SGLT2 inhibitors reliably raise haemoglobin - and in a minority the rise matters

A systematic review pooled 10 studies of the effect of sodium-glucose co-transporter 2 (SGLT2) inhibitors on blood counts in adults with type 2 diabetes, with follow-up from 8 weeks to 24 months. People with primary polycythaemia, end-stage kidney disease or heart failure were excluded. Across three randomised trials totalling only 171 patients, SGLT2 inhibitors raised haematocrit by 2.29 percentage points (95% CI 1.48 to 3.09) and haemoglobin by 0.51 g/dL (95% CI 0.28 to 0.74) against placebo. Real-world data on 9646 patients showed the prevalence of polycythaemia rising from 2.4% to 9.7%, with cases judged severe in 1.4%. One small observational study of 100 patients with erythrocytosis and no JAK2 mutation reported a 10% rate of thrombotic events.

This rise in haemoglobin is not a side effect in the ordinary sense - it is thought to be part of how these drugs work, reflecting recovered kidney erythropoietin production and a fall in plasma volume, and it tracks with their cardiovascular benefit. The point is not to be alarmed by it but to recognise it, because the alternative is a patient sent down an unnecessary haematology pathway, or occasionally a genuine erythrocytosis missed.

Practically: expect the rise, do not investigate a haematocrit that moves within the normal range, and set a threshold at which you will act. Reversibility on stopping the drug is a useful diagnostic step when the picture is unclear. The thrombotic signal here rests on a single study of 100 people and is not a reason to stop SGLT2 inhibitors in anyone - it is a reason to look harder in the small group with a markedly raised haematocrit, particularly smokers, people with untreated obstructive sleep apnoea, those on testosterone, and men living at altitude, where the causes stack.

• Check a full blood count before starting an SGLT2 inhibitor if you do not have a recent one, and again at around 6 months
• Expect haemoglobin to rise by roughly 0.5 g/dL and haematocrit by about 2 percentage points - this is the drug working, not a complication
• Investigate rather than accept a haematocrit above about 54% in men or 48% in women, or any rise well beyond the expected range
• Look for the co-existing causes first: smoking, untreated sleep apnoea, testosterone therapy, high altitude
• If the picture is unclear, a trial of stopping the drug is informative - the change reverses

The statistics, in plain English

The randomised evidence here rests on just 171 patients across three trials, so although the confidence intervals for haematocrit (1.48 to 3.09) and haemoglobin (0.28 to 0.74) both sit clearly above zero, the pooled estimate is fragile and could shift with one more trial. The jump in polycythaemia prevalence from 2.4% to 9.7% comes from real-world records rather than a randomised comparison, so some of it reflects who gets tested and when, not the drug alone. The 10% thrombosis rate comes from a single observational study of 100 selected patients with no control group - it establishes that thrombosis happens in such patients, not that SGLT2 inhibitors cause it.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.