Aernoud Fiolet

219 Colchicine in chronic coronary disease in relation to history of acute coronary syndrome Event rates in prior ACS subgroups Cox proportional hazard models were used to investigate univariable relationships between prior ACS status and the primary composite end point. These analyses were performed in two separate ways: no prior ACS versus prior ACS, and recent ACS versus remote or very remote ACS. To adjust for baseline differences between these groups, a multivariate analysis was performed including all baseline covariates reported in the baseline table, without prior ACS status, and using the forward Wald method with entry at p<0.10 and removal at p>0.15. Predictors of the primary end point were incorporated in an adjusted cox proportional hazard model that included prior ACS status, thereby to assess whether prior ACS status independently affected the hazard of the primary composite end point. Treatment effect of colchicine by prior ACS status Cumulative event rates following randomization were plotted by treatment group using the Kaplan-Meier method and according to prior ACS status. Cox proportional hazard models were used to investigate colchicine treatment effect in each subgroup for the primary composite end point. The interaction between colchicine treatment effect and prior ACS status was assessed by addition of treatment allocation and treatment-by-‘prior ACS status’ variable interaction to the Cox proportional hazard models. Hazard ratios (HR) and 95% confidence intervals (CI) were generated and all reported p values are two-sided. Efficacy analyses were performed on an intention- to-treat basis. Analyses were performed using SPSS version 26.0 (IBM SPSS Statistics, Chicago, USA). RESULTS A total of 5,522 patients were randomized to placebo (n = 2,760) or colchicine 0.5 mg (n = 2,762) once daily and followed for a median of 28.6 months (interquartile range: 20.5 to 44.4). A total of 864 (16%) patients had no prior ACS, 1,479 (27%) had a recent ACS within six to twenty-four months, 1,582 (29%) had a remote ACS within two to seven years, and 1,597 (29%) had a very remoteACS more than seven years prior to randomization. Of those with prior ACS, the median time from the most recent acute event to randomization was 4 years (interquartile range: 2 to 10), with 68% of randomizations occurring >2 years after the last acute event. Baseline characteristics of patients by prior ACS status are shown in Table 1. Patients in the very remote ACS subgroup were older, had more comorbidities, and were more