It is commonly considered that there are associations between potassium and hydrogen ion metabolism. One such is an association between K+ states and alkalosis. It is often stated as a cause and effect relationship, i.e. that the low K+ states will produce an metabolic alkalosis. This hypothesis has been refuted a number of times. An alkalosis can occur without a K+ abnormality (acute pyloric obstruction) and a stable K+ abnormality without an alkalosis (Kassirer et al, (b) 1966; Jones et al, 1982). In a K+ deficient state an alkalosis can develop as the K deficiency is corrected (Belich et al, 1966) and conversely a metabolic acidosis can be induced by means of K+ deprivation with a resulting K+ deficiency (Burnell et al, 1974). An alkalosis causes renal loss of K but the low K state (Kassirer et al, 1966a) does not per se seem to cause an alkalosis. (This is well discussed in review articles by Schwartz et al, 1968 and 1978).

In any case, if an alkalosis and a low K state co-exist, both have to be corrected. Potassium salts of organic acids will not correct the alkalosis associated with a potassium deficiency (Aber et al, 1962). Sodium chloride will correct the alkalosis (Kassirer et al (b) 1966) but will not correct the potassium deficiency, i.e. the potassium deficiency must be corrected and the alkalosis must be corrected. Potassium chloride is, therefore, usually the correct management together with sodium chloride.

Another common idea is that pH changes produce or are consistently associated with changes in serum potassium concentrations. A fall in pH is said to raise the potassium concentration and visa versa. Such alterations can be induced by for example some acids but not others. Changes which do occur differ in magnitude. Factors, which may alter potassium balances will often act at the same time as pH changes making it impossible to decide if the pH alteration or something else has altered serum potassium. In their concluding remarks in a review of the subject Adrogue and Madias (1981) state that:

a) Wide differences in plasma potassium (occur) among various acute acid-base disorders.

b) Blood pH appears to be but one of many factors altering plasma potassium. See Appendix A8.


Appendix A.8 Potassium and pH

In the review by Schwartz et al, 1968, it is stated that there are approximately ten (10) well documented cases of saline-resistant alkalosis in the literature. I have examined the references quoted, but I did not think that any of the cases would now be considered to be well documented. They were obviously very ill patients with severe nutritional disorders but without a preconceived belief that K+ and pH are causally related, the idea would not be suggested by the data.

Garella et al (1970) present four more cases but again these are very complex cases. There are large gaps in the data due to failure of collecting specimens, assessment of electrolyte contents of stools, evidence of adequacy of inherent renal function as distinct from any functional and possibly reversible effect of K+ deficiency etc. They speculate that "Four cases observed in a one (1) year period at a single hospital" also suggests that this condition may be commoner than previously appreciated. I know a case in which it was estimated by balance studies that a K+ deficiency of the order of 3000meq was present. In this case there was a small acisosis.

Another example of the non-interdependence of K+ and pH changes is presented by Oster et al (1978). They infused HCl and lactic, b hydroxybutyric and methylmalonic acid into animals. The HCl acidosis was associated with a rise in K+ concentration in plasma whereas the organic acid infusions were not. The authors concluded that any high K+ in diabetic acidosis must be due to something other than a direct effect of keto acids.

I think that there is a common idea following the reports in the 1940's and 1950's that all hypokalaemias cause metabolic alkalosis. I think that the balanced view at present should be that low potassium states are essentially independent of pH states. The concept of the casual connection between the two was accepted blindly (i.e. on the say-so of well known authorities). When data appeared that should have demolished the idea, ad hoc hypotheses were created to save the concept.

The process of making ad hoc hypotheses is of great interest in the philosophy of science. Following is a summary of an amusing example which appears in Lakatos and Musgrave, 1974, p.100. There is an apocryphal planet which deviates from its calculated path so a new small planet is postulated. This is too small to be seen so a new larger telescope is financed by a grant. When the planet is still not seen, a cloud of cosmic dust is postulated as hiding the new planet. This is looked for by a new satellite. When the cloud also is not seen, new suggestions such as inadequacy of the instrumentation are made etc. etc.