Delftia tsuruhatensis, which was first isolated in environmental samples, was rarely associated with human infections. We report on pneumonia caused by D. tsuruhatensis in an infant who underwent cardiac surgery. Retrospective analyses detected 9 other isolates from 8 patients. D. tsuruhatensis is an emergent pathogen, at least for immunocompromised patients.

Thrombus formation is one of the main issues in the development of blood-contacting medical devices. This article focuses on the modeling of one aspect of thrombosis, the coagulation cascade, which is initiated by the contact activation at the device surface and forms thrombin. Models exist representing the coagulation cascade by a series of reactions, usually solved in quiescent plasma. However, large parameter uncertainty involved in the kinetic models can affect the predictive capabilities of this approach. In addition, the large number of reactions of the kinetic models prevents their use in the simulation of complex flow configurations encountered in medical devices. In the current work, both issues are addressed to improve the applicability and fidelity of kinetic models. A sensitivity analysis is performed by two different techniques to identify the most sensitive parameters of an existing detailed kinetic model of the coagulation cascade. The results are used to select the form of a novel reduced model of the coagulation cascade which relies on eight chemical reactors only. Then, once its parameters have been calibrated thanks to the Bayesian inference, this

No abstract

Computational models of the coagulation cascade are used for a wide range of applications in bio-medical engineering such as drug and bio-medical device developments. However, a lack of robustness of numerical models has been highlighted when studying clinically relevant scenarios. In order to develop more robust models, numerical simulations need to be confronted with realistic situations relevant to clinical practice. In this work, two well-established numerical representations of the coagulation cascade initiated by the intrinsic and extrinsic systems, respectively, were compared with thrombin generation assays considering realistic pathological conditions. Proper modifications were needed to align the in vitro and in silico data, namely; adapting initial conditions to the thrombin assay system, omitting reactions irrelevant to our case study, and improving the fitting of some reaction rates. The modified models were able to capture the experimental trends of thrombin generation for a range of concentrations of factors XII, XI, and VIII for cases in which the coagulation cascade is triggered through the extrinsic and intrinsic systems. Our work emphasizes that when existing coagulation cascade models are extrapolated to experimental settings for which they were not calibrated, careful adjustments must be made. We show that the two coagulation models used in this work can predict physiological conditions, but when studying pathological conditions, proper modifications are needed to improve the numerical results.

No abstract