Early weight loss is an independent risk factor for shorter survival and increased side effects in patients with metastatic colorectal cancer undergoing first‐line treatment within the randomized Phase III trial FIRE‐3 (AIO KRK‐0306)

Body weight loss is frequently regarded as negatively related to outcomes in patients with malignancies. This retrospective analysis of the FIRE‐3 study evaluated the evolution of body weight in patients with metastatic colorectal cancer (mCRC). FIRE‐3 evaluated first‐line FOLFIRI (folinic acid, fluorouracil and irinotecan) plus cetuximab or bevacizumab in mCRC patients with RAS‐WT tumors (ie, wild‐type in KRAS and NRAS exons 2‐4). The prognostic and predictive relevance of early weight loss (EWL) regarding patient outcomes and treatment side effects were evaluated. Retrospective data on body weight during first 6 months of treatment were evaluated (N = 326). To correlate with efficacy endpoints and treatment side effects, patients were grouped according to clinically significant EWL ≥5% and <5% at Month 3. Age constituted the only significant predictor of EWL following a linear relationship with the corresponding log odds ratio (P = .016). EWL was significantly associated with the incident frequencies of diarrhea, edema, fatigue, nausea and vomiting. Further, a multivariate analysis revealed EWL to be an independent negative prognostic factor for overall survival (32.4 vs 21.1 months; hazard ratio [HR]: 1.64; 95% confidence interval [CI] = 1.13‐2.38; P = .0098) and progression‐free survival (11.8 vs 9.0 months; HR: 1.72; 95% CI = 1.18‐2.5; P = .0048). In conclusion, EWL during systemic treatment against mCRC is significantly associated with patient age. Patients exhibiting EWL had worse survival and higher frequencies of adverse events. Early preventative measures targeted at weight maintenance should be evaluated, especially in elderly patients being at highest risk of EWL.

regarding patient outcomes and treatment side effects were evaluated. Retrospective data on body weight during first 6 months of treatment were evaluated (N = 326).
To correlate with efficacy endpoints and treatment side effects, patients were grouped according to clinically significant EWL ≥5% and <5% at Month 3. Age constituted the only significant predictor of EWL following a linear relationship with the corresponding log odds ratio (P = .016). EWL was significantly associated with the incident frequencies of diarrhea, edema, fatigue, nausea and vomiting. Further, a multivariate analysis revealed EWL to be an independent negative prognostic factor for overall survival (32.

What's new
When patients with metastatic colorectal cancer (mCRC) rapidly lose weight early in the course of treatment, that often forebodes a negative outcome. Here, the authors examined changes in body weight in the first 3 months of treatment. Older patients had the highest risk of extreme early weight loss (greater than 5%). This weight loss was correlated with adverse events such as nausea, vomiting, and diarrhoea, and also with an 11-month reduction in overall survival. These results should increase oncologists' awareness of patients' body weight change early in treatment and encourage intervention from dietitians to help prevent weight loss.

| BACKGROUND
With over 1.8 million newly diagnosed cases in 2018, colorectal cancer (CRC) is the second most common malignancy for females and third most common malignancy for males worldwide. 1 The 5-year survival rate for the metastatic colorectal cancer (mCRC) is estimated at less than 12.5%. 2 With the introduction of modern targeted therapy, median overall survival (OS) times exceeding 30 months have been reached in mCRC. [3][4][5] However, side effects occur in almost all patients and do compromise quality of life and impair physical performance. [6][7][8] Literature shows that patients exhibiting loss of body weight during antineoplastic treatment have been identified being at higher risk for treatment side effects. [9][10][11] The frequency of weight loss prior to chemotherapy reported in the literature ranges from 31% for  Patients' inclusion and exclusion criteria are presented in Figure S1.
Regarding the design, conduct of the trial, the full study population, treatment schedules, concordance with the Declaration of Helsinki and approval of ethics committees were reported previously. 25

| Patients
In light of the adoption of RAS analyses as an improved biomarker of response to cetuximab therapy and its evaluation in FIRE-3, we decided to perform the present analyses in the RAS-WT population with unresectable mCRC as previously described. 3   lost an average of 7.8 kg (95% CI = 6.8-8.7, P < .001). The difference between the two groups was 9.1 kg from baseline to Month 3 (95% CI = 8-10.1, P < .001) ( Figure 1B).  (Figure 2A). Of note, patient age exhibited a linear effect on log-odds ratio regarding the occurrence of EWL (P = .016) ( Figure 2B).

| Adverse events
Among all patients with available body weight data, the number of patients receiving full 3 months of treatment was 307 (93.9%). Only these patients were evaluated to allow for comparison of adverse event rates.
A significant relationship between EWL and side effects after 3 months of treatment was observed as follows: diarrhea, edema, fatigue, nausea and vomiting ( Table 2). Of note, comparable results were observed for side effects after 1 month of treatment (Table S2). From baseline to Month 1, EWL was associated with a higher risk of diarrhea, edema and fatigue (Table S2).

| The prognostic relevance of weight loss
To control for guarantee-time bias, only patients who had completed at least 3 months of treatment were considered. In Kaplan-Meier  Figure 3B; PFS: 9.0 vs 11.8 months, P = .0022, Figure 4). showed that EWL was not significantly associated with overall response rate (ORR) (HR 0.5, 95% CI = 0.21-1.24, P = .12, Figure S3), most probably reflecting the disadvantages of this parameter in the assessment of targeted first-line treatment in mCRC patients.

| The predictive relevance of weight loss
To evaluate the relevance of EWL to predict a treatment benefit of FOLFIRI plus either bevacizumab or cetuximab, we compared EWL subgroups within each treatment arm. Here, no formal interaction of treatment arm with EWL was detected (P = .65) ( Figure S4).

| DISCUSSION
We investigated the evolution of body weight during standard firstline treatment for mCRC and evaluated the prognostic and predictive relevance of EWL, that is, weight loss evaluated after 3 months. To this end, we used data from the large Phase III trial FIRE-3 comparing FOLFIRI plus cetuximab with FOLFIRI plus bevacizumab in RAS-WT mCRC patients. 3,25 An important finding of FIRE-3 was prolonged OS favoring FOLFIRI/cetuximab in the absence of significant differences in PFS and ORR.
Body weight loss according to CTCAE assessment is different from nutritional assessment in the clinic, which contains more evaluations regarding the overall nutritional status with consideration of quality of life. In our cohort, all patients were categorized as Grade 1 (body weight loss 5% to <10%) or Grade 2 (body weight loss 10% to <20%) according to CTCAE. No patients were categorized as Grade 3 (body weight loss ≥20%). We first examined the evolution of body weight during the first 6 months of treatment within FIRE-3. Here, we found that patients lost most weight during the first month of treatment (an average of 0.7 kg), whereas patients slowly recovered hereafter with a weight gain of average 0.38 kg/mo ( Figure 1A). To evaluate the impact of weight loss on treatment side effects and patient outcome, we divided patients according to early and clinically relevant weight loss ≥5% or <5% after 3 months of treatment (EWL). 15,19,31,32 Patients with EWL ≥5% showed an average maximum weight loss of 1.1 kg/mo during first 6 month of treatment ( Figure 1B Our study is certainly limited by its retrospective nature. The patient number in our cohort gradually decreased due to discontinuation of treatment. Patients' dietary behaviors, situations or environments that could promote WC were not recorded. Furthermore, baseline and follow-up data regarding body weight were evaluable among 81.5% of the patients (326 out of 400). In consideration of the guarantee-time-bias, we did a landmark analysis to rule out that EWL merely indicated treatment duration. Here, we decided to focus our investigation on impact of weight loss at Month 3 on survival since this is the most recognized time point. 20,39 In addition, we admitted that dose intensity could be the confounding factor associated with prognosis. Further prospective study with consideration of dose intensity is needed to validate our results. 40 In conclusion, EWL ≥5% from baseline to Month 3 is an independent prognostic biomarker for patient survival and adverse events in