Key points are not available for this paper at this time.
The quality of the epidemiological data on diabetic neuropathies remains poor for a variety of reasons. They include variability in 1) ascertainment of diabetes, 2) the clinical varieties of diabetic patients studied, 3) characterization of neurological dysfunction, 4) abnormal limits for neurological examinations and tests, 5) minimal criteria for neuropathy, 6) correct attribution of nondiabetic neurological disease, 7) correct attribution of type of neuropathy, 8) estimating neuropathy from use of multiple tests, and 9) estimating severity of polyneuropathy. We have tried to remedy these short-comings in the Rochester Diabetic Neuropathy Study (RDNS). It was not possible to adequately characterize and quantitate diabetic polyneuropathies using only one or two clinical or test abnormalities. To estimate severity of diabetic polyneuropathy, the results of the neurological examination and abnormalities of nerve conduction, quantitative sensory tests, and quantitative autonomic tests were combined into a composite score. One begins by scoring a standard test of neurological deficits (impairments) of the lower limbs (NISLL) and adds to this transformed numbers for percentile abnormality of seven good functional tests. This NIS(LL)+7 tests score appears to provide a much more comprehensive and stable numeric score by which to diagnose and grade severity of diabetic polyneuropathy than does the use of individual clinical or test results. This test score should be useful as a measure of change in diabetic polyneuropathy for purposes of medical practice, epidemiology studies, and controlled clinical trials. The staging approach that we introduced previously continues to provide an important measure of overall severity of diabetic polyneuropathy, taking into account both symptoms and impairments.
Dyck et al. (Mon,) studied this question.