Original Article

Logistic Regression Model Based on Ultrafast Pulse Wave Velocity and Different Feature Selection Methods to Predict the Risk of Hypertension


Background: Hypertension is the main reason why the incidence of cardiovascular disease has increased year-by-year and early diagnosis of hypertension is necessary to reducing the incidence of cardiovascular disease. This also puts forward higher requirements for the accuracy of diagnosis. We tried a variety of feature selection methods to improve the accuracy of logistic regression (LR).

Methods: We collected 397 samples from Nanjing, Jiangsu, China between Jan 2016 and Dec 2017, including 178 hypertension samples and 219 control samples. It includes not only clinical and laboratory data, but also imaging data. We focused on the difference of imaging attributes between the control group and the hypertension group, and analyzed the correlation coefficients of all attributes. In order to establish the optimal LR model, this study tried three different feature selection methods, including statistical analysis, random forest (RF) and extreme gradient boosting (XGBoost). The area under the ROC curve (AUC) and accuracy were used as the main criterion for model evaluation.

Results: In the prediction of hypertension, the performance of LR with RF as the feature selection method (accuracy: 0.910; AUC: 0.924) was better than the performance of LR with XGBoost as the feature selection method (accuracy: 0.897; AUC: 0.915) and the performance of LR with statistical analysis as the feature selection method (accuracy: 0.872; AUC: 0.926).

Conclusion: LR with RF as the feature selection method may provide accurate results in predicting hypertension. Carotid intima-media thickness (cIMT) and pulse wave velocity at the end of systole (ESPWV) are two key imaging indicators in the prediction of hypertension.

1. Wang TJ, Gona P, Larson MG, et al (2006). Multiple biomarkers for the prediction of first major cardiovascular events and death. N Engl J Med, 355:2631-2639.
2. Laurent S, Boutouyrie P, Asmar R, et al (2001). Aortic stiffness is an independent predictor of all-cause and cardiovascular mortality in hypertensive patients. Hyper-tension, 37:1236-1241.
3. Bérard E, Bongard V, Ruidavets JB, Amar J, Ferrières J (2013). Pulse wave velocity, pulse pressure and number of carotid or femoral plaques improve prediction of cardiovascular death in a population at low risk. J Hum Hypertens, 27:529–534.
4. O'Brien E, Atkins N, Stergiou G, et al (2010). European society of hypertension international protocol revision 2010 for the validation of blood pressure measur-ing devices in adults. Blood Press Monit, 15:23-38.
5. Cruickshank K, Riste L, Anderson Simon G, et al (2002). Aortic pulse-wave velocity and its relationship to mortality in diabe-tes and glucose intolerance: an integrated index of vascular function? Circulation, 106:2085-2090.
6. Townsend RR, Wimmer NJ, Chirinos JA, et al (2010). Aortic PWV in chronic kidney disease: a CRIC ancillary study. Am J Hy-pertens, 23:282-289.
7. Lorenz MW, Polak JF, Kavousi M, et al (2012). Carotid intima-media thickness progression to predict cardiovascular events in the general population (the PROG-IMT collaborative project): a me-ta-analysis of individual participant data. Lancet, 379:2053-2062.
8. Williams B, Mancia G, Spiering W, et al (2018). 2018 ESC/ESH Guidelines for the management of arterial hypertension: The Task Force for the management of arterial hypertension of the European So-ciety of Cardiology (ESC) and the Euro-pean Society of Hypertension (ESH). Eur Heart J, 39:3021-3104.
9. Coude M, Pernot M, Messas E, et al (2011). Ultrafast imaging of the arterial pulse wave. IRBM, 32(2): 106-108.
10. Robertson CM, Gerry F, Fowkes R, et al (2012). Carotid intima-media thickness and the prediction of vascular events. Vasc Med, 17:239-248.
11. Vermeersch SJ, Dynamics B, Society L (2010). Determinants of pulse wave ve-locity in healthy people and in the pres-ence of cardiovascular risk factors: estab-lishing normal and reference values. Eur Heart J, 31:2338-2350.
12. Wilkinson IB, McEniery CM, Schillaci G, et al (2010). ARTERY Society guidelines for validation of non-invasive haemodynam-ic measurement devices: Part 1, arterial pulse wave velocity. Artery Res, 4:34-40.
13. Zhu ZQ, Chen LS, Jiang XZ, et al (2021). Absent atherosclerotic risk factors are as-sociated with carotid stiffening quantified with ultrafast ultrasound imaging. Eur Ra-diol, 31:3195-3206.
14. Zhu ZQ, Chen LS, Wang H, et al (2019). Carotid stiffness and atherosclerotic risk: non-invasive quantification with ultrafast ultrasound pulse wave velocity. Eur Radiol, 29:1507-1517.
15. Mirault T, Pernot M, Frank M, et al (2015). Carotid stiffness change over the cardiac cycle by ultrafast ultrasound imaging in healthy volunteers and vascular Ehlers-Danlos syndrome. J Hypertens, 33:1890-1896.
16. Yang W, Wang Y, Yu Y, et al (2020). Estab-lishing normal reference value of carotid ultrafast pulse wave velocity and evaluat-ing changes on coronary slow flow. Int J Cardiovas Imaging, 36:1931-1939.
17. Breiman L (2001). Random Forests. Mach Learn, 45:5-32.
18. Chen TQ, Guestrin C (2016). XGBoost: A scalable tree boosting system. ACM, New York, pp785-794.
19. Hermeling E, Reesink KD, Kornmann LM, et al (2009). The dicrotic notch as alterna-tive time-reference point to measure local pulse wave velocity in the carotid artery by means of ultrasonography. J Hypertens, 27:2028-2035.
IssueVol 51 No 9 (2022) QRcode
SectionOriginal Article(s)
DOI https://doi.org/10.18502/ijph.v51i9.10565
Hypertension Ultrafast pulse wave velocity Feature selection Logistic regression

Rights and permissions
Creative Commons License This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
How to Cite
Bai X, Liu W, Huang H, You H. Logistic Regression Model Based on Ultrafast Pulse Wave Velocity and Different Feature Selection Methods to Predict the Risk of Hypertension. Iran J Public Health. 2022;51(9):2099-2107.