Construction and application of predictive model of secondary mild cognitive impairment in patients with painful diabetic neuropathy

Abstract

Objective To build predictive model of secondary mild cognitive impairment (MCI) in patients with painful diabetic neuropathy (PDN), and analyze its apply. Methods The patients with PDN were consecutively selected from March 2013 to March 2016. The relevant clinical data were recorded, and the patients were followed up for 1 year. According to the results of follow-up, secondary MCI risk indicators were predicted, and the time window of adverse outcomes event was validated. Results A total of 82 PDN patients completed the study, and secondary MCI occurred in 16 cases. Sixty-six cases had not secondary MCI. The Cox regression model multivariate analysis results showed that the independent influencing factors of secondary MCI was course of PDN, brief pain inventory (BPI) score and neuron-specific enolase (NSE) in patients with PDN (HR= 1.238, 1.336 and 1.450; P<0.05). The secondary time window of the MCI in PDN patients with the course of PDN ≥ 3.367 years, BPI score ≥ 4.704 scores and NSE ≥ 7.420 μg/L was shorter, in whom BPI score and NSE had a higher evaluation ability. Conclusions The courses of PDN, BPI score and NSE are independent influencing factors of secondary MCI in PDN patients, and the BPI score ≥ 4.704 scores and NSE ≥ 7.420 μg/L have a higher evaluation ability. Key words: Diabetes mellitus, type 2; Diabetic neuropathies; Cognition disorders; Forecasting