Wound Infections of Snakebites from the Venomous Protobothrops mucrosquamatus and Viridovipera stejnegeri in Taiwan: Bacteriology, Antibiotic Susceptibility, and Predicting the Need for Antibiotics—A BITE Study

Abstract

Snakebites from Taiwan habus (Protobothrops mucrosquamatus) and green bamboo vipers (Viridovipera stejnegeri) account for two-thirds of all venomous snakebites in Taiwan. While there has been ongoing optimization of antivenin therapy, the proper management of superimposed bacterial wound infections is not well studied. In this Bacteriology of Infections in Taiwanese snake Envenomation (BITE) study, we investigated the prevalence of wound infection, bacteriology, and corresponding antibiotic usage in patients presenting with snakebites from these two snakes. We further developed a BITE score to evaluate the probability of wound infections and guide antibiotic usage in this patient population. All snakebite victims who presented to the emergency departments of seven training and research hospitals and received at least one vial of freeze-dried hemorrhagic antivenin between January 2001 and January 2017 were identified. Patient biodata, laboratory investigation results, and treatment modalities were retrieved. We developed our BITE score via univariate and multiple logistic regression analyses. The receiver operating characteristic (ROC) curve was plotted to evaluate the predictive performance of the BITE score. Out of 8,295,497 emergency department visits, 726 patients presented with snakebites from a Taiwan habu or a green bamboo viper. The wound infection rate was 22.45%, with seven positive wound cultures, including six polymicrobial infections. Morganella morganii, Enterococcus spp., Bacteroides fragilis, and Aeromonas hydrophila were most frequently cultured. There were no positive blood cultures. A total of 33.0% (n = 106) of snakebite patients who received prophylactic antibiotics nevertheless developed wound infections, while 44.8% (n = 73) of wound infection patients were satisfactorily treated with one of the following antibiotics: amoxicillin/clavulanic acid, oxacillin, cefazolin, and ampicillin/sulbactam. With the addition of gentamicin, the success of antibiotic therapy increased by up to 66.54%. The prognostic factors for the secondary bacterial infection of snakebites were white blood cell counts, the neutrophil lymphocyte ratio, and the need for hospital admission. The area under the ROC curve for the BITE score was 0.839. At the optimal cut-off point of 5, the BITE score had a 79.58% accuracy, 82.31% sensitivity, and 79.71% specificity when predicting infection in snakebite patients. Our BITE score may help with antibiotic stewardship by guiding appropriate antibiotic use in patients presenting with snakebites. It may also be employed in further studies into antibiotic prophylaxis in snakebite patients for the prevention of superimposed bacterial wound infections.