By Deborah Borfitz

January 16, 2024 | Things could start looking up for randomized controlled clinical trials (RCTs) in Alzheimer’s disease (AD) prevention, which have a 25-year track record of failure, once broad biomarkers of the brain-robbing condition are found that reflect real-life risk factors, according to Sandrine Andrieu, M.D., Ph.D., professor of public health and clinical epidemiology at Toulouse University Hospital (France). Studies today are less often measuring the impact of pharmaceuticals than they are multidomain interventions which, while promising, introduce new challenges in establishing validity and utility, she says. 

Andrieu, a leader in the field of age-related decline who has led several major AD trials, was speaking at the recent Clinical Trials on Alzheimer’s Disease conference in Boston. To date, 112 AD prevention trials have been published variably testing pharmacological, physical activity, nutrition, cognitive, and multidomain interventions, and she offers an overview of progress across those lines of investigation.  

The interventions have notably included antihypertensive and anti-amyloid drugs, antioxidants, and omega-3 fatty acids, aerobic training, and cognitive stimulation. Results of multidomain intervention trials have not appeared frequently in the literature, but it’s a popular tactic in many ongoing studies, Andrieu says. 

Many of the AD trials conducted to date have shown negative results on the primary endpoint—decreasing the incidence of dementia or, more recently, decreasing cognitive decline. Positive results have been seen with a secondary outcome or subgroup analysis, says Andrieu, but many of the large trials were stopped early due to adverse events from an investigative drug. Fewer than 20% of the prevention trials have had positive results and they were predominantly small-scale, short-term studies assessing efficacy based on changes in cognition.  

The high number of trial failures has many possible methodological causes, she points out, including choosing the wrong window of exposure to risk factors or an unsuitable outcome for a lifestyle intervention. Some trials may have lost power due to selection bias or the intervention or outcome being analyzed, and others may have used the wrong statistical analysis tools for detecting late-occurring events or tiny signals. Another reality is that many of the trialists didn’t have the advantage of previously published RCTs upon which to base their initial assumptions. 

Life Exposures

Risk factors in epidemiological studies are presented as life exposures, some of which have a negative impact during midlife and others during late life, says Andrieu. AD prevention trials have attempted to capture the accumulated exposures at a specific time point.  

In one of the earliest dementia prevention trials (The Lancet, DOI: 10.1016/s0140-6736(98)03086-4), and one of the only ones to have a positive result on the primary outcome, an anti-hypertensive drug was found to decrease the incidence of dementia by 50%. But very few of the subjects in the trial had AD, she says.    

Most recently, a 2,389-subject trial looking at the combined use of an antihypertensive drug and a statin, plus or minus aspirin, showed no effect on cognition but a positive effect on functional decline (JAMA Neurology, DOI: 10.1001/jamaneurol.2022.5088). The 2019 SPRINT-MIND trial, which was assessing the efficacy of intensive versus standard blood pressure treatment likewise had no impact on dementia incidence but when the population was followed for three years showed an effect on the incidence of mild cognitive impairment and probable dementia (JAMA, DOI: 10.1001/jama.2018.21442). That trial stopped early when another study was published showing the benefit was much higher in the population at risk of developing the disease, says Andrieu.   

A multidomain intervention trial for managing cardiovascular risk factors also found no effect on dementia incidence but after multiple years of follow-up showed a positive effect on a subgroup at risk for developing AD (The Lancet, DOI: 10.1016/S0140-6736(16)30950-3), she notes.  

The inclusion criteria of these large trials nicely cover individuals in their life course from mid- to late life, continues Andrieu. But the final assessment is done on the population in late life, which may not be the critical window for intervening. There is also a “competitive risk” for death from cardiovascular causes and dementia. 

Drug Trials

Anti-amyloid drugs are a high-interest area in the pharmacological treatment arena, Andrieu reports. Unfortunately, as covered in a study recently published in The New England Journal of Medicine (DOI: 10.1056/NEJMoa2305032) solanezumab failed to slow cognitive decline relative to placebo in people with preclinical AD. That’s likely because the drug was not strong enough to impact the amyloid loads, she remarks. 

Trials for BACE-1 inhibitors have been stopped due to adverse effects, says Andrieu, quickly adding that a promising trial is on the way that will assess the efficacy of lecanemab in asymptomatic subjects with an elevated amyloid level in their brain. Results of that trial are expected in the next few years. 

Two large trials have assessed the efficacy of low-dose aspirin, one enrolling over 15,000 people with diabetes and another 19,000 individuals without any risk factors. “Both trials were negative on cognitive outcomes, including dementia and cognitive impairment,” she says. 

A trio of trials assessing the efficacy of anti-inflammatory drugs reported “huge adverse events in the [study] population,” continues Andrieu. Safety, as well as reduction in amyloid load, is particularly important in prevention studies since most participants will never develop the disease. 

Non-pharmacological Treatments

Four small, short-term trials of physical activity interventions have been published showing a positive effect on cognition. The same can’t be said for the big LIFE trial (JAMA, DOI: 10.1001/jama.2015.9617), where researchers found physical activity had no effect on cognition among 1,635 subjects followed for two years. The problem with this trial, and others like it, is that over time there is a decline in the self-reported walking activities of people, Andrieu says.  

Unlike in a drug trial, not all subjects are naïve to the intervention. “Too little attention was paid to the level of physical activity exposure before the trial,” says Andrieu. Additionally, the positive cognitive effects of physical activity, or increasing levels of it, may well take many years to accrue.  

In 2023 alone, a nutrition intervention was the focus of three big trials, including the MIND-Diet trial finding no significant difference in cognitive effect from following the MIND diet—a hybrid of the Mediterranean diet and the DASH (Dietary Approaches to Stop Hypertension) diet—and a control diet with mild caloric restriction after three years (The New England Journal Of Medicine, DOI: 10.1056/NEJMoa2302368).  The failure, she reasons, is that the two diets are remarkably similar.  

Results from a pair of ancillary COSMOS studies were more encouraging. The first (COSMOS-Mind) assessed, by phone, the efficacy of cocoa extract (including flavanols) and multivitamin-mineral supplements on cognitive outcomes over three years (Alzheimer’s & Dementia, DOI: 10.1002/alz.12767). Only the latter had a positive effect, improved global cognition, episodic memory, and executive function. 

The second study (COSMOS-WEB) focused on multivitamin-mineral supplements in a larger and more diverse cohort and moved cognitive testing to the internet (The American Journal of Clinical Nutrition, DOI: 10.1016/j.ajcnut.2023.05.011). It showed a positive effect on the primary outcome of improving memory in older adults. 

In each of the COSMOS trials, the positive effects were greater in the population at risk of developing AD, Andrieu notes. It is also “much easier to take a supplement than to change the behavior of subjects.” 

Important information about cognitive stimulation intervention comes primarily from a trial published in 2006 (JAMA, DOI: 10.1001/jama.296.23.2805). Researchers found that memory, reasoning, and speed-of-processing training had no effect on cognition and activities of daily living after three years, but at five years and even more so at 10 years they found a positive effect from the approach, she shares.  

Multidomain Interventions 

Multidomain interventions represent a new and important field, says Andrieu, referencing the DoHealth trial assessing the efficacy of vitamin D supplementation, omega-3 fatty acids, and physical activity that failed to show any effect on cognition after three years. But another study looking at the value of “behavioral activation” in preventing cognitive decline in Black individuals with mild cognitive impairment found a positive effect on episodic memory (JAMA Neurology, DOI: 10.1001/jamaneurol.2018.2513). 

Andrieu was lead investigator on the Multidomain Alzheimer Preventive Trial (MAPT) trial assessing the efficacy of cognitive stimulation, physical activity, and nutritional advice plus omega-3 fatty acid, which showed no effect on the primary cognitive outcome after three years (The Lancet: Neurology, DOI: 10.1016/S1474-4422(17)30040-6). “Since that time, we analyzed a subgroup that was more at risk to develop AD based on positive amyloid PET or... the biomarker Aβ42/40 ratio... [who] have a benefit from the intervention,” she says. 

Then there was the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) trial, the first large study of a multidomain intervention looking at the effectiveness of nutrition, physical activity, cognitive training, social activity—together with intensive monitoring of cardiovascular risk factors by a nurse—on cognition in a population at risk for AD. The trial showed a positive result after two years based on a change in a neuropsychological test battery Z score (The Lancet, DOI: 10.1016/S0140-6736(15)60461-5). 

Such trials typically select extremely healthy participants because of the rigors of participation that can make adherence difficult to maintain over time, Andrieu says. An analysis of adherence in the MAPT and FINGER trials revealed a positive association with more passive components of the intervention (e.g., taking a supplement and the evaluation of cardiovascular risk factors) that fell away when it came to the more demanding components (e.g., cognitive stimulation).  

Study Design

AD prevention studies don’t have many of the standard features of the gold-standard RCTs, says Andrieu. “We don’t have a double-blind intervention,” she offers as an example. “At the best, we have a blind assessment of the primary outcome.” The interventions themselves are often not very standardized, she continues. “[But] at least we changed a few points on cognitive decline,” a clinically relevant effect. Contrary to most RCTs, the population selected for AD prevention trials also tends to be individuals less likely to need clinical management due to a decline in cognitive function, Andrieu adds.  

What such trials should do, she says, is analyze the impact of exposure to risk factors as they happen in real life. It remains unknown how the exposures impact cognition and if the effects are cumulative or people are more sensitive to them during a certain part of their life. Epidemiological studies follow a population for a decade or two to understand such associations, so perhaps it is unsurprising that AD prevention trials that attempt to change a lifetime of habits over a much shorter period so often fail. 

Future Perspectives

The way forward should become clearer with forthcoming results from 116 AD prevention trials now underway, many testing a multidomain intervention and a growing number using a digital device to monitor cognitive function over time, reports Andrieu. Larger trials are needed to allow for statistically significant subgroup analysis and to, at minimum, do a long-term assessment of the outcome to detect disease-modifying effects of an intervention. 

Importantly, she adds, studies need to stop conflating the maintenance of brain health and AD prevention, which are two different things. Up to now, enrolled study populations have included older people with cognitively normal aging, 30% of whom have amyloid plaques in their brain. 

Andrieu advocates for a move to broad biomarkers and using biological age as a “sort of summary” of various risk exposures and resiliency to them. She is currently involved in an AD prevention trial analyzing cognition together with multiple other cognition-dependent functions. 

A multidomain intervention is being analyzed for its ability to postpone not just loss of cognition but also mobility, hearing, vision, and mental health, says Andrieu. The study will utilize brain biomarkers and assess efficacy based on changes in biological age. 

“We have made tremendous progress in understanding the disease since the 1980s,” she concludes, drawing a contrast with cardiovascular research where breakthroughs came only after decades of study.  AD prevention trials need to reach the target population and interventions must be deployed at the population level with adherence being assessed over time. “It is more than science,” she says.