The approval by FDA of the first two thrombopoietin-receptor agonists (TPO-RA), the peptibody romiplostim followed by the small non-peptide molecule eltrombopag, represented a paradigm shift in the treatment of immune thrombocytopenia (ITP), from reducing the destruction of autoantibody opsonized platelets to maximally increasing platelet production beyond the potential of the endogenous thrombopoietin homeostatic mechanism. The experience accumulatedf so far indicates that TPO-RA can induce a variably durable response in more than 60–70% of patients nonresponsive to initial therapy. In addition, the response may be maintained even for several months after stopping treatment in 10–15% of cases. The initial fears of favoring myelofibrosis or myeloid clonal disorders have been largely dismissed but still require clinical surveillance. Their use is not exempted from an increased risk of thrombosis, particularly in the elderly or in subjects with additional risk factors. This risk seems higher with avatrombopag, another small nonpeptide molecule TPO-RA approved in 2019, but no definite conclusions can be drawn. Switching from one TPO-RA to another may be effective in 20–30% of cases. Understandably, TPO-RA are the most widely used agents for maintenance treatment in patients nonresponsive to first-line therapy, followed by rituximab and splenectomy. Quite unexpectedly, due to the rarity of ITP, many new agents are undergoing clinical development. These agents are targeted to specific steps of the pathogenic pathway, from quenching altered autoimmunity to reducing platelet destruction. BTK inhibitors, various anti-CD38 monoclonal antibodies, and monoclonal antibodies directed against BAFF-R or Fc fusion proteins eliminating BAFF and APRIL are aimed at reducing or aborting the altered autoimmunity. SYK inhibitors, neonatal Fc receptor antagonists, and complement inhibitors have been designed to reduce peripheral platelet destruction. Among SYK inhibitors, fostamatinib, approved in 2018 for chronic ITP, is of proved efficacy and without an apparent risk of increasing the risk of thrombosis. It could be a valuable choice for a patient failing previous treatments including TPO-RA and/or splenectomy. Because of the lack of trials directly comparing the different agents, selecting a patient-tailored therapy could become very difficult, with the risk of patients’ prolonged medicalization. It is appropriate to remember that splenectomy remains the only approach that can achieve substantial improvement or a life-long complete response in up to 70% of patients. Consequently, its role should be evaluated in patients with an ITP unresponsive to several different treatments.