Alzheimer’s disease (AD) is the most common cause of dementia – a chronic progressive decline in global cognitive ability. It is an actual disease, in that it is a specific pathophysiological entity. In other words, it is not defined only by a clinical syndrome, but also by specific pathological findings in the brain. The two most important pathological findings are the presence of beta-amyloid plaques and neurofibrillary tangles. (As an aside, in practice most patients with this type of dementia are diagnosed clinically with Alzheimer’s-type dementia. The Diagnosis of AD requires a brain biopsy or autopsy, the former is rarely done because there is no benefit to doing so.)

Reports about new findings in AD come out almost weekly – it is a complex and well-studied disease. Very few of the findings translate into a change in medical practice – the incorporation of a new diagnostic tool or treatment. It remains to be seen, therefore, if the latest findings will remain a curiosity or change the way AD is treated.

Lennart Mucke and his colleagues published in Neuron their research on the possible role of seizures in AD. They looked at a mouse model of AD (human amyloid precursor protein (hAPP) transgenic mice) to answer some questions to which careful observation and thoughtful reflection had led them. They noticed that AD brains show some of the same pathological changes that can occur following a seizure. It was known that one of the consequences of the destruction of AD is that remaining neurons may form aberrant networks – as neurons die the wiring becomes faulty. They speculated that maybe these aberrant networks lead to seizures that in turn cause further damage and contribute to the progression of AD.

They found that the AD transgenic mice did indeed have hyperexcitability in the hippocampus – a brain structure critical for short term memory. They also found pathological changes in the hippocampus of such mice consistent with the kind of changes that occur to compensate for this hyperexcitability, namely “GABAergic sprouting, enhanced synaptic inhibition, and synaptic plasticity deficits” – these would have inhibitory effects.

This is an intriguing study, especially since it may lead to an entirely new approach to the treatment of AD, using off-the-shelf anti-seizure medications.

But this study also shows the complexity of diseases like AD and why we see so many hopeful headlines but few treatments down the road. As a clinical neurologist and educator, here are the questions that occur to me in response to this interesting data:

Of course, as the authors admit, we need to see if the same is true in humans (and we won’t be able to sacrifice human test subject and slice up their brains, so we won’t have the pathological data unless we eventually get it from autopsies, which will take years). But we need to find out if the same seizures or hyperexcitability is present in humans with AD.

The authors argue that the seizures they are seeing in the mouse do not have any motor manifestation and that is why they are not recognized. This is a fair hypothesis, even though it is explaining the lack of clinical evidence for seizures. However, electroencephalograms (EEGS), which could pick up evidence for seizure activity, are routinely (if not always) performed as a standard part of the dementia workup. So how come clinicians are not already finding seizures or the potential for seizures in AD patients?

Further, even so-called “silent” seizures (without any obvious outward motor manifestation) are possible to detect in people – much more easily than in a mouse. During such a seizure a person would not contact and not respond to their environment. Clinicians see this too, and when we encounter this pattern of behavior we suspect silent seizures as a possibility and specifically look for them. We may even treat empirically to see if the episodes go away. So again, my impression is that we would have encountered this more clinically already, if seizures were playing a major role is AD.

On the other side you could argue that AD patients do fluctuate in their symptoms and that this is often written off as just part of the disease and therefore many clinicians might not look for seizures in AD patients. This is reasonable, and could explain why it would be sometimes missed – but not universally missed.

Also, we have a vast experience with epilepsy. Patients with even severe epilepsy do not necessarily develop AD. Whether or not seizures damage the brain remains controversial – which means any effects are subtle and hard to nail down. Part of the problem is that anti-seizure medication can inhibit cognition, and of course an underlying neurological problem can cause both seizures and cognitive impairment. So it is difficult to entirely separate out these other factors. But any dementing effect of seizures is mild at best since it has eluded direct search without a definitive answer for decades.

So while the results of this study are very interesting, and this line of research will likely tell us something new about AD and the brain, I think it is not likely that it will have a huge impact on AD clinically (at least no directly). I base this largely on the fact that current clinical assessment of AD would have picked up seizures as a significant contributor before. But negative arguments are always inherently weak – our statements about the absence of a phenomenon are only as reliable as the thoroughness with which we looked. So it would be reasonable to study AD patients to look more carefully and thoroughly for subtle epileptic activity.

Also, keep in mind that no one is claiming seizures cause AD. What the authors are suggesting is that a subtle form of seizures is the consequence of AD but then secondarily contributes to the progression of AD. Which further means that at best anti-seizure treatments would slow down the progression of AD, not cure it. Still, any benefit could have a huge impact on the lives of people with AD and their families. I hope this does lead to something useful.

And finally I think the larger perspective to take is that in very complex diseases like AD once something goes wrong it sets off a chain reaction of secondary effects. It often seems that no matter what parameter you choose to look at, it is abnormal. This leads to copious headlines about everything “causing” the disease under study, when rather researchers are just finding a bunch of secondary markers for the presence of the disease or the damage it causes. A chain of evidence is required before we can conclude that the abnormal finding is actually causative, or that it represents a possible target for intervention.

So while I am hopeful at every new such finding, I will await for the chain of evidence to unfold before I get excited, knowing that most of the time these new findings do not pan out and result in new treatments.