Abstract
BACKGROUND: Continuous renal replacement therapy (CRRT) is routinely used in human patients with acute kidney injury (AKI) but studies in dogs are scarce. OBJECTIVE: To describe CRRT in dogs and assess the utility of a previously validated scoring system for dogs with AKI undergoing hemodialysis, and the Acute Patient Physiological and Laboratory Evaluation (APPLE(Full)/APPLE(Fast)) scores, for outcome prediction. ANIMALS: Thirty, client-owned dogs. METHODS: Cases were retrospectively reviewed. Prognostic scores were calculated upon admission and before CRRT initiation. The CRRT effluent dose followed the KDIGO guidelines. Receiver operating characteristic curves (ROCC) were constructed to evaluate the prognostic utility of these scores. RESULTS: Median (IQR) serum creatinine (mg/dL) at CRRT initiation, at discharge, and 3 months after discharge were 9.4 (7.4), 3.4 (1), and 1.3 (0.3) respectively. Median (IQR) treatment duration and total number of treatments were 24 (18.5) h and 2 (2) treatments, respectively. The prescribed median (IQR) CRRT effluent dose was 29 (18.5) mL/kg/h. Median (IQR) overall time-average concentration for urea and creatinine were 92 (60) mg/dL and 3.7 (1.7) mg/dL, respectively. The normalized weekly median (IQR) standardized Kt/V was 2.41 (2.29). Eleven dogs (37%) survived to discharge/3-months after treatment. Areas under the ROCC for the APPLE(Full)/APPLE(Fast) scores before CRRT initiation were 0.99 (95% CI, 0.99-1.00) and 0.91 (95% CI, 0.81-1.00), respectively. Optimal cutoff points were < 35 for the APPLE(Full) and < 23 for the APPLE(Fast), yielding sensitivities/specificities of 100% (95% CI, 74.12%-100.0%)/94.7% (95% CI, 75.36%-99.73%) and 90.9% (95% CI, 62.26%-99.53%)/78.95% (95% CI, 56.67%-91.49%), respectively. CONCLUSION: The APPLE scores, unlike clinicopathological findings or the Segev score, proved to be a highly discriminatory prognostic tool. Additionally, the human-derived, KDIGO guideline-based CRRT protocol proved safe and efficacious in dogs undergoing CRRT.