Skip to content
live well bell

SGLT2 medicines: why normal glucose can miss ketoacidosis

SGLT2 medicines can be linked to ketoacidosis without very high glucose. What symptoms, fasting precautions, and sick-day plans deserve attention?

By Julian Vance6 min read
Glass of water and tablets shown as a medication illustration; tablets are not identified as SGLT2 medicines.
Photo: SHVETS production on Pexels. Illustrative medication photograph.

A reassuring glucose reading can answer the wrong question. For people taking an SGLT2 inhibitor, diabetic ketoacidosis can develop without the very high blood sugar many people associate with it. The distinction matters when illness, reduced food intake, or fasting enters the picture. What we have is a recognised medication safety issue, not evidence that every rise in ketones is dangerous. The 2024 international consensus report on hyperglycaemic crises explains that distinction.

Glucose is only part of the metabolic picture

Diabetic ketoacidosis, usually shortened to DKA, involves excess ketones alongside metabolic acidosis: a dangerous disturbance in the blood’s acid–base balance. Insufficient insulin action and increased stress hormones allow ketone production to accelerate. A glucose reading does not measure that acidity, and it does not measure ketones. This is why the 2024 consensus report in Diabetes Care requires more than a glucose result to establish the diagnosis.

The report describes euglycaemic DKA, where the glucose concentration is below the level traditionally expected in this emergency. “Euglycaemic” should not be read as “safe”. It describes the glucose concentration, not the severity of the acid–base disturbance. Reduced food intake and SGLT2 treatment are among the recognised contexts. The clinical implication is narrow but important: symptoms and medication history still matter when the glucose number looks relatively ordinary.

For a reader using a glucose monitor, the useful question is therefore not simply whether the trace is high. It is whether the information being displayed can answer the safety question at hand. A glucose result cannot stand in for a clinical assessment of possible ketoacidosis.

Why this medicine deserves a separate conversation

SGLT2 inhibitors include dapagliflozin and empagliflozin. They increase the amount of glucose passed in urine. They also have uses beyond lowering blood sugar, including treatment for some people with heart failure or chronic kidney disease. The NHS overview of dapagliflozin describes these different indications. Someone may therefore be taking one without thinking of it primarily as a diabetes medicine.

That makes the name on the prescription more useful than an assumption about why it was prescribed. A medication review should include the actual drug name, any combination tablet, and the reason for treatment. The question is not whether a reader recognises themselves as a typical diabetes patient. It is whether the safety information for their particular prescription has been explained.

The benefits should remain in view. A warning about ketoacidosis is not an argument for abandoning a useful prescription. It is an argument for understanding the circumstances in which the usual plan needs clinical review. The March 2026 NHS England leaflet on dapagliflozin discusses both treatment benefits and temporary interruptions during illness.

Fasting changes the context, not just the timetable

Skipping food deliberately and being unable to eat during an infection are different experiences. Both can nevertheless be relevant to a medicine whose safety advice includes reduced intake and dehydration. The UK Kidney Association’s 2023 guideline in BMC Nephrology advises against ketogenic diets during SGLT2 treatment and addresses intermittent fasting separately. Its fasting recommendation is conditional and highlights people who are older, take diuretics, or have chronic kidney disease.

That is guidance for an individual treatment conversation, not a ready-made instruction to miss tablets on fasting days. The same guideline recommends sick-day advice and temporary withholding during acute illness. It does not establish a universally safe fasting window. Nor does it justify replacing prescribed treatment with a dietary experiment.

What we have is enough evidence to take the interaction seriously. What we do not have is a simple calculation in which a particular number of fasting hours cancels out a particular drug dose. A proposed fast, a very low-carbohydrate diet, or preparation for a procedure belongs in the conversation before the change happens.

Which symptoms need urgent attention?

Possible DKA symptoms include stomach pain, nausea or vomiting, deeper breathing, unusual tiredness, sleepiness or confusion, thirst, and fruity-smelling breath. These symptoms are not a checklist that must be completed before seeking help. The NHS guidance on diabetic ketoacidosis describes it as a potentially life-threatening condition requiring urgent hospital treatment.

If DKA is suspected, call 999 or go to A&E. Do not drive yourself. Do not wait to obtain a ketone meter or for glucose to become markedly raised. In particular, a person who is confused, unusually sleepy, or breathing abnormally needs prompt assessment rather than another attempt to finish a fast. Tell the assessing team about the SGLT2 prescription and any recent change in eating, drinking, illness, or medication.

Diabetes UK’s SGLT2 safety guidance explicitly warns that DKA can occur with normal or only mildly raised glucose. It advises temporary stopping of the SGLT2 medicine when unwell, checking glucose and ketones if the person has been taught and has the equipment, and speaking to the healthcare team. Symptoms suggesting DKA require emergency help; testing must not delay it.

A written sick-day plan is more useful than guesswork

There are two separate decisions: recognising a situation that needs help, and managing treatment through recovery. They should not be collapsed into a general instruction to stop every medicine. Insulin is especially important: the NHS advises people not to stop insulin even when they are not eating, and to follow the adjustment plan agreed with their diabetes team.

The NHS England dapagliflozin leaflet lists dehydration, illness that stops eating, and surgery among circumstances when an SGLT2 medicine may need temporary interruption. A useful personal plan identifies whom to contact, what to do if food or fluids cannot be kept down, and who confirms when treatment can resume. An article cannot supply that individual decision.

Useful questions are concrete: does the plan cover vomiting as well as fever? What happens if symptoms begin outside surgery hours? Who should be contacted if recovery takes longer than expected? The March 2026 NHS England leaflet advises contacting a GP or NHS 111 if illness continues after 48 hours. That is separate from the emergency response to suspected DKA. Asking for the instructions in writing gives the reader something more dependable than memory when unwell.

What this means in practice

  • Check the actual names on your medication list. Ask a pharmacist whether a combination tablet contains an SGLT2 inhibitor if the label is unclear.
  • Before starting fasting or a ketogenic diet, discuss the proposed change with the prescriber. Do not create your own schedule of skipped doses.
  • Ask for a written sick-day and procedure plan, including contact details and advice on restarting treatment. Mention any fluid restriction.
  • If you have been given ketone-testing equipment, ask the team to confirm when to use it and what action each result requires. A glucose monitor alone does not answer the ketone question.
  • For symptoms suggesting DKA, seek emergency care even if glucose looks reassuring. Do not delay for home testing, and do not stop insulin independently.

These points bring together the NHS and specialist guidance linked above. They are questions and safety boundaries for care, rather than a fasting prescription.

What we don’t know

The evidence is firmer on recognising ketoacidosis and known risk contexts than on defining the safety of every voluntary fasting pattern during SGLT2 treatment. The UK Kidney Association grades its intermittent-fasting recommendation 2D, reflecting a conditional suggestion supported by very low-quality evidence. That uncertainty should remain visible; a guideline precaution is not a trial proving which eating window is safe.

Individual risk cannot be read directly from a single glucose value or from whether someone has tolerated fasting before. Nor should the possibility of this complication be used to imply that a medicine’s established benefits disappear. The appropriate balance depends on the indication, the person’s health, and the clinical circumstances. Those are reasons for a specific medication conversation, rather than stronger promises about fasting.

A reassuring glucose result is useful information. In someone who is unwell while taking an SGLT2 inhibitor, it is not the whole safety assessment.

Photo: SHVETS production on Pexels. Illustrative medication photograph; the tablets shown are not identified as SGLT2 inhibitors.

About the author

Fasting and metabolic health researcher exploring autophagy and cellular repair.

Leave a Comment