ARTICLES

See REVIEWER COmmENTARY page 1073

nature publishing group

Pulse Wave Analysis Is a Reproducible Technique for Measuring Central Blood Pressure During Hemodynamic Perturbations Induced by Exercise
David J. Holland1,2, Julian W. Sacre2, Sarah J. McFarlane2, Jeffrey S. Coombes2 and James E. Sharman1,2
BACKGROUND Central blood pressure (BP) and markers of wave reflection (augmentation index; AIx) measured by radial tonometry have prognostic value independent from brachial BP. The measurement of the central waveform is increasingly used during altered hemodynamics, including exercise, but reliability of the test has not been reported under changed loading conditions. This study aimed to test the techniques reproducibility during major hemodynamic perturbations induced by exercise. METHODS Radial waveforms were recorded (SphygmoCor) in 28 healthy subjects (aged 53 11 years) at rest, during submaximal exercise (cycling at 50, 60, and 70% of maximal age-predicted heart rate (HR)) and immediately after maximal treadmill exercise on two occasions separated by 9 5 days. Data were compared between testing days. Waveforms were calibrated with brachial BP measured using a mercury sphygmomanometer. Pulse pressure amplification (PPAmp) was defined as the ratio of brachial to central pulse pressure. RESULTS There was very good reproducibility between visits at all exercise intensities for all waveform measures, including AIx, central pulse pressure, and PPAmp (intraclass correlations at 50% exercise were 0.93, 0.89, and 0.89, respectively; P < 0.001). The mean difference between tests at this intensity was 0 4% for AIx, 4 6mmHg for central pulse pressure, and 0.02 0.09 for PPAmp. There were no significant differences between visits for HR, PPAmp, or AIx at rest or with exercise (P > 0.05 for all). CONCLUSiONS Radial tonometry is a reproducible technique for measurement of central waveform indices during perturbations induced by exercise. It should, therefore, be suitable for use in intervention studies in which hemodynamics are altered.
Am J Hypertens 2008; 21:1100-1106 2008 American Journal of Hypertension, Ltd.

INTRODUCTiON

Although brachial (peripheral) blood pressure (BP) is an established method for assessing cardiovascular risk, it is now accepted that a major discrepancy may occur between the value of the systolic BP (SBP) recorded at the upper arm and that at the ascending aorta because of pressure wave amplification.1 The significance of this central to peripheral pressure difference has been highlighted by the finding from recent studies that central pulse pressure and augmentation index (AIx) are independently associated with cardiovascular morbidity and mortality.27 Therefore, an individuals risk related to BP may be inadequately assessed by singular consideration of brachial BP.8 Noninvasive technology (SphygmoCor; AtCor Medical, Sydney, Australia), which is now in widespread use, allows the estimation of central BP and waveform indices using radial applanation tonometry.
2School of Human Movement Studies, The University of Queensland, Brisbane, Australia. Correspondence:James E. Sharman (j.sharman@uq.edu.au) 1Department of Medicine, Princess Alexandra Hospital, Brisbane, Australia;

Several reports have shown radial tonometry to have good reproducibility at rest.911 The technique is increasingly being used under various research and medical interventions that are known to alter SBP amplification and AIx. These have included vasoactive drugs,1215 caffeine,16,17 cigarette smoking,18,19 and 2023 exercise. We have also found that monitoring the central BP response to aerobic exercise may provide additional insight relating to an individuals cardiovascular risk.24,25 Despite this, the reproducibility of central waveform indices in response to BP and heart rate (HR) changes has never been assessed. The purpose of this study was to assess the reproducibility of central waveform parameters (including AIx) and pressure amplification in normotensive subjects during hemodynamic perturbations induced by exercise.
METHODS

Received 21 April 2008; first decision 10 May 2008; accepted 19 July 2008; advanceonline publication 21 August 2008. doi:10.1038/ajh.2008.253 2008 American Journal of Hypertension, Ltd.
1100

Subjects. Thirty-one consecutive subjects (20 men) were recruited from the community by local advertisement. Subjects were excluded if they had a history of treated hypertension, type II diabetes, cardiovascular or renal disease, were aged >75 years, or were not in sinus rhythm. One woman was excluded because of poor quality waveforms caused by

OCTOBER 2008 | VOLUME 21 NUMBER 10 | 1100-1106 | AMERICAN JOURNAL OF HYPERTENSION

Exercise Waveform Reproducibility

ARTICLES
position immediately after maximal treadmill exercise within 42 23s of test termination. The testing appointment times for each individuals visits were made for the same time of day to reduce the potential confounding influence of diurnal variation. Before attendance, subjects were instructed to abstain from heavy exercise, caffeine, and smoking on the days of testing and to maintain similar food intake. Brachial BP and pressure waveform acquisition. After 5min of seated rest, brachial BP was recorded as the average of duplicate measurements using mercury sphygmomanometer, separated by 1min. During submaximal exercise, brachial BP was also measured in duplicate (consecutive measures); however, only one measure of brachial BP was recorded at peak exercise. The central (ascending aortic) pressure waveform was derived by radial applanation tonometry and application of a generalized transfer function to the radial pressure waveform using a commercial device (SphygmoCor 7.1; AtCor Medical, Sydney, Australia). The transfer function has been validated under different hemodynamic perturbations including light exercise.26,27 Resting and postmaximal exercise waveforms were recorded in duplicate by a hand-held tonometer (SPC-301; Millar Instruments, Houston, TX), whereas submaximal waveforms were acquired using a servo-controlled device (Colin CBM7000; Colin, Komaki-city, Japan). Hand-held tonometry is commonly used under resting conditions, and we also chose to use this technique during postmaximal exercise because of rapid acquisition of waveforms that is not possible with the Colin device because of the time taken to set up and calibrate the system. The servo-controlled system was used to achieve optimal waveform quality during cycle exercise in conjunction with a molded wrist brace which provided support for the hand and reduced movement artifact by restricting muscular movement. The radial pressure waveforms at rest and submaximal cycle exercise were calibrated by brachial SBP and diastolic BP (DBP) measured within 1min of waveform recordings. Postmaximal exercise waveforms were calibrated by the brachial BP obtained at peak exercise. Hemodynamic indices. On the central pressure waveform, AIx was calculated as the difference between the first (P1) and second (P2) systolic peaks (also termed the augmented pressure) as a percentage of the central pulse pressure. Pulse pressure (both brachial and central) was defined as SBP DBP. Pulse pressure amplification (PPAmp) was calculated as the ratio of the peripheral to central pulse pressure. Aortic pulse wave timing (Tr) was defined as the time between the foot of the pressure wave and the first inflection point, P1. The tonometric software has inbuilt indicators for assessing the quality of radial waveforms, including an operator index rating out of 100. Our aim was to achieve high-quality waveforms indicated by a pulse height of >100, with pulse length and diastolic variation 5; we used the operator index as a guide to achieve this. Statistics. All data are presented as mean s.d. and P < 0.05 is considered significant. The BlandAltman method was used to
1101

Table 1| Baseline characteristics of study population

Male (n = 18) Age (years) Height (m) Weight (kg) Body mass index (kg/m2) Total cholesterol (mg/dl) Triglycerides (mg/dl) Glucose (mg/dl) Smoker: current/ former/never
Data are presented as mean s.d.

Female (n = 10) 58 8 1.64 0.05 62 8 22.9 2.4 186 32 82 29 91 6 0/7/3

P value 0.07 <0.001 <0.001 0.004 0.106 0.004 0.072 <0.001

51 11 1.76 0.08 90 21 29.0 5.8 211 35 141 43 98 10 0/3/15

requent ventricular ectopic beats and two men were excluded f because of the commencement of antihypertensive therapy after completion of the study. Clinical characteristics of the studied population are presented in Table1. All subjects were apparently healthy; none were taking antihypertensive medication and all were free from a clinical history of hypertension and other cardiovascular or renal disease as defined by medical screening questionnaire completed by telephone before the first visit. One patient regularly took medication to control reflux. Nine of the ten women were postmenopausal and two of these were taking hormone replacement therapy. A fasted finger-prick sample was used to analyze blood biochemistry (Cholestech LDX; Cholestech, Hayward, CA). Four subjects had total cholesterol >240mg/dl and six separate subjects were on statin therapy. There were no current smokers, with 11 having previously smoked. The study protocol adhered to the principles of the Declaration of Helsinki and was approved by the local ethics committee. Informed written consent was obtained from each subject before testing. Study protocol. Subjects attended the clinic on two occasions: at the same time of day and separated by 1 week (mean follow-up: 9 5 days). At each visit, brachial BP and central pressure waveforms were sequentially recorded under resting and exercise conditions. Resting BP and radial waveforms were recorded after 5min of seated rest in a quiet, temperature- controlled room. Subjects were then seated on an upright, electronically braked cycle ergometer (Ergometrics 800; Ergoline, Bitz, Germany) where they were instructed to cycle at a cadence of 50rpm to complete a submaximal exercise protocol at increasing intensities. Resistance on the ergometer was adjusted to maintain a steady-state HR at 50, 60, and 70% of age-predicted maximal HR (220 age target %). Brachial and central BP measures were recorded while the subject was exercising after at least 1-min of steady-state cycling at each intensity. After a period of rest, each subject then underwent a maximal treadmill exercise test to volitional fatigue in accordance with the Bruce protocol (with continuous electrocardiogram monitoring). Brachial BP was measured at peak exercise and applanation tonometry was used to measure central BP in the supine
AMERICAN JOURNAL OF HYPERTENSION | VOLUME 21 NUMBER 10 | OCTOBER 2008

ARTICLES
Table 2| Comparison of resting and exercise hemodynamic variables between visits
Seated Baseline (n = 28) PSBP (mmHg) Visit 1 Visit 2 ICC CSBP (mmHg) Visit 1 Visit 2 ICC MAP (mmHg) Visit 1 Visit 2 ICC PPP (mmHg) Visit 1 Visit 2 ICC Tr (ms) Visit 1 Visit 2 ICC HR (bpm) Visit 1 Visit 2 ICC AIx at 75 bpm Visit 1 Visit 2 ICC 18 13 20 11 0.75 24 9 23 10 0.90 25 11 24 10 0.83 64 9 65 8 0.80 84 6 84 5 0.91 100 7 100 6 0.93 117 7 117 8 0.96 146 10 148 15 0.75 139 6 138 8 0.74 133 7 135 9 0.80 129 11 131 9 0.53 45 9 43 9 0.66 62 14* 58 15 0.85 77 14 74 16 0.67 91 18 92 17 0.87 94 10 93 12 0.74 100 12 97 12 0.85 109 11 106 12 0.88 116 14 115 14 0.87 113 12 110 15 0.74 126 19* 120 20 0.88 136 17* 131 17 0.85 145 21 144 19 0.88 124 13 122 16 0.79 142 20* 135 20 0.88 161 19* 155 20 0.84 176 24 176 23 0.92 50% Exercise (n = 20) 60% Exercise (n = 28)

Exercise Waveform Reproducibility

70% Exercise (n = 25) Maximal Exercise (n = 28) 186 25 183 23 0.92 147 16 145 16 0.89 119 14 116 14 0.86 101 20 100 18 0.80 128 8 130 9 0.65 132 18 133 19 0.98

Data are presented as mean s.d. AIx at 75 bpm is the heart ratecorrected augmentation index, which is valid to only 110 bpm (hence missing data at higher heart rates). The ICCs for all comparison were significant at P < 0.001 except Tr at 70% exercise where P = 0.003. Asterisk indicates significant difference from visit 2 (P < 0.05). AIx, augmentation index; bpm, beats/min; CSBP, central systolic blood pressure; HR, heart rate; ICC, intraclass correlation; MAP, mean arterial pressure; PPP, peripheral pulse pressure; PSBP, peripheral systolic blood pressure; Tr, aortic pulse wave timing.

assess intraobserver variability, with the difference between two measures plotted against their mean. Baseline characteristics between genders were assessed by independent t-tests or the 2-test for independence as indicated. Differences between visits and exercise intensities were assessed by repeated meas ures analysis of variance, Bonferroni corrected for multiple comparisons. Intraclass correlation coefficients were used to assess the relationship between variables. Statistical analyses were performed using SPSS software version 14.0 (SPSS, Chicago, IL).
RESULTS

Data at 50% exercise intensity were not recorded in eight subjects because their resting HRs were higher than the age- predicted values. Data from three patients who failed to reach the target HR at the 70% exercise intensity were also not presented. Three subjects had a resting brachial
1102

BP>140/90mmHg on both visits (two of these were excluded as previously described) and another two subjects presented with a BP >140/90mmHg on at least one occasion. Resting and exercise data for both visits are presented in Table2. There were no differences in HR, PPAmp, or AIx between visits (P >0.05 for all; AIx: rest P=0.241, 50% P = 0.671, 60% P = 0.562, 70% P = 0.397, maximal P = 0.062). There were statistical differences in brachial and central SBP at 50 and 60% exercise, and augmented pressure at maximal exercise (P < 0.05 for all). HR from rest to maximal exercise ranged from 47 to 160 beats/min, brachial SBP from 88 to 226mmHg, central SBP from 80 to 184mmHg, and central pulse pressure from 16 to 84mmHg. Example brachial and central waveforms from rest to maximal exercise are displayed in Figure1, whereas Bland Altman plots and Pearson correlations are presented for AIx, central pulse pressure, and PPAmp in Figures 24, respectively. Subgroup analysis on combined resting and exercise
OCTOBER 2008 | VOLUME 21 NUMBER 10 | AMERICAN JOURNAL OF HYPERTENSION

Exercise Waveform Reproducibility

ARTICLES
b
Peripheral blood pressure (mm Hg)

Peripheral blood pressure (mm Hg)

180 160 140 120 100 80

180 160 140 120 100 80

Difference between measurements

a
Legend Maximal exercise 70% Exercise 60% Exercise 50% Exercise Seated baseline

30 20 10 0 10 20 30 10 20 30 40 50 60 70 2SD Overall mean diff.: 1.8 7.1 80 90 Average central pulse pressure (mm Hg) +2SD

Central pulse pressure (mm Hg) visit 1

Figure 1 | Example (a) radial and (b) central pressure waveforms from a 48-year-old male subject at rest and increasing exercise intensities.
Difference between measurements

90 80 70 60 50 40 30 20 10 10 20 30 Overall ICC: 0.88 40 50 60 70 80 90 Central pulse pressure (mm Hg) visit 2 Visit 1 Baseline 50% 60% 70% Maximal Visit 2 Diff. 26 46 39 16 07 ICC 0.75 0.89 0.72 0.87 0.70

20 10 0 10

+2SD

2SD 20 Overall mean diff.: 0.03 6.3 30 20 10 0 10 20 30 40 Average augmentation index (%)

Augmentation index (%) visit 1

40 30 20 10 0 10 20 30 30 20 10 0 Overall ICC: 0.91 10 20 30 40 Augmentation index (%) visit 2 Visit 1 Baseline 50% 60% 70% Maximal Visit 2 Diff. ICC 0.84 0.93 0.89 0.87 0.91

34 9 31 9 44 13 41 13 51 12 47 12 57 14 56 12 57 9 57 9

Figure 3 | BlandAltman plots and intraclass correlations (ICCs) between visits one and two for central pulse pressure. Data are compared at rest, during submaximal (50, 60, and 70% of age-predicted maximal heart rate) and maximal exercise. All ICCs are significant (P < 0.001). Diff.; mean difference between visits (s.d.).

20 13 19 13 2 7 18 11 18 12 0 4 12 13 12 12 1 6 8 14 6 13 1 7 3 14 0 14 2 6

in Table 3. Across both visits there were strong correlations between resting and exercise AIx for all submaximal intensities (50% r = 0.87, 60% r = 0.87, 70% r = 0.82; P < 0.001 for all).
Maximal exercise

Figure 2 | BlandAltman plots and intraclass correlations (ICCs) between visits one and two for augmentation index. Data are compared at rest, during submaximal (50, 60, and 70% of age-predicted maximal heart rate) and maximal exercise. All ICCs are significant (P < 0.001). Diff.; mean difference between visits (s.d.).

data showed similar intraclass correlations for subjects with risk factors (total cholesterol >240mg/dl, BMI >30 kg/m2 and those with a hypertensive response to exercise) compared to those without (data not shown).
Submaximal exercise

Each increment of submaximal exercise from baseline induced a significant stepwise increase in HR, as well as brachial and central SBP (P < 0.05 for all). There was excellent reproducibility of all waveform variables during all exercise intensities (Table2, Figure2), with the possible exception of Tr at 70% exercise intensity, where the intraclass correlation was 0.53 (P = 0.003). Nonetheless, the reproducibility of submaximal waveform parameters appeared to be most stable at 60% exercise intensity, as shown in the power calculations presented
AMERICAN JOURNAL OF HYPERTENSION | VOLUME 21 NUMBER 10 | OCTOBER 2008

Four subjects (three men, one women; two with resting BP >140/90 on at least one occasion) had a hypertensive response to maximal exercise on both occasions (210/105mmHg in men or 190/105mmHg in women). As with submaximal exercise, there was good reproducibility of pressure waveform characteristics for all variables (Table2 and Figure2). The correlation between resting and maximal exercise AIx was lower than that for submaximal exercise (r = 0.52; P = 0.004). Heart rate increased significantly from 70% exercise intensity to maximal exercise (P < 0.001), with a significant rise in brachial SBP (8.4mmHg) (P = 0.021). However, central SBP appeared to plateau and increased by only 1.3mmHg (P = 0.491).
Quality control

Waveforms recorded at rest had the highest operator index and lowest variation in other parameters of quality control, indicating the highest quality (Table 4). All resting and maximal exercise waveforms had average pulse heights >100, whereas submaximal waveforms acquired by the servo-controlled system were more susceptible to variation, showing smaller average pulse heights (<100) and lower operator indexes (Table 4).
1103

ARTICLES
There were no significant differences between visitsin the quality control parameters for baseline and all exercise intensities. At rest, AIx varied by >10% between visits in three subjects. On the other hand, no subjects had >10% difference in AIx between
Difference between measurements

Exercise Waveform Reproducibility

0.50 0.25 0.00 0.25 0.50 1.00 2SD +2SD

isits during exercise at 50% intensity. For exercise at 60, 70%, v and maximal intensity, AIx was >10% between visits in two, three, and three subjects, respectively. The average difference in AIx between visits was highest at rest (6.0 4.2%, P= 0.241 between visits), but the difference was not significant compared with the average difference for all exercise intensities (range: 4.1 2.6% at 50% exercise; P = 0.075, to 5.5 3.5% at maximal exercise; P = 0.708).
DiSCUSSiON

Overall mean diff.: 0.02 0.10 1.20 1.40 1.60 1.80 2.00 Average pulse pressure amplification (ratio)

Pulse pressure amplification (ratio) visit 1

2.00 1.80 1.60 1.40 1.20 1.00 Overall ICC: 0.91 1.00 1.20 1.40 1.60 1.80 2.00 2.20 Pulse pressure amplification (ratio) visit 2 Visit 1 Baseline 50% 60% 70% Maximal 1.38 0.19 1.44 0.21 1.57 0.19 1.63 0.17 1.78 0.16 Visit 2 1.43 0.23 1.46 0.19 1.59 0.19 1.66 0.18 1.77 0.15 Diff. 0.04 0.15 0.02 0.09 0.02 0.07 0.03 0.08 0.01 0.06 ICC 0.76 0.89 0.93 0.90 0.93

Figure 4 | BlandAltman plots and intraclass correlations (ICCs) between visits one and two for pulse pressure amplification. Data are compared at rest, during submaximal (50, 60, and 70% of age-predicted maximal heart rate) and maximal exercise. All ICCs are significant (P < 0.001). Diff.; mean difference between visits (s.d.).

This study aimed to determine the reproducibility of central waveform parameters and pressure amplification during exercise as measured by radial tonometry. We found good reproducibility of waveform indices at all intensities of submaximal exercise and also after maximal exercise. Although many investigators have used radial tonometry during hemodynamic perturbations, this is the first study to test the reproducibility with changed loading conditions. We were particularly interested in the reliability of AIx, central pulse pressure, and amplification of pulse pressure, because these measures have prognostic value under resting conditions25,28,29 and are known to be significantly modified with changes in HR,30 mean arterial pressure,15 and during exercise.22 Overall, our findings support the use of radial tonometry under circumstances of large hemodynamic shifts. Wilkinson et al.9 reported good intra- and intertester reliability of resting AIx recorded in duplicate at one laboratory visit. Others have studied patients or healthy individuals on more than one occasion at rest and also found AIx to be highly reproducible despite variation in brachial BP.11,31,32 Papaioannou etal.10 also recently found AIx to remain relatively steady with repeated measures at rest, extending from 1h to several weeks. In support of this work, our own findings suggest that although

Table 3| Number of participants required to detect a significant difference between two groups for hemodynamic variables at rest and during exercise
Variable Augmentation index (%) Peripheral SBP (mmHg) Central SBP (mmHg) Pulse pressure amplification (ratio) Seated baseline 26 22 25 36 50% Exercise 22 33 38 33 60% Exercise 25 25 26 23 70% Exercise 30 29 31 19 Maximal exercise 31 26 20 12

Data have been derived from the average of two visits. Calculations are based on detecting a 10% relative difference (absolute difference for augmentation index) between two independent groups with 80% power and = 0.05. SBP, systolic blood pressure.

Table 4| Quality control parameters for resting and exercise pressure waveforms
Pulse height Baseline 50% Exercise 60% Exercise 70% Exercise Maximal exercise 143 35 76 22 86 26 92 35 178 67 Pulse height variation 41 42 43 53 73 Pulse length variation 21 21 21 11 22 Diastolic variation 31 52 52 53 52 Operator index 91 11 67 23 72 25 67 27 79 19

Data are presented as mean s.d. and represent the average of both visits.

1104

OCTOBER 2008 | VOLUME 21 NUMBER 10 | AMERICAN JOURNAL OF HYPERTENSION

Exercise Waveform Reproducibility

ARTICLES
Reproducibility of waveform indices during wide variations in HR and mean arterial pressure was in agreement with that of previous literature reported in resting individuals. These findings are of importance because radial tonometry for estimating central BP is regularly used in intervention studies as well as randomized clinical trials in which cardiovascular loading is altered.
Acknowledgment: This study was supported in part by a Centre for Clinical Research Excellence award, National Health and Medical Research Council (NHMRC), Canberra, Australia. Dr Sharman was supported by an NHRMC Australian Clinical Research Fellowship (reference 409940). Disclosure: Dr Sharman has received research equipment and funding from AtCor Medical, the manufacturers of the SphygmoCor device. The authors declared no conflict of interest.
1. ORourke MF. From theory into practice: arterial haemodynamics in clinical hypertension. J Hypertens 2002; 20:19011915. 2. Safar ME, Blacher J, Pannier B, Guerin AP, Marchais SJ, Guyonvarch PM, London GM. Central Pulse Pressure and Mortality in End-Stage Renal Disease. Hypertension 2002; 39:735738. 3. London GM, Blacher J, Pannier B, Guerin AP, Marchais SJ, Safar ME. Arterial wave reflections and survival in end-stage renal failure. Hypertension 2001; 38:434438. 4. Roman MJ, Devereux RB, Kizer JR, Lee ET, Galloway JM, Ali T, Umans JG, Howard BV. Central pressure more strongly relates to vascular disease and outcome than does brachial pressure: the Strong Heart Study. Hypertension 2007; 50:197203. 5. Chirinos JA, Zambrano JP, Chakko S, Veerani A, Schob A, Perez G, Mendez AJ. Aortic pressure augmentation predicts adverse cardiovascular events in patients with established coronary artery disease. Am J Hypertens 2005; 18(Suppl 1):A11A12. 6. Ueda H, Hayashi T, Tsumura K, Yoshimaru K, Nakayama Y, Yoshikawa J. The timing of the reflected wave in the ascending aortic pressure predicts restenosis after coronary stent placement. Hypertens Res 2004; 27:535540. 7. Williams B, Lacy PS, Thom SM, Cruickshank K, Stanton A, Collier D, Hughes AD, Thurston H, ORourke M. Differential impact of blood pressure-lowering drugs on central aortic pressure and clinical outcomes: principal results of the Conduit Artery Function Evaluation (CAFE) Study. Circulation 2006; 113:12131225. 8. Izzo J, Joseph L. Pulse contour analysis and augmentation index: its time to move beyond cuff blood pressure measurement. Am J Hypertens 2005; 18(Suppl 1):12. 9. Wilkinson IB, Fuchs SA, Jansen IM, Spratt JC, Murray GD, Cockcroft JR, Webb DJ. Reproducibility of pulse wave velocity and augmentation index measured by pulse wave analysis. J Hypertens 1998; 16:20792084. 10. Papaioannou TG, Karatzis EN, Karatzi KN, Gialafos EJ, Protogerou AD, Stamatelopoulos KS, Papamichael CM, Lekakis JP, Stefanadis CI. Hour-to-hour and week-to-week variability and reproducibility of wave reflection indices derived by aortic pulse wave analysis: implications for studies with repeated measurements. J Hypertens 2007; 25:16781686. 11. Filipovsky J, Svobodova V, Pecen L. Reproducibility of radial pulse wave analysis in healthy subjects. J Hypertens 2000; 18:10331040. 12. Kelly RP, Millasseau SC, Ritter JM, Chowienczyk PJ. Vasoactive drugs influence aortic augmentation index independently of pulse-wave velocity in healthy men. Hypertension 2001; 37:14291433. 13. Jiang XJ, ORourke MF, Jin WQ, Liu LS, Li CW, Tai PC, Zhang XC, Liu SZ. Quantification of glyceryl trinitrate effect through analysis of the synthesised ascending aortic pressure waveform. Heart 2002; 88:143148. 14. Wilkinson IB, MacCallum H, Hupperetz PC, van Thoor CJ, Cockcroft JR, Webb DJ. Changes in the derived central pressure waveform and pulse pressure in response to angiotensin ii and noradrenaline in man. J Physiol 2001; 530:541550. 15. Wilkinson IB, MacCallum H, Cockcroft JR, Webb DJ. Inhibition of basal nitric oxide synthesis increases aortic augmentation index and pulse wave velocity in vivo. Br J Clin Pharmacol 2002; 53:189192. 16. Mahmud A, Feely J. Acute effect of caffeine on arterial stiffness and aortic pressure waveform. Hypertension 2001; 38:227231. 17. Vlachopoulos C, ORourke MF. Caffeine alters arterial wave reflection: a new insight into its cardiovascular effects [abstr]. Circulation 2000; 18(Suppl II):519. 18. Vlachopoulos C, Kosmopoulou F, Panagiotakos D, Ioakeimidis N, Alexopoulos N, Pitsavos C, Stefanadis C. Smoking and caffeine have a synergistic detrimental effect on aortic stiffness and wave reflections. J Am Coll Cardiol 2004; 44: 19111917. 19. Mahmud A, Feely J. Effect of smoking on arterial stiffness and pulse pressure amplification. Hypertension 2003; 41:183187.
1105

AIx, central pulse pressure, and PPAmp are known to undergo sizeable fluctuations in response to common daily stimulus, such as postural changes33 and light exercise,22 they appear to have a remarkably stable response within an individual exposed to consistent, albeit altered, hemodynamic conditions. This argument is strengthened by the highly significant correlations between resting and exercise AIx measures. To our knowledge this is the first study to report this. Waveform quality: According to the inbuilt quality control parameters, the highest quality waveforms were those recorded using hand-held tonometry at rest. Conversely, there was more variation in waveform morphology and lower quality of submaximal (recorded by the automated Colin device) and maximal exercise waveforms (recorded by hand). Despite the lower quality, all exercise variables remained highly reproducible. Indeed, based on the intraclass correlations, BlandAltman plots, and power calculation data, the waveforms recorded at submaximal exercise intensities appeared to have equivalent, or even better, reproducibility than those recorded at rest. The study by Wilkinson et al.9 reported that 25 people would be required to detect significant differences between two independent groups (based on resting AIx measures). Our own power calculations (Table 3) for exercise waveform parameters are similar to the resting data of Wilkinson et al.9 Thus, the technique may be confidently applied in studies of people undertaking exercise. Limitations: We have tested the SphygmoCor device during different intensities of exercise and inferred that if the technique is reproducible in this context it may, therefore, be reliable under other altered conditions such as that induced by drugs. However, this is speculative because reproducibility has not been examined with drug infusions or other hemodynamic alterations other than exercise. Nevertheless, it could be argued that cardiovascular changes in response to exercise far exceed those of daily-use therapeutic drugs, because exercise causes large increases in respiratory rate and cardiac output as well as vasodilatory responses that may differ between vascular beds.34 Also, movement during exercise can create difficulty in achieving quality recordings. For these reasons, the domain of exercise is one of the most difficult to record radial tonometry, yet we still found good reproducibility of the technique. The possible confounding influence of menstrual cycle hormones on waveform morphology was not controlled for. This study limitation is unlikely to have significantly influenced findings because there was only one premenopausal woman. On another note, our subject population was normotensive and apparently healthy. Nonetheless, the study group included people with risk factors such as those undergoing treatment for hypercholesterolemia, as well as obese individuals and those with a hypertensive response to exercise. Although the presence of cardiovascular risk factors did not appear to affect the overall between-test intraclass correlations, subject heterogeneity may be a confounding factor. Summary and conclusion: This study assessed the reliability of pressure waveform parameters obtained on separate days by radial tonometry during different intensities of exercise.
AMERICAN JOURNAL OF HYPERTENSION | VOLUME 21 NUMBER 10 | OCTOBER 2008

ARTICLES
20. Murakami T. Squatting: the hemodynamic change is induced by enhanced aortic wave reflection. Am J Hypertens 2002; 15:986988. 21. Munir SM, Jiang B, Guilcher A, Brett S, Redwood S, Chowienczyk PJ. Effects of inhibition of nitric oxide synthase on the peripheral arterial waveform response to exercise. Br J Clin Pharmacol 2007; 63:778 (abstract). 22. Sharman JE, McEniery CM, Campbell R, Coombes JS, Wilkinson IB, Cockcroft JR. The effect of exercise on large artery hemodynamics in healthy young men. Eur J Clin Invest 2005; 35:738744. 23. Casey DP, Nichols WW, Braith RW. Impact of aging on central pressure wave reflection characteristics during exercise. Am J Hypertens 2008; 21:419424. 24. Sharman JE, McEniery CM, Dhakam ZR, Coombes JS, Wilkinson IB, Cockcroft JR. Pulse pressure amplification during exercise is significantly reduced with age and hypercholesterolemia. J Hypertens 2007; 25:12491254. 25. Scott JA, Coombes JS, Prins JB, Leano RL, Marwick TH, Sharman JE. Patients with type 2 diabetes have exaggerated brachial and central exercise blood pressure: relation to left ventricular relative wall thickness. Am J Hypertens 2008; 21: 715721. 26. Pauca AL, ORourke MF, Kon ND. Prospective evaluation of a method for estimating ascending aortic pressure from the radial artery pressure waveform. Hypertension 2001; 38:932937. 27. Sharman JE, Lim R, Qasem AM, Coombes JS, Burgess MI, Franco J, Garrahy P, Wilkinson IB, Marwick TH. Validation of a generalized transfer function to

Exercise Waveform Reproducibility

28.

29. 30. 31. 32. 33. 34.

noninvasively derive central blood pressure during exercise. Hypertension 2006; 47:12031208. Weber T, Auer J, ORourke MF, Kvas E, Lassnig E, Lamm G, Stark N, Rammer M, Eber B. Increased arterial wave reflections predict severe cardiovascular events in patients undergoing percutaneous coronary interventions. Eur Heart J 2005; 26:26572663. Hashimoto J, Imai Y, ORourke MF. Monitoring of antihypertensive therapy for reduction in left ventricular mass. Am J Hypertens 2007; 20:12291233. Wilkinson IB, Mohammad NH, Tyrrell S, Hall IR, Webb DJ, Paul VE, Levy T, Cockcroft JR. Heart rate dependency of pulse pressure amplification and arterial stiffness. Am J Hypertens 2002; 15:2430. Siebenhofer A, Kemp C, Sutton A, Williams B. The reproducibility of central aortic blood pressure measurements in healthy subjects using applanation tonometry and sphygmocardiography. J Hum Hypertens 1999; 13:625629. Savage MT, Ferro CJ, Pinder SJ, Tomson CR. Reproducibility of derived central arterial waveforms in patients with chronic renal failure. Clin Sci 2002; 103:5965. Kroeker EJ, Wood EH. Comparison of simultaneously recorded central and peripheral arterial pressure pulses during rest, exercise and tilted position in man. Circ Res 1955; 3:623632. Naka KK, Tweddel AC, Parthimos D, Henderson A, Goodfellow J, Frenneaux MP. Arterial distensibility: acute changes following dynamic exercise in normal subjects. Am J Physiol Heart Circ Physiol 2003; 284:970978.

1106

OCTOBER 2008 | VOLUME 21 NUMBER 10 | AMERICAN JOURNAL OF HYPERTENSION

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.