Significance of Bacteria Detection with Filter Paper Method on Diagnosis of Diabetic Foot Wound Infection
X H Zou,Y P Zhu,G Q Ren,G C Li,J Zhang,L J Zou,Z B Feng,B H Li
DOI: https://doi.org/10.3760/cma.j.issn.1009-2587.2017.02.005
2017-01-01
Abstract:Objective: To evaluate the significance of bacteria detection with filter paper method on diagnosis of diabetic foot wound infection. Methods: Eighteen patients with diabetic foot ulcer conforming to the study criteria were hospitalized in Liyuan Hospital Affiliated to Tongji Medical College of Huazhong University of Science and Technology from July 2014 to July 2015. Diabetic foot ulcer wounds were classified according to the University of Texas diabetic foot classification (hereinafter referred to as Texas grade) system, and general condition of patients with wounds in different Texas grade was compared. Exudate and tissue of wounds were obtained, and filter paper method and biopsy method were adopted to detect the bacteria of wounds of patients respectively. Filter paper method was regarded as the evaluation method, and biopsy method was regarded as the control method. The relevance, difference, and consistency of the detection results of two methods were tested. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of filter paper method in bacteria detection were calculated. Receiver operating characteristic (ROC) curve was drawn based on the specificity and sensitivity of filter paper method in bacteria detection of 18 patients to predict the detection effect of the method. Data were processed with one-way analysis of variance and Fisher's exact test. In patients tested positive for bacteria by biopsy method, the correlation between bacteria number detected by biopsy method and that by filter paper method was analyzed with Pearson correlation analysis. Results: (1) There were no statistically significant differences among patients with wounds in Texas grade 1, 2, and 3 in age, duration of diabetes, duration of wound, wound area, ankle brachial index, glycosylated hemoglobin, fasting blood sugar, blood platelet count, erythrocyte sedimentation rate, C-reactive protein, aspartate aminotransferase, serum creatinine, and urea nitrogen (with F values from 0.029 to 2.916, P values above 0.05), while there were statistically significant differences among patients with wounds in Texas grade 1, 2, and 3 in white blood cell count and alanine aminotransferase (with F values 4.688 and 6.833 respectively, P<0.05 or P<0.01). (2) According to the results of biopsy method, 6 patients were tested negative for bacteria, and 12 patients were tested positive for bacteria, among which 10 patients were with bacterial number above 1×10(5)/g, and 2 patients with bacterial number below 1×10(5)/g. According to the results of filter paper method, 8 patients were tested negative for bacteria, and 10 patients were tested positive for bacteria, among which 7 patients were with bacterial number above 1×10(5)/g, and 3 patients with bacterial number below 1×10(5)/g. There were 7 patients tested positive for bacteria both by biopsy method and filter paper method, 8 patients tested negative for bacteria both by biopsy method and filter paper method, and 3 patients tested positive for bacteria by biopsy method but negative by filter paper method. Patients tested negative for bacteria by biopsy method did not tested positive for bacteria by filter paper method. There was directional association between the detection results of two methods (P=0.004), i. e. if result of biopsy method was positive, result of filter paper method could also be positive. There was no obvious difference in the detection results of two methods (P=0.250). The consistency between the detection results of two methods was ordinary (Kappa=0.68, P=0.002). (3) The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of filter paper method in bacteria detection were 70%, 100%, 1.00, 0.73, and 83.3%, respectively. Total area under ROC curve of bacteria detection by filter paper method in 18 patients was 0.919 (with 95% confidence interval 0-1.000, P=0.030). (4) There were 13 strains of bacteria detected by biopsy method, with 5 strains of Acinetobacter baumannii, 5 strains of Staphylococcus aureus, 1 strain of Pseudomonas aeruginosa, 1 strain of Streptococcus bovis, and 1 strain of bird Enterococcus. There were 11 strains of bacteria detected by filter paper method, with 5 strains of Acinetobacter baumannii, 3 strains of Staphylococcus aureus, 1 strain of Pseudomonas aeruginosa, 1 strain of Streptococcus bovis, and 1 strain of bird Enterococcus. Except for Staphylococcus aureus, the sensitivity and specificity of filter paper method in the detection of the other 4 bacteria were all 100%. The consistency between filter paper method and biopsy method in detecting Acinetobacter baumannii was good (Kappa=1.00, P<0.01), while that in detecting Staphylococcus aureus was ordinary (Kappa=0.68, P<0.05). (5) There was no obvious correlation between the bacteria number of wounds detected by filter paper method and that by biopsy method (r=0.257, P=0.419). There was obvious correlation between the bacteria numbers detected by two methods in wounds with Texas grade 1 and 2 (with r values as 0.999, P values as 0.001). There was no obvious correlation between the bacteria numbers detected by two methods in wounds with Texas grade 3 (r=-0.053, P=0.947). Conclusions: The detection result of filter paper method is in accordance with that of biopsy method in the determination of bacterial infection, and it is of great importance in the diagnosis of local infection of diabetic foot wound.