# Genetic drivers of progression in Alzheimer's disease are distinct from disease risk.

- **DOI:** [10.1186/s13195-026-02036-1](https://doi.org/10.1186/s13195-026-02036-1)
- **Rating:** Limited (2.8 out of 5)
- **Verdict:** Amyloid-confirmed longitudinal genetics offers useful progression hypotheses, but unresolved internal inconsistencies prevent firm conclusions about APOE and progression-specific biology.
- **Framework version:** v5.9
- **Pipeline version:** v0.4
- **Evaluation date:** 2026-10-08
- **Evaluation page:** https://www.nabu.science/paper/10.1186%2Fs13195-026-02036-1

## Key claims

1. Established AD risk factors, including APOE‑ε4 dosage and polygenic risk scores, were not associated with disease progression in amyloid positive individuals who are actively declining.
   - Support: Tentative
   - Type: Primary finding
   - Basis: model-dependent APOE results and limited power prevent a firm null conclusion
   - Location: Abstract, Results; Results, APOE and Polygenic Risk Scores; Tables 2–3
2. A mixed effects GWAS meta-analysis revealed one genome-wide significant locus on chromosome 22 (rs78369883) and several nominally significant loci linked with AD progression.
   - Support: Supported
   - Type: Secondary
   - Basis: the reported associations are directly identified, though the small unreplicated discovery sample limits certainty
   - Location: Abstract; Results, Meta-analysis; Table 4
3. These progression-related genes were significantly enriched in neuronal-interferon-microglial signalling pathways and normal homeostatic processes of neuronal networks, with specific enrichment in dopaminergic and inhibitory neuronal populations.
   - Support: Tentative
   - Type: Secondary
   - Basis: enrichment rests on genes mapped from nominal discovery loci without independent biological validation
   - Location: Abstract; Results, Implicated Pathways; Figures 7–9
4. the genetic drivers of AD pathogenesis are, at least mostly, distinct from those modulating progression.
   - Support: Tentative
   - Type: Secondary
   - Basis: the comparison relies on underpowered nulls rather than a direct, replicated risk-versus-progression analysis
   - Location: Discussion, Uncoupling Genetic Factors of AD Risk and Progression; Title

## Strengths

- **Amyloid-confirmed longitudinal progression phenotype** Participants were selected using cohort-specific amyloid criteria and analysed through repeated MMSE measurements, focusing the study on decline after pathological confirmation.
  - Dimension: contribution
  - Location: Methods, Patient Filtering; Figure 1
- **Broad longitudinal genetics analysis pipeline** The analysis combines random-intercept and random-slope models, cohort-specific GWAS meta-analysis, fine-mapping, colocalisation, and cell-type enrichment.
  - Dimension: methodological rigour
  - Location: Methods, Mixed-effects Modelling through Co-localisation Analysis
- **Contrasting literature directly addressed** The Discussion compares its APOE and PRS findings with prior progression analyses and identifies amyloid restriction and baseline-stage adjustment as possible explanations for disagreement.
  - Dimension: positioning
  - Location: Discussion, Uncoupling Genetic Factors of AD Risk and Progression

## Limitations

- **Headline APOE result remains contradictory** The Abstract reports no APOE association with progression, while the nonlinear model reports significantly improved fit and approximately 16% faster decline per ε4 allele. Table 2 also does not clearly display the stated APOE×time coefficient.
  - Dimension: reporting
  - Location: Abstract; Results, APOE; Table 2; Figure 4
- **Reported statistics do not reconcile** Table 6 includes beta, standard-error, and p-value combinations that are arithmetically incompatible, while PRS counts and locus totals vary across methods, results, and tables.
  - Dimension: methodological rigour
  - Location: Methods, Polygenic Risk Scores; Results, Meta-analysis and Fine-mapping; Tables 3, 4, and 6
- **Exploratory biology receives strong interpretation** Nominal, unreplicated loci support pathway and cell-type hypotheses, but the Discussion describes distinct neuronal drivers and validation across annotation datasets more firmly than the evidence permits.
  - Dimension: positioning
  - Location: Results, Implicated Pathways; Discussion, Cell Types Implicated and Study Limitations

## Reliability checks

Caveats (4 of 4 checks).

- **Results check:** Noted
  - Established AD risk factors, including APOE‑ε4 dosage and polygenic risk scores, were not associated with disease progression in amyloid positive individuals who are actively declining: "Established AD risk factors, including APOE‑ε4 dosage and polygenic risk scores, were not associated with disease progression in amyloid positive individuals who are actively declining." (Abstract, Results) vs "The inclusion of APOE‑ε4 dosage as a predictor of the decline-rate parameter significantly improved model fit (likelihood ratio test p = 0.014), corresponding to an estimated ~ 16% acceleration in MMSE decline per ε4 allele." (Results, APOE). The unqualified absence-of-association claim is contradicted by the paper's nonlinear model, which reports an APOE-ε4 decline-rate association at p = 0.014 and approximately 16% acceleration per allele. → Not supported by the paper's own data
  - A mixed effects GWAS meta-analysis revealed one genome-wide significant locus on chromosome 22 (rs78369883) and several nominally significant loci linked with AD progression: "Table 4Summary statistics of meta-analysis independent signals at p < 5 × 10–05." (Table 4, caption) vs "22:26,417,956rs761670TC0.4550.0436.85 × 10–05MYO18B, SEZ6L, TFIP11" (Table 4, rs761670 row). Counting Table 4 gives 23 rows, of which rs761670 has p = 6.85 × 10–05 above the stated 5 × 10–05 cutoff, leaving one genome-wide significant row and 21 additional rows below that cutoff rather than 22 additional qualifying independent signals. → Conclusion holds
- **Reporting and transparency:** Noted. Reporting gaps concern the APOE interaction label, power intervals at different sample sizes, PRS counts, fine-mapping statistic definitions, the colocalisation sample size, and an unnamed GitHub repository; these do not establish additional corrected results.
  - Conflict of interest: not stated · Ethics approval: declared · Preregistration: not stated · Data availability: declared · Code availability: declared
- **Reference integrity:** Clean
  - 105 references in manuscript
  - 105 of 105 checkable references found in an index
  - References not verified automatically are resolved manually within 24 hours.

## Post-publication record

- **Status:** Clean
- **Summary:** No retraction or expression of concern found in Crossref (includes Retraction Watch).
- **Last checked:** 2026-10-17
