Acid-Base Physiology

Case 10 : A semi-comatose diabetic on diuretics


A 55 year old insulin dependent diabetic woman was brought to Casualty by ambulance. She was semi-comatose and had been ill for several days. Past history of left ventricular failure. Current medication was digoxin and a thiazide diuretic.

Results include: K+ 2.7, glucose 67, anion gap 34 (all in mmol/l)

Arterial Blood Gases

pH  7.41

pCO2 32 mmHg

pO2   82 mmHg

HCO3   19 mmol/l


Firstly: Initial clinical assessment

This lady is a known diabetic and she presents with mental obtundation and severe hyperglycaemia. The clinical diagnosis suggested by the history is diabetic ketoacidosis (DKA) or hyperosmolar non-ketotic coma (NKHC). There are several other points to always be aware of in diabetic patients:

A patient with an "obvious diabetic ketoacidosis" should always be approached with these considerations in mind.

Secondly: The acid-base diagnosis

  1. pH: The pH is normal. So we conclude that there are 3 possibilities:
    • No acid-base disorder is present
    • There are compensating disorders present (ie acidosis and alkalosis together)
    • This is a rare case of a fully compensated disorder (This is considered very unusual but may occur in some circumstances eg during recovery of the primary disorder and before the level of compensation can adjust, and perhaps during pregnancy where the compensation for the respiratory alkalosis can return the pH into the normal range in some cases.)
  2. Pattern: Both the pCO2 & the bicarbonate are lower than normal so this disproves the option of no acid-base disorder. Considering the option of at least 2 compensating disorders then we would initially suspect a metabolic acidosis (low bicarbonate) and a respiratory alkalosis (low pCO2).
  3. Clues: The glucose is very high confirming diabetic decompensation, either DKA or NKHC). The absence of any results from urine testing is very unfortunate. The presence or absence of ketones would be valuable information. The high anion gap indicates that there is a severe high anion gap acidosis. Indeed this is much higher then expected given the relatively modest decrease in bicarbonate. This alerts us to the possibility of a pre-existing metabolic alkalosis related to the diuretic therapy. The delta ratio here is [(34 - 12) / (24 - 19)] = 22/5 = 4.4, and this very high value is consistent with this suggestion of a high anion gap acidosis in association with a pre-existing metabolic alkalosis.
  4. Compensation: As we know there is a metabolic acidosis lets us the "one and a half plus 8 rule" (rule 5) to assess the compensatory response: The expected pCO2 is (1.5 x 19 + 8) = 36.5mmHg. This is not that much different from the actual value of 32mmHg so there is only a minor & insignificant respiratory component present.
  5. Formulation: There is a high anion gap acidosis (probably diabetic ketoacidosis and/or lactic acidosis) in a patient who had a pre-existing metabolic alkalosis (due to thiazide therapy). Even though originally considered a possibility therew is no real evidence of any significant respiratory acid-base disorder.
  6. Confirmation: The nature of the high anion gap acidosis needs urgent sorting out, in particular urine testing for ketones, a serum lactate level and urea & creatinine levels. The hypokalaemia may be chronic and due to the metabolic alkalosis. A more detailed history would have been helpful here. Vomiting if present could have contributed to the metabolic alkalosis.

Finally: The Clinical Diagnosis

This lady has a severe diabetic ketoacidosis complicating a pre-existing metabolic alkalosis (due to thiazide use). Further investigations are necessary to exclude a lactic acidosis. Results of urinalysis, lactate and electrolytes are urgently required.


This is certainly an "interesting case". The first thing to notice here is that a quick scan of the acid-base results suggests only a minor acid base problem Disorders with compensating effects on the pH are present so pH is normal. The initial suggestion of a mild metabolic acidosis balancing a mild respiratory alkalosis turned out to be wrong. The big clue to understanding this acid-base problem was the large anion gap - a gap of this size always suggests a severe metabolic acidosis. So if a severe metabolic acidosis is present why is the bicarbonate not low. Conclusion: the compensating alkalosis must be a metabolic one (which primarily affects the bicarbonate) and not a respiratory one (which primarily affects the pCO2). Additionally, as it is balancing a severe acidosis, the metabolic alkalosis itself must be a significant one.

The high delta ratio (4.4) also suggests a coexistent metabolic alkalosis or a previous compensated respiratory acidosis. There is no evidence of severe respiratory disease. Also as the pCO2 is less than 40, this provides further support that a second primary disorder is raising the [HCO3], that is, a metabolic alkalosis is present. The history suggests that this would have been caused by the thiazide diuretic therapy. Inadequate chloride intake is usually necessary for diuretic use to result in a metabolic alkalosis.

The overall picture suggested is a woman with a pre-existing metabolic alkalosis (due to thiazides) who has developed an acute metabolic acidosis (diabetic ketoacidosis and/or lactic acidosis).

Assuming the normal anion gap is 12 mmols/l, then the increase in the anion gap is 22 mmol/l and should be approximately matched by the decrement in the [HCO3] as the average delta ratio found in DKA is typically about 1.0. This predicts that a [HCO3] of up to 41 mmols/l was present before the onset of the acute ketoacidosis (as 41 - 22 = 19 which is the measured [HCO3]). Vomiting was not mentioned in the history but is common in DKA and may have contributed to the alkalosis and K+ loss.

As both primary disorders are metabolic ones, we can also use the [HCO3] to see if the pCO2 value is an appropriate one. The only caveat here would be the usual one that sufficient time (12-24 hours) had passed for the maximal respiratory response to occur. More than this amount of time has passed since onset, so this is not a problem. The predicted pCO2 of 36mmHg is close to the measured value of 32 so respiratory compensation is appropriate and there is little evidence of the presence of a primary respiratory acid-base disorder.

Do you think you could have diagnosed this acid-base problem merely by inspecting the blood gas results? Perhaps you might have considered a mild metabolic acidosis and a slight hyperventilation due to anxiety from the arterial puncture. Remember also if the acid-base disorder or hyperosmolality seems too small to explain the degree of obtundation, other causes of coma should always be considered in a diabetic patient (eg trauma, epilepsy, drug overdose in the younger adult, stroke in the older patient). The potential mistake here is to diagnose only a minor acid-base disorder. If the blood glucose or urine tests (glucose and ketones) were not checked, this patient could have been admitted to a medical ward as ‘?CVA’ without appropriate management. If an elderly patient presents with mental obtundation and marked hyperglycaemia but without evidence of ketoacidosis, then the diagnosis is hyperosmolar non-ketotic coma rather then diabetic ketoacidosis. In this case the elevated anion gap has indicated the presence of a severe acidosis.

Two situations giving rise to a mixed metabolic acidosis and alkalosis are: