GRADE Working Group grades of evidence: High quality, further research is very unlikely to change our confidence in the estimate of effect; Moderate quality, further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate; Low quality, further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate; Very low quality, we are very uncertain about the estimate. The corresponding risk (and its 95% confidence interval) is based on the assumed risk (eg, the median/most representative control group risk across studies) in the comparison group and the relative effect of the intervention (and its 95% CI). For this, we used the median event rate in the control group when the number of studies was odd for an outcome and the average of the middle 2 studies when the number of studies was even.

Studies were at low risk of bias (Figure 2).

For all outcomes, visual inspection of the forest plot, the test for heterogeneity, and the I² results of 0% all suggested consistent relative effects.

All studies enrolled relevant representative populations, but we rated down adverse events for indirectness because the studies used very different evaluations of adverse events. Adverse events were measured very differently across trials: old blood adverse events varied from 0.5% to 51%.

Confidence intervals are wide enough that they include for mortality and adverse events the possibility of important benefit and important harm and, for infections, no impact or important harm for younger blood.

We identified no reason to suspect publication bias.