Objective: To appraise the performance of a new point-of-care wound infection detection kit in diabetic foot ulcers (DFUs), using clinician opinion as the primary comparator. The proprietary swab-based chromatic Glycologic (Glycologic Ltd., UK) detection kit used in this study is designed to detect host response to pathogenic levels of bacteria in wounds. Method: In high-risk podiatry clinics, patients with DFUs were recruited and infection detection kit test results compared with initial clinician opinion. Chi-squared tests, principal component analysis (PCA) and multiple regression analysis were performed to determine which variables were possibly associated with infection. The variables considered were patients' wound parameters, wider vascular comorbidity and demographics. Results: A total of 136 patients, providing 383 wound swabs, were included in the study. Total agreement in terms of DFU wound assessment for infection—between podiatrists' clinical opinion and Glycologic kit test result—was observed in 79% of cases (301/383). For 56 of the 349 negative infection detection kit test results (16%), podiatrists identified a ‘possible’ or ‘definite’ infection. Conversely, in 14 of the 307 cases (4.6%) where podiatrists deemed the wound ‘not infected’, the infection detection kit test showed a colour change. Regression analysis and PCA showed that clinical signs of wound infection, namely erythema, purulence and odour, were all significantly associated with both a positive clinical opinion and infection detection kit test result. However, in the case of the infection detection kit, a patient's number of lesions and vascular comorbidities were also significantly correlated with a positive test result. Conclusion: A host response to critical pathological levels of bioburden in a wound—as detected with the infection detection kit—may partly be determined by an individual patient's (vascular) health and therefore be person-specific. Further research is indicated to determine the relationship between an infection detection kit test result and the microbiological status of the wound.
Self-Taught Learning Based on Sparse Autoencoder for E-Nose in Wound Infection Detection