Introduction
The traditional risk factors for stroke have been identi-fied by various large clinical trial studies conducted over many years1,2. In addition, stroke prevention with var-ious medications such as aspirin has been shown to be effective1,3,4. However, stroke remains one of the top ten causes of death in Taiwan5. A key factor is lack of effective and portable screening tools. Advances in technology and new inventions, such as electron beam
C
ORRELATION
B
ETWEEN
A
RTERIAL
S
TIFFNESS
I
NDEX AND
A
RTERIAL
W
AVE
P
ATTERN AND
I
NCIDENCE OF
S
TROKE
Cheuk-Sing Choy1,2,3, David Yen-Ju Wang1, Tu-Bin Chu4,5†, Wai-Mau Choi6*, Hung-Yi Chiou3 1Department of Emergency Medicine, Taipei Medical University Hospital, 2Department of Medicine, School of Medicine, Taipei Medical University, 3Graduate Institute of Public Health, Taipei Medical University,
4Department and Graduate Institute of Business Administration, National Taiwan University, 5Taipei Medical University Hospital, Taipei, and 6Department of Emergency Medicine,
Mackay Memorial Hospital, Hsinchu Branch, Taiwan.
SUMMARY
Background:Although there are several extensive studies on stroke risks and prevention, stroke has remained one of the top ten causes of death in Taiwan for many years. The key factor is a lack of an effective and nonin-vasive screening tool for stroke risk. The arterial stiffness index (ASI), which measures artery distensibility and is correlated with arthrosclerosis, is gaining popularity nowadays for stroke risk assessment. In our study, we investigated the relationship between stroke and ASI as well as other noninvasive screening tools such as the ankle brachial index (ABI) and arterial wave pattern.
Methods:This community-based prospective study was conducted between August 2005 and June 2006. The control group consisted of 629 volunteer adults above 30 years of age living in the northern Taipei area. The stroke group consisted of 266 newly diagnosed stroke patients. Participants completed a structured question-naire, and blood samples were collected. In addition, a validated oscillometric automated digital blood pres-sure device was used to meapres-sure the participant’s ASI, ABI and arterial wave pattern. The odds ratio was then computed to evaluate the association between each factor and stroke.
Results:Our study showed that abnormal ASI (ASI <70) was associated with a six times higher risk of stroke. Even after adjustment, the adjusted odds ratio was still 1.8. We also found a significant association between stroke and both ABI and arterial wave pattern (odds ratios, 2.15 and 2.98, respectively). In addition, when we employed ASI and arterial wave pattern together, the adjusted odds ratio for stroke was 1.87. The odds ratio increased significantly to 10.53 when all three factors were taken into consideration.
Conclusion:Our study showed that ASI, arterial wave pattern and ABI are correlated with stroke risk. In addition, when all these factors are taken into consideration, they create a synergistic effect in evaluating the risk of stroke risk. [International Journal of Gerontology 2010; 4(2): 75–81]
Key Words:ankle brachial index, atherosclerosis, cardiovascular diseases, stroke
*Correspondence to:Dr Wai-Mau Choi, Department of Emergency Medicine, Mackay Memorial Hospital, Hsinchu Branch, No. 690, Section 2, Guangfu Road, Husinchu 300, Taiwan.
E-mail: [email protected] Accepted: November 18, 2009
†Tu-Bin Chu and the corresponding author
computed tomography, magnetic resonance coronary angiography, carotid ultrasonography, and high sensi-tivity C-reactive protein (CRP) and lipid profiling, all have their pitfalls when used as screening tools on a large scale. This includes high cost, poor accessibility, inade-quate accuracy, or invasiveness6–9. Because of these inconveniences, patients often receive an examination after presentation of symptoms, resulting in a delay in treatment. Thus, it is essential to find a cost-effective method with good accessibility to the public for stroke risk assessment.
Among the current methods for noninvasive screen-ing tools, the arterial stiffness index (ASI) shows promise. ASI is a measurement of the compliance and distensi-bility of the arteries, and is directly correlated with atherosclerosis10. However, there is still some debate about the optimal method for ASI calculation and mea-surement. Among them, aortic pulse wave velocity (PWV) and 24-hour ambulatory ASI are the most popular for ASI calculation at the present time. However, there are extensive studies demonstrating a strong correlation between stroke risk and both indices11–14. Measure-ment requires extensive effort, which violates our goal of simplicity and accessibility. In our study, we chose a simple technique in which ASI is measured by a blood pressure (BP) cuff on the subject’s arm. To our knowl-edge, current research implementing this technique either is limited in sample size or was not tailored for stroke risk assessment. Furthermore, we have measured the ankle brachial index (ABI), because it is considered to be an indicator for atherosclerosis as well as being easy to obtain by simply applying an additional BP cuff to the lower leg15. We have conducted a large prospec-tive study aimed at investigating the correlation be-tween ASI and stroke risk and whether other risk factors such as ABI have an impact on this assessment.
Materials and Methods
Study population
Two hundred and sixty-six patients, aged 26–98 years and with stroke newly diagnosed by experienced neurol-ogy doctors, were recruited from the emergency depart-ment of Shin Kong Wu Ho–Su Memorial Hospital from August 2005 to June 2006. A total of 629 control sub-jects with no evidence of stroke were recruited from community-based health examinations in the same area. The study was approved by the institutional review
board for human subjects of Taipei Medical University, Taipei, Taiwan, and each subject provided written informed consent.
Data collection
Well-trained personnel carried out standardized per-sonal interviews based on a structured questionnaire. Information collected included demographic and so-cioeconomic characteristics, and personal medical his-tory. We measured each participant’s body weight and height. These data were used to calculate body mass index. In addition, we asked participants to fast for more than 8 hours before blood collection. Total cholesterol, high-density lipoprotein cholesterol, low-density lipo-protein cholesterol, glucose and CRP were measured.
After the subjects had rested for at least 5 minutes, BP was measured by a validated oscillometric auto-mated digital BP device (CardioVision MS 1900; Medmax Medical Technology Corp. Co., Ltd., Taichung, Taiwan). If a BP reading between two measurements differed by more than 10 mmHg, an additional BP measurement was carried out. The average BP was calculated from the two nearest BP readings. Based on the collected data, the computer would calculate an ASI and arterial wave pattern. In addition, another BP cuff was placed on the participant’s lower leg for ABI measurement. We de-fined the normal values as ASI >7014, ABI of 0.9–1.3, and arterial wave pattern of type A. Data outside this range were considered an indication of high risk of stroke.
Statistical methods
Statistical analyses were performed using SAS 9.1 soft-ware (SAS Institute Inc., Cary, NC, USA). The χ2test was used to test the differences between categorical vari-ables and the Student ttest to examine the differences in continuous variables between cases with and with-out stroke. Logistic regression models were further used to estimate the multivariate-adjusted odds ratios (ORs) and their 95% confidence intervals for stroke risk.
Results
Of the total 895 participants, 266 were stroke patients. We found that stroke patients were older and had a higher percentage were smokers and had diabetes mellitus (DM) (Table 1). In addition, they tended to have elevated sys-tolic BP, blood sugar, triglyceride and CRP, but lower high-density lipoprotein and total cholesterol levels. Sex, body
mass index, diastolic BP and low-density lipoprotein were not significant characteristics among stroke patients.
Table 2 illustrates that only DM and age showed statistical significance on ASI, while other factors such as cholesterol had little impact on calculation of ASI. However, we found that cholesterol along with sex, DM, age and body mass index did show varying degrees of significance regarding arterial wave pattern.
To investigate the correlation between ASI, ABI and arterial wave pattern with stroke events, we calculated their ORs against stroke. In Table 3, we show that patients with abnormal ASI were six times more likely to have a stroke compared with patients with normal ASI. Even after adjustment for age, sex, smoking, DM and cholesterol (for factors that showed a statistically significant impact in Table 2), the adjusted OR was still 1.8 times higher than in patients with a normal ASI. Therefore, abnormal ASI is a significant risk factor for stroke. With the same methodology, we found that abnormal arterial wave pattern and ABI had adjusted OR values of 2.98 and 2.15, respectively. This indicates
that all three measurements were significantly associ-ated with the risk of stroke.
Although we found that all three factors were indi-vidually associated with stroke, we explored whether adding them together created a synergistic effect. We set an ASI <70, type A arterial wave pattern and ABI be-tween 0.9 and 1.3 as our norm. As shown in Table 3, we found that the adjusted OR for abnormal ASI was 1.81. When we added an abnormal arterial wave pattern to our model, the adjusted OR increased to 1.87 (Table 4). Furthermore, the adjusted OR increased dramatically to 10.53 when all three factors were abnormal. This indi-cated that a patient’s stroke risk increased by about 10 times when all three factors were present.
Discussion
Nowadays, most studies focus on measuring the 24-hour ambulatory ASI after continuously measuring BP for 24 hours or aortic PWV via the femoral artery. These
Table 1. Basic characteristics of stroke patients and urban volunteers*
Variable Total no. of subjects Stroke patients (n=266) Urban residents (n=629) p†
Sex 0.06 Male 454 148 (32.60) 306 (67.40) Female 441 118 (26.76) 323 (73.24) Smoking‡ <0.01 No 633 165 (26.07) 468 (73.93) Yes 254 93 (36.61) 161 (63.39) DM§ No 699 140 (20.03) 559 (79.97) <0.01 Yes 190 125 (65.79) 65 (34.21) Age (yr) 65.80±0.83 56.72±0.43 <0.01 BMI (kg/cm2) 24.58±0.41 24.23±0.13 0.41 SBP (mmHg) 161.75±1.92 125.73±0.77 <0.01 DBP (mmHg) 77.55±1.10 78.73±0.44 0.32 Glucose (mg/dL) 116.93±2.95 98.19±0.91 <0.01 Triglyceride (mg/dL) 144.74±5.64 123.78±3.33 <0.01 Cholesterol (mg/dL) 185.53±3.31 204.04±1.51 <0.01 HDL-Cho (mg/dL) 47.74±1.01 50.46±0.57 0.02 LDL-Cho (mg/dL) 129.38±3.73 129.95±1.39 0.89 CRP (mg/dL) 1.74±0.24 0.19±0.01 <0.01 *Data are presented as n (% of total number of subjects) or mean±standard error; †c2test for categorical variables and Student t test for
continuous variables; ‡eight subjects’ data were missing; §six subjects’ data were missing. DM=diabetes mellitus; BMI=body mass index;
SBP=systolic blood pressure; DBP=diastolic blood pressure; HDL-Cho=high-density lipoprotein cholesterol; LDL-Cho=low-density lipoprotein cholesterol; CRP=C-reactive protein.
Table 2. Significance of basic characteristics and lipid profiles against arterial stiffness index (ASI) and arterial wave pattern* Arterial wave pattern
p‡ ASI p† A AC Other Sex 0.88 0.01 Male 62.13±1.43 316 (69.60) 87 (19.16) 51 (11.23) Female 61.79±1.84 299 (67.80) 65 (14.74) 77 (17.46) Smoking§ 0.56 0.18 No 61.40±1.47 442 (69.83) 96 (15.17) 95 (15.01) Yes 62.74±1.78 170 (66.93) 51 (20.08) 33 (12.99) DM <0.01 <0.01 No 56.87±0.98 526 (75.25) 83 (11.87) 90 (12.88) Yes 80.95±3.82 84 (44.21) 68 (35.79) 38 (20.00) Age (yr) <0.01 <0.01 <65 55.50±0.98 458 (77.63) 71 (12.03) 61 (10.34) ≥65 74.47±2.70 157 (51.48) 81 (26.56) 67 (21.97) BMI (kg/cm2) 0.09 0.02 <27 57.02±1.17 445 (76.86) 65 (11.23) 69 (11.92) ≥27 61.47±2.45 106 (72.11) 29 (19.73) 12 (8.16) Triglyceride (mg/dL) 0.41 0.20 <200 61.59±1.27 523 (69.00) 123 (16.23) 112 (14.78) ≥200 64.47±3.31 82 (70.09) 24 (20.51) 11 (9.40) Cholesterol (mg/dL) 0.25 <0.01 <240 62.47±1.28 501 (66.98) 133 (17.78) 114 (15.24) ≥240 58.63±3.02 104 (81.25) 14 (10.94) 10 (7.81) HDL (mg/dL) 0.95 0.78 ≥40 61.10±1.42 434 (70.68) 91 (14.82) 89 (14.50) <40 61.27±2.31 155 (72.77) 31 (14.55) 27 (12.68) LDL (mg/dL) 0.65 0.21 <160 60.86±1.33 471 (70.09) 103 (15.33) 98 (14.58) ≥160 62.30±3.04 112 (77.24) 16 (11.03) 17 (11.72)
*Data are presented as mean±standard error or n (%); †Student t test; ‡c2test; §eight subjects’ data were missing; ||six subjects’ data were missing.
DM=diabetes mellitus; BMI=body mass index; HDL=high-density lipoprotein; LDL=low-density lipoprotein. measurement methods either are time consuming or
require invasive procedures. Other noninvasive tech-niques such as brachial/ankle PWV have been attempted with mixed results for stroke risk assessment16–20. In our study, ASI was measured by connecting a sphygmo-manometer to a laptop. The computer-assisted oscillo-metric sensor detected a change in BP at the brachial artery and recorded its pulse wave amplitude pattern to formulate the ASI. We found that the frequency of an abnormal ASI increased with stroke incidence. After adjustment for other risk factors, subjects with an ASI level>70 had a 1.8-fold increase in the risk of stroke. In addition, the risk increased exponentially to tenfold when the arterial wave pattern and ABI were also abnormal.
Our study also showed that ABI has a great influence on stroke risk. Traditionally, ABI is used to evaluate pe-ripheral arterial obstructive disease, which is also an atherosclerotic disease15. However, the relationships between ABI, ASI and atherosclerosis are still contro-versial. One study compared ABI and brachial/ankle PWV in patients receiving hemodialysis or those who had chronic renal disease, and the authors found that ABI<0.9 was associated with increased age, DM and increased pulse pressure, while brachial/ankle PWV was positively associated with age, DM, and systolic and dia-stolic BP21. Another study found that a group with high ABI (ABI≥1.3) had the characteristics of male sex, larger waist circumference, and lower high-density lipoprotein and blood glucose, but this did not correlate with PWV22.
Furthermore, one investigator found that there was a strong association between ABI and ASI15while another found that a high ABI value might not be a marker for atherosclerosis risk in the general population22. How-ever, few studies have investigated whether a combi-nation of ABI and ASI creates a synergistic effect in stroke risk prediction. Regarding our findings, further research may help determine if ABI will increase the power to predict stroke risk.
Although our research is still in its primitive stages, we are able to conclude that ASI is a good screening tool for stroke risk, especially if further analysis is still re-quired. Compared with other highly sensitive tech-niques such as sonography or magnetic resonance coronary angiography9,23,24, we believe that this tech-nique has good accessibility, portability and cost effec-tiveness, which will be more influential for stroke prevention. For instance, the public has been aware of the danger of both stroke and hypertension for a long time. Similarly, the clinical course for both illnesses is silent until the advanced stages. However, over 60% of relevant subjects receive antihypertensive treatment,
while the number participating in a primary prevention program for stroke is much fewer2,25. We believe this is because sphygmomanometers are widely available in Taiwan. They can be found in hospitals, clinics, pharma-cies, and many public spaces. Thus, the elderly usually have an awareness of their BP levels and seek medical attention before hypertension becomes uncontrollable. However, most people lack awareness about stroke and often receive treatment only after severe symptoms have presented25. Despite a minor reduction in accu-racy, an inexpensive method such as ASI can screen out lower-risk patients and redirect higher-risk patients to hospital for extensive workup and treatment, which would not only use scarce medical resources more effi-ciently, but would also increase public awareness about stroke when such a tool is widely available.
There are several limitations in our study including measuring error, reoccurrence of stroke in patients, and the effects of drugs. We calculated ASI based primarily on BP readings, which may vary with a participant’s condition on the examination date. However, we tried to eliminate this error by asking participants to take a
Table 3. Odds ratios (ORs) and 95% confidence interval (CI) for associations of arterial wave pattern, arterial stiffness index (ASI) and ankle brachial index (ABI) with the risk of stroke*
Variable Total no. Stroke patients Urban residents Crude OR (95% CI) OR† (95% CI) of subjects (n=266) (n=629) ASI <70 698 145 (20.77) 553 (79.23) 1.0 1.0 ≥70 197 121 (61.42) 76 (38.58) 6.07 (4.32–8.53)‡ 1.80 (1.12–2.89)§ Pattern A 615 109 (82.28) 506 (17.72) 1.0 1.0 Non-A 280 157 (56.07) 123 (43.93) 5.93 (4.33–8.11)‡ 2.98 (1.86–4.49)‡ ABI 0.9–1.3 714 178 (24.93) 536 (75.07) 1.0 1.0 <0.9 or >1.3 88 47 (53.41) 41 (46.59) 3.45 (2.20–5.42)‡ 2.15 (1.12–4.14)§ Missing 93 41 (44.09) 52 (55.91) 2.37(1.52–3.70)‡ 1.79 (0.92–3.47) *Data are presented as n (% of total number of subjects); †adjustment for age, sex, smoking status, systolic blood pressure, diabetes mellitus,
cholesterol in the model; ‡p<0.001; §p<0.05; 0.05<p<0.1.
Table 4. Multivariate-adjusted odds ratios (ORs) for stroke risk with regard to arterial wave pattern, arterial stiffness index (ASI), and ankle brachial index (ABI)
ASI Pattern ABI Crude OR (95% CI) OR* (95% CI)
<70 A 1.3–0.9 1.0 1.0
≥70 Non-A 7.28 (4.87–10.89)† 1.87 (1.05–3.31)‡
≥70 Non-A >1.3 or <0.9 36.42 (10.59–125.20)† 10.53 (2.09–52.97)† *Adjustment for age, sex, smoking status, systolic blood pressure, diabetes mellitus, cholesterol in the model; †p<0.01; ‡p<0.05. CI=
break before the examination and we rechecked the patient’s BP if a large fluctuation was found. However, the error may still have been present. Secondly, we ex-cluded prior experience of stroke by questionnaire or initial physical examination, and some subjects who had experienced a transient ischemic attack or minor stroke may not have been identified. Finally, the aver-age aver-age for both groups was around 60 years, and many were taking several drugs for various reasons. Nevertheless, there are many studies denying an effect of most antihypertensive drugs on calculation of ASI, and evidence is lacking of an effect of many other medications26.
In conclusion, our study showed that subjects with higher ASI are likely to have a higher incidence of stroke. When ASI, arterial wave pattern and ABI are considered together, this creates a synergistic effect and enhances stroke risk assessment. However, further studies are required to evaluate if these methods can serve as effective screening tools for stroke risk prediction in practice.
References
1. Sacco RL, Adams R, Albers G, et al. Guidelines for pre-vention of stroke in patients with ischemic stroke or transient ischemic attack: a statement for healthcare professionals from the American Heart Association/ American Stroke Association Council on Stroke: co-sponsored by the Council on Cardiovascular Radiology and Intervention: the American Academy of Neurology affirms the value of this guideline. Circulation 2006; 113: e409–49.
2. Lloyd-Jones D, Adams R, Carnethon M, et al. Heart disease and stroke statistics—2009 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2009; 119: e21–181.
3. Gillum RF, Gorelick PB, Copper ES. Stroke in Blacks: A Guide to Management and Prevention. Basel: Karger, 1999.
4. Wood DA. Clinical reality of coronary prevention guide-lines: a comparison of EUROASPIRE I and II in nine countries. Lancet 2001; 357: 995–1001.
5. Hu HH, Sheng WY, Chu FL, et al. Incidence of stroke in Taiwan. Stroke 1992; 23: 1237–41.
6. Arad Y, Spadaro LA, Goodman K, et al. Predictive value of electron beam computed tomography of the coronary arteries: 19-month follow-up of 1173 asymptomatic subjects. Circulation 1996; 93: 1951–3.
7. Pitt B, Rubenfire M. Risk stratification for the detection of preclinical coronary artery disease. Circulation 1999; 99: 2610–2.
8. Smith SC Jr, Greenland P, Grundy SM. Prevention confer-ence V: beyond secondary prevention: identifying the high-risk patient for primary prevention: executive summary. Circulation 2000; 101: 111–6.
9. Vogel RA, Benitez RM. Noninvasive assessment of car-diovascular risk: from Framingham to the future. Rev Cardiovasc Med 2000; 1: 34–42.
10. Safar ME, Levy BI, Struijker-Boudier H. Current perspec-tives on arterial stiffness and pulse pressure in hyper-tension and cardiovascular diseases. Circulation 2003; 107: 2864–9.
11. Hansen TW, Staessen JA, Torp-Pedersen C, et al. Ambu-latory arterial stiffness index predicts stroke in a general population. J Hypertens 2006; 24: 2247–53.
12. Dolan E, Thijs L, Li Y, et al. Ambulatory arterial stiffness index as a predictor of cardiovascular mortality in the Dublin Outcome Study. Hypertension 2006; 47: 365–70. 13. Muxfeldt ES, Fiszman R, Castelpoggi CH, et al. Ambulatory arterial stiffness index or pulse pressure: which correlates better with arterial stiffness in resistant hypertension? Hypertens Res 2008; 31: 607–13.
14. Inoue N, Maeda R, Kawakami H, et al. Aortic pulse wave velocity predicts cardiovascular mortality in middle-aged and elderly Japanese men. Circ J 2009; 73: 549–53. 15. Maldonado J, Pereira T, Resende M, et al. Usefulness
of the ankle-brachial index in assessing vascular func-tion in normal individuals. Rev Port Cardiol 2008; 27: 465–76.
16. Choi JC, Lee JS, Kang SY, et al. Limitation of brachial-ankle pulse wave velocity in assessing the risk of stroke: importance of instantaneous blood pressure. Cerebrovasc Dis 2009; 27: 417–25.
17. Kim DH, Kim J, Kim JM, et al. Increased brachial-ankle pulse wave velocity is independently associated with risk of cerebral ischemic small vessel disease in elderly hypertensive patients. Clin Neurol Neurosurg 2008; 110: 599–604.
18. Huck CJ, Bronas UG, Williamson EB, et al. Noninvasive measurements of arterial stiffness: repeatability and interrelationships with endothelial function and arterial morphology measures. Vasc Health Risk Manag 2007; 3: 343–9.
19. Nichols WW, Denardo SJ, Wilkinson IB, et al. Effects of arterial stiffness, pulse wave velocity, and wave reflections on the central aortic pressure waveform. J Clin Hypertens (Greenwich) 2008; 10: 295–303.
20. Woodman RJ, Kingwell BA, Beilin LJ, et al. Assessment of central and peripheral arterial stiffness: studies indicat-ing the need to use a combination of techniques. Am J Hypertens 2005; 18: 249–60.
21. Chen SC, Chang JM, Hwang SJ, et al. Comparison of ankle-brachial index and brachial-ankle pulse wave velocity between patients with chronic kidney disease and hemodialysis. Am J Nephrol 2009; 29: 374–80. 22. Tabara Y, Igase M, Kido T, et al. Composition of lower
extremity in relation to a high ankle-brachial index. J Hypertens 2009; 27: 167–73.
23. Simons PC, Algra A, Bots ML, et al. Common carotid intima-media thickness and arterial stiffness: indicators of cardiovascular risk in high-risk patients. The SMART Study (Second Manifestations of ARTerial disease). Circulation 1999; 100: 951–7.
24. Hundley WG, Hamilton CA, Clarke GD, et al. Visualiza-tion and funcVisualiza-tional assessment of proximal and middle left anterior descending coronary stenoses in humans with magnetic resonance imaging. Circulation 1999; 99: 3248–54.
25. Goldstein LB, Adams R, Becker K, et al. Primary preven-tion of ischemic stroke: a statement for healthcare pro-fessionals from the Stroke Council of the American Heart Association. Circulation 2001; 103: 163–82.
26. Davies JI, Struthers AD. Pulse wave analysis and pulse wave velocity: a critical review of their strengths and weaknesses. J Hypertens 2003; 21: 463–72.
