Not a one-way street: diabetic polyneuropathy can regress even in long-standing type 1 diabetes
Aims: To evaluate the 5-year course of diabetic polyneuropathy (DPN) in type 1 diabetes (T1D), to determine whether DPN can regress, and to identify predictors of progression and regression.
Methods: This single-centre retrospective study included 43 adults with T1D (mean age 38.9 years, mean diabetes duration 23.1 years). Each participant had at least two neuropathy assessments, a median of about 5 years apart. DPN was diagnosed using the Toronto Consensus criteria. Symptoms were assessed with the MNSI questionnaire and signs with the clinical portion of the MDNS, vibration perception threshold (VPT) and thermal perception thresholds. DPN was classified as possible or probable, and probable DPN was further graded as mild, moderate or severe. Cardiovascular autonomic reflex tests were also performed. Participants were classified as progressors, regressors or unchanged according to the change in DPN stage.
Results: The proportion of participants without DPN fell from 34.9% to 20.9%, and possible DPN increased from 27.9% to 41.9%. Probable DPN remained stable at 37.2%. DPN progressed in 37.2% of participants, regressed in 16.3% and was unchanged in 46.5%. VPT and MDNS worsened significantly, and abnormal thermal thresholds increased from 35.1% to 53.8%. In multivariate analysis adjusted for age, sex, BMI and diabetes duration, higher follow-up HbA1c (OR 1.79) and a smaller increase in HDL cholesterol (OR 0.88) independently predicted progression. Lower baseline total cholesterol predicted regression (OR 0.91).
Conclusions: Over 5 years the overall prevalence of DPN increased, but a meaningful proportion of patients improved, some to the complete absence of DPN. Glycaemic control and lipids appear to shape the trajectory of DPN even decades after diabetes onset.
Comments. The dogma that DPN is an inevitably progressive and irreversible condition has recently been challenged by prospective cohorts in type 2 diabetes and in newly diagnosed diabetes. This study extends that observation to adults with long-standing T1D, a population in whom little room for improvement is usually assumed. The clinical message is encouraging: continued metabolic intervention after a DPN diagnosis is still worthwhile. The finding that HDL dynamics were related to VPT deterioration supports the growing interest in lipids as modifiable drivers of nerve damage, beyond glycaemic control. A particular strength of the study is the careful, multimodal phenotyping performed by an experienced team. Validated, standardised tools and age-adjusted reference values were used for quantitative sensory testing, and gold-standard cardiovascular autonomic reflex tests complement the characterization. The approach of grading DPN severity, and not only its presence, makes it possible to capture subtle changes in both directions that would be missed by a simple yes/no diagnosis. The study also provides a concise and useful overview of the available longitudinal data on the natural history of DPN in both types of diabetes. These promising observations open an important perspective, and future larger prospective studies will help to confirm them and to identify biomarkers of different DPN trajectories.
Aleksandra Araszkiewicz
Reference. Pertile P, D'Ippolito I, Mascambroni S, D'Amato C, Andreadi A, Lauro D, Spallone V. Heterogeneous course of diabetic polyneuropathy in type 1 diabetes over 5 years: is regression possible? Endocrine. 2026 Apr 27;91(1):147. doi: 10.1007/s12020-026-04613-8. PMID: 42043615; PMCID: PMC13121219.
