Corneal confocal microscopy for diagnosis of diabetic peripheral neuropathy: an analysis of patients with diabetes screened as part of the South Manchester Diabetic Retinopathy Screening Service

Abstract

Background and Aims: Quantitative assessment of small nerve fibre damage is key to the early diagnosis of diabetic peripheral neuropathy (DPN) and assessment of its progression. Corneal confocal microscopy (CCM) is a non-invasive, in-vivo diagnostic technique that provides an accurate surrogate biomarker for small fibre neuropathy. Its diagnostic efficacy has been previously validated in several studies. This thesis uses CCM images obtained, for the first time, in a large cohort of patients whose CCM examinations were undertaken during retinopathy screening in primary care. The following were the primary aims of the study: 1. To determine the prevalence of diabetic peripheral neuropathy, as defined by CCM parameters in a cohort of people with diabetes 2.To assess whether abnormalities in corneal nerve fibre morphology are present during the first two years following diabetes diagnosis. 3. To assess whether abnormalities in corneal nerve morphology are present prior to any retinopathy, defined as grade 1 or more. 4. To assess whether abnormalities in corneal nerve morphology are present prior to clinical evidence of diabetic neuropathy, as defined by diabetic neuropathic symptom (DNS) scoring of 1 or more The hypotheses for these main aims were that firstly, the prevalence of diabetic peripheral neuropathy, defined using CCM parameters would be lower in this population in comparison to previous CCM studies using patients under the hospital eye service to determine prevalence of DPN. There will be evidence of abnormalities in corneal nerve fibre morphology in some, but not all, patients with diabetic disease duration of less than or equal to 2 years, patients with retinopathy and maculopathy grade 0 and patients with a DNS score of 0. Methods: In this retrospective, primary care, cross-sectional study, 427 patients with diabetes (18 T1DM, 407 T2DM, 2 unknown) and 40 healthy controls underwent quantification of corneal nerve parameters using both automated and semi-automated analysis software. Clinical levels of neuropathy were assessed via diabetic neuropathy symptom score (DNS). Diabetic Retinopathy (DR) was graded using the Early Treatment Diabetic Retinopathy Study (ETDRS) grading scale. Results: Patients with diabetes demonstrated significant differences in all nerve parameters in comparison to healthy control subjects (p<0.05). CCM detected significant differences in nerve parameters of patients with diabetes in the first two years after diagnosis and in those who had no evidence of DR (grade 0) or symptomatic DPN (DNS score 0) (p<0.05), in comparison to heathy control subjects. Corneal nerve parameters were significantly altered in patients with proliferative DR compared to non-proliferative and no DR (p<0.05), however no relationship was observed between DNS score and changes to corneal nerve fibres (p>0.05). There was no significant difference in any CCM parameters between white and black patients with diabetes (p>0.05). Automated software showed poor agreement with semi-automated results, with a general underestimation for CNFD, CNFL and CNBD. Conclusion: In patients attending primary care screening, CCM in a sensitive biomarker for DPN. Semi-automated CCM quantification reliably detected corneal nerve abnormalities soon after diagnosis of diabetes. Changes in corneal nerve morphology were present prior to any neuropathy symptoms or retinopathy. CCM measured using automatic software requires development to improve agreement with semi-automated analysis.