← Back
NeuroTrials.ai
Neurology Clinical Trial Database

REWIND-Dementia

Dulaglutide and neurodegeneration biomarkers: REWIND post hoc analysis

Year of Publication: 2026

Authors: Wilson JM, Dage JL, Qian HR, ..., Bethel MA

Journal: Alzheimer's & Dementia

Citation: Wilson JM, et al. Dulaglutide and neurodegeneration biomarkers: REWIND post hoc analysis. Alzheimer's Dement. 2026;22:e71391.

Link: https://doi.org/10.1002/alz.71391


Clinical Question

Does dulaglutide reduce plasma ADRD biomarkers (NfL, p-tau217, GFAP) and substantive cognitive impairment in adults with type 2 diabetes?

Bottom Line

In a post hoc analysis of REWIND, dulaglutide did not significantly change ADRD biomarkers or SCI overall, but in participants with elevated baseline NfL (≥56 pg/mL) it produced a significant 2-year reduction in NfL, and in those with elevated baseline p-tau217 (≥0.25 pg/mL) it was associated with a 22% reduction in SCI — suggesting potential benefit in biomarker-defined high-risk subgroups.

Major Points

  • Post hoc analysis of 7368 REWIND participants (dulaglutide n=3741; placebo n=3627) with baseline and 2-year plasma biomarkers
  • Overall, dulaglutide did not significantly change NfL, p-tau217, GFAP, or SCI
  • Significant treatment-by-baseline NfL interaction (p < 0.001): in those with NfL ≥56 pg/mL, dulaglutide reduced NfL at 2 years (−27.4% vs −14.2%; p = 0.003)
  • In participants with baseline p-tau217 ≥0.25 pg/mL (29% of cohort), dulaglutide reduced SCI by 22% (HR = 0.78; p = 0.0064)
  • Baseline NfL was negatively associated with baseline standardized MoCA and DSST scores after adjustment for age/sex (both p < 0.001)
  • p-tau217 and GFAP showed limited 2-year treatment-related changes overall
  • Findings support exploring incretin-related cognitive effects in biomarker-defined high-risk subgroups in future trials

Design

Study Type: Post hoc analysis of a multicenter, international, randomized, double-blind, placebo-controlled cardiovascular outcomes trial

Randomization: 1

Blinding: Double-blind; biomarkers measured without knowledge of clinical data or treatment assignment

Allocation: 1:1 dulaglutide vs placebo

Enrollment Period: August 2011 through August 2013

Follow-up Duration: Median 5.4 years (biomarkers at baseline and 2 years; cognitive testing at baseline, 2 years, 5 years, and end of study)

Centers: 371

Countries:

Sample Size: 9901

Analyzed: 7368

Analysis: Analysis of covariance for 2-year biomarker changes; Cox proportional hazards models for SCI; subgroup analyses by predefined biomarker cut-points; p-values not adjusted for multiplicity (exploratory)

Registration: NCT01394952


Inclusion Criteria

  • Age ≥50 years
  • Type 2 diabetes
  • Previous cardiovascular event OR cardiovascular risk factors
  • HbA1c ≤9.5% (≤80 mmol/mol)
  • BMI ≥23 kg/m²
  • Taking up to two oral glucose-lowering drugs, with or without basal insulin, at stable doses for ≥3 months
  • For this analysis: baseline + Year 2 biomarker data and ≥1 follow-up MoCA or DSST score

Exclusion Criteria

  • Coronary or cerebrovascular event within the previous 2 months
  • eGFR <15 mL/min/1.73 m² or renal dialysis
  • Severe hypoglycemia within the past year
  • Cancer within the past 5 years
  • History of pancreatitis
  • History of bariatric surgery
  • Known abnormal gastric emptying
  • Previous participation in any dulaglutide study

Baseline Characteristics

CharacteristicDulaglutide (n=3741)Placebo (n=3627)
Median Age (IQR), years65.4 (61.5–70.1)65.6 (61.4–70.0)
Female1790 (48%)1724 (48%)
White ethnic origin2884 (77%)2760 (76%)
Education ≤12 years2338 (63%)2267 (63%)
Current tobacco use527 (14%)508 (14%)
Diabetes duration, years (mean ± SD)10.1 ± 7.010.3 ± 6.9
Cardiovascular disease1151 (31%)1132 (31%)
Stroke or TIA342 (9%)329 (9%)
Atrial fibrillation240 (6%)231 (6%)
Heart failure335 (9%)342 (9%)
Hypertension3482 (93%)3378 (93%)
HbA1c, %7.3 ± 1.17.3 ± 1.1
BMI, kg/m²32.4 ± 5.732.3 ± 5.7
eGFR, mL/min/1.73 m²78.5 ± 23.677.3 ± 23.8
Albuminuria1249 (33%)1242 (34%)
Systolic BP, mm Hg137.0 ± 16.4138.0 ± 17.0
Diastolic BP, mm Hg78.8 ± 9.778.8 ± 9.9
LDL cholesterol, mmol/L2.6 ± 1.02.6 ± 1.0
SCI events735 (20%)780 (22%)
DSST score (median, IQR)36 (25–49)36 (25–48)
MoCA score (median, IQR)25 (23–28)25 (22–28)
NfL, pg/mL (median, IQR)26.1 (19.9–37.4)26.8 (19.7–37.7)
p-tau217, pg/mL (median, IQR)0.20 (0.16–0.26)0.20 (0.16–0.26)
p-tau217 ≥0.25 pg/mL1063 (28%)1042 (29%)
GFAP, pg/mL (median, IQR)148.1 (107.9–203.9)145.9 (107.4–199.2)

Arms

FieldDulaglutideControl
N37413627
InterventionDulaglutide 1.5 mg subcutaneously once weeklyMatching placebo subcutaneously once weekly
DurationMedian 5.4 yearsMedian 5.4 years

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Change in plasma ADRD biomarkers (NfL, p-tau217, GFAP) from baseline to 2 years and treatment-by-baseline-biomarker interactionsPrimaryNfL LS mean 30.1 ± 0.03 pg/mL (10.2% increase from baseline)NfL LS mean 29.8 ± 0.03 pg/mL (9.1% increase from baseline)Overall nonsignificant; significant treatment-by-baseline-NfL interaction (p < 0.001)0.47 (overall NfL change)
2-year change in NfL among participants with baseline NfL ≥56 pg/mLSecondaryLS mean 75.4 ± 3.0 pg/mL (−14.2%)LS mean 63.8 ± 2.5 pg/mL (−27.4%)0.003
Substantive cognitive impairment (SCI) in subgroup with baseline p-tau217 ≥0.25 pg/mLSecondary0.780.0064
Previously published overall SCI in REWIND (reference)Secondary0.860.0018
Association of baseline NfL with baseline standardized MoCA and DSST scores (adjusted)SecondaryNegative association<0.001 for both
Association of baseline p-tau217 and GFAP with baseline standardized MoCA/DSST scores after adjustmentSecondaryNonsignificant

Subgroup Analysis

Pre-specified biomarker cut-points: NfL ≥56 pg/mL (90th percentile); p-tau217 ≥0.25 pg/mL (AD-associated threshold); GFAP tertiles (<121, 121–180, >180 pg/mL). Significant dulaglutide effects were limited to NfL reduction in the high-NfL subgroup and SCI reduction in the high-p-tau217 subgroup.


Criticisms

  • Post hoc, exploratory analysis with p-values not adjusted for multiplicity
  • Biomarker measurements limited to baseline and 2-year timepoints
  • Overall biomarker and SCI effects were nonsignificant; significance limited to subgroups
  • Population restricted to T2D patients with cardiovascular risk; generalizability to non-diabetic AD populations unclear
  • SCI defined by cognitive test performance thresholds rather than clinical dementia diagnosis

Funding

Eli Lilly and Company

Based on: REWIND-Dementia (Alzheimer's & Dementia, 2026)

Authors: Wilson JM, Dage JL, Qian HR, ..., Bethel MA

Citation: Wilson JM, et al. Dulaglutide and neurodegeneration biomarkers: REWIND post hoc analysis. Alzheimer's Dement. 2026;22:e71391.

Content summarized and formatted by NeuroTrials.ai.