Abstract
BACKGROUND: Diabetic sensorimotor polyneuropathy is an important risk factor for foot ulceration and amputation. Thus, patients with diabetes should be screened for this disorder according to local guidelines. An obstacle to the diagnosis of this disease may be the lack of unified diagnostic criteria due to the lack of properly validated scales used for assessment. AIM: To validate both sections (A and B) of the Michigan Neuropathy Screening Instrument (MNSI) in Polish (PL) patients with diabetes. METHODS: A cross-sectional study using a test (A1, B1) and re-test (A2, B2) formula was performed in 80 patients with diabetes. The gold standard used for neuropathy detection was a nerve conduction study (NCS) which was performed in all participants. Reliability of the MNSI-PL was assessed using the Cronbach's alpha, Kuder-Richardson formula 20 (KR-20), split-half reliability, the Gottman split-half tests, and correlation between first and second half was accessed. Stability was assessed using an intraclass correlation coefficient (ICC). For external validation, we used simple linear correlation, binomial regression, and agreement between two different tools using a Bland-Altman plot analysis. RESULTS: The scale was internally consistent (Cronbach's alpha for the full scale: 0.81 for A and 0.87 for B). MNSI-PL scores in test/retest showed high stability (ICC = 0.73 for A and ICC = 0.97 for B). The statistically important correlations between MNSI-PL and NCS were found for B1, B2, and A1 (P < 0.005). The cut-off points of ≥ 3 for section A (sensitivity of 90%-100%; specificity of 33%-40%) and ≥ 2 for section B (sensitivity of 81%-84%; specificity of 60%-70%) were obtained during neuropathy detection. CONCLUSION: The MNSI-PL is a reliable and valid instrument in screening for diabetic neuropathy.