11015 Background: Fatigue in patients with cancer is common and negatively impacts quality of life. Patient-reported outcomes (PROs) assess a patient’s direct report of their symptoms and well-being; research has highlighted their potential to predict cancer outcomes. We examined whether fatigue prior to treatment was associated with AEs. Methods: We analyzed 18 phase II and III cancer treatment clinical trials with baseline fatigue data conducted by SWOG between 1988-2018. The Common Terminology Criteria for Adverse Events was used. Symptomatic AEs were defined as those aligned with the NCI’s PRO-CTCAE; laboratory-based or measurable AEs were designated as objective (hematologic v nonhematologic). Thirteen symptomatic and 14 objective AE categories were examined. Fatigue was derived from multiple scales (e.g., FACT, EORTC QLQ-C30 and PROMIS-Fatigue 7a). Each measure was mapped to a 5-point Likert scale. Binary categories were compared (e.g., < some vs. > = some fatigue). Generalized estimating equations were used with a binomial distribution and logit link, adjusting for age, sex, race, and BMI, with clustering by study. Results: We examined N = 8,040 patients receiving systemic therapy. Among 18 trials (prostate, 5; lung, 3; colorectal, 2; lymphoma, 2; breast, 2; bladder, 1; melanoma, 1; ovarian, 1, pancreas, 1), n = 108,059 AEs were examined. Patient fatigue was present at levels of “a little” or greater in 75.8%, “some” or greater in 42.8%, and “quite a lot” or “very much” in 17.7%. Patients with some or greater fatigue were nearly twice as likely to have severe or worse toxicity (OR = 1.92, 95% CI, 1.53-2.40, p < .001) or life-threatening toxicity (OR = 1.93, 95% CI, 1.44-2.59, p < .001) and nearly threefold more likely to have fatal toxicity (OR = 2.76, 95% CI, 1.37-5.53, p = .004). Similar patterns were seen at a threshold of “quite a lot”. A dose response pattern was evident; patients reporting quite a lot/very much compared to no fatigue were > 5 times more likely to experience fatal toxicity (OR = 5.63, p = .002). These patterns were consistent in symptomatic, objective hematologic, and objective non-hematologic AEs. Conclusions: Baseline fatigue was highly predictive of subsequent risk of AEs. These findings suggest that in an era of precision medicine, patient-reported fatigue may be an important component of determining toxicity risk and could aid in treatment decision making.[Table: see text]