New study raises alarm over Alzheimer’s blood tests

Research from Rutgers Health indicates that blood tests for Alzheimer’s disease need to be interpreted with caution—particularly for Black patients.

“These tests are currently geared towards primary care physicians and directly to older adults concerned about cognitive performance,” said William Hu, director of the Rutgers Center for Healthy Aging Research, and Chief of Cognitive Neurology & Alzheimer’s Disease Clinic and senior author of the paper published in Alzheimer’s & Dementia.

“They may provide some value to primary care doctors who understand their limitations, but there is no reason for patients who have concerns about their memory to buy these tests, which cost $1,200 to $2,000 and are almost never covered by insurance.”

Concentrations of an Alzheimer’s biomarker are significantly lower in blood than in spinal fluid, where they have been validated as a diagnostic tool in white patients. This difference is particularly large for Black patients, who naturally have lower levels of the biomarker in their spinal fluid. Worse memory function itself also raises blood biomarker levels.

“The blood tests have many caveats,” Hu said. “Poor transfer of these proteins from spinal fluid to blood means many patients will go undiagnosed, especially if their disease is mild, while the ability of other ailments to increase these proteins in the blood mean tests produce false positives.”

The researchers examined test results from patients at Rutgers and Emory University. Both cohorts included Black and non-Hispanic white participants. The Rutgers cohort also included ethnic Chinese participants.

“For Black patients, it’s a double whammy,” Hu said. “Not only will you have a harder time using these proteins in the spinal fluid to diagnose Alzheimer’s disease in Black patients, you will further reduce the detection rate by relying on the blood tests.”

All 221 participants underwent detailed clinical evaluations and brain imaging. Most had both blood and spinal fluid samples collected on the same day. The research team used automated testing systems to measure levels of specific proteins—particularly p-Tau₂₁₇—in both blood and spinal fluid samples.

The researchers found a 70% correlation between protein levels in spinal fluid and blood. The racial disparities, moreover, were significant. When using a blood p-Tau₂₁₇ threshold optimized for detecting Alzheimer’s, the test showed 90.3% sensitivity (chance of diagnosing patients with disease) and 81.1% specificity (chance of correctly identifying someone as not having the disease) in white participants but 73.7% sensitivity and 72.5% specificity in Black participants.

The positive predictive value—the chance of actually having the disease when the test is positive—was much higher in white participants (87%) than Black participants (58%). These differences persisted even after accounting for known racial differences in spinal fluid protein levels, suggesting the blood test introduced additional disparities beyond what was already known.

Among patients of Chinese ancestry, test performance roughly matched its performance in white patients.

Hu added the tests may eventually prove useful but need significant refinement.

“In another five to 10 years, these tests may get to the point where we can reliably use them, but right now, they are similar to some of the home COVID tests that had accuracy issues,” he said.

The researchers emphasized that anyone with memory concerns should see a neurologist for a comprehensive evaluation rather than relying on blood test results alone. They said roughly half of all Americans with Alzheimer’s are thought to be undiagnosed, particularly those with early-stage cases that can benefit most from treatment.

Research from Rutgers Health indicates that blood tests for Alzheimer’s disease need to be interpreted with caution—particularly for Black patients.

“These tests are currently geared towards primary care physicians and directly to older adults concerned about cognitive performance,” said William Hu, director of the Rutgers Center for Healthy Aging Research, and Chief of Cognitive Neurology & Alzheimer’s Disease Clinic and senior author of the paper published in Alzheimer’s & Dementia.

“They may provide some value to primary care doctors who understand their limitations, but there is no reason for patients who have concerns about their memory to buy these tests, which cost $1,200 to $2,000 and are almost never covered by insurance.”

Concentrations of an Alzheimer’s biomarker are significantly lower in blood than in spinal fluid, where they have been validated as a diagnostic tool in white patients. This difference is particularly large for Black patients, who naturally have lower levels of the biomarker in their spinal fluid. Worse memory function itself also raises blood biomarker levels.

“The blood tests have many caveats,” Hu said. “Poor transfer of these proteins from spinal fluid to blood means many patients will go undiagnosed, especially if their disease is mild, while the ability of other ailments to increase these proteins in the blood mean tests produce false positives.”

The researchers examined test results from patients at Rutgers and Emory University. Both cohorts included Black and non-Hispanic white participants. The Rutgers cohort also included ethnic Chinese participants.

“For Black patients, it’s a double whammy,” Hu said. “Not only will you have a harder time using these proteins in the spinal fluid to diagnose Alzheimer’s disease in Black patients, you will further reduce the detection rate by relying on the blood tests.”

All 221 participants underwent detailed clinical evaluations and brain imaging. Most had both blood and spinal fluid samples collected on the same day. The research team used automated testing systems to measure levels of specific proteins—particularly p-Tau₂₁₇—in both blood and spinal fluid samples.

The researchers found a 70% correlation between protein levels in spinal fluid and blood. The racial disparities, moreover, were significant. When using a blood p-Tau₂₁₇ threshold optimized for detecting Alzheimer’s, the test showed 90.3% sensitivity (chance of diagnosing patients with disease) and 81.1% specificity (chance of correctly identifying someone as not having the disease) in white participants but 73.7% sensitivity and 72.5% specificity in Black participants.

The positive predictive value—the chance of actually having the disease when the test is positive—was much higher in white participants (87%) than Black participants (58%). These differences persisted even after accounting for known racial differences in spinal fluid protein levels, suggesting the blood test introduced additional disparities beyond what was already known.

Among patients of Chinese ancestry, test performance roughly matched its performance in white patients.

Hu added the tests may eventually prove useful but need significant refinement.

“In another five to 10 years, these tests may get to the point where we can reliably use them, but right now, they are similar to some of the home COVID tests that had accuracy issues,” he said.

The researchers emphasized that anyone with memory concerns should see a neurologist for a comprehensive evaluation rather than relying on blood test results alone. They said roughly half of all Americans with Alzheimer’s are thought to be undiagnosed, particularly those with early-stage cases that can benefit most from treatment.