WEIRD SCIENCE

ARE ONE-THIRD OF DEMENTIA CASES REALLY PREVENTABLE? LEND ME AN EAR.

Science news comes to most of us as a sort of fog of headlines. The medical journal Lancet, that ever-living expression of the 19th-century English reforming spirit, published a new “commission” this week that reviewed the best existing evidence on the subject of dementia prevention. A “commission” is an exercise in which the Lancet pulls together a couple dozen top experts on a topic; this one tackled the question of whether the tremendous global burden of age-related cognitive decline can be reduced, and, if it can, by how much.

The article might have come to you second-hand, out there in the fog, as a headline like, “About one-third of dementia cases could be prevented” (which comes from the L.A. Times), or, “Nine lifestyle changes can reduce dementia risk, study says” (which was BBC News’s take.) Editors do love a list.

Reports on the Lancet article tended to mention smoking and obesity as major factors in dementia prevalence. This might be unfortunate, since everybody already knows both these things are bad for you, and also that they tend to be risk factors for almost any disease you can imagine. The fog is mostly created by doctors who really, really want you to put down the french fries and the unfiltered Camels.

The authors of the commission started out with a list of known, potentially preventable risk factors for dementia — high blood pressure, social isolation, diabetes, things like that. (That’s where some headline writers are getting their “nine lifestyle changes.”) They then made careful estimates of how much the risk is believed to be intensified by each factor; they estimated how much of each factor exists around the world; and they came up with a “population attributable fraction” (PAF) for each one. The PAF is the number that (very loosely) answers questions like this: if we could eliminate cigarette smoking, how much would that reduce the incidence of dementia?

So let me quiz you: what would you guess was the preventable risk factor with the highest PAF? Or, put another way: what preventable syndrome, characteristic, or unhealthy activity is contributing the most to worldwide dementia in the elderly?

I think many doctors might answer “smoking” first, because it’s their Great Satan of preventable disease. Smoking is a fairly big one, but it is not the right answer: it ranked third. Depression (in later life) came fourth, on the statistical premise that 100 per cent of depression is preventable, rather than more like zero per cent. Physical inactivity and social isolation, which I suppose everyone knows instinctively to be contributors to mental decline, came further down the list — and so did high blood pressure.

You should feel pretty good if you guessed the Lancet’s second most prominent preventable cause of dementia, because it is, like the winner, a weird one: lack of formal education in childhood. This is a “lifestyle factor” that people in the First World don’t have to worry about. The risk (thought to represent an increased likelihood of around 60 per cent) comes from not having any secondary education at all, which means you left school before age 11 or thereabouts. It has a high PAF in the table because 40 per cent of the world’s population still has this risk factor. But the highest PAF of all belongs to...

... peripheral hearing loss. That’s right: poor hearing may be the single largest preventable contributing factor to dementia. If this was your answer, you must be pretty darn good at keeping up with dementia research. But there’s a catch.

Studies suggest that persons who suffer peripheral hearing loss in late midlife, at around age 55, are at almost double the ordinary person’s risk of becoming demented. The doubling of risk is independent of other health and social factors, including isolation, as far as anyone can tell. (“Peripheral” deafness is hearing loss that is without an obvious cause in the brain or central nervous system. In plain English, it is hearing loss that a hearing aid would fix.)

This is, as of 2017, just an observed correlation. The first study to find the connection was published as recently as 1989. In the small but increasing literature on the mystery, it doesn’t yet seem to make any difference what hearing test is used, or what measure of dementia. Most importantly, as the Lancet’s group confesses, no one knows whether hearing aids can help stop cognitive decline. No one is even sure how to find out: it is easy to prescribe a hearing aid, and real hard to be certain a patient is using it.

This means many explanations for the correlation are still possible. Midlife deafness might indeed contribute to dementia: some neurologists speculate that the work of trying to interpret speech steals “cognitive resources” from memory. Others more or less suggest that being hard of hearing makes you sad and the sadness makes you old more quickly. But hearing loss might also just be an early sign of dementia. If mild deafness in middle age and later senility have a cause in common, this would mean you can’t fix one by fixing the other.

And this would, in turn, mean the Lancet’s guess that one-third of dementia cases are preventable might be way off. They are using headline-grabbing numbers to gain attention for what might be an incredibly important public health phenomenon. (Family physicians are already being encouraged to diagnose hearing loss less passively.)

Medical journals, in general, should probably not get up to this sort of thing, but... well, Lancet invented this tactic, and did it almost two centuries ago. Professional readers, knowing this, give the journal a whole lot of latitude and a little eye-rolling. Those of us out in the fog have to be careful.