Association between Early Neuroretinal Dysfunction and Peripheral Motor Unit Loss in Patients with Type 1 Diabetes Mellitus

Abstract

Objectives: It has been already confirmed that retinal neurodegeneration has a predictive value in the development of microvascular alterations in diabetic retinopathy. However, no data are available on the association between neuroretinal dysfunction and peripheral motor unit loss. Our study, therefore, was aimed at investigating the hypothesis that retinal neurodegeneration could be considered an early marker of diabetic peripheral neuropathy (DPN).

Methods: 20 T1DM patients with no symptoms/signs of peripheral polyneuropathy, without DR or with very mild nonproliferative DR, and 14 healthy controls (C) age- and gender-matched were enrolled. The following electrophysiological tests were performed: standard nerve conduction studies (NCS) and incremental motor unit number estimation (MUNE) from the abductor hallux (AH) and abductor digiti minimi (ADM). Neuroretinal function was studied by multifocal electroretinogram (MfERG) recordings, measuring response amplitude density (RAD) and implicit time (IT) from rings and sectors of superior (S)/inferior (I)/temporal (T)/nasal (N) macular sectors up to 10 degrees of foveal eccentricity.

Results: MfERG RADs from rings and sectors were significantly reduced in T1DM (p < 0.05) vs. C. ADM MUNE and AH MUNE were significantly decreased in T1DM (p = 0.039 and p < 0.0001, respectively) vs. C. A positive correlation between mean MfERG RADs from the central 5 degrees of the four (S, I, T, and N) macular sectors and lower limb motor unit number (r = 0.50, p = 0.041; r = 0.64, p = 0.005; r = 0.64, p = 0.006; and r = 0.61, p = 0.010, respectively) was observed in T1DM patients. No abnormalities of NCS were found in any subject.

Conclusions: The motor unit loss on the one hand and neuroretinal dysfunction on the other hand are already present in T1DM patients without DPN. The relationship between neuroretinal dysfunction and motor unit decline supports the hypothesis that neuroretina may represent a potential "window" to track the early neurogenic damage in diabetes.

Figure 1

Combined sector analysis of mean MfERG RADs from S1 + S2 (0–5°) in the superior, temporal, and nasal sectors showed significantly reduced values in T1DM subjects vs. C (p < 0.05). There was also a reduced but not statistically significant RAD value from the inferior sector (0.055). In addition, significant sector X group interaction was found (F(3,81) = 6.07; p = 0.001), since the difference between DM and C was larger in the temporal and nasal sectors with respect to the superior and inferior sectors (consistently, p < 0.005) N1-P1 RADs are defined as amplitude densities between the first negative peak, N1, and the first positive peak, P1. S: superior; I: inferior; T: temporal; N: nasal macular quadrants.

Figure 2

Motor unit number estimation (MUNE) (mean ± SEM) of abductor digiti minimi (ADM) in the diabetic groups and controls. ADM MUNE was significantly decreased in T1DM vs. C (p < 0.05).

Figure 3

Motor unit number estimation (MUNE) (mean ± SEM) of abductor hallux (AH) in the diabetic groups and controls. AH MUNE was significantly decreased in T1DM vs. C (p < 0.001).

Figure 4

Scatter plot between MfERG RADS of S1 + S2 of the N (nasal), S (superior), I (inferior), and T (temporal) sectors and motor unit number estimation of the abductor hallux (AH MUNE) in type 1 DM patients. A positive correlation between the mean MfERG RADs of the four quadrants (S, I, T, and N) and AH MUNE (r = 0.50, p = 0.041; r = 0.64, p = 0.005; r = 0.64, p = 0.006; and r = 0.61, p = 0.010, respectively) was observed in T1DM patients. Since such correlations could be significantly affected by outliers, influence statistics (standardized DfBetas) were computed. After eliminating cases with values higher than the cutoff (2/sqrt(n)), r correlations resulted similar (r = 0.49, p = 0.065; r = 0.64, p = 0.015; r = 0.64, p = 0.013; and r = 0.52, p = 0.048, respectively).