Abstract
Diabet. Med. 27, 1271–1279 (2010) Aims  Randomized clinical trials have frequently shown improvement in diabetic sensorimotor polyneuropathy in placebo‐treated participants, counter to the prevailing concept that it deteriorates with time. We aimed to determine the variables associated with this paradoxical nerve function improvement.Methods  Participants with diabetic sensorimotor polyneuropathy randomized to placebo in a multi‐centre, double‐blind study were evaluated for the primary outcome of 1‐year change in the summed sensory nerve conduction velocity of the bilateral sural and non‐dominant median nerves. Association with clinical and biochemical variables measured at 13 time points were examined.Results  The 134 participants had mild to moderate diabetic sensorimotor polyneuropathy of 4.6 years’ duration and mean 1‐year improvement of 2.0 ± 8.0 m/s. Primary outcome measures were available for 122 participants (91%). In multivariate analyses, the change in HbA1c and serum triglycerides from baseline to 2 months demonstrated the strongest association, even independent of baseline and end‐of‐study levels. According to quintiles of change, we determined thresholds: participants with salutary improvement in HbA1c (exceeding a drop of –0.8%) or whose triglycerides did not increase (by 0.32 mmol/l or more) experienced significant improvement (2.9 m/s), while those with salutary levels of both these variables had an exaggerated improvement (5.1 m/s). In comparison, those with non‐salutary changes in both variables experienced a loss of –4.9 m/s (ANOVA P = 0.0014).Conclusions  In mild to moderate diabetic sensorimotor polyneuropathy, short‐term improvements in glycaemic control and serum triglyceride levels have an independent, additive and durable effect on restoration of nerve function.

This publication has 28 references indexed in Scilit: