Prediction of diabetic foot amputation using newly revised DIRECT coding system: Comparison of accuracy with that of five existing classification systems

Abstract

Diabetes mellitus (DM) causes various complications over time, one such complication is diabetic foot ulcers (DFU), which are challenging to treat and can lead to amputation. Additionally, a system for accurate prediction of amputation has yet to be developed. In total, 131 patients were included in the study after retrospectively collecting data from 2016 to 2020 about DFU. The collected data were used for comparison of the accuracy between five existing classification systems and the newly revised DIRECT coding system, and investigation of risk factors for lower extremity amputation (LEA). The existing five classification systems and DIRECT system can effectively predict LEA. The DIRECT3 system has three elements, C-reactive protein (CRP), ulcer history (UH), and hypertension (HTN) in addition to those of the DIRECT system. It had a high predictive value and accuracy similar to that of Wagner and University of Texas (UT) on depth among the five classification systems. Among the statistically significant risk factors, duration of DM and HTN, haemoglobin (Hb), CRP, and UH showed an association with LEA. The DIRECT coding system is effective for predicting LEA and explaining appropriate treatment methods for DFU, and is widely applicable because of its user accessibility and convenience.