Exercise Training In Healthy Type A

Exercise Training in Healthy Type A Middle-Aged Men: Effects on Behavioral and Cardiovascular Responses JAMES A. BLUMENTHAL, P H D , CHARLES F. EMERY, P H D , MARGARET A. WALSH, MS, DAVID R. COX, PHD, CYNTHIA M. KUHN, PHD, REDFORD B. WILLIAMS, MD, AND R. SANDERS WILLIAMS, MD Thirty-six healthy Type A men (x = 44.4 years) were randomly assigned to either an aerobic exercise training group or a strength and flexibility training group. Subjects completed a comprehensive psychological assessment battery before and after the exercise programs consisting of behavioral, psychometric, and psychophysiological testing. The behavioral assessment consisted of repeated Type A interviews that were videotaped for subsequent component analyses. The psychometric testing included two self-report questionnaires to assess Type A behavior. The psychophysiological test consisted of a standard behavioral challenge, a mental arithmetic task, performed while cardiovascular responses were monitored. Aerobic exercise (AE) training consisted of 12 weeks of continuous walking or jogging at an intensity of at least 70% of subjects' initial maximal oxygen consumption (VO2max) as determined by an initial treadmill test. Strength and flexibility (SF) training consisted of 12 weeks of circuit Nautilus training with no aerobic exercise. After 12 weeks of exercise, the AE group increased their V02max by 15%, while the SF group did not change. Both groups experienced decreases in overt behavioral manifestations of the Type A behavior pattern and self-reported Type A traits. However, the AE group showed an attenuation of heart rate, systolic and diastolic blood pressure, and estimated myocardial oxygen consumption (MVO2) during the task and had lower blood pressure, heart rate, and (MV02) during recovery. In contrast, the SF group showed a significant reduction only in DBP during the task, which was likely due to habituation. These results support the use of aerobic exercise as a method for reducing cardiovascular risk among healthy Type A men.

INTRODUCTION There has been growing acceptance of the role of behavioral factors in the etiology, prevention, and treatment of coronary heart disease (CHD) (1). In particular, the Type A behavior pattern has been widely studied, and may be a risk

iongitudinai for C H D (2) In a study of almost 3,200 healthy men, Type

factor

A s w e r e s h o w n to exhibit twice the rigk of

CHD

of

counterparts

easy.going Type B (3). Although recent stud-

their

ies h a v e not f o u n d

Type A behavior

to

From the Departments of Psychiatry 0A.B.,C.F.E., M.A.W., D.R.C., R.B.W.), Medicine (J.A.B., R.B.W., R.S.W.), and Pharmacology (C.M.K.), Duke University Medical Center, Durham, North Carolina. Address reprint requests to: James A. Blumenthal, Ph.D., Department of Psychiatry, Duke University Medical Center, Box 3119. Durham, NC 27710. Received for publication June 26, 1987; revision received January 26,1988.

i n c r e a S ed risk of CHD i n p a t i e n t s w i t h established CHD or severe risk factors (4), Type A behavior is still regarded as an important behavioral risk factor, particularly among healthy a numm i ddle-aged men. Consequently, M . , , / , b, e r o f , , ° , behavioral treatments have been proposed to modify the behavior of Type A individuals. Treatment Strategies have , , , . . .° . . . included anxiety management training (5), cognitive-behavioral stress manage-

418

Psychosomatic Medicine 50:418^133 (1988)

Copyright © 1988 by the American Psychosomatic Society, Inc. Published by Elsevier Science Publishing Co., Inc. 52 Vanderbilt Avenue, New York, NY 10017

be associated with

0033-3174/88/$3 50

EXERCISE AND TYPE A BEHAVIOR

ment (6-9), and short-term psychodynamic psychotherapy (8). To date, results have been equivocal. On the positive side, Friedman and colleagues recently demonstrated that a comprehensive social learning-based behavior modification program reduced overt manifestations of Type A behavior in cardiac patients (10). Furthermore, reductions in Type A behaviors were associated both with reduced rates of recurrent nonfatal CHD events (11) and with lowered mortality rates in low-risk cardiac patients (12). A similar behavior modification program was adapted for healthy subjects and was successful in reducing Type A behaviors among Army personnel (13). However, other intervention studies have failed to find changes in either self-reported Type A behaviors (5, 8) or biologic factors known to be related to increased risk of CHD, such as elevated serum cholesterol or blood pressure (6, 7). It has been suggested that Type As may be more vulnerable to CHD because of their propensity to exhibit exaggerated cardiovascular and neuroendocrine responses to everyday challenges (15). Although numerous sutides have shown that Type As display this pattern of enhanced reactivity, few studies have monitored changes in cardiovascular reactivity as a result of treatment. Those that have assessed reactivity have not demonstrated that the intervention significantly modified cardiovascular responses (7, 9, 10). The studies cited above also have a number of methodological and practical limitations. For example, some studies did not include a no-treatment control group (8) or relied solely on self-report measures of Type A behavior such as the Jenkins Activity Survey (JAS) (5, 8), the Framingham Type A Scale (7), or the Bortner Type A Rating Scale (6). Self-report measures of Type A behavior have been widely

criticized because of their imprecision relative to the direct behavioral assessment used in the Type A Structured Interview (SI) (14). Finally, most studies have practical limitations, since behavioral interventions are often expensive and timeconsuming. Type A modification typically involves groups of eight to 12 people and may require at least 1 year of treatment (10). Since the prevalence of Type A behavior is at least 50% in the general population (3) and may reach 75% in highrisk populations, such as the Multiple Risk Factor Intervention Trial Group (16), Type A behavior modification may not be costeffective. Aerobic exercise has been proposed as a relatively simple and inexpensive intervention that also may be effective in reducing the risk of CHD among healthy Type A individuals. Laboratory studies have shown that exercise training reduces standard risk factors for CHD (17), and epidemiological studies have reported that active individuals have lower rates of CHD morbidity and mortality than those who are sedentary (18, 19). In a previous report from our laboratory (20), we observed that a 10-week program of aerobic exercise training was successful in reducing blood pressure and body weight and in increasing high density lipoprotein (HDL) cholesterol in a sample of healthy adults. Moreover, we reported that individuals who were classified as Type A, on the basis of their scores on the JAS, obtained lower Type A scores following the completion of the exercise training. This study was limited, however, by reliance on the JAS as the measure of Type A behavior and by the lack of a no-exercise control group. More recently, Roskies and colleagues (10, 21) used the SI as a measure of Type A behavior but failed to observe any reduction in overt Type A be-

Psychosomatic Medicine 50:418^33 (1988)

419

J. A. BLUMENTHALetal.

haviors or psychophysiological reactivity in Type A men who participated in an aerobic exercise program. However, the Roskies et al. findings are inconclusive since their method of measuring aerobic fitness was based upon estimates from submaximal exercise testing. Changes in heart rates at submaximal workloads may reflect greater efficiency but not necessarily improved maximal oxygen consumption (V02max) (22). Lake and associates (23) found that unfit Type As exhibited greater cardiovascular reactivity than did the fit Type As and Type Bs of both fitness levels. Since their experimental design was cross-sectional (i.e., fitness was not experimentally modified), the results could be interpreted as providing only partial support for a direct causal relationship between fitness and reactivity. The present study was designed to improve upon the methodological limitations of previous research in several ways by: a) using a randomized longitudinal design; b) precisely measuring changes in aerobic fitness with direct measurement of oxygen consumption; and c) evaluating changes in overt behavioral characteristics and cardiovascular responses, as well as changes in self-reported Type A behav-

METHODS

Subjects Thirty-six men who were initially rated as Type A by the SI served as subjects for this study. All subjects were judged to be free of clinical manifestations of CHD by medical history, physical examination, and graded exercise treadmill testing under continuous electrocardiographic monitoring. The subjects were all employed and had at least a high school education. They ranged in age from 31 to 59 years (x = 44.4 years).

420

Procedure Subjects were randomly assigned to one of two exercise programs. One group (N = 18) participated in aerobic exercise training (AE), while a comparison group (N = 18) participated in strength and flexibility (SF) training. Subjects in the AE group attended three supervised exercise sessions per week for 12 consecutive weeks. Aerobic exercise sessions began with a 1015-min warm-up period, followed by 35 min of continuous walking, jogging, and stair climbing at an intensity of at least 70% of subjects' initial VOzmax. As conditioning occurred, subjects systematically increased their jogging distance and speed to maintain their heart rate within the training range for the full duration of the 35-min training session. Subjects were instructed to monitor their heart rate by taking their radial pulses. Subjects also wore a heart rate monitor (Exersentry) twice a week during the first 2 weeks of the training program and once per week thereafter to ensure that their heart rates were kept within the prescribed training ranges. Subjects in the SF group participated in 20 min offlexibility exercises, followed by 30 min of circuit Nautilus training. Subjects participated in these supervised group sessions two to three times per week for 12 weeks. The SF group provided a control for the effects of social stimulation and attention from trainers without producing cardiovascular training effects (24). Subjects in the SF group were requested not to engage in any aerobic exercises during their participation in this study. All participants underwent comprehensive physiological evaluations as well as assessments of Type A behavior. Evaluations were conducted prior to the beginning of exercise training and after 3 months of exercise conditioning. Physiological Measures. Blood pressure was obtained by standard cuff sphygmomanometry while the subject was in the sitting position. Plasma triglycerides, total serum cholesterol, and HDL cholesterol were determined from blood samples drawn between 0700 and 0900 hours following a 14-hr fast. Blood was withdrawn by a one-syringe 15-cc vacutainer tube, anticoagulated with 3.5% sodium citrate, and centrifuged at 6,000 x g for 15 min prior to analysis by a commercial laboratory. In order that cardiorespiratory fitness could be measured, subjects underwent treadmill exercise testing using a modification of the Balke protocol (25). Graded exercise treadmill testing began at 2.0 mph and 0% grade. Speed and grade were system-

Psychosomatic Medicine 50:418-433 (1988)

EXERCISE A N D TYPE A BEHAVIOR atically increased at the rate of 1 Met (3.5 ml/kg/min 02) per minute—i.e., minute 2: 2.5 mph, 0% grade; minute 3: 2.5 mph, 2.0% grade; minute 4: 3 mph, 2.5% grade; minute 5: 3 mph, 5% grade, etc. (26). Fasting subjects exercised until exhaustion (mean Borg (27) rating = 19.6) during continuous electrocardiographic monitoring.1 Heart rate was obtained using a Hewlett Packard 4685-A digital cardiotachometer. Expired air was collected by a facemask for quantification of oxygen consumption (VO2), expired ventilation, and respiratory exchange ratio at 15-sec intervals with the use of an MMC Horizon System 2 Beckman Metabolic Measurement Cart. An identical protocol was used for retesting subjects after 3 months of participation in the exercise program. Assessment of Type A Behavior. Three independent assessment procedures were employed to evaluate different behavioral manifestations of the Type A behavior pattern. These procedures included direct behavioral observations, self-reported attitudes and habits, and measurements of cardiovascular responses to a standard behavioral challenge. Behavioral Assessment. All subjects underwent repeated versions of the standard Type A SI in counterbalanced order. All Sis were videotaped for subsequent blind review by a trained rater. Component ratings were performed to assess behavioral subcomponents that were previously found to distinguish Type As from Type Bs and were also judged to be stable over time (28). These components included speed and volume of speech, uneven speech, response latency, plosive words, number of interruptions, and potential for hostility during the SI. Both the interviewer and rater were unaware of the subject's group assignment. Videotapes were reviewed following the completion of all (i.e., Times 1 and 2) data collection. Psychometric Assessment. Subjects completed two psychometric instruments to assess self-reported Type A behaviors before and after the 3-month exercise program. The JAS (29) yields a global Type A score and three factor analytically derived subscales:

'One subject experienced a >lmm ST-segment depression on his ECG. A subsequent multigated angiogram (MUGA) was performed with no evidence of myocardial ischemia or wall motion abnormalities. Consequently he was included in the study.

Psychosomatic Medicine 50:418-433 (1988)

speed and impatience, job involvement, and hard driving and competitive. Subjects also completed the Type A Self-Rating Inventory (TASRI) (30), a 28-item questionnaire that yields a single measure of global Type A. PsychophysiologicoJ Assessment. All subjects were tested individually in a sound-attenuated, temperature-controlled (80°F) chamber. After a 45 min period in which subjects were requested to "sit quietly without falling asleep," a 15-min baseline period was used for collection of resting cardiovascular measures. Following this procedure, subjects underwent a mental arithmetic task (MAT) consisting of a series of serial additions. Subjects first added the digits of a three-digit number (e.g., 111 = 1 + 1 + 1 = 3), which was then added to the sum of the original number (e.g., I l l + 3 = 114) to form a new three-digit number. Subjects repeated this procedure for a total of 15 min (e.g., saying aloud "111, 114, 120," etc.). Three blocks of 5-min serial additions were conducted with a 1-min break between each block. Subjects were encouraged to "go a little faster" 2 1/2 min into each block. Each of the two MAT testing sessions (initial and after 3 months) was divided into resting (15 min preceding the MAT), task (15 min of serial additions), and recovery (15 min following completion of the MAT) phases. Heart rate and blood pressure were monitored at 1-min intervals with a Dinamap Monitor No. 845. Subjects were tested at the same time each afternoon and were asked to refrain from coffee, tea, cola, chocolate, vanilla extract, and exercise during the 24 hours preceding the procedure. Data Analysis. Data initially were analyzed by a 2 (Group) x 2 (Time) repeated-measures multivariate analysis of variance (MANOVA). Group (AE or SF) served as a between-subjects variable and time (pre-exercise/Time 1 and postexercise/Time 2) served as a within-subjects variable. Separate MANOVAs were performed for the physiological variables and for the three independent response domains of Type A behavior: behavioral, psychometric, and psychophysiologic data. In order that changes in levels of cardiovascular response could be evaluated, the psychophysiologic data were analyzed by a 2 (Group) x 2 (Time) x 3 (Phase) MANOVA followed by an analysis of covariance (ANCOVA), with resting values for each session serving as covariates. In addition, change scores from rest to task phase were determined to assess cardiovascular reactivity. Finally, recovery change scores (i.e., rest-recovery) were ex-

421

J. A. BLUMENTHAL et al. amined to assess the rate of recovery from the behavioral challenge.

peated-measures analyses of variance (ANOVAs) were performed for each variable separately. Examination of the blood lipid values shown in Table 2 indicates RESULTS that neither form of exercise altered levels of total plasma cholesterol, but low denThirty-one out of 36 subjects (86%) sity lipoprotein (LDL) cholesterol tended completed the 3-month program. Three to increase between Time 1 and Time 2. subjects from the SF group and two sub- There was a significant Time x Group injects from the AE group dropped out be- teraction for triglycerides and HDL chofore completing their Time 2 assessments lesterol. Triglycerides decreased and HDL and were omitted from subsequent data cholesterol increased significantly in the analysis. Compliance with the exercise AE group, but the triglyceride and HDL programs was excellent. All but two sub- levels in the SF group did not change. jects attended at least 92% of the aerobic exercise sessions, and all but three subBehavioral Observations jects completed 100% of the strength training sessions. Type A component ratings for Time 1 and Time 2 are presented in Table 3. A MANOVA for the Type A component ratPhysiological Measures ings revealed a significant time main effect In order to document cardiovascular [F(7,21) = 3.91, p < 0.007]. The multivartraining , a set of three variables were se- iate group main effect and the lected from an array of measurements ob- Time x Group interaction were not statistained during exercise treadmill testing tically significant. Examination of the uni(Table 1). These variables included heart variate time main effects revealed that both rate (HR) at a submaximal workload (3 mph groups experienced significant reductions at 5% grade or 5 Mets), VO2max, and total in the number of interruptions, emitted treadmill time (in minutes). The MAN- more plosive words, and achieved lower OVA of these three variables yielded a sig- ratings of potential for hostility at Time 2 nificant main effect for time than at Time 1. [F(3,23) = 13.72, p < 0.0002] and a significant Time x Group interaction Psychometric Scores [F(3,23) = 15.73, p < 0.0001]. The groups did not differ on any variThe mean scores for the JAS and TASRI able at Time 1. At Time 2, the AE group are presented in Table 4. The MANOVA had significantly lower HR at submaximal of these scores revealed a significant time workloads, achieved a higher peak VO2, or main effect [F(5,25) = 3.35, p < 0.01]. The V02max, and exercised longer on the multivariate group main effect and the treadmill than did the SF group. The AE Group X Time interaction did not reach group increased their average V02max by statistical significance. Examination of Ta15%, while the SF group remained essen- ble 4 shows that both groups tended to tially unchanged. report a reduction in some components of Since not all lipid values were expected Type A behavior. There was no change in to change with exercise, a series of re- global Type A behavior as measured either 422

Psychosomatic Medicine 50:418-433 (1988)

EXERCISE A N D TYPE A BEHAVIOR

TABLE 1. Mean Cardiorespiratory (HR, BP, MVO2, and V02max) Responses During Exercise Treadmill Testing" Aerobic Exercise

Heart rate (bpm) Resting Submaximal Maximal Blood pressure (mm Hg) Resting SBP Submaximal SBP Maximal SBP Resting SBP Submaximal DBP Maximal DBP Rate pressure product(MVO2) Submaximal HR x SBP x 10"2 Maximal HR x SBP x 10"2 Maximal oxygen consumption VC^max (ml/kg/min) (1/min) Time Minutes

Strength and Flexibility

Time 1

Time 2

Time 1

Time 2

74.6 ± 14.4 116.6 ± 17.6 179.6 ± 14.7

65.2 ± 11.46 105.6 ± 14.1 d 178.9 ± 8.7

76.3 ± 9.3 116.1 ± 10.9 180.9 ± 9.6

71.6 ± 7.4 116.1 ± 14.3 177.5 ± 8.1

126.5 151.6 190.6 86.2 84.3 85.0

125.7 144.6 198.5 84.6 79.2 81 7

130.6 165.2 194.3 85.0 84.3 83.1

138.9 170.3 200.7 90.0 85.5 82.8

± ± ± ± ± ±

11.3 16.2 18.4 10.4 10.5 12.0

± ± ± ± ± ±

11.7 19.4 25.7 9.2 9.9 13.5

± ± ± ± ± ±

14.8 14.2 16.9 9.6 8.4 9.8

£ 18.3 ± 26.7 ± 26.6 ± 9.2 ± 10.2 ± 12.5

178.2 ± 39.7 342.6 ± 45.0

154.5 ± 38.T 354.0 ± 38.9

192.4 ± 28.7 351.8 ± 37.9

200.4 ± 53.5 353.5 ± 56.1

34.5 ± 5.2 3.09 ± .4

39.8 ± 6.8b 3.42 ± .4 b

34.6 ± 5.7 2.97 ± .3

34.5 ± 4.8 3.05 + .3

12.7 ± 1.6

14.8 ± 1.6d

13.0 ± 1.6

12.9 ± 1.6

"HR, heart rate; SBP, systolic blood pressure; DBP, diastolic blood pressure; MVO 2 , myocardial oxygen consumption; VC^max, maximal oxygen consumption. b Significance levels of differences between Time 1 and Time 2. p < 0.05. c p < 0.001. d p < 0.01.

by the JAS or TASRI; however, examination of univariate ANOVAs revealed that scores on the Hard Driving component of the JAS were lower at Time 2 than at Time 1. Psychophysiologic Responses In order that changes in cardiovascular response during the MAT could be assessed, data were first analyzed by a threeway MANOVA. The independent variables were group, time, and phase (consisting of resting, task, and recovery phases of the MAT). Dependent variables included systolic blood pressure (SBP), diastolic blood pressure (DBP), and HR. Significant main effects were observed for time Psychosomatic Medicine 50:418-433 (1988)

[F(3,27) = 7.10, p < 0.001] and phase [F(6,24) = 22.33, p < 0.001]. The phase main effect was due to the greater HR and BP values during the task than during rest and recovery, reflecting the effectiveness of the MAT in eliciting elevations in HR and BP. The Group x Time x Phase interaction also was significant [F(6,24) = 2.56, p < 0.04]. Significant univariate Group x Time x Phase interactions for SBP [F(2,28) = 4.20, p < 0.04] (Fig. 1) and HR [F(2,28) = 3.63, p < 0.03] (Fig. 2) were observed, but not for DBP [F(2,28) = 1.16, p = ns]. In order that the triple interaction could be further explored, two-way interactions were next considered. The Group x Phase interaction was not significant at either 423

EXERCISE AND TYPE A BEHAVIOR

X a

o. 3

O

O

vC

CO

i— LD O

Ln

E 2 P u

a.

o.

g

u.

E

u_

C

3

2 o b-

1

IO O

\6 r-^

N

O

d

d

s for

W

o o o o rN O +1 +1 +1 +1 +1 +1 +1

>-

S roro

nerit Scor

§ f) c aj £ 55

Time

-a c to CO ro i ^* o O o O .— o +1 +1 +1 +l + 1 +1 +1

06 01 13 26 66 55 66

Uni •o

lex ibillity

is *E S3

IN

89 06 46 80

s "3

o 0.

S O o O O O •— O

I

Bel

cd 01

i Type

ibic Exei

5 20)

1

TABLE

+1 +1 +1 +1 +1 +1 +1

r-

CO

ij

CO

U )

rs| u )

S

t —

ai

rv IN. CO •
s

O

o O O

t —

o

+1 +1 +1 +1 +1 +1 +1 0^ 00 (N

CO

00

O O ro ro

•2 a,

O "1 i—

.l| oo

— o d9d o V V V

EC X

Psychosomatic Medicine 50:418-433 (1988)

425

EXERCISE A N D TYPE A BEHAVIOR

T|

Aerobic D

Strength A

85--

150-Systolic 140--

75-130--

:c en

-

65

120

55

Diasto lie

Rest

Task

Recovery

85--

Fig. 2. Mean heart responses across the resting, task, and recovery phases of the mental arithmetic task for subjects in the aerobic exercise training group (AE) and the strength and flexibility training group (SF).

75--

65

Rest

Task Recovery

Fig. 1. Mean blood pressure responses across the rest, task, and recovery phases of the mental arithmetic task for subjects in the aerobic exercise training group (AE) and the strength and flexibility training group (SF).

Time 1 or Time 2. When the AE and SF groups were considered separately, however, different patterns emerged. The MANOVA for the SF group revealed only a significant phase main effect [F(6,10) = 16.97, p < 0.001]. The time main effect and the Phase x Time interaction were not significant. In contrast, the MANOVA for the AE group revealed significant main effects for phase [F(6,9) = 7.18, p < 0.004] and time [F(3,12) = 4.91, p < 0.01] and a marginally significant Time x Phase interaction [F(6,9) = 2.74, p < 0.08]. These results indicate that the exercise treatments led to differential changes over time in patterns of cardiovascular responses. The cardiovascular responses did not change Psychosomatic Medicine 50:418-433 (1988)

over time for the SF group; however, the cardiovascular responses were significantly lower for the AE group at Time 2 than at Time 1. For the AE group, significant univariate time main effects were observed for SBP [F(l,14) = 10.91, p < 0.005], DBP [F(l,14) = 15.23, p < 0.001], and HR [F(l,14) = 3.71, p < 0.07], and there was a significant Time x Phase interaction for SBP [F(2,13) = 8.44, p < 0.004]. Since the intervention was designed to affect resting cardiovascular levels, an ANCOVA was performed to compare the groups using the same-session rest values as covariates. This technique has been suggested to be useful when regression to the mean or habituation occur with repeated measurements (31]. These analyses revealed a pattern of results consistent with the MANOVA. That is, in addition to a significant Group x Time x Phase interaction for SBP and myocardial oxygen consumption (MV02), the Time x Phase interaction was significant for SBP, HR, 427

EXERCISE A N D TYPE A BEHAVIOR

130--

70

Aerobic a

Strength

T,

h

•

A

Rest

Task

A

Recovery

Fig. 3. Mean rate pressure product (MV02) responses across the resting, task, and recovery phases of the mental arithmetic task for subjects in the aerobic exercise training group (AE) and the strength and flexibility training group {SFJ.

J. A. BLUMENTHAL et al.

During the recovery phase of the MAT, blood pressures were significantly lower at Time 2 than at Time 1 for the AE group for SBP [F(l,13) = 24.55, p < 0.001], DBP [F(l,13) = 6.24, p < 0.05], and for HR [F(l,13) = 5.06, p < 0.05]. The AE group also experienced lower levels of estimated MV02 at Time 2 than at Time 1 [F(l,13) = 15.97, p < 0.001]. On the other hand, the recovery levels of SBP (F < 1), DBP (F < 1.1), HR (F < 1.4) and MV02 (F < 1) were not significantly different at Time 2 than at Time 1 for the SF group. Analysis of Change Scores

and MV02 for the AE group but not for the SF group. Covariate-adjusted BP levels during the task phase were significantlylower for the AE group at Time 2 than at Time 1 for SBP [F(l,13j = 29.94, p < 0.001] and DBP [F(l,13) = 12.29, p < 0.01]. The HR during the task phase also tended to be lower at Time 2 than at Time 1 for the AE group [F(l,13] = 3.23, p < 0.10]. In contrast, task phase HR was the same at Time 1 and Time 2 for the SF group (F < 1), although SBP (p < 0.10) and DBP (p < 0.01) also tended to be lower at Time 2 for the SF group, reflecting habituation effects. In addition, the double product (HR x SBP) was calculated as an indirect measure of myocardial oxygen consumption (MV02). The pattern of changes are displayed in (Figure 3). The ANCOVA revealed that the estimated MV02 during the task was lower for the AE group at Time 2 than at Time 1 [F(l,13) = 7.24, p < 0.05]. However, MV02 did not change for the SF group between Time 1 and Time 2, (F < 1).

In addition to the analyses described above, a series of ANCOVAs were performed on the change scores (i.e., task-rest) for each of the dependent variables (with testing values for each session serving as covariates) to assess changes in reactivity to the MAT (Table 5). A 2 (Time) x 2 (Group) ANCOVA of the change scores, covarying baseline values, indicated significant time main effects for SBP [F(l,28) = 22.83, p < 0.001] and DBP [F(l,28) = 26.47, p < 0.001]. Table 5 shows that both groups experienced significant reductions in BP reactivity at Time 2, although the SBP decrease of the AE group was slightly greater than that of the SF group (p < 0.11). Although there were no significant HR effects, HR reactivity tended to decrease from Time 1 to Time 2 for the AE group (p < 0.09), but HR reactivity remained virtually unchanged for the SF group (F < 1). The reduced HR response for the AE group was significantly greater than that of the SF group (p < 0.05). There was a significant time main effect for MVO2 [F(l,28) = 4.53, p < 0.05] and a marginally significant Time x Group interaction [F(l,28) = 3.71, p < 0.07]. Tests of simple effects revealed that MVO2 re-

428

Psychosomatic Medicine 50:418-433 (1988)

EXERCISE A N D TYPE A BEHAVIOR

TABLE 5.

Mean Adjusted Reactivity Scores During MAT*

Aerobic Exercise

SBP DBP HR MVO 2

14.8 9.1 12.0 26.6

± ± ± ±

Strength and Flexibility Time 2

Time 1 7.8 4.9 6.8 14. 7

9.3 6.1 9.6 19.3

Time 1

± 7.5b ± 4.5" ± 7.1 e ± 14.6d

12.9 9.8 9.4 22.4

± 7.0 ± 5.8 ± 5.2 ±11.0

Time 2 11.3 7.0 9.4 21.7

± ± ± ±

6.2" 4.1 C 6.9 13.8

"MAT, mental arithmetic task. Mean adjusted change scores (task-rest), p values represent within-group comparisons across time, p < 0.001. p < 0.01. d p < 0.05. e p < 0.10. b c

sponse was significantly reduced following aerobic training (p < 0.02). However, the MV02 response did not change for the SF group (F < 1). The reduced MVO2 response was significantly greater for the AE group than for the SF group (p < 0.05). Change scores were also analyzed for return to baseline levels (i.e., rest-recovery), again with resting levels serving as covariates in the two-way ANCOVA. These results revealed significant time main effects for SBP [F (1,28) = 17.98, p < 0.001], DBP [F(l,28) = 7.95, p < 0.01] and MV02 [F(l,28) = 11.23, p < 0.01]. Significant Time x Group interactions were also observed for SBP [F(l,28) = 8.83, p < 0.01], DBP [F(l,28) = 3.15, p < 0.08], HR [F(l,28) = 6.72, p < 0.05], and MV02 [F(l,28) = 17.12, p < 0.001]. As shown in Table 6, TABLE 6.

tests of simple effects revealed that cardiovascular responses returned to baseline more quickly (i.e., the difference between rest and recovery levels was significantly reduced) after training for the AE group but not for the SF group. DISCUSSION

The results of this investigation suggest that a program of aerobic exercise training may help to reduce the risk of CHD among healthy Type A men. First, aerobic exercise, but not strength and flexibility training, increased levels of cardiovascular fitness. Peak or maximum oxygen consumption increased 15% in men participating in the 12-week AE

Mean Adjusted Change Scores During Recovery

Aerobic Exercise Time 1 SBP DBP HR MVO 2

5.18 2.02 0.72 4.34

± ± ± ±

3.8 3.4 2.3 4.0

Strength and Flexibility Time 2

-0.04 -0.09 -1.72 -2.29

± ± ± ±

4.6a 4.1b 3.2b 5.7C

Time 1 1.97 1.04 -0.77 -0.38

± ± ± ±

3.4 3.9 2.5 4.6

Time 2 1.78 1.02 -0.10 1.19

± ± ± ±

4.2 3.4 2.9 5.3

"Mean adjusted change scores (rest-recovery), p values represent within-group comparisons across time, p < 0.001. b p < 0.05. c p<0.01.

Psychosomatic Medicine 50:418-433 (1988)

429

I. A. BLUMENTHAL et al.

training program but was unchanged in the SF group. Since physical fitness is predictive of CHD mortality and morbidity (32, 33), and even moderate exercise may reduce risk of future CHD events (19), this finding may be especially important. Moreover, aerobic exercise was also associated with other favorable changes in risk for CHD, including a significant increase in HDL cholesterol. Participation in both exercise groups was associated with reduced overt behavioral manifestations of Type A behavior and selfreported Type A mannerisms. However, the reductions were not greater for aerobic training than for strength training. In contrast, aerobic exercise appeared to have an advantage over strength and flexibility training when responses to the psychophysiological task were examined. Subjects in the AE group generally showed lower levels of cardiovascular response to mental arithmetic after exercise training than before training. Although both groups had lower BPs, the AE group also had a lower HR and lower estimated MVO2 during the task and lower BP, HR, and estimated MVO2 during recovery. The SF group also exhibited lower DBP during the task, which was most likely a result of habituation; however, neither SBP, HR, nor estimated MVO2 was significantly lower at Time 2 for the SF group during the MAT or during recovery. Several features of these data deserve comment. The Time x Phase multivariate interaction was significant only for the AE group. Nevertheless, we conducted simple effect tests for both groups and caution the reader to interpret the results conservatively. Although both groups tended to show an attenuation of cardiovascular responses during the MAT at Time 2, only the AE group experi-

enced significantly lower SBP, HR, and MVO2 levels during the task. Furthermore, differences in the levels of cardiovascular responses between the AE and SF groups were even more pronounced during the recovery phase. The AE group had significantly lower SBP, DBP, HR, and MVO2 levels at Time 2 than at Time 1. In contrast, there were no differences in the levels of any of these measures for the SF group at Time 2 compared with Time 1. Analysis of change scores was complicated by the tendency for the AE group to be more reactive than the SF group initially, and by reduced reactivity due to habituation in both groups. However, with baseline values covaried the AE group experienced a significantly greater attenuation of HR, SBP, and MVO2 responses during the task than did the SF group. This procedure is actually conservative since the treatments inherently exert a differential effect on the baseline values. Moreover, the AE group had a clear advantage over the SF group following the MAT, returning to resting levels significantly more quickly. The present data suggest that the autonomic adaptations associated with aerobic exercise exert a beneficial influence on cardiovascular responses to mental stress that cannot be attributed simply to familiarity with the task or to compliance with the "demand characteristics" of the testing situation. These longitudinal data extend results from previous cross-sectional studies, which showed that fit individuals display smaller increases in SBP (23, 33, 34), DBP (23, 35), and HR (36) responses, smaller T-wave amplitude attenuation (37), and faster HR recovery (38, 39) during psychosocial stressors. The results of this study indicate that these cardiovascular responses may be successfully modified by

430

Psychosomatic Medicine 50:418-433 (1988)

EXERCISE AND TYPE A BEHAVIOR

a relatively brief program of aerobic exercise training. It should be noted that our results are discrepant from those of Roskies and colleagues (9, 21), who reported a trend only for reduced HR and SBP from baseline to peak response in their aerobic group. Although the reason for this inconsistency is not clear, there are numerous methodological differences in the two studies (e.g., determination of aerobic capacity, exercise prescription, duration of baseline periods, selection of subjects, stress tasks, and statistical approach) that could account for the difference in findings. Though these data indicate that aerobic exercise training reduces cardiovascular reactivity to stress among healthy Type A men, direct demonstration of the clinical significance of this effect remains to be documented. Several considerations do provide support for the beneficial effects of reduced cardiovascular reactivity. First, the relationship between Type A personality and CHD may be mediated by frequent and sustained activation of the sympathetic-adrenomedullary and pituitaryadrenocortical systems (40). Second, animal studies have shown that increased cardiovascular reactivity to behavioral challenges is associated with increased levels of coronary atherosclerosis (41). Moreover, it has been demonstrated that surgically lowering HR is associated with slower progression of CAD (42). Third, hu-

man studies also have suggested that reactivity may be important in the development of hypertension and CHD (43). Our data suggest that exercise training not only reduces HR and BP responses during submaximal exercise, but also may reduce levels of BP and HR to psychosocial stressors. Although our findings are encouraging, prospective studies, with a more extended follow-up, are needed to document the clinical benefits of reduced cardiovascular responses to psychosocial stressors in persons who undergo aerobic training. Further studies are also needed to investigate the mechanisms by which these responses are attenuated. The authors wish to express their appreciation to Robin H. Pomeroy, Julie Whitaker, Sally Schnitz, flick Friedrich, and Jaye Efland for technical support; to Dr. Lars-Goran EkeJund for advice about the exercise testing protocol; to Drs. David Madden and James Lane for suggestions on an earlier version of this manuscript; to Dr. Kathleen Light for her statistical advice; and to Janet Ivey for secretarial assistance. This research was supported in part, by a grant from the John D. and Catherine T. MacArthur Foundation and by grants from the National Heart, Lung, and BJood Institute (HL30675) and the National Institute on Aging (AGO4238).

REFERENCES 1. Blumenthal JA, Levenson RM: Behavioral approaches to secondary prevention of coronary heart disease. Circulation 76(Suppl I):I-130, 1987 2. Review Panel on Coronary-Prone Behavior and Coronary Heart Disease. Coronary-prone behavior and coronary heart disease: A critical review. Circulation 63:1199-1215, 1981 3. Rosenman RH, Brand RJ, Jenkins D, Friedman M, Straus R, Wurra M: Coronary heart disease in the Western Collaborative Group Study: Final follow-up experience of 8.5 years. JAMA 233:872-877,1975

Psychosomatic Medicine 50:418-433 (1988)

431

J. A. BLUMENTHAL et al. 4. Matthews KA, Haynes SG: Type A behavior and coronary risk: Update and critical evaluation. Am J Epidemiol 123:923-960, 1986 5. Suinn RM, Bloom LJ: Anxiety management training for Pattern A behavior. J Behav Med 1:25-35,1978 6. Jenni MA, Wollersheim JP: Cognitive therapy, stress management training and the Type A behavior pattern. Cogn Ther Res 28:279-287, 1979 7. Levenkron JC, Cohen JD, Mueller HS, Fisher EB Jr: Modifying the Type A coronary-prone behavior pattern. J Consult Clin Psychol 51:192-204, 1983 8. Roskies E, Spevack M, Surkis A, Cohen C, Gilman S: Changing the coronary-prone Type A behavior pattern in a non-clinical population. J Behav Med 1:201-216, 1978 9. Roskies E, Seraganian P, Oseasohn R, Hanley JA, Collu R, Martin N, Smilga C: The Montreal Type A intervention project: Major findings. Health Psychol 5:45-69, 1986 10. Powell LH, Friedman M, Thoresen CE, Gill JJ, Ulmer D: Can the Type A behavior pattern be altered after myocardial infarction? A second year report for the Recurrent Coronary Prevention Project. Psychosom Med 46:293-313, 1984 11. Friedman M, Thoresen CE, Gill JJ, Powell L, Ulmer D, Thompson L, Price VA, Rabin DD, Breall SW, Dixon T, Levy RA, Bourg E: Alteration of Type A behavior and reduction in cardiac recurrences in post-myocardial infarction patients. Am Heart J 108:237-248, 1984 12. Friedman M, Thoresen CE, Gill JJ, Ulmer D, Powel LH, Price VA, Brown B, Thompson L, Rabin DD, Breall WS, Bourg E, Levy R, Dixon T: Alteration of type A behavior and its effect on cardiac recurrences in post myocardial infarction patients: Summary results of the recurrent coronary prevention project. Am Heart J 112:653, 1986 13. Gill JJ, Price VA, Friedman M, Thoresen CE, Powell LH, Ulmer D, Brown B, Brews FR: Reduction in Type A behavior in healthy middle-aged American military officers. Am Heart J 110:503-514, 1986 14. MacDougall JM, Dembroski TM, Musante L: The Structured Interview and questionnaire methods of assessing coronary prone behavior in male and female college students. J Behav Med 2:71-84, 1979 15. Krantz DS, Manuck SB Acute psychophysiologic reactivity and risk of cardiovascular disease: A review and methodologic critique. Psychol Bull 96:435-464, 1984 16. The Multiple Risk Factor Intervention Trial Group. The MRFIT Behavior Pattern Study-l. Study design, procedures, and reproducibility of behavior pattern judgments. J Chron Dis 32:293-305, 1979 17. Claussen JP: Effect of physical training on cardiovascular adjustments to exercise in men. Physiol Rev 57:779-816, 1977 18. Paffenbarger RS, Hyde RT, Irving AL, Steinmetz CH: A natural history of athleticism and cardiovascular health. JAMA 252:491-495, 1984 19. Leon AS, Connett J, Jacobs DR, Rauramaa R: Leisure-time physical activity levels and risk of coronary heart disease and death: The Multiple Risk Factor Intervention Trial. JAMA 258:2388-2395, 1987 20. Blumenthal JA, Williams RS, Williams RB Jr, Wallace AG: Effects of exercise on the Type A (coronaryprone) behavior pattern. Psychosom Med 42:289-296, 1980 21. Seraganian P, Roskies E, Hanley JA, Oseasohn R, Collu R: Failure to alter psychophysiological reactivity in type A men with physical exercise or stress management programs. Psychol Health 1:195-213,1987 22. Astrand PO, Rodahl K: Textbook of Work Physiology, New York, McGraw Hill, 1977 23. Lake BW, Suarez EC, Schneiderman N, Tocci N: The Type A behavior pattern, physical fitness, and psychophysiological reactivity. Health Psychol 4:169-187, 1985 24. Hurley BF, Seals DR, Ehsani AA, Cartier LJ, Dalsky GP, Hagberg JM, Holloszy JO: Effects of highintensity strength training on cardiovascular function. Med Sci Sports Exer 16:483—488, 1984 25. Balke B, Ware RW: An experimental study of physical fitness of Air Force personnel. US Armed Forces Med J 10:675-688, 1959 26. Blumenthal JA, Rejeski J, Walsh-Riddle M, Emery CF, Miller H, Roark S, Ribisl PM, Morris PB, Brubaker P, Williams RS: Comparison of high and low intensity exercise training early after acute myocardial infarction. Am J Cardiol, 61:26-30, 1988 27. Borg G: Physical Performance and Perceived Exertion. Lund, Sweden, Gleerup, 1962 28. Blumenthal JA, O'Toole LC, Haney T: Behavioral assessment of the Type A behavior pattern. Psychosom Med 46:415^123, 1984 29. Jenkins CD, Zyzanski SJ, Rosenman RH: Progress toward validation of a computer-scored test for the Type A coronary prone behavior pattern. Psychosom Med 33:193—201, 1971 432

Psychosomatic Medicine 50:418^133 (1988)

EXERCISE A N D TYPE A BEHAVIOR 30. Blumenthal JA, Herman S, OToole LC, Haney TL, Williams RB Jr, Barefoot JC: Development of a brief self-report measure of the Type A (coronary prone) behavior pattern. J Psychosom Res 29:265—274, 1985 31. Lee J: A note on the comparison of group means based on repeated measurements of the same subject. J Chron Dis 33:673-675, 1980 32. Ekelund LG, Haskell WH, Johnson JL, Whaley FS, Sheps DS: Physical fitness as predictor of cardiovascular (CVD) mortality in asymptomatic men. The lipid research clinics (LRC) prevalence follow-up study. Circulation 74(Suppl 11-397), 1986 33. Light KC, Obrist PA, James SA: Self-reported exercise levels and cardiovascular responses during rest and stress [Abstract]. Psychophysiology 21:586, 1984 34. Cantor JP, Zillman D, Day KD: Relationship between cardiovascular fitness and physiological responses to films. Percept Mot Skills 46:1123-1130, 1978 35. Hull EM, Young SH, Ziegler MG: Aerobic fitness affects cardiovascular and catecholamine responses to stressors. Psychophysiology 21:353-360, 1984 36. Holmes DS, Roth DL: Association of aerobic fitness with pulse rate and subjective responses to psychological stress. Psychophysiology 22:525-529, 1985 37. Shulhan D, Seer H, Furedy JJ: Phasic cardiac reactivity to psychological stress as a function of aerobic fitness level. Psychophysiology 23:562-566, 1986 38. Sinyor D, Schwartz SG, Peronnet F, Brisson G, Seraganian P: Aerobic fitness level and reactivity to psychosocial stress: Physiological, biochemical, and subjective measures. Psychosom Med 45:205-217, 1983 39. Cox JP, Evans JF, Jamieson JL: Aerobic power and tonic heart rate responses to psychosocial stressors. Pers Soc Psychol Bull 5:160-163, 1979 40. Williams RB, Lane JD, Kuhn CM, Melosh W, White AD, Schanberg SM: Type A behavior and elevated physiological and neuroendocrine responses to cognitive tasks. Science 218:483—485, 1982 41. Manuck SB, Kaplan JR, Clarkson TB: Behaviorally-induced heart rate reactivity and atherosclerosis in cynomolgus monkeys. Psychosom Med 45:95-108, 1986 42. Beere PA, Glagov S, Zarsins CK: Retarding effect of lowered heart rate on coronary atherosclerosis. Science 226:180-182, 1984 43. Matthews KA, Weiss SM, Detre T, Dembroski TM, Falkner B, Manuck SB, Williams RB Jr: Handbook of Stress, Reactivity, and Cardiovascular Disease. New York, Wiley, 1986

Psychosomatic Medicine 50:418-433 (1988)

433