Towards Timely Alzheimer’s Diagnosis: A Clinical Data-Driven Approach to Biomarker Discovery
Smita Sahay1, Ali Imami1, Abdul Hamoud1, Robert McCullumsmith1, Sinead O'Donovan1
1Department of Neurosciences, The University of Toledo College of Medicine and Life Sciences
Objective:

Determine lab-based biomarkers for early Alzheimer's disease diagnosis based on retrospectively collected critical care medical data.

Background:
Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by a progressive decline in cognitive function and memory. Early symptoms are often non-specific. Traditional diagnosis relies on clinical and cognitive assessments, which are subjective and imprecise. Thus, there is a critical need for objective and reliable biomarkers for the detection of AD. Our study proposes lab-based biomarkers that may aid in the early diagnosis and management of AD.
Design/Methods:

We utilized the MIMIC-IV database that includes critical care data from 40,000 racially and ethnically diverse patients admitted to intensive care units at the Beth Israel Deaconess Medical Center. For each lab (n=1,600+), MIMIC-IV was filtered for patients with a primary diagnosis of AD and age- and sex-matched non-psychiatrically ill control (CTL) patients. Each case was matched to four CTLs. Means and standard deviations were calculated. Welch’s two sample t-test was used to compare means in AD vs. CTL groups. Effect sizes (ES) were calculated to further assess the practical significance of our results.

Results:
Of 477 labs, 51 were significantly different between AD and CTL groups (p<0.05). Of these, 22 had a moderate-large ES (range 0.27-1.16). Of 227 labs exclusive to females, 22 were significantly different between female AD and CTL groups (p<0.05). Of these, 6 had a moderate ES (0.35-0.48). Lastly, of 210 labs exclusive to males, 22 were significantly different between male AD and CTL groups (p<0.05). Of these, 9 had a moderate ES (0.26-0.49).
Conclusions:
Potential lab-based biomarkers for AD were in the following broader diagnostic categories: kidney function (urea nitrogen), cardiovascular (cholesterol), and electrolyte/metabolic balance (chloride, urine glucose). Interestingly, while females and males had overlapping labs, males were distinguished by hematological labs, indicating the importance of considering sex in biomarker analysis for AD.
10.1212/WNL.0000000000205096