Meta-Analysis of Randomized Trials: Efficacy and Safety of Colchicine for Secondary Prevention of Cardiovascular Disease

Background Colchicine has shown potential cardioprotective effects owing to its broad anti-inflammatory properties. We performed a meta-analysis to assess its safety and efficacy in secondary prevention in patients with established coronary artery disease (CAD). Methods We searched Ovid Healthstar, MEDLINE, and Embase (inception to May 2022) for randomized controlled trials (RCTs) evaluating the cardiovascular effects of colchicine compared with placebo or usual care in patients with CAD. Study-level data on efficacy and safety outcomes were pooled using the Peto method. The primary outcome was the composite of cardiovascular (CV) death, myocardial infarction (MI), or stroke. Results A total of 8 RCTs were included with a follow-up duration of ≥1 month, comprising a total of 12,151 patients. Compared with placebo or usual care, colchicine was associated with a significant risk reduction in the primary outcome (odds ratio (OR) 0.70, 95% CI 0.60 to 0.83, P < 0.0001; I2 = 52%). Risks of MI (OR 0.75, 95% CI 0.62 to 0.91, P = 0.003; I2 = 33%), stroke (OR 0.47, 95% CI 0.30 to 0.74, P = 0.001; I2 = 0%), and unplanned coronary revascularization (OR 0.67, 95% CI 0.55 to 0.82, P = 0.0001; I2 = 58%) were all reduced in the colchicine group. Rates of CV and all-cause mortality did not differ between the two groups, but there was an increase in noncardiac deaths with colchicine (OR 1.54, 95% CI 1.10 to 2.15, P = 0.01; I2 = 51%). The occurrence of all other adverse events was similar between the two groups, including GI reactions (OR 1.06, 95% CI 0.94 to 1.20, P = 0.35; I2 = 42%) and infections (OR 1.04, 95% CI 0.84 to 1.28, P = 0.74; I2 = 53%). Conclusions Colchicine therapy may reduce the risk of future cardiovascular events in patients with established CAD; however, there remains a concern about non-CV mortality. Further trials are underway that will shed light on non-CV mortality and colchicine NCT03048825, and NCT02898610.


Introduction
1.1.Background and Rationale.Patients with established coronary artery disease (CAD) remain at high risk for recurrent cardiovascular events despite current evidence-based secondary prevention that includes antithrombotic and lipid therapies alongside lifestyle changes [1].Te residual incidence of such events is estimated at 3-5% per year in patients treated with guideline-directed medical therapy [2,3].Improvements to current treatment options are still needed.
Persistent subclinical coronary infammation is perceived as a key driver of residual risk.A number of studies suggested that infammatory biomarkers play a pivotal role in the development and progression of atherosclerosis [4,5], and research has now progressed into clinical trials investigating whether specifcally targeting infammation prevents cardiovascular events [6].Colchicine is a low-cost drug that has been used for many years for the treatment of infammatory conditions such as gout, pericarditis, and familial Mediterranean fever (FMF).It interferes with several steps in the infammatory process and has an antitubulin efect that inhibits neutrophil function.In trials of gout [7] and FMF [8], retrospective observations suggested a cardioprotective efect with continued use of colchicine [9].More recently, large randomized controlled trials (RCTs) showed potential beneft with colchicine in patients with stable coronary disease [10] and post-ACS [11].We therefore conducted a meta-analysis of RCTs to examine the efcacy and safety of colchicine for secondary prevention in patients with CAD.

Methods
2.1.Protocol and Registration.Tis meta-analysis was conducted in accordance with the Cochrane Handbook for Systematic Reviews and Interventions [12] and reported following the PRISMA checklist for meta-analysis in healthcare interventions [13].A protocol specifying the objectives, inclusion criteria, and analysis methods was submitted to PROSPERO [14] on January 21, 2021, and registered with the number CRD42021227630.

Information Sources.
We conducted systematic searches for relevant studies comparing colchicine versus placebo or usual care for secondary cardiovascular prevention up to May 30, 2022.Tis included searching Ovid Healthstar (from 1966), MEDLINE (from 1948), and Embase (from 1980) as well as the Cochrane Central Register of Controlled Trials.Reference lists of included studies, relevant articles, and related systematic reviews were also reviewed.To identify any ongoing or recently completed studies that have not been published, we searched conference abstracts from the American College of Cardiology (ACC), the American Heart Association (AHA), Transcatheter Terapeutics (TCT), and the European Society of Cardiology (ESC) from the last three years.Web-based registries were also searched, including ClinicalTrials.govand PROSPERO to identify completed studies that have not been published.
A search strategy was created for Ovid Healthstar and modifed for application to the other databases (Figure S1).Te key concepts of the search strategy included the intervention ("colchicine"), disease/morbidity subject terms ("coronary artery disease," "coronary heart disease," "myocardial ischemia," "cardiovascular diseases," "acute coronary syndrome," and "myocardial infarction"), and study type (randomized trials).Te search was limited to the English language.Review articles, editorials, duplicates, and post hoc analyses of original RCTs were excluded.

Eligibility Criteria.
To be included in this meta-analysis, all RCTs had to meet the following criteria: (1) studies that compared colchicine versus placebo or usual care for secondary cardiovascular prevention; (2) studies that reported at least one of the following outcomes: cardiovascular death, myocardial infarction (MI), and stroke; (3) colchicine was administered at any dose for a minimum of 30 days; and (4) follow-up duration was at least 30 days.

Outcome Measures and Certainty of Evidence.
Information on study outcomes was abstracted for the longest available follow-up.Te primary outcome was the composite of cardiovascular (CV) death, MI, or stroke.Secondary analyses were conducted on the individual components of the primary outcome, as well as all-cause mortality and unplanned coronary revascularization.Safety outcomes included noncardiovascular mortality, gastrointestinal events, infection, cancer, myalgia, and myelotoxicity.Te Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach was used to evaluate the certainty of the evidence presented in this study [15].Te GRADE approach incorporates evaluations on trial risk of bias, inconsistency, indirectness, imprecision, and publication bias.

Study Selection, Data Extraction, and Bias Assessment.
Study selection was performed by two independent review authors (E.A. and N.S.).Titles and abstracts were screened for eligibility followed by full-text review.Reasons for exclusion were documented.Disagreements in each stage were resolved by discussion and consensus.An electronic database was developed to document study characteristics and outcome data on the intent-to-treat population.One reviewer entered the data (E.A.), which was subsequently validated by a second reviewer (N.S.).
Te risk of bias was assessed in the 5 domains as recommended in the Cochrane Handbook for Systematic Reviews of Interventions.Tese domains included risk of bias (1) arising from the randomization process; (2) due to deviations from the intended interventions; (3) related to missing outcome data; (4) in the measurement of the outcome; and (5) in the selection of the reported result (Figure S2).

Statistical Analysis.
Peto odds ratios (ORs) were calculated to display dichotomous outcomes.Peto's fxed-efect method was used to calculate the pooled Peto ORs and corresponding 95% confdence intervals (CIs) when two or more studies provided combinable data.Heterogeneity was evaluated with the Cochrane Q test (signifcant at P < 0.10) and the Higgins I 2 statistic.Published guidelines for low (I 2 � 25-49%), moderate (I 2 � 50-74%), and high (I 2 ≥ 75%) heterogeneity were used [16].Subgroup analyses for the primary outcome were performed according to (1) treatment duration (30 days, >30 days), (2) colchicine dosage (<1 mg/d, ≥1 mg/d), and (3) CAD phenotype (stable CAD vs ACS presentation).We expected a greater beneft as treatment duration increased and with a larger dosage of colchicine.We also expected heterogeneity by baseline CAD phenotype since the risk for cardiovascular events would difer.Publication bias was evaluated using the inverted funnel plot techniques (signifcant at P < 0.10) [17,18].As a sensitivity analysis, we also performed a Bayesian metaanalysis using a noninformative prior distribution.Tis allowed us to calculate the posterior probability and estimate the probability of any beneft or of a clinically signifcant beneft for each of the specifed endpoints.All analyses were performed using Review Manager (RevMan), version 5

Search and Selection of Studies.
A total of 599 abstracts were retrieved from electronic databases and handsearching conference proceedings and reference lists.Of the 33 studies selected for full-text review, 25 were eliminated after applying the eligibility criteria (Figure 1).[10,24].Te majority of trials used a colchicine dose of 0.5 mg daily.Tree RCTs had a follow-up duration of 1 month [19][20][21], one RCT had a follow-up duration of 6 months [23], and all 4 other trials had follow-up durations of at least 1 year [10,11,22,24].Figures 2(a) and 2(b) summarize the risk of bias for the included RCTs and individual assessments.Te LoDoCo [24] and COLIN [20] trials were the only open-label RCTs comparing colchicine to usual care, but the selection bias was minimized in both trials through allocation concealment, and detection bias was reduced in LoDoCo through blinded outcome assessment.

Patient Characteristics.
Patient characteristics are listed in Table 2. Te mean age of patients ranged between 57 and 67.Over 75% of enrolled patients were male in all trials.Te prevalence of hypertension, diabetes, and smoking was well balanced between the intervention and control arms of the studies.Use of secondary prevention drugs was high across trials, and more than 90% of patients were on concomitant statin therapy.S1 summarizes the fndings for the primary and secondary outcomes including the GRADE quality assessment.

Efcacy Outcomes.
Compared with placebo or usual care, the addition of colchicine to standard medical therapy in patients with CAD was associated with a signifcant reduction in the primary outcome of CV death, MI, or stroke (OR 0.70, 95% CI 0.60 to 0.83, P < 0.0001; I 2 � 52%) (Figure 3).Tere was also a reduction in MI (OR 0.75, 95% CI 0.62 to 0.91, P � 0.003; I 2 � 33%), stroke (OR 0.47, 95% CI 0.30 to 0.74, P � 0.001; I 2 � 0%), and unplanned coronary revascularization (OR 0.67, 95% CI 0.55 to 0.82, P � 0.0001; I 2 � 58%).Te reduction in the primary outcome was consistent between subsets of trials that difered according to treatment duration (P for interaction � 0.81), colchicine dose (P for interaction � 0.65), and CAD phenotype at baseline (P for interaction � 0.37) (Figure 4).Excluding one study at a time in a "one-out" sensitivity analysis did not neutralize the pooled Peto OR of the primary outcome.Excluding results from the LoDoCo and COLIN open-label trials did not alter primary outcome results, although the magnitude of efect and heterogeneity (I 2 ) between studies were decreased (Figure S3).Te Bayesian posterior probability of colchicine reducing the primary endpoint was 99%.

Risk of Publication Bias.
Visual inspection of the funnel plots did not show asymmetry and suggested no signifcant risk of publication bias (Figure S4).

Discussion
Te present meta-analysis included eight RCTs of patients receiving colchicine for secondary prevention of cardiovascular diseases.Colchicine was associated with a signifcant risk reduction in the primary outcome of CV death, MI, or stroke as compared with placebo or usual care.Tere was also a signifcant reduction in the individual outcomes of MI, stroke, and unplanned coronary revascularization but not CV death.Tese fndings were consistent between subsets of trials that difered according to CAD phenotype at baseline (ACS vs stable CAD).However, there was an increase in noncardiovascular mortality in the colchicine group.Tis fnding is uncertain given the signifcant heterogeneity.Colchicine did not increase the risk of all other major adverse events, including gastrointestinal reactions and infections.
Te plausibility of reduction of cardiovascular events by targeting coronary infammation is well supported in the literature.Te role of proteolytic enzymes released as part of the chronic infammatory process leading to atherosclerotic plaque erosion or rupture leading to recurrent coronary events has been well described [25].Statins have been shown to reduce cardiovascular events through the reduction of infammation in addition to their low-density lipoprotein (LDL)-lowering properties [26].In the CANTOS trial, canakinumab signifcantly reduced the rate of recurrent cardiovascular events compared with placebo by targeting the interleukin-1β innate immunity pathway in patients with previous MI and elevated high-sensitivity C-reactive protein levels [6].It is now appreciated that extracellular cholesterol crystals trigger infammatory processes upstream and downstream of the interleukin-1β pathway, suggesting that the broad anti-infammatory properties of colchicine may be required to adequately tackle the atherosclerotic infammatory process [27].Te mechanisms of action of colchicine are complex and include microtubule assembly, infammasome activation, infammatory cell chemotaxis, leukotriene and cytokine generation, and phagocytosis.Recently, colchicine has been found to attenuate the nucleotide-binding oligomerization domain-, leucine-rich repeat-, and pyrin domain-containing protein-3 (NLRP3) infammasome-mediated crystal-induced infammation [9].A proteomic analysis of the LoDoCo-2 trial performed after 30 days of colchicine treatment revealed a reduction of interleukin IL-18, IL-1 receptor antagonist, and IL-6, consistent with an attenuation of the NLRP3 infammasome pathway.Tere was also a reduction of the upstream NF-κB essential modulator, required in NLRP3 activation [28].In a large meta-analysis investigating the efects of colchicine on infammatory markers in patients with CAD across 11 clinical trials, colchicine led to a signifcant reduction in hs-CRP (weighted mean diferences (WMDs), −0.81 mg/L; 95% confdence interval (CI), −1.34 to 0.28 mg/L; P � 0.003) and IL-6 levels (WMD, −1.28 pg/mL; 95% CI, −2.35 to −0.21 pg/ mL; P � 0.02) compared with placebo [29].From a clinical standpoint, colchicine is inexpensive and taken orally.On the other hand, widespread use of canakinumab has been limited due to increased risk of fatal infections, elevated cost, and inconvenience related to subcutaneous administration.
Although the risk of ischemic cardiovascular events in colchicine-treated patients was signifcantly reduced, it did not translate into an overall survival beneft, and noncardiovascular mortality was increased with colchicine therapy.Tese fndings, along with the CANTOS trial results, perhaps indicate that immunomodulating therapy in patients with established CAD may bring about cardiovascular benefts at a cost of noncardiac deaths.Future research should focus on evaluating the safety of colchicine in this patient population, especially its impact on noncardiac mortality and serious infections.Such research includes the Colchicine and Spironolactone in Patients With MI/SYNERGY Stent Registry (CLEAR SYNERGY) trial, which uses a 2 × 2 factorial design to examine the efect of colchicine and spironolactone in patients presenting within  Te risk reduction in stroke with colchicine therapy was substantial and reinforced the role of colchicine as a treatment of infammation-related atherothrombosis.It may be especially relevant to neurologists and other healthcare professionals involved in secondary stroke prevention.Te colchicine for prevention of vascular infammation in noncardio embolic stroke (CONVINCE) will further inform us on the role of lowdose colchicine for secondary prevention of recurrent stroke and major vascular events in patients who have already sufered a nonembolic ischemic stroke or transient ischemic attack (ClinicalTrials.govidentifer: NCT02898610).
Te lack of diference between groups for myalgia is especially reassuring given some existing concerns about   As per authors, all patients received aspirin at the time of primary PCI, but the percentage of patients that remained on aspirin in each arm is not stated.Journal of Interventional Cardiology increasing the risk of drug-induced myopathy when coprescribing statins and colchicine.Over 94% of patients in both groups were on statin therapy and yet such phenomenon was not observed.

Journal of Interventional Cardiology
Overall, our results are consistent with previous metaanalyses with smaller sample sizes [30,31].In a metaanalysis of four RCTs in which endpoints were harmonized across trials, treatment with colchicine was associated with a 32% reduction in the incidence of major cardiovascular events, but a trend towards increased noncardiovascular mortality was observed (pooled RR 1.38 (95% CI 0.99 to 1.93); I 2 � 0.0%) [31].Similarly, in a meta-analysis by the Colchicine Cardiovascular Trialists Collaboration, low-dose colchicine reduced the risk of MACE by 25% (RR 0.78, 95% CI 0.64 to 0.94; P � 0.010) and there was no diference in all-cause death, with fewer cardiac deaths (RR 0.82, 95% CI 0.55 to 1.23; P � 0.34) counterbalanced by a trend for more noncardiovascular deaths (RR 1.38, 95% CI 0.99 to 1.92; P � 0.060) [32].Compared with these two meta-analyses, ours included a larger number of patients and events and showed a statistically signifcant increase in non-CV death with colchicine.Tis fnding remains uncertain given the signifcant heterogeneity, the fact that four of the trials were not evaluable due to lack of reported data and unexplained as rates of cancer and infections were not increased in the colchicine group.

Exploration of Heterogeneity.
With the exclusion of the two open-label trials, a signifcant drop in heterogeneity was observed for the primary outcome of CV death, stroke, or MI (I 2 � 0.0%) and its individual components of CV death (I 2 � 0.0%), stroke (I 2 � 12%), and MI (I 2 � 7.0%) (Figure S5).Tis was not observed for non-CV mortality, so our fndings related to this outcome should still be interpreted with caution.Also, our data do not provide sufcient evidence to determine clinical features that identify patients with CAD that are most likely to beneft or be harmed from colchicine therapy, such as patient age, comorbidities, and duration of therapeutic time window.Te other major determinants of the heterogeneity observed in this analysis may remain uncertain until patient-level data analysis is performed.

4.2.
Limitations.Tis meta-analysis does have limitations.First, we did not have access to individual patient data from the trials which would have allowed detailed analysis of subgroups.Second, our study may be underpowered to detect any potential diference in rare events such as cardiovascular death.Indeed, <1% of patients in both the colchicine and control groups in this meta-analysis died from cardiovascular causes.Tird, the inclusion of the LoDoCo and COLIN trials may have introduced randomization bias due to their open-label design.However, bias was likely reduced thanks to allocation concealment in both trials.Furthermore, excluding these two trials did not alter our primary outcome results as shown in the sensitivity analysis (Figure S3).Fourth, duration of follow-up exceeded two years in only two trials (n � 6,054), preventing us from drawing any frm conclusion about the safety and efcacy of colchicine for secondary cardiovascular prevention on very long-term follow-up.Tree trials (n � 363) had a short 30day course of colchicine therapy, which poses challenges to    Journal of Interventional Cardiology interpretation of mortality outcomes.Finally, our metaanalysis was underpowered to detect the interaction between colchicine dose or duration and the benefcial efect of colchicine over placebo.

Conclusion
In randomized trials of patients receiving colchicine for secondary prevention of cardiovascular disease, colchicine led to a signifcant risk reduction in the composite outcome of CV death, MI, or stroke.However, there remains a concern for noncardiovascular mortality with colchicine in the CAD population.Terefore, although these fndings may support the concept of targeting infammation in atherosclerosis, further clinical research is required before colchicine can be safely administered routinely for secondary cardiovascular prevention.

Figure 2 :
Figure 2: Risk of bias graph and summary.(a) Risk of bias graph: review authors' judgments about each risk of bias item presented as percentages across all included studies.(b) Risk of bias summary: review authors 'judgment about each risk of bias item for each included study.
Characteristics and Risk of Bias within Studies.

Table 2 :
Patient characteristics at baseline.