Monotonicity of nerve tests in diabetes: subclinical nerve dysfunction precedes diagnosis of polyneuropathy.

OBJECTIVE The objective of this study was to test whether monotone worsening of nerve function, attributable to diabetes, can be demonstrated before criteria for diabetic sensorimotor polyneuropathy (DSPN) have been met. Which nerve tests are best? RESEARCH DESIGN AND METHODS From a prevalence cohort of 504 individuals in the Rochester Diabetic Neuropathy study (RDNS), we identified 238 individuals (group 1) who at first examination were without polyneuropathy (DSPN) by a sum score of the normal deviates (from percentiles) of five attributes of nerve conduction of the legs (i.e., their five nerve conduction normal deviate values were <97.5th percentile) and were followed longitudinally two or more times. Of these 238, 90 (group 2) were followed six or more times at yearly or bi-yearly intervals. We compared different nerve tests for the ones most sensitive and reliable in showing latent nerve dysfunction and monotone (the extent to which a variable measured repeatedly over time reveals a significant trend of worsening or improvement). RESULTS In group 1 patients, the mean sum score of five attributes of nerve conduction (sigma 5 NC nds) at baseline was 1.08 and at the last examination (only patients with Sigma 5 NC nds <97.5th percentile) was 3.63, markedly higher than that in healthy subjects (only of individuals with Sigma 5 NC nds <97.5th percentile) (-0.12), indicating a subtle latent shift of nerve conduction tests toward abnormality. Serial evaluations of many individual and especially sum scores of nerve conduction tests in group 2 patients showed statistically significant worsening with time, even when nerve conduction tests were still well within normal limits. Neurologic signs also worsened but barely to significant levels; however, symptoms and quantitative sensation tests did not. Considering the composite score sigma 5 NC nds, 42 (of 90 group 2 patients) showed significant worsening, 22 were still without DSPN by nerve conduction test criteria, and some were even below the 50th percentile at the last evaluation. CONCLUSIONS Subtle and latent functional worsening of nerve conduction can be demonstrated even before nerve conduction test criteria for DSPN have been met. For demonstrating monotone worsening, the order (from best to worst) of tests was: some composite scores of nerve conduction and individual attributes of nerve conduction. We did not show monotone worsening of symptoms or of quantitative sensation test results. In multivariate analysis of risk factors and their association with worsening sigma 5 NC nds, 24-h microalbuminuria (a marker of microvessel disease) was found to be a significant covariate, an indication that the asymptomatic alterations of nerve conduction are meaningful.