Investigators queried the Korean National Health Insurance Service database to identify patients with type 2 diabetes mellitus and gout who had no prior coronary artery disease. The source reported 48,295 eligible patients between 2015 and 2019. Among those, 28,964 were recorded as colchicine users and 19,331 as non-users prior to matching. The study team applied 1:1 propensity score matching and generated 17,337 matched pairs for outcome comparisons. This matching approach was intended to balance baseline covariates between users and non-users; the abstract did not list the full set of covariates used for matching in the provided text.
The investigators predefined the primary outcome as a composite of all-cause death, nonfatal myocardial infarction, and nonfatal stroke. In the matched cohort, the composite outcome occurred in 12.52% of colchicine users versus 15.03% of non-users. The authors reported a hazard ratio (HR) of 0.71 with a 95% confidence interval (CI) of 0.66–0.75 (P < 0.001). Simple arithmetic shows the absolute difference favored the colchicine group; statistical estimates reported by the authors indicated a relative risk reduction in the composite outcome.
All-cause mortality, evaluated as a standalone endpoint, was lower among colchicine users. The reported event rates were 8.01% in the colchicine group and 11.07% in the non-user group. The corresponding HR was 0.58 (95% CI 0.53–0.62; P < 0.001). The authors noted that the observed reduction in the composite endpoint was driven by this mortality difference rather than by differences in nonfatal ischemic events.
When the components of the composite were analyzed individually, investigators reported no statistically significant difference for nonfatal myocardial infarction: HR 1.15 (95% CI 0.96–1.37; P = 0.121). For nonfatal stroke, the HR was 0.91 (95% CI 0.81–1.01; P = 0.076), which also did not reach conventional significance. These point estimates imply no clear reduction in nonfatal myocardial infarction or stroke associated with colchicine in this cohort; the mortality signal therefore appears to be the dominant driver of the composite outcome result.
The study team conducted a landmark analysis at 365 days to address potential immortal time bias. In that analysis, the associations with lower all-cause mortality and with the composite outcome remained statistically significant. Specifically, the landmark HR for mortality was 0.67 (95% CI 0.62–0.73; P < 0.001) and for the composite outcome was 0.76 (95% CI 0.71–0.81; P < 0.001). By contrast, in the 365-day landmark analysis neither myocardial infarction (HR 1.15; 95% CI 0.95–1.40; P = 0.144) nor stroke (HR 0.99; 95% CI 0.87–1.12; P = 0.823) differed significantly between colchicine users and non-users. The investigators therefore report a persistent association between colchicine exposure and lower mortality that is not mirrored by reductions in nonfatal ischemic events in both primary and landmark analyses.
The paper was authored by Shin et al. and published in PLoS One; the affiliations listed include departments of cardiology in Korea and a data science department at a pharmaceutical company. The analysis used a nationwide administrative claims database, which enabled the investigators to capture a large real-world sample of patients with type 2 diabetes mellitus and gout without prior coronary artery disease. The source abstract did not provide several methodological details in the text provided here: dosing regimens for colchicine, duration of exposure, and adverse event or safety data were not reported in the abstract. The authors did perform propensity score matching and a 365-day landmark analysis, which they used to address confounding and immortal time bias respectively, but the abstract did not list all covariates or sensitivity analyses that might further clarify residual confounding.
Readers should recognize that the outcome differences reported in the abstract were driven by lower mortality rather than reductions in myocardial infarction or stroke. The provided source text did not report cause-specific mortality breakdowns, nor did it report colchicine dosing or adverse-event rates—those details were not available in the abstract. Further scrutiny of the full manuscript would be required to evaluate safety signals, exposure definitions, and the specific covariates included in the propensity matching. These unanswered, documentable gaps in the abstract represent the primary limitations evident from the provided source material.