GVHD-specific survival (GSS) has been investigated as a potential study end point to describe the clinical course and outcome of chronic GVHD (cGVHD). However, reaching this end point requires a long observation time. We hypothesized that the time to the first flare-up (FFU) of cGVHD (TTF) can be an alternative statistical end point to GSS. This retrospective study included 96 patients with a diagnosis of cGVHD from a cohort of 119 patients with a prior history of acute GVHD. The median TTF was 73 days after the diagnosis of cGVHD. The 2-year cumulative incidences of first, second and third episodes of flare-up (FU) during courses of cGVHD were estimated as 69.5, 46.4 and 22.1%. Those patients who did not have an episode of FU of cGVHD had 96.0% of 2-years GSS rate, while those with 1 and > or =2 episodes had 50.8 and 46.8%, respectively (P=0.001). Shorter TTF was associated with poor GSS and decreased overall survival. The shorter TTF during the course of cGVHD was significantly associated with extensive cGVHD (P=0.002), Hopkins' risk category (P=0.022) and progressive-type cGVHD (P<0.001) in multivariate analysis. We propose that TTF can be an alternative end point to GSS in cGVHD trials.