6525 Background: Ethical clinical research requires fair subject selection and a favorable risk-benefit ratio. Enrollment criteria can unfairly exclude patients if they are not based in safety. There are few data regarding variability within criteria, which can also represent unfair exclusion, or for acute myeloid leukemia (AML). Methods: We assessed enrollment criteria use and variability for frontline phase II and III therapeutic AML trials from 2010-2019 listed on Clinicaltrials.gov. Drug safety profiles were obtained from protocols, publications, and FDA labels. FDA publications guided a standardized abstraction and coding process. Associations between criteria and safety profiles were identified using X2 and t-tests. Regression modeling identified differences in infectious disease (ID) exclusion by sponsor adjusted for trial phase, year, size, and safety profiles. Results: 190 trials were eligible; 159 (84%) were phase II and 32 (17%) phase III. Trials were sponsored by the National Institutes of Health (NIH; 50; 26%), industry (109; 57%), and academic investigators (AI; 108; 57%). Exclusions and variability are shown in the table. ID exclusions were for HIV (96; 51%) and hepatitis B (60; 32%) and C (64; 34%). Older adults were considered as over 60 (42 trials; 39%), 65 (23; 21%), 70 (13; 12%), or 75 years (18; 17%). There was no association between QTc limits and the use of QTc prolonging drugs (criteria presence p=0.2; mean 450 vs 474 ms, p=0.3) or renal function limits and nephrotoxic drugs (presence p=0.2; mean 1.65 vs 1.67 ULN, p=0.8). Hepatitis B and C exclusions were associated with the use of drugs linked to viral reactivation (OR 2.1 [95% CI 1.1,3.9], 2.0 [95% CI 1.1,3.7]). This was not the case for HIV (OR 1.2 [95% CI 0.7,2.2]). Industry sponsorship was independently associated with hepatitis B and C exclusion (OR 3.0 [95% CI 1.2,7.3], 3.3 [95% CI 1.4,4.7]) and AI sponsorship with HIV exclusion (OR 2.5 [95% CI 1.4,5.5]). Conclusions: There is substantial variability in enrollment criteria for AML clinical trials. Many criteria are related to risks, but some exclusions and variability are not clearly based in safety. Avoiding arbitrary exclusions will likely enhance trial representativeness and enrollment rates. [Table: see text]