DKA and HHS — Viva

Long Case Viva Defence

Candidate's opening statement (model answer)

"This is a 19-year-old man with a 10-year history of type 1 diabetes who presents with two days of vomiting, abdominal pain and drowsiness, having stopped his insulin two days ago because he was not eating. On examination he is drowsy with a GCS of 13, has Kussmaul respiration at 28 breaths per minute, is tachycardic and hypotensive, has dry mucous membranes and epigastric tenderness, and a fruity odour to his breath. His capillary glucose is 31, his venous pH is 6.98 with a bicarbonate of 6 and a beta-hydroxybutyrate of 7.8, his potassium is 5.8, his anion gap is 32, and his chest X-ray shows a left lower lobe consolidation." [1]

"His main problems are:

1. Severe diabetic ketoacidosis (pH below 7.0), precipitated by insulin omission and complicated by a lobar pneumonia
2. Severe dehydration with an acute kidney injury
3. High risk of cerebral oedema given his age, severe acidosis and high urea
4. Longstanding poorly controlled type 1 diabetes (HbA1c 94) with likely microvascular complications, and a pattern of insulin omission that suggests a psychosocial or mental health issue." [1]

"My immediate priorities are to resuscitate with fluid, start a fixed-rate insulin infusion, manage his potassium, and begin antibiotics for the pneumonia, while monitoring closely for cerebral oedema and hypokalaemia in ICU." [1]

Examiner probing questions and model answers

Q1: "Take me through your diagnostic reasoning and the severity grade." [1]

"The diagnosis is severe diabetic ketoacidosis. All four criteria are met: glucose above 11 at 31, ketones above 3 at a beta-hydroxybutyrate of 7.8, pH below 7.3 at 6.98, and bicarbonate below 15 at 6. The anion gap is 32, confirming a high-anion-gap metabolic acidosis. The severity grade is severe, because the pH is below 7.0 and the bicarbonate is below 10 — this warrants ICU admission. The precipitant is insulin omission, with a lobar pneumonia as a possible additional trigger." [1]

Q2: "Walk me through your fluid and insulin protocol with specific doses for a 70 kg patient." [1]

"Fluid first. I start 0.9% saline — 1 litre in the first hour, then 1 litre over 2 hours, then 1 litre over 2 hours, then 1 litre over 4 hours, reassessing at each stage for cardiac and renal tolerance. His potassium is 5.8, so no potassium in the first bag. After the first litre, I start the fixed-rate intravenous insulin infusion at 0.1 units per kilogram per hour — for 70 kg that is 7 units per hour of soluble human insulin, made up as 50 units in 50 mL of saline. I do not give a loading bolus. The metabolic targets are a fall in ketones of at least 0.5 mmol per litre per hour and a rise in bicarbonate of 3 per hour. When the glucose falls below 14, I switch to 10 per cent dextrose at 125 mL per hour and continue the insulin unchanged. I check glucose and potassium hourly, ketones and venous gas two-hourly. Once the ketones are below 0.6, the pH above 7.3 and the bicarbonate above 18, I transition to subcutaneous insulin with an overlap of 30 to 60 minutes." [1]

Q3: "His potassium is 5.8. Would you give insulin?" [1]

"Yes — a potassium of 5.8 is above 5.5, so it is safe to start insulin. The concern is the reverse: if the potassium were below 3.3, I would delay insulin and replace potassium first, because insulin drives potassium into cells and can precipitate a fatal arrhythmia. At 5.8, he is safe to proceed, but I know the level will fall rapidly once insulin starts, so I will check it hourly and add 40 mmol per litre of potassium to the fluid the moment it drops below 5.5. The target range is 4 to 5.5." [1]

Q4: "What about bicarbonate? His pH is 6.98." [1]

"I would not give bicarbonate routinely. The Glaser study showed that bicarbonate administration is an independent risk factor for cerebral oedema in DKA. The acidosis corrects with fluid and insulin as the ketones are metabolised to bicarbonate. I would reserve bicarbonate for the exceptional case of a pH below 6.9 with haemodynamic instability that is not responding to standard therapy, and only in consultation with ICU — but even then, the evidence is weak and the cerebral oedema risk is real." [1]

Q5: "He is 19. How does that change your monitoring?" [1]

"It makes cerebral oedema a real concern. Cerebral oedema occurs in about 0.5 to 1 per cent of children and young adults with DKA, and it carries a 20 to 40 per cent mortality. The Glaser study identified three risk factors — low initial PaCO2 (his is very low at 14), high initial urea (his is 12), and bicarbonate administration. I will not give bicarbonate, I will keep the fluid within the protocol to avoid too rapid a correction, and I will monitor his neurological observations hourly. The moment he develops a headache, vomiting, drowsiness, or bradycardia, I will treat it as cerebral oedema: reduce the fluids, give mannitol 0.5 to 1 gram per kilogram or hypertonic 3 per cent saline, and move to ICU for neuroprotection." [1]

Q6: "What is the role of the SGLT2 inhibitor in this presentation?" [1]

"It is not relevant here — he has type 1 diabetes and is not on an SGLT2 inhibitor. But if he were, the mechanism of euglycaemic DKA would be the SGLT2 inhibitor causing glucosuria that lowers glucose, suppresses insulin, and raises glucagon, driving ketogenesis with a near-normal glucose. The defence is to check ketones in any acidotic diabetic regardless of the glucose, and to teach the sick-day rule — stop the SGLT2 inhibitor during acute illness with vomiting, diarrhoea, or reduced intake." [1]

Short Case Discussion

Scenario: "Interpret this blood gas and discuss your management"

Provided data: A 72-year-old woman with type 2 diabetes, drowsy. Capillary glucose 48 mmol/L. Venous blood gas: pH 7.32, bicarbonate 21, PaCO2 40. Beta-hydroxybutyrate 0.6. Sodium 150, potassium 4.5, urea 20, creatinine 180. Calculated osmolality 2 x 150 + 48 + 20 = 368 mOsm/kg. [1]

Candidate presentation (model): [1]

"The picture is a hyperosmolar hyperglycaemic state, not DKA. The glucose is markedly elevated at 48, the calculated osmolality is 368 — well above the diagnostic threshold of 320 — but the ketones are only 0.6, the pH is 7.32 (above 7.3), and the bicarbonate is 21 (above 15). This is HHS. The severe hyperosmolality explains her drowsiness. I note the urea of 20, which is an independent marker of severity and mortality in HHS, and the sodium of 150, which is high — the corrected sodium is about 150, so I will use 0.9% saline for initial resuscitation and consider 0.45% if the sodium rises further. The potassium is 4.5, which is acceptable." [1]

"My management priorities are aggressive but cautious fluid resuscitation — she is elderly and I do not know her cardiac function — low-dose insulin (0.05 units per kilogram per hour or fluid alone initially), prophylactic low-molecular-weight heparin given the high thrombotic risk, and an urgent search for the precipitant. The most common precipitant of HHS is infection, so I will send blood and urine cultures, a chest X-ray, and a troponin for a possible silent infarction. The precipitant is the primary determinant of her survival." [1]

Examiner: "How does your fluid and insulin management differ from DKA?" [1]

"In three ways. First, the fluid deficit is larger in HHS — 6 to 9 litres versus 3 to 6 in DKA — so I will need more fluid, but I will give it more cautiously because she is elderly and may have cardiac or renal comorbidity. I will reassess frequently for fluid overload. Second, the insulin dose is lower — 0.05 units per kilogram per hour or even fluid alone initially — because she retains residual endogenous insulin that is suppressing ketogenesis. I do not need to chase a ketone target because there is no significant ketosis. Third, I will give prophylactic enoxaparin because the severe dehydration and hyperviscosity create a high risk of venous and arterial thrombosis. My glucose fall target is 3 mmol per litre per hour — too rapid risks cerebral oedema." [1]

Examiner: "What is her prognosis, and what determines it?" [1]

"HHS carries a mortality of 10 to 20 per cent — substantially higher than DKA. The prognosis is determined less by the metabolic derangement and more by the precipitant and the patient's comorbidities. If her precipitant is a treatable urinary tract infection and she has no major cardiac failure, she has a good chance. If it is a silent myocardial infarction in a woman with established ischaemic heart disease and CKD, the outlook is poorer. My job is to treat the precipitant as aggressively as the metabolic emergency." [1]

Examiner: "She becomes agitated and then drowsier at hour 6. What do you think, and what do you do?" [1]

"I am concerned about two possibilities. First, a too-rapid osmolality correction causing cerebral oedema — I would check the osmolality trend and slow the fluids if it has fallen faster than 3 to 8 mOsm per kilogram per hour. Second, the underlying precipitant — a stroke, a septic embolus, or a metabolic derangement such as a sodium shift. I would check a repeat gas and electrolytes urgently, examine her neurologically for focal signs, and arrange a CT head. I would not assume it is simply the HHS resolving — a change in neurological status demands investigation." [1]

Examiner: "Name three exam traps in hyperglycaemic emergencies." [1]

"First, treating DKA as a glucose problem — the insulin runs until ketones clear, not until glucose normalises; the glucose is managed with dextrose. Second, missing euglycaemic DKA because the glucose is reassuring — always check ketones in an acidotic diabetic, especially on an SGLT2 inhibitor. Third, starting insulin before checking the potassium — if it is below 3.3, insulin will precipitate a fatal arrhythmia. A fourth, if you will allow: failing to overlap the transition from intravenous to subcutaneous insulin — the DKA recurs if there is a gap." [1]