Neurontin (gabapentin) spends plenty of time in our patients medicine cabinets, but lately it has spent nearly as much time splashed across news sections of daily papers. Parke-Davis, the company that used to market Neurontin before it merged with Pfizer, has been accused of improperly promoting its use for a variety of off-label indications (1).

As psychiatrists, we know a lot about off-label uses of Neurontin, since it is only approved for two indications, neither of them psychiatric: epilepsy and postherpetic neuralgia. This doesnt stop us from using it an awful lot, though, current company included. Common psychiatric uses include: bipolar disorder, anxiety disorders, insomnia, alcohol detox, and cocaine addiction. Ask almost any psychiatrist on the street, and he or she will swear that it is effective treatment for at least some patients with these problems. Unfortunately, placebo-controlled studies of Neurontin have rarely corroborated the results of open label trials or anecdotal experiences.

As an example, consider the tumultuous relationship between Neurontin and bipolar disorder. A plethora of letters to major journals, small case series, and uncontrolled clinical trials in the late 1990s appeared to glowingly endorse Neurontin as an effective treatment for acute mania, mixed mania, bipolar depression, and schizoaffective disorder (2). However, we all sustained a harsh reality check when the placebo-controlled trials started rolling in. First, a ParkeDavis funded trial found that Neurontin performed worse than placebo when it was added to pre-existing mood stabilizers in bipolar disorder (3). Then, an NIMH study found it to be no more effective than placebo as monotherapy for refractory bipolar disorder and unipolar mood disorders; in this study, hotshot upstart Lamictal (lamotrigine) handily beat both Neurontin and placebo (4).

Getting back to this months TCR focus, what about Neurontin for panic disorder and other anxiety disorders? Theoretically, Neurontin would be an ideal agent for anxiety. It is structurally similar to GABA, which is the main inhibitory neurotransmitter in the central nervous system. Recall that those two legendary anti-anxiety agents, benzodiazepines and ethyl alcohol, both exert their primary action by stimulating GABA receptors in different ways (5). Neurontins mechanism of action is less clear, but it appears to modulate GABA without causing tolerance or withdrawal, unlike its anti-anxiety cousins. But is it effective? Unfortunately, the evidence is scant. There is a small case series of 4 patients (6) with either panic disorder or generalized anxiety disorder, all of whom responded to relatively low doses of Neurontin (ranging from 100 mg t.i.d. to 300 mg t.i.d.). And then there are two randomized controlled studies, one for social phobia, and the other for panic disorder. Both were funded by Parke-Davis, were well-designed, and were rather underwhelming in their results. The social phobia study (7) randomized 69 socially phobic patients to either Neurontin (average dose a very high 2868 mg per day) or placebo. Neurontin-treated patients had a 32% response rate, significantly higher than the 14% placebo response rate. Not terribly impressive, especially when compared to typical response rates of 50% or more seen in studies of SSRIs and benzodiazepines (8).The panic disorder study was even more dismal: no difference at all between Neurontin and placebo (9). However, using some statistical sleight of hand, the authors were able to show some separation from placebo in the 53 patients defined as having more severe panic symptoms. As in the social phobia study, Neurontin doses were high (up to 3600 mg per day) although the average dose was not reported.

So what to say about Neurontin for anxiety? It has a great side effect profile, it causes no drug-drug interactions, it isnt addictive, and most clinicians reading this have seen robust anxiolytic responses with their own eyes. If only the data would catch up!

TCR VERDICT: The Data is Lukewarm, at Best