Evaluating acid-base disorders

Sir, In their review of acid-base disturbances in cholera, Zalunardo et al. illustrate the limitations of the conventional approach when dealing with complex clinical situations.1 The patient with cholera was thought to have both contraction metabolic alkalosis and metabolic acidosis from bicarbonate losses in diarrhoea, with the combination leading to a relatively normal arterial pH, PCO2, and plasma bicarbonate concentration. A different approach to these issues was introduced by Stewart2,3 and supported by Fencl4 and Kellum,5 among others. In this development, acid-base disorders can be considered to arise from alterations in three independent variables: the plasma strong ion difference (SID), the plasma concentration of weak acids (in most situations the plasma concentration of albumin and phosphate), and the arterial partial pressure of carbon dioxide (PCO2). From the data provided, an alternative explanation for the clinical problem presented would be that the relatively normal pH, PCO2, and plasma bicarbonate concentration result from a combination of metabolic alkalosis resulting from an elevated SID, offset by a metabolic acidosis from the markedly increased plasma protein concentration. While the marked reduction in ECF helps to explain the acid-base abnormalities identified, it is by no means central to the Stewart approach. When there is a reduction of water in plasma the SID will increase: this high SID metabolic alkalosis may be independent of extracellular volume, so the term ‘contraction alkalosis’ is not entirely correct in all situations. I thank Dr Cole for his interesting letter. Each of his points concerning the Stewart approach, however, merits comment. It is true that the anion gap approach will not reveal the pathophysiology of the critical acid-base disturbances in this patient. The point to emphasize, however, is that all approaches that rely on information derived solely from measurements on arterial blood cannot overcome this limitation—this includes the Stewart approach. As regards the alternative explanations offered by the Stewart approach: (i) ‘Metabolic alkalosis resulted from an elevated strong ion difference’. The strong ion difference describes concentration differences; this cannot reveal whether the most important mechanism for metabolic alkalosis (higher plasma \(\mathrm{HCO}_{3}^{{-}}\) concentration and a higher pH) was a surplus of \(\mathrm{HCO}_{3}^{{-}}\) in the ECF compartment or a decrease in the ECF volume. Clearly this patient had a deficit, not a surplus, of \(\mathrm{HCO}_{3}^{{-}}\) in his ECF compartment. (ii) The ECF volume is not central to the Stewart approach. This is very unfortunate because the major factor elevating this patient's plasma \(\mathrm{HCO}_{3}^{{-}}\) concentration was a contracted ECF volume. Hence, contrary to Dr Cole's last line in his letter, the Stewart approach is incorrect if the interpretation was that this patient's metabolic alkalosis was ‘independent of his ECF volume’. (iii) Metabolic acidosis was due to the increased plasma protein concentration. Again the problem here is that one relies solely on concentrations in the Stewart approach. There are two ways to raise the concentration of albumin in plasma. First, if isoelectric albumin were added, this would cause a modest H+ load (∼80 mEq), if one added enough albumin to double its content in the ECF compartment (this virtually never occurs in clinical medicine). Second, if the basis for the rise in albumin concentration were a lower ECF volume (a much more frequent occurrence), one would not have an important H+ load (minor changes are possible due to changes in ionic strength). Hence it is not correct to attribute metabolic acidosis to a rise in albumin concentration, unless one knew that this rise was due to the administration of a sufficient quantity of salt-poor albumin.1,2 This leaves one other issue—to point out what was missing in the traditional and Stewart approaches, due to the fact that they do not include measurements of venous blood. These latter data permit the clinician to recognize when a given H+ load is particularly dangerous, as described below. When one only examines the arterial PCO2, one knows about the stimulus to alveolar ventilation and the capacity to increase alveolar ventilation. This information, however, fails to reveal whether buffering of H+ was optimal by the bicarbonate buffer system (BBS) in cells.3 When the BBS does not operate optimally because of a high cellular PCO2, more H+ will bind to intracellular proteins, changing their charge, shape and possibly function, especially in vital organs: we called this a ‘tissue form of respiratory acidosis’. The PCO2 in cells of a given venous drainage system is revealed by examining the PCO2 in venous blood.4 In summary, there are two major deficits in the traditional and the Stewart approach, but both could be markedly improved with two simple, additional measurements. First, one needs to obtain a quantitative estimate of the ECF volume at the bedside. This information could be obtained with sequential values for the haematocrit4 and/or total proteins.5 Second, venous PCO2 should be measured and compared to arterial PCO2. Not only will this help in diagnosis, but it can also be used to guide therapy.4 Moreover, a quantitative estimate of the ECF volume was particularly valuable in this patient, because it revealed that there was a large deficit of \(\mathrm{HCO}_{3}^{{-}}\) in the ECF compartment, which was a danger sign about a decision to infuse isotonic saline at a rapid rate to improve haemodynamics. The danger was that the PHCO3 would decline markedly for three important reasons (dilution, enhanced diarrhoea, and reversal of the production of \(\mathrm{HCO}_{3}^{{-}}\) secondary to the anticipated fall in the venous PCO2, Figure 9 in our paper). The increased severity of the acidosis might exacerbate the peripheral arterial and venous vasoconstriction, leading to a long delay in the distribution of the infused saline into the entire ECF compartment. This could cause central blood volume expansion and predispose the patients to develop pulmonary oedema before the ECF was re-expanded to normal values.6