Abstract
Background: Diabetic sensorimotor polyneuropathy (DSP) is a common chronic diabetic complication. Traditionally, DSP was once considered irreversible with a typical loss of axon. However, the superimpose of acquired demyelination on axonal loss in DSP patients has been observed, implying that DSP may be preventable or reversible, particularly within a subgroup of patients exhibiting early-stage acquired demyelination, underscoring the critical importance of identifying early prognostic markers. Methods: We systemically review the literature on the roles of biomarkers in predicting DSP and monitoring the progress. The underlying mechanisms of biomarkers were also discussed. Results: The pathogenesis of DSP is multifaceted, with various pathological mechanisms contributing to its development. Key mechanisms include aberrant glucose metabolism and induction of oxidative stress and inflammation. Several pathological processes, such as disrupted glucose metabolism, nerve damage, impaired microcirculation, genetic variants, and microRNA dysregulation, lead to molecular and protein changes that may be detectable in blood and other biological compartments, thus serving as potential biomarkers for DSP progression. However, the utility of a biomarker depends on its predictive accuracy, practicality, and ease of measurement. Conclusions: Most biomarkers for predicting DSP have demonstrated suboptimal predictive value, and many lack established accuracy in forecasting DSP progression. Consequently, the diagnostic utility of any single biomarker remains limited. A comprehensive combination of biomarkers from various categories may hold incredible promise for accurate detection. As artificial intelligence (AI) techniques, especially machine learning, rapidly advance, these technologies may offer significant potential for developing diagnostic platforms to integrate and interpret complex biomarker data for DSP.