Abstract Background and Aims Central venous lines (CVLs) used for haemodialysis (HD) in children are associated with a high complication rate, leading to inadequate dialysis and need for a line replacement in nearly 45% of children on HD. This high complication rate may be due to poor CVL configurations, designed without considering specific anatomical differences between children and adults. Using a bioengineering approach, we aim to provide a fluid dynamics characterization of different CVL models commonly used in children to establish an association between CVL design and clinical complications, and to improve CVL performance by optimising their design. Method Computational fluid dynamics (CFD) using modelling and simulations is a powerful engineering tool that use computers to study complex systems, applying mathematics, physics and computer science. Four models of CVLs of varying designs and sizes (6.5Fr, 8Fr, 10Fr and 14Fr) were scanned by microCT to reconstruct computational models. CFD analyses were set up to simulate blood flow within the CVLs under both ideal and realistic models of superior vena cava. A variety of flow rates, routinely used in clinical practice, were applied to study the whole range of working conditions. Haemodynamic features that cannot be measured in vivo such as high shear stresses and large areas of stagnation, were analysed. CFD findings were compared to clinical outcomes (n = 26 patients with 57 lines) and to the blood clots distribution of CVLs removed from patients (n = 8). Fluid dynamics characteristics of the CVLs studied were used to set up a preliminary design optimisation process of a 6.5F CVL model. Using sampling methods, a variety of CVL designs were created starting from the original model and results were compared. Results In all the simulated CVLs, the arterial lumens showed the highest stagnation and recirculation areas. The role of arterial tip configuration was negligible for the blood flows and the proximal side holes played a major role for blood aspiration being also subject to the highest levels of shear stress, regardless of the design. In all the anatomical models, blood velocity increased after catheter insertion (Fig. 1a) together with wall shear stresses. CFD results were in accordance with the clinical data which showed a higher recurrence rate of thrombosis for the 8Fr and 10Fr CVLs. Also, microCT of the CVLs removed due to complications confirmed presence of thrombosis at the catheter tips. Results suggest that shear stress is the parameter most strongly related to clinical complications. The preliminary design optimisation study was used to find the parameters which have the highest influence on the CVL performance. Preliminarily optimised designs of the 6.5F model showed a better haemodynamic performance, reducing stagnation areas (Fig. 1b). Conclusion Characterisation of the fluid dynamics of commercially available CVLs explains some of the clinical outcomes experienced by patients. Identification of potential design-related complications in CVLs has allowed to set up the optimisation process which will lead to new CVL designs with improved blood flow distribution and minimal damage to blood cells. Funding Kidney Research UK, PhD fellowship and MedTech grants.
Read full abstract