Editorial on: ‘Lowering the C‐Reactive Protein Threshold Improves Risk Stratification in Acute Ulcerative Colitis: A Propensity‐Matched Analysis’. Authors' Reply

We thank Bradbury and Tyrrell‐Price for their considered Editorial [1]. Their commentary complements our original analysis, although we welcome the opportunity to clarify the interpretation of the reduced CRP threshold and address its wider implications for clinical practice [2].

We are pleased the editorial affirms our central message: that non‐acute severe ulcerative colitis (NASUC) is not benign. Once matched for objective severity and treatment exposure, colectomy risk was indistinguishable from that of acute severe ulcerative colitis (ASUC) at every timepoint to one year. This is why we favour acute ulcerative colitis as a more inclusive term, with ASUC and NASUC contained within this broader category of severe disease requiring inpatient care.

We have previously reported that a CRP threshold of 12 mg/L is a better correlate of the original Truelove and Witts Criteria (TWC) erythrocyte sedimentation rate (ESR) > 30 mm/h than the currently used CRP 30 mg/L [3, 4, 5]. This is why in the current study we examined whether applying this lower threshold captured high‐risk NASUC patients by reclassifying them as ASUC. In doing so, we found that this simple change substantially improved risk stratification across the NASUC population: it reclassified 26% of NASUC admissions as ASUC and captured 43% of those who ultimately required colectomy.

However, as our paper and the Editorial highlights, among prevalent ulcerative colitis (UC) cases admission CRP is muted by immunosuppressive treatment and neither the CRP 12 nor CRP 30 mg/L threshold separated the risk of early or late colectomy. In our exploratory analysis (Supporting Information Figures S2A and S2B), the highest risk of 30‐day colectomy occurred among ASUC patients irrespective of biologic exposure, together with biologic‐exposed NASUC patients, while beyond 6 months colectomy risk remained highest in biologic‐exposed ASUC and NASUC patients alike and biologic‐naive ASUC patients fared substantially better. Among bio‐exposed patients (n = 67), endoscopic severity (Mayo endoscopic subscore) and treatment history (prednisolone on admission) discriminated 90‐day colectomy risk, whereas CRP did not (Table 1). This dissociation between biochemistry and biology is why no single biomarker threshold can capture the most treatment‐experienced patients, and why endoscopic severity and treatment history remain essential complements to CRP. It is this recognition that motivates our ongoing work towards a composite predictive index integrating biochemistry, disease severity and treatment context, in the tradition of the Risk of Rescue score [6].

TABLE 1.

Multivariable Cox proportional hazards analysis for risk of colectomy (90 day).

Variable	Standard Cox HR (95% CI)	p	Firth penalised Cox HR (95% CI)	p
CRP at admission (per 5 mg/L)	1.02 (0.96–1.08)	0.537	1.03 (0.96–1.07)	0.352
Mayo endoscopic score (MES 3 vs. 1 + 2)	4.32 (1.17–15.98)	0.028	3.80 (1.22–15.15)	0.020
Prednisolone on admission	7.13 (1.43–35.64)	0.017	6.07 (1.61–34.36)	0.006

Note: Multivariable Cox proportional hazards analysis of time to colectomy within 90 days of admission, restricted to biologic‐exposed patients (n = 67, of whom 13 underwent colectomy). C‐reactive protein (CRP) was modelled as a continuous variable, with the hazard ratio (HR) expressed per 5 mg/L increment; endoscopic severity was defined by the Mayo endoscopic subscore (MES 3 vs. 1–2). Adjusted HRs are shown with 95% confidence intervals (CI). Given the small number of events (events per variable [EPV] = 4.3 across three predictors), Firth penalised estimation was applied to reduce small‐sample bias, and the proportional hazards assumption was assessed using Schoenfeld residuals. With only 4.3 events per variable, the model is underpowered; the wide confidence intervals reflect this, and the estimates should be interpreted as exploratory.

We are grateful to Bradbury and Tyrrell‐Price for advancing this discussion. Their observation that specialists may already apply Truelove and Witts criteria loosely is itself an argument to move towards a graded, evidence‐based estimate of risk that serves therapy‐naïve and therapy‐exposed patients alike.

Author Contributions

Amirah Etchegaray: conceptualization, investigation, writing – original draft, writing – review and editing, methodology, validation, visualization, formal analysis, data curation, project administration. Naeman Goetz: writing – review and editing. Katherine Hanigan: writing – review and editing, data curation. Jennifer Phillips: writing – review and editing. Rina Kumar: writing – review and editing. George Tambakis: writing – review and editing. Graham Radford‐Smith: writing – original draft, writing – review and editing, methodology, supervision, data curation. Gareth Walker: supervision, writing – original draft, writing – review and editing, visualization, methodology, data curation. Anthony Croft: data curation, conceptualization, investigation, writing – original draft, writing – review and editing, visualization, methodology, supervision, resources.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Linked Articles

This article is linked to Etchegaray et al. papers. To view these articles, visit https://doi.org/10.1111/apt.70849 and https://doi.org/10.1111/apt.70825.

Supporting information

Data S1: Supplementary Material.

Figure S2: Time to Colectomy by Severity Group and Biologics on Admission.

Etchegaray A., Goetz N., Hanigan K., et al., “Editorial on: ‘Lowering the C‐Reactive Protein Threshold Improves Risk Stratification in Acute Ulcerative Colitis: A Propensity‐Matched Analysis’. Authors' Reply,” Alimentary Pharmacology & Therapeutics 64, no. 7 (2026): 1007–1008, 10.1111/apt.70885.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.