THE EFFECT OF CAFFEINE ON CARDIAC VULNERABILITY TO TACHYARRHYTHMIA IN THE RAT

• • • • AC KNOl~L EDGEMENTS • • TABLE OF CONTENTS. LIST OF TABLES • LIST OF FIGURES. TABLE OF CONTENTS

In this study the term caffeine administration on the the heart to ventricular fibrillation Vulnerabil tty to cardiac arrhythmia in the adult male rat was determined by measuring the ventricular tachycardia threshold (VTT). The VTT was used as an index of the vulnerabil tty of larger hearts to fibrillation. It was defined as the minimum current necessary to generate sustained ventricular tachycardia, and was tested by applying trains of pulses directly to the right ventricular epicardium through bipolar platinum electrodes.
The tratn consisted of constant current, 70 Hz, monophaslc, 2 msec rectangular pulses. It was delivered 10 msecs after the right atrial pactng pulse and lasted 90 msecs so as to end before the apex of the T wave of the electrocardiogram.
In rats receiving a stngle oral injection of caffeine (30 or 90 mg/kg) VTT was 25 per cent lower than the VTT of water injected rats (p <0.05). The oral injection of water reduced VTT by 35% when compared with untreated controls. The caffeine related reductlon was in addition to that associated wtth the tnjectlon process.
Corresponding tachycardia suggested that the injection process, in addition to the caffeine administration, was associated with sympathetic discharge which presumably medl~ted the reduction in threshold.
Daily or.al treatment with caffeine (in the same doses) for 2,4,8 or 10 weeks caused a time related reversal of the acute effects of caffeine injection.
Following 2 or 4 weeks of treatment, acute caffeine administration no longer was associated with a reduction in VTT. After 8 or 10 weeks of chronlc caffeine administration, acute caffeine injection was associated with thresholds more than twice as great as water treated controls. This relative increase In VTT following chronic caffeine was dose dependent. The increase in VTT following chronic caffeine admini~tration was the same magnitude whether the rat was injected just prior to VTT I i testing with water or with caffeine.
Caffeine injection (90 mg/kg/day, p.o.) for a period of 10 weeks attenuated the maximal chronotropic effect of isoproterenol. This result suggested a reduction in cardiac beta adrenerglc receptor concentration. The reduction Jn the vulnerabll Jty of the rat ventricle to tachyarrhythmia associated with chronlc caffeine administration, may be due to the same reduction In the cardiac adrenergic receptor population.  Since then, there has been a continufng controversy over the adverse effects of coffee lntake on the incidence of CHO and its subsequent lethality. Problems inherent to clinical studies such as subject selection and uncontrolled varfables seem to be responsible, at least in part, for the often conflicting results concernfng this relat i onship.
In the more than half dozen clinical studies found i n the 1 iterature since the early 1960's, the risk factor of CHD associated with coffee consumption has ranged from no increased risk (Dawber, Kanne'l and Gordon, 1974) to a two fold increase In the risk of acute myocardial fnfarction ( Ml) seen in coffee drinkers (greater than s i x cups per day) as compared to non coffee drinkers (Jick et~ 1973). The motivation for this laboratory study was to examine, in part, the overall question of the effect of coffee consumption on mortality due to sudden card i ac death.
Strict controls used in laboratory studfes were adhered to, with the intent of avoiding some of t he difficulties confounding the clinical studies.
It is known that the majorfty of persons dying from page 2 acute Ml do so suddenly prior to hospitalization (Kannel g,t. ~ 1975). Furthermore, in studying this disease state, Lown and his co-workers (1 969) have suggested that ventricular fibrillation (VF) is one of the primary causes of early sudden death associated with Ml.
It has also been reported that sudden death due to arrhythmia (generally V.F. or cardiac arrest) is often seen even in those subjects with no overt pathological evidence of acute Ml (Roberts and Buja, 1972). As a result it susceptible is not possible to predict which people are to this terminal event based solely on the status of their coronary circulation.
To understand the magnitude of this problem, it must be pointed out that estimates of the incidence of sudden death, without regards to etiology, ranges as high as 30 per cent of all deaths (Kannel .fil:. ~ 1975).
Accordlngly it seemed important to determine what effect caffeine, the often touted pharmacologically active agent in coffee, would have on the susceptibility of the heart to ventricular fibrillation.
Bellet anrl his co-workers studied caffeine's acute effect on cardiac vulnerabll ity in t he do g (Bel let  Brooks t l .a.L.. Surawlcz (1971).
Part of the reason for the continued acceptance of both of these constructs is the severe difficulty in identifying specific mechanisms causing fibrillation. Surawicz (1971) lists three reasons for this difficulty: (1) the lmpossibil ity of local izlng the fiber or group of fibers in which the first wave of reentry begins; (2)  Later,  demonstrated that under 1 f ght chloroform anesthesia, epinephrine could initiate VF.
It was not until the late 1 930 1 s, however, that fibrillatory agents were used to quantltatively assess the vulnerability of the heart to arrhythmia. Meek and his co-workeri (1937) used test doses of epinephrine to assess vulnerability of dog heart to arrhythmia after treating the animals with several different volatile anesthetics. Shen and Simon (1938) seem to be the first investigators, using non-electrical technlques, to demonstrate a reduction in susceptibil tty to VF after treatment with an antiarrhythmlc agent, namely procaine. Blumenthal and Tribe -Oppenheimer (1939) su ggested that a quantitative assessment of vulnerability could be made by determinin g the amount of barium chloride whtch was just sufficient to fibrillate the perfused cat's heart.
Since the work of these early investigators, the use of non-electrical techniques have proved to be very useful as a means of studying both the pathophysiology of VF and the heart's susceptibility to the arrhythmia.
Mechanistically, the use of these non-electrical techniques relate to their capabillty to increase cardiac automaticity and/or to increase the abil fty of the heart to sustain reentry. Aconotine, first used by , increases automaticity when topically appl led to the epicardium (Schmidt, 1960 andMatsuda~~ 1959 a matter of minutes to decrease that susceptlbil ity (Han _g_t ~ 1964 andMoore and. Finally, It is the nature of the non-electrical techniques that the arrhythmias they initiate are not readf ly reversible.
Again, depending upon the question under investigation, this characterlstic can either be an advantage or dlsadvantage.

JL. Electrical Fibrillation Threshold Technlques
The electrical fibrillation threshold ls the minimal current required to induce cardtac flbrillatfon (Szekeres and Papn, 1971 There are at least three characterf stics of premature beats which contribute to the reentry process.
The first is their relatively slow conduct i on velocity as compared to normal sinus beats. Secondly, premature beats tend to cause an increase in the normal temporal dispersion of refractoriness  This shortened ERP is, most probably, one reason for the prolonged vulnerable per i od duration of premature beats.
In summary, deliver i ng a sin g le electrlc pulse to page 27 the myocardium el I cits a premature extrasystole. As the stimulation current increases, the electrical recovery from the premature response becomes more asynchronous.
At some critical stage this asynchrony reaches the point where multiple reentry is made possible and VF ensues.
Many of the early workers in this field (i.e. prior to the 1gso's) did not pace the heart via a second stimulator. This difficulty is particularly relevent to this thesis since Lubbe and co-workers (1975) reported that isolated perfused rat heart had a mean vulnerable period of only 3 msec.
Another disadvantage of the single shock method is that the thresholds are somewhat higher than those measured using the trains-of-pulses method (Tamargo~ ~ 1975). High intensity shocks, especially when applied in each cycle, are capable of injuring cardiac tissue.

) The train of pulses method
The most recently developed method of electrical fibrillation threshold testing has been the train of pulses method. Han (1969) was the first to use this technique of applying a gated train of monophasic, rectangular pulses to the heart to assess vulnerability.
In that first use, the train was initiated just after the R wave of the cardiogram and was sustained just to the peak of the T wave. Each pulse was 2-4 msec in duration, and the pulses were delivered at a rate of 100 cycles per second. Current was fibrillation ensued; progressively increased until that current was then taken to be the fibrillation threshold.
One Interesting result from this study was that the time lnterval between elicfted extrasystoles became progressively shorter at increasing stfmulatlon currents.
As a result, it was possible to correlate the ventricular fibr'illation threshold (VFT) wi th a minimal "critical" coupling interval. This finding had practical applicability in that it was used in that same study as a measure of vulnerability.
page 35 Using trafns of increasing duratfon, but constant current, susceptJbil tty to VF was estimated by the duration of the train, the number of acceleratfng premature beats required to induce fibrlllation and the critical cycle length of the premature beat just preceding fibrJllation (i.e. the interval from the previous beat to that extrasystole just precedJng VF).
Because of its ease of application and, most notably the rapidity with which VFT can be measured, the traJn of pulses method has found widespread use Jn the past seven years (Moore and Spear, 1975 Although not a train of pulses method, the sequential R on T technique developed by Thompson and Lown (1972) and separately by Gamble and Cohn (1972)  shown that beta blockade is ineffective in regucing reperfusion associated arrhythmias (Lown _g_t ~ 1977).
lschemic and infarcted tissue also directly alters conduction velocity and causes disparity in refractory period duration within adjacent areas of myocardium.
Such effects directly predispose the heart to arrhythmfa (see Vulnerability To Ventricular Fibrillation).
Development of collateral circulation prevents these effects by 1 imiting the ischemic zone slze.
Other factors reported to increase vulnerability to VF are 1 isted in Table I    Tea, an infusion of tea leaves (Thea sinensis) also contains caffeine as well as theophyll ine. Both caffeine and theobromine are obtalned from cocoa, a product of Theobroma cacao seeds. Caffeine is also found in certain soft-drink sodas, particularly those ln the cola family (from the nuts of the Cola acuminata tree) (Ritchie, 1975 Burg, 1975 (Galli and Spagnuolo, 1975). The liver concentrates caffeine to a slight extent in that the ratio of tissue water caffeine to plasma water caffeine is 1.11 (Axelrod and Reichenthal,19S3). This concentration effect presumably indicates the site of biotransformation of the compound.
In man, 99 per cent of ingested caffeine is biotransformed prior to excretion. The major products of biotransformation are 1-methyluric acid and 1-methylxanthine. Excretion of these products are in relatively equal amounts (Ritchie, 1975 As a result of decreased catecholamine synthesis, MAO inhibition ls perceived to reduce turnover rates (Carlson g_t ~ 1960 andCosta, 1966  The effect of caffeine on cardiac catecholamine content and turnover has been studied by Berkowitz and Spector (1971 Possible molecular mechanisms of this activity will be presented in subsection C of this discussion of caffeine's pharmacology.
In 5 normal patients, coffee consumptfon has been shown to produce cardiac effects which are very much in accord with the action of caffeine described above. The elevation of Intracellular c AMP levels following phosphodiesterase Jnhibltion have been correlated with the positive inotropic effects of the methylxanthines (Kukovetz and Poch, 1975). Phosphodi~sterase inhibition by the methylxanthines also potentiate the cardiac inotropic effects of epinephrine, histamine and glucagon all of which stimulate the synthetic enzyme, adenyl cyclase (Rall andWest, 1963 andMc Neill andMuschek, 1972 The catecholamines do not produce this inhibition.
Caffeine inhibition of adenosine's effects ls a third molecular mechanism by which it may act. Adenosine shortens action potential duration, reduces contractile force, slows heart rate and causes coronary vasodllation (Chiba ~ .aL.....c. 1973 andThorp andCobbin, 1967). The effects of adenosine and other purines have been suggested to be related to stlmulatlon of a 'purnerglc receptor' (see review by Burnstock, 1972 were to be tested that day, the order of testing would be varied in one of the following ways: water-caffeine-water-caffeine, caffeine-water-caffeine-water, caffeine-water -water-caffeine or water-caffeine-caffeine-water. A similar sequence would be followed if 2 rats treated with 30 mg/kg caffeine and 2 rats treated with 10 mg/kg caffeine were tested on a sTngle day. Wf thin ten minutes of oral injection the rat was anesthetized with urethan (1300 mg/kg, i.p. in a 20% w/v solution). A surface lead I I electrocardiogram was recorded on a Grass Model 7 Polygraph and displayed on a Tektronix model 5103 N oscilloscope. Tracheotomy was then performed and positive pressure respiration instituted using a Narco V-100 Kg respirator. Bilateral thoracotomy at the fourth and fifth lntercostal space followed, with removal of the fourth rib bilaterally.
The pericardium was incised, the heart exposed and the page 66 lungs retracted. Room temperature normal saline (0.9% Na Cl) was dripped directly onto the epicardium at a rate of 40-60 drops/minute throughout the experiment.
Bipolar silver-silver chloride electrodes were then placed on the right atrial appendage to provlde right atrial pacing. These electrodes were 0.025 Inches in diameter and were plated prior to use by placing several pure silver electrodes In a 5% Na Cl solution and passing 3 volts of cathodal D.C. current through them. The anode was a single silver electrode also placed In the beaker of salt water (Sidowski, 1966). Following the plating of these electrodes, they were soldered to electrically shielded teflon wires and coated with epoxyl ite for insulation. Just the tips of the electrodes were sanded so that electrode surface area could be estimated to be 0.0032 square centimeters. They were pressed a gainst the myocardium so that there was no appreciable movement about the epicardial surface. However, the pressure was not enough to displace the entire heart.  Han, 1969). Figure 1 provides a block diagram illustrating the electrical components for VTT testing. Right atrial pacing was instftuted using a Grass S 9 stimulator. The RF isolated output from the pacing stimulator was connected to a custom built, semiautomatic shut off device, then to a constant voltage to constant current converter and current measuring junction box. The output from the junction box then paced the right atrium through the bipolar silver-silver ch 1 o rf de e 1 e c t rode s p 1 aced on t he a n i ma 1 1 s r i g h t a t r i um • The test pulses of the t ra 1 n we re generated by a Grass SD 9 stimulator which was ultimately triggered from the pacer 'synch out' jacks ( The design is discussed in the text. With each pacing pulse, a train would be generated by th~ Grass SD 9 stimulator. However, Its output to the rat's right ventricle was short circuited to ground by a normally closed hand-held switch ( #2 )(see Figure 3).
As with the pacing pulse, the test train was converted from a constant voltage to a constant current source using a 1000 ohm series resistance (Shumway~~ shut off device (Figures 1,2 and 4). Activation was initiated by depressing hand held switch #1. The next pacing pulse and test train was then delivered to the rat.
Output from the Grass S4, "shut off" stimulator was The pacing pulse was a 3 msec, monophasic rectan g ular pulse usually 0.5 marnps i n stren g th.
Following 6-10 paced beats, the test train was delivered to the right ventricle by depress i ng hand held switches page 74 #l and #2 In rapid succession.
The pulse train consisted of 2 msec, monophasrc rectangular pulses delivered at a frequency of 70 Hz. It was delayed 10 msec after the pacing pulse and continued for 90 msec. Thus, 6 pulses were delivered during the train.
In relation to the electrocardiogram, the train was initiated just after the P wave and term?nated just prior to the T wave ( Figure 5). This placement assured that the train intersected the vulnerable period of the cardiac cycle, but did not encroach upon the protective zone found just after the vulnerable period.
Initially, the train was set at 2 mamp. If no sustained ventricular tachyarrhythmia developed at this current strength, pacine was reinstituted. Test current was increased by 2 mamp and following 6-10 paced beats, the train was again delivered to the ventricular epicardium. Sustained ventricular tachyarrhythmia was defined electrocardiographically, by a substantial reduction in mean arterial blood pressure, and by visual identification of a fibr?llatory quiverin g of t he ventricles.
The arrhythmia had to be sustalned for two to three seconds ( Figure 6). The ventricular tachycardia The upper trace in each pair is the lead II electrocardiogram.
The lower trace is a moniror of the output of the SD 9 Test Pulse Generator. The pulse is only delivered to the rat following the third paced beat.
Train characteristics are described in the text.
It is delayed 10 msec after the pacing pulse and ends well before the T wave. In the lower pair of traces current was Increased to 12 mamps and VT was produced. Current in the top trace was 6 mamp. The rhythm~c ventrtcular complexes of rapid rate and the fall t-n mean arterlal blood pressure from 90 to 35 mm Hg ts charactertsttc of VT. The VT threshold was the min rmum current wh i'ch generated this cardtographic , pattern, sustained for a mtnimum of 2 seconds: The middle trace indicates time (in seconds). Furthermore, barhitone has been shown to reverse and prevent cardiac arrhythmia assocf ated wlth intracerebroventricular caffeine injection  In one group caffeine was administered for 10 weeks (90 mg/kg/day), but the rats were challenged with distilled water immediately before vulnerability testing.

Effect of Chronic Caffetne
Finally, one group was given no treatment prior to anesthesia and VTT determination.
Of the 138 rats initially included in the VTT study, no data was recorded from eight. Two chronically treated rats died immediately following an oral administration of caffeine. Death followed violent major motor seizures. Three other rats died from anesthesia at the time of VTT testing, and two animals who had severe respiratory infection were discarded from the sturly. All those animals who were not included in the study were The data presented in Figure 8 shows that there was no significant correlation between these parameters. Since these results suggest that threshold did not increase with time following anesthesia, no detailed analysis was subsequently performed · of the time at which threshold was measured. a.. Initial VT thresholds following an oral injectfon and subsequent anesthesia are shown not to be correlaterr with the interval between anesthesia administration and VTT determination. This data is from rats treated with water for 10 wee~s.  There was no dose dependent trend in t he effect of chronic caffeine treat ment on heart rate. Two way analysis of variance on all of t he heart rate data also VTT thresholds of rats treated with caffeine (90 mg/kg/day) were compared with rats treated with water for the same periods of time. Single acute treatment is labelled as 1 0 weeks 1 • Caffeine treated rats were challen ged with caffeine. W ater treated rats were challenged with water, Means+ SEM for groups of 6 rats. VT thresholds of rats administered a sin gl e injection of water or caffeine just prior to VTT determin ation.
Each point is the mean threshold of 6 rats.  * Water or caffeine (10, 30,   and that they also interact in some additive fashion. The least significant difference was calculated from the analysis of variance data and applied to analyze differences in VTT between groups (Table V) between elevation of sympathetic tone to the periphery and caffeine -induced reduction in VTT. This assessment is also in agreement with the finding that beta adrenergic blockade prevented the acute reduction in threshold by caffeine (Bel let et~ 1972).
The oral injection of water was also associated with a reduction in VTT. The magnitude of the reduction was significantly less than that associated with caffeine ingestion. Thus the caffeine related reduction in threshold was in addition to that caused by the injection process.
The acute injection process may be associated with repetitive response was reduced to 67% of control values when dogs were placed in a stressful envlronment.
Selective beta-1 adrenergic blockade with tolamolol hydrochloride prevented this reduction in threshold associated with psychologic stress (Lown ~nd Verrier, 1976). Increased ventricular vulnerability in the setting of sympathetic increased stress thus appears to also be mediated by nervous system activity. Sympathetlc tone, as a result of the injection process, would be further augmented by caffeine administration. This combined activation of the sympathetic 1 imb of the autonomic nervous system seems to mediate the additive reduction in VTT by caffeine and the Injection process.
The reduction in VTT following a single caffeine injection is maximal at a dose of 30 mg/kg. The tachycardic effect of acute oral caffeine administration is also maximal at this same dose. At 10 ~g/kg, caffeine does not significantly alter either of these parameters.
Rough conversions of these doses to human equivalent doses (with rat considered one tenth as sensitive as man) suggest that the caffeine content of as 1 ittle as 2 cups of coffee may increase the vulnerability of the human heart to arrhythmia. A more conservative estimate of the conversion factor Ci .e. no difference in sensitivity) would predict that this dose in rat represents that found in 21 cups of coffee (30 mg/kg X 70 kg body wei gh t /100 mg/cup).
Based by Vitello and page 105 on the rough conversion factor reported Woods (1975), this dose represents the caffeine content in 10-15 cups of coffee.
Daily subacute administration of caffeine (for 2 or 4 weeks) prevented the reduction in VTT associated with caffeine injection just prior to threshold measurement.
Tolerance occurs to many effects of the drug (see Literature Review).
In a similar manner subacute administration of caffeine may make the rat tolerant to the VTT reducing action of caffeine. Four week treatment with oral water injections produce rats whose thresholds are the same as untreated controls (Table V). Thus tolerance has developed to the injection process as well.
Stress adaptation may play a part in this demonstration of tolerance to the effect of the injection process and caffeine administration on ventr i cular vulnerability. Following 32 days of restraint stress, male albino rats demonstrated adaptation to the effects of stress on brain serotonin content, norepinephrine content, blood corticosterone levels, and various organ weights (Rosecrans and De Feo, 1965). The physical and psychologic stress related to the inject i on process may well be the same as that associated wi th restraint stress. The action of caffeine administrat i on to increase brain norepinephrine turnover rates and decrease brain serotonin rates (see Literature Revie w), would be page 106 construed as Increases in brain serotonin levels and a reduction in NE levels -i.e. the same effects associated with restraint stress (Rosecrans and De Fee, 1965). Thus caffeine administration may, to a certain extent, be interpretted as producing a stress reaction. Tolerance to the effect of caffeine on VTT may therefore be due to pharmacologic tolerance (i.e. the increased metabolism and excretion or decreased effect of the drug at the receptor) and/or adaptation to stress associated with caffeine administration.
Subchronic caffeine treatment (90 mg/kg/day) for 8 or 10 weeks was associated with a t wo fold increase in VTT when compared with that of control rats treated with oral water injection for the same periods of time (Fi g ure 10). Smaller doses of caffeine produced correspon d in g ly less elevation in threshold ( Figure 11). This reversal of the VTT reducin g effect of acute caffeine i nj ection suggests that chronic caffeine administration causes some radically different effect than does the acute or even subacute administration of the dru g . This effect of chronic caffeine administration on vulnerability cannot be interpreted merely as tolerance, since to l erance implies compensation which prevents t he a gent from producing its normal effect. In this case caffeine produces an elevation in VTT, re gard l ess of t he challen ge dose, thus su gg estin g an effect which is just the page 107 opposite of that seen with single injection. Chronic administration may reduce cardiac vulnerability by altering certain receptor sites.
Chronic caffeine treatment was shown to reduce the maximal chronotropic effect of lsoproterenol, but not to shift the dose-response curve ( Figure 13).