P656 Superior treatment persistence with ustekinumab in biologic-experienced luminal Crohn’s disease: real-world registry data from the Persistence Australian National IBD Cohort (PANIC) study

Abstract

Abstract Background Ustekinumab has been shown equivalent to adalimumab in the treatment of biologic-naïve patients with Crohn’s disease (CD) at 1 year. However the real world persistence of these medications, and particularly when used sequentially in biologic-experienced patients is unknown. Prescription data provides real-world evidence about treatment efficacy, tolerability and prescriber and patient acceptability. Methods Persistent of medication was analysed from the Australian Pharmaceutical Benefits Scheme (PBS) registry data 2005-2019 for luminal Crohn’s disease, using a 10% representative sample. Results There were 2029 patients over the 15 year period with 5618 years of patient follow up. The median age was 43 years (interquartile range (IQR): 34-58), with 891/2029 (43.9%) male. Via therapy, 915/2029 (45.1%) were on adalimumab, 722/2029 (35.6%) were on infliximab 155/2029 (7.6%) were on vedolizumab, and 237/2029 (11.7%) were on ustekinumab. Used as a first-line biologic (bio-naïve), there was no difference between any agent for persistence (p>0.05). Used after first-line (bio-experienced), ustekinumab had significantly better persistence than TNFa inhibitors (TNFi) (p= 0.006) and vedolizumab (p<0.001), additionally vedolizumab had poorer persistence than TNFi (p=0.028). In bio-experienced patients, ustekinumab persistence was greater than non-ustekinumab biologic agents (p= 0.0018). TNFi persistence was significantly poorer in bio-experienced vs naïve patients (p=0.042), with a non-significant trend for poorer outcomes in vedolizumab (p= 0.059), but did not differ for ustekinumab (p=0.509). Used after a TNFi agent, ustekinumab had superior persistence to an alternative TNFi agent (p=0.033). Using a propensity score matched analysis for age, immunomodulator use and line of therapy, ustekinumab had superior to persistence to non-ustekinumab agents (p=0.0044) and to TNFi (p=0.01). Multivariate predictors of loss of persistence were biologic used (adjusted hazard ratios (aHR): 0.84, p=0.001), and line of therapy (aHR: 1.16, p=0.003). Conclusion The PANIC cohort with real-world data of non-hierarchical prescribing of biological agents does not support the superiority of any agent in biologic-naïve luminal CD. However for biologic experienced, the persistence is greater with ustekinumab than TNFa inhibitors, and vedolizumab persistence is significantly poorest. These findings will assist in the positioning of biological therapies in luminal CD.