Background: Preclinical data suggest that adiposity activates pro-inflammatory and insulin-dependent pathways which may lead to resistance to TT, e.g., bevacizumab (Bev) or cetuximab (Cet). Only two retrospective trials have studied the relationship between BMI and outcome for mCRC pts who received TT, with conflicting results. Our aim was to compare OS across BMI groups for mCRC pts treated with TT. Methods: Retrospective data, pertaining to clinical characteristics and outcome, were obtained from the South Australian Registry for mCRC from Feb 2006-Oct 2012. The BMI at first treatment was grouped as Normal (N) = 18.5- <25, Overweight (OW) = 25- <30, Obese I(Ob1) = 30- <35, Obese II (Ob2)≥35. Results: Of 1,174 pts, 39% were OW, 15% Ob1 and 7% Ob2. 352 pts received chemo+TT (Bev, Cet, panitumumab (Pan) and/or regorafenib) and 814 chemo alone. Baseline characteristics were similar across all BMI groups except for type of mCRC: N pts were more likely than obese pts to have synchronous CRC (77.9% vs 56-69.7% for obese). On adjustment for age, sex, synchronous disease, metastatic sites, number of lines of chemo and TT, median OS was longer for N versus OW or Ob1 pts with chemo+TT (35.4 vs 24.9 or 22.7 mons, Table) with no difference in OS for chemo alone. Only N gp pts had an improvement in OS on the addition of TT to chemo. On breakdown by type of TT, OW and Ob1 pts had a poorer outcome with Bev but not with EGFR TT. Conclusions: The BMI is an independent predictor for a poorer outcome for OW and OB1 pts with chemo+TT, specifically for pts receiving Bev. The OW and OB1 patients may be a target group for lifestyle and nutrition advice to improve OS with TT. Prospective studies are required to validate this finding.