Risk factors for antibiotic resistance and mortality in patients with bloodstream infection of Escherichia coli.

This study aimed to investigate the risk factors for bloodstream infection (BSI) caused by extended-spectrum beta-lactamase (ESBL)-producing Escherichia coli (E. coli) and related mortality. The clinical data of 388 patients with E. coli BSI were analyzed. Blood cultures were performed and the antimicrobial susceptibility profiles of the resulting isolates were determined. Four single-nucleotide polymorphisms (rs231775, rs12343816, rs16944, and rs2233406) were genotyped using real-time PCR. ESBL were detected by disk diffusion confirmatory testing. Univariate and multivariate regression analyses were applied to identify the risk factors for ESBL-producing isolates and the BSI-induced mortality. The prevalence of ESBL-producing E. coli in BSI patients was 40.98%. E. coli isolates were commonly susceptible to carbapenem and ß-lactam/ß-lactamase inhibitor combinations. The major ESBL genes were CTX-M-14, CTX-M-55, CTX-M-15, and CTX-M-27. The proportion of CTX-M-15 was significantly higher in patients over 70 years and those receiving stomach tube catheterization. Nosocomial infection, biliary tract infection, stomach tube catheterization, and previous cephalosporin administration were independent risk factors for ESBL-producing isolates. ESBL positivity, nosocomial infection, and cancer were independent risk factors of mortality. Two genetic polymorphisms associated with inflammation activation, rs231775 A allele and rs2233406 T allele, significantly increased the mortality risk of E. coli BSI with a risk ratio (95% CI) of 1.93 (1.05-3.55) and 4.38 (2.07-9.29), respectively. For patients with nosocomial infection, biliary tract infection, and cancer, the monitor of BSI and antibiotic susceptibility should be enhanced. The invasive procedures should be minimized. rs231775 and rs2233406 are promising prognostic markers for E. coli BSI patients.

Research and Verification of Convolutional Neural Network Lightweight in BCI.

With the increasing of depth and complexity of the convolutional neural network, parameter dimensionality and volume of computing have greatly restricted its applications. Based on the SqueezeNet network structure, this study introduces a block convolution and uses channel shuffle between blocks to alleviate the information jam. The method is aimed at reducing the dimensionality of parameters of in an original network structure and improving the efficiency of network operation. The verification performance of the ORL dataset shows that the classification accuracy and convergence efficiency are not reduced or even slightly improved when the network parameters are reduced, which supports the validity of block convolution in structure lightweight. Moreover, using a classic CIFAR-10 dataset, this network decreases parameter dimensionality while accelerating computational processing, with excellent convergence stability and efficiency when the network accuracy is only reduced by 1.3%.