Aims: To prospectively validate the existing classifications to stratify subjects with diabetes mellitus (DM) by their risk of diabetic foot ulcer (DFU), in high and low risk settings.
Methods: A prospective multicentre cohort study was conducted, including 446 subjects with DM without active DFU followed in the hospital or primary care setting. Demographic, clinical characterization variables, and those included in the classifications were collected at baseline. Subjects were followed for 1year, until DFU or death.
Results: In our sample, with a mean age of 65years, 52% were male; 32 developed a DFU, 7 required an amputation and 18 died. Differences were found between participants' characteristics and classifications' accuracy according to the setting. The great majority of the variables were associated with higher DFU risk. Globally, classifications were highly and equally valid, positive predictive values (PV) were inferior to 40%, negative PV superior to 90% and area under the receiver operating characteristic curve superior to 0.75.
Discussion: All the existing classifications are valid to be applied in high risk clinical context and have a very high capacity to categorize as low risk those subjects that will not develop a DFU. Further research is needed in the primary care setting.
Keywords: Classifications; Diabetic foot; Foot ulcer; Risk.
Copyright © 2017 Elsevier B.V. All rights reserved.