Corneal confocal microscopy reflects axonal degeneration but not neuropathic pain in (pre)diabetes: insights from the Maastricht study
Aims: Corneal confocal microscopy (CCM) is increasingly used as a non-invasive biomarker of small-fibre damage in diabetes. However, its ability to distinguish painful from painless neuropathy remains uncertain. This study aimed to investigate, in a large population-based cohort, whether wide-field CCM (WF-CCM) parameters are associated with (1) electrophysiological nerve abnormalities and (2) neuropathic pain in individuals with normal glucose metabolism, prediabetes and type 2 diabetes.
Methods: Cross-sectional data from 3,425 participants of the Maastricht Study were analysed. WF-CCM was performed using fully automated image analysis to quantify corneal nerve fibre length (CNFL), fibre density (CNFD), branch density (CNBD), fractal dimension and tortuosity. Peripheral nerve function was assessed by nerve conduction studies, and composite scores were derived for general electrophysiological abnormalities and for axonal degeneration (based on motor and sensory amplitudes). Neuropathic pain was defined as a DN4 interview score ≥3. Multivariable linear regression analyses were performed, including adjustment for demographic and metabolic covariates.
Results: WF-CCM parameters were significantly associated with the axonal degeneration composite score. In fully adjusted models, lower CNFL, reduced fractal dimension and altered tortuosity remained independently associated with axonal degeneration. In contrast, no consistent associations were observed between CCM parameters and the general EMG composite score. Importantly, no significant associations were found between neuropathic pain and CCM parameters in the total cohort. Subgroup analyses across glucose metabolism categories yielded similar results. Sensitivity analyses confirmed the robustness of the association between axonal degeneration and CCM parameters, whereas the lack of association with neuropathic pain persisted.
Conclusions: In this study, WF-CCM parameters reflect axonal degeneration but are not associated with neuropathic pain. These findings suggest that CCM captures structural correlates of peripheral nerve loss but has limited value in differentiating painful from painless neuropathy.
Comments: This study provides important clarification in an area of ongoing debate. While CCM has been established as a sensitive marker of small-fibre damage in diabetes, its relationship to neuropathic pain has remained inconsistent. By leveraging a large, well-phenotyped population, this analysis substantially strengthens the evidence base. A key novelty is the distinction between different electrophysiological constructs. The lack of association between CCM and a broad EMG composite score, contrasted with the association with an axonal degeneration–focused score, suggests that CCM primarily mirrors axonal loss. This refines our understanding of what CCM actually measures within the spectrum of diabetic neuropathy. Equally relevant is the robust absence of association between CCM metrics and neuropathic pain. These findings align with several prior reports but contradict others that suggested increased branching or tortuosity in painful neuropathy. The present data indicate that CCM metrics are insufficient to explain pain phenotypes. This reinforces the concept that diabetic neuropathic pain reflects functional alterations (e.g. hyperexcitability, central sensitisation, ion channel dysregulation) rather than simply the extent of small-fibre loss. Strengths of the study include the large sample size, inclusion of prediabetes, fully automated WF-CCM analysis reducing observer bias, and comprehensive adjustment for confounders. Limitations are the cross-sectional design precluding causal inference. Neuropathic pain was defined by DN4 rather than phenotyping or quantitative sensory testing, and the cohort had short diabetes duration with few cases of advanced neuropathy restricting generalisability. This study refines the positioning of CCM as a structural biomarker of axonal degeneration across dysglycaemia, but not useful to stratify painful versus painless neuropathy.
Zoltan Kender
Reference. Borbjerg MK, Mokhtar S, Sutedja N, Koster A, Mørch CD, Berendschot TTJM, Schaper N, Ejskjaer N, Røikjer J. Association of corneal nerve parameters with nerve abnormalities and neuropathic pain in prediabetes and type 2 diabetes: the Maastricht Study. Diabetologia. 2026 Feb 10. doi: 10.1007/s00125-026-06676-8. Epub ahead of print. PMID: 41665664.
