Oxford Criteria for Acute Severe Ulcerative Colitis
The Oxford criteria predict steroid failure in acute severe ulcerative colitis (ASUC) by assessing patients on day 3 of intravenous corticosteroid therapy: >8 stools per day OR 3-8 stools per day with CRP >45 mg/L indicates an 85% risk of colectomy and should prompt immediate consideration of rescue therapy. 1111
The Oxford Criteria Defined
The Oxford criteria remain the simplest and most widely validated prognostic tool for ASUC 1. Assessment occurs on day 3 of intravenous corticosteroid treatment:
- >8 stools per day alone, OR
- 3-8 stools per day with CRP >45 mg/L
Meeting either criterion historically predicted an 85% colectomy rate 112. However, recent evidence suggests this rate has decreased to approximately 36% in the modern biologic era, though these patients still require escalation of therapy 3. A 2025 study confirmed that day 3 Oxford criteria remains an independent predictor of steroid non-response (OR 4.70,95% CI 1.06-20.80) 4.
Initial Management Framework
Day 0-3: Intravenous Corticosteroids
Start intravenous hydrocortisone 100 mg three to four times daily or methylprednisolone 40-60 mg daily equivalent immediately upon diagnosis of ASUC 115. Do not delay treatment for infectious workup results, though stool studies should be sent 1.
Critical supportive measures:
- VTE prophylaxis with subcutaneous low molecular weight heparin plus compression stockings (rectal bleeding is NOT a contraindication) 11
- Nutritional assessment and enteral feeding (no role for routine bowel rest or parenteral nutrition) 1
- Early surgical consultation to discuss potential colectomy 1
Avoid adjunctive antibiotics unless documented gastrointestinal or extra-intestinal infection is present 5.
Day 3: Critical Decision Point
Formally assess the Oxford criteria on day 3 1111. This is the most important prognostic checkpoint. Count stool frequency over 24 hours and measure CRP.
If Oxford criteria are met (steroid failure):
- Initiate rescue therapy planning immediately
- Options include infliximab, cyclosporine, or surgery 11
- Infliximab is preferred if patient has previously failed thiopurine therapy 11
- Both infliximab and cyclosporine show equivalent short-term efficacy 11
Additional poor prognostic indicators on imaging/endoscopy:
- Colonic dilatation >5.5 cm on plain abdominal X-ray 11
- Mucosal islands on X-ray 11
- Deep ulceration on flexible sigmoidoscopy 11
Day 7: Secondary Assessment
If rescue therapy was not initiated on day 3, reassess on day 7. Steroid failure at this point is defined as:
This predicts a 40% colectomy rate in subsequent months 11. Continuing intravenous corticosteroids beyond 7 days without response is not effective 5.
Rescue Therapy Selection
For steroid-refractory ASUC, use either infliximab or cyclosporine 5. The choice depends on:
Choose infliximab if:
- Prior thiopurine failure (cyclosporine should NOT be used in this setting) 11
- Patient preference for biologic therapy
- Availability and institutional experience
Cyclosporine (2 mg/kg IV) may be considered if:
- Thiopurine-naïve patient 1
- Infliximab unavailable or contraindicated
- Plan to bridge to thiopurine maintenance 1
Surgery is indicated when:
- Toxic megacolon develops 11
- Medical rescue therapy fails 11
- Medical rescue therapy is contraindicated 11
Important Caveats
The Oxford criteria's predictive value has evolved: While originally associated with 85% colectomy rates, modern studies show rates of 36% when rescue therapy is available 3. This doesn't diminish its utility—it identifies patients who need treatment escalation, not necessarily those who will undergo surgery 4.
Alternative scoring systems exist (Swedish/Fulminant Colitis Index, Edinburgh score, Lindgren score) but Oxford criteria remain the simplest to apply in clinical practice 11. A 2023 study developed a new admission-based model using CRP ≥100 mg/L, albumin ≤25 g/L, and UCEIS scores, but this requires validation 6.
Long-term outcomes: Even patients who respond to steroids initially have substantial colectomy risk after discharge (11-13.5% at 1 year) 4. Initial steroid response does not predict long-term colectomy rates 4.