The clinical identification of peripheral neuropathy among older persons

To identify simple clinical rules for the detection of a diffuse peripheral neuropathy among older outpatients.

Observational, blinded, controlled study.

A tertiary-care electrodiagnostic laboratory and biomechanics laboratory.

One hundred research subjects, 68 with electrodiagnostic evidence of peripheral neuropathy, between the ages of 50 and 80 years.

Not applicable.

One examiner, unaware of the results of electrodiagnostic testing, evaluated Achilles' and patellar reflexes, Romberg testing, semiquantified vibration, and position sense at the toe and ankle in all subjects, and unipedal stance time and the Michigan Diabetes Neuropathy Score in a subset of subjects.

Significant group differences were present in all clinical measures tested. Three signs, Achilles' reflex (absent despite facilitation), vibration (128Hz tuning fork perceived for <10s), and position sense (<8/10 1-cm trials) at the toe, were the best predictors of peripheral neuropathy on both univariate and logistic regression (pseudo R(2)=.744) analyses. The presence of 2 or 3 signs versus 0 or 1 sign identified peripheral neuropathy with sensitivity, specificity, and positive and negative predictive values of 94.1%, 84.4%, 92.8%, and 87.1%, respectively. Values were similar among subgroups of subjects with and without diabetes mellitus. When other clinicians applied the technique to 12 more subjects, excellent interrater reliability regarding the presence of peripheral neuropathy (kappa=.833) and good to excellent interrater reliability for each sign (kappa range,.667-1.00) were shown.

Among older persons, the presence of 2 or 3 of the 3 clinical signs strongly suggested electrodiagnostic evidence of a peripheral neuropathy, regardless of etiology. Age-related decline in peripheral nerve function need not be a barrier to the clinical recognition of a diffuse peripheral neuropathy among older persons.