New research from Vanderbilt University Medical Center, reported in Alzheimer’s & Dementia, shows that Alzheimer’s disease (AD) blood tests now arriving on the market might produce inaccurate results in the one-third of older adults who have kidney impairment, notably including those with only mild impairment.
“We found that even mild levels of kidney dysfunction affected the levels of several blood-based biomarkers for Alzheimer’s disease,” said study leader Corey Bolton, PsyD, assistant professor of Medicine in the Division of Geriatric Medicine and a member of the Vanderbilt Memory and Alzheimer’s Center. “The effect was most pronounced for a biomarker of neurodegeneration called neurofilament light, where we saw that the biomarker was no longer predictive of changes in cognition or brain imaging when we looked only in people who had mild to moderate kidney disease.”
Blood testing for AD is expected to soon become widespread. When an AD diagnosis is in doubt, clinicians sometimes take recourse to brain imaging using PET (positron emission tomography) or the analysis of cerebral spinal fluid collected by lumbar puncture (also known as spinal tap). Both tests are expensive. An initial AD blood test was approved in May by the Food and Drug Administration, and additional candidate tests are now under regulatory review. The Alzheimer’s Association has issued a call to AD researchers to examine potential confounders of these arriving tests, and their appeal makes specific reference to various diseases that may accompany AD.
According to the Centers for Disease Control and Prevention, 34% of Americans age 65 or older have chronic kidney disease (CKD).
The VUMC observational study involved data collected over an average follow-up of 6.4 years from 333 participants in the Vanderbilt Memory and Aging Project. The team began by replicating results in the literature showing the usefulness of four blood biomarkers in predicting AD neuropsychological and neuroimaging results: amyloid beta, phosphorylated tau, glial fibrillary acidic protein, and neurofilament light chain (NfL). The team proceeded to look for significant effects on these predictions from CKD as measured by a common clinical test called estimated glomerular filtration rate, or eGFR. (Lacking participants with stages 4 or 5 CKD, the study measured effects in mild to moderate CKD.)
NfL’s ability to predict cognitive test results was generally lost in participants with mild to moderate CKD. NfL wasn’t the only AD blood biomarker found to broadly interact with kidney function when predicting cognitive outcomes, but the many effects seen in the other three biomarkers became statistically insignificant when results were pooled across the study’s seven cognitive measures.
While effects from CKD on blood biomarker prediction of white and gray matter results from MRI (magnetic resonance imaging) were again broad, after correction for multiple comparisons, NfL again emerged as the sole AD biomarker whose predictive power was significantly compromised.
“Blood testing has great potential to advance clinical evaluation of AD, but this calls for careful estimation of the effects of kidney dysfunction and other potential confounders of these tests,” Bolton said. “In future studies we hope to better understand what exactly causes kidney dysfunction to affect AD biomarker levels in the blood, and why some biomarkers are more affected than others. Using this information, we hope to be able to improve our biomarkers or suggest new guidelines to ensure that the tests are equally effective for all patients, regardless of medical comorbidities.”
Others on the study from VUMC include Panpan Zhang, PhD, Devika Nair, MD, MSCI, Dandan Liu, PhD, L. Taylor Davis, MD, Kimberly Pechman, PhD, Niranjana Shashikumar, Sydney Wilhoite, Dominic Roby, Carlie Beeson, Haley Komorowski, MS, Timothy Hohman, PhD, and Angela Jefferson, PhD.
The study was supported in part by the Alzheimer’s Association, the Veterans Health Administration Office of Research and Development, and the National Institutes of Health under awards R01AG034962, R01AG056534, K24AG046373, K23AG045966, F32AG076276 and K23AG084850.
