Table 3.

Analysis of the primary endpoint

Objective: Demonstrate superiority of programmed endoscopic therapy compared to standard of care on the composite endpoint of all-cause mortality, liver transplantation and cholangiosepsis 6 months after randomization
Target population: Adult patients with by ERC confirmed SSC–CIP
Variable/endpoint: Occurrence of any of the events
Primary estimand:
Intercurrent events	Early discontinuation of the study
Strategy	Treatment policy
Population-level summary	A Mantel–Haenszel estimate for the risk difference (experimental minus control group) with center as stratification factor will be used to compare treatment failures in all randomized patients
Estimand description	Superiority of the experimental group on the composite endpoint of death, liver transplantation or cholangiosepsis will be measured using the Mantel–Haenszel estimate for the risk difference irrespectively from the early discontinuation of the study
Imputation/data/censoring rule(s)	Patients who early discontinue the study without assessment of the primary endpoint will be counted as treatment failures
Sensitivity analyses	1)Estimation of the risk difference and the corresponding 95% CI without adjustment for center 2)Estimation of the risk difference and the corresponding 95% CI in a model where patients who early discontinue the study without assessment of the primary endpoint are counted as treatment responders 3)Estimation of the risk difference and the corresponding 95% CI in the per-protocol population