It has been known for many years that hemorrhagic shock causes metabolic acidosis. In the present model, a prolonged metabolic acidosis associated with a transient increase in AG after shock induction was observed but was not adequately accounted for by the concomitant hyperlactatemia. In addition, the SIG increased significantly after induction of shock.
The physicochemical approach to acid-base balance originally described by Stewart [6
] and subsequently modified by Watson [15
], Fencl and Rossing [16
], and Figge and colleagues [13
] has become common in the last decade [18
]. According to this approach, the dissociation equilibrium is supplemented with equations incorporating the necessity for electrical neutrality and the principles of conservation of mass. Weak acid concentrations (albumin and phosphate), the pCO2
, and the SID have been identified as variables with independent effects on pH [6
]. Two different methods of calculating the SID exist. The first, leading to the apparent SID (SIDa
), relies on simply measuring as many strong cations and anions as possible and then summing their charges. The second, yielding the effective SID (SIDe
), estimates the SID from the pCO2
and the concentrations of the weak acids [27
]. The difference between SIDa
has been termed SIG and attains a positive value when unmeasured anions are present in excess of unmeasured cations and attains a negative value when unmeasured cations exceed unmeasured anions [7
In the present study, a negative SIG obtained at baseline indicates an excess of unmeasured cations. However, it should be noted that the baseline values were established after surgical preparation and infusion of large amounts of a crystalloid solution, resulting in electrolyte concentrations with particularly high serum chloride levels. Therefore, for graphical depiction, we used relative values representing increments and decrements in SIG and AG.
The data from the present study strongly suggest that large amounts of unmeasured anions, expressed either as the AG or as the SIG, are likely to be generated during states of global tissue hypoxia. This finding is in line with results of Kaplan and Kellum [28
], who reported increases in SIG in patients with major vascular injury, a condition generally associated with global tissue hypoperfusion. Also, in a study investigating the cause of the metabolic acidosis after cardiac arrest, Makino and colleagues [29
] showed that increases in SIG contributed approximately 33% to the metabolic acidosis.
With regard to the source of unmeasured anions, one can only speculate. An increased SIG appears to occur in patients with renal [30
] and hepatic [7
] impairment, and unexplained anions have been shown experimentally to arise from the liver in animals challenged with bolus intravenous endotoxin [31
]. In our canine model of hemorrhagic shock, serum concentrations of citrate were significantly increased after shock induction. This is in accordance with a recent finding of Forni and colleagues [32
], who found elevated levels of anions usually associated with the Krebs cycle in patients with large AG acidosis. Citric acid, a tribasic acid, is reported to be 97% ionized at a pH of 7.0 [33
]. Thus, each molecule of citric acid adds three protons to a solution upon ionization, and the contribution of citrate to the generation of unmeasured anions is of much greater significance than is apparent from its molarity.
We believe that the source of citrate is the mitochondria. The rate of oxygen delivery to respiring tissue plays a role in generating citrate, with several authors suggesting that tissue hypoxia can cause an increase in intermediates of the citric acid cycle [33
]. In further support of the mitochondrial origin of many of the unmeasured anions, fumarate and α-ketoglutarate, both metabolites of the Krebs cycle (like citrate), were identified in the dog sera in this study. Though not detectable by means of capillary electrophoresis at baseline, both metabolites were found in all dog sera after the induction of shock and until the end of the experiment. All of the Krebs cycle intermediates investigated as possible candidates for unmeasured anions have acidic dissociation constants guaranteeing full dissociation at a pH of 7.4 (Table ).
pK values of Krebs cycle intermediates
Another potential source of citrate might have been the stepwise reinfusion of whole blood during the standardized induction of shock given that the blood was stored with a CPDA solution. This blood contained approximately 12 mmol citrate per liter, and amounts ranging from 0 to 360 mL were given. However, there was absolutely no correlation between the individually reinfused volume of blood and the level of citrate found afterward in blood samples taken following induction of shock (results not shown). Since citrate changes in serum are a balance between endogenous production, exogenous load, and liver metabolism, a contribution of exogenous citrate to the changes in SIG cannot be ruled out totally.
The rise of acetate during induction of shock is not really surprising. Irrespective of the type of energy-yielding substrate (sugars, amino acids, and fats), oxidative utilization always passes via degradation to acetate, which is then coupled to coenzyme A. Hydrolytic cleavage of acetyl-coenzyme A back to acetate will occur when there is a block in mitochondrial consumption of this thiol-ester. Thus, increased acetate also supports the assumption that mitochondrial dysfunction was caused by the hemorrhagic shock.
Serum concentration of urate also increased significantly after shock induction. This is in excellent agreement with induction of a state of catabolism of high-energy adenine and guanine nucleotides during shock. The rise in urate supports the presumed damage to hepatic metabolism because urate is normally degraded to allantoin in the dog liver. However, the concentrations of this metabolite in dog serum, as befits a non-primate species, were much too low to account for the changes in SIG.
The healthy vascular endothelium is coated by a large variety of extracellular domains of membrane-bound molecules, which together constitute the glycocalyx. Heparan sulfate is a polysulfated polysaccharide that is linked to core molecules of the endothelial glycocalyx. Shedding of these polyanionic heparan sulfates might be another potential source of unmeasured anions, and, indeed, our group has recently demonstrated acute destruction of the endothelial glycocalyx in humans experiencing ischemia and reperfusion injury [35
]. The present study also indicates shedding of heparan sulfate after hemorrhagic shock. However, this did not parallel the changes in SIG (Figures and ). After alkaline hydrolysis of serum, sulfate anions, already present in canine serum at levels of approximately 0.7 mM, did not change enough to account for much of the changes in SIG.
The impact of different variables on the acid-base status during induction of shock is shown in Figure . The values represent the difference between the time points at baseline and in shock. Interestingly, changes in AG and SIG were the strongest determinants of acidemia, accounting for -11.0 and -7.1 mEq/L of acidifying effect, respectively. An increase in lactate concentration contributed -4.0 mEq/L to acidemia. Changes in phosphate, magnesium, sodium, and chloride each accounted for less than -0.5 mEq/L of acidifying effect. The acidemia was attenuated by alkalinizing changes of several variables. A decrease in albumin concentration had the strongest alkalinizing effect (+1.3 mEq/L). Increases in potassium and calcium concentration were of minor importance (less than +0.4 mEq/L). The increases in citrate (-2.2 mEq/L), acetate (-2.2 mEq/L), and sulfate (-0.1 mEq/L) concentration together accounted for approximately 63% (-4.5 mEq/L) of the increase in SIG during induction of shock. The net balance yields a deficiency of unidentified anions amounting to approximately 2.6 mEq/L.
Figure 6 Impact of different variables on the acid-base state during induction of shock. Values (mean ± standard error of the mean) (n = 8) are presented as the difference between the time points of baseline and shock. A negative value represents an increase (more ...)
Outcome prediction based on the quantitative approach remains controversial. Some investigators have found that the pH and the standard base excess are better outcome predictors than the SIG [36
]. However, other investigators have found that the SIG is a powerful predictor of outcome in acutely ill or injured patients. In critically ill patients, SIG was a strong independent predictor of mortality when it was the major source of acidosis [8
]. Also, in patients with major vascular injury [28
] and in children following cardiopulmonary bypass surgery [27
], an elevated SIG appeared to be superior to other conventional mortality predictors.
Growing evidence suggests that extracellular acidosis itself has profound effects on the host, particularly in the area of immune function. It is now becoming apparent that different forms of acidosis and even different types of metabolic acidosis produce different effects [38
], and SIG generation may be one feature.
Fluid resuscitation might have affected the SIG in the present model of hemorrhagic shock, although only hydroxyethyl starch solutions were given. The colloid molecule itself may be a weak acid. Albumin and gelatin preparations contain a weak acid activity [20
]. Gelatins have been shown to increase both AG and SIG, most likely due to their negative charge and relatively long circulating half-life [40
There are several limitations in this study. First, we were not able to find a strict correlation between SIG and the serum concentrations of citrate and/or acetate. This is not entirely unexpected since the generation of the SIG is most probably multifactorial. Second, we have used capillary electrophoresis for identification of potential candidates, and concentrations of still-unknown metabolites may be below the level of detection of this method. Third, using Stewart's approach to acid-base balance has some limitations. A major criticism is a possible inaccuracy of determinations of plasma electrolyte concentrations. Such inaccuracy means that the calculation of the SIDa
, AG, and SIG can be erroneous [41
]. If, as in our study, mean values of larger collectives are used, always with utilization of the same measurement techniques for determinations of electrolytes, these limitations should be insignificant.