Over the past couple of decades, psychiatry has adopted a number of anticonvulsants that effectively treat psychiatric conditions. The kindling hypothesis has provided a rationale for their increasing use, but what is the evidence behind this theory, and is it actually applicable for psychiatric practice?

The phenomenon of kindling was first discovered in 1967 by a scientist in Halifax, Nova Scotia, named Graham Goddard. Goddard was a neuroscientist interested in the neurobiology of learning. In one series of experiments, he electrically stimulated various regions of rats brains to observe the effects on their ability to learn tasks. In repeating these stimulations daily, he discovered something unexpected: the rats began having seizures in response to stimuli that would normally be too low to provoke seizures. Ultimately, many of the rats began having unprovoked seizures. Somehow, Goddard had created epileptic rats.

He eventually called this phenomenon kindling (Goddard GV, Development of epileptic seizures through brain stimulation at low intensity, Nature 1967;214:1020). Just as a large log will not burn unless kindled by the combined action of small twigs burning, it appeared that epilepsy required a similar kind of kindling by a sequential series of small electrical stimuli.

How does this relate to psychiatry? The most common analogy is between an epileptic seizure and a manic episode of bipolar disorder. Like seizures, manic episodes can occur without obvious triggers, and have fairly abrupt beginnings and endings. In the case of bipolar disorder, the kindling is theoretically provided by stressful life events, which may produce certain kinds of electrical brain stimulations. At first, these events are not sufficient to cause a manic episode, but over time, they may accumulate to trigger such an episode. Furthermore, episodes may beget episodes, meaning that the manic episodes themselves may damage the brain in some way, making it more vulnerable, so that eventually the episodes may begin to occur spontaneously, without a trigger.

The evidence for kindling in bipolar disorder is indirect. The most eloquent spokesperson indeed, the person who initially applied the idea of kindling to psychiatric illnesses is Robert Post, who is currently a professor of psychiatry at George Washington University. In a recent paper, he concisely reviews the evidence for kindling in affective disorders (Post R, Neuroscience and Biobehavioral Reviews 31 (2007) 858-873). He cites studies showing that patients who have had a number of affective episodes are more vulnerable to future episodes and that later episodes are less likely to require an environmental trigger than earlier episodes. But he acknowledges that some studies disagree, and that many patients do not follow these patterns.

Skeptics would argue that studies cited as evidence of kindling may simply be identifying a subset of patients with severe affective illness who get worse over time, as do many severely ill patients in all of medicine. True, one possible explanation of worsening over time is that the prior episodes do some cumulative damage (episodes begetting episodes) but there are many other equally plausible explanations: an underlying disease of neurotransmitters may worsen with time and be unrelated to kindling; severely psychiatrically ill patients make a series of poor life decisions that lead to vicious cycles of more stress triggering more illness, and so on.

If the kindling hypothesis were true, what are the clinical implications? The major one is that you should treat early and aggressively, in order to prevent the pathological affective episodes. But again, this clinical wisdom is hardly dependent on the kindling hypothesis, and most clinicians would agree that aggressive treatment of psychiatric illness is warranted, regardless of the hypothesized cause.

Perhaps the most misunderstood aspect of kindling is that it implies that we should treat affective disorders with the same medications as are used for epilepsy. In fact, in the words of Dr. Post, Weuse the kindling model only for its heuristic value in asking questions regarding the longitudinal course of illness and response to treatment. The utility of this model must ultimately rest on its indirect or clinical predictive validity (Post RM, et al., Clinical Neuroscience Research, 2001;1:69-81). In an email to me, Post pointed out that another big misunderstanding of the kindling hypothesis is that it means that affective illness progresses relentlessly. Not true, he said. If you treat it aggressively enough any point in its course, you can hopefully stop it.

TCPR VERDICT: Kindling: Not a roadmap for treatment decisions