Heterogeneous course of diabetic polyneuropathy in type 1 diabetes over 5 years: is regression possible?

To evaluate the course of diabetic polyneuropathy (DPN) and changes in DPN severity over 5 years in type 1 diabetes (T1D) and the possibility of regression in DPN. A retrospective analysis was conducted on 43 participants with T1D (age 38.9 ± 12.5, duration 23.1 ± 13.1 years, 15 males). DPN was diagnosed using symptoms (MNSI-Q ≥ 4) and signs [MDNS ≥ 7 or abnormal vibration (VPT) or thermal perception thresholds (TPT)] and classified as probable when both symptoms and signs were present and possible when only symptoms or signs were observed. Probable DPN was further graded as mild, moderate or severe according to the number of signs identified. Participants were categorized as progressors, regressors, or unchanged based on worsening, improvement or stability of DPN stage. Participants without DPN decreased from 34.9% to 20.9%, those with possible DPN increased from 37.2% to 41.9%, while probable DPN remained at 37.2%. DPN stage progressed in 37.2%, regressed in 16.3%, and was unchanged in 46.5%. Progressors exhibited higher baseline cholesterol compared to regressors (P = 0.027), higher follow-up (FU) HbA1c compared to regressors and unchanged participants (P = 0.037), and a lower FU-baseline change (∆) in HDL than both regressors (P = 0.029) and unchanged (P = 0.013). Multivariate analysis showed that FU HbA1c (OR 1.79, P = 0.047) and ∆ HDL (OR 0.88, P = 0.030) were predictors of progression, while baseline cholesterol predicted regressors (OR 0.91, P = 0.030). Over 5 years the overall prevalence of DPN increased; however, regression to less severe DPN or complete absence of DPN occurred. These findings indicate that regression of DPN is still possible even in longstanding T1D. Changes in lipids and HbA1c may predict DPN course, suggesting that continued intervention remains beneficial well after a DPN diagnosis.