Introduction: Aortic stiffness, assessed through carotid-femoral pulse wave velocity (PWV), has been associated with an increased risk of cardiovascular events and mortality. Measurements of PWV are based on the proper identification of the foot of the pulse waveform by either the maximum of the second-derivative method (as used in Complior) or the intersecting tangents algorithms (as used in SphygmoCor). These approaches can give different results, especially at higher PWV ranges. However, these devices also differ by signal acquisition technology, signal filtering, and quality control algorithms, making the true contribution of analytical algorithms uncertain. The aim of the present study was to identify the differences in pulse transit time (PTT) and PWV calculated by these two algorithms when provided with the same input signal. Methods: In 113 subjects, 346 recordings of 10 s were obtained using the Complior Analyse system (PWVComp-2nd). The pulse waves were imported into MATLAB and filtered (n = 4,102 pairs of pulse waves), where after inspection 3,770 pairs were available for determination of PTT using second-derivative and intersecting tangents algorithms (PTTMat-2nd and PTTMat-IT) and the respective PWVMat-2nd and PWVMat-IT for each pair. Additionally, the same pulse wave recordings were analyzed using the SphygmoCor system in simulation mode, employing the intersecting tangents algorithm (PWVSphyg-IT). Results: The mean beat-by-beat PTTMat-2nd and PTTMat-IT were 54.55 ± 18.55 ms (range 15.00–129.00) and 54.61 ± 18.61 ms (range 15.00–126.00) (p = 0.09), respectively. The mean per participant PWVMat-2nd and PWVMat-IT were 9.67 ± 3.46 m/s and 9.66 ± 3.4 m/s with a mean difference of 0.01 ± 0.32 m/s (p = 0.35). The PWVComp-2nd and PWVSphyg-IT were 9.48 ± 3.25 m/s and 9.59 ± 3.25 m/s with a mean difference of 0.11 ± 0.66 m/s (p = 0.04). Conclusion: The present study shows that the difference between the two algorithms is negligible across a wide range of PTT and hence does not support the need for adjusting PWV according to the algorithm used for determining PTT.

Large artery stiffness is an important determinant of systolic hypertension and it is associated with a higher risk of cardiovascular events in patients with hypertension, diabetes, and/or chronic kidney disease [1‒6]. Carotid-femoral pulse wave velocity (PWV) increases with arterial wall stiffness and has been identified as the gold standard for assessing aortic stiffness due to its robust association with morbidity and mortality [7‒11].

PWV is assessed by measuring pulse transit time (PTT) from the carotid to the femoral artery. Determination of PTT relies on the proper detection of the foot of the wave as this is the area of the wave, which is least affected along the arterial tree [12, 13]. Among various mathematical algorithms proposed for identifying the foot of the pressure waves, the maximum second-derivative and intersecting tangents algorithms have emerged as the most reliable techniques for pressure waveform analysis [12].

Two frequently used systems for PWV assessment, namely, SphygmoCor® (AtCor Medical Pty, Sydney, Australia) and Complior® Analyse (Alam Medical, France), differ in sensor technologies and algorithms used for PTT calculation [14]. The SphygmoCor® device uses an arterial tonometer to asynchronously record pressure waveforms of carotid and femoral arteries, which are then realigned with the ECG signal, and uses intersecting tangents algorithm to determine the foot of the wave and derive the PTT. In contrast, the Complior Analyse system records carotid and femoral waveforms simultaneously using mechanotransducers and references timing to the point of maximum systolic upstroke. Preliminary investigations have highlighted significant differences in PWV values between the SphygmoCor and Complior devices, emphasizing the crucial role of algorithms in PTT determination. The Reference Values for Arterial Stiffness Collaboration group has endorsed the intersecting tangents method as the reference standard for establishing aortic stiffness reference values, accompanied by a conversion formula for the maximum of the second-derivative method [15]. However, concerns have been raised regarding the reliability of this formula, particularly in patients with elevated arterial stiffness due to chronic diseases [16].

The aim of the present study was to eliminate variabilities in PTT that are related to biological factors, methodological differences in arterial waveforms recordings, and differences in device-based algorithms and to perform a head-to-head comparison of the two methods in determining PTT. More specifically, the objectives of the present study were to (1) calculate beat-by-beat PWV from synchronized pairs of carotid and femoral arterial pressure waveforms using two custom MATLAB algorithms, (2) compare PWV results obtained from our algorithms and those obtained by Complior’s system second-derivative algorithm, as well as those obtained by the processing of the same arterial waveforms through the SphygmoCor intersecting tangents algorithm, and (3) evaluate the validity of the proposed conversion formula in a population at a high risk of aortic stiffness.

Study Design and Patient Population

This is a cross-sectional study of heterogeneous participants comprising healthy individuals and patients with hypertension, chronic kidney disease, diabetes, and/or cardiovascular disease. Patients were excluded from the study if they had any clinical conditions that could impact arterial pressure wave recording such as impalpable arterial pulse at the site of measurement, carotid stenosis, or systolic blood pressure (SBP) under 80 mm Hg. Clinical, pharmaceutical, laboratory, and aortic stiffness data were collected through the review of medical records between September 2016 and June 2021.

Hemodynamic Assessment

Participants were instructed to refrain from engaging in intensive physical activity and abstain from consuming coffee or alcohol within 12 h prior to testing. Hemodynamic parameters were assessed in a quiet room after 10 min of rest in supine position [17]. Brachial artery blood pressure was measured using a validated automatic sphygmomanometer (BpTRU, Coquitlam, Canada). Blood pressure was recorded three times on the right arm within 2-min intervals between each measurement and an average of these measurements was calculated to determine brachial SBP and diastolic blood pressure. Carotid and femoral arterial pressure waveforms were simultaneously recorded using the piezoelectric sensors of the Complior® Analyse system in triplicates, with a sampling rate of 1,000 Hz. Three 10 s-long acquisitions were performed for each participant, and device-based algorithm (second derivative) was used to determine PTT (PTTComp-2nd) and PWV (PWVComp-2nd). As recommended by guidelines on measurement of PWV [18], 80% of the direct distance between carotid and femoral sites was used to obtain carotid-femoral PWV.

We then exported and used in-house algorithms (MATLAB; MathWorks, Natick, MA, USA) to determine PTT using both second-derivative (PTTMat-2nd) and intersecting tangents (PTTMat-IT) algorithms to calculate the respective PWVs (PWVMat-2nd and PWVMat-IT). Finally, we developed and used a Python-based script (Python Software Foundation; Wolfeboro Falls, NH, USA) to transform raw waveforms generated by the Complior® system into a format compatible with the SphygmoCor® CvMS system. This allowed us to determine PWV (PWVSphyg-IT) and calculate PTT (PTTSphyg-IT) using the SphygmoCor internal intersecting tangents algorithm.

In-House MATLAB-Based Algorithms

Raw carotid-femoral waveforms were extracted from the Complior system and imported into an in-house MATLAB software to determine PTT only for valid pressure waveforms. As illustrated in online supplementary Figure S1 (for all online suppl. material, see https://doi.org/10.1159/000543354), poor quality signals were excluded during PWV recalculation using MATLAB (cross zone denotes waves identified as nonvalid). To eliminate artifacts that are due to movements, raw waveforms were filtered using a moving average filter. Importantly, the same procedure of filtering and smoothing were applied on both the carotid and the femoral signals to prevent phase shifts that could result from differences in signal processing.

  • Maximum of second derivative: first, the waveform is passed through a low-frequency filter characterized by a linear phase, which allows low frequency signals to pass through while attenuating high frequency signals. The second derivative is then smoothed by triangle moving average filter before locating the time corresponding to the maximum value of the second derivative (Fig. 1a).

  • Intersecting tangent: the tangential method calculates the intersection of two tangents by geometric techniques. The first tangent is parallel to the time axis and reflects the diastolic minimum level; then the first derivative of the pulse wave is computed. The point where the first derivative reaches its maximum value is identified. This point corresponds to the steepest upward slope of the pulse wave, indicating the most rapid increase in the pulse wave amplitude. The second tangent passes tangentially through the maximum of the first derivative at five equidistant points. The intersection of these tangents is identified as the foot of the arterial pulse wave (Fig. 1b).

Fig. 1.

Foot wave detection algorithms. Detection of the foot waveform using the maximum of the second derivative in a and intersecting tangents in b. The black circle represents the foot of the wave.

Fig. 1.

Foot wave detection algorithms. Detection of the foot waveform using the maximum of the second derivative in a and intersecting tangents in b. The black circle represents the foot of the wave.

Close modal

PWV by SphygmoCor CvMS System

Given that PWV was measured using only the Complior® system (PWVComp-2nd), a script using Python software was used to develop a function that transforms raw carotid and femoral waveforms generated by the Complior® system into a format compatible with the SphygmoCor® CvMS system. Instead of generating simultaneous measurements as done by the Complior® system, a proxy ECG was generated to sequentially record carotid and femoral waveforms as required by SphygmoCor® CvMS software. The “simulation mode” in SphygmoCor® software was employed to calculate PWV values using the integrated intersecting tangent algorithm as used in the SphygmoCor® device. Overall, 80% of the carotid-femoral distance was used to calculate PWVSphyg-IT and PTTSphyg-IT. This process is applied to each acquisition conducted on every participant in the study.

Transformation of Second-Derivative PTT into Calculated Intersecting Tangents PTT

The conversion formula [15] used by the Reference Values for Arterial Stiffness Collaboration group was used to determine the calculated PTT intersecting tangents algorithm (PTTITc) using the PTT second-derivative algorithm obtained by the in-house MATLAB-based algorithm (PTTMat-2nd) or PTTComp-2nd as shown in the formula below:

Calculated PWV intersecting tangents algorithm (PWVITc) values were determined and compared to PWVMat-IT.

Statistical Analysis

All values are expressed as mean ± standard deviation. To assess the relationship between the two custom-made and device-based algorithms, linear regression analysis and Pearson correlation coefficients were used. To determine agreement between the two methods, Bland-Altman graphs were generated using PWV values, with calculation of the mean difference and limits of agreement [19]. A two-tailed p value <0.05 was considered statistically significant for all analyses. Data analyses were conducted using the SPSS 29 statistical software (SPSS Inc., Armonk, NY, USA).

Aortic stiffness assessments were performed in 113 participants (72 male) with a mean age of 59 ± 18 years (range 20–88). Table 1 presents baseline information on clinical characteristics, metabolic parameters, and medication taken by the studied population. On average, the brachial SBP and diastolic blood pressure were 118 ± 12 and 81 ± 10 mm Hg, respectively, as shown in Table 2.

Table 1.

Clinical characteristics for the study population (n = 113)

Age, years 59±18 (20–88) 
Male 72 (64%) 
Weight, kg 78.3±14.8 
Height, cm 167.8±8.8 
BMI, kg/m2 27.8±4.0 
Risk factors and comorbidities 
 Diabetes 19 (17%) 
 CVD 13 (12%) 
 Hypertension 68 (60%) 
 Smoking 8 (7%) 
 CKD 37 (33%) 
 Serum creatinine, µmol/L 214 (52–1,000) 
 Cholesterol, mmol/L 4.4±1.1 
 HDL, mmol/L 1.4±0.5 
 LDL, mmol/L 2.2±0.9 
 TG, mmol/L 1.76±1.15 
Medication 
 ACEi/ARB 46 (41%) 
 ß-Blocker 26 (23%) 
 Calcium channel blocker 37 (33%) 
 Diuretic 25 (22%) 
 Statin 50 (44%) 
Age, years 59±18 (20–88) 
Male 72 (64%) 
Weight, kg 78.3±14.8 
Height, cm 167.8±8.8 
BMI, kg/m2 27.8±4.0 
Risk factors and comorbidities 
 Diabetes 19 (17%) 
 CVD 13 (12%) 
 Hypertension 68 (60%) 
 Smoking 8 (7%) 
 CKD 37 (33%) 
 Serum creatinine, µmol/L 214 (52–1,000) 
 Cholesterol, mmol/L 4.4±1.1 
 HDL, mmol/L 1.4±0.5 
 LDL, mmol/L 2.2±0.9 
 TG, mmol/L 1.76±1.15 
Medication 
 ACEi/ARB 46 (41%) 
 ß-Blocker 26 (23%) 
 Calcium channel blocker 37 (33%) 
 Diuretic 25 (22%) 
 Statin 50 (44%) 

Values are mean ± SD, median (range), or n (%).

ACEi, angiotensin-converting enzyme inhibitors; ARB, angiotensin II receptor blockers; BMI, body mass index; CKD, chronic kidney disease; CVD, cardiovascular disease; HDL, high-density lipoprotein; LDL, low-density lipoprotein; TG, triglyceride.

Table 2.

Hemodynamic parameters for population

Mean per participant±SDRange
SBP, mm Hg 118±12 106–130 
DBP, mm Hg 81±10 70–88 
HR, bpm 56±12 45–72 
PWVMat-2nd, m/s 9.67±3.46 5.10–23.43 
PTTMat-2nd, ms 55.00±17.23 22.7–102.07 
PWVMat-IT, m/s 9.66±3.4 5.28–23.21 
PTTMat-IT, ms 54.92±17.03 22.92–102.12 
PWVComp-2nd, m/s 9.48±3.25 4.21–21.49 
PTTComp-2nd, ms 55.99±18.31 19.35–119.30 
PWVSphyg-IT, m/s 9.59±3.25 4.80–20.90 
PTTSphyg-IT, ms 55.07±17.21 19.90–105.00 
PWVITc, m/s 17.43±13.57 5.08–84.44 
PTTITc, ms 37.37±17.23 6.41–104.74 
Mean per participant±SDRange
SBP, mm Hg 118±12 106–130 
DBP, mm Hg 81±10 70–88 
HR, bpm 56±12 45–72 
PWVMat-2nd, m/s 9.67±3.46 5.10–23.43 
PTTMat-2nd, ms 55.00±17.23 22.7–102.07 
PWVMat-IT, m/s 9.66±3.4 5.28–23.21 
PTTMat-IT, ms 54.92±17.03 22.92–102.12 
PWVComp-2nd, m/s 9.48±3.25 4.21–21.49 
PTTComp-2nd, ms 55.99±18.31 19.35–119.30 
PWVSphyg-IT, m/s 9.59±3.25 4.80–20.90 
PTTSphyg-IT, ms 55.07±17.21 19.90–105.00 
PWVITc, m/s 17.43±13.57 5.08–84.44 
PTTITc, ms 37.37±17.23 6.41–104.74 

DBP, diastolic blood pressure; HR, heart rate; PWVMat-2nd, pulse wave velocity using MATLAB in-house second-derivative algorithm; PTTMat-2nd, pulse transit time using MATLAB in-house second-derivative algorithm; PWVMat-IT, pulse wave velocity using MATLAB in-house intersecting tangents algorithm; PTTMat-IT, pulse transit time using MATLAB in-house intersecting tangents algorithm; PWVComp-2nd, pulse wave velocity using Complior device-based second-derivative algorithm; PTTComp-2nd, pulse transit time using Complior device-based second-derivative algorithm; PWVSphyg-IT, pulse wave velocity using SphygmoCor device-based intersecting tangents algorithm; PTTSphyg-IT, pulse transit time using SphygmoCor device-based intersecting tangents algorithm; PWVITc, calculated pulse wave velocity intersecting tangents algorithm using the conversion formula; PTTITc, calculated pulse transit time intersecting tangents algorithm using the conversion formula; range, minimum value – maximum value; SBP, systolic blood pressure; SD, standard deviation.

Figure 2 shows the study flowchart. Overall, 346 10-s recordings of waveforms were available for the entire cohort. These waveforms were exported into MATLAB and overall, 4,102 eligible pairs of pulse waves were available for analysis. After visual inspection, 3,770 (92%) pairs of pulse waves were considered reliable for head-to-head comparison of both PWVMat-2nd and PWVMat-IT. Among the 346 original recordings, 295 waveforms (85%) generated by the Complior® Analyse were valid and compatible for analysis by the SphygmoCor® CvMS software.

Fig. 2.

Study flowchart describing the distribution of the studied techniques used to analyze PTT and PWV.

Fig. 2.

Study flowchart describing the distribution of the studied techniques used to analyze PTT and PWV.

Close modal

Beat-By-Beat PTTs and PWVs

The mean beat-by-beat TTMat-2nd was 54.55 ± 18.55 ms (range 15.00–129.00) and mean beat-by-beat PTTMat-IT was 54.61 ± 18.61 ms (range 15.00–126.00). For 3,770 beats analyzed, the ensuing mean beat-by-beat PWVMat 2nd was 10.00 ± 4.18 m/s (range 3.58–33.25) and mean beat-by-beat PWVMat-IT was 10.01 ± 4.23 m/s (range 3.70–34.13). The beat-by-beat correlation between PTTMat2nd and PTTMat-IT is excellent as shown in Figure 3a (r2 = 0.98, p < 0.001) with a slope of 0.99 and an intercept of 0.51 ms. Correspondingly, the Bland-Altman plot in Figure 3b shows the mean average of beat-by-beat PTTs. Similarly, in Figure 3c the beat-by-beat PWVMat-2nd and PWVMat-IT showed a strong agreement (r2 = 0.95, p < 0.001) with a slope of 0.98 and an intercept of −0.16 m/s. Figure 3d presents the corresponding Bland-Altman plot.

Fig. 3.

PTT according to foot detection algorithm. a Correlation of beat-by-beat PTT using MATLAB algorithms (PTTMat-2nd vs. PTTMat-IT). The broken line represents the line of identity. b Bland-Altman plot between PTTMat-2nd and PTTMat-IT. c Correlation of beat-by-beat PWV using MATLAB algorithms (PWVMat-2nd vs. PWVMat-IT). The broken line represents the line of identity. d Bland-Altman plot between PWVMat-2nd and PWVMat-IT.

Fig. 3.

PTT according to foot detection algorithm. a Correlation of beat-by-beat PTT using MATLAB algorithms (PTTMat-2nd vs. PTTMat-IT). The broken line represents the line of identity. b Bland-Altman plot between PTTMat-2nd and PTTMat-IT. c Correlation of beat-by-beat PWV using MATLAB algorithms (PWVMat-2nd vs. PWVMat-IT). The broken line represents the line of identity. d Bland-Altman plot between PWVMat-2nd and PWVMat-IT.

Close modal

PTTs and PWVs per Participant

The average PWVMat-2nd and PWVMat-IT per participant are shown in Table 2. The difference between PWVMat-2nd and PWVMat-IT was 0.01 ± 0.33 m/s (p = 0.35). Figure 4a shows the average PTTMat-2nd against the average of PTTMat-IT per participant, with excellent correlation and agreement as shown in the Bland-Altman plot (Fig. 4b). This is also the case for the average PWVMat-2nd and PWVMat-IT per participant (Fig. 4c, d).

Fig. 4.

Mean PTT and mean PWV per participant. a PTT correlation between mean PTTMat-2nd per participant and mean PTTMat-IT per participant. The broken line is the line of identity. b Bland-Altman plot between PTTMat-2nd and PTTMat-IT. c PWV correlation between mean PWV per participant using MATLAB in-house second-derivative algorithm (PWVMat-2nd) and mean PWV per participant using MATLAB in-house intersecting tangents algorithm (PWVMat-IT). The broken line is the line of identity. d Bland-Altman plot between PWVMat-2nd and PWVMat-IT.

Fig. 4.

Mean PTT and mean PWV per participant. a PTT correlation between mean PTTMat-2nd per participant and mean PTTMat-IT per participant. The broken line is the line of identity. b Bland-Altman plot between PTTMat-2nd and PTTMat-IT. c PWV correlation between mean PWV per participant using MATLAB in-house second-derivative algorithm (PWVMat-2nd) and mean PWV per participant using MATLAB in-house intersecting tangents algorithm (PWVMat-IT). The broken line is the line of identity. d Bland-Altman plot between PWVMat-2nd and PWVMat-IT.

Close modal

Using the system-based algorithms, the average PWVComp-2nd and PWVSphyg-IT per participant are reported in Table 2. Figure 5a shows a strong correlation between both device-based algorithms (r2 = 0.96, p < 0.001), a slope of 0.98, and an intercept of 0.3 m/s. Figure 5b shows an excellent agreement between the two device-based algorithms.

Fig. 5.

Mean PWV by device-based algorithms. a PWV correlation between mean PWV per participant using Complior device-based second-derivative algorithm (PWVComp-2nd) and mean PWV per participant using SphygmoCor device-based intersecting tangents algorithm (PWVSphyg-IT). b Bland-Altman plot between PWVComp-2nd and PWVSphyg-IT.

Fig. 5.

Mean PWV by device-based algorithms. a PWV correlation between mean PWV per participant using Complior device-based second-derivative algorithm (PWVComp-2nd) and mean PWV per participant using SphygmoCor device-based intersecting tangents algorithm (PWVSphyg-IT). b Bland-Altman plot between PWVComp-2nd and PWVSphyg-IT.

Close modal

Calculated Intersecting Tangents PTT and PWV

Transformation of the PTTMat-2nd into PTTITc using the proposed formula resulted in a shorter mean scatter plot (Fig. 6), showing that at low PWV values, the formula used to calculate PWVITc estimates PWV values quite well. However, the stiffer the aorta becomes, the greater the formula overestimates PWVITc. The difference between PWVITc and PWVMat-IT is statistically significant (p < 0.001).

Fig. 6.

Calculated PWV correlation. PWV correlation between PWVMat-IT and calculated PWV intersecting tangents algorithm using the conversion formula (PWVITc). The broken line is the line of identity.

Fig. 6.

Calculated PWV correlation. PWV correlation between PWVMat-IT and calculated PWV intersecting tangents algorithm using the conversion formula (PWVITc). The broken line is the line of identity.

Close modal

The present study shows that assessment of aortic stiffness by determination of carotid-femoral PWV is not significantly different between second-derivative and intersecting tangents approaches, two of the most commonly used algorithms to identify the location of the foot of the arterial pressure waveform. These results have been confirmed both by using in-house custom MATLAB algorithms and device-based algorithms.

The beat-by-beat custom-made MATLAB algorithms demonstrated high agreement in PWV values and PTT measurements. The correlation between the two algorithms was strong, following the identity line with negligible intercept, indicating consistency in the assessment of arterial stiffness using different mathematical methods. The Bland-Altman plots further supported the agreement between the two algorithms, with minimal differences between observed PWV values. These findings suggest that both of our MATLAB-based algorithms can provide reliable and consistent measurements of PWV in a beat-by-beat analysis.

One potential explanation for the previously reported differences between device-based algorithms might have been related to the signal processing system rather than the analytical method used. Our study also addressed the potential impact of device-based algorithms on PWV measurements. Processing Complior raw waveforms in the SphygmoCor system required developing a compatible function in Python software. The comparison between system-based algorithms (second-derivative algorithm from the Complior system and the intersecting tangents algorithm from the SphygmoCor system) revealed a strong correlation and agreement in PWV values. This indicates that despite differences in calculation methods, both device-based algorithms yielded comparable results.

Therefore, this leaves only one other potential explanation for the previously observed differences, which is the difference in signal acquisition between devices. Indeed, the sensors were different in the earlier generations of the Complior II and Complior SP, which used a sensor to generate distension waveforms, whereas the Complior Analyse system (used in the current study) used different piezoelectric sensors to record pressure signals sampled at 1 kHz [20]. Nevertheless, in the validation study of the Complior Analyse system, in which it is possible to use either second-derivative or the intersecting tangents algorithm, the PWV values were slightly lower for the second derivative (8.3 m/s ± 2.4 m/s vs. 8.6 ± 2.6 m/s) [21].

Due to concerns that PWVComp-2nd might underestimate PWVSphyg-IT [19], the Reference Values for Arterial Stiffness Collaboration group used a conversion formula to convert PTTComp-2nd to PTTITc [15]. This was used for the harmonization of the database and establishment of reference values but not abnormal values. As such, our findings demonstrate that the differences between PTTITc and PTTMat-IT were minimal when PWVMat-2nd values are in the low to normal range. As the equation subtracts 14.96 ms from PTTMat-2nd before dividing it by 0.8486, it results in an exponential increase in PWVITc for short transit times and could even result in negative values of PWV for PTTMat-2nd values below 14.96 ms (encountered in shorter subjects and increased aortic stiffness). Therefore, the use of calculated intersecting tangents PTT from a second-derivative PTT should be used with caution, and its validity should be restricted to normal values of PWV.

The study has several strengths and limitations. First, the heterogeneous nature of the study population and inclusion of beat-by-beat PTT analysis increased the range of PTT (range 15–129 ms). Second, the study results are based on a very large number of pairs of pulse waves that were simultaneously recorded, giving the same results either by custom-based or device-based filtering and algorithms. Finally, the PTT output from our custom MATLAB-based algorithms was in discrete format and this explains the Bland-Altman plot for PTT (Fig. 3b). However, the mean PWV per participant did not differ significantly between methods.

In conclusion, this study shows that second-derivative and intersecting tangents algorithms provide quasi-identical results across a wide range of PTT and underlines the limitations of the current conversion formula, which should be used with caution, especially in subjects at risk of high degree of aortic stiffness. As both methods yield similar results, the use of either algorithm is unlikely to significantly impact clinical practice. However, standardizing device-based quality control algorithms is crucial to ensure consistent performance across methods. These efforts will enhance the robustness and expand the applicability of PWV across a broader range of devices, thereby improving its utility in cardiovascular risk assessment in clinical practice.

We are grateful for the cooperation of all participants in this study.

The study was conducted in accordance with the Declaration of Helsinki. The protocol of this study was reviewed and approved by Research Ethics Committee (CER) of the CHU de Québec-Université Laval (Approval No.: 2012-163, 5.3.04.01; MP-32-2021-2173; MP-20-2021-5282). All participants enrolled voluntarily and provided explicit written informed consent before undertaking aortic stiffness assessments in Hotel-Dieu de Québec hospital (Canada).

All authors have reported that they have no relationships relevant to the contents of this paper to disclose.

A.T.: Master’s training scholarship for M.D.-M.Sc. – Fonds de recherche du Québec – Santé (FRQS) and Fondation du CHU de Québec, Université Laval. H.O.: scholarship from MITACS acceleration program and Fondation du CHU de Québec, Université Laval. Data used in this study were generated through studies funded by the Canadian Institute of Health Research.

Hasan Obeid and Mark Butlin were involved in algorithm development and provided software codes; Amira Tairi, Hasan Obeid, and Catherine Fortier were involved in data acquisition; Amira Tairi, Hasan Obeid, and Mohsen Agharazii were involved in data analysis and wrote drafts of the manuscript; Saliha Addour, Mark Butlin, Alberto P. Avolio, Catherine Fortier, and Mohsen Agharazii were involved in experimental supervision of studies and provided critical feedback.

The datasets generated and analyzed during the current study are not publicly available due to privacy or ethical restrictions but can be made available upon reasonable request. Access to the data requires approval from the Data Sharing Committee, which oversees the distribution of research materials in accordance with institutional guidelines. The committee consists of Doctor Mohsen Agharazii ([email protected]), the corresponding author, and Miss Amira Tairi ([email protected]), the first author. Requests for data sharing must include a detailed rationale, and any release of data will be subject to applicable ethical approvals and the signing of a data-sharing agreement to ensure confidentiality and appropriate use. Researchers interested in accessing the data are encouraged to contact the committee for further information and to discuss potential collaborations.

1.
Mitchell
GF
,
Hwang
SJ
,
Vasan
RS
,
Larson
MG
,
Pencina
MJ
,
Hamburg
NM
, et al
.
Arterial stiffness and cardiovascular events: the Framingham Heart Study
.
Circulation
.
2010
;
121
(
4
):
505
11
.
2.
Laurent
S
,
Boutouyrie
P
,
Asmar
R
,
Gautier
I
,
Laloux
B
,
Guize
L
, et al
.
Aortic stiffness is an independent predictor of all-cause and cardiovascular mortality in hypertensive patients
.
Hypertension
.
2001
;
37
(
5
):
1236
41
.
3.
Laurent
S
,
Katsahian
S
,
Fassot
C
,
Tropeano
AI
,
Gautier
I
,
Laloux
B
, et al
.
Aortic stiffness is an independent predictor of fatal stroke in essential hypertension
.
Stroke
.
2003
;
34
(
5
):
1203
6
.
4.
Zhang
Y
,
Lacolley
P
,
Protogerou
AD
,
Safar
ME
.
Arterial stiffness in hypertension and function of large arteries
.
Am J Hypertens
.
2020
;
33
(
4
):
291
6
.
5.
Zanoli
L
,
Lentini
P
,
Briet
M
,
Castellino
P
,
House
AA
,
London
GM
, et al
.
Arterial stiffness in the heart disease of CKD
.
J Am Soc Nephrol
.
2019
;
30
(
6
):
918
28
.
6.
Blacher
J
,
Guerin
AP
,
Pannier
B
,
Marchais
SJ
,
Safar
ME
,
London
GM
.
Impact of aortic stiffness on survival in end-stage renal disease
.
Circulation
.
1999
;
99
(
18
):
2434
9
.
7.
Terquem
A
.
D.-J. KORTEWEG. – Ueber die Fortpflanzungsgeschwindigkeit des Schalles in elastischen Röhren (Sur la transmission du son par les fluides renfermés dans les tubes à parois élastiques); Ann. der Physik und Chemie, t. V, p. 525; 1878
.
J Phys Theor Appl
.
1880
;
9
(
1
):
127
34
.
8.
Isebree Moens
A
,
Korteweg
DJ
.
On the speed of propagation of waves in elastic tubes [Internet]
. [cited 2024 Jul 20]. Available from: https://www.researchgate.net/publication/280623781_A_Isebree_Moens_and_DJ_Korteweg_On_the_speed_of_propagation_of_waves_in_elastic_tubes
9.
Fortier
C
,
Mac-Way
F
,
Desmeules
S
,
Marquis
K
,
De Serres
SA
,
Lebel
M
, et al
.
Aortic-brachial stiffness mismatch and mortality in dialysis population
.
Hypertension
.
2015
;
65
(
2
):
378
84
.
10.
Expert consensus document on arterial stiffness: methodological issues and clinical applications
.
Eur Heart J
. [cited 2024 Jul 19]. Available from: https://academic.oup.com/eurheartj/article/27/21/2588/2887386?login=false
11.
Prognostic value of aortic pulse wave velocity as index of arterial stiffness in the general population
.
Circulation [Internet]
. [cited 2024 Jul 19]. Available from: https://www.ahajournals.org/doi/10.1161/CIRCULATIONAHA.105.579342
12.
Chiu
YC
,
Arand
PW
,
Shroff
SG
,
Feldman
T
,
Carroll
JD
.
Determination of pulse wave velocities with computerized algorithms
.
Am Heart J
.
1991
;
121
(
5
):
1460
70
.
13.
O’Rourke
MF
,
Hayward
CS
.
Arterial stiffness, gender and heart rate
.
J Hypertens
.
2003
;
21
(
3
):
487
90
.
14.
Pannier
BM
,
Avolio
AP
,
Hoeks
A
,
Mancia
G
,
Takazawa
K
.
Methods and devices for measuring arterial compliance in humans
.
Am J Hypertens
.
2002
;
15
(
8
):
743
53
.
15.
The Reference Values for Arterial Stiffness’ Collaboration
.
Determinants of pulse wave velocity in healthy people and in the presence of cardiovascular risk factors: ‘establishing normal and reference values’
.
Eur Heart J
.
2010
;
31
(
19
):
2338
50
.
16.
Côté
N
,
Fortier
C
,
Jafari
K
,
Paré
M
,
Addour
S
,
Goupil
R
, et al
.
Estimated versus measured aortic stiffness: implications of diabetes, chronic kidney disease, sex and height
.
J Hypertens
.
2024
;
42
(
12
):
2115
21
.
17.
Fortier
C
,
Garneau
CA
,
Paré
M
,
Obeid
H
,
Côté
N
,
Duval
K
, et al
.
Modulation of arterial stiffness gradient by acute administration of nitroglycerin
.
Front Physiol
.
2021
;
12
:
774056
.
18.
Van Bortel
LM
,
Laurent
S
,
Boutouyrie
P
,
Chowienczyk
P
,
Cruickshank
JK
,
De Backer
T
, et al
.
Expert consensus document on the measurement of aortic stiffness in daily practice using carotid-femoral pulse wave velocity
.
J Hypertens
.
2012
;
30
(
3
):
445
8
.
19.
Millasseau
SC
,
Stewart
AD
,
Patel
SJ
,
Redwood
SR
,
Chowienczyk
PJ
.
Evaluation of carotid-femoral pulse wave velocity: influence of timing algorithm and heart rate
.
Hypertension
.
2005
;
45
(
2
):
222
6
.
20.
Validation of the new complior sensor to record pressure signals non-invasively
.
J Clin Monit Comput [Internet]
. [cited 2024 Jul 21]. Available from: https://link.springer.com/article/10.1007/s10877-013-9477-y
21.
Stea
F
,
Bozec
E
,
Millasseau
S
,
Khettab
H
,
Boutouyrie
P
,
Laurent
S
.
Comparison of the Complior Analyse device with Sphygmocor and Complior SP for pulse wave velocity and central pressure assessment
.
J Hypertens
.
2014
;
32
(
4
):
873
80
.