PHARMACODYNAMICS AND DRUG ACTION Dose-de~endentinhibition of the hemodynamic response to dipyridamole by caffeine I

We investigated the effects of the adenosine antagonist caffeine on the hemodynamic response to dipyridamole infusion (0.4 mgkg for 4 minutes). According to a randomized, placebo-controlled doubleblind protocol, eight normotensive volunteers each participated in five tests: placebo after placebo, dipyridamole after placebo, and dipyridamole after 1, 2, and 4 mgkg caffeine. Infusion of caffeine alone (4 mg/kg) induced an increase in mean arterial pressure of 6.1 2 0.5 mm Hg versus 1.5 2 0.9 mm H g after placebo (p < 0.05). Infusion of dipyridamole alone exerted a characteristic hemodynamic response with an increase in systolic blood pressure (+8.4 2 2.4 mm Hg), pulse pressure (+ 7.0 r 2.4 mm Hg), heart rate (+25.7 2 3.8 beatslmin) and calculated rate-pressure product (+3419 mm H g x beats per minute), all being significantly different from the changes induced by placebo. Caffeine induced a dosedependent attenuation of the response to dipyridamole, with a significant negative correlation between the dose of caffeine on the one hand (0, 1, 2, and 4 m&) and the dipyridamole-induced increments in systolic blood pressure (r = - 0.53), pulse pressure ( r = - 0.50), heart rate ( r = - 0.95), and ratepressure product (r = - 0.93) on the other hand. We conclude that caffeine attenuates the hemodynamic response to dipyridamole infusion in humans in a dose-dependent fashion. Because of the widespread use of caffeine, this pharmacologic interaction may be of clinical importance, for example, in the THER diagnostic use of dipyridamole in thallium-201 myocardial imaging. (CLIN PHARMACOL 1991;50:529-37.)

Paul Smits, MD, Carin Straatman, MD, Evelien Pijpers, MD, and Theo Thien, MD Nijmegen, The Netherlands

Dipyridamole is a widely prescribed drug in Westem society, especially because of its putative beneficial effect on platelet aggregation and on the coronary blood flow.' In the last decade the intravenous formulation of dipyridamole has gained a growing interest, not as a therapeutic agent but for diagnostic purposes in clinical cardiology.2 In patients with a structural From the Division of General Internal Medicine, Department of Medicine, University Hospital Nijmegen. Received for publication March 7, 1991; accepted July 15, 1991. Reprint requests: Paul Smits, MD, Division of General Internal Medicine, Department of Medicine, University Hospital Nijmegen, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. 1311132418

coronary stenosis, dipyridamole infusion may result in a reduction of subendocardial flow mediated by a coronary steal phenomenon or by a decreased perfusion pressure.3 The regional dipyridamole-induced reductions in coronary flow can subsequently be visualized by thallium-201 myocardial scintigraphy. These socalled dipyridamole-thallium-201 tests can be considered as a valuable diagnostic aid for the evaluation of coronary artery disease in patients who are unable to perform physical exercise t e ~ t i n ~ . ~ . ~ In a case report, we have recently shown that administration of a relatively high dose of caffeine (250 mg, roughly equivalent to 2 to 3 cups of coffee) before dipyridamole- thallium-20 1 scintigraphy induced a false negative test r e ~ u l t We . ~ suggested that this

CLIN PHARMACOL THER NOVEMBER 1991

530 Snzits e t al. Schedule of measurements infusion 1

infusion 2

caffeine

placebo1 dipy ridarnole

n placebo I

0

v

v

Blood sample

..... .

Blood pressure 1 1 1 1 1 1 1 Heart rate

equilibration

I

o

I

I

I

20

LO

60

I

ao

I

loo

I

120

Time ( m i n )

Fig. 1. The schedule of measurements of the tests in which dipyridamole (0.4 mglkg in 4 minutes) or placebo was infused after previous intravenous administration of 0 (placebo), 1, 2, or 4 mg/kg caffeine.

may be caused by a pharmacologic interaction at the level of vascular adenosine receptors because the mechanism of action of dipyridamole may be liased on an increase in endogenous plasma adenosine levels5 and because caffeine has been shown to antagonize vascular adenosine receptor-mediated processes in hum a n ~ . ~ Because ,' caffeine is a widely used stimulant, the aforementioned interaction between caffeine and dipyridamole may be of importance in clinical practice. Therefore we decided to explore this pharmacologic interaction more thoroughly in a randomized, double-blind placebo-controlled design. In this study, we addressed three main questions: What are the hemodynamic effects of dipyridamole infusion? Are the hemodynamic effects of dipyridamole infusion attenuated by caffeine and, if so, what is the dose-response relationship of this caffeine-mediated inhibition? Do low dosages of caffeine already induce a relevant reduction in the dipyridamole-mediated effects?

SUBJECTS AND METHODS After approval of the protocol by the local ethics committee of the University Hospital Nijmegen, eight normotensive nonsmoking volunteers (five men, three women) gave their written informed consent for participation in the study. The mean values ( k S D ) for age, weight, height, and body mass index amounted to24.4 3.3years,76.1 2 9 . 9 k g , 1.79 k 0 . 1 1 m, and 23.6 1.9 kg/m2, respectively. All individuals

*

*

were accustomed to daily caffeine intake, and their mean use of coffee averaged 4.1 k 2.3 cups of coffee per day. Each volunteer participated in five tests. Because of the relatively high terminal plasma half-life of dipyridamole,' these tests were separated from each other by at least 4 days. In four of the five tests, dipyridamole was infused in a dose of 0.4 mg/kg for 4 minutes after previous intravenous administration of placebo (NaCl 0.9%) or 1, 2, or 4 mg/kg caffeine (dissolved in NaCl 0.9%) for 10 minutes. In one of the five tests, only placebo (NaC1 0.9%) was infused instead of caffeine and instead of dipyridamole. All tests were performed in a randomized order, and caffeine and placebo were administered in a double-blind fashion. Because of the yellow color, the administration of dipyridamole was blinded for the subjects only by covering the infusion system, but this was not sufficient to keep the procedure blinded for the investigator. The subjects were instructed to abstain from caffeine for 36 hours (no coffee, tea, cacao, chocolate, or cola-flavored drinks) and from alcohol for at least 12 hours before each test. The main study parameters were systolic and diastolic blood pressure, recorded noninvasively by means of an Arteriosonde 1225 blood pressure device (Roche Medical Electronics Division, Hoffmann-La Roche Inc., Orangeburg, N.Y .), and heart rate, calculated from an electrocardiographic registration. Further, plasma caffeine concentrations were measured by a reverse-phase H P L C . ~The sub-

VOLUME SO NUMBER 5, PART 1

Cafeine inhibits dipyridavnole response

53 1

Blood pressure (mmHg )

Heart rate (bpm 1

00 PL-PL Qe

PL-DIP

I 0 4 8 12 26 34 Time (min)

Fig. 2. The course of systolic and diastolic blood pressure and heart rate before and after placebo

infusion (PL), and before, during, and after subsequent infusion of placebo (PL) or dipyridamole (DIP, 0.4 mgikg). Data are mean values 5 SEM.

jects remained in the supine position throughout all measurements. In the left antecubital vein, a cannula was inserted for both drug infusion and blood sampling. Fig. 1 presents the schedule of measurements during the tests. In short, baseline recordings of blood pressure and heart rate were made after an equilibration period, and blood was sampled for determination of the plasma caffeine concentration. Subsequently, placebo or caffeine (1, 2, or 4 mg/kg) was infused, and 15 minutes after ending this infusion, blood pressure and heart rate recordings were started at 2-minute intervals. Thirty minutes after the placebolcaffeine infusion was ended, the administration of 0.4 mglkg dipyridamole (or placebo) was started. Blood pressure and heart rate recordings continued until 30 minutes after this second infusion. During and just after the in-

fusion of dipyridamole, the electrocardiogram was registered continuously to calculate the heart rate four times per minute. A second blood sample was taken after both infusions to determine the plasma caffeine concentrations reached. This sample was taken 50 minutes after completion of the caffeine infusion; meanwhile, the infusion system was perfused with dipyridamole and/or saline (Fig. 1). Therefore we can be sure that this second blood sample for the determination of the plasma caffeine concentrations has not been contaminated by the previously administered caffeine solution. Calculations and statistical analysis. For statistical analysis, the baseline recordings and the measurements after the first but before the second infusion were both averaged to one mean value (Fig. 1). Because the heart rate response to dipyridamole appeared

CLIN PHARMACOL THER NOVEMBER 1991

532 Snzitsetal.

Table I. Mean ( 2 S E ) of the systolic (SBP) and diastolic blood pressure (DBP) and of the heart rate at each representative time period of the study: (1) baseline, before the first infusion (placebotcaffeine), (2) after the first infusion but before the second infusion (placebo/dipyridamole) and (3) after the second infusion Drugs

Before jirst infusion

After jrst infusion

After second infusion

Placebo-placebo SBP (mm Hg) DBP (mm Hg) Heart rate (beatslmin) Placebo-dipyridamole SBP (mm Hg) DBP (mm Hg) Heart rate (beatslmin) Caffeine (1 mg1kg)-dipyridamole SBP (mm Hg) DBP (mm Hg) Heart rate (beatslmin) Caffeine (2 mg1kg)-dipyridamole SBP (mm Hg) DBP (mm Hg) Heart rate (beatslmin) Caffeine (4 mg1kg)-dipyridamole SBP (mm Hg) DBP (mm Hg) Heart rate (beatslmin) For the heart rate, the data after the second infus~onrefer to averaged values from the fourth to the eighth minute after starting the dipyridamole mfusion. For the SBP and DBP, these data refer to the averaged values over the entire post-dipyridamole infusion (fourrh to th~rty-fifthminute after the start of dipyridamole). The results are given for all five tests.

to reach maximal values between the fourth and the eighth minute after the dipyridamole infusion was started and because this maximal response was of rather short duration (Fig. 2), the data on heart rate were averaged to one mean value during this particular period for each test. For analysis of blood pressure, the data from the fourth to the thirty-fifth minute after the start of dipyridamole infusion were all averaged to one mean value for each test because the effects of dipyridamole on blood pressure lasted until the end of the test (Fig. 2 ) . The pulse pressure was calculated as the difference between the systolic blood pressure and the diastolic blood pressure. Mean arterial pressure was calculated as the sum of the diastolic blood pressure and one third of the pulse pressure. The rate pressure product, a parameter that may reflect the oxygen demand of the myocardium, was calculated by the product of systolic blood pressure and heart rate. The effects of the several dosages of caffeine alone were analyzed by comparison of the caffeine-induced changes of the five tests in an overall nonparametric analysis of variance (ANOVA, Friedman analysis).

''

Further, Spearman rank correlation coefficients between the dose of caffeine (0, 1, 2 , and 4 mglkg) and the caffeine-induced effect were calculated for each individual, and subsequently compared with zero by a paired Wilcoxon test. The hemodynamic response to dipyridamole alone was analyzed by comparison of the results of the placebo-placebo test with those of the placebodipyridamole test by means of the paired Wilcoxon test. The interaction between caffeine and dipyridamole was analyzed by comparison of the dipyridamole-induced effects (corrected for placebo) of the four tests in which dipyridamole was administered after 0, 1, 2 , or 4 mglkg caffeine by means of a nonparametric ANOVA (Friedman analysis) and again by analysis of the correlation between the dose of caffeine and the dipyridamole-induced effect. In general, a dose-dependent effect of caffeine was thought to exist if both the overall Friedman analysis and the analysis of the Spearman correlation coefficients showed significant results. For all tests, differences were considered to be statistically significant at p values of less than 0.05 (two

VOLUME SO N ~ M B E R5, PART 1

Cufeine inhibits dipyidurnole response 533

ASBP (mmHg1

APP ImmHg)1 -

ADBP (mmHg 1 101

AHR (bpm 1

APPlmmHg) ----

ASBP (mmHg 1

1

AMAP (mmHg ) lo

I AMAP (mmHg 1

APRP(mmHg.bpm)

rNS i

(NSI

Fig. 3. Mean (?SEM) changes (A) of systolic and diastolic blood pressures (SBP, DBP), mean arterial pressure (MAP), pulse pressure (PP), and heart rate (HR) as a result of the infusion of placebo (PL, and PL,) or 1, 2, or 4 mglkg caffeine (IC, 2C, and 4C) before the administration of dipyridamole. p Values refer to the nonparametric analysis of variance (Friedman analysis) on the changes induced by the several interventions. NS, Not significant.

PLPL

PLDIP

1 CDIP

2CDIP

LCDIP

PLPL

PLDIP

1CDIP

2 C - LCDIP DIP

RESULTS Hemodynamic effects of caffeine alone. Table I

Fig. 4. Mean (kSEM) changes (A) of systolic and diastolic blood pressure (SBP, DBP), mean arterial blood pressure (MAP), pulse pressure (PP), heart rate (HR), and rate pressure product (RPP) induced by infusion of placebo (PL) or dipyridamole (DIP, 0.4 mglkg) after previous administration of placebo (PL) or 1, 2, or 4 mglkg caffeine (IC, 2C, and 4C). p Values refer to the nonparametric analysis of variance (Friedman analysis) on the dipyridamole-induced changes after 0, 1, 2, or 4 mglkg caffeine. * p < 0.05 (paired Wilcoxon test referring to the comparison between the effects of dipyridamole alone versus placebo). NS, Not significant.

presents the averaged values of systolic blood pressure, diastolic blood pressure and heart rate at each representative period for the five tests. Fig. 3 summarizes the hemodynamic changes induced by placebo (two times) and by the several dosages of caffeine. After the highest dose of caffeine, systolicldiastolic blood pressure was increased by 2.9 t 1.317.8 t 0.9 mm Hg versus -0.5 2 1.312.5 2 1.3 mm Hg after placebo, whereas heart rate changed with -4.2 t 1.3 versus - 1.9 t 0.9 beatslmin. Significant differences between the five tests (Friedman analysis) were only observed for the diastolic blood pressure and the mean arterial pressure. The baseline plasma caffeine concentrations were

below the level of detection (