A paper which changed clinical practice (slowly). Jacob Holler on potassium deficiency in diabetic acidosis (1946).

It is often said that the introduction of insulin into clinical medicine made a 'dramatic' difference to the mortality resulting from diabetic coma. This is true in the sense that before 1922 it was almost uniformly fatal, but until the 1950s the mortality in many large hospitals was as high as 30-50%. Often autopsy did not establish a cause of death. Many may have been a result of hypokalaemia, a complication which was not recognized until 1946; in that year in the Journal of the American Medical Association, Jacob Holler described a patient who developed respiratory paralysis 12h into treatment that, after several hours in an iron lung, was cured by potassium infusion. In the 5 years after Holler's paper there were many reports of deaths resulting from hypokalaemia, as well as several 'near misses', but clinicians were extremely cautious about early replacement probably, as an editorialist in The Lancet suggested, because 'the frightening effects of intravenous injections of potassium made clinicians reluctant to believe in a lack of potassium as a cause of trouble, except in very rare conditions such as familial periodic paralysis'. It had been known since 1923 that insulin lowered serum potassium, but this was not of great interest because the symptoms of hypokalaemia were not known. Also, potassium was not an electrolyte with which clinicians were familiar. Until the introduction of flame photometry in 1950, it was only measured in research studies as chemical methods took several hours to complete.