cartouche ECN WORKSHOP
-

Risk of cardiovascular events in patients with gout starting urate-lowering therapy with or without flare prophylaxis with colchicine: an emulated target trial using english primary-care, hospitalisation, and mortality data

 

Dr Edoardo Cipolletta, Dr Georgina Nakafero, Prof Anthony J. Avery, Prof Mamas A. Mamas, Dr Laila J. Tata, Prof Abhishek Abhishek

 

Affiliation(s):

Academic Rheumatology, University of Nottingham

 

 

Background: Gout flares are associated with cardiovascular events (CVEs). Clinical trials show that low-dose colchicine reduces the risk of CVEs in patients with atherosclerotic cardiovascular diseases [1,2]. A recent study reported an increased risk of myocardial infarction (MI) in people with gout starting urate-lowering therapy (ULT) with colchicine [3].

Objectives: To estimate the risk of CVEs among patients with gout initiating ULT with colchicine flare prophylaxis compared with no prophylaxis.

Methods: Study design and population. English patients with a new diagnosis of gout starting ULT for the first time were included in a new-user cohort study using an emulated target trial design with propensity score overlap weighting.

Exposure. Colchicine prophylaxis vs no prophylaxis. Patients co-prescribed colchicine for ≥21 days were compared with patients not prescribed colchicine or NSAIDs for ≥21 days on the same date as ULT initiation.

Main outcome. First CVE (fatal and non-fatal stroke or MI) after cohort entry. Secondary outcomes. First-ever CVEs, fatal CVEs, MI, stroke. Negative control outcomes: respiratory tract infections, peptic ulcer disease. Positive control outcome: diarrhoea.

Follow-up. We performed intention-to-treat (ITT) and per-protocol (PP) analyses. In ITT, follow up was from ULT initiation to 180 days. In PP, follow-up was censored earlier if exposed patients ended prophylaxis, switched to NSAID prophylaxis, or if controls were prescribed prophylaxis.

Statistical analysis. We used Cox proportional hazards models to obtain adjusted hazard ratios (aHR). We calculated adjusted incidence rates (aIR), adjusted risk differences (aRD), and E-values. In the PP analyses, inverse probability of censoring weighting was applied.

Results: 99,800 patients were included (Figure): mean (SD) age 62.8 (15.5) years, 25,511 (25.6%) female, 4,063 (4.1%) with prior CVE, median (IQR) disease duration 0.2 years (0-2.4) and mean (SD) serum urate 516.4 micromol/l (94.1). Nearly all patients started allopurinol (99.3%).

Controls and exposed patients were followed-up for 176.9 (26.7) and 175.5 (29.7) days in the ITT analysis and 161.1 (49.5) and 45.7 (33.1) days in PP analysis.

Patients with colchicine prophylaxis had lower risks of CVEs compared to those without prophylaxis in both the ITT (aHR: 0.82, 95%CI: 0.69-0.94) and PP analyses (aHR: 0.79, 95%CI: 0.58-0.99) (Table).

There was no evidence of effect modification by age, gender, cardiovascular risk category, and year of ULT initiation in stratified analyses (not shown). Colchicine prophylaxis was not associated with negative control outcomes (not shown).

The E-values (95%CI lower bound) for ITT and PP analyses were 1.74 (1.32) and 1.85 (1.11).

Conclusion: Colchicine prophylaxis is associated with a reduced risk of CVEs when initiating ULT.

References

[1] 10.1056/NEJMoa1912388 

[2] 10.1056/NEJMoa2021372

[3] 10.1136/ard-2023-224154

 

 

Back