K. M. Hocking1,2, E. Wise1, B. Evans2, C. Duvall2, C. M. Brophy1 1Vanderbilt University Medical Center,Nashville, TN, USA 2Vanderbilt University Medical Center,Biomedical Engineering,Nashville, TN, USA
Introduction:
An underappreciated feature of healthy saphenous vein (SV) is its ability to conform to viscoelastic parameters. This feature of biologic tissue may be lost in pathological processes such as calcification or particularly, in surgical preparation when used as graft. The consequences of deformation, while not fully understood, include promotion of intimal hyperplasia and perhaps development of a thrombogenic surface. In fact, external stents are becoming increasingly studied as a method to maintain normal flow conditions in autografted human SV. In this study, the viscoelasticity of unmanipulated SV is compared to that of graft that was fully prepared prior to implantation into CABG patients, manipulations that include supraphysiologic distension, marking with a toxic surgical skin marker and preservation in normal saline.
Methods:
Unmanipulated (UM) human saphenous vein (HSV) and HSV that had been surgically prepared (SP) were obtained from CABG patients, dissected free of fat and cut into rings. Rings were then suspended on a muscle bath in bicarbonate buffer at 37 degrees celsius and were challenged with 110 mM potassium chloride (KCl) to ensure viability of the tissue. The rings were also contracted with phenylephrine and relaxed with sodium nitroprusside, a traditional smooth muscle vasodilator. Data was recorded on LabChart and imported into Eureqa where it was fit to solutions based on the nonlinear least square error of the equation. The equation with a complexity under 10 that had the best fit was used to model the equation. A correlation coefficient of above 0.99 was achieved for each modeled set with a mean absolute error of under 0.03.
Results:
In response to KCl, the UM tissue responded with a force equation of (a+x)(b+mx) resembling a Hill elastic muscle model for 12 out of 12 samples. SP tissue responded to KCl contraction with a force equation of (a+mx) resembling an elastic model for 10 out of 12 samples and 2 out of 12 samples exhibited the Hill elastic model (χ2=17.1, P<0.0001, Figure 1 A,B) . In response to sodium nitroprusside the UM tissue had a relaxation profile of a/(b+mx) whilst the SP tissue responded with a relaxation profile of (a-mx) (χ2=7.273, P=0.007, Figure 1 C,D). The Hill elastic muscle model is represented by a contractile element and two non-linear spring elements where one is in series with the contractile element and the other in parallel.
Conclusion:
It is shown that healthy unmanipulated SV corresponds to a viscoelastic model both when contracting and relaxing. Surgical preparation causes a loss of confirmation to the Hill elastic muscle model in human SV. The use of less deformative techniques may mitigate intimal hyperplasia in autografts.
