Why Blood Tests Are a Poor Screening Tool for Alzheimer’s

As a primary care doctor who has been in the same practice for almost 30 years, my patients have aged with me. Those who were in their 60s when I started are now, shockingly, in their 90s. This reality makes the concern for dementia a common clinical problem. We are making strides in the diagnosis and treatment of Alzheimer’s disease, but these strides have created important questions regarding clinical application.

This post addresses one very important issue. It reminded me of a great post we had back in October 2025 by Alex Burns on Decontextualized Risk Information.

Adam Cifu

There is a growing effort to frame Alzheimer’s disease as a problem of late detection. The proposed solution is straightforward: test earlier, ideally in primary care, using blood-based biomarkers that can identify Alzheimer’s pathology before symptoms become obvious. It is an appealing idea. It is also a fundamental inversion of how this disease is responsibly evaluated.

The difficulty in Alzheimer’s care is not that the pathology is hard to detect. It is that pathology does not map cleanly onto symptoms, prognosis, or treatment decisions. Detecting a biological signal is the final step in a reasoning process, not the first.

In clinical practice, cognitive evaluation begins with phenotype and trajectory: which domains are affected, how function has changed, the tempo of decline, and what collateral history reveals. Structural imaging follows, not to “confirm Alzheimer’s,” but to define pattern—regional vulnerability, vascular burden, and anatomic concordance with the clinical syndrome. Only after this groundwork is complete does biological testing become meaningful, because only then is there a specific question the test is asked to answer.

Blood-based biomarkers reverse this sequence. When ordered upfront, they generate a biological signal without clinical anchoring. A positive result immediately raises questions it cannot resolve: Is this signal causal or incidental? Is it sufficient to explain the presentation, or merely one contributor among several? Should discordant findings prompt escalation or restraint? These questions are unavoidable, and answering them requires precisely the context that screening bypasses.

This is not a controversial claim: clinicopathologic and longitudinal cohort studies have repeatedly shown substantial discordance between Alzheimer’s biomarkers, clinical syndrome, and functional trajectory, particularly in older adults. Discordance is not unusual; it is the norm. Alzheimer’s pathology is common with aging and frequently coexists with vascular disease, sleep disorders, mood disorders, and other neurodegenerative processes. Biomarkers often fail to align with one another, and even when they do, they do not determine the rate of decline, the degree of neuronal injury, or the likelihood of treatment benefit. In practice, it is common to see patients referred after early blood-based testing who are functioning independently yet now navigating anxiety, referrals, and confirmatory studies triggered by a result that cannot, on its own, explain symptoms or predict course. The biological signal does not eliminate clinical uncertainty; it often amplifies it.

The limitations of a screening mindset become even clearer when treatment is considered. Anti-amyloid monoclonal antibody therapy is often portrayed as the logical endpoint of early detection. In reality, initiating therapy requires confirmatory testing with PET or cerebrospinal fluid, careful MRI review for hemorrhagic risk, and frank discussion of modest benefit weighed against real harm. Ongoing surveillance is required to distinguish disease progression from treatment effect or complication. These decisions are not triggered by test results; they are sustained by clinical judgment.

Seen this way, Alzheimer’s disease resembles conditions that require staged subspecialty evaluation—oncology, epilepsy surgery, movement disorder phenotyping—far more than diseases amenable to broad screening. In those fields, identifying a biological signal does not decentralize care; it concentrates it. Alzheimer’s is no different simply because blood tests now exist.

The industry logic behind early testing is understandable. When a biomarker becomes measurable, availability is easily mistaken for obligation. Screening feels proactive; caution feels like delay. But medicine has been here before. Broad cancer blood tests followed a similar trajectory: early enthusiasm, unclear actionability, and patients left with results that generated anxiety without guiding care. (See Dr. Cifu’s excellent series on the Galleri test.) Alzheimer’s biomarker evangelism risks repeating that mistake—confusing detection with understanding.

Primary care clinicians are essential to cognitive care. Recognizing concern, performing and documenting brief screening (MMSE or Mini-Cog) when appropriate, addressing reversible contributors, and referring patients for specialized evaluation are all critical steps. What undermines care is reversing the order—introducing complex biological data before the clinical framework exists to interpret it. In practice, this rarely accelerates diagnosis. More often, it creates interpretive debt that must later be unwound. Early detection is only valuable when it leads to early understanding or treatments that are more effective when given early than late. In Alzheimer’s disease, understanding emerges from the synthesis of clinical presentation, trajectory, imaging, biology, and patient priorities. That synthesis is the work. Blood tests can support it, but they cannot replace it.

If we want to improve Alzheimer’s care, we should be skeptical of narratives that frame it as a screening problem waiting for better tools. The harder truth is that it is a sequencing problem, and doing it backwards does patients no favors.

---

Travis S. Allen, MD, is a PGY-3 neurology resident at HCA Healthcare TriStar Skyline Medical Center in Nashville, Tennessee. The views expressed are the author’s own and do not necessarily reflect those of his employer or affiliated institutions.

Photo Credit: Steven HWG