Sensitivity and Specificity of a New Test for Thermographic Evaluation of the Foot in the Diagnosis of Diabetic Peripheral Polyneuropathy

Abstract

OBJECTIVE: This work determined the sensitivity, specificity, and predictive values of a test evaluating the thermal perception threshold to cold and heat and established a thermal interval with the participant’s basal temperature in order to detect diabetic peripheral neuropathy in 172 people with diabetes, 86 with and 86 without diabetic neuropathy.

DESIGN: This was a cross-sectional, diagnosis accuracy study.

SETTING: This research was performed from January 2010 to February 2012 in Seville, Spain, and Córdoba, Spain.

PARTICIPANTS: One hundred seventy-two people with diabetes, 86 with and 86 without diabetic neuropathy, were recruited to participate in the study.

MAIN OUTCOME MEASURES: The thermal sensitivity was examined in the participants’ feet with an instrument designed to assess thermal discrimination and sensitivity based on the Peltier principle. The skin temperature was recorded, and it was performed the following tests: (1) determining the thermal discrimination threshold, used to identify the lowest noticeable temperature difference, and (2) examining the thermal sensitivity, used to determine the absolute thermal threshold. The receiver operating characteristic curve technique was used to calculate the sensitivity, specificity, and predictive values.

MAIN RESULTS: The area under the curve that showed the best sensitivities and specificities bilaterally was for the warm temperatures under the first and fifth metatarsal heads. In these zones, the temperature difference needed to predict whether a patient was neuropathic was 2°C. Sensitivities were greater than 75%, and the positive predictive values in all cases exceeded 60%.

CONCLUSIONS: This technique enabled the authors to quantify the degrees of colder or warmer temperature differences that participants with diabetes (with and without neuropathy) needed to be able to detect a difference, with acceptable sensitivity, specificity, and predictive values.