Study	Reason for exclusion
Fortune 1987	This study measured oxygen utilisation. The trial planned no clinical outcomes of interest.
Robertson 2014	The trial used a 2 x 2 factorial design in which participants received erythropoietin or transfusion at different thresholds. We were unable to isolate the effect of transfusion from erythropoietin.
Vichinsky 1995	The transfusion trigger was based on the level of sickle haemoglobin (HbS), not the haemoglobin or haematocrit level.
Zygun 2009	This trial measured oxygen utilisation and planned no clinical outcomes of interest.

Abbreviation

HbS = sickle haemoglobin