Introduction
Therapeutic response in Crohn’s disease (CD) varies considerably, and identifying predictors of response remains a clinical challenge. A clinical decision support tool (VDZ-CDST) has been developed to estimate the likelihood of vedolizumab response in CD. While the tool has previously been evaluated in clinical trials and site-based observational cohorts, its performance in routine care remains unclear.
Aims & Methods
We conducted a register-based real-world cohort study of adults with CD who initiated vedolizumab in Sweden between 2014 and 2024 and had available data at time of vedolizumab initiation to calculate VDZ-CDST scores. The primary aim was to evaluate the ability of the VDZ-CDST score to predict clinical remission at 52 weeks after vedolizumab initiation (w52), assessed by remission rates across score categories and discrimination (AUC). Secondary aims included prediction of clinical remission at week 26 (w26), biochemical remission and changes in quality of life at w26 and w52, as well as treatment persistence and steroid- and surgery-free survival at w26 and w52. Trends across ordered groups were evaluated using Jonckheere–Terpstra and discrimination was assessed using receiver operating characteristic analysis. Kaplan–Meier methods were also used to assess treatment persistence, steroid-free survival, and surgery-free survival across ordered VDZ-CDST groups.
Results
Among 3191 adults treated with vedolizumab, 816 had complete data for score calculation (anti-TNF exposure, serum/plasma albumin and CRP, previous surgery, fistulizing disease) and were included (median age 30 years [IQR 20–46], median disease duration 9 years [3.7–17.7]). Of these, 84 were classified as low, 447 as intermediate, and 285 as high probability of remission at w52.
Clinical remission status at w52 was available for 448 patients (55%). Baseline VDZ-CDST categories were not significantly associated with clinical remission at w52 (p-trend = 0.253), and discrimination was low (AUC = 0.56). By contrast, categories were associated with clinical remission at w26: among 372/816 assessable patients (46%), remission rates were 6% (3/52), 17% (36/214), and 26% (28/106) in the low-, intermediate-, and high-probability groups, respectively (p-trend <0.001; AUC = 0.66).
VDZ-CDST categories also showed stepwise associations with biochemical remission. At w26, among 593/816 assessable patients (73%), remission rates were 3% (2/74), 30% (100/339), and 35% (63/180) (p-trend <0.001; AUC = 0.64). At w52, corresponding rates among 588/816 assessable patients (72%) were 8% (6/75), 22% (75/336), and 29% (51/177) (p-trend <0.001; AUC = 0.61).
Treatment persistence showed clear stepwise separation across VDZ-CDST categories, with median time on vedolizumab of 8.3, 27.0, and 44.2 months in the low-, intermediate-, and high-probability groups, respectively (log-rank p<0.001). Both treatment persistence and surgery-free survival remained separated beyond 5 years after vedolizumab initiation, whereas steroid-free survival showed a similar numerical trend but did not reach statistical significance.
Conclusion
In this real-world cohort, VDZ-CDST did not predict clinical remission at week 52. However, it showed modest performance for other vedolizumab-related outcomes: biochemical remission, treatment persistence, and surgery-free survival. Although individual-level discrimination was limited in this setting, the score provides a framework for further refinement toward individualized risk assessment.
Disclosure
Dr. Olén has served as the principal investigator (PI) on projects at the Karolinska Institutet, unrelated to the current abstract, funded by grants from Janssen, AbbVie, Takeda, Ferring, Alfasigma, Bristol Myers Squibb, and Pfizer. The institute has received fees for lectures and for participation on advisory boards from Janssen, Ferring, Bristol Myers Squibb, Alfasigma, and Takeda.
Dr. Ludvigsson has led an unrelated study for the Swedish IBD Register (SWIBREG), funded by Janssen. He has also received financial support from MSD to develop a paper reviewing national healthcare registers in China and to fund an unrelated study on IBD and fibrosis. Additionally, Dr. Ludvigsson is involved in a research collaboration with Takeda on celiac disease and is currently in discussions about a collaborative effort on liver disease with the same company.
C. R. H. Hedin served as a speaker and/or advisory board member for AstraZeneca, Abbvie, Celltrion, Dr Falk Pharma and the Falk Foundation, Galapagos, Janssen, Lilly, Pfizer, Pharmacosmos, Ferring, Takeda, Tillotts Pharma, and received grant support from Tillotts and Takeda, outside the submitted work.
Jonas Halfvarson served as a speaker and/or advisory board member for AbbVie, Alfasigma, BMS, Celgene, Celltrion, Dr Falk Pharma and the Falk Foundation, Eli Lilly, Ferring, Alfasigma, Gilead, Hospira, Index Pharma, Janssen, Johnson & Johnson, MEDA, Medivir, Medtronic, Merck, MSD, Novartis, Pfizer, Prometheus Laboratories Inc., Sandoz, Shire, STADA, Takeda, Thermo Fisher Scientific, Tillotts Pharma, Vifor Pharma, and UCB. He received grant support from Janssen, MSD, and Takeda.
N-P. Ekholm is an Employee at Takeda Pharma AB.
The remaining authors disclose no conflicts.
