Background: It is generally accepted that arterial cross-section is an exponential function of pressure. We hypothesized that the order of non-linearity varies substantially between individuals.

Methods: We obtained simultaneously intra-arterial pressure waveforms and common carotid artery (CCA) diameter waveforms by ultrasound in 10 patients (age 70±8 yrs) undergoing coronary catheterisation. We extracted beat-to-beat diastolic (D), systolic (S), and dicrotic notch (N) values from both pressure and diameter waveforms by an automated algorithm. We reconstructed pressure-cross-section curves of each patient by fitting the p=p_d·exp(a(A–A_d)/A_d)) model to D, S, and N; with: pressure (p), diastolic pressure (p_d), artery cross-section (A), and diastolic cross-section (A_d). Based on a we recalculated systolic blood pressure (p_s).

Results: Eight patients had systolic hypertension and two were normotensive. P_s was 144±25 (mean±SD), p_d was 75±12, and pulse pressure was 69±25 mmHg. In 7/10 the exponential model (a=9.4±3.9) fitted the three point data well (r²>0.99, difference in p_s=0±1 mmHg), but in 3/10 (one normotensive) the fit was less good because the non-linearity was of a higher order than contained in the model. In those particular patients, the model underestimated real p_s by 4 to 10 mmHg.

Conclusions: Non-invasive methods to estimate local pulse pressure or characterize arterial stiffness based on the above exponential model are less suitable in 30% of patients with cardiovascular disease due to the presence of a higher order non-linearity in arterial wall distensibility.