Prediction of outcome after ligation or thin film banding of extrahepatic shunts, based on plasma albumin concentration and hematologic expression of 8 target genes in 85 dogs

Abstract Background In dogs with a congenital extrahepatic portosystemic shunt (EHPSS), outcome after surgical attenuation is difficult to predict. Objectives Develop a minimally invasive test to predict outcome after surgical EHPSS attenuation and establish risk factors for postattenuation seizures (PAS). Animals Eighty‐five client‐owned dogs referred for surgical attenuation of a single EHPSS. Methods mRNA expression of 8 genes was measured in preoperatively collected venous blood samples. Outcome was determined at a median of 92 days (range, 26‐208) postoperatively by evaluating clinical performance, blood test results and abdominal ultrasonography. Multivariable logistic regression was used to construct models predicting clinical and complete recovery. The associations between putative predictors and PAS were studied using univariable analyses. Results Five of 85 dogs developed PAS. Risk factors were age, white blood cell (WBC) count and expression of hepatocyte growth factor activator and LysM and putative peptidoglycan‐binding domain‐containing protein 2. Clinical recovery was observed in 72 of 85 dogs and complete recovery in 51 of 80 dogs (median follow‐up, 92 days). The model predicting clinical recovery included albumin, WBC count, and methionine adenosyltransferase 2 alpha (MAT2α) expression, whereas the model predicting complete recovery included albumin, and connective tissue growth factor precursor and MAT2α expression. The areas under the receiver operating characteristic curves were 0.886 (95% confidence interval [CI]: 0.783, 0.990) and 0.794 (95% CI: 0.686, 0.902), respectively. Conclusions and Clinical Importance Two models were constructed for predicting outcome after EHPSS attenuation using venous blood samples. The model predicting clinical recovery showed the best diagnostic properties. Clinical application requires further validation.