Diagnostic Accuracy of Clinical Methods for Detection of Diabetic Sensory Neuropathy

To determine the accuracy of clinical methods for detection of sensory neuropathy as compared to biothesiometry.

Cross-sectional analytical study.

1 Mountain Medical Battalion, Azad Kashmir, from October 2013 to September 2014.

Patients with type 2 diabetes were enrolled by convenience sampling. Exclusion criteria included other identifiable causes of neuropathy, extensive ulceration of feet, amputated feet, those on treatment for neuropathy and unwilling patients. Average of 3 vibration perception threshold values measured with a biothesiometer on distal hallux was calculated. Ten gm monofilament was used to examine touch sensation over dorsal surfaces of great toes. Vibration sensation was checked over the tips of great toes using 128Hz tuning fork. Ankle jerks were checked bilaterally.

Neuropathy (vibration perception threshold > 25 volts) was present in 34 (21.12%) out of 161 patients and 93 (57.76%) were symptomatic. Measures of diagnostic accuracy for monofilament, tuning fork and ankle jerks were: sensitivity 41.18%, 55.88% and 64.71%; specificity 92.91%, 93.70% and 80.31%; positive predictive value (PPV) 60.87%, 70.37% and 46.81%; negative predictive value (NPV) 85.51%, 88.81% and 89.47%; and, diagnostic accuracy 81.99%, 85.71% and 77.02%, respectively. Values for any 1 positive sign, any 2 positive signs or all 3 positive signs were: sensitivity 35.29%, 14.71% and 32.35%; specificity 81.89%, 93.70% and 99.21%; PPV 34.29%, 38.46% and 91.67%; NPV 82.54%, 80.41% and 84.56%; and, diagnostic accuracy 72.05%, 77.02% and 85.09%, respectively.

Clinical methods are high accurate and specific diagnosis in picking up symmetric sensory neuropathy in patients with type 2 diabetes, particularly when multiple signs are combined.