due to NaCl infusion could also explain the OEC rightward shift. To check this specific point, complementary in vitro experiments were performed, using the method described by Murphy et al. (13) for intracellular pH determination. By increasing the plasma chloride concentration in tonometered blood with a pH of 7.4, at a level similar to that recorded in vivo (Δ = +20 mmol/l plasma) or higher (Δ = +60 mmol/l plasma), the intracellular pH values measured after chloride addition were, respectively, 7.261 ± 0.007 and 7.244 ± 0.001 vs. 7.255 ± 0.003 in control (n = 4). These data suggest that the effect of NaCl infusion on the standard P50 is not related to a Bohr effect. Finally, hyperchloremia itself could explain the OEC rightward shift recorded in groups B andC. Several in vitro studies support this hypothesis. Using hemoglobin solutions, Fronticelli et al. (8) showed that bovine hemoglobin is more sensitive to chloride ions than is human hemoglobin, in terms of the concentration-dependent effect of this ion on P50. This doubtless reflects the different amino acid composition of the two hemoglobins (7, 8). According to Perutz and Imai (15), a valine and a histidine present in the human β-chain are replaced by a methionine in bovine hemoglobin. In ruminant β-chains, moreover, an arginine at the place of a lysine present in human hemoglobin should be a better ligand for chloride than lysine (7). Although the OEC rightward shift induced by chloride in bovine and human hemoglobin has been confirmed by Perutz et al. (16), the mechanism proposed to explain this effect was different from that suggested by Fronticelli (7). Indeed, it has been shown that chloride reduced the oxygen affinity of mammalian hemoglobin by acting as an allosteric effector, neutralizing the electrostatic repulsion in the cavity through the center of the molecule but without binding to any specific sites (1, 16, 17). Because hemoglobin functions differently in an artificial solution than in red blood cells, we recorded the whole blood OEC under standardized conditions with and without added chloride. We thus demonstrated that chloride significantly shifts the OEC to the right, likely by an action on hemoglobin (9). A higher chloremia level appears to cause a more sustained rightward shift of the OEC. The influence of the nature of the cation associated to chloride was not investigated in this experiment. However, in a previous study, it has been shown that KCl and NaCl induced a similar shift of the OEC in vitro (9).

The OECs for arterial and venous blood were calculated from the standard OEC corrected for the effects of pH, PCO2 , and temperature, which can be accurately predicted by appropriate mathematical models (4). It is well known that 7.5% NaCl induces acidosis in all investigated animal species (2), a change that can reinforce the OEC rightward shift via the Bohr effect. The increase in PCO2 also contributes to the rightward shift through its direct influence on hemoglobin. This effect was limited in this study, however, because hypercapnia was not significant at most sampling times.

Although hyperchloremia significantly affected the venous and arterial P50, it did not significantly alter the OEF between the arterial blood and the venous blood collected in the jugular vein. Moreover, because of the fall in hemoglobin concentration, mainly due to the osmotic action of hypertonic saline, some water moving from the extravascular space into the vascular compartment, the OEF vol% tended to decrease. Because one can assume that the oxygen consumption in resting healthy calves remains constant, this tendency must be likely counterbalanced by a moderate increase in the cardiac output, as previously described in endotoxemic calves and healthy pigs and dogs infused with hypertonic saline (2). As a consequence of the new position of the OEC, a new equilibrium between oxygen dissolved in plasma and oxygen bound to hemoglobin occurs, leading to a tendency for PvO2 to be increased. Experiments should be conducted on hypoxic calves to assess the overall influence of hypertonic saline infusion on oxygen consumption, taking into account the effects on blood oxygen affinity, cardiac output, and hemoglobin concentration. Our data demonstrate, however, that in vivo administered hypertonic saline causes the OEC to shift to the right and could thus be a good pharmacological tool for modulating blood oxygen binding in calves.


This work was supported by the Fonds National de la Recherche Scientifique and by the Ministère des Technologies Nouvelles de la Région Wallonne (grant nos. 3012 and 2555).


  • Address for reprint requests: P. Gustin, Dept. of Pharmacology and Toxicology, Faculty of Veterinary Medicine, Univ. of Liège, Boulevard de Colonster B 41, B-4000 Liège, Belgium.


View Abstract