Aim: To study the associations between baseline blood glucose levels (BGL), glycemic variability and clinical outcomes in patients with severe acute kidney injury (AKI) receiving continuous renal replacement therapy (CRRT). Methods: We performed a secondary analysis of the Randomized Evaluation of Normal versus Augmented Level of RRT (RENAL) study. A multivariate Cox regression model was used to assess the association between baseline BGL, glycemic variability and clinical outcomes. The primary outcome was all-cause mortality, and secondary outcomes were duration of hospital and intensive care unit (ICU) stay. Results: Baseline BGL data were available in 1404 out of 1508 patients from the RENAL study. Among them, 627 patients died within 90 days of randomization. Compared to patients in the second quartile (BGL 5.8-7.2 mmol/L), patients in the first quartile (BGL < 5.8 mmol/L) had increased mortality rate (90-day HR 1.48; p = 0.001; 28-day HR 1.47; p = 0.042). However, there were no significant differences in ICU and hospital length of stay (LOS) (p = 0.82 and p = 0.33, respectively). Glycemic variability data were from 1345 out of 1404 patients who had data for BG values within 28 days. Higher coefficient of variation (CV) (HR 1.02; P trend = 0.002) and standard deviation value (SD) (HR 1.29; P trend = 0.027) were associated with higher risk of death at day 90. Conclusions: We identified a low BGL within the normal physiological range at baseline and greater CV and SD values as significant modifiable risk factors for mortality in severe AKI patients in ICU, which may be a target for intervention.