Underestimation of Murine Cardiac Hemodynamics Using Invasive Catheters: Errors, Limitations, and Remedies

Abstract

This study confirms the existence and quantifies the magnitude of a compound bias error in the estimation of hemodynamic indices of global cardiac function from the use of invasive, miniature, commercially available catheters in mice. Correspondingly, a new catheter design alternative is proposed that minimizes underestimation errors of such indices, according to simulations using end-diastolic (ED) and end-systolic (ES) computational myocardial models. Computational surface and finite element ED and ES mouse cardiac left ventricular (LV) models were constructed from MR images from five C57BL/6J mice. A commercially available miniature catheter and a modified improve design were also developed and imported in XFdtd for simulation of the electric field excitation pattern in composite catheter-LV blood and myocardium models. Straight (ED) and bent (ES) catheter orientations were tested and global functional indices of hemodynamics were estimated. Comparison of ED and ES estimates from catheter simulations, and direct MRI quantification, shows mean underestimation errors from five C57BL/6J mice that range between −8.3 ± 52.9 and −68.9 ± 10.4%. Propagation of such errors for estimating stroke volume, cardiac output, and ejection fraction, yields compound errors (for the commercially available catheter), of the order of 85.4 ± 29.1, 182.3 ± 117.4, 132.9 ± 171.1%, respectively, assuming a catheter electric field threshold sensitivity of 5%. Elicited response improvements (peak amplitude and spatial extent) in the generated ED and ES electric fields are quantified to be 405% and 52% (3.8 vs. 2.5 mm at 1% electric field fall-off) in ED, and 934 and 80% in ES, for the new catheter, compared to the commercially available design. Significant reductions in absolute hemodynamic error estimates were computed (of the order of 4.8–60.8%) for the new catheter design based on the computational models of one of the studied C57BL/6J mice. Overall, implementation of the newly proposed design can be realized at a reduced cost, with manufacturing processes that do not deviate from the currently employed methods, thereby leading to significantly improved accuracies in the estimation of hemodynamic indices of cardiac function.