By Gopalakrishna Prabhu

In diabetic patients, Neuropathy (loss of sensation) is the primary etiologic factor. Due to the sensory neuropathy, patients are unable to sense pressure pain or micro trauma on the foot. An injury or infection in the neuropathic foot results in a serious medical condition often leading to amputation. Therefore foot problems in patients with diabetes mellitus are a major public health concern these days. The present study is aimed at using new diagnostic pressure parameters and correlating hem to quantified plantar sensation to understand pressure distribution patterns under the soles of diabetic feet in different levels of sensation loss, possibly responsible for ulcer formation. Sensations of plantar surface of the diabetic foot (in ten standard foot sole areas) are quantified using Semmes-Weinstein mono filaments and classified into four different categories. Standing and walking foot pressure measurements are carried out using optical pedobarograph.

In the space domain analysis of walking foot pressures, studies are performed on 20 normal and 40 diabetic feet indifferent levels of sensation loss, using new diagnostic parameters, normalised peak pressure (NPP), pressure contact ratio (PCR) and their gradients (relative difference). It is found that the values of these foot pressure parameters of diabetic patients are much higher than those of normal subjects in general and these values increase with degree of neuropathy, in all the areas of the foot. But these pressure parameters in the space domain analysis could not distinguish clearly the diabetic feet in different levels of sensation loss.

A detailed analysis is performed first on standing foot pressure images in frequency domain. A new parameter, power ratio (PR, ratio of power in higher spatial frequency components to the total power in an image) is used to distinguish between the foot pressure image patterns of diabetic neuropathy subjects (in different levels of sensation loss) and those of normal foot. The results could help to detect the early stages of neuropathy responsible for plantar ulcers only in the heel of diabetic feet.

The study is carried out further to analyse walking foot pressure images of diabetic subjects in different levels of neuropathy and use the same parameter power ratio, PR (defined earlier) to distinguish normal from the diabetic feet. The parameter is able to distinguish clearly the normal form of the diabetic feet and also make clear distinctions without overlap between diabetic feet in different levels of loss of sensation. The result could help to detect the early stages of neropathy responsible for plantar ulcers in both the heel and fore-foot areas of diabetic feet.

The comprehensive foot models are developed using artificial neural networks with pressure parameters namely, NPP, PCR and PR are used as input variables and sensation levels as output variable. The linear model could help in better understanding the degree of Neuropathy in terms of foot pressure parameters and changes in functionality of the foot. The nonlinear model helps in classifying the foot in the category of normal diabetic in different levels of sensation loss. The neural network model can be trained with sufficient number of data sets of foot pressures parameters corresponding to the ulcer at early stage neuropathy. This model then could help in detecting the plantar surface of foot likely to have ulcer and the information can be utilised by orthopaedic surgeons to device early corrective methods to prevent the foot-sole ulcers and enhance the quality of life by avoiding the lower limb complications of diabetes.