Little is known about the process of deciding to discontinue continuous renal replacement therapy (CRRT) in patients with acute kidney injury (AKI) and the impact of CRRT duration on outcomes. We report the clinical parameters of prolonged CRRT exposure and predictors of doubling of serum creatinine or need for dialysis at 90days after CRRT with propensity score matching, including covariates that were likely to influence patients in the prolonged CRRT group. Among 104 survey responders, most use urine output (87%) to guide CRRT discontinuation, 24% use improvement in clinical or hemodynamic status. In the cohort study, of 854 included patients, 465 participated in the assessment of kidney recovery. Patients with prolonged CRRT had higher SOFA scores (11.9 vs. 11.2) and were more likely to be mechanically ventilated (99% vs. 84%) at CRRT initiation compared to patients without prolonged CRRT, p-value < 0.05. In multivariable logistic regression, daily urine output and cumulative fluid balance leading to CRRT discontinuation or day seven were independently associated with lower [OR 0.87 per 200ml/day increase] and higher odds [OR 1.03 per 1-L increase] of requiring prolonged CRRT, respectively. After propensity score matching, prolonged exposure to CRRT was independently associated with increased risk of doubling serum creatinine or dialysis at 90days, OR 3.1 (95% CI 1.23-8.3 p = 0.017). Resolution of critical illness and signs of kidney recovery are important factors when considering CRRT discontinuation. Prolonged CRRT exposure may be associated with less chance of kidney recovery among survivors.