Thrombocytopenia is common (40-65%) and potentially serious in myelodysplastic syndromes (MDS). A systematic review was conducted to determine the safety and efficacy of adding a thrombopoietin-receptor (THPO-R) agonist to standard MDS treatment. MEDLINE, EMBASE and CENTRAL databases were searched. We included randomized controlled trials comparing a THPO-R agonist to placebo. A meta-analysis of the effects was performed. Endpoints included bleeding and platelet transfusion rates, risk of progression to acute myeloid leukaemia (AML) and mortality. Three hundred and eighty four patients from five trials were included, four using romiplostim and one using eltrombopag. Overall, the relative risk (RR) of bleeding with romiplostim versus placebo was 0·84 [95% confidence interval (CI): 0·57-1·24]. However, compared to placebo, romiplostim significantly decreased the exposure-adjusted bleeding rate (RR 0·92; 95% CI: 0·86-0·99), as well as the exposure-adjusted platelet transfusion rate (RR 0·69; 95% CI: 0·53-0·88). The RR of AML progression with romiplostim was 1·36 (95% CI: 0·54-3·40), however the outcome data were judged as higher risk of bias. Romiplostim is promising in its ability to decrease patient-important outcomes: bleeding and platelet transfusion need. Although the risk of AML progression was not increased, due to unclear risk of bias in the data, this safety concern is difficult to assess. Therefore, romiplostim cannot yet be routinely recommended. Early eltrombopag data is promising.