Aerobically trained (N = 15) and untrained (N = 15) Caucasian preadolescent males possessing no history of cardiovascular or cardiopulmonary disorders, and not taking any medication acted as subjects. The effect size of the HR response of adult males to the Stroop task using a similar protocol (6) was 0.75. Based on a medium-large effect size, it was predicted that sample sizes of 11–12 preadolescent males would provide a statistical power of 0.75. Trained subjects were recruited from local swimming and running clubs, and reported that they took part in a minimum of 4 h of aerobic training each week (range = 4–7 h·wk−1) and had done so for at least 32 wk before the study. The majority of trained subjects had been competitive within their respective sports for several years. Training patterns, assessed through questionnaire, were verified by the subject’s coaches. Untrained subjects were recruited from local schools and reported that they did not engage in any organized or unorganized aerobic exercise. Average hours involved in aerobic exercise training per week for trained subjects (5.48; SE = 0.25 h) was significantly greater (P < 0.05) than hours involved in aerobic training exercise per week for control subjects (0; SE = 0 h). Before testing, subjects were screened for normal cardiac function at rest. All subjects and their parents read and signed a consent form before engagement in the study in accordance with a university’s human ethics committee. Trained and untrained subjects’ age, height, and mass were as follows: 10.3 (SE = 0.4 yr), 9.5 (SE = 0.2 yr), 142.2 (SE = 2.1 cm), 137.3 (SE = 2.1 cm), 34.4 (SE = 0.4 kg), and 31.1 (SE = 1.6 kg).
Impedance cardiography, using a tetrapolar aluminum band electrode configuration (16), was used to measure reactivity (Minnesota Impedance Cardiograph Model 304bB). This technique has been validated for use with children (1). The inner two measuring electrodes were located at the base of the neck and at the level of the xiphisternal point of the thorax (16). The two outer currents were placed 3–5 cm outside of the measuring electrodes, imposing a sinusoidal current of 4 mA with a frequency of 100 kHz. The electrocardiogram (ECG) was recorded through an Amlab Physiograph (Model 1.7) using three spot electrodes. ECG, basal thoracic impedance (Zo), and the first derivative of the pulsatile impedance (dZ/dt) were measured and processed using a computer based system (COP, Microtronics Inc., Chapel Hill, NC). Ensemble averaging software was used to remove respiratory influence on the impedance cardiogram. This system was used to measure HR, cardiac output, stroke volume, and cardiac contractility. The equation used to calculate stroke volume was that developed by Kubicek et al. (16) and a fixed value of 135 ohm·cm was used for blood resistivity. Cardiac contractility was assessed by first measuring preejection period (PEP) and left ventricular ejection time (LVET), and then calculating the ratio between these two variables (PL ratio). The Heather index of cardiac contractility was derived by dividing the dZ/dt(max) amplitude (ohm·s−2) by the time before its occurrence after the EKG Q-wave (the Q-Z interval, in seconds). Blood pressure was recorded on-line, every beat, using the Ohmeda Finapres continuous blood pressure monitor (Model 2300). A cuff was attached to the third digit on the left hand that was positioned at heart level. The reliability and validity of the Ohmeda blood pressure monitor has been previously established (14,20). Total peripheral resistance (TPR) was calculated by using the equation: TPR (dyn·s·cm−5) = mean arterial pressure (MAP)/cardiac output·80. Data were collected for 30-s periods for both mental challenge tasks. Interbeat interval (IBI) was determined on a beat-by-beat basis as the difference in the time of the peak voltage between R-waves. Data were stored on hard disk and Amlab software was used to calculate IBI.
Heart period assessment.
During the baseline data collection period, two spontaneous (supine and reclined sitting) and one paced epoch (supine) of cardiovascular data were recorded and averaged for statistical analysis. Heart period as 3-min epochs of R-R data (IBI) were collected during spontaneous (2 × 3 min) and paced breathing (3 min) at a fixed rate (3 s in and 3 s out; 10 breaths·min−1). Participants practiced at the beginning of the session until their breathing was synchronized with a tape that instructed subjects when to inhale and exhale.
Calculation of vagal activity by time series analysis.
Vagal influence on the heart was assessed using time series analysis of heart period variability (HPVts; MXEDIT, Delta-Biometrics, Inc., Bethesda, MD). A band-pass filter was used to remove sources of variance below the two major oscillatory HR spectral components. The high-frequency component is synchronized with respiration and typically occurs at frequencies between 0.12 and 0.40 Hz. The other component, called Mayer waves, is termed medium frequency and is centered between 0.06 and 0.11 Hz. The slow-frequency characteristics of the Mayer waves correspond to the slow oscillations present in arterial pressure variability (12,23). The natural logarithm of the band-passed variance (ms2) was then calculated and used as high (0.12–0.40 Hz) and medium-frequency (0.06–0.11 Hz) measures of HPVts. These estimates of HPVts are reported as a linear scale ranging from 0 (minimal) to 10 (maximal;23). Owing to technical problems experienced during data collection the sample size was reduced for the HPVts measures only (trained, N = 10; untrained, N = 10).
The Stroop task.
The Stroop task (28) is a word-color information-processing task to which participants are asked to make a verbal response. Words printed in opposing colors are displayed at random on slides (e.g., the word “red” may be in green text). Participants must verbally respond to the color in which the text is written and not the word it spells. Words were displayed on a screen using a slide projector at a rate of one slide every second for 2 min, during which the absolute number of errors were recorded.
The Tetris computer task.
A PC-based version of the computer game Tetris (version 1.5 Lite) was used as a second mental challenge. The game, which demands spatial awareness and rapid decision-making, requires the participant to fit randomly shaped falling blocks into spaces left by the preceding blocks. As the game ensues, it becomes progressively harder. The game was run at speed 6 with a clear screen at the game onset. No subject reported prior experience with the Tetris computer task. All subjects finished the game between 90 s and 115 s.
Pretest anxiety was assessed using the State Trait Anxiety Inventory (STAI;26) immediately before testing.
Subjects were requested to abstain from ingesting stimulants (e.g., smoking, or drinking coffee or other caffeinated substances) or suppressants (e.g., alcohol) for at least 4 h before testing and from eating for at least 2 h before the study but not to refrain from eating for more than 4 h. No subject was on medication of any kind. Before the test onset participants (with parental assistance) were asked to complete an informed consent form. Ambient room temperature was maintained between 22° and 24°C during all testing sessions.
Baseline data was collected over three, 3-min epochs. The first epoch involved participants breathing spontaneously and lying relaxed in a supine position. The second epoch involved subjects sitting in a reclined position (legs horizontal on a plinth and back inclined at 30°) while breathing spontaneously. During the final epoch, subjects breathed in synchronization with instructions from a tape timed at 10 breaths·min−1 while sitting in a reclined position. After baseline data collection, all participants were exposed to the Stroop task for a 2-min bout followed by a 3-min period of recovery; this was repeated once. Participants’ HR were then allowed to return to their resting values before they completed up to 115 s of the Tetris task followed by a 3-min recovery period. This task was also repeated.
Initial data processing of impedance cardiograms used ensemble averaging to filter respiratory influence. Impedance waves were then manually edited using the edit mode of the COP software. Vagal influence on the heart was averaged over a 3-min period to satisfy the assumptions underlying spectral analysis of heart period variability.
The SPSS statistical package for the PC (SPSS Inc., Chicago, IL) was used to perform a repeated measures mixed analysis of variance. The between factor was group (trained and untrained), whereas the repeated factor was time for each of the variables. Absolute scores refer to the mean values of each variable, whereas change scores refer to the mean values minus the baseline values during paced breathing while sitting reclined. Change score data were examined by a two-factor ANOVA. As no significant difference existed between the two trials of Stroop and the two trials of Tetris, the data were collapsed and reported as one trial for Stroop and one trial for Tetris. Analysis was conducted on data averaged every 60 s during the Stroop and every 90 s during recovery, and every 90 s during and recovering form Tetris.
There were no significant differences observed for trained (mean = 33.1 ± 2.3) and untrained (mean = 32.5 ± 2.1) subjects’ state anxiety. During the Stroop task, untrained subjects (mean = 10.5 ± 1.9) made significantly (P < 0.05) more errors than trained subjects (mean = 5.5 ± 1.0), whereas during Tetris there was no differences in performance.
Cardiovascular measures during rest.
During paced breathing while sitting reclined, trained compared with untrained subjects demonstrated significantly lower HR (P < 0.05). During paced breathing, trained (mean = 8.09 ± 0.3 ms2) compared with untrained (mean = 7.37±0.2 ms2) subjects possessed significantly greater high-frequency HPVts (P < 0.05). Trained (mean = 3.97 ± 0.2 ms2) and untrained (mean = 3.34 ± 0.3 ms2) subjects possessed similar resting medium-frequency HPVts. During unpaced breathing while sitting reclined, trained compared with untrained subjects again demonstrated significantly lower HR (Table 1). During unpaced breathing, trained (mean = 9.02 ± 0.2 ms2) compared with untrained (mean = 8.03 ± 0.2 ms2) subjects also possessed significantly greater (P < 0.05) high-frequency HPVts (see Fig. 2). Trained (mean = 4.24 ± 0.2 ms2) and untrained (mean = 3.53 ± 0.3 ms2) subjects possessed similar resting medium-frequency HPVts. During rest, trained compared with untrained subjects demonstrated significantly greater stroke volume (P < 0.05), and significantly lower PL ratio and rate pressure product (Table 1). All other variables were similar (see Table 1).
Cardiovascular response to the two mental challenge tasks.
For both groups, combined exposure to the Stroop task resulted in significant increases in HR, TPR, systolic and diastolic blood pressure, and rate pressure product (P < 0.05) and significant decreases in stroke volume (P < 0.05;Table 1). For both groups, combined exposure to Tetris resulted in significant increases in diastolic blood pressure (P < 0.05;Table 2).
Cardiovascular response during and recovering from the Stroop task.
The means and standard errors for all cardiovascular variables at rest (unpaced breathing while sitting reclined) and during and after Stroop for all variables are shown in Table 1. For absolute HR during and recovering from Stroop, the untrained group demonstrated significantly higher absolute values compared with the trained (P < 0.05;Fig. 1). During Stroop, the trained compared with the untrained had a significantly greater increase in HR change scores (P < 0.05;Table 2). Trained (mean = 7.0 ± 0.3 ms2) compared with untrained (mean = 6.73 ± 0.2 ms2) subjects demonstrated similar absolute high-frequency HPVts, during and recovering (trained: mean = 7.66 ± 0.3 ms2; untrained: mean = 7.64 ± 0.4 ms2) from the Stroop task (Fig. 2). For the trained (mean = −2.02 ± 0.4 ms2) compared with the untrained (mean = −1.30 ± 0.3 ms2), there was also a significantly greater decrease (P < 0.05) in high-frequency HPVts change scores during Stroop. For the trained (mean = −1.36 ± 0.4 ms2) compared with the untrained (mean = 0.39 ± 0.3 ms2), there was also a significantly greater decrease (P < 0.05) in high-frequency HPVts change scores recovering from Stroop. Trained (mean = 3.51 ± 0.2 ms2) compared with untrained (mean = 3.27 ± 0.3 ms2) subjects demonstrated similar absolute medium-frequency HPVts during and recovering from the Stroop task (trained: mean = 3.38 ± 0.2 ms2; untrained: mean = 3.51 ± 0.2 ms2). All other variables were similar.
Cardiovascular response during and recovering from the Tetris task.
The means and standard errors for all cardiovascular variables at rest (unpaced breathing while sitting reclined) and during and recovering from Tetris are shown in Table 3. For absolute HR during and recovering from Tetris, the untrained group demonstrated significantly higher absolute values compared with the trained (P < 0.05;Table 2). Trained (mean = 7.64 ± 0.4 ms2) compared with untrained (mean = 7.58 ± 0.3 ms2) subjects demonstrated similar absolute high-frequency HPVts, during and recovering (trained: mean = 7.88 ± 0.5 ms2; untrained: mean = 7.08 ± 0.5 ms2) from Tetris. Trained (mean = 3.77 ± 0.2 ms2) compared with untrained (mean = 3.74 ± 0.2 ms2) subjects demonstrated similar absolute medium-frequency HPVts during and recovering from Tetris (trained: mean = 3.63 ± 0.3 ms2; untrained: mean = 3.53 ± 0.2 ms2). During Tetris, there were no significant differences in absolute or change scores for any other variable.
The primary purpose of this study was to compare the cardiovascular response of aerobically trained and untrained preadolescent subjects to mental challenge. For both tasks, absolute HR of the trained subjects at rest, during, and recovering from mental challenge were significantly lower than that of untrained subjects. Also HR and vagal activity change from baseline to Stroop was significantly greater in the trained subjects.
That maximal oxygen uptake was not assessed in this study is a limitation. The activity patterns of the trained subjects, however, indicate that they were all involved in regular aerobic exercise and competing in regular competition. The resting measures of these physically active preadolescent subjects also demonstrate exercise-training characteristics. For example, the larger resting stroke volumes (Table 1) are typical of aerobically trained adults (13). The lower resting HR and greater vagal activity of the trained compared with the untrained subjects (Table 1) further extends the relationship between aerobic training and resting HR that has consistently been found in adult trained males (5). However, bradycardia (HR < 60 beats·min−1) was not evident for trained subjects.
These results support prior research showing that low resting HR (bradycardia) are not typically found in aerobically trained preadolescent children (24,29). Although it is unclear why children do not acquire training-induced bradycardia, it is possible that because they sweat less during exercise, (24) they are unable to expand blood plasma volume after exercise. When adults exercise, heat is primarily lost through the evaporation of sweat, which originates from blood plasma. When fluid is replaced after exercise, blood plasma expands beyond its original volume, causing an increase in stroke volume and the low resting HR that are typical of aerobically trained athletes (9). Also these preadolescent boys’ smaller body size may contribute to their higher HR (25). Even though the trained subjects’ resting HR (67 beats·min−1) were higher than that usually recorded in trained adult male athletes (typically 40–55 beats·min−1), their HR was still significantly lower than that of the untrained subjects (80 beats·min−1). Thus, trained preadolescent subjects “bradycardia” may be some 20 beats·min−1 higher than their trained adult counterparts.
Both groups exhibited a significant cardiovascular response to the Stroop task. The HR increase to Stroop (Fig. 1) was comparable to that of adult males in prior Stroop reactivity studies (6,7). However, the pattern of cardiovascular response of the boys to the Stroop differed to that of adults. Previous research with adults has shown that the Stroop typically brings about increases in both beta- and alpha-adrenergic activity resulting in increased HR, cardiac contractility, systolic and diastolic blood pressure, and TPR (7). A decrease in vagal activity when exposed to the Stroop has also been shown to occur (5). In contrast, subjects in the present study failed to demonstrate increases in cardiac contractility during the Stroop or Tetris as the PEP/LVET ratio and the Heather Index responses remained similar to baseline levels (Tables 1 and 2). Similar to adults, subjects recorded significant increases in systolic and diastolic blood pressure, which was brought about by large increases in TPR. Subjects’ HR increase was primarily a result of vagal withdrawal. The lack of beta-adrenergic activity and heightened alpha-adrenergic and vagal withdrawal may reflect a unique cardiovascular response pattern to this form of mental challenge. This pattern extends Berntson and colleagues’ (4,8) taxonomy for autonomic response to mental challenge established in adults.
Less HR reactivity was exhibited to the computer game than that exhibited to the Stroop. The Tetris task is a game requiring spatial awareness and the ability to make quick decisions. Although the Tetris game is similar to computer games used by preadolescent children, it may be less visually and audibly stimulating than modern computer games, and thus participants may have found this task less arousing. Less reactivity to the Tetris may also have been a result of behavioral habituation because it was presented after the Stroop task.
The trained group demonstrated significantly lower absolute HR and greater HR change during and recovering from the Stroop task (Fig. 1). The lower absolute HR responses reflect the differences in resting HR between the trained and untrained groups, whereas the larger HR change suggests greater training-induced HR reactivity. The vagal activity results indicate that withdrawal of cardiac parasympathetic influence was greater for trained subjects and suggest greater training-induced parasympathetic reactivity to mental challenge. This lower absolute HR response and greater decrease in vagal activity to mental challenge has been found in prior research that has compared trained and untrained adult males (5). The medium-frequency HPVts levels of both groups did not change significantly during or recovering from stressors. This is surprising as it has been suggested that these frequencies are mediated by sympathetic mechanisms (27). However, Pomeranz et al. (22) suggest that the medium frequencies, while supine, depend on parasympathetic activity, whereas during standing the medium frequencies are mediated by sympathetic mechanisms. As the subjects performed mental challenge in a reclined sitting position, which was more supine than standing, it is feasible that the medium frequencies results of the present study represent mainly parasympathetic influences.
The trained subjects’ greater vagal reactivity could be a function of their significantly higher resting vagal activity. Thus, greater vagal withdrawal could have occurred because the trained subjects had a greater physiological range to respond. However, the trained subjects’ greater vagal reactivity could also reflect a difference in autonomic balance. The trained subjects’ HR response to mental challenge was increased primarily by parasympathetic influence to the heart, whereas the untrained subjects’ HR response may have been increased more by sympathetic innervation. Using an animal model, Manuck et al. (17) have shown that stress-induced catecholamine excretion caused increased coronary arteriosclerosis. If individuals who display greater vagal reactivity to mental challenge also respond with lower blood catecholamine levels, then regular exercise may have a cardio-protective effect by reducing catecholamine excretion to daily stressors. Studies measuring both cardiac vagal activity and catecholamine excretion to mental challenge are needed to verify this relationship.
Although the trained compared with untrained subjects demonstrated greater HR and vagal reactivity to the Stroop, they made fewer errors. Thus, it is unlikely that the trained subjects’ greater reactivity was due to differences in performance. It is feasible, however, that trained subjects may have possessed a more competitive attitude toward the tasks.
It is possible that trained subjects’ cardiovascular response to both aerobic performance and mental challenge was influenced by genetic endowment. For example, it is possible that the trained boys possessed low resting HR, large stroke volumes, and high resting vagal activity before they started aerobic training. Previous research has shown that possessing a genetically low resting HR influences reactivity to Stroop challenge (5). Consequently, inherited cardiovascular characteristics may have an influence on cardiovascular reactivity to mental challenge. The cross-sectional nature of this study did not control for potential subject differences such as genetic endowment and personality differences.
During baseline, it was noted that the vagal activity of both trained and untrained subjects was significantly greater during unpaced compared with paced breathing; however, this difference has not been noted with adults (5). An explanation may be that because children’s ventilatory efficiency is less than that of adults, children may ventilate at a greater frequency. During paced breathing, however, subjects were instructed to breathe at a rate of 10 breaths·min−1, which may be a slower rate than normal for children. Thus, these findings indicate that breathing rate has an important influence upon vagal influence on the heart of preadolescent males.
In conclusion, it was found that trained compared with untrained subjects exhibited significantly lower absolute HR levels but greater relative HR change during mental challenge. Trained compared with untrained subjects possessed greater absolute resting vagal activity but recorded greater vagal withdrawal during the Stroop task. Cardiac contractility of both groups failed to increase during either stressor.
1. Alpert, B. S., K. R. Bloom, D. Gilday, and P. M. Olley. The comparison between non-invasive and invasive methods of stroke volume determination in children. Am. Heart J. 98: 763–765, 1979.
2. Beere, P. A., S. Glagov, and C. K. Zarins. Retarding effect of lowered heart rate on coronary atherosclerosis. Science 226: 180–182, 1984.
3. Berenson, G. S., G. C. Frank, S. M. Hunter, S. R. Srinivasan, A. W. Voors, and L. S. Webber. Cardiovascular risk factors in children. Am. J. Diseases Children 136: 855–862, 1982.
4. Berntson, G. G., J. T. Cacioppo, P. F. Binkley, B. N. Uchino, K. S. Quigley, and F. Fieldstone. Psychological stress and cardiac response in autonomic space as revealed by pharmacological blockade. Psychophysiology 31: 599–608, 1994.
5. Boutcher, S. H., F. W. Nugent, P. F. Mclaren, and A. L. Weltman. Heart period variability of trained and untrained men at rest and during mental challenge. Psychophysiology 35: 16–22, 1998.
6. Boutcher, S. H., F. W. Nugent, and A. L. Weltman. Cardiac response to psychological stressors of individuals possessing low resting heart rate. Behav. Med. 21: 42–46, 1995.
7. Boutcher, S. H., and D. Stocker. Cardiovascular response of young and older males to mental challenge. J. Gerontol. Psychol. 51: 261–267, 1996.
8. Cacioppo, J. T., B. N. Uchino, and G. G. Berntson. Individual differences in the autonomic origins of heart rate reactivity: the psychometrics of respiratory sinus arrhythmia and preejection period. Psychophysiology 31: 412–419, 1994.
9. Convertino, V. A. Blood volume: its adaptation to endurance training. Med. Sci. Sports Exerc. 23: 1338–1348, 1991.
10. Costill, D. L. Inside Running: Basics of Sports Physiology. Indianapolis: Benchmark Press, 1986, p. 19.
11. Falkner, B., G. Onesti, and E. T. Angelakos. The effect of salt loading on the cardiovascular response to stress in adolescents. Hypertension 3: (Suppl.) II195–II199, 1981.
12. Furlan, R., S. Piazza, S. Dell’Orto, et al. Early and late effects of exercise and athletic training on neural mechanisms controlling the heart. Cardiovasc. Res. 27: 482–488, 1993.
13. Gledhill, N., D. Cox, and R. Jamnik. Endurance athletes’ stroke volume does not plateau: major advantage is diastolic function. Med. Sci. Sports Exerc. 26: 1116–1121, 1994.
14. Gerin, W., C. Pieper, L. Marchese, and T. G. Pickering. The reliability of cardiovascular change scores: a comparison of intermittent vs. continuous methods. J. Psychosom. Res. 37: 493–501, 1993.
15. Jiang, W., W. Bab Jiang, M. Babyak, et al. Mental stress-induced myocardial ischemia and cardiac events. JAMA 275: 1651–1656, 1996.
16. Kubicek, W. G., F. J. Kottke, M. M. Ramos, and D. A. Witsoe. Minnesota impedance cardiography. Biomed. Eng. 9: 410–416, 1974.
17. Manuck, S. B., J. R. Kaplan, and T. B. Clarkson. Behaviorally induced heart rate reactivity and atherosclerosis in Cynomolgus monkeys. Psychosom. Med. 45: 95–108, 1983.
18. Matthews, K. A., K. L. Woodall, and C. M. Stoney. Changes in and stability of cardiovascular responses to behavioural stress: results from a four-year longitudinal study of children. Child Dev. 61: 1134–1144, 1990.
19. Murphy, J. K., B. S. Alpert, E. S. Willey, and G. W. Somes. Cardiovascular reactivity to psychological stress in healthy children. Psychophysiology 25: 144–152, 1988.
20. Parati, G., R. Casadei, A. Groppelli, M. D. Rienzo, and G. Mancia. Comparison of finger and intra-arterial blood pressure monitoring at rest and during laboratory testing. Hypertension 13: 647–655, 1989.
21. Parker, F., J. Croft, J. Cresanta, et al. The association between cardiovascular response tasks and future blood pressure levels in children: Bogalusa Heart Study. Am. Heart J. 113: 1174–1179, 1987.
22. Pomeranz, B., R. J. B. Macaulay, and M. A. Caudill. Assessment of autonomic function in humans by heart rate spectral analysis. Am. J. Physiol. 248: H151–H153, 1985.
23. Porges, S. Cardiac vagal tone: a physiological index of stress. Neurosci. Biobehav. Rev. 9: 225–233, 1995.
24. Rowland, T. W. Developmental Exercise Physiology. Champaign, IL: Human Kinetics, 1986, pp. 43–48.
25. Silvetti, M. S., F. Drago, and P. Ragonese. Heart rate variability in healthy children and adolescents is partially related to age and gender. Int. J. Cardiol. 81: 169–174, 2001.
26. Speilberger, C. D., R. L. Gorsuch, and R. E. Lushene. Manual for the State Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press, 1972, pp. 9–23.
27. Stein, P. K., and R. E. Kleiger. Insights from the study of heart rate variability. Annu. Rev. Med. 50: 249–261, 1999.
28. Stroop, J. P. Studies of interference in serial verbal reactions. J. Exp. Psychol. 18: 643–662, 1935.
29. Weymans, M. L., T. M. Reybrouck, H. J. Stijns, and J. Knops. Influence of habitual levels of physical activity on the cardiorespiratory endurance capacity of children. In:Children and Exercise,
Vol. IV. Champaign, IL: Human Kinetics, 1989, pp. 149–156.
30. Williams, J. E., C. C. Paton, I. C. Siegler, M. L. Eigenbrodt, F. J. Nieto, and H. A. Tyroler. Anger proneness predicts coronary heart disease risk: prospective analysis from the atherosclerosis in communities (ARIC) study. Circulation 101: 2034–2039, 2000.
31. Wood, D. L., S. G. Sheps, L. R. Elveback, and A. Schirger. Cold pressor tests as a predictor of hypertension. Hypertension 6: 301–306, 1984.