Effects of caffeine on pressor regulation during rest and exercise in men at risk for hypertension

Effects of caffeine on pressor regulation during rest and exercise in men at risk for hypertension Caffeine-induced blood pressure elevations are well documented in habitual consumers, occurring through both vasconstrictive and cardiostimulatory actlons. Whether caffeine hinders pressor regulation during exercise has been uncertain, particularly in those at risk for hypertension. Thus effects of caffeine versus placebo were studied during supine bicycle exercise In healthy men (ages 20 to 35). Hypertension risk was defined during screening: high risk (HAISK) = 135 to 154/85 to 94 mm Hg plus parental hypertension (n = 20); low risk (LRISK) = 1132/84 mm Hg and no parental hypertension (n = 14). Exaggerated pressor responses (r230/100 mm Hg) seen during exercise after placebo identified a subgroup of seven HRlSKs indistinguishable at rest from the remaining HRISK men. This subgroup showed a larger resting diastolic response to caffeine (p < 0.05) than LRISKs and other HRISKs. Compared with placebo, caffeine increased the number of LRISK (0% to 36%) and HRISK (35% to 50%) men reaching abnormal exercise blood pressures, and blunted normal increments in cardiac index at higher workloads among HRISK men (p = 0.05). Thus restrictlon of caffeine before exercise might benefit persons with either risk for hypertension or unusual sensitivity to caffeine. (AM HEART J 1991;122:1107.)

Gwendolyn A. Pincomb, PhD? b,c Michael F. Wilson, MD> c,d Bong Hee Sung, PhD, b*d,f Richard B. Passey, PhDF and William R. Lovallo, PhD.” b! c Oklahona City, Okla.

Caffeine has been shown to produce reliable blood pressure (BP) elevations in regular users.lV8 These pressure increases may be produced through either vascular or cardiac mechanisms.‘, 2 Several studies conducted in our laboratory11 2y4 confirm that caffeine’s pressor response at rest may be attributed to heightened systemic vascular resistance. In contrast, during work on challenging psychomotor tasks, caffeine elevated BPS by potentiating the cardiac stimulatory properties of the task.2 Sustained and episodic elevations in both vascular toneg-‘I and cardiac drive I2 have been reported in those at risk for essential hypertension, although their relative contributions in the course of disease progression have been disputed. Evidence for early vascular dysregulation in persons with hypertensive FromaMedical Research Service and bthe Behavioral Sciences Laboratories, Veterans Affairs Medical Center; and ‘the Departments of Psychiatry and Behavioral Sciences, dMedicine, ePathology, and ‘College of Pharmacy, Oklahoma Health Sciences Center. Supported by National Heart, Lung, and Blood Institute grant HL320.50, and by Department of Veterans Affairs Medical Research Service. Received for publication May 18, 1990; accepted April 5, 1991. Reprint requests: Gwendolyn A. Pincomb, PhD, Behavioral Sciences Laboratories _ 151A, Veterans Affairs Medical Center, 921 NE 13th St., Oklahoma City, OK 73104. 4/1131113

parents has been elicited by some vasoconstrictive substances.il> I3 Whether such vascular sensitivity generalizes to caffeine has not been previously examined. Furthermore, how caffeine use affects basic mechanisms underlying BP control during challenge in hypertension risk groups has not been well defined, particularly during natural physical stressors such as exercise. Thus the current study examined caffeine’s influence on BP regulation at rest and during graded, supine bicycle exercise in healthy young men differing in risk for essential hypertension. Hypertension risk was defined by a parental history of essential hypertension14-20 and high normal resting BPS with only transient elevations into borderline ranges (HRISK). Low risk controls had no known parental hypertension and consistently had low-normal blood pressures (LRISK). A secondary classification within the HRISK group emerged during the course of the study. We observed that a subgroup of HRISK men, not discriminable by any resting variable, displayed a hypertensive response to supine bicycle exercise on their placebo day, reaching systolic pressures of ~230 mm Hg and/or diastolic pressures ~100 mm Hg.” Preliminary evidence from other laboratories22-24 suggests that provocation of abnormal ex1107

1108

Pincomb

et al.

ercise BP levels also may be a useful predictor of hypertension risk. METHODS Subjects.

Normotensive, white male volunteers (n = 35, age 20 to 35 years) were recruited with their full informed consent and were paid for participation. One subject was omitted from data analysesafter identification of a dosing error (final n = 34). All subjectscamefrom a sampleof 40 men used in a more extensive study.21 Inclusion criteria were: regular caffeine use(40 to 937mg/day), weight within 20% of normal ranges (defined by the Metropolitan Life Insurance Company tables) and no aerobic functional impairment exhibited during an exercisetreadmill test (Bruce protocol). Exclusion criteria were: known cardiovascular disease,previous treatment for hypertension, other selfreported illnesses,regular smoking, use of recreational or prescription drugs, or caffeine intolerance. During eachof two screeningsessions, three casualBP measurementswere made at 2-minute intervals following 5 minutes of rest in a seatedposition. Volunteers were then classified as controls (LRISK: n = 15, all pressures1132/84 mm Hg) or high risk (HRISK: n = 20, with two or more pressuresbetween 135/85 and 154/94 mm Hg). All HRISK men displayed normotensive pressureswithin 15 minutes of further rest and reported having at least one parent with essentialhypertension. LRISK men were retained if neither parent was hypertensive. Confirmation of parental hypertension status wassought by phoneor letter from the parents’ physicians. Percent body fat wasmeasuredduring screeningprior to exercisetreadmill testing with calibrated skinfold calipers. The sum of skinfold thicknesses (in millimeters) was obtained from four siteson the left sideof the body (biceps, triceps, subscapular,and suprailiac) and was converted to percentagefat using norms adjusted for gender and age.2j Instrumentation and measurements. Habitual caffeine consumption was estimated by structured recall using a comprehensivequestionnaire of typical daily, weekly, or monthly intake from dietary and medicinal sources.Blood was withdrawn for labeling of red cells and measuring plasmacaffeine concentrations via an intravenous line attached to a 20-gaugeindwelling angiocatheter inserted in the left median cubital vein. Plasmawasseparatedimmediately following eachsession(1 ml eachfor pre- and postdosing caffeine concentrations) and was stored at -70” C until batched assayswere performed. Then caffeine was isolated from the plasma using acetonitrile precipitation and was quantified by high-pressure liquid chromatography.26 A cuff for automated measurementof BPS (Paramed, Palo Alto, Calif.) was attached to the subject’sright arm. Cardiac volumes were measured by electrocardiogramgated radionuclide ventriculography at rest and during each stage of exercise.27,2sCardiac imageswere acquired from the left anterior oblique view that offered optimal discrimination of the left ventricle from other structures. A mobile camerawith a low-energy, all-purpose collimator

October American

Heart

1991 Journai

and a computer interface alloweddigitization of multigated imagesto be acquired at rest and during the last 2 minutes of each exercisestagefor later calculation of end-diastolic and end-systolic volumes (in milliliters). Stroke volume wascalculated asthe difference betweenend-diastolic and end-systolic volumes.Cardiac output (in liters per minute) wasderived asthe product of stroke volume and heart rate obtained from thoseportions of the electrocardiogramthat were time-locked with each 2-minute period of nuclear imaging.Stroke volume and cardiac output were then indexed to body size by dividing these values by body surface area (in square meters). Systemic vascular resistance index (dynes. sec.cmw5me2)wascalculatedfrom simultaneously acquired BPS and cardiac volumes by the standard formula: (mean arterial pressure+ cardiac index) I 80. Study protocol. Volunteers reported to the Nuclear Cardiac Studies laboratory on 2 mornings at 8:00 AM following 12 hours’ abstinencefrom caffeine. Placebo and caffeine sessionswere held at least 2 days apart. Caffeine (3.3 mg/kg) wasadministered orally in 6 oz of unsweetened grapefruit juice; placebo drinks were juice alone. Each sessionrequired instrumentation, withdrawal of blood (5 ml) for labeling red cells with 25 mCi of 99mtechnetium (Tc-99m), and supine rest (30 minutes), followed by reinjection of labeled cells in the right median cubital vein, predrug measurements (2 minutes), drug administration (2 to 3 minutes), postdrug absorption period (40 minutes), postdrug baseline measurements (2 minutes), submaximal supine bicycle exercise (9 minutes), recovery rest period (1 hour), maximal supine bicycle exercise (12 to 18 minutes) and final recovery period (15 minutes). The full protocol lasted approximately 3 hours. Submaximal exerciserequired 3 minutes at eachworkload of 100, 200, and 400 kpm. A 5-minute isometric handgrip exercisetask wasperformed at 33SOmaximum grip strength halfway through the l-hour recovery rest period (unpublished data). Maximal exercise was initiated 30 minutes later, beginning with 3 minutes each at 100, 200, and 400 kpm and was continued at 200 kpm increments. Exercise wasstoppedfor severefatigue, shortnessof breath, or when the bicycle ergometerlimits had beenreachedat 1000kpm. Analyses. The primary study designcontrasted cardiovascular responsesto double-blinded, within-subject manipulation of caffeine and physical challengebetweenthree groupsvarying in hypertension risk. Assignmentof caffeine versus placebo was made according to a counterbalanced schedule.Subject characteristics measuredduring screening were compared in a one-way seriesof analysesof variance (ANOVA) for the LRISK and two subgroups of HRISK men (EXG = men whose exercise diastolic BP reached 2100 mm Hg and/or systolic BP ~230 mm Hg; and EXN = men with normal exercise pressure levels). When significant group effects werefound, further pairwise mean comparisonsbetween the three groups were computed using Tukey HSD tests. Subject characteristics, other than their primary risk classifications that might differ betweengroups,were further explored by examining their relative contributions to caffeine and exercise re-

Volume Number

Table

122 4, Part

Caffeine,

1

Group

p Value group effects)

(main

hypertension

1109

risk

Height (cm) Weight (kg) Percent body fat* Treadmill time (min) Caffeine use (mg/day) Caffeine dose (mg) Screening (after 5-min rest) Systolic BP+ (mm Hg) Diastolic BP? (mm Hg) Resting baseline (after 30-min rest) Systolic BPt (mm Hg) Diastolic BP (mm Hg) Heart rate (beats/min) Max exercise (placebo day only) Systolic BP3 (mm Hg) Diastolic BP$ (mm Hg) Heart rate (beats/min)

HRISK fEXGJ

HRISK (EXN)

LRISK


14 27.5 (1.3) 181.3 (1.9) 75.5 (2.4) 20.5 (0.9) 14.9 (0.7) 334 (103) 249.2 (8.0)

13 26.4 (1.4) 180.7 (2.2) 81.1 (3.7) 22.4 (1.5) 15.5 (0.9) 235 (62) 267.6 (12.3)

31.6 176.0 82.3 24.9 13.1 477 271.5

<0.0001 <0.002

116.9 (2.3) 65.9 (2.5)

134.1 (2.1) 76.2 (1.6)

131.9 (3.4) 77.3 (2.4)

<0.004 NS NS

116.2 (2.5) 70.8 (1.8) 57.6 (2.1)

126.6 (2.4) 75.8 (2.3) 62.6 (3.2)

126.8 (3.1) 75.4 (2.6) 54.4 (3.2)

<0.03 <0.004 NS

199.6 (5.0) 90.0 (1.5) 165.9 (3.2)

201.0 (4.1) 90.2 (1.5) 165.8 (2.9)

224.4 (10.8) 100.6 (3.8) 155.8 (6.1)

size (no.)

Age(~1)

Table values are mean (standard error). LRISK, No parental hypertension and screening BP 5 132/84 mm Hg; HRISK, parental hypertension Hg; EXG, maximal exercise diastolic BP 2 100 mm Hg or systolic BP z 230 mm Hg; EXN, maximal Tukey HSD results: *LRISK
7 (1.6) (2.4) (3.4) (1.3) (0.6) (130) (11.3)

and screening BP between 135/85 and 154/94 mm exercise BP below EXG levels; BP, blood pressure.

II. Resting values obtained at predrug baselineand at 40 minutes postdosing LRZSK Placebo Time

Mean

SE

Heart rate (beats/min) Systolic BP (mm Hg) Diastolic BP

Pre Post Pre Post Pre Post Pre Post Pre Post

59.1 58.7 116.9 118.6 71.3 72.2 2598 2760 2.71 2.60

2.1 2.1 2.2 2.2 1.5 2.0 101 122 0.10 0.11

(mmBg) Vascular resistance Cardiac index

SE, Standard error of the mean. Units for vascular resistance (dynes

EXN Caffeine

Variable

set

Mean 58.3 56.6 117.9 127.1 73.2 77.6 2659 2904 2.72 2.67

Placebo Mean

SE

1.3 1.9 2.7 2.6 1.7 2.2 150 186 0.11 0.13

62.2 66.1 129.4 130.0 77.7 78.0 2913 2730 2.68 2.85

3.3 3.3 1.8 2.0 1.9 1.8 168 128 0.11 0.11

cm-s) and cardiac output

(L/min)

at

rest

were

examined

using

repeated

measures

ANOVA having three groups (LRISK, EXN, and EXG), two drug conditions (placebo and caffeine), and two peri-

EXG Caffeine

SE

sponsesusing stepwise multiple regressionmodels. Risk group differences in plasmacaffeine levels were tested by repeated measuresANOVA for three groups (LRISK, EXN, and EXG) by three periods (40 minutes postdrug, submaximal, and maximal exercise). Relationships between self-reported habitual caffeine use(in milligramsper day) and pre- to postdosingBP changeswere tested using Pearson product-moment correlations. Responsesto caffeine

and

I. Characteristics of men varying in risk for hypertension Subject characteristics

Table

exercise,

are expressed

Mean 61.8 60.6 130.3 136.2 80.8 86.7 2905 3125 2.72 2.71

Placebo

Caffeine

SE

Mean

SE

Mean

SE

3.4 3.3 2.8 2.9 2.0 2.5 123 168 0.10 0.12

53.6 55.4 126.9 127.0 79.3 74.9 2897 2882 2.69 2.60

3.4 3.0 3.1 2.8 2.8 4.0 217 191 0.16 0.13

55.7 54.1 126.9 136.9 75.0 86.4 2794 3180 2.68 2.68

3.2 3.6 2.9 2.6 4.2 2.3 208 250 0.12 0.19

as indices of body surface area (m2).

ods (pre- and postdrug resting baselines).Responsesfor eachdependent variable during exercisewere examined in two repeated measuresANOVA having three groups, two drug conditions and repeated measuresfor three to four exercisework loads(submaximal = 100,200, and 400kpm; maximal = 100, ZOO,400, and 600 kpm). Work loadsthat werenot completedby all subjects(800and 1000kpm) were omitted from the analyses of maximal exercise periods. Tukey HSD tests and Pearsoncorrelations were tested using SYSTAT (IBM-PC compatible version4.1, Systat, Inc., Evanston, Ill). SYSTAT’s multivariate general linear hy-

1110

Pincomb

et al.

Amencan

. -~ 1 .l

-L----

~~..

EXG Risk Groups Fig. 1. Standard box plot of individual differences between caffeine and placebo days for the changein diastolic BP from pre- to postdosing.ADH’, Post - predosingDBP. Difference scores= Caffeine ADBP - Placebo ADBP. Vertical lines denote ranges for each risk classification. Middle horizontal line representsthe group median,while upper and lower box boundaries depict the 75th and 25th percentiles, respectively. No statistical outliers were detected in any group defined by a distance from the box boundariesof 1.5 x the box height. The two highest values in the EXG group depict men with habitual caffeine use >800 mglday.

pothesisprogramwasusedfor stepwisemultiple regression modelsand for an ANOVA on plasmacaffeine concentrations. All other ANOVAs were performed using BMDP P2V (IBM-mainframe version, BMDP Statistical Software, 1988,Los Angeles, Calif.). RESULTS Subject

characteristics. Table I displays characteristics of LRISK and HRISK men subclassified according to the presence (EXG) or absence (EXN) of a clinically abnormal pressor response to maximal exercise on the placebo day. Within the HRISK group, EXG men were not distinguishable from EXN men based on any resting variable measured prior to drug administration (Tables I and II). A tendency for group-related differences was seen only for age

October 1991 Heart Journal

(F = 2.76; df = 2,31; p = 0.079) and percent body fat (F = 2.99; df = 2,31; p = 0.065). Subsequent Tukey HSD tests indicated small pairwise differences only between LRISK versus EXG men (age: p = 0.069; percent body fat: p = 0.054), whereby the EXGs tended to be older and have more body fat than LRISKs. The influence of subject characteristics on responsiveness to caffeine and exercise manipulations will be discussed later. Caffeine dosage and plasma concentrations. Oral caffeine dosages, shown in Table I, did not differ significantly between risk groups; the slightly higher weight-adjusted doses in HRISK men reflect their tendency to be heavier. Plasma samples were not obtainable in five men. Mean plasma caffeine concentrations for the remaining subjects on placebo days were
times larger than that observed in the other men (group x drug x period: F = 3.53; df = 2,31; p = 0.042). Net differences* between caffeine and placebo days for the pre- to postdosing changes in diastolic BP are shown in Fig. 1 for each individual in the three risk groups. The range (+4 to +39 mm Hg) and distribution of individual diastolic responses attributable to caffeine was shifted upward for EXG men. Neither vascular resistance (group X drug X period: F = 1.02, df = 2,31; p = 0.37) nor cardiac output “See

Addendum.

P. 1114.

Volume Number

122 4, Part

Cafjeine,

1

LRlSK

EXN

EXG

LRISK

exercise, and hypertension

EXN

risk

1111

EXG

2. Responses during maximal exerciseperiod after placeboversuscaffeine. BL, Pre-exercisebaseline. 1,2,4, and 6 denote exerciseworkloads (x100) in kilopound/meters (kpm) that were completed by all subjects. SBP, Systolic blood pressure;DBP, diastolic blood pressure;VRI, vascular resistanceindex; CI, cardiac index. Fig.

(F = 1.47; df = 2,31; p = 0.25) responses alone were significantly more elevated in the EXGs than in other men. Since the EXG men tended to be older and have a higher percentage of body fat, we examined the influence of these two factors on resting BP responses to caffeine. While age and percent body fat were correlated with each other (r = 0.507, p = 0.003), age was not correlated with any resting cardiovascular response to caffeine. However, percent body fat was correlated with the resting diastolic BP response to caffeine (r = 0.539, p = 0.001). Since caffeine was administered as a weight-adjusted dose (3.3 mg/kg), we then examined whether a confounding influence of weight on dosage accounted for the association between percent body fat and diastolic BP responses to caffeine. While percent body fat was highly correlated with dosage (r = 0.658, p < O.OOl), the more critical correlations between plasma concentrations and body fat or any cardiovascular response to caffeine were all nonsignificant (7s < /fI.Z7/,ps > 0.36). Finally, several sources of individual difference that might be related to diastolic pressor responses to

caffeine were considered together in a series of stepwise multiple regression models (e.g., age, percent body fat, diastolic and systolic BPS observed during screening, caffeine dosage, plasma concentrations, and exercise BP responses). Only two factors contributed significantly to a model for predicting the resting diastolic pressor response to caffeine. Percent body fat (multiple r = 0.51, /3 = 0.569, p = 0.005 using a two-tailed t test) accounted for 26% of the variance, and maximum diastolic BP change from rest to exercise (/3 = 0.443, p = 0.003) accounted for an additional 19%. No correlation was found between percent body fat and the maximum exercise response (r = -0.140, p = 0.436), so each served as essentially independent predictors without overlapping variance (tolerances = 0.98) in a multiple regression model based on only these two variables. Thus a combination of these two hypertension risk factors accounted for 45% of the total variance in resting diastolic responsivity to caffeine and significantly improved prediction beyond that permitted by either factor alone (multiple r = 0.67, squared multiple F = 0.45, F = 12.23, df = 2,30, p < 0.001).

1112

Pincomb et al.

Amerlcan

Percent

100

200

Reaching

SBP

400

! 230

or DBP

100

Exercise c:I

1 100

200

400

Stages

Placebo

m

October 1991 Heart Journal

600

800

1000

(kpm) Caffeine

Fig. 3. Percent of men by risk group who displayed abnormal BP elevations during supine bicycle exer-

ciseafter placeboversus caffeine. Abnormal responseswere defined by occurrence of a systolic BP ~230 mm Hg and/or diastolic BP L 100mm Hg. Submaximal exerciseincluded the initial stages(100to 400kpm). Symptom-limited maximal exercisewasconducted after a l-hour rest and included stages100to 600 kpm, with continued 200 knm increments in work load every 3 minutes until either fatigue or completion of the 1000kpm load. -

Caffeine effects during exercise. The pressor effects of caffeine seen at rest persisted during the moderate work loads required for submaximal exercise across all risk groups for both systolic (F = 11.60; df = 1,31; p = 0.0018) and diastolic (F = 5.83; df = 1,31; p = 0.022) BPS. R esponse patterns similar to those seen during submaximal exercise were obtained 1 hour later during maximal exertion, so comprehensive analyses of other hemodynamic measures were deferred to the latter exercise period. On both days the group-related pressure differences seen at pre-exercise baseline persisted during maximal exercise for systolic BP (F = 4.80; df = 2,31, p = 0.015) and diastolic BP (F = 5.91; df = 2,31; p = 0.0067). Thus EXGs sustained higher BP levels at all workloads than either EXN or LRISK men (Fig. 2). Higher exercise BPS in EXGs were associated at each stage with maintenance of their elevated vascular resistance (F = 3.38; df = 2,31; p = 0.047) initially seen at rest. Caffeine continued to have an additive effect during exercise for systolic BP (F = 6.57; df = 1,31; p = 0.015), while drug-related elevations in diastolic BP (F = 3.62; df = 1,31; p = 0.067) and vascular resistance (F = 3.02; df = 1,31; p = 0.092) were somewhat weakened (Fig. 2). Cardiac index was the only measured variable that showed an interaction between work load, risk classification, and caffeine. When risk classification was based on only the two initial groups (LRISK versus HRISK), this interaction reached higher significance (F = 2.99; df = 3,96; p = 0.035; after GreenhouseGeisser correction p = 0.059) than when the HRISK

group were subdivided by EXN and EXG (F = 2.01; df = 6,93; p = 0.073; after correction, p = 0.092). Thus the typical increases in cardiac output associated with increasing work loads on the placebo day were diminished in HRISK men after caffeine ingestion. Neither age nor percent body fat alone, or in combination, was a significant predictor of maximum diastolic BP levels seen during exercise on either placebo or caffeine days. Similarly, these two factors were unrelated to caffeine-placebo differences in the change from. rest to maximum exercise for diastolic BP. Habitual caffeine use. Tolerance to caffeine’s influences on BP regulation from habitual use was not detected during rest or exercise. Habitual use was not negatively correlated with pre- to postdosing changes at rest in systolic or diastolic BP (KS 5 0.24, p’s 1 0.22), as would be expected if tolerance to caffeine’s vascular actions had occurred. In fact, the two EXG men who exhibited the largest caffeine-related diastolic BP changes at rest (Fig. 1) reported habitually consuming >800 mg/day of caffeine (190th percentile for all subjects), while the remaining EXGs consumed 139 to 606 mg/day. Maximum systolic and diastolic BP levels achieved during exercise with caffeine also were unrelated to habitual use (Irl’s I 0.15, p’s L 0.44). Caffeine and frequency of abnormal pressor responses. Fig. 3 reveals that the presence of caffeine in

EXNs resulted in BPS that met or exceeded the threshold BPS used to define EXGs under placebo

Volume Number

122 4, Part

1

conditions. Thus the cumulative percentage of men exhibiting clinically abnormal pressor responses during exercise after placebo and caffeine was plotted by risk group in Fig. 3. All stages of exercise were included to encompass each individual’s maximum performance. Caffeine induced additional abnormal pressor responses in three HRISK (15% ) and five LRISK (36%) men who displayed otherwise normal BP regulation during exercise on the placebo day. Thus the overall prevalence of abnormal pressor responses on the caffeine day was 36% for LRISK and 50% for HRISK men. DISCUSSION

Two key findings emerge from the current study. First, individual differences were observed in the resting diastolic BP response to caffeine. The greatest caffeine sensitivity was found in men with four potential risk factors for hypertension: a positive parental history, high-normal resting BP, an exaggerated pressor response to exercise, and a higher percentage of body fat. Second, caffeine appeared to enhance provocation of abnormal BP responses during exercise, even among men possessing fewer or none of the preceding risk factors. These results bear clinical and theoretical implications for caffeine use in persons with elevated risk for developing cardiovascular disease. A group of men was identified who showed heightened sensitivity to caffeine’s vascular effect, evidenced by an increase in resting diastolic BP two to three times larger than that which was seen in other subjects. Since this heightened sensitivity to caffeine during resting conditions was seen predominantly in EXGs, a subgroup of our HRISK subjects, its occurrence may yield useful information for elucidating physiologic mechanisms underlying their hypertension risk. No differential findings concerning caffeine dosing, plasma concentrations, or regular use accounted for their caffeine sensitivity. Also, despite the “co-occurrence” of larger diastolic BP responses to maximum exercise and higher percent body fat among EXG men, these two variables were not systematically correlated with each other. Instead, each apparently contributed in an independent, additive fashion to the caffeine-induced diastolic BP elevations. The latter finding may suggest that each variable may exert its own influence through different mechanisms. Current available studies were not intended to explore differences in BP regulation to caffeine challenge as a function of body fat; thus future studies will be required to describe relevant underlying hemodynamic relationships. A tendency toward vascular dysregulation might be consistent with our findings in which EXG men,

Caffeine, exercise, and hypertension

risk

1113

classified by their exaggerated exercise responses, showed higher vascular resistances prior to initiating exercise. They also maintained higher vascular resistances in the face of maximum bicycle exercise, normally associated with robust vasodilation. Mechanisms for potentially disturbed vascular regulation in EXG men have not been determined. Heightened vascular reactivity to noradrenergic stimulation11*13 and behavioral challengelo has been reported in normotensive offspring of hypertensive patients. However, vasoconstriction following caffeine ingestion under resting conditions appears to result from antagonism of peripheral adenosinerelated vasodilation2g> 3o rather than from excitation of adrenergic mechanisms. In addition to known robust adrenergic and local metabolic responses, adenosine-induced vasodilation has also been implicated as a possible autoregulatory response for maintaining muscle blood flow during vigorous exercise.31 Further studies will be required to fully characterize the pattern of vasomotor behavior seen in our EXG men and to explain how these potential mechanisms for disrupting vasodilatory reserve through short-lived interferences with vasodilation and/or exaggerated vasoconstriction might interact with more long-term structural adaptations such as vascular thickening and left ventricular hypertrophy, both of which can precede the onset of clinically recognizable hypertension. Caffeine’s pressor effects persisted during both submaximal and maximal exercise. However, caffeine enhanced exercise BPS more in the LRISK and EXN men than in the EXG men. It is noteworthy that both HRISK groups evidenced reduced cardiac output responses at higher work loads in the presence of caffeine. On placebo days, HRISK men had equivalent cardiac output levels and work load-related changes as LRISK men. Yet on the caffeine day, with identical work requirements and presumably the same oxygen demands during exercise, HRISK men displayed smaller increments in cardiac output. Such a response, despite serving as a potential limiter to BP elevations during exercise, is of questionable benefit, since it might also increase the likelihood of exercising under ischemic conditions. Finally, the persistence of pressor responses to caffeine during exercise may have implications for physicians who supervise exercise programs. Findings about caffeine effects on exercise performance are mixed. Some report improved exercise endurance in healthy athletes32 and extended exercise durations before angina1 pain in patients with coronary disease.33 Yet other studies found no caffeine-related increases in exercise duration or performance34-38 and one study3s reported that regular caffeine use

I 114

Pincomb et al.

American

was associated with decreased tolerance to an exercise treadmill test. We observed that caffeine provoked abnormal BP responses during exercise among some LRISK and EXN men whose BP changes after placebo were within the normal range. Thus compared to the placebo day, the total number of men displaying abnormal exercise BPS after caffeine rose from 0% to 36 % in LRISK men and from 35 % to 50 % in HRISK men. The popular view that “caffeine induces dilation of the peripheral vasculature during exercise . . . an obvious advantage during muscular work,“3g appears premature for persons susceptible to caffeine’s pressor effects. In such cases, caffeine may actually compromise an otherwise normal vasodilatory reserve. Another practical issue raised by our findings concerns the possible presence of tolerance to caffeine in habitual users. The present study provides no evidence that caffeine’s effects at rest or during exercise are diminished in men who consume larger daily quantities. It is noteworthy that the largest caffeinerelated changes in diastolic BP (+28 and +39 mm Hg) at rest occurred among EXG men whose regular use exceeded 800 mg/day. Thus it seems that regular use may not confer sufficient protection from caffeine’s vasoconstrictive actions in some individuals. While complete tolerance to pressor effects with regular use has been suggested,40-“’ no correlation between habitual intake and response to an acute dose was observed in the current study. Persistence of BP responses in our habitual users after only an overnight abstinence is consistent with the results of several other recent studies.* In summary, the lack of evidence for beneficial effects of caffeine during exercise, and its contribution to higher resting and exercise BP levels seen in the current study, argue for increased attention to its effects in specific subgroups. Our findings suggest that even in habitual users, caffeine may hinder accurate diagnosis and treatment of hypertension in some patients or risk groups. Relevant clinical questions remain: (1) whether occurrence of such high pressure states exposes these persons to significantly greater or more prolonged cardiac demand, afterload, and hemodynamic stress on vessel walls and (2) whether frequent exposure to such effects during exercise training might encourage undesired structural adaptations, such as the ventricular hypertrophy and vascular wall thickening seen in hypertensive patients.

*References 2, 3, 5, 6, 8, 43, and

44.

The authors thank Kathy McDaniel with data collection.

October 1991 Heart Journai

for her skillful assistance

ADDENDUM

Individual difference scores were used here in preference to absolute values for pre- and postdrug diastolic BPS for the following reasons: (1) the number of EXG men was limited (n = seven); (2) the predrug diastolic BPS in EXGs were lower and more variable on caffeine than on placebo days; and (3) diastolic BP variability was generally greater at all measurement periods among EXG men than among LRISK and EXN men. Given the limitations induced by post hoc identification of EXGs rather than deliberate selection of larger and equal numbers of EXG and EXN men, it is not possible to tell whether the greater variability among EXGs is random noise or reflects a natural lability that may occur in the early phases of borderline hypertension. These sources for potential confounding variance occurring prior to drug treatment were thus statistically controlled by focusing on within-subject change scores from pre- to postdosing for comparisons across risk groups and drug conditions, rather than by reliance on absolute values.

REFERENCES

1. Pincomb GA, Lovallo WR, Passey RB, Whitsett TL, Silverstein SM, Wilson MF. Effects of caffeine on vascular resistance, cardiac output and myocardial contractility in young men. Am J Cardiol 1985;56:119-22. 2. Pincomb GA, Lovallo Wk, Passey RB, Wilson MF. Effect of behavior state on caffeine’s ability to alter blood pressure. Am J Cardiol 1988;61:798-802. 3. Pincomh GA, Lovallo WR, Passey RB, Brackett DJ, Wilson MF. Caffeine enhances the physiological response to occupational stress in medical students. Health Psycho1 1987;6:10112. 4. Sung BH, Lovallo WR, Pincomb GA, Passey RB, Wilson MF. Effects of caffeine on blood pressure response during exercise in normotensive young men, Am J Cardiol 1990;65:909-13. 5. Whitsett TL, Manion CV, Christensen HD. Cardiovascular effects of coffee and caffeine. Am J Cardiol 1984;53:918-22. 6. Lane JD, Williams RB. Cardiovascular effects of caffeine and stress in regular coffee drinkers. Psychophysiology 1987; 24:157-64. Robertson D, Frolich JC, Carr K, Watson JT, Hollilield JW, Shand DG, Oates JA. Effects of caffeine on plasma renin activity, catecholamines, and blood pressure. N Engl J Med 1978;298:181-6. Smits P, Thien T, van’t Laar A. Circulatory effects of coffee in relation to the pharmacokinetics of caffeine. Am J Cardiol 1985;56:958-63. Doyle AE, Fraser JRE. Essential hypertension and inheritance of vascular reactivity. Lancet 1961;2:509-11. 10. Ewart CK, Harris WL, Zeger S, Russell GA. Diminished pulse pressure under mental stress characterizes normotensive adolescents with parental high blood pressures. Psychosom Med 1986;48:489-501. 11. Weidmann P, Beretta-Piccoli C. Dysregulation of blood pres-

Volume

122

Number

4, Part

12. 13.

14. 15.

16. 17.

18.

19.

20.

21.

22. 23. 24. 25.

26.

27.

1

sure8 in normotensive offspring of hypertensive parents. J Cardiovasc Pharm 1988;12(Suppl 3):S134-S48. Julius S. Transition from high cardiac output to elevated vascular resistance in hypertension. AM HEART J 1988;166:600-6. Aalkjaer C, Heagerty AM, Bailey I, Mulvany MJ, Swales JD. Studies of isolated resistance vessels from offspring of essential hypertensive patients. Hypertension 1987;9(Suppl III): 111-115-58. Kaplan NM. Clinical hypertension. Baltimore: The Williams & Wilkins Co, 1982:1-41. Stamler R, Stamler J, Riedlinger WF, Algera G, Roberts RH. Family (parental) hypertension and prevalence of hypertension: results of a nationwide screening program. JAMA 1979;241:43-6. Cinciripini PM. Cognitive stress and cardiovascular reactivity. I. Relationship to hypertension. AM HEART J 1986;112:104450. Falkner B, Onesti G, Angelakos ET, Fernandes M, Langmann C. Cardiovascular response to mental stress in normal adolescents with hypertensive parents: hemodynamics and mental stress in adolescents. Hypertension 1979;1:23-30. Pickering TG, Harshfield GA, Kleinert HD, Blank S, Laragh JH. Blood pressure during normal daily activities, sleep and exercise: comparisons of values in normal and hypertensive ._ subjects. JAMA 1982;247:992-6. Matthews KA. Rakaczkv CJ. Familial asnects of the Tvne A behavior pattern and physiological reactivity to stre& In: Schmidt TH, Dembroski TM, Blumchen G, eds. Biological and psychological factors in cardiovascular disease. Berlin: Springer-Verlag, 1986:228-45. McCann BS, Matthews KA. Influences of potential for hostility, Type A behavior, and parental history of hypertension on adolescents’ cardiovascular responses during stress. Psychophysiology 1988;25:503-11. Wilson MF, Sung BH, Pincomb GA, Lovallo WR. Exaggerated blood pressure response to exercise in normotensive men at high risk for essential hypertension. Am J Cardiol 1990;66: 731-6. Erikssen J, Jervell J, Karfang K. Blood pressure responses to bicycle exercise testing in apparently healthy middle-aged men. Cardiology 1980;66:56-63. Wilson NV, Meyer BM. Early prediction of hypertension using exercise blood pressure. Prev Med 1981;10:62-8. Jackson AS, Squires WG, Grimes G, Beard EF. Prediction of future resting hypertension from exercise blood pressure. J Cardiac Rehabil 1983;3:263-8. Durnin VGA, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974;33:77-97. Christensen HD, Whitsett TL. Measurement of xanthines and their metabolites by means of high pressure liquid chromatography. In: GL Hawk, ed. Biological and biomedical application of liquid chromatography science. vol 10. New York: Marcel Dekker, 1979507-37. Wilson MF, Sung BH, Herbst CP, Lee RL, Brackett DJ. Evaluation of left ventricular contractility indexes for the detection of symptomatic and silent myocardial ischemia. Am J Cardiol 1988;62:1176-9.

Caffeine, exercise, and hypertension

risk

1115

28. Wilson MF, Lovallo WR, Pincomb GA. Noninvasive measurement of cardiac functions. In: Schneiderman N, Weiss SM, Kaufmann PE, eds. Handbook of research methods in cardiovascular behavioral medicine. New York: Plenum Press, 1989;23-50. 29. Smits P, Boekema P, Abreu RD, Thien T, van’t Laar A. Evidence for an antagonism between caffeine and adenosine in the human cardiovascular system. J Cardiovasc Pharmacol 1987;10:136-43.

30. Smits P, Schouten J, Thien T. Cardiovascular effects of two xanthines and the relation to adenosine antagonism. Clin Pharmacol Ther 1989;45:593-9. 31. Berne RM, Knabb RM, Ely SW, Rubio R. Adenosine in the local regulation of blood flow: a brief overview. Fed Proc 1983;42:3136-42. 32. Costill DL, Dalsky GP, Fink WJ. Effects of caffeine ingestion on metabolism and exercise performance. Med Sci Sports 1978;10:155-8.

33. Piters KM, Colombo A, Olson HG, Butman SM. Effect of coffee on exercise-induced angina pectoris due to coronary artery disease in habitual coffee drinkers. Am J Cardiol 1985;55:27780. 34. Toner MM, Kirkendall DT, Delio DJ, Chase JM, Cleary PA, Fox EL. Metabolic and cardiovascular responses to exercise with caffeine. Ergonomics 1982;25:1175-83. 35. Schade D, Simonelli C, Standefer JC, Eaton RP. Caffeine-induced hormonal changes during running [Abstract]. Fed Proc 1979; 38:944. 36. Sasaki H, Takaoka I, Ishiko T. Effects of sucrose or caffeine ingestion on running performance and biochemical responses to endurance running. Int J Sports Med 1987;8:203-7. 37. Hirsch AT, Gervino EV, Nakao S, Come PC, Silverman KJ, Grossman W. The effect of caffeine in exercise tolerance and left ventricular function in patients with coronary artery disease. Ann Intern Med 1989;110:593-8. 38. Leon AS, Jacobs DR, DeBacker G, Taylor, HL. Relationship of physical characteristics and life habits to treadmill exercise capacity. Am J Epidemiol 1981;113:653-60. 39. Gordon NF, Myburgh JL, Kruger PE, Kempff PG, Cilliers JF, Moolman J, Grobler HC. Effects of caffeine ingestion on thermoregulatory and myocardial function during endurance performance. S Afr Med J 1982:52:644-7. 40. Ammon HPT, Bieck PR, Mandalaz D, Verspohl EJ. Adaptation of blood pressure to continuous heavy coffee drinking in young volunteers. A double-blind crossover study. Br J Clin Pharmacol 1983;15:701-6. 41. Robertson D, Hollister AS, Kincaid D, Workman R, Goldberg MR, Tung C, Smith B. Caffeine and hypertension. Am J Med 1984;77:54-60.

42. Robertson D, Wade D, Workman R, Woosley RI, Oates JA. Tolerance to the humoral and hemodynamic effects of caffeine in man. J Clin Invest 1981;67:1111-7. 43. Ratliff-Crain J, O’Keeffe MK, Baum A. Cardiovascular reactivity, mood, and task performance in deprived and nondeprived coffee drinkers. Health Psycho1 1989;8:427-47. 44. van Dusseldorp M, Smits P, Thien T, Katan MB. Effect of decaffeinated coffee on blood pressure: a 12-week, doubleblind trial. Hypertension 1989;14:563-9.