“…Uchihara M, et al [44] Benci A, et al [45] Chan T.M, et al [43] Izopet J, et al [42] Campistol J.M, et al [48] Hanrotel C, et al [52] Espinosa M, et al [49] Fernandez J.L, et al [41] Huraib S, et al [51] Mean overall estimate Koenig P, et al [33] Casanovas-Taltavull T, et al [50] Raptopoulou-Gigi M, et al [39] Pol S, et al [38] Degos F, et al [53] Patients discontinued treatment (%) -10% 10% 20% 30% 40% 50% 60% 70% 80% 90% 0% be the best descriptive analysis, since they assume that all estimates from individual studies are from the same population. The sources of between study heterogeneity are unclear.…”