Colchicine in cardiac disease: a systematic review and meta-analysis of randomized controlled trials

Background Colchicine has unique anti-inflammatory properties that may be beneficial in various cardiovascular conditions. This systematic review and meta-analysis of randomized controlled trials (RCTs) examines this issue. Methods We searched MEDLINE, EMBASE, and the Cochrane Database from inception to June 2014 for RCTs using colchicine in adult patients with cardiac diseases. Results were pooled using random effects. Results 15 RCTs (n = 3431 patients, median treatment 3 and follow-up 15 months) were included. All but 2 used colchicine 1 mg/day. In 5 trials, n = 1301) at risk for cardiovascular disease (coronary artery disease, acute coronary syndrome or stroke, post-angioplasty [2 RCTs], or congestive heart failure), colchicine reduced composite cardiovascular outcomes by ~60 % (risk ratio [RR] 0.44, 95 % confidence interval [CI] 0.28-0.69, p = 0.0004; I2 = 0 %) and showed a trend towards lower all-cause mortality (RR 0.50, 95 % CI 0.23-1.08, p = 0.08; I2 = 0 %). In pericarditis or post-cardiotomy, colchicine decreased recurrent pericarditis or post-pericardiotomy syndrome (RR 0.50, 95 % CI 0.41-0.60, p < 0.0001; I2 = 0 %; 8 RCTs, n = 1635), and post-pericardiotomy or ablation induced atrial fibrillation (RR 0.65, 95 % CI 0.51-0.82, p = 0.0003; I2 = 31 %; 4 RCTs, n = 1118). The most common adverse event was diarrhea. Treatment discontinuation overall and due to adverse events (RR 4.34, 95 % CI 1.70-11.07, p = 0.002; I2 = 29 %; 7 RCTs, 83/790 [10.5 %] vs. 11/697 [1.6 %]) was higher in colchicine-assigned patients. Conclusions Current RCT data suggests that colchicine may reduce the composite rate of cardiovascular adverse outcomes in a range of patients with established cardiovascular disease. Furthermore, colchicine reduces rates of recurrent pericarditis, post-pericardiotomy syndrome, and peri-procedural atrial fibrillation following cardiac surgery. Further RCTs evaluating the potential of colchicine for secondary prevention of cardiovascular events would be of interest.


Background
Colchicine is used to treat gout and other inflammatory diseases such as familial Mediterranean fever and Behçet's syndrome [1][2][3][4][5][6]. Randomized controlled trials (RCTs) have evaluated colchicine in a wide spectrum of cardiac disease [7]. There is now tantalizing evidence that it may prove to be a useful adjunct to current therapy in the treatment and prevention of pericarditis, post procedural atrial fibrillation, atherosclerosis, and stent related disease. In addition, new insights into the role of cholesterol crystal-induced, neutrophil-mediated inflammation in atherosclerosis add plausibility to the clinical benefits observed with its use in secondary prevention of cardiovascular disease [8,9]. Consequently, we set out to conduct a meta-analysis of RCTs to evaluate the benefits of colchicine in patients with both coronary artery and other cardiac diseases.

Data sources
We systematically searched OVID versions of MEDLINE (1946 through to July 2014, week 1), EMBASE Classic and EMBASE (1947 through 2014 week 28), and the Cochrane Central Register of Controlled Trials (Issue 5, June 2014) for relevant studies using "colchicine" and "coronary artery" or "cardiovascular" search terms, and previously-published sensitive filters to identify randomized controlled trials (RCTs) [10,11] (see Fig. 1 for detailed search strategy). We also searched bibliographies of included studies and personal files. We did not impose language restrictions.

Study selection
We included prospective clinical trials randomizing patients to colchicine vs. placebo or no therapy. Eligible studies included adult patients with cardiac disease including cardiovascular disease, cardiomyopathy or congestive heart failure, pericardial disease, or arrhythmias. Studies were excluded if patients were not randomized between treatment groups. Citations were screened and full text review was conducted to determine eligibility when a screening reviewer felt a citation potentially met inclusion criteria.

Data extraction and quality assessment
Details of the publication (i.e. trial authors and acronym, enrolment period, year of publication), inclusion/exclusion criteria, demographics and cardiac risk factors of the enrolled patients, description of the interventions used, and outcome definitions and events were collected and collated. Risk of bias in RCTs (including blinding of participants, method of sequence generation and allocation concealment, intention-to-treat analysis, early trial stopping for efficacy before the planned enrollment was completed, and loss to follow-up) was also assessed.

Data analysis
The primary outcome was the composite cardiovascular outcome as defined by each RCT enrolling patients with cardiovascular diseases, and was recurrent pericarditis or post-pericardiotomy syndrome for RCTs enrolling patients with pericarditis or post pericardiotomy. (Given the different cardiac diseases in the included trials, we only pooled data for trials enrolling patients with relevant conditions for each outcome). Secondary outcomes were all-cause mortality, acute coronary syndrome or non-fatal myocardial infarction, revascularization, and non-fatal stroke (cardiovascular disease trials); and arrhythmias (pericarditis, post-pericardiotomy syndrome, and post radiofrequency ablation trials). Adverse events, and treatment discontinuation data were pooled from all trials as we felt these may be independent of patient condition and to provide the most robust estimate of overall adverse event rates. All analyses were performed using Review Manager (RevMan version 5.2; Cochrane Collaboration, Oxford, UK). Random effects models [12] which incorporate between-trial heterogeneity and give wider and more conservative confidence intervals (CI) when heterogeneity is present were used for all analyses. Statistical heterogeneity among trials was assessed using the I 2 statistic, defined as the percentage of total variability across studies attributable to heterogeneity rather than chance, and using published guidelines for low (I 2 = 25 %-49 %), moderate (I 2 = 50 %-74 %) and high (I 2 ≥ 75 %) heterogeneity [13]. Relative risks (RR) were used to pool outcomes with a two-sided significance level of 5 %. Individual trial and summary results are reported with 95 % CIs. For the pericarditis RCTs, we use Z-tests of interaction to calculate interaction p-values comparing RRs between separate subgroups (e.g. acute pooled RR vs. recurrent pericarditis pooled RR, etc.). To assess for publication bias, a funnel plot comparing effect measure to study precision was examined for evidence of asymmetry.
Results from two of these RCTs [24,27] were each published in two publications [28,29].
Description of included studies and quality assessment All but 2 RCTs used a colchicine dose of 1 mg/day. Almost half the trials, mainly those enrolling patients with pericarditis [18][19][20][21][22] or post cardiac surgery [24,25], reduced the dose to 0.5 mg/day in patients with signs of medication intolerance or body weight <70 kg. Median treatment duration was 3 months. Patient follow-up ranged from 7 days to 3 years (median 15 months).

Quantitative data synthesis
Cardiovascular Outcomes: Colchicine was associated with a >50 % reduction in the composite cardiovascular outcome (RR 0.44, 95 % CI 0.28-0.69, p = 0.0004, I 2 = 0 %; 5 trials [1301 patients]) (Fig. 2). As indicated in Fig. 2, two of these trials reported only deaths. [16,17]. This result was driven by the single-centre RCT in stable CAD which was the only RCT to individually demonstrate a statistically significant benefit (RR 0.33, 95 % CI 0.19-0.59, p = 0.0001; n = 532). It made up 65 % of the weighting in the pooled estimate and included mortality, myocardial infarction, stroke, and cardiac arrest in its composite outcome [14]. This trial was open   label but blinded outcome assessors and treated patients for a median of 3 years [14]. Visual inspection of the funnel plot for this outcome showed no evidence of asymmetry.
All-cause mortality rates were low with only 4 trials having a total of 10 deaths in the colchicine group and 18 deaths in the control group with a nonsignificant trend of benefit in the colchicine group (RR 0.50, 95 % CI 0.23-1.08, p = 0.08; I 2 = 0 %) (Fig. 3). There were no differences in rates of acute coronary syndrome or non-fatal myocardial infarction (RR 0.59, 95 % CI 0.09-3.90, p = 0.58; I 2 = 46 %; 2 trials with  [RCTs enrolling patients with pericarditis [18][19][20][21][22], post-pericardiotomy syndrome [23][24][25][26], or post-RF ablation [27] are excluded. Only two of these RCTs [24,25] reported any deaths: 2/169 vs. 2/167 patients [24], and 6/180 vs. 2/180 [25]. Including these RCTs does not significantly change the pooled result: RR 0.74, 95 % CI 0.37-1.49, p = 0.40. The pooled RRs with 95 % CI were calculated using random-effects models. Weight refers to the contribution of each study to the overall pooled estimate of treatment effect. Each square and horizontal line denotes the point estimate and 95 % CI for each trial's RR. The diamonds signify the pooled RR; the diamond's centre denotes the point estimate and width denotes the 95 % CI   (Fig. 6). The pericarditis patient trials [18][19][20][21][22] which reduced the dose of colchicine for intolerance or low body weight (<70 kg) and enrolled generally younger patients (mean age 48-57), showed no significant increase in gastrointestinal side effects. This was different than the results from the other trials which showed a doubling of gastrointestinal side effects (interaction p = 0.01). However, including data from all 7 RCTs that reduced the dose of colchicine for intolerance or body weight <70 kg [18-22, 24, 25] (Fig. 7). This was the case even including data only from the 7 RCTs that reduced the dose of colchicine for intolerance or body weight <70 kg [18-22, 24, 25] . There were no differences in medication discontinuation overall or discontinuation due to adverse events between subgroups of cardiovascular disease, pericarditis, and post-pericardiotomy or radiofrequency ablation patient trials (interaction p = 0.16-0.18) (results not shown).

Discussion
As suggested by this meta-analysis, in different populations of patients with established cardiovascular disease, colchicine reduces the composite cardiovascular outcome by approximately 60 %. Although the pooled results are consistent between trials with no heterogeneity, the estimates are dominated by one open-label RCT (though with blinded outcome assessors) enrolling patients with stable coronary artery disease [14]. Our meta-analysis also demonstrates that colchicine reduces the rates of pericarditis recurrence, post-pericardiotomy syndrome, and atrial fibrillation post cardiac surgery or radiofrequency ablation by about 50 %. These pooled results are supported by numerous RCTs providing consistent results. Pooled results from all RCTs demonstrated that side effects, primarily diarrhea and other gastrointestinal symptoms, are increased contributing to higher medication discontinuation rates compared to placebo/control groups; however, no serious adverse events were reported.
Our systematic review and meta-analysis is the first to include RCTs testing the effect of colchicine in all cardiac diseases and is the largest and most comprehensive published to date. Previous reviews have included only RCTs enrolling patients with pericarditis, or the post pericardiotomy syndrome [30][31][32][33][34], and even the most recent of these did not include the most recently published post pericardiotomy [25] or atrial fibrillation [25][26][27] prevention RCTs. A protocol for a systematic review of colchicine for secondary prevention of Fig. 6 Forest Plot for Gastrointestinal Adverse Events. Individual and pooled risk ratios (RR) with 95 % confidence intervals (CI) for randomized controlled trials (RCTs) comparing colchicine to placebo or control in patients with cardiovascular diseases, pericarditis, and post pericardiotomy or radiofrequency (RF) ablation. The pooled RRs with 95 % CI were calculated using random-effects models. Weight refers to the contribution of each study to the overall pooled estimate of treatment effect. Each square and horizontal line denotes the point estimate and 95 % CI for each trial's RR. The diamonds signify the pooled RR; the diamond's centre denotes the point estimate and width denotes the 95 % CI. The pericarditis patient trials [18][19][20][21][22] which reduced the dose of colchicine for intolerance or low body weight (<70 kg) and enrolled generally younger patients (mean age 48-57), showed no significant increase in gastrointestinal side effects. This was different than the results from the other trials which showed a doubling of risk (interaction p = 0.01). Including data from all 7 RCTs that reduced the dose of colchicine for intolerance or body weight <70 kg [18-22, 24, 25], the increase in gastrointestinal adverse events was lower but still statistically significant (RR 1.56, 95 % CI 1.09-2.24, p = 0.01, I 2 = 0 %; 7 RCTs, 1524 patients) suggesting that dose reduction by itself is not sufficient to eliminate gastrointestinal side effects. For two RCTs that reported non-diarrhea gastrointestinal side effects separately from the diarrhea side effects we assumed that the 5/130 vs. 4/67 patients with nausea or vomiting and 0/130 vs. 1/67 patients with dyspepsia were different than the 36/130 vs. 3/67 patients with diarrhea [16] and the 6/103 vs. 3/103 patients with nausea were different than the 10/103 vs. 2/103 patients with diarrhea [27]. Results are similar if one assumes that these events occurred in the same patients for these 2 RCTs (overall RR 2.11, 95 % CI 1.54-2.89, p < 0.0001, I 2 = 26 %; 216/1559 [13.9 %] vs. 77/1463 [5.3 %]) cardiovascular disease [35] has been published by the Cochrane Collaboration.
Plausible mechanisms have been proposed to explain the apparently beneficial effects of colchicine for prevention of cardiovascular events [36]. Colchicine has been shown to inhibit neutrophil chemotaxis, ingress and activation within a pro-inflammatory environment. These anti-inflammatory effects appear to be important for preventing pericarditis recurrence; however, activated neutrophils are also present in atherosclerotic plaques and appear to play a key role in the transformation of a stable to an unstable plaque. By suppressing neutrophils, colchicine may play a role in stabilizing plaques and preventing fissuring or rupture that can result in the clinical manifestations of acute coronary ischemia or stroke. It is reasonable to speculate that colchicine has greater potential for action in stable coronary disease, where inflammatory mechanisms play a major role in disease progression, and that it may have less potential to provide any benefit in acute coronary syndromes in which atherothrombosis predominates. This may explain the large benefit observed in the only RCT that evaluated colchicine in patients with stable CAD [14], and lower or absent benefits observed in RCTs trying to prevent angioplasty induced vascular injury [16], stent-related disease [7] and acute plaque instability [15], which are different cardiovascular disease entities with unique pathophysiology. Numerous additional trials evaluating the use of colchicine in patients with stable cardiovascular disease, acute coronary syndrome, atrial fibrillation, and post operatively are listed on trial registries suggesting that more data regarding colchicine's effectiveness in specific cardiovascular conditions will be forthcoming ( Table 4). In addition several other trials of inflammation reduction in secondary prevention of cardiovascular events are currently underway, some with Fig. 7 Forest Plot for All Medication Discontinuation and Discontinuation Due to Side Effects. Individual and pooled risk ratios (RR) with 95 % confidence intervals (CI) for randomized controlled trials (RCTs) comparing colchicine to placebo or control in patients with various cardiac conditions. The pooled RRs with 95 % CI were calculated using random-effects models. Weight refers to the contribution of each study to the overall pooled estimate of treatment effect. Each square and horizontal line denotes the point estimate and 95 % CI for each trial's RR. The diamonds signify the pooled RR; the diamond's centre denotes the point estimate and width denotes the 95 % CI. Including only data from the 7 RCTs that reduced the dose of colchicine for intolerance or body weight <70 kg [18-22, 24, 25], the rates of medication discontinuation were still increased: discontinuation overall, RR 1.40, 95 % CI 1.04-1.89, p = 0.03, I 2 = 0 %; 7 RCTs, n = 1524; and discontinuation due to adverse events, RR 2.22, 95 % CI 1.08-4.56, p = 0.03, I 2 = 0 %; 4 RCTs, n = 684. There were no differences in medication discontinuation overall or discontinuation due to adverse events between subgroups of cardiovascular disease, pericarditis, and post-pericardiotomy or radiofrequency ablation patient trials (interaction p = 0.16-0.18) (results not shown) non-specific anti-inflammatory agents (like methotrexate), and others with specific anti-inflammatory approaches (such as cannakinumab) [37].
This meta-analysis demonstrated that side effects were minimal; however, the short duration of therapy and follow up in most of the included RCTs meant that primarily early side effects, best described as intolerance that leads to early discontinuation of the medication, were recorded. In applications requiring only short term use such as pericarditis and atrial fibrillation prevention there is a need to develop preparations that address (mostly gastrointestinal) intolerance. In contrast, the use of colchicine for primary or secondary prevention of cardiovascular events will require longer term use. Based upon the experience in the treatment of familial Mediterranean fever, the incidence of late side effects of very long term continuous use of colchicine doses of up to 2-3 mg/day appear infrequent [2,4]; however, these incidences of late side effects mostly relate to the use in younger people without vascular disease who are not on statins or at risk of multiple drug use and renal or hepatic dysfunction over years. Long-term studies in people with cardiovascular disease will be important in proving the safety of longterm lower dose therapy in this population.

Study limitations
Although we used rigorous systematic review and meta-analytic methods consistent with PRISMA guidelines including a reproducible and comprehensive literature search strategy, clearly defined inclusion criteria, citation review, data abstraction, and quality assessment of individual studies, and a pre-defined analysis plan, we pooled results from studies enrolling patients with a variety of cardiac diseases. Except in the case of patients with pericarditis, most of the RCTs enrolled patients with related but non-identical cardiovascular diseases (e.g. single or at most two RCTs enrolled patients with stable CAD, acute coronary syndrome or acute stroke, post PTCA, etc.). This illustrates the paucity of RCT data in patients with cardiovascular diseases, and limits the ability to draw definitive conclusions regarding the effectiveness of colchicine in non-pericarditis cardiac patients. However, we minimized pooling data from disparate RCTs by including only RCTs with similar characteristics for specific outcomes. For example, for the primary outcomes, only patients in the cardiovascular disease RCTs were included in the composite cardiovascular event or mortality analyses, and only patients in the pericarditis RCTs had recurrent pericarditis events. Similarly the low number of non-gastrointestinal adverse events limits the ability of our systematic review to rule out increases in other, potentially more serious, adverse events since the upper limit of the 95 % confidence intervals of the RR's of the non-gastrointestinal adverse event pooled analyses all included increased risks that would be clinically significant.

Conclusions
In a wide range of patients with established cardiovascular disease colchicine reduces the composite rate of adverse cardiovascular outcomes. Furthermore, colchicine is associated with a significant reduction in the rates of recurrent pericarditis, post-pericardiotomy syndrome, and peri-procedural atrial fibrillation following cardiac surgery and atrial fibrillation ablation. This suggests that the diverse anti-inflammatory effects of colchicine appear to have benefits in a wide range of cardiovascular diseases. These data support the need for future RCTs of colchicine especially for the secondary prevention of cardiovascular disease.

Combined intervention
Abbreviations: BMS bare metal stent, CABG coronary artery bypass grafting, CAD coronary artery disease, MI myocardial infarction, n/a not available, PTCA percutaneous coronary angioplasty, RCT randomized controlled trial