A massive Vanderbilt-led study of electronic health records has pinpointed dozens of medical conditions that tend to show up years before an Alzheimer’s diagnosis. The findings could help doctors spot risk earlier and test new ways to delay or prevent the disease.
Alzheimer’s disease may leave clues in a person’s medical history a decade or more before memory problems appear, according to a large new study led by Vanderbilt Health.
By mining millions of electronic health records and genetic data, researchers identified dozens of conditions that tend to show up more often in people who later develop Alzheimer’s. The work, published in Alzheimer’s Research & Therapy, could help doctors recognize risk earlier and guide future efforts to delay or prevent the disease.
Alzheimer’s is a progressive brain disorder that develops slowly over many years, often long before symptoms like memory loss and confusion become obvious. Scientists have known that some midlife conditions, such as high blood pressure and high cholesterol, are linked to higher Alzheimer’s risk later on. But the full range of medical issues that might signal trouble ahead has not been clear.
The goal was to build a much more complete picture of those early warning signs, according to co-corresponding author Xue Zhong, a research assistant professor of medicine in the Division of Genetic Medicine and Clinical Pharmacology.
“If we know the full inventory of medical conditions that predict Alzheimer’s disease development 10 or more years later, we can potentially intervene before clinical symptoms of memory and/or cognitive impairment become apparent,” Zhong said in a news release.
To do that, the team turned to two huge troves of de-identified health records.
First, they used MarketScan, a U.S. claims database that includes information on more than 150 million people, as a discovery set. From that resource, they identified 43,508 people with an Alzheimer’s diagnosis and 419,455 similar people without the disease, matched by age and sex.
Then they checked their findings in Vanderbilt Health’s own electronic health record system, which includes about 3 million patients. In that second group, they studied 1,320 people with Alzheimer’s and 12,720 matched controls.
For each person with Alzheimer’s, the researchers looked back over a 10-year window before diagnosis and compared their medical histories to those of people who did not develop the disease. They focused on which conditions appeared more often in the group that eventually received an Alzheimer’s diagnosis.
Across both databases, the team found more than 70 conditions that consistently showed up more frequently in people who later developed Alzheimer’s.
Many of those conditions fell into four broad categories:
- Mental health conditions, including depression and severe neuropsychiatric symptoms such as paranoia, psychosis and suicidal thoughts.
- Neurologic and sleep-related conditions, such as insomnia, excessive sleepiness and sleep apnea.
- Cardiovascular and circulatory conditions, including essential hypertension, cerebral atherosclerosis and reduced blood flow in the brain.
- Endocrine and metabolic conditions, including Type 2 diabetes.
The researchers then went a step further, asking whether some of these conditions might share genetic roots with Alzheimer’s. They used data from two large DNA biobanks, Vanderbilt’s BioVU and the UK Biobank, to examine how individual genetic variants and an overall polygenic risk score for Alzheimer’s related to the conditions they had identified.
That analysis highlighted 19 conditions that were associated either with specific Alzheimer’s-linked genetic variants or with a higher overall genetic risk score for the disease. While those links do not prove that any one condition causes Alzheimer’s, they suggest that certain health problems and the disease may be connected at the genetic level.
The authors emphasized that associations in electronic health records cannot, by themselves, establish cause and effect. But they can reveal patterns that might otherwise be missed and help set priorities for future research.
“Longitudinal EHRs offer a powerful view into the decades-long development of Alzheimer’s disease,” Zhong added. “By identifying medical patterns that consistently precede Alzheimer’s disease, we can unlock new opportunities for risk reduction, early intervention and improved patient outcomes.”
The study also reinforced some messages that doctors already share with patients. Zhong noted that the results confirm high blood pressure and high cholesterol as risk factors for late-life Alzheimer’s, suggesting that managing these conditions in midlife — through healthier lifestyles or medications — could help lower risk.
In addition to risk factors, the team found a pattern that has puzzled scientists for years: people with cancer diagnoses were less likely to develop Alzheimer’s, and vice versa. Zhong noted this inverse association appeared in both health record datasets and lined up with earlier population studies.
“We also observed an inverse association between cancer and Alzheimer’s disease across both EHR datasets, replicating prior epidemiologic findings,” added Zhong. “We are now investigating the mechanisms underlying this phenomenon, with the goal of generating insights that could inform novel therapeutic strategies for Alzheimer’s disease.”
The findings offer a data-driven roadmap for earlier risk recognition. In the future, clinicians might use combinations of common conditions — such as depression, sleep disorders and cardiovascular problems — to flag patients who could benefit from closer monitoring, lifestyle counseling or enrollment in prevention trials, long before memory issues arise.
For now, the study underscores a hopeful message: what happens in midlife matters for brain health decades later. Keeping blood pressure, cholesterol and blood sugar under control, treating mental health and sleep problems, and staying engaged with regular medical care may all play a role in reducing the chances of developing Alzheimer’s.
As researchers continue to probe the links between these conditions, genetics and brain changes over time, the Vanderbilt-led work shows how the medical records already being collected in clinics and hospitals can be turned into powerful tools in the fight against one of aging’s most feared diseases.
Source: Vanderbilt Health