Table 2

Summary of findings table

Anticipated absolute effects
OutcomesNo of participants (studies) follow-upQuality of the evidence (GRADE)Relative effect (95% CI)Risk with any control treatmentRisk difference with colchicine (95% CI)
Mortality (all-cause)4174 (30 studies) 0.5–14 years⊕⊕⊕⊝ MODERATE* due to imprecisionRR 0.94 (0.82 to 1.09)193 per 100012 fewer per 1000 (from 37 fewer to 17 more)
Subgroup: Patients with high cardiovascular risk1230 (4 studies) 0.5–3 years⊕⊕⊕⊝ MODERATE* due to imprecisionRR 0.54 (0.26 to 1.14)32 per 100015 fewer per 1000 (from 24 fewer to 4 more)
Mortality (cardiovascular)1132 (7 studies) 0.5–14 years⊕⊕⊕⊝ MODERATE* due to imprecisionRR 0.34 (0.09 to 1.21)27 per 100018 fewer per 1000 (from 24 fewer to 6 more)
Subgroup: Patients with high cardiovascular risk754 (2 studies) 0.5–3 years⊕⊕⊝⊝ LOW†‡ due to imprecisionRR 0.25 (0.02 to 2.66)31 per 100031 fewer per 1000 (from 30 fewer to 51 more)
Myocardial Infarction (total)§652 (2 studies¶) 3 years⊕⊕⊕⊝ MODERATE** due to imprecisionRR 0.20 (0.07 to 0.57)58 per 100046 fewer per 1000 (from 25 fewer to 54 fewer)
Subgroup: Patients with high cardiovascular risk§532 (1 study) 3 years⊕⊕⊕⊝ MODERATE** due to imprecisionRR 0.20 (0.07 to 0.57)Study population
72 per 100058 fewer per 1000 (from 31 fewer to 67 fewer)
Extrapolated 1-year risk
25 per 100020 fewer per 1000 (from 11 fewer to 23 fewer)
Adverse event (total)‡‡‡1313 (11 studies) 0.5–14 years⊕⊝⊝⊝ VERY LOW*,‡‡,§§ due to risk of bias, imprecision, publication biasRR 1.52 (0.93 to 2.46)Study population
89 per 100046 more per 1000 (from 6 fewer to 130 more)
Assumed 1-year risk
89 per 100046 more per 1000 (from 6 fewer to 130 more)
Adverse event (gastrointestinal)1258 (11 studies) 0.5–14 years⊕⊕⊝⊝ LOW§§,¶¶ due to inconsistency, publication biasRR 1.83 (1.03 to 3.26)Study population
132 per 1000110 more per 1000 (from 4 more to 299 more)
Assumed 1-year risk
132 per 1000110 more per 1000 (from 4 more to 298 more)
Adverse event (serious)472 (4 studies) 824 patient-years⊕⊕⊝⊝ LOW***,††† due to imprecision, publication biasNot estimableNo illustration of comparative risks due to very uncertain assumed risks.
Heart failure (total)426 (3 studies) 0.5–3 years)⊕⊕⊝⊝ LOW‡,¶,** due to imprecision, publication biasRR 0.62 (0.1 to 3.88)No illustration of comparative risks due to very uncertain assumed risks.
Stroke (total)874 (3 studies) 0.5–3 years⊕⊕⊝⊝ LOW‡,¶,** due to imprecision, publication biasOR 0.38 (0.09 to 1.7)No illustration of comparative risks due to very uncertain assumed risks.
  • The basis for the assumed risk is the control group risk across studies. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). For balancing benefits and harms using absolute risk measures (1-year risk), we extrapolated the 1-year risk of myocardial events assuming that the risk is constant over the entire follow-up and we assumed that all adverse events observed over the entire follow-up accumulate within the 1st year of treatment.

  • *CI is compatible with patient-relevant benefit and harm.

  • †Small number of events therefore downgraded for imprecision.

  • ‡Substantial imprecision because CI compatible with major harm and major benefit.

  • §Most evidence provided by a single study.

  • ¶One study without events.

  • **Effect based on small number of events.

  • ‡‡High risk for attrition bias in 5 of 10 studies.

  • §§Visual inspection of funnel plot shows asymmetry; lack of small studies reporting lower adverse event rates with colchicine than with comparator (the rate of adverse events with colchicine may appear too high due to bias).

  • ¶¶Substantial between-study heterogeneity (I2=74%) without plausible explanation.

  • ***Only four studies reported on serious adverse events (zero events in approximately 800 patient-years).

  • †††No indication for publication bias, but reporting quality very limited. In many other studies events occurred (eg, deaths) that could be regarded as serious adverse events.

  • ‡‡‡No study in patients with high cardiovascular risk reported on total adverse events.

  • GRADE, Grading of Recommendations Assessment, Development and Evaluation; RR, relative risk.