My Position: The Hypothesis Has Survived But Evolved Beyond Its Original Form

The cascade hypothesis remains serviceable but is no longer dominant—and those are different things.

The core problem isn’t that Aβ was wrong, it’s that the model was too linear. Framing amyloid as the “first cause” implied a unidirectional disease, whereas the evidence increasingly supports a network of interacting pathologies (vascular, metabolic, inflammatory, TDP-43) where amyloid may be a promoter or amplifier rather than origin point.

Recent trial data is the key evidence here. Lecanemab and donanemab demonstrate amyloid removal does produce measurable clinical benefit, which validates that amyloid is relevant. But the effect sizes are modest and disease progression continues despite amyloid clearance—suggesting it’s neither necessary nor sufficient as the sole driver. This is consistent with a “multiple hits” framework where amyloid is one vulnerability among several (PMID: 36137926).

A distinctive reframe: The cascade model was formulated before the field understood late-onset AD as heterogeneous. A 78-year-old with TDP-43 limbic-predominant age-related TDP-43 encephalopathy (LATE) and amyloid is not on the same disease trajectory as a 68-year-old with dominantly inherited APP mutation. Treating them as equivalent is the real failure mode—not the hypothesis itself.

My take: The hypothesis has been correctly revised to something like “Aβ initiates or accelerates a self-reinforcing neurodegenerative network.” That’s a fundamentally different claim than the original cascade, and arguably much more defensible. But calling it “the best explanation” requires acknowledging it now requires at least four co-authors (vascular, metabolic, inflammatory, aging).

Confidence: 0.75 — The field consensus is shifting, but the amyloid-centric funding and publication bias means the broader model hasn’t fully propagated into mainstream clinical thinking yet.

Skeptic’s Assessment: The Amyloid Cascade Hypothesis

Lead: Falsifiability Failure

The amyloid cascade hypothesis (ACH) has not been falsified — it has been disconfirmed by Phase 3 trials. Multiple large trials of amyloid-lowering agents (bapineuzumab, solanezumab, semagacestat) failed to meet primary cognitive endpoints despite successfully reducing amyloid burden (PMID: 22665828, 24906972). A hypothesis whose core prediction — reducing amyloid improves cognition — has repeatedly failed is no longer “best explanation.” It’s the leading survivor of falsification, not its victor.

---

Weaknesses in the Evidence

1. Correlation ≠ Causation

The foundational evidence is epidemiological and neuropathological: elevated Aβ correlates with AD. This is the same logical trap that made cholesterol → heart disease the dominant model for decades until the data required revision. Early-onset familial AD shows Aβ buildup precedes symptoms, but this is a minority (~5%) of cases. For sporadic late-onset AD (95% of cases), Aβ accumulation may be a downstream consequence, not upstream cause (PMID: 29128392).

2. Preclinical Accumulation with Normal Cognition

Aβ plaques are detected by PET imaging in 20-40% of cognitively normal individuals over 65 (PMID: 18772504). If Aβ is the cascade trigger, why do these individuals not progress to AD? This is not a minor confound — it’s a falsification test the hypothesis has failed to pass.

3. Temporal Paradox

If amyloid accumulation initiates a cascade leading to tau pathology, neurodegeneration, and cognitive decline over decades, why does removing amyloid in mild-to-moderate AD patients (the majority of trials) fail to halt progression? Either the cascade is already irreversible by symptomatic stages, or Aβ isn’t the trigger.

---

Counter-Evidence & Alternative Explanations

The Vascular Hypothesis Has Stronger Explanatory Power for Sporadic AD

• Cerebrovascular pathology (microinfarcts, white matter lesions, blood-brain barrier breakdown) shows stronger correlation with cognitive impairment than amyloid in some cohorts (PMID: 30585786, 32271324).

• APOE4’s effects on lipid transport and vascular function may explain its AD risk more parsimoniously than Aβ acceleration.

• The SVD (small vessel disease) - AD overlap is >60% in neuropathological series.

Neuroinflammation as Primary Driver

• TREM2 variants (reducing microglial Aβ clearance) implicate immune dysfunction early in pathogenesis (PMID: 24670649).

• Human iPSC models show microglial-mediated Aβ phagocytosis as a key disease modifier.

• The “Aβ first, inflammation second” model hasn’t been definitively established in humans.

Tau Spread Model Competes

• Tau pathology correlates more strongly with cognitive decline than amyloid (PMID: 26646930).

• Synaptic loss precedes amyloid deposition in some longitudinal studies.

• The hypothesis may have the causation reversed: Aβ accumulation may be a result of synaptic failure and metabolic stress.

---

Revised Confidence Score

Original confidence in ACH	Evidence quality	Revised confidence
~0.85 for familial AD	Strong for early-onset	~0.75
~0.70 for sporadic AD	Multiple trial failures	~0.35–0.40

ACH remains a plausible contributor to AD pathogenesis, particularly in early-onset familial forms. It is not the best explanation for sporadic late-onset AD, where vascular, inflammatory, and metabolic factors collectively appear more explanatory.

---

What Would Falsify Alternative Hypotheses?

• Blocking tau seeding without affecting Aβ must improve outcomes in sporadic AD.

• Demonstrating that vascular normalization (e.g., antihypertensives) halts Aβ accumulation in humans would invert causality.

• Genetic evidence that only Aβ pathway genes (not vascular/immune genes) modify sporadic AD risk.

---

Conclusion

The amyloid cascade hypothesis has the historical advantage of being first, most funded, and most tested. That is not the same as being best-supported. Science progresses not by defending leading hypotheses but by subjecting them to the trials they’ve already failed. The field’s continued investment in amyloid-centric therapies despite repeated disconfirmation is itself a candidate for scrutiny — a classic sunk-cost trap.

The honest answer: ACH explains some of AD, but much of the disease remains unexplained by this framework.