Patient-specific blood pressure correction technique for arterial stiffness: evaluation in a cohort on anti-angiogenic medication.

Abstract

Arterial pulse wave velocity (PWV) depends on blood pressure (BP). Correction of PWV for BP is commonly performed using a statistical approach, requiring a patient cohort. We recently developed a mechanistic, model-predictive approach to assess BP-independent changes in carotid PWV (cPWV) at the level of the individual. The goal of the present study is to compare our novel technique to conventional statistical correction, in the context of anti-cancer therapy using anti-angiogenic drugs (AADs). AADs frequently lead to a PWV increase, but also to hypertension, underlining the need for BP correction of PWV measurements. We obtained carotid artery systolic and diastolic cross-sectional areas (echotracking) and corresponding BPs (tonometry) in 48 patients before starting AAD treatment (sorafenib/sunitinib), and at four follow-up visits spaced 2 weeks apart. For each patient, we derived cPWV and a baseline single-exponential BP cross-sectional area curve. Based on these baseline curves and follow-up BPs, we predicted cPWV at follow-up due to BP. By comparing predicted and measured cPWVs at follow-up, we assessed the BP-independent cPWV increase. In the same way, we assessed whether diastolic cross-sectional area (Ad) changed beyond the BP-induced amount. The AAD-induced BP-independent increase in cPWV was 0.43(0.09,0.77) m s-1 (mean (95%CI), P=0.014, mechanistic approach) and 0.48(0.14,0.82) m s-1 (P=0.006, statistical approach). Ad increased with 1.92(0.93,2.92) mm2 (P<0.001) and 2.14(1.06,3.23) mm2 (P<0.001), respectively. In conclusion, the present study demonstrates the feasibility and potential of our mechanistic, model-predictive approach to quantify BP-independent effects on arterial stiffness at the level of the individual, in a clinically relevant setting of AAD therapy.