Vitals Knowledge Vault

GLP-1 Non-Responder Genetic Variants

7 min read

GLP-1 Non-Responder Genetic Variants

TL;DR

Some GLP-1 receptor (GLP1R) genetic variants are associated with modest differences in weight-loss response to GLP-1 agonist drugs. The largest study (Nature 2026, n=27,885) found a GLP1R missense variant associated with approximately 0.76 kg additional weight loss per copy of the effect allele — meaning a two-copy carrier loses roughly 1.5 kg more than a non-carrier. Average weight loss on semaglutide or tirzepatide is 15–20 kg. The genetic effect is therefore <10% of total treatment response and does not explain clinical non-response. A second large GWAS (DIRECT, Lancet Diab Endocrinol 2023) identified ARRB1 — not GLP1R — as the strongest HbA1c-response signal. Genetic testing is not clinically indicated for GLP-1 prescribing by any guideline body.

Why it matters for Vitals

  • Coaches: Genetic variants explain only a small fraction of individual response variation. Adherence, dose titration, diet, and baseline metabolic health remain far more important levers.
  • Users on GLP-1s: A 23andMe or similar DTC report of a GLP1R variant does not mean you are or are not a responder. The effect size is too small for individual prediction.
  • Users experiencing poor response: Early weight trajectory (4–8 weeks) is the most evidence-backed early non-response signal. No wearable biomarker has been validated for genetic non-response detection.
  • Tirzepatide users with GI side effects: A GIPR variant is associated with increased nausea/vomiting specifically in tirzepatide users. This may be relevant if someone tolerates tirzepatide poorly relative to semaglutide.
  • Clinical boundaries: Vitals coaches should not recommend or discourage GLP-1 therapy based on genetic variants. Escalate any genetic-data-related questions to the prescribing clinician.

Key Facts

ClaimEffect Size / DetailEvidence Grade
GLP1R missense variant (rs1030559) → more weight loss−0.76 kg/copy (P = 2.9×10⁻¹⁰)High — but single source
Two-copy carrier advantage over non-carriers~1.5 kg more weight lossConfirmed
Average semaglutide/tirzepatide weight loss15–20 kg (~15–20% body weight)Confirmed
Genetic contribution to response<10% of total treatment effectConfirmed
GIPR variant → more nausea on tirzepatidetirzepatide-specific; not semaglutideHigh — single source
ARRB1 (β-arrestin-1) variant → better HbA1c response0.25%/copy (P = 5.2×10⁻⁶); strongest DIRECT GWAS signal — not GLP1RHigh — multi-cohort
TCF7L2 variant → weight signal in liraglutide RCTp=0.015Reported (small RCT)
Clinical non-response explained by geneticsNo — effect sizes far too smallConfirmed
Clinical utility trial for genetic testingNoneGap
Any guideline recommends genetic testing for GLP-1 prescribingNo — not in ADA, AACE, FDA, EMA, MHRAConfirmed

What the current evidence suggests

GLP1R Missense Variant (Nature 2026)

The variant (rs1030559) sits at the exendin(9-39) binding site in the third extracellular loop of the GLP-1 receptor — a region critical for ligand binding and receptor activation. In vitro functional assays suggest enhanced downstream signaling per unit ligand, consistent with the weight-loss signal. The effect is pharmacodynamic, not pharmacokinetic — receptor sensitivity is altered, not drug clearance.

Study characteristics: n=27,885, self-reported GLP-1 RA users (semaglutide, tirzepatide, others), predominantly European ancestry, all 23andMe employees or research participants.

ARRB1 as Primary HbA1c Signal (DIRECT GWAS, 2023)

The DIRECT GWAS (n=4,571, multi-cohort) found the strongest pharmacogenetic signal for HbA1c response was an ARRB1 (β-arrestin-1) low-frequency missense variant (Thr370Met): 0.25%/copy greater HbA1c reduction (P = 5.2×10⁻⁶). The GLP1R Gly168Ser variant showed only a 0.08% HbA1c effect. This directly contradicts the Nature 2026 paper’s framing of GLP1R as the primary signal — the two largest GWAS point to different primary genes for different outcomes (weight vs. HbA1c).

GIPR Variant and Tirzepatide Nausea

A GIPR variant is associated with increased nausea/vomiting specifically in tirzepatide users — not in semaglutide users. This is the strongest evidence that the finding reflects GIPR agonism (unique to tirzepatide), not a GLP-1 mechanism. Tolerability differences between tirzepatide and semaglutide in variant carriers may warrant agent-switching discussions with a prescribing clinician.

Conflicting and Null Findings

  • Chinese exenatide cohort (n=285): rs10305420 T allele associated with less weight loss (+1.27 kg) — opposite direction to Nature 2026.
  • Greek cohort (n=191): No carriers of GLP1R rs367543060 found; TCF7L2/CTRB1/2 non-significant.
  • Liraglutide 3 mg RCT (n=136): TCF7L2, not GLP1R, as weight signal (p=0.015).
  • Diabetologia 2021 review (22 studies): Most prior candidate-gene studies were underpowered and inconsistent.

Effect Size Context

The −0.76 kg/copy GLP1R effect means:

  • Zero-copy (no effect allele): baseline response
  • One-copy carrier: ~0.76 kg more weight loss
  • Two-copy carrier: ~1.5 kg more weight loss

Average weight loss on semaglutide 2.4 mg: ~15–20 kg (15–20% body weight in a 100 kg person). The genetic effect represents <10% of total treatment response — far too small to explain the ~10% of patients who are clinical non-responders.

Mechanism Summary

  • GLP1R variant: Alters receptor conformation at the exendin(9-39) binding site → enhanced downstream signaling per unit ligand → modestly greater efficacy.
  • ARRB1 variant: β-arrestin-1 is involved in GLP1R desensitization and β-arrestin-dependent signaling. A low-frequency missense variant (Thr370Met) may alter β-arrestin recruitment efficiency — a mechanistically distinct pathway from direct ligand binding.
  • GIPR variant: Modulates GI side-effect sensitivity specifically in the presence of GIPR agonism (tirzepatide, not semaglutide).
  • PK: No evidence that any of these variants affect pharmacokinetics. GLP-1 agonists are large peptides cleared primarily by proteolysis, not CYP enzymes subject to genetic variation.

Risks and Uncertainty

High-severity limitations

  • Single-source for GLP1R finding: Nature 2026 is the only study reporting the −0.76 kg/copy effect. No independent replication exists.
  • All authors are 23andMe employees: Conflict-of-interest concern; self-selection in cohort.
  • Self-reported outcomes: Weight and side effects were self-reported in the primary GWAS — recall and social desirability bias concerns.
  • ARRB1 vs. GLP1R contradiction: Two largest GWAS point to different primary genes for different outcomes — the pharmacogenomic landscape is not resolved.
  • Effect size too small to explain non-response: ~1.5 kg difference for two-copy carriers is not clinically meaningful for explaining the ~10% who are non-responders.
  • No clinical utility trial: No study has tested whether genotype-guided prescribing leads to better outcomes.
  • European ancestry overrepresentation: Findings may not generalize to other ancestries.

Evidence grades by claim

ClaimGrade
GLP1R rs1030559 → weight loss signal in large GWASHigh (statistically robust; single source)
GIPR variant → tirzepatide nauseaHigh (biologically plausible; single source)
ARRB1 Thr370Met → strongest HbA1c signalHigh (multi-cohort DIRECT GWAS)
Polygenic model for patient stratificationModerate (proof-of-concept; not replicated)
Genetics explains clinical non-responseFalse — effect size too small
Genetic testing clinically indicated for GLP-1 prescribingFalse — no guideline supports this
Genetics applies equally across ancestriesUnknown — no data

Wearable / Vitals Monitoring Implication

Genetic non-response detection via wearables: not actionable. No wearable-derived metric (HRV, RHR, sleep architecture, CGM) has been shown to detect or predict GLP1R or ARRB1 genetic non-response.

Early weight trajectory monitoring (evidence-backed):

  • Patients losing <2–3% body weight at 12 weeks are likely poor long-term responders.
  • This is implementable now with connected scales — no genetic data required.
  • Action: if trajectory suggests non-response, consider dose escalation or agent switching in consultation with the prescribing clinician.
def assess_glp1_response(weight_baseline, weight_12wk):
    pct_loss_12wk = ((weight_baseline - weight_12wk) / weight_baseline) * 100
    return "possible_non_response_signal" if pct_loss_12wk < 2.5 else "on_track"

What Vitals Can and Cannot Say

✅ Can say:

  • “Early weight trajectory is the best available predictor of long-term GLP-1 response”
  • “The GLP1R genetic variant has a modest effect on weight loss (~1.5 kg for two-copy carriers)”
  • “If you have concerns about response, discuss your genetic data with your prescribing clinician”

❌ Cannot say:

  • “Genetics explains why you may not respond to GLP-1 therapy”
  • “This genetic test can guide your GLP-1 treatment choice”
  • “Your wearable data shows you are a genetic non-responder”
  • “This variant means you should switch GLP-1 agents”

Sources

  • PMID 41951734 — Ashenhurst et al., Nature, April 8, 2026: “Genetic predictors of GLP1 receptor agonist weight loss and side effects”
  • PMID 36528349 — DIRECT GWAS, Lancet Diabetes Endocrinol, 2023: ARRB1 Thr370Met as strongest HbA1c pharmacogenetic signal
  • PMID 30883264 — Chinese exenatide cohort, Pharmacogenomics, 2019
  • PMID 38453649 — Greek cohort, Postgrad Med, 2024
  • PMID 35894080 — Liraglutide 3mg RCT, Obesity, 2022
  • PMID 33594477 — Diabetologia systematic review of 22 GLP-1 pharmacogenetic studies, 2021

Last updated: 2026-04-21 (Batch 85, vault promotion)

Graph View

Backlinks