Forecasting Alzheimer’s: With a tsunami of new diagnoses expected, experts calculate the best ways to shift the tide

The forecast looms like a portent of doom. From 5.8 million people today, the number of Americans with Alzheimer’s disease is projected to reach 13.8 million by 2050, overwhelming caregivers and the health care system — a prospect that has produced alarm bordering on panic about an unstoppable Alzheimer’s tsunami.

Reality, however, is far more nuanced: Medical breakthroughs and other factors could dramatically reduce that number — though, paradoxically, such advances could also increase the prevalence of this most common form of dementia.

The ominous “13.8 million people with Alzheimer’s by 2050,” publicized most prominently by the Alzheimer’s Association, has driven home to policymakers and others the urgent need for more Alzheimer’s research: The National Institutes of Health more than tripled its spending on the disease since 2015, to $2.4 billion this year.

But there has been little discussion of how different ways of spending that money —  improving cardiovascular health, increasing education, developing drugs that act at the earliest stages of disease, or at later ones —  might affect the Alzheimer’s toll decades hence. With little quantitative research on the impact of different interventions, there has been scant effort to set spending or research priorities. Instead, the attitude has been, “everything is important.”

That’s true. But not everything is equally important.

STAT therefore asked eight experts on the demographics of Alzheimer’s which interventions, or even unintended changes, might affect the projected prevalence of the disease, by how much, and when. We also used software made available by biostatisticians at UCLA to calculate how developments ranging from a generally healthier population to the discovery of an effective drug would affect future Alzheimer’s prevalence.

What emerged is that “13.8 million by 2050” is far from locked in. Changes made today can change tomorrow.

The most impactful intervention, according to the mathematical model created by UCLA biostatisticians Ron Brookmeyer and Nada Abdalla and confirmed by others, would be primary prevention — that is, keeping healthy brains from taking the first pathological steps toward Alzheimer’s. The benefits of primary prevention are largely long-term, however, with a significant reduction in cases not evident for at least 15 years.

In contrast, the discovery of a drug that stops mild cognitive impairment from progressing to full-on Alzheimer’s would have a more immediate effect. But such “secondary prevention,” at least in the UCLA analysis, “results in the smallest decrease in Alzheimer’s cases decades from now,” Abdalla said.

In fact, while desperate patients and their families ardently wish for a drug that slows the progression of the disease, such an intervention would have the counterintuitive effect of increasing the total number of Alzheimer’s patients: More people would remain longer in the early stage of the disease rather than progressing to the later, often fatal stage.

The future is not all bleak, however. If people in middle age or even young adulthood adopted a slew of healthy habits, and if scientists developed drugs that kept brains with the first hints of Alzheimer’s pathology from becoming functionally impaired, and if other drugs could stop the tragic march from slight cognitive impairment to full-on Alzheimer’s, and if that triple whammy were instituted in 2025, then the number of cases of Alzheimer’s in 2050 would not be 13.8 million. (See graph below.) It would be 3.15 million.

(Some of the difference reflects the fact that the UCLA model counts 2 million fewer Alzheimer’s cases in 2019 than the Alzheimer’s Association does, and projects fewer cases in a nothing-changes scenario, too, with just over 8 million cases in 2050. The three-part intervention would reduce that by more than 5 million.)

“Rigorous estimates of future Alzheimer’s incidence, prevalence, and costs … can guide the research enterprise into putting resources toward the highest priority areas,” said Matthew Baumgart, vice president of health policy at the Alzheimer’s Association. In 2015, the association estimated that if a hypothetical treatment postponed by just five years when people developed Alzheimer’s, the number of cases in 2050 would be 7.8 million, not more than 13 million.

An ounce of prevention …

Some measures would take far longer than others to affect the case count. For instance, adopting at least four healthy habits (such as exercising both mind and body and not smoking) reduces the risk of Alzheimer’s by 60%, researchers led by Dr. Klodian Dhana of Rush University Medical Center reported at this year’s Alzheimer’s Association annual conference. Healthy habits don’t reverse the disease in people who already have it. But they can keep healthy brains from developing the pathology that causes Alzheimer’s.

Similarly, there are hints that antiviral drugs can also stave off the brain pathology that precedes the clinical symptoms of Alzheimer’s. A 2018 study found that such drugs reduced by 90% the risk of developing the disease.

That sounds like a godsend. But even if so many Americans lived such a healthy life that the number of brains developing Alzheimer’s pathology fell by the full 60%, let alone 90%, Alzheimer’s cases would nevertheless increase for 15 years after this public health miracle, according to the UCLA model.

The reason is that primary prevention, by definition, keeps healthy brains healthy: They do not accumulate the protein fragment called amyloid, their neurons do not die, and their synapses aren’t lost. People whose brains already have those signs of early disease, but are not yet showing symptoms (which can take a decade or more), would unfortunately remain on the path to full-blown Alzheimer’s despite primary prevention; the horse has already galloped out of the barn. Only a decade or longer after healthy living or a preventive drug has kept Alzheimer’s pathology from developing in millions of elderly people who would otherwise have it would the disease’s toll be lower than it would have been.

“It’s years before you get the full benefit of delaying the brain damage that causes Alzheimer’s because of the long lag between [that damage] and clinical Alzheimer’s,” said Brookmeyer, a longtime expert in Alzheimer’s projections who led the modeling project.

Realistic primary prevention — assuming it started now and cut the rate at which healthy brains become amyloid-ridden brains by 75% — would reduce Alzheimer’s prevalence in 2050 to 4.96 million, as shown in the graph below:

26 million or 8 million?

Even the most basic inputs can change prevalence projections. One might expect that, absent a sudden unexpected change in mortality rates, future population should be easy to calculate: Today’s 40-year-olds are 2050’s 71-year-olds. But surprises can pop up. In 2002 demographers projected that in 2010 there would be 39.4 million Americans 65 or older. The 2010 census counted 40.3 million. That changed the 2050 projection of elderly from 81.7 million to 88.5 million, researchers at the Chicago Health and Aging Project reported in 2013.

That increase in turn boosted the 2050 Alzheimer’s projection by 600,000. Think of that as a need for 6,000 more nursing homes and 500,000 more licensed nurses than previously thought.

The Chicago projection reflects the percentages of people in different age groups who have Alzheimer’s today. But because the disease can be diagnosed definitively only after death, by examining the brain, the number of current patients — and therefore those key percentages — is only an approximation.

Experts estimate that about 10% of people over 65 have Alzheimer’s. But among those 65 to 74, the rate is 3%. It’s 17% in people 75 to 84, and 32% in those 85 and older. Applying the current prevalence rates to the projected elderly population produces that widely cited 13.8 million in 2050.

Overall, Brookmeyer and his colleagues estimated in a 2018 study, there were nearly 6 million people in the U.S. with Alzheimer’s, including the earliest mild cognitive impairment, in 2017. But because of diagnostic uncertainty, it could be as few as 3.1 million or as many as 11.6 million. Present uncertainty produces future uncertainty: The Alzheimer’s prevalence in 2060, Brookmeyer calculated, could be as low as 8.2 million but as high as 26.1 million, a huge range that’s almost never mentioned.

“There are well-established criteria for an Alzheimer’s diagnosis, but the trouble is they can be interpreted differently by different clinicians,” said Dr. Denis Evans of Rush University Medical Center, who has studied Alzheimer’s prevalence for decades. “The number of people with Alzheimer’s at any one time is therefore indefinite.”

The challenge of predicting Alzheimer’s prevalence decades hence has, oddly, become more difficult as both demographers and neurologists learn more about the disease and its risk factors. For instance, the generation born in the 1950s is developing Alzheimer’s at a rate 44% lower than those born in the 1910s through 1930s, according to the long-running Framingham Heart Study. In a group of United Kingdom counties that have been studied for decades, the prevalence of Alzheimer’s fell from 8.3% in 1991 to 6.5% in 2011.

How the decreasing risk across generations plays out depends on why people who are 75 today are less likely to have Alzheimer’s than people who were 75 in the 1990s. (It also depends on whether the Framingham and U.K. findings, based mostly on white populations, apply to other ethnic groups. If not, they might be too non-representative of the U.S. population to offer much hope that Alzheimer’s rates will fall significantly as today’s 50- and 60-somethings age.)

If the hopeful phenomenon is real, one possible explanation is that today’s just-barely-elderly have better cardiovascular health than earlier generations at the same age, thanks to lower rates of smoking and better blood pressure and cholesterol control. Those factors also reduce the risk of Alzheimer’s. But better cardiovascular health lowers the risk of premature death, too. That means more people survive into their 80s and 90s, ages with the highest rates of Alzheimer’s, points out epidemiologist Andrew Patton, a graduate student at the Johns Hopkins Bloomberg School of Public Health who projected dementia rates in Marin County, Calif.

Better heart health therefore has countervailing effects on Alzheimer’s prevalence: It reduces the percentage of people who develop it, but extends lives, producing a larger base population at high risk for the disease.

Another possibility is that the rate of Alzheimer’s is dropping in later generations because they have higher levels of education. That protects against Alzheimer’s, scientists believe, because it creates a larger “cognitive reserve”: Two brains may be equally ravaged by neuronal and synaptic loss, but the brain with a greater store of knowledge has a larger cushion before memory loss and cognitive impairment become apparent. Despite such brains’ biological wreckage, Alzheimer’s is delayed.

If education is the main protective effect explaining the generational decline, that’s most likely to affect the rate at which someone with mild cognitive impairment, or the even-earlier asymptomatic accumulation of amyloid and loss of synapses, progresses to Alzheimer’s. But if the main influence is better cardiovascular health, that would likely act earlier, Brookmeyer said, perhaps reducing the formation of amyloid deposits.

“If something changes to delay the progression from mild cognitive impairment to Alzheimer’s, or reduce the number of people [who worsen in that way], we’ll see that immediately in prevalence numbers,” he said.

Billion-dollar breakthrough

The biggest unknown is whether an effective Alzheimer’s drug will ever arrive, and if so how well it works and on which patients — those with mild cognitive impairment or more severe dementia. Halving the rate at which amyloid-ridden brains progress to mild cognitive impairment would markedly reduce Alzheimer’s cases starting almost immediately, and would decrease Alzheimer’s prevalence in 2050 to 4.96 million, as shown in the graph below:

The effect of a multibillion-dollar breakthrough — a drug that keeps mild cognitive impairment from becoming full-on Alzheimer’s — would also have a near-immediate effect: If the drug prevents a 65-year-old with MCI from developing Alzheimer’s, that’s one less case of the disease almost right away. A drug that becomes available in 2025 and lowered by a mere 25% the number of people going from mild cognitive impairment to Alzheimer’s — a realistic estimate for what an “effective” Alzheimer’s drug could accomplish — would reduce Alzheimer’s prevalence in 2050 by about 1 million, the UCLA model shows:

Perhaps the most significant advance in research on future Alzheimer’s prevalence is recognizing and quantifying the effects of cross-cutting changes that, at first glance, seem like they should be purely positive. Health economist Julie Zissimopoulos of the University of Southern California has led the development of a mathematical model that captures the opposing effects of risk factors.

Two of the leading factors with such Janus-like effects are diabetes and hypertension. Both raise the risk of Alzheimer’s. But both can also kill people early, before they reach their peak Alzheimer’s risk years. “Many people die of diabetes before they progress into dementia,” said economist Michael Hurd of RAND. But if diabetes cuts short fewer lives, then more people will live to an age when their chances of developing Alzheimer’s peaks. The rate of Alzheimer’s dementia soars after age 65.

For instance, a 2018 study found that adopting five healthy habits (not smoking, drinking little to no alcohol, exercising 30 minutes or more daily, eating healthily, and maintaining a normal weight) increases the life expectancy of a 50-year-old woman from 29 years (that is, living to 79) to 43 years. For men, life expectancy at 50 increased from 25.5 years to 37.6. Of every 1,000 79-year-olds, almost 24 will develop Alzheimer’s in the year they reach that age; of every 1,000 93-year-olds, the comparable figure is just under 86.

Because of these longevity effects, the USC model finds, decreasing the rate of diabetes (which is largely driven by obesity) starting now actually increases the number of people projected to have Alzheimer’s in 2040: 11.78 million if diabetes rates drop by 50%, and 11.66 million if they remain at today’s levels. Halving rates of hypertension would have similar consequences, Zissimopoulos and her colleagues calculated: 11.86 million people with Alzheimer’s in 2040, as the Alzheimer’s-increasing effects of living longer outweigh the Alzheimer’s-reducing effects of blood pressure control, as shown in the graph below:

“For policymakers, clinicians, and others, it’s important to think about these trade-offs and interactions,” Zissimopoulos said. “If we do this, what happens?”

That’s not to say public health and other efforts to combat diabetes and high blood pressure have no value. Reducing the incidence of either would reduce suffering and premature deaths. But society should not expect a reduction in rates of diabetes or hypertension to also avert an Alzheimer’s boom. In general, the great unknown for future prevalence is, which is greater: the reduced risk of Alzheimer’s from better health in middle and old age, or the increased risk of Alzheimer’s as better health keeps people alive longer, and well into the most Alzheimer’s-prone years?

“Let’s say you’re 80,” said RAND’s Hurd. “Exercising decreases your risk of Alzheimer’s. But does it change your lifetime risk?” That is, is the lower risk of Alzheimer’s at age 80 and even 85, due to healthier habits, swamped by the higher risk that comes with older old age? “The effect is going in two different directions,” he said.

Modelers struggle to capture that because the magnitude of each effect — how much does exercising reduce the risk of Alzheimer’s vs. how many years of life does it buy? — is not known precisely. That is one of many challenges to projecting the number of Americans with Alzheimer’s disease in the years to come.