5 Composite outcome: Was the primary outcome a combination of outcomes?
By combining several individual outcomes into a composite outcome a trial can increase its ability to detect a difference between groups. However, composite outcomes require careful interpretation of the individual components to avoid making erroneous conclusions.
Checklist Questions
Are the components of the composite outcome all of similar importance to patients? |
Did the components occur with similar frequencies? |
Are the point estimates of treatment effect (HR, OR, RR) similar between each component? Do the 95% CIs overlap? Are they sufficiently narrow? |
Do the components share a similar underlying biological mechanism? |
Clinical importance: Are the components of the composite outcome all of similar importance to patients?
E.g. #1 The primary outcome of DAPA-HF (McMurray JJV et al.), a trial comparing dapagliflozin vs. placebo in patients with heart failure with reduced ejection fraction (HFrEF), was a composite of:
- Hospitalization for heart failure (HF) resulting in intravenous therapy
- Urgent HF visit resulting in intravenous therapy
- Death from cardiovascular (CV) causes
These are all outcomes of significant importance to patients.
E.g. #2 The primary outcome of CONDOR (Chan FKL et al.), a trial comparing celecoxib vs NSAID+PPI, was a composite of:
- Gastrointestinal (GI) bleed
- GI obstruction
- GI perforation
- Clinically significant anemia (decrease in hemoglobin ≥ 20 g/L or decrease in hematocrit ≥10%)
The latter of which was notably less important than the other components.
Statistical contribution: Did the component outcomes occur with similar frequencies?
E.g. #1 Components of the primary outcome in DAPA-HF (McMurray JJV et al.) and their rates for dapagliflozin vs. placebo:
- Hospitalization or urgent visit for HF (10% vs. 14%)
- Death from CV causes (10% vs. 12%)
The most important endpoint (death from CV causes) occurred only slightly less frequently than the other component.
E.g. #2 Components of the primary outcome in CONDOR (Chan FKL et al.) and their rates for celecoxib vs NSAID+PPI:
- GI bleed (0.2% vs. 0.2%)
- GI obstruction (0% for both groups)
- GI perforation (0% for both groups)
- Clinically significant anemia (0.7% vs. 3%)
The greatest contributor of events (clinically significant anemia) drove the difference between groups was also the least clinically important.
Consistency in effect of therapy: Are the point estimates of treatment effect between each component consistent? Do the 95% CIs overlap? Are they sufficiently narrow?
E.g. #1 HRs in DAPA-HF (McMurray JJV et al.) for dapagliflozin vs. placebo:
- Composite HR = 0.74 (95% CI 0.65-0.85)
- Hospitalization or urgent visit for heart failure HR = 0.70 (95% CI 0.59-0.83)
- Cardiovascular death HR = 0.82 (95% CI 0.69-0.98)
Since all the CIs overlap and are sufficiently narrow, there can be greater confidence that the composite outcome is not misleading.
E.g. #2 Relative risk reductions (RRRs) in CONDOR (Chan FKL et al.) for celecoxib vs NSAID+PPI:
Since there is a very large difference in the point estimates, it is better to consider the individual endpoints rather than the composite endpoint.
Biologic rationale: Do the components of the composite outcome share a similar underlying biological mechanism?
An outcome which consists of multiple component endpoints. For example, a cardiovascular composite may include stroke, myocardial infarction, and death.
Hazard ratios are a relative measure of effect. Hazards refer to average instantaneous incidence rate at every point during the trial. This differentiates it from other measures, such as relative risk, which rely only on cumulative event rates. See here for a more detailed discussion.
Odds ratios are the ratio of odds (events divided by non-events) in the intervention group to the odds in the comparator group. For example, if the odds of an event in the treatment group is 0.2 and the odds in the comparator group is 0.1, then the OR is 2 (0.2/0.1). See here for a more detailed discussion.
Relative risk (or risk ratio) is the risk in one group relative to (divided by) risk in another group. For example, if 10% in the treatment group and 20% in the placebo group have the outcome of interest, the relative risk in the treatment group is 0.5 (10% ÷ 20%; half) the risk in the placebo group. See here for a more detailed discussion.
A primary outcome is an outcome from which trial design choices are based (e.g. sample size calculations). Primary outcomes are not necessarily the most important outcomes.
The difference between two relative risks (RRs). If the intervention has a RR of 70% and the comparator a risk of 100%, then the relative risk reduction is 30% (100% - 70%).