348 Background: Maintaining weight (wt) and adequate nutrition during systemic treatment in advanced gastric cancer (G/GEJ) therapy remains a challenge. We investigated the impact of early wt-loss on survival in three phase 3 studies of ramucirumab (R); REGARD (RG), RAINBOW (RB), and RAINFALL (RF) in G/GEJ. Methods: ITT pts were categorized into 2 groups based on their body wt change from start to end of cycle 1 (C1; C = 28 days in RG, RB; C = 21 days in RF): wt-loss < 3% vs ≥3%. Univariate Cox PH models were performed in each individual study to evaluate the effects of body wt change from the start to end of C1 on OS. A pooled meta-analysis stratified by study and a sensitivity analysis of the subgroup of responders was also performed. Results: A total of 311 (RG: 212 in R+BSC; 99 in Placebo (PB)+BSC), 591 (RB: 306 in the R+Paclitaxel (P); 285 PB+P), and 562 (RF: 279 in R+Cape/Cis (CC); 283 in PB+CC) pts with body wt data during C1 were evaluated. The number of pts with wt-loss of ≥3% and < 3% are shown in Table. Pts with wt-loss < 3% during C1 experienced longer OS compared to those with wt-loss ≥3%, irrespective of treatment arms across studies (Table). In pooled treatment arms within each study, the HR for wt-loss group ( < 3% vs ≥3%) was 0.359 (95% CI = 0.254, 0.507), 0.632 (0.497, 0.804), 0.752 (0.608, 0.930) in RG, RB, RF, respectively. In the meta-analysis that combined the 3-studies, univariate Cox PH model stratified by study showed consistent effect of early wt-loss on OS regardless of treatment arm, HR ( < 3% vs ≥3%) = 0.632 (0.546, 0.732). Conclusions: Analysis from three phase 3 studies demonstrates early wt-loss ≥3% during C1 is an important negative prognostic factor for survival in gastric/GEJ cancer. Prospective studies of the relationship of weight preserving nutritional interventions on OS are warranted. Clinical trial information: NCT00917384, NCT01170663, NCT02314117. [Table: see text]