Volunteering was linked to slower cognitive decline in APOE4 Alzheimer's gene carriers among 17,972 adults

Genes are the part of dementia risk that nobody gets to choose. One variant in particular, APOE4, raises the odds of Alzheimer’s disease more than any other common stretch of DNA, and people who learn they carry it often ask what, if anything, they can do about it.

Sae Hwang Han of the University of Texas at Austin and Jessica Faul of the University of Michigan have tested one candidate that costs nothing: helping other people. Their analysis, published in Alzheimer’s & Dementia in September, followed 17,972 Americans for as long as 22 years. Carriers of the variant lost ground on cognitive tests more slowly after they took up volunteering, and more quickly after they gave it up.

What is APOE4, and who carries it?

APOE is a gene involved in moving fats around the body and brain, and it comes in three common versions. A large genetic study in JAMA Neurology describes two of them as the strongest protective and risk-increasing, common genetic variants for late-onset Alzheimer disease. APOE4 is the risk-increasing one.

Carrying it is common. In the new study, 72.5% of participants did not carry the APOE variant, about a quarter had one copy and roughly 1 in 40 had two. Risk rises with each copy, and so, the authors confirmed, does the speed at which test scores fall with age.

What did volunteering change for APOE4 carriers?

The data come from the Health and Retirement Study, a long-running survey that interviews Americans over 50 every two years. Each time, participants take a short cognitive test scored out of 27, covering word recall, mental subtraction and counting backward. They are also asked a plain question: “Have you spent any time in the past 12 months doing volunteer work for religious, educational, health-related or other charitable organizations?”

Han and Faul did not simply compare volunteers with everyone else. They tracked the moments when individuals started or stopped, and asked what happened to each one’s own trajectory afterward. The survey recorded thousands of transitions into and out of formal volunteering and informal helping over the two decades.

For carriers, starting to volunteer went with a gentler slope of decline. The reverse held too: withdrawing from formal volunteering was associated with steeper decline tied to the variant. The amounts involved were small. Benefits appeared in people giving under 100 hours a year and in those giving 100 to 199, which works out at no more than about four hours a week.

More was not better. Those who jumped to 200 hours a year or beyond showed no such link, a result the authors call consistent with prior evidence that excessive helping may generate role strain that offsets cognitive benefit.

The differences on the test are fractions of a point per survey wave, so the paper offers an illustration to give them scale. For a 60-year-old with two copies of the variant who starts volunteering and keeps it up, sustained volunteering corresponds to reaching the cognitive impairment threshold about two years later than the same person would without it.

Everyday help outside any organization, such as driving a neighbor to an appointment, told a narrower story. Across the full sample it showed no clear association. Among Hispanic participants it did, and the authors write that informal helping was uniquely protective among Hispanic older adults. That group was the smallest in the study, which makes the estimate less certain.

Is volunteering protecting the brain, or do sharper people volunteer?

This is the obvious objection, and other researchers have put numbers on it. Anyone whose thinking is already slipping is less likely to sign up for a shift at the food bank, which would make volunteers look sharper without volunteering doing anything. Ben Lennox Kail and Dawn Carr examined exactly that in the same survey. They found that selection explained between 4.9% and 29% of the effect of volunteering on cognitive function, which leaves most of the association standing but confirms that part of it is an illusion.

Han and Faul guard against the problem in two ways. They compare each person with themselves over time, and they check whether a change in volunteering predicts scores at a later survey wave. On the second test the authors write that the pattern argues against reverse causation as a sole explanation. That wording stops short of ruling it out.

Independent work points the same way. A separate team using the same survey, led by Frank Infurna at Arizona State University, reported in 2016 that volunteering at the initial assessment and volunteering regularly over time independently decreased the risk of cognitive impairment over 14 years. Han’s own earlier analysis of 31,303 people found that transitioning into volunteering and informal helping were both associated with a higher level of cognitive function. Speaking about that work last year, Han also described the other direction: “Conversely, our data show that completely withdrawing from helping is associated with worse cognitive function.”

Only one study has assigned people to volunteer at random, the design that can show cause. The Baltimore Experience Corps Trial placed older adults in elementary schools as tutors, and its imaging arm randomized 111 men and women. After two years, men in the Experience Corps arm showed a 0.7% to 1.6% increase in brain volumes in regions vulnerable to dementia, while the same regions shrank with age in men in the comparison group. The result was weaker in women, and the trial measured brain size, not thinking or genes.

A selected sample, a short test and no dementia diagnoses: limits of the volunteering study

Nobody was told to volunteer. Health, energy, transport and free time all shape who can, and the analysis adjusts for wealth, work, marital status, disability and depressive symptoms without being able to capture everything.

The outcome was a score on a brief telephone-style test, so the analysis tracks decline in thinking and has no data on who went on to a dementia diagnosis. Volunteering was reported by the participants themselves, once every two years. And the people analyzed were those who gave a saliva sample between 2006 and 2012; the authors caution that the analytic sample is a selected subset of the HRS core sample, which leaves out anyone who died before DNA collection began.

The finding also applies to a gene’s effect on a slope. It says carriers who volunteered declined more slowly than carriers who did not, with no claim that volunteering cancels the risk.

What the study changes is the range of things known to travel with slower decline in the people at highest inherited risk. A few hours a week of organized helping now sits on that list as an association, with one small randomized trial pointing in the same direction and no trial yet in APOE4 carriers.