Alcohol Withdrawal Kindling

Abstract

Animal and clinical studies supporting the kindling hypothesis of alcohol withdrawal suggest the need to revisit current treatment concepts. While traditional approaches have emphasized symptom reduction and prevention of complications, novel approaches include slowing progress of clinical severity associated with multiple withdrawals. Currently, it is unclear if medications can halt cumulative neurotoxicity associated with multiple withdrawals. However, the ability of anticonvulsants to improve the course of alcohol withdrawal and their neuroprotective effects may be of interest. The use of anticonvulsants as probes in animal models of kindling and controlled trials examining the efficacy of newer anticonvulsants in the treatment of alcohol withdrawal may improve understanding of alcohol withdrawal kindling and its treatment.

Introduction

Alcohol dependence, a chronic relapsing disorder, affects 14 percent of Americans causing significant morbidity and mortality.¹ Chronic heavy drinking leads to central nervous system (CNS) neuroadaptations. Discontinuation of alcohol following chronic use may result in CNS over-excitation clinically manifested by the signs and symptoms of alcohol withdrawal.² For many alcohol-dependent patients, detoxification is the first step of treatment. Alcoholics with adequate social support, mild to moderate alcohol withdrawal, and stable medical and psychiatric status may be suitable candidates for outpatient management.³

In the United States, benzodiazepines have become the gold standard pharmacological treatment of alcohol withdrawal.⁴ However, the potential for benzodiazepine misuse and the interactions with alcohol complicate its use in outpatient detoxification.⁵ To protect against unnecessary prescriptions of benzodiazepines for withdrawal, decisions can be based upon quantitative ratings of alcohol withdrawal symptom severity.^6,7 The approach of monitoring symptoms while withholding benzodiazepines has been widely endorsed for cases of mild alcohol withdrawal.^4,8 Yet, the potential for subclinical neurotoxicity during each episode of untreated alcohol withdrawal suggests that there may be a role for a more aggressive pharmacological intervention.^9,10

According to the kindling hypothesis, multiple episodes of alcohol withdrawal not only cause immediate neurotoxicity, but also cumulative changes in neuronal excitability.¹¹ This phenomenon of neural sensitization, referred to as “kindling,” appears to clinically manifest itself by progressive worsening of subsequent withdrawal episodes, particularly the predisposition to experience alcohol withdrawal seizures.⁹ The question of whether benzodiazepines can prevent neurotoxicity associated with repeated alcohol withdrawal by blocking early withdrawal reactions has recently received some attention.^12,13 There has also been growing interest in the potential neuroprotective effects of anticonvulsants in preclinical models with cerebral ischemia and traumatic brain injuries.¹⁴ These medications may exert neuroprotective effects by either interacting directly with neurotransmitter receptors or via ion channels (e.g., voltage sensitive calcium channels).^15,16 Clinical trials demonstrating that anticonvulsants improve the course of alcohol withdrawal suggest neuroprotective effects may be relevant to the pathophysiology of alcohol withdrawal states.²⁰

This article examines the potential for increasing the use of anticonvulsants in the outpatient treatment of alcohol withdrawal in view of the kindling hypothesis. First, we will discuss the empirical basis for the kindling hypothesis of alcohol withdrawal. Second, we will review the effects of alcohol and selected anticonvulsants on major excitatory and inhibitory neurotransmitter systems, as well as possible neurobiological correlates of the kindling hypothesis. Third, we will examine clinical trials evaluating the efficacy of selective anticonvulsants for mild to moderate alcohol withdrawal. Lastly, we will conclude by identifying directions for future research on the kindling model and highlighting therapeutic implications.

Alcohol Withdrawal and Kindling

The kindling concept is derived from epilepsy animal models and refers to a sensitization phenomenon by which stimuli of neural sites with no initial epileptogenic effect will, if repeated over time, result in the development of seizures.²¹ In alcohol withdrawal, the kindling model is used to demonstrate the association between the number of withdrawal episodes and the severity of subsequent withdrawal reactions.^9,20 Specifically, repeated episodes of ethanol withdrawal are thought to promote changes in neuronal excitation, leading to cumulative neuronal damage clinically manifested by an increasing severity of the withdrawal syndrome in subsequent episodes; this includes a higher risk for seizures.

Mice studies comparing the severity of withdrawal and pharmacologically induced seizures after single and multiple withdrawal reactions are of interest in this regard.^23,24,26 Specifically, withdrawal seizures are significantly more severe in mice tested after multiple withdrawals compared to mice tested after a single withdrawal.²³ Similarly, mice exposed to multiple cycles of ethanol withdrawal display significantly enhanced sensitivity to the pharmacologically induced convulsions in comparison to mice the experience a single withdrawal episode.^24–26 Animal electroencephalogram (EEG) studies also show that repeated alcohol withdrawal results in epileptiform abnormalities that worsen during successive withdrawals.^27,28

Specifically, at baseline, all mice display low levels of brief spindle episode activity. During periods of ethanol withdrawal, mice exhibit a significant increase in activity, which is potentiated during subsequent withdrawal episodes.²⁹ Human studies showing a positive relationship between the number of previous withdrawal reactions and the risk of developing seizures during subsequent episodes of alcohol withdrawal provide further support for the kindling hypothesis.^30–32

Although the precise mechanisms involved in kindling need further elucidation, long-term changes in N-methyl-D-aspartate (NMDA) and gamma aminobutyric acid (GABA-±) receptor functioning following repeated episodes of alcohol withdrawal may be implicated.²⁹ The development of experimental models sensitive to the effects of repeated withdrawals has shed some light on the neurobiology of sensitized withdrawal seizures and neurotoxicity resulting from repeated alcohol withdrawal. Next, we will review alcohol effects on NMDA and GABA systems and speculate how it may relate to kindling. We will also discuss candidate mechanisms of anticonvulsants putative neuroprotective effects on these systems.

Ethanol Affects the Glutamate NMDA Receptor System

The glutamate NMDA receptor complex is a ligand-gated ion channel that mediates the influx of calcium ions when activated. The NMDA receptor complex and voltage-gated calcium currents are both involved in controlling neuronal excitability.³³ Animal studies show that acute administration of ethanol disrupts glutamatergic neurotransmission, causing potent and selective inhibition of the function of the NMDA subtype of glutamate receptor.³⁴

Chronic alcohol exposure results in prolonged NMDA inhibition and compensatory increases in the number of NMDA receptors (upregulation) in several areas of the brain.^34,35 Similarly, acute ethanol exposure inhibits voltage-sensitive calcium channels while chronic ethanol exposure results in an increase in the number of voltage sensitive calcium channels.^36,37

Animals withdrawing from ethanol exhibit a marked increase in glutamatergic function, including an increase in glutamate release combined with functional overactivity of the NMDA receptors.^38–40 This glutamatergic overactivity is likely to mediate withdrawal symptoms (e.g. seizures) as well as the neurotoxic effects of ethanol.⁴¹ Interestingly, the anticonvulsant carbamazepine has been shown to depress NMDA-receptor-mediated responses.^42,43 Several other anticonvulsants including valproate and gabapentin also attenuate glutamatergic neurotransmission.⁴⁴ Glutamatergic-induced neurotoxicity occurring during ethanol withdrawal is thought to be due in part to excessive amounts of calcium ions entering the hyperexcitable neurons and has been implicated in cumulative neuronal damage associated with repeated ethanol withdrawal.⁴⁵

Evidence for involvement of the glutamate NMDA receptors in kindling derives from animal studies indicating that a history of multiple withdrawal episodes correlates with increased sensitivity to NMDA-induced seizures.⁴⁶ Efforts to inhibit glutamatergic excitability and neurotoxicity associated with alcohol withdrawal focused on acamprosate.⁴⁷ This medication, which was approved for the treatment of alcohol dependence in Europe and recently approved in the US, exerts a weak inhibitory effect on the glutamatergic system via the glutamate/NMDA receptors resulting in reduced NMDA receptor function.⁴⁸ In at least two in-vitro studies, acamprosate had protective effects against glutamate-induced neurotoxicity in ethanol-withdrawn cultures.^49,50 While these effects do not correlate with acamprosate's effects on calcium entry, anticonvulsants may exert their neuroprotective effects by interacting with voltage-gated calcium currents.^49,51

One in-vitro study has shown that the anticonvulsants oxcarbamazepine and lamotrigine consistently inhibit voltage-activated calcium currents.¹⁵ The anticonvulsant gabapentin inhibits neuronal calcium influx in a concentration-dependent manner.⁵² If decreased calcium influx indeed reduces excitatory glutamate release, anticonvulsants may indeed exert neuroprotective effects during alcohol withdrawal. Because current understanding of the full spectrum of the biological effects of alcohol, anticonvulsants, and neurotransmitter systems are incomplete, both animal and human models of kindling require further development. Perhaps the use of anticonvulsants as pharmacological probes in animal kindling models may improve the understanding of kindling neurobiological substrates as well as the development of alternative pharmacological treatments for alcohol withdrawal.

Ethanol Effects on GABA

Acutely, ethanol consumption acts at the benzodiazepine GABA-±-chloride receptor complex, potentiating its inhibitory effects and suppressing neuronal activity in several areas of the brain.⁵³ Specifically, activation of GABA-± receptors results in the entrance of chloride ions into the neuron. This increase in intraneuronal chloride concentration dampens neural activity. Conversely, chronic ethanol administration may alter GABA-± receptor functioning by decreasing either the number or the expression of the GABA-± receptor subunit.^54,55 Both GABA-± downregulation and conformational changes in the configuration GABA-± subunits associated with chronic ETOH exposure may influence neuronal responsivity to alcohol and the development of hyperexcitability during ethanol withdrawal.^36,37

The ability of benzodiazepines and certain anticonvulsants to substitute for the GABA-enhancing effects of ethanol probably accounts for their effectiveness in suppressing acute symptoms of alcohol withdrawal.⁵⁶ Anticonvulsants' neuroprotective effects may also occur by potentiation of GABAergic inhibitory pathways. Valproic acid, for example, has multiple GABAergic actions. It appears to mediate GABA-± receptor activity, decrease GABA catabolism, and increase GABA release.⁴⁴ The anticonvulsant topiramate, recently shown to decrease relapse in previous heavy drinking, also has GABAergic actions.⁵⁷ Similarly, the anticonvulsant gabapentin is an amino acid structurally related to GABA. Although it neither adheres to the GABA receptor nor modulates GABA functioning, it may indirectly interact with the GABA transporter and increase GABA levels in a dose-related manner.^58,59 Its mechanism of action remains elusive, but inhibition of a neuronal calcium channel (presumably L-type) and potentiation of GABA synthesis may play a role.⁶⁰

While the effects of anticonvulsants on GABA may vary, it has been postulated that anticonvulsants may inhibit alcohol withdrawal kindling by potentiating GABA-± receptor activity.^61,62 Specifically, enhancement of GABAergic inhibition is thought to protect neurons during insults, which induces excessive neuronal depolarization. Of interest, however, is the discovery that medications targeting alcohol withdrawal may have differential effects on acute versus long-term effects of alcohol withdrawal.⁶³

Efforts to determine whether GABAergic modulators can prevent or blunt the development of sensitized CNS hyperexcitability during a subsequent withdrawal episode are underway. At least two animal studies have evaluated the effects of benzodiazepines in attenuating CNS hyperexcitability associated with multiple cycles of ethanol withdrawal.^12,13 These studies have employed a mouse model sensitive to the effects of multiple cycles of ethanol withdrawal. This experimental model consists of exposing mice to continuous or chronic intermittent ethanol treatment (CIE) followed by various withdrawal patterns. For instance, one group may receive four cycles of 16 hours of ethanol exposure separated by eight hours of withdrawal periods. The second group may receive a single 16-hour exposure period while a third group consists of ethanol-naïve controls. The presence, frequency, and severity of seizures or epileptiform abnormalities on EEG are compared among these groups and the effects of various medications on seizures during acute and repeated withdrawal can be evaluated.

In one study employing this model, lorazepam attenuated the development and expression of sensitized handling-induced convulsions (HIC) during the acute withdrawal phase and initial subsequent unmedicated withdrawals in a dose-dependent fashion.¹² At later points, however, lorazepam exacerbated the development and expression of sensitized seizures.

Another study of similar design showed that diazepam suppressed acute withdrawal symptoms, but when administered during intermittent withdrawal did not alter seizure sensitivity during a subsequent unmedicated withdrawal.¹³ Together, these findings suggest that suppression of acute withdrawal symptoms by benzodiazepines may either not prevent or worsen seizures resulting from repeated exposures to ethanol withdrawal.^12,13 Similar animal studies employing anticonvulsants that interfere with GABAergic systems combined with clinical studies evaluating the efficacy of such medications on the treatment of alcohol withdrawal may improve understanding of the kindling-GABAergic relationship. Advancements in the field could be expected if both antikindling and improvement of acute symptoms could be demonstrated for these medications.

Anticonvulsants for Alcohol Withdrawal

Carbamazepine (CBZ) is the most studied anticonvulsant for the treatment of alcohol withdrawal.^17–20,65 It has been shown to be superior to placebo and comparable to benzodiazepines in improving symptoms of mild to moderate alcohol withdrawal.^18–20,64 In addition, at least two double-blind trials support its utility in comparison to other sedative-hypnotic drugs (clomethiazole and tiapride) in ameliorating withdrawal symptoms.^17,65 In a few of these studies, CBZ has also shown beneficial effects on protracted withdrawal symptoms, such as sleep disturbances, anxiety, and psychological distress.^18,64,66

Because alcohol-dependent individuals often have protracted withdrawal symptoms, which may contribute to relapse shortly after detoxification, these post-detoxification benefits may be of clinical value. Indeed, at least two placebo-controlled studies suggest that CBZ may reduce some alcohol use in alcohol-dependent patients in the post-withdrawal period.^20,67 One of these trials (n=136) compared lorazepam and carbamazepine for the outpatient treatment of moderate alcohol withdrawal and drinking behaviors in the immediate post-detoxification period.²⁰ Patients received either CBZ (600–800mg) or lorazepam (6–8mg) on the first day and were tapered to either 200mg of CBZ or 2mg of lorazepam over a five-day period. Although both CBZ and lorazepam caused a comparable decrease in alcohol withdrawal symptoms, CBZ was superior to lorazepam in reducing post-treatment drinking, particularly among patients with multiple withdrawal episodes. In this study, the relative risk of having a first drink was at least three times more likely in lorazepam-treated patients than in those receiving CBZ. Despite these potential advantages, the use of carbamazepine is limited due to hematological and hepatic toxicities, in addition to multiple drug interactions. Oxcarbamazepine, a newly improved derivative of CBZ, appears to have significantly less hepatotoxicity and less drug interactions.⁶⁸ It may also have neuroprotective effects that need to be examined in the context of alcohol withdrawal.¹⁴

Preliminary studies suggest that various formulations of the anticonvulsant valproate can modify the course of alcohol withdrawal.^62,69,70 In one such study, 400 patients were randomly assigned to either valproic acid plus conventional treatment or conventional treatment alone.⁶⁹ Patients receiving adjunctive valproic acid demonstrated a faster decline in symptoms and required less conventional treatment. More recently, researchers evaluated the efficacy of divalproex sodium for alcohol withdrawal and relapse prevention in an open-label study.⁷⁰ Sixteen patients in mild to moderate alcohol withdrawal were randomly assigned to receive benzodiazepine detoxification, divalproex sodium detoxification, or divalproex sodium detoxification with maintenance. Symptom reduction occurred more rapidly in the patients on divalproex sodium than in the patients treated with benzodiazepines. In a double-blind, controlled, seven-day study (n=36), researchers evaluated the adjunctive use of divalproex sodium to a benzodiazepine in the treatment of moderate alcohol withdrawal.⁶² Thirty-six patients in moderate alcohol withdrawal were randomized to receive either divalproex sodium (500mg TID) or placebo. All patients received a dose of oxazepam (30mg) at the time they first received the study medication and additional oxazepam in accordance with a standard, symptom triggered detoxification protocol. Divalproex sodium-treated patients required significantly less oxazepam than placebo-treated patients. In addition, the progression of withdrawal severity was significantly less among patients treated with divalproex sodium. Although replication is required, these pilot studies suggest that both formulations of valproate may alter the course of alcohol withdrawal by decreasing its severity. In addition, two small pilot studies suggest that divalproex sodium may decrease measures of heavy drinking in alcohol-dependent patients.^70,71

Finally, the anticonvulsant gabapentin has also been studied for the treatment of alcohol withdrawal states. It is reported that gabapentin's protective effects against convulsant and anxiogenic dimensions of alcohol withdrawal syndrome in rodents.⁷² Subsequently, Myrick, et al., used gabapentin in six alcohol-dependent patients experiencing mild to moderate alcohol withdrawal for a period of five days.⁷³ Gabapentin was both effective and well-tolerated in these subjects. Bonnet, et al., reported its potential value as an adjunct to chlormethiazole in treating four alcohol-dependent patients undergoing detoxification.⁷⁴ Gabapentin was found to be an effective treatment for insomnia associated with alcohol dependence in 15 detoxified patients at average doses of 600mg/d (range 200–1,500mg/d). In general, each patient showed improvements in the sleep pattern questionnaire.⁷⁵

Voris, et al., conducted a retrospective chart review in 49 inpatient (18) and outpatient (31) veterans receiving gabapentin for mild to moderate alcohol withdrawal.⁷⁶ Gabapentin dosing in general was 400mg three times a day for three days, 400mg twice a day for two days and 400mg once a day for one day. This report indicated the potential efficacy of gabapentin for mild to moderate alcohol withdrawal. Controlled trials are currently being conducted to determine if indeed gabapentin is useful in mild to moderate alcohol withdrawal treated in outpatient settings. In sum, most available studies support the efficacy of CBZ in the treatment of mild to moderate alcohol withdrawal. Similarly, preliminary trials suggest the positive effects of various valproate formulations and gabapentin on the course of alcohol withdrawal. Oxcarbamazepine, a derivative of CBZ, has not yet been studied and may yield favorable results.

Conclusion

The kindling concept has increasingly challenged current notions of the optimal management of alcohol withdrawal. Novel therapeutic approaches focus not only on symptom reduction, but also on prevention of cumulative neurotoxicity and progression of clinical severity associated with multiple withdrawal episodes. While it is unclear if early pharmacological interventions can slow this process, putative neuroprotective effects of anticonvulsants may influence alcohol withdrawal induced neurotoxicity. Animal models designed to study neurobiological changes associated with multiple cycles of alcohol withdrawal may clarify the molecular basis of kindling. Perhaps the use of anticonvulsants as pharmacological probes in these models may help determine if anticonvulsants can halt the progression of withdrawal severity associated with multiple alcohol withdrawal episodes.

Thus far, clinical studies evaluating anticonvulsants for alcohol withdrawal have focused on carbamazepine, which has the strongest empirical basis supporting its use. Carbamazepine has also shown beneficial effects on both protracted withdrawal and in reducing alcohol consumption measures in the immediate post-detoxification period. Although not studied, the anticonvulsant oxcarbazepine is a likely candidate for examination based on pharmacological similarities with carbamazepine and its potential neuroprotective effects. Pilot studies of both valproic acid and gabapentin have shown that these anticonvulsants may truly improve the course of alcohol withdrawal. To date however, large scale double-blind placebo-controlled trials are lacking and more information on the efficacy, safety, and neuromechanisms of these medications are needed.