Aims: The LDIFLARE method has been previously shown to be a sensitive measure of small fibre function. In this longitudinal 5-year study, we evaluated the LDIFLARE as a method to predict the development of DPN.
Methods: From the cohort of 160 people with type 1 and 2 diabetes recruited at baseline, 58 patients who did not have DPN at onset and completed 5 years of follow up were studied. Incident DPN was defined as progression to a Neurology Disability Sore (NDS) of ≥3. All subjects underwent comprehensive neurological review including electrophysiology. We used duration-dependent receiver operating characteristic (ROC) curves to calculate the predictive validity of the baseline LDIFLARE for the development of a NDS of ≥3.
Results: Of the 58 subjects without DPS who completed 5yr follow up, 34 had Type-1 and 24 had Type-2 diabetes; 35 females. Mean age (±SD) was 47.1±8.8 years. Mean follow up time was 5.4 years. Of these, 24 (T1Dm & 14 T2DM) progressed to a NDS of ≥3 (41%; 7.7 events per 100 subject-years) . There was a strong correlation with older age, longer duration of diabetes, raised HbA1c and raised triglycerides (significance: p<0.05) but no effect of BMI or blood pressure. The linear rate of fall of LDIFLARE from baseline in newly diagnosed DPN was 0.16 cm2/yr for T1DM and 0.22cm2/yr for T2DM. The area under the ROC curve for LDIFLARE ranged between 0.66 and 0.75. The optimal diagnostic threshold for a baseline LDIFLARE of 5.05cm2 was associated with 71% sensitivity, 74% specificity, and a hazard ratio of 3.73 (95% CI 2.14-6.73; P < 0.01) for new-onset DPN.
Conclusions: We conclude that the LDIFLARE method with a cut-off of 5.cm2 showed good predictive validity for identifying people with diabetes who are at risk of developing DPN. We suggest that as a measure of small fibre function, the LDIFLARE method may be preferred over conventional large fibre methods in predicting the later development of clinical DPN.
S.Sharma: None. J.Cross: None. G.Rayman: None.
