Mutual Interaction of Clinical Factors and Specific microRNAs to Predict Mild Cognitive Impairment in Patients Receiving Hemodialysis

Abstract

Cognitive impairment (CI) is not uncommon in dialysis patients. Various factors have been implicated. This study aims to examine mutual interaction of various clinical factors for CI in patients receiving hemodialysis. A total of 48 hemodialysis patients in outpatient clinic were recruited from 2015 to 2017. Demographics, circulating uremic toxin concentrations, miRNA concentrations, and nerve injury protein concentrations were collected. Clinical dementia rating (CDR) scores were used to stratify the functional scores of the patients. Receiver operating characteristic (ROC) analysis was used to evaluate diagnostic test performance for predicting dichotomous results, and cumulative ROC analysis was used to examine the combined contribution of clinical factors. CDR scale 0 included 15 patients (mean age, 59.1 years); CDR > 0.5 included 33 patients (mean age, 64.0 years). On cumulative ROC analysis, the major predictors of mild CI were hemoglobin, age, sex, homocysteine, neuron-specific enolase (NSE), and miR-486. The cumulative area under the curve (AUC) on combining hemoglobin, age, and miR-486 was the highest (0.897, 95% confidence interval 0.806–0.988). Two dichotomized variables reached 81.82% sensitivity and 86.67% specificity, with the likelihood ratio for positive and negative results being 6.14 and 0.21, respectively. In conclusion, hemoglobin, age, and miR-486 display high-degree combined effects on mild CI in patients receiving hemodialysis.