Study reveals new ways to predict onset of Alzheimer’s disease

LOS ANGELES - Alzheimer’s disease may not begin with a single red flag—but rather with patterns that unfold in a predictable sequence. That’s the finding of a new UCLA Health study, which offers fresh insight into how the disease develops and how clinicians could better predict who’s most at risk.

Instead of looking at isolated health conditions, researchers used advanced data modeling to map how various diagnoses accumulate over time before Alzheimer’s onset. Their work, published July 7 in eBioMedicine, could lead to more personalized prevention strategies and earlier interventions.

The study analyzed nearly 25,000 patients in the University of California Health system and confirmed its findings using data from the All of Us Research Program—a national cohort representing diverse populations across the U.S.

What are the four pathways to Alzheimer’s disease?

The backstory:

Researchers identified four distinct diagnostic sequences that frequently lead to Alzheimer’s disease. Each reflects a different type of progression:

• Mental health pathway: Psychiatric diagnoses like depression increase the risk of cognitive decline.
• Encephalopathy pathway: Brain dysfunction conditions accumulate and intensify over time.
• Mild cognitive impairment pathway: Early memory and thinking problems gradually worsen.
• Vascular disease pathway: Heart-related issues, including hypertension, contribute to dementia risk.

Each pathway had unique clinical and demographic traits, suggesting that different groups may face different risks depending on their medical history.

What we know:

About 26% of Alzheimer’s progressions followed identifiable diagnostic sequences.

These patterns, such as hypertension followed by depression, predicted Alzheimer’s onset more reliably than single diagnoses.

Advanced methods—including machine learning and network analysis—helped identify these progression models.

What we don't know:

It’s unclear how soon these diagnostic trajectories can be translated into clinical screening tools.

More research is needed to test whether interrupting these sequences can delay or prevent Alzheimer’s.

The pathways show correlation, not causation—so other factors may still be at play.

How this research could change early Alzheimer’s care

What they’re saying

"We found that multi-step trajectories can indicate greater risk factors for Alzheimer’s disease than single conditions. Understanding these pathways could fundamentally change how we approach early detection and prevention," said Mingzhou Fu, first author and pre-doctoral student in medical informatics at UCLA.

Lead author Dr. Timothy Chang, assistant professor of Neurology at UCLA Health, said:

"Recognizing these sequential patterns rather than focusing on diagnoses in isolation may help clinicians improve Alzheimer’s disease diagnosis."

What's next:

Researchers say these trajectory models could support:

• Risk stratification, identifying who’s most likely to develop Alzheimer’s
• Targeted intervention, disrupting dangerous diagnostic chains early
• Personalized care, based on the pathway a patient is following

The next step will be translating these insights into tools doctors can use in real-time to flag risk earlier and possibly intervene before symptoms appear.