Search

Background Dietary protein and phosphorus (P) restriction is the mainstay for nutritional management of chronic kidney disease (CKD). However, adequate restriction levels for cats with early CKD remain unclear. Objectives To investigate responses in cats with early CKD to varying dietary protein, P, and calcium (Ca) : P ratio. Animals Nineteen research colony cats with International Renal Interest Society stages 1‐2 CKD. Methods In an opportunistic longitudinal case study, cats were fed a low protein (59 g/Mcal), low P (0.84 g/Mcal) dry diet (LP‐LP; Ca : P = 1.9) for 18 months and later transitioned onto a moderate protein (76‐98 g/Mcal), moderate P (1.4‐1.6 g/Mcal) dry‐wet diet regimen (MP‐MP; Ca : P = 1.4‐1.6) for 22 months. Fold‐changes in serum creatinine, total Ca (tCa) and P (primary outcomes) and fibroblast growth factor 23 (FGF23) were assessed by linear‐mixed models. Results While feeding LP‐LP, mean serum creatinine decreased (0.87‐fold, 95% confidence interval [CI] 0.81, 0.93, P

Author(s): Bradley, Richard; Tagkopoulos, Ilias; Kim, Minseung; Kokkinos, Yiannis; Panagiotakos, Theodoros; Kennedy, James; De Meyer, Geert; Watson, Phillip; Elliott, Jonathan | Abstract: BackgroundAdvanced machine learning methods combined with large sets of health screening data provide opportunities for diagnostic value in human and veterinary medicine.Hypothesis/objectivesTo derive a model to predict the risk of cats developing chronic kidney disease (CKD) using data from electronic health records (EHRs) collected during routine veterinary practice.AnimalsA total of 106 251 cats that attended Banfield Pet Hospitals between January 1, 1995, and December 31, 2017.MethodsLongitudinal EHRs from Banfield Pet Hospitals were extracted and randomly split into 2 parts. The first 67% of the data were used to build a prediction model, which included feature selection and identification of the optimal neural network type and architecture. The remaining unseen EHRs were used to evaluate the model performance.ResultsThe final model was a recurrent neural network (RNN) with 4 features (creatinine, blood urea nitrogen, urine specific gravity, and age). When predicting CKD near the point of diagnosis, the model displayed a sensitivity of 90.7% and a specificity of 98.9%. Model sensitivity decreased when predicting the risk of CKD with a longer horizon, having 63.0% sensitivity 1 year before diagnosis and 44.2% 2 years before diagnosis, but with specificity remaining around 99%.Conclusions and clinical importanceThe use of models based on machine learning can support veterinary decision making by improving early identification of CKD.