PREDICTION OF INFLAMMATION IN HEMODIALYSIS PATIENTS USING NEURAL NETWORK ANALYSIS

Background. Numerous hemodialysis patients (HD) suffer from severe, life-threatening inflammation that must be treated to prevent further complications. Early diagnosis of inflammation in HD is highly needed. The present study used matrix metalloproteinase-1 (MMP3) and tissue inhibitor of metalloproteinases-1 (TIMP1) to differentiate between patients with/without inflammation using the neural network analysis (NN). Methods. The positive results of C-reactive protein were used as a criterion for the presence of inflammation in the patients (HD+CRP) versus the negative group (HD-CRP). The NN analysis was used to discriminate between groups using the measured biomarkers. Results. HD+CRP patients have a higher duration of disease, MMP3 and lower calcium than the HD-CRP level is significantly higher, while vitamin D is significantly lower in the HD+CRP group compared with both other groups (all p0.05). TIMP1 is significantly correlated with inorganic phosphate and CRP. In NN#1, the model for the prediction of HD+CRP from HD-CRP has an area under the curve (AUC) of the receiver operating characteristic (ROC) of 0.907 with a sensitivity and specificity 89.2% and a specificity of 100.0%. The top predicting variable for the prediction of HD+CRP is MMP3 (100%), followed by creatinine (87.1%). MMP3 is linked to the pathophysiology of HD, at least through their correlation with the inflammation in HD. In NN#2, the AUC of the ROC for predicting the kidney disease and subsequent HD was 98.9%, with a sensitivity of 100.0% and a specificity of 97.1%. The top four predicting variables for the prediction of high risk of inflammation in HD patients are urea (100%), creatinine (100%), MMP3 (59.7%), and vitamin D (57.1%). Conclusion. The NN analysis may differentiate between HD patients with inflammation from the HD without inflammation. Also, the measured parameters, especially MMP3, TIMP1, and vitamin D are useful as a diagnostic tools for the kidney diseases and inflammation linked with the disease.