Factors predicting the early mortality of trauma patients.

BACKGROUND: The aim of this study was to identify factors predicting early mortality in trauma patients. METHODS: This was a study of 6288 trauma patients admitted to the hospital between July 2011 and June 2016. Among the variables recorded for a prospective trauma registry, the following were selected for analysis: sex; age; a combination of the Glasgow Coma Scale score, age, and systolic blood pressure (SBP) (GAP); a combination of the mechanism of injury, the Glasgow Coma Scale score, age, and SBP (MGAP); SBP; respiratory rate; peripheral oxygen saturation (SpO2 value); the Glasgow Coma Scale score; laboratory variables; and presentation time. Logistic regression analysis was used to explore associations between these variables and early mortality. RESULTS: In total, 296 (4.6%) patients died within 24 hours. Univariate regression analysis indicated that age, the GAP, the MGAP, SBP, SpO2, the Glasgow Coma Scale score, base excess, hemoglobin level, platelet count, INR, and presentation time predicted early mortality. Multivariate regression showed that the GAP, the MGAP, SpO2, base excess, platelet count, and INR were independently predictive. The areas under the receiver operator curve comparisons for the GAP and MGAP models revealed the superiority of the GAP-based model. CONCLUSION: The GAP model, SpO2, base excess, platelet count, and INR predicted the early mortality of trauma patients.

Evaluation of four methods of assigning species and genus to medically important bacteria using 16S rRNA gene sequence analysis.

The four methods for assigning bacterial species are the Clinical and Laboratory Standards Institute (CLSI), modified CLSI (mCLSI), phylogenetic analysis (PA) and closest match (CM) methods, these are used to identify the genus and species using 16S rRNA gene sequence results. In this study, the results of identification by these four methods of 37 aerobic reference strains, 30 anaerobic reference strains, 15 Acinetobacter reference strains and 167 Acinetobacter clinical strains were compared. The rates of accurate identification to the species level using the CLSI, mCLSI, PA and CM methods were as follows: 24.3, 86.5, 86.5 and 89.2%, respectively, for the 37 aerobic reference strains; 73.3%, 96.7%, 90.0% and 93.3%, respectively, for the 30 anaerobic reference strains; 40.0%, 93.3%, 100% and 93.3%, respectively, for the 15 Acinetobacter reference strains; and 53.9%, 90.4%, 95.8% and 90.4%, respectively, for the 167 Acinetobacter clinical strains. The rates of accurate identification to the genus level using the CLSI, mCLSI, PA, and CM methods were as follows: 91.9%, 91.9%, 94.6% and 91.9%, respectively, for the 37 aerobic reference strains; 100%, 100%, 100% and 100%, respectively, for all of the 30 anaerobic reference strains, 15 Acinetobacter reference strains and the 167 Acinetobacter clinical strains. The mCLSI is the most practical and pragmatic method for identification of species based on 16S rRNA sequences for hospital, research or industry laboratories because it performs well and involves a simple procedure.

Evaluation of VITEK 2, MicroScan, and Phoenix for identification of clinical isolates and reference strains.

To compare the identification accuracies of VITEK 2 (bioMérieux), MicroScan (Siemens Healthcare), and Phoenix (Becton Dickinson), microbial identification was performed on 160 clinical isolates and 50 reference strains on each of these 3 systems, using the appropriate identification kit provided by each system. Of the 142 clinical isolates that were identified at the species level, VITEK 2, MicroScan, and Phoenix correctly identified 93.7%, 82.4%, and 93.0%, and incorrectly identified 2.1%, 7.0%, and 0%, respectively. In the reference strain tests, VITEK 2, MicroScan, and Phoenix correctly identified 55.3%, 54.4%, and 78.0% of the reference strains at the species level and incorrectly identified 10.6%, 13.0%, and 6.0% of the reference strains, respectively. In conclusion, the identification rate of VITEK 2, Phoenix, and MicroScan was high or acceptable on clinical isolates. Phoenix showed a significantly higher performance than VITEK 2 or MicroScan in identifying the reference strains.