Objective To test for an association between duration of epistaxis and clinically relevant bleeding in the following 12 months in children with immune thrombocytopenia (ITP).

Design Prospective cohort study.

Setting The national UK Paediatric ITP registry, a multicentre prospective clinical registry of new cases of ITP between 2006 and February 2020.

Patients All children aged between 2 months and 16 years in participating UK centres.

Exposure Epistaxis at presentation defined as none, <10 min, 10–30 min and >30 min.

Main outcome measures Incident severe bleeds, combined moderate and severe bleeds and drop in haemoglobin (Hb) by ≥20 g/L.

Results The sample included 1793 patients, of which 334 had epistaxis <10 min, 88 lasting 10–30 min and 97 >30 min. In the 12 months following presentation, 19 had a severe bleed, 140 had a moderate bleed and 54 had a drop in Hb ≥2 g/dL. Epistaxis >30 min duration was associated with increased odds of severe bleeds (OR 1.43–15.67), moderate or severe bleeding (OR 1.33–4.2) and drop in Hb (OR 1.23–6.91). Shorter duration epistaxis was not associated with increased odds of any outcome. A trend for increased odds with longer duration epistaxis was significant for all outcomes.

Conclusions The longer the duration of epistaxis at presentation with ITP, the higher the risk of a clinically significant bleeding event in the 12 months following. This should inform clinical severity ratings and treatment decisions.