- Top of page
- MATERIALS AND METHODS
Summary: Purpose: To investigate the consequences of caffeine consumption on epileptic seizures, we used the pilocarpine and the kainate models of epilepsy. We hypothesized that prolonged caffeine consumption or its withdrawal would alter adenosine levels and hence alter seizure susceptibility.
Methods: We administered a 0.1% caffeine solution in the drinking water of adult male Wistar rats over a 2-week period. We challenged another group of animals with the same doses of pilocarpine or kainate 12 h after the withdrawal of the same caffeine-administration protocol.
Results: This did not alter the threshold for the induction of seizures by a subconvulsant dose of pilocarpine (200 mg/kg, i.p.) or kainic acid (8 mg/kg, i.p.). Similarly, challenging another group of animals with the same doses of pilocarpine or kainate 12 h after the withdrawal of the same caffeine-administration protocol did not lead to any significant changes in seizures.
Conclusions: With the pilocarpine model of epilepsy, we were not able to find any significant difference in seizure profile that could stem from either caffeine administration or its withdrawal. Despite the extensive laboratory evidence on the convulsant properties of xanthine derivatives in animal models of epilepsy, such strong evidence is lacking in clinical settings. Our current findings with the administration of caffeine at doses similar to those of daily life both support and confirm the clinical experience.
Caffeine is one of the central nervous system (CNS) stimulants most widely present in human diet (1). This methylxanthine is found in appreciable concentrations in our common beverages such as coffee, tea, and soft drinks (2). It has proconvulsant effects on seizures in rats induced by kainic acid or pilocarpine (3–5), induces seizures in genetically epilepsy-prone rats (1), and prolongs kindled seizures in rats (6,7). In humans, scarce evidence of the convulsive effects of caffeine exists, with only two case reports indicating a clear association between excessive caffeine ingestion and increased seizure frequency (8) and caffeine withdrawal and the occurrence of a tonic–clonic seizure (9).
Several studies demonstrated that the CNS effects of methylxanthines might be linked to their ability to antagonize the actions of endogenous adenosine by blocking its receptors (10–12). The population of receptors involved on adenosine activities is heterogeneous, but the main mechanism of action of caffeine is represented by the antagonism of adenosine A1 and A2a receptors (13). A1 receptors are plentiful in the hippocampal complex, especially at CA1 and CA3 sectors and molecular layer of dentate gyrus, sites related to spontaneous recurrent seizures in pilocarpine models (14–16). Evidence exists that adenosine causes an inhibition of the release of different types of excitatory transmitters, such as glutamate and aspartate (17), and it has been suggested that adenosine is released during seizures, providing an inhibitory tone in the mammalian nervous system (18). Thus adenosine may function as an endogenous anticonvulsant (18–23).
Caffeine consumption is extremely common, yet surprisingly little attention has been paid to the brain biochemical effects of its administration and mainly its withdrawal. Caffeine intake varies worldwide according to the population studied. Overall, caffeine consumption can be estimated at ∼70–76 mg/person/day [i.e., 1–2 cups of coffee per day (24)], but reaches 210 to 238 mg/day (3–4 mg/kg/day) in the United States and >400 mg/person/day in Sweden and Finland (25–27).
Prolonged dietary consumption for 2 weeks of 0.1% caffeine ad libitum increased the plasma adenosine concentration in rats compared with that in control animals drinking tap water. Caffeine discontinuation had the opposite effect, that is, when the caffeinated solution was discontinued and replaced with tap water on the evening before the measurement, the plasma adenosine concentration declined compared with that in a second control group consuming tap water. Intravenously administered caffeine also increased plasma adenosine concentration in a dose-related and saturable manner, but long-term caffeine administration was of greater magnitude than that after acute intravenous administration: 10-fold as opposed to twofold (28). These data show that adenosine antagonists such as caffeine influence plasma adenosine concentrations, by an unknown mechanism, at doses similar to those in the range of medium to heavy coffee drinkers. Thus sudden changes in methylxanthine consumption could alter plasma adenosine concentrations and could be involved to the genesis of seizures in epilepsy patients.
Because long-term consumption of caffeine increases plasma adenosine levels in rats and its abrupt discontinuation results in a decrease of this purine nucleoside (28), and adenosine may act as an endogenous anticonvulsant (18–23), we expected a proconvulsant effect on seizures after caffeine discontinuation. Therefore the purpose of the present study was to evaluate the convulsive effects of long-term exposure to doses of caffeine, similar to the regular daily human intake, and after its abrupt discontinuation, in three groups of animals: (a) rats subjected to acute subconvulsant doses of pilocarpine; (b) rats subjected to acute subconvulsant doses of kainic acid; and (c) rats subjected to status epilepticus (SE), induced by systemic administration of cholinergic agonist pilocarpine, that underwent a latent period and subsequently developed a state of “chronic” epilepsy, characterized by the emergence of spontaneous recurrent seizures (SRSs).
These findings may have implications for individuals who cavalierly increase or withdraw from consumption of caffeine, a compound generally considered to be a useful but innocuous stimulant (28).
- Top of page
- MATERIALS AND METHODS
Our study was based on the hypothesis of Conlay and colleagues (28) that sudden changes in caffeine consumption—and thus in plasma adenosine concentrations—could influence seizure thresholds. Unexpectedly, given that adenosine receptors are known to be involved in seizure threshold in humans and laboratory animals (36,37), we showed that prolonged administration of 0.1% caffeine in drinking water for a period of 2 weeks and its abrupt withdrawal did not influence the later development of seizures in rats subjected to an acute subconvulsant dose of pilocarpine or kainic acid (experiment 1) and did not change the frequency of SRSs in the pilocarpine model epilepsy (experiment 2).
How can we reconcile the data and hypothesis forwarded by Conlay and colleagues (28) and our current findings? In addition, how can we reconcile the discrepancy regarding the convulsive effects of caffeine over animal models of epilepsy and the scarcity of human clinical observations on the convulsive effects of caffeine?
To extrapolate information derived from animal experiments to humans is not a trivial task, particularly regarding frequency (short term and long term) and dose (low and high) of caffeine consumption in relation to physiologic and toxic effects. The only known biochemical mechanism that is significantly affected by clinically relevant doses of caffeine, similar to those attained during normal human consumption, is the blockade of adenosine A1 and A2 receptors without any selectivity (38).
A brief exposure to methylxanthines decreases the threshold to various convulsants (1,4–7,39), whereas prolonged exposure to low doses of caffeine leads to decreased susceptibility to seizures induced by γ-aminobutyric acid (GABA)A-receptor antagonists such as bicuculline and pentylenetetrazol (40), or by the glutamatergic agonist N-methyl-d-aspartate (NMDA) (41,42). These data indicate that the effects of prolonged caffeine use are not related to any specific form of seizure model but are rather more general and occur in the complete absence of any change in the number of adenosine A1 receptors (41,43) or GABAA/benzodiazepine receptors (40). In contrast, Boulenger and colleagues (44) reported an increase in the number of central adenosine receptors and also a transitory increase in the number of benzodiazepine receptors in the mouse brain as a result of the prolonged blockade of adenosine receptors by long-term caffeine consumption in high doses. Dose size might be the critical issue underlying some of these discrepancies. Indeed, seizures and altered seizure susceptibility by means of caffeine or methylxanthine are possible only by means of unrealistic very high doses of these compounds (1,2).
Studies investigating the effects of prolonged caffeine administration at doses similar or equal to that used in the current study report either decreased seizure susceptibility in mice (40,41) or unaltered frequency of spontaneous seizures in rats (45). Similarly, we did not find any evidence for an effect of long-term caffeine administration in the pilocarpine and kainate models of epilepsy. In addition, our data indicate that at the currently used doses, abrupt withdrawal of caffeine also did not increase seizure susceptibility and did not alter the frequency of spontaneous seizures. In this sense, not even the abrupt biochemical changes likely be associated to caffeine withdrawal seem to interfere with seizure susceptibility in these models of epilepsy.
The main hypothesis underlying the current experiments was that altered adenosine levels associated with abrupt caffeine withdrawal (28) would be able to affect seizure susceptibility. Evidence exists that the release of excitatory transmitters is more strongly inhibited by adenosine than is that of inhibitory transmitters (46). In addition, when the effect of endogenous adenosine on potassium channels via A1 receptors on glutamatergic neurons is blocked by caffeine, it leads to epileptiform activity (19,20). However, recent evidence suggests that field potentials in CA3 pyramidal cells evoked by mossy fiber stimulation increase by only 20–30%, even when A1 adenosine receptors are fully blocked (47).
Although caffeine acts at the level of adenosine receptors and caffeine consumption and/or cessation influence adenosine concentrations, it clearly appears that this does not seem to change brain excitability enough to trigger seizures or increase their occurrence in an epileptic brain. It is known that long-term caffeine also alters the coupling of receptors to G proteins and the phosphorylation of DARPP-32 (48). Thus, first, whether or not the serum concentrations of adenosine reflect its central level and/or action must be proven, and second, the molecular cascade involved in caffeine action is far from being entirely clear, which most likely explains the paradoxical results observed in the present study.
It is surprising that the available evidence about the effects of caffeine on seizure susceptibility on animal models is not matched by clinical evidence over different forms of epilepsy in humans. Indeed, none of the major textbooks and clinical information databases on human epilepsy devotes more than a few sentences, if any, to describing an association of caffeine consumption and seizure threshold. A survey made to compare the distribution of seizures precipitants among epilepsy syndromes showed that caffeine was infrequently noted by patients as a precipitant (49). The only two reports of an association between caffeine consumption and seizures are the description of a single patient who often ingested a high volume (>2 L) of a caffeinated beverage over a short period (8) and a 45-year-old woman who had a tonic–clonic seizure after caffeine withdrawal (9).
In conclusion, these data suggest that caffeine administration and withdrawal, even though affecting plasma adenosine levels, are also associated with additional compensatory changes and had no influence on seizures in two different models of temporal lobe epilepsy in rats. Despite the extensive laboratory evidence on the convulsant properties of xanthine derivatives in animal models of epilepsy, such strong evidence is lacking in clinical settings. Our current finding with the administration of caffeine at doses similar to those in the range of medium to heavy coffee drinkers (28) supports and confirms the clinical experience. It is possible that long-term treatment with caffeine, at doses similar to those provided to humans, induces important adaptive changes in the brain (50), which may be beneficial rather than detrimental.