GLP-1 Medications in 2026: What New Clinical Data Reveals About Tirzepatide, Semaglutide, Weight Regain, and Long-Term Use

Medical Disclaimer: This article is for informational and educational purposes only. Semaglutide and tirzepatide are FDA-approved prescription medications. All prescribing decisions should be made in partnership with a licensed healthcare provider. This article does not constitute medical advice.

When semaglutide and tirzepatide earned FDA approval for weight management, the clinical evidence came primarily from trials measuring outcomes at 68–72 weeks. That was enough to establish efficacy and initial safety. It was not enough to answer the questions that matter most to patients and prescribers: How do the drugs compare directly? What happens when you stop? How durable are the cardiovascular benefits? Do these drugs protect the kidneys? Are there safety signals emerging from broader real-world use?

The 2024–2026 data cycle has provided substantive answers to each of these questions — and expanded the recognized benefit profile of GLP-1 drugs well beyond weight loss. This post synthesizes the most clinically important new findings.

The Head-to-Head: SURMOUNT-5

For years, patients and prescribers have asked which drug is “better” — semaglutide or tirzepatide. Until late 2025, that question had no direct answer: the approval trials used placebo comparators, and any comparison between drugs required indirect, cross-trial extrapolation that methodologists appropriately cautioned against.

The SURMOUNT-5 trial changed that. This head-to-head RCT enrolled adults with obesity (BMI ≥30 or ≥27 with a weight-related comorbidity) and randomized them to tirzepatide (10 mg or 15 mg) or semaglutide 2.4 mg — the doses that reflect actual clinical use — over 72 weeks.

Results published in late 2025:

Tirzepatide 15 mg: mean weight loss of approximately 20.2% of body weight
Tirzepatide 10 mg: mean weight loss of approximately 17.8%
Semaglutide 2.4 mg: mean weight loss of approximately 13.7%

The difference was statistically significant and clinically meaningful. Responder rates — defined as achieving at least 15% weight loss — were significantly higher with tirzepatide. The proportion achieving 20% or more weight loss was more than twice as high with tirzepatide 15 mg compared to semaglutide.

Tolerability: Gastrointestinal side effects (nausea, vomiting, diarrhea) were similar between the groups, though semaglutide had slightly higher early nausea rates during dose escalation. The counterintuitive direction of this difference — the dual agonist producing less early nausea despite its greater appetite-suppressant and weight-loss efficacy — has a mechanistic explanation rooted in GIP receptor pharmacology at the brainstem. GLP-1R-driven nausea and emesis are mediated primarily through the area postrema (AP) and nucleus tractus solitarius (NTS), where GLP-1R/Gs/cAMP/PKA signaling activates the chemoreceptor trigger zone and drives vagal efferent output to the gastrointestinal tract. The GIP receptor (GIPR) is also expressed in the AP, but its Gs/cAMP activation in this context operates on a neuron population that is functionally distinct from and inhibitory to the GLP-1R nausea circuit. Rodent and non-human primate studies have demonstrated that direct GIPR agonism in the AP/NTS reduces GLP-1R-driven emesis — coadministered GIPR agonist blunts the nausea response to GLP-1R agonist alone, and this interaction is not simply a dose-dilution effect but a genuine circuit-level inhibitory modulation. The proposed mechanism: GIPR-expressing AP interneurons, when activated by the GIP component of tirzepatide, release inhibitory neurotransmitters onto the GLP-1R-expressing AP principal neurons driving nausea output — GIP-mediated cAMP in the interneuron activates K_ATP channels and CREB-driven inhibitory neuropeptide expression that tonically suppress the GLP-1R effector neurons’ ability to drive the full chemoreceptor trigger zone cascade. This GIPR/GLP-1R interaction at the AP is mechanistically distinct from the additive appetite suppression the two receptors generate in the arcuate nucleus (where GIP suppresses NPY/AgRP while GLP-1 activates POMC — two non-overlapping orexigenic and anorexigenic nodes). The result is a pharmacological profile where tirzepatide simultaneously achieves greater hypothalamic appetite suppression and attenuated brainstem nausea drive relative to a pure GLP-1R agonist — explaining the SURMOUNT-5 tolerability finding that greater efficacy did not come with worse GI side effects.

What This Means in Practice

Tirzepatide’s superior weight loss is mechanistically expected: it is a dual agonist (GLP-1 and GIP receptors), while semaglutide is a single agonist. The GIP receptor adds an incremental effect on appetite suppression, energy expenditure, and fat mobilization — each through distinct molecular substrates that do not overlap with GLP-1R signaling.

Appetite suppression — central GIP receptor signaling: GIP receptors are expressed on neurons in the arcuate nucleus of the hypothalamus, including a population of NPY/AgRP neurons that are the primary drivers of hunger signaling. GIPR is a Gs-coupled receptor; its activation raises intracellular cAMP, which activates PKA. In NPY/AgRP neurons, cAMP/PKA activation reduces NPY and AgRP neuropeptide release through a pre-synaptic modulation mechanism — the elevated cAMP directly reduces vesicular release probability at the axon terminal, blunting the orexigenic neuropeptide output from the arcuate nucleus to downstream targets including the paraventricular nucleus and lateral hypothalamus. This NPY/AgRP suppression is additive to GLP-1R-driven POMC neuron activation because the two mechanisms target opposite arms of the arcuate hunger circuit simultaneously: GLP-1R increases the anorexigenic signal (POMC → α-MSH → MC4R), while GIPR suppresses the orexigenic counter-signal (NPY/AgRP → NPY receptors and inverse MC4R agonism). When both arms are engaged concurrently by tirzepatide, the net anorexigenic output from the hypothalamus is substantially greater than either receptor achieves alone.

Energy expenditure — GIP receptor in brown adipose tissue: GIP receptors are expressed on brown adipocytes and on beige adipocyte precursors in subcutaneous fat. GIPR/Gs/cAMP/PKA in brown adipocytes phosphorylates CREB at Ser133 → CREB/CBP drives transcription of PGC-1α (peroxisome proliferator-activated receptor gamma coactivator 1-alpha) and UCP1 (uncoupling protein 1). UCP1 is the molecular effector of non-shivering thermogenesis: it dissipates the mitochondrial proton gradient generated by fatty acid oxidation by allowing proton flux across the inner mitochondrial membrane without coupling to ATP synthase, converting oxidative energy directly to heat rather than ATP. GIPR-driven UCP1 upregulation — layered on top of the BAT thermogenesis activated by the β3-adrenergic system — increases the fraction of metabolized substrate that generates heat rather than ATP, raising basal energy expenditure beyond what caloric restriction alone would produce. GLP-1 receptors are expressed at low density in adipose tissue; GIP receptors are expressed at much higher density in both white and brown adipocytes, making adipose thermogenesis a GIP-dominant rather than GLP-1-dominant mechanism.

Fat mobilization — GIPR in white adipose tissue: In white adipocytes, GIPR/Gs/cAMP/PKA activates hormone-sensitive lipase (HSL) via PKA phosphorylation of HSL at Ser563 and Ser660, increasing its catalytic rate for triacylglycerol hydrolysis → fatty acid and glycerol release from lipid droplets into the circulation. Concurrently, cAMP/PKA phosphorylates perilipin-1 (the lipid droplet surface coat protein that normally limits HSL access), causing perilipin-1 to be displaced from the droplet surface and clearing the way for HSL and adipose triglyceride lipase (ATGL) to access triacylglycerol substrate. The released fatty acids are available for β-oxidation in skeletal muscle, liver, and activated brown adipose tissue — the coordinated mobilization-to-oxidation partnership. GLP-1R on adipocytes is expressed at substantially lower levels than GIPR, making the GIP pathway the predominant receptor-mediated fat mobilization signal of the two components in tirzepatide.

For patients who have tried semaglutide and achieved partial but inadequate response, this data supports a rationale for trialing tirzepatide. For new patients, the data supports tirzepatide as the more effective agent for weight loss specifically — though insurance coverage, cost, availability, and individual tolerability remain practical considerations.

One important caveat: SURMOUNT-5 used semaglutide 2.4 mg, the maximum approved dose for weight management. For cardiovascular risk reduction specifically, semaglutide has more mature long-term outcomes data (see SELECT below). The head-to-head comparison for cardiovascular outcomes awaits SURPASS-CVOT.

Long-Term Cardiovascular Benefit: SELECT 5-Year Extension

The SELECT trial was the pivotal cardiovascular outcomes trial for semaglutide 2.4 mg in adults with overweight or obesity and established cardiovascular disease (without diabetes). The primary analysis, published in the New England Journal of Medicine in 2023 (PMID: 37952131), showed a 20% reduction in major adverse cardiovascular events (MACE: cardiovascular death, non-fatal heart attack, non-fatal stroke) with semaglutide over a mean follow-up of approximately 3 years.

The 5-year extension data, presented at AHA 2025, addressed a key question: does this benefit hold over longer follow-up, or attenuate as the weight loss plateau effect diminishes?

What the extension showed:

Cardiovascular benefit is durable: The approximately 20% MACE reduction was maintained through 5-year follow-up without meaningful attenuation
Benefit appears partially weight-independent: Prespecified analyses examining cardiovascular outcomes relative to weight loss magnitude found that a portion of the cardiovascular benefit occurred in patients with modest weight loss — suggesting direct vascular effects beyond weight reduction. Three parallel intracellular pathways operating in the vascular wall independent of adiposity provide the mechanistic basis. First, GLP-1 receptors expressed on vascular endothelial cells couple to Gs → adenylyl cyclase → cAMP → PKA, which phosphorylates eNOS at Ser1177, activating the enzyme to generate nitric oxide (NO); NO diffuses into adjacent vascular smooth muscle → soluble guanylyl cyclase → cGMP → PKG → myosin light-chain phosphatase activation → vascular relaxation and vasodilation. Second, GLP-1Rs on macrophages and foam cells within atherosclerotic plaques engage the same Gs/cAMP/PKA axis → PKA phosphorylates and inactivates IKKβ → IκBα is not phosphorylated or degraded → NF-κB p65/p50 heterodimer remains sequestered in the cytoplasm → reduced transcription of scavenger receptors SR-A (MSR1) and CD36 → less oxidized LDL endocytosis → attenuated foam cell formation and plaque lipid accumulation → plaque stabilization. Third, GLP-1Rs on platelets → Gs/cAMP → PKA phosphorylates thromboxane A₂ TP receptors → reduced TP receptor sensitivity → diminished platelet aggregation in response to TXA₂ → anti-thrombotic effect. These three vascular-wall mechanisms do not require weight loss to operate and provide a coherent molecular explanation for the persistent cardiovascular benefit observed across the full range of weight-loss response in SELECT
All-cause mortality signals: A trend toward reduced all-cause mortality emerged but did not reach statistical significance in this population analysis

Clinical significance: For patients with established cardiovascular disease or high cardiovascular risk, the SELECT 5-year data strengthens the case for semaglutide as a cardiovascular intervention, not just a weight management tool. The FDA has approved Wegovy for this specific indication: cardiovascular risk reduction in patients with established CVD and BMI ≥27.

Tirzepatide cardiovascular data pending: The SURPASS-CVOT trial is evaluating tirzepatide’s cardiovascular outcomes; interim data presented at AHA 2025 showed directionally positive signals, with final results expected in 2027.

Kidney Protection: The FLOW Trial

One of the most consequential GLP-1 trial results of 2024 received less consumer-facing attention than the weight loss data but may ultimately affect a larger number of patients: the FLOW trial demonstrated that semaglutide protects the kidneys.

The FLOW trial (NCT03819153) enrolled over 3,500 patients with type 2 diabetes and chronic kidney disease (CKD; eGFR 50–75 mL/min/1.73m², with albuminuria). These are patients at high risk of kidney failure — the leading cause of dialysis initiation globally. Participants were randomized to semaglutide 1 mg weekly or placebo on top of standard CKD care.

Primary outcome: Composite kidney failure endpoint (kidney failure requiring dialysis/transplant, sustained ≥50% eGFR decline from baseline, or kidney or cardiovascular death).

Results (published NEJM 2024, PMID: 38587991):

24% reduction in the primary kidney composite outcome with semaglutide
The trial was stopped early for efficacy — a strong signal that the benefit was clear enough that continuing placebo was no longer ethical
Kidney-protective effects appeared independent of the degree of glucose lowering, suggesting a direct kidney mechanism (anti-inflammatory, hemodynamic, or tubular effects) beyond metabolic benefit. The proximal tubule mechanism is the most characterized: GLP-1 receptors are expressed on proximal tubule cells, where Gs/cAMP/PKA signaling reduces the activity of sodium-hydrogen exchanger 3 (NHE3) — the major apical transporter responsible for sodium reabsorption in the proximal tubule. Reduced NHE3 activity lowers tubular sodium reabsorption, increasing distal sodium delivery to the macula densa; via the tubuloglomerular feedback mechanism, the macula densa responds by triggering afferent arteriolar constriction, reducing intraglomerular hydrostatic pressure and thereby reducing glomerular hyperfiltration — the sustained mechanical stress that damages the glomerular filtration barrier in diabetic nephropathy. GLP-1R/Gs/cAMP/PKA in proximal tubular epithelial cells adds two mechanistically independent arms beyond NHE3-mediated hemodynamic protection. Tubular NF-κB suppression: PKA phosphorylates IKKβ at regulatory serine residues in its activation-loop region, reducing IKKβ’s own kinase activity toward its substrate IκBα → IκBα Ser32/Ser36 remains unphosphorylated → IκBα is not polyubiquitinated or proteasomally degraded → NF-κB p65/p50 remains cytoplasmic → transcription of IL-6, MCP-1 (CCL2), and TGF-β1 in tubular epithelium falls. Reduced tubular TGF-β1 is mechanistically significant beyond its pro-inflammatory role — TGF-β1 secreted from tubular cells signals in paracrine fashion to peritubular fibroblasts via TGFβRII/ALK5/SMAD2/3/SMAD4 → COL1A1 and COL3A1 transcription → interstitial collagen deposition → tubulointerstitial fibrosis, the histological hallmark of progressive CKD. Suppressing this NF-κB → TGF-β1 → SMAD2/3 fibrotic axis at the tubular source provides structural preservation independent of the glomerular hemodynamic protection driven by NHE3. Tubular Nrf2/HO-1 induction: cAMP/PKA → Nrf2 phosphorylation at Ser40 in its Neh2 regulatory domain → reduced Nrf2 binding affinity for the Keap1–Cullin3–Rbx1 E3 ubiquitin ligase complex that normally retains Nrf2 in the cytoplasm → Nrf2 nuclear accumulation → Nrf2/sMaf heterodimer binds antioxidant response elements (ARE) in promoters of HO-1 (heme oxygenase-1), NQO1 (NAD(P)H quinone oxidoreductase-1), and GCLM (rate-limiting subunit of glutamate-cysteine ligase, controlling glutathione synthesis) → HO-1 converts filtered heme from albumin-bound hemoglobin to biliverdin/CO/Fe²⁺ (sequestered by ferritin); elevated glutathione via GCLM scavenges ROS generated by the albumin endocytosis burden on hyperfiltration-stressed brush border cells. This multi-pathway direct renal mechanism — hemodynamic (NHE3-driven intraglomerular pressure reduction) + anti-fibrotic (PKA/IKKβ/NF-κB/TGF-β1/SMAD2/3 suppression) + antioxidant (PKA/Nrf2/HO-1/GCLM induction) — explains why the FLOW benefit occurred independently of glycemic control magnitude

FDA action: The FDA has approved semaglutide (Ozempic) for the additional indication of reducing the risk of kidney disease worsening and kidney failure in patients with type 2 diabetes and CKD with albuminuria.

Why this matters for patients: The FLOW trial positions semaglutide alongside established CKD treatments (ACE inhibitors, SGLT2 inhibitors like empagliflozin) as a kidney-protective agent with its own primary trial evidence. For patients with both obesity and CKD, this strengthens the case for semaglutide specifically — where the kidney-protective data is now established — over tirzepatide (for which equivalent CKD data does not yet exist).

Heart Failure: STEP-HFpEF

Heart failure with preserved ejection fraction (HFpEF) — a form of heart failure where the heart pumps normally but stiffens abnormally — has historically had very limited treatment options. Obesity is a major contributor to HFpEF pathophysiology, and the STEP-HFpEF trial tested whether semaglutide could help.

The STEP-HFpEF trial (NCT04788511) enrolled patients with HFpEF (ejection fraction ≥45%), BMI ≥30, and no type 2 diabetes. Co-primary outcomes were symptom burden (KCCQ-CSS score, a validated heart failure quality-of-life measure) and body weight.

Results (published NEJM 2023, PMID: 37350574):

13.3% body weight loss with semaglutide vs. 2.6% placebo
KCCQ-CSS improvement: 7.8 points with semaglutide vs. 4.3 points placebo — a clinically meaningful improvement in heart failure symptoms
6-minute walk test: Improved significantly with semaglutide
CRP (inflammation marker): Reduced substantially, suggesting anti-inflammatory effects independent of weight loss
Exercise capacity and quality of life both improved

FDA action: The FDA has approved Wegovy for the treatment of heart failure with preserved ejection fraction in adults with obesity.

Clinical significance: HFpEF has few disease-modifying therapies. Semaglutide’s benefits in STEP-HFpEF — particularly the inflammation reduction and functional improvement — suggest effects beyond simple weight loss in cardiac physiology. For patients with both obesity and HFpEF, semaglutide now has primary trial evidence supporting use as a heart failure treatment, not just a weight loss adjunct.

The Cardiac Mechanisms: How GLP-1 Drugs Help HFpEF

HFpEF is mechanistically complex — a syndrome where the heart pumps normally but relaxes abnormally (diastolic dysfunction), with obesity-specific drivers including elevated cardiac preload from fluid retention, ventricular hypertrophy from sustained pressure-volume stress, and pro-inflammatory signaling from surrounding and infiltrating adipose tissue. Understanding the multiple cardiac mechanisms explains why the STEP-HFpEF benefit exceeded what weight loss alone would be expected to produce.

Direct cardiomyocyte effects via GLP-1 receptors: GLP-1 receptors are expressed on ventricular cardiomyocytes. Activation → cAMP → PKA → phosphorylation of calcium handling proteins, the most important of which is phospholamban (PLN). PLN is a 52-amino acid transmembrane protein that resides in the sarcoplasmic reticulum (SR) membrane and, in its unphosphorylated state, physically associates with SERCA2a (sarco/endoplasmic reticulum Ca²⁺-ATPase 2a) — the pump responsible for actively transporting Ca²⁺ from the cardiomyocyte cytoplasm back into the SR at the end of each systolic contraction. Unphosphorylated PLN binds SERCA2a’s transmembrane domain and reduces SERCA2a’s apparent Ca²⁺ affinity (increases the K0.5 for Ca²⁺), slowing Ca²⁺ reuptake. When GLP-1R/Gs/cAMP/PKA is activated, PKA phosphorylates PLN at Ser16 → the PLN-SERCA2a binding interface undergoes a conformational change → PLN dissociates from SERCA2a’s transmembrane domain → SERCA2a operates at its full intrinsic Ca²⁺ affinity → faster Ca²⁺ reuptake from the cytoplasm into the SR during diastole → faster, more complete cytoplasmic Ca²⁺ clearance → cardiomyocyte relaxes more quickly and completely (positive lusitropic effect). CaMKII (calcium/calmodulin-dependent kinase II) also phosphorylates PLN at Thr17 through a parallel Ca²⁺-dependent pathway, providing an additional route to SERCA2a disinhibition during high heart rate states. In HFpEF hearts, PLN-SERCA2a coupling is dysregulated — reduced phosphorylation of PLN (due to elevated phosphatase PP1/PP2a activity in hypertrophied myocardium) keeps more PLN in the inhibitory, SERCA2a-bound state, impairing diastolic Ca²⁺ reuptake and prolonging the relaxation phase. GLP-1R-driven PKA/PLN phosphorylation partially reverses this dysregulation, directly addressing the underlying Ca²⁺ handling deficit that produces diastolic dysfunction. In animal models of HFpEF, GLP-1R activation reduces myocardial collagen deposition (cardiac fibrosis) and decelerates the development of ventricular stiffness through pathways downstream of cAMP signaling.

Epicardial adipose tissue (EAT) reduction: Epicardial fat — the visceral adipose tissue immediately surrounding and infiltrating the myocardium — is disproportionately elevated in obesity-related HFpEF. EAT secretes pro-inflammatory cytokines (TNF-α, IL-6, IL-1β) that penetrate the underlying myocardium paracrinally. The molecular pathways through which these cytokines drive fibrosis, diastolic dysfunction, and atrial remodeling are distinct and mechanistically important. Fibrosis: TNF-α from EAT binds TNFR1 on cardiac fibroblasts → TRAF2 recruitment → IKKβ → IκBα degradation → NF-κB p65/p50 nuclear translocation → NF-κB drives transcription of TGF-β1 in fibroblasts; TGF-β1 then signals in autocrine/paracrine fashion: TGF-β1 → TGFβRII → transphosphorylates ALK5 (TGFβRI) → ALK5 phosphorylates SMAD2/3 → SMAD2/3/SMAD4 nuclear complex → COL1A1 and COL3A1 promoter activation → collagen I/III deposition in the interstitium → reactive interstitial fibrosis → increased myocardial passive stiffness → elevated left ventricular end-diastolic pressure (the hemodynamic hallmark of diastolic dysfunction). Cardiomyocyte Ca²⁺ handling impairment: EAT-derived IL-6 binds IL-6Rα/gp130 on cardiomyocytes → JAK1/STAT3 phosphorylation → STAT3-driven transcription of CAMK2D (the gene encoding CaMKII delta — the dominant CaMKII isoform in ventricular and atrial cardiomyocytes): phospho-STAT3 (pTyr705) homodimerizes and binds GAS elements (5′-TTCNnNGAA-3′) in the CAMK2D gene promoter → CAMK2D mRNA and CaMKII delta protein accumulate in cardiomyocytes; elevated CaMKII delta, sustained by IL-6-driven chronic Ca²⁺ cycling disturbance, achieves persistent Thr286 autophosphorylation → constitutive, Ca²⁺/calmodulin-independent CaMKII (Ca²⁺/calmodulin-dependent protein kinase II) activity; constitutively active CaMKII then hyperphosphorylates the ryanodine receptor 2 (RyR2) at Ser2814 → RyR2 channels open spontaneously between beats → SR Ca²⁺ leak into the cytoplasm between contractions → elevated diastolic Ca²⁺ → slowed SERCA2a-dependent Ca²⁺ reuptake (because SERCA2a is working against a higher cytoplasmic Ca²⁺ load) → prolonged relaxation phase → diastolic dysfunction that compounds the fibrosis-driven stiffness. Atrial remodeling: Epicardial adipocyte infiltration is anatomically concentrated in the interatrial groove and the right ventricular free wall, where adipocytes physically replace coupled cardiomyocytes with electrically resistive adipose and fibrotic tissue; simultaneously, IL-6/STAT3 in atrial myocytes upregulates miR-21 (hsa-miR-21-5p), whose mature sequence (UAGCUUAUCAGACUGAUGUUGA) contains a seed sequence (nucleotides 2–8: AGCUUAU) that base-pairs with complementary target sites at positions 479–485 and 1,108–1,115 in the GJA1 3’UTR (the gene encoding Cx43). The Ago2-RISC complex loaded with miR-21 binds these GJA1 3’UTR sites → mRNA deadenylation via CCR4-NOT complex (progressive poly-A tail shortening → decapping → 5’→3’ exonuclease degradation) with secondary translational repression (blocked 60S subunit joining at the 80S assembly step) → reduced Cx43 protein at atrial intercalated discs. Cx43 is the predominant atrial gap junction protein, forming hexameric connexon hemichannels that dock head-to-head between adjacent atrial cardiomyocytes to establish low-resistance electrical coupling (~100–120 pS conductance) — the physical pathway that allows the action potential upstroke (Nav1.5 Na⁺ influx) to propagate rapidly and uniformly across the atrial wall. When Cx43 density falls, intercellular coupling resistance increases heterogeneously across the atrial myocardium: regions retaining adequate Cx43 propagate rapidly while Cx43-depleted zones propagate slowly. This conduction velocity dispersion is the electrophysiological substrate for micro-reentrant circuits: a wavefront entering a slow-conducting zone curves around the zone’s exit boundary and re-excites the adjacent fast-conducting region from behind before its refractory period expires — creating a self-perpetuating re-entrant loop → atrial fibrillation that is disproportionately elevated in obesity-related HFpEF. GLP-1 and GLP-1/GIP agonists preferentially reduce visceral adiposity; multiple cardiac imaging studies confirm significant EAT volume reduction with semaglutide and tirzepatide. The marked CRP reduction in STEP-HFpEF — substantially greater than placebo — is consistent with this TNF-α/TGF-β/IL-6 signaling cascade being interrupted at the source by EAT mass reduction rather than through direct receptor antagonism.

Natriuretic and hemodynamic effects: GLP-1R activation has natriuretic properties independent of weight loss, increasing renal sodium and water excretion through two parallel mechanisms. The first operates via atrial GLP-1Rs: GLP-1Rs on atrial cardiomyocytes → Gs/cAMP → PKA → PKA phosphorylates pre-formed ANP (atrial natriuretic peptide) secretory granules, enhancing their exocytosis → ANP is released into the circulation → ANP binds NPR-A (natriuretic peptide receptor A, a particulate guanylyl cyclase) on renal collecting duct principal cells → cGMP elevation → PKG → (1) phosphorylation and inactivation of ENaC (epithelial sodium channel) at the apical membrane → reduced Na⁺ reabsorption from tubular lumen → reduced water co-reabsorption via aquaporin-2 → natriuresis and diuresis; (2) NPR-A activation also suppresses aldosterone secretion from adrenal zona glomerulosa cells (ANP → cGMP → PKG → inhibits StAR-mediated cholesterol import into adrenal mitochondria → reduced aldosterone synthesis → further reduction in collecting duct Na⁺ reabsorption). The second natriuretic mechanism operates directly in the renal proximal tubule: GLP-1Rs are expressed on brush-border proximal tubule cells, where Gs/cAMP/PKA → PKA phosphorylates and reduces the activity of NHE3 (sodium-hydrogen exchanger 3) — the primary apical Na⁺ reabsorber in the proximal tubule — delivering more sodium to the distal nephron for excretion. This direct tubular NHE3 effect is the same mechanism underlying GLP-1’s kidney-protective action in the FLOW trial (reducing intraglomerular hyperfiltration via the tubuloglomerular feedback loop); in the HFpEF context, its consequence is reduced extracellular fluid volume and therefore reduced cardiac preload. In HFpEF patients, both arms — ANP-driven collecting duct natriuresis and proximal tubule NHE3 inhibition — combine to reduce total body sodium and water burden, lowering left ventricular end-diastolic filling pressure and reducing pulmonary venous hypertension. The subjective improvement in dyspnea with STEP-HFpEF, and particularly the 6-minute walk test improvement, likely partly reflect improved pulmonary decongestion through this mechanism rather than only improved cardiac contractile function.

Indirect weight loss effects: A 13% reduction in body weight substantially decreases systemic vascular resistance, venous return, and the mechanical work imposed on the left ventricle — improving diastolic filling pressures through reduced preload and afterload. This is likely the largest single contributor to symptom improvement. The mechanism operates through three parallel axes, each distinct from the direct GLP-1R cardiac and natriuretic pathways described above. First, adipocyte-derived angiotensinogen (AGT) reduction: adipocytes constitutively express and secrete AGT — the precursor substrate for the renin-angiotensin-aldosterone system (RAAS). Adipocyte-derived AGT is cleaved by renin → angiotensin I → ACE (angiotensin-converting enzyme) → angiotensin II (AngII) → AngII binds AT1R on vascular smooth muscle cells → Gq/PLCβ → IP3-driven SR Ca²⁺ release + DAG/PKC-ε → MLCK activation → vasoconstriction; AngII simultaneously stimulates adrenal zona glomerulosa → StAR/CYP11B2 → aldosterone → mineralocorticoid receptor → distal tubule ENaC + AQP2 → Na⁺ and water retention → elevated plasma volume → increased cardiac preload. As fat mass decreases with weight loss, adipocyte AGT secretion falls in proportion to adipocyte lipid content → less tissue AngII generation → reduced AT1R-driven smooth muscle contraction → SVR falls. Second, leptin-mediated sympathetic deactivation: adipose tissue secretes leptin in direct proportion to adipocyte lipid content; elevated leptin → hypothalamic melanocortin pathway (arcuate POMC → α-MSH → MC4R on pre-sympathetic neurons in the PVN and rostral ventrolateral medulla/RVLM) → sympathetic outflow to peripheral vasculature → norepinephrine release onto α1-adrenergic receptors on vascular smooth muscle → Gq/IP3/Ca²⁺ → vasoconstriction and elevated heart rate. As visceral fat mass and plasma leptin fall, the melanocortin-sympathetic tone reduces → less α1-adrenergic vasoconstriction → reduced SVR and cardiac chronotropic drive independently of direct GLP-1R receptor pharmacology. Third, mechanical venous decompression: visceral and intrathoracic adipose tissue exerts direct compressive load on the inferior vena cava and hepatic venous outflow; as abdominal and thoracic fat mass reduces, this mechanical impediment to venous return diminishes → right atrial filling pressure falls → reduced preload independently of neurohormonal effects. The convergence of all three mechanisms — reduced RAAS-driven vasoconstriction and fluid retention, reduced leptin-sympathetic peripheral resistance, and mechanical venous decompression — explains why 13% body weight reduction produces a hemodynamic improvement in HFpEF that substantially exceeds what direct GLP-1R receptor effects on the heart would achieve in isolation.

The combination of direct cardiac mechanisms (cardiomyocyte calcium handling, fibrosis reduction), EAT inflammation reduction, natriuresis, and weight-related hemodynamic improvement explains why the STEP-HFpEF CRP and functional changes exceeded what simple caloric restriction studies typically show. Ongoing trials are combining semaglutide with SGLT2 inhibitors (empagliflozin, dapagliflozin) to test whether dual therapy provides additive cardiac benefit. The SGLT2 inhibitor contribution operates through at least three mechanisms non-overlapping with GLP-1R biology. First, SGLT2 (SLC5A2) is the primary sodium-glucose cotransporter in the renal proximal tubule S1/S2 segment; its inhibition reduces tubular glucose and sodium co-reabsorption → osmotic diuresis → reduced plasma volume and extracellular fluid burden → reduced cardiac preload, complementing (but mechanistically parallel to) the GLP-1-driven NHE3/ANP natriuresis already described. Second, SGLT2 inhibitors suppress NLRP3 inflammasome activation in cardiac macrophages and cardiomyocytes — independent of glycemia and of GLP-1R signaling — by lowering intracellular Na⁺ and ketone body availability (β-hydroxybutyrate is an endogenous NLRP3 inhibitor elevated by SGLT2 inhibition): NLRP3 → caspase-1 → IL-1β/IL-18 maturation suppression reduces myocardial inflammation and fibrosis through a distinct pathway from the EAT TNF-α/TGF-β/IL-6 axis that semaglutide addresses. Third, cardiomyocyte NHE1 (sodium-hydrogen exchanger 1) — distinct from renal NHE3 — is directly inhibited by SGLT2 inhibitors in a glucose- and transporter-independent manner; reduced NHE1 activity → lower intracellular Na⁺ → the NCX (Na⁺/Ca²⁺ exchanger, NCX1/SLC8A1) operates in forward mode (Na⁺ in, Ca²⁺ out) rather than reverse mode → reduced diastolic cytoplasmic Ca²⁺ accumulation → faster cardiomyocyte relaxation and improved lusitropy. This NHE1/NCX1 diastolic improvement is mechanistically additive to the GLP-1-driven PLN/SERCA2a Ca²⁺ re-uptake enhancement, because PLN acts on SERCA2a’s reticulum re-uptake rate while NHE1/NCX1 inhibition reduces the cytoplasmic Ca²⁺ load that SERCA2a must clear — two independent Ca²⁺ management strategies that could combine to produce diastolic relaxation improvements exceeding either agent alone.

The Weight Regain Problem: SURMOUNT-4

This is perhaps the most clinically consequential finding from the recent data cycle — and the one that most directly affects how patients and prescribers should frame expectations.

The SURMOUNT-4 trial enrolled adults who had completed a 36-week lead-in of tirzepatide (achieving substantial weight loss) and then randomized them to continue tirzepatide or switch to placebo. Follow-up extended to 88 weeks post-randomization.

What happened off medication:

Participants who switched to placebo regained approximately 14 percentage points of body weight over 88 weeks
By the end of follow-up, much of the metabolic benefit (improved blood pressure, glucose, lipids) had also partially reversed
Participants who continued tirzepatide maintained their weight loss and metabolic improvements

The absolute weight regain numbers are striking: a participant who had lost 25% of body weight on tirzepatide might regain half of that — roughly 12–13% of their original weight — within two years of stopping.

Why This Matters — and Why Regain Is Biologically Inevitable Without Medication

This pattern — robust weight loss while on medication, progressive regain off it — mirrors what has been observed with other anti-obesity medications historically. The mechanistic reason for this predictable reversal is important for patients to understand: GLP-1 drugs correct the symptoms of obesity’s neurobiological dysregulation, but they do not correct the underlying state.

The leptin resistance problem re-exposes itself. As detailed in the gut-brain pathways guide, chronic obesity creates persistent leptin resistance — SOCS3-mediated depletion of JAK2 kinase in hypothalamic POMC neurons and saturable leptin transport impairment at the blood-brain barrier. GLP-1 agonists bypass this resistance by directly activating arcuate POMC neurons through GLP-1R/Gs/cAMP signaling that operates independently of the leptin receptor. When the drug is removed, the bypass is removed: POMC neurons revert to receiving the chronically attenuated leptin signal → AgRP/NPY neurons are relatively disinhibited → hypothalamic appetite-promoting NPY (Neuropeptide Y) release at the paraventricular nucleus → NPY1R/NPY5R (Gi-coupled) → reduced cAMP → attenuated anorexigenic output → hyperphagia relative to caloric need. The patient’s brain is, in effect, perceiving near-starvation despite adequate fat stores.

Adaptive thermogenesis reduction is not reversed. During the weight-loss phase on GLP-1 therapy, the body reduces resting metabolic rate by approximately 10–20% per 10% of body weight lost — partly from lean mass reduction, partly from neuroendocrine adaptation (reduced free T3 and sympathetic drive to brown adipose tissue, mediated by reduced hypothalamic TRH/TSH output under caloric restriction). Critically, this metabolic rate suppression persists after weight loss is complete and is not immediately restored by caloric reintroduction. The persistence is mechanistically rooted in adipocyte biology rather than CNS habituation to the drug. During the weight-loss phase, adipocytes deplete their lipid content as triglycerides are mobilized for energy. Leptin secretion is proportional to adipocyte lipid content per cell — specifically, intracellular lipid volume drives leptin mRNA transcription via C/EBPα and PPARγ acting on the leptin gene promoter; a shrunken, lipid-depleted adipocyte secretes substantially less leptin than its pre-loss hypertrophic counterpart, independent of total fat mass. The hypothalamus therefore continues to receive a low-leptin signal that it interprets as near-starvation even though the body is now at a lower but adequate fat mass: persistently low circulating leptin means attenuated JAK2/STAT3 signaling on TRH neurons in the hypothalamic paraventricular nucleus (PVN), which normally require leptin-driven JAK2/STAT3 to maintain TRH transcription and secretion → reduced TRH → suppressed pituitary TSH → lower free T4/T3 → reduced T3/thyroid hormone receptor (TR) transcriptional drive on the UCP1 gene promoter in brown adipocyte nuclei → blunted thermogenic capacity in BAT. Concurrently, low leptin reduces sympathetic outflow to BAT via the ventromedial hypothalamus (VMH) → intermediolateral cell column (IML) → preganglionic sympathetic → norepinephrine → β3-adrenergic receptor/PKA/HSL axis that normally amplifies lipolysis and UCP1 activation within existing brown adipocytes. Both arms of thermogenesis — transcriptional (UCP1 gene expression) and acute (β3-AR/PKA/HSL) — remain suppressed as long as adipocyte leptin output is low. This suppression normalizes only as the patient re-gains adipose lipid content and adipocyte leptin secretion rebuilds — explaining the asymmetry between rapid pharmacological weight loss (GLP-1R bypasses leptin resistance at arcuate POMC neurons to achieve caloric deficit) and slow post-cessation metabolic recovery (adaptive thermogenesis tracks adipocyte lipid content, not the pharmacological state), and why weight regain accelerates before metabolic rate is fully restored. At the lower body weight post-treatment, the patient has both a higher appetite drive (from leptin resistance re-exposure) and a lower resting energy expenditure — creating a strong positive energy balance that drives weight regain toward the pre-treatment adiposity level.

Adipocyte hypercellularity creates fat-storage efficiency. A history of obesity is associated with elevated adipocyte cell number (hyperplasia in addition to hypertrophy). Weight loss reduces adipocyte size but cannot eliminate adipocytes (lipolysis depletes lipid but not the cell itself). Smaller, lipid-depleted adipocytes have elevated lipoprotein lipase (LPL) surface expression and enhanced GLUT4 trafficking, making them highly efficient at recapturing circulating fatty acids and glucose from any caloric surplus — accelerating fat mass reconstitution once dietary intake returns to pre-treatment levels. The molecular basis of this enhanced uptake efficiency operates through two convergent pathways. For glucose: hypertrophic adipocytes accumulate ceramide and diacylglycerol (DAG) — lipid second messengers that activate PKC-θ and IKKβ to phosphorylate IRS-1 at Ser307, uncoupling IRS-1 from PI3K and generating insulin resistance. Weight-loss-induced lipid depletion reverses this: reduced ceramide/DAG → less IRS-1 Ser307 phosphorylation → restored PI3K/PIP3 generation → PDK1 phosphorylates Akt at Thr308 → Akt phosphorylates AS160 (TBC1D4, Akt substrate of 160 kDa) at Thr642. AS160 is a Rab-GAP (GTPase-activating protein) that in its unphosphorylated state drives GTP hydrolysis on Rab8A and Rab10, keeping them GDP-bound and inactive → GLUT4 storage vesicles (GSVs) are retained in their intracellular perinuclear depot. Akt-mediated AS160 phosphorylation blocks its Rab-GAP activity → Rab10 accumulates as Rab10-GTP → Rab10-GTP engages GSV motor and tethering complexes → SNARE-dependent fusion at the plasma membrane (VAMP2/syntaxin-4/SNAP-23) → GLUT4 surface density increases two- to fourfold above basal → markedly enhanced insulin-stimulated glucose influx. For fatty acids: in lipid-replete adipocytes, ANGPTL4 (angiopoietin-like protein 4) is secreted constitutively and inhibits LPL by disrupting the LPL homodimer at Asn45 and Asn117 glycosylation sites, reducing its catalytic efficiency on capillary endothelium. Lipid-depleted adipocytes, now highly insulin-sensitive via the restored Akt pathway, suppress ANGPTL4 secretion: Akt → FoxO1 nuclear exclusion → reduced FoxO1-driven ANGPTL4 transcription → capillary LPL is disinhibited → active LPL hydrolyzes circulating triglycerides from VLDL and chylomicrons → free fatty acids and 2-monoacylglycerols are transported across the endothelium → CD36/FATP1 transporters on the adipocyte surface deliver them to cytoplasmic FABP4 → fatty acyl-CoA synthetases (ACSLs) generate fatty acyl-CoA → DGAT1/DGAT2 re-esterify to triacylglycerol → lipid droplet reconstitution. The consequence is that post-weight-loss adipocytes have simultaneously restored GLUT4 trafficking (enhanced glucose capture) and reduced ANGPTL4-mediated LPL inhibition (enhanced fatty acid capture) — creating a metabolic state more efficient at refilling fat stores than the original hypertrophic state, explaining the clinical observation of disproportionately rapid fat mass reconstitution relative to calorie surplus during weight regain.

The cumulative effect of these three mechanisms — reinstated leptin resistance driving hyperphagia, persistent metabolic rate suppression, and hypercellular adipocytes ready to refill — produces near-complete weight regain without continued pharmacotherapy. This is not a patient behavior failure; it is an expected biological response to removing pharmacological correction from a chronically dysregulated system.

This has two important implications:

1. GLP-1 therapy for weight management is likely a long-term or indefinite commitment for most patients. The framing of these medications as a finite “course of treatment” is not supported by the data. A more accurate analogy is antihypertensive or statin therapy: effective while taken, with meaningful reversal of benefit upon discontinuation.

2. The muscle loss question becomes more urgent. If patients who stop GLP-1 therapy regain weight primarily as fat (rather than rebuilding the lean mass lost during the weight-loss phase), their body composition after regain may be worse than before they started.

Body Composition: The Lean Mass Concern

Weight loss from GLP-1 therapy is not purely fat loss — and the body composition data deserves careful attention.

Body composition analyses from the SURMOUNT trials using DEXA or MRI imaging found that approximately 25–40% of total weight lost on GLP-1 and GLP-1/GIP agonists is lean mass (muscle, bone, and other non-fat tissue). The remaining 60–75% is fat mass.

How does this compare to other weight loss approaches?

Lean mass loss is not unique to GLP-1 drugs — it occurs with caloric restriction by any method, and the underlying mechanism is conserved across all approaches that produce a sustained energy deficit. During caloric restriction, circulating insulin falls as post-meal glucose excursions diminish (less duodenal glucose → less GIP/GLP-1 co-stimulation → reduced β-cell insulin secretion → lower basal and postprandial insulin). Reduced insulin → reduced IRS-1/PI3K/PIP₃/PDK1/Akt signaling in skeletal muscle → less mTORC1 activation (Akt → TSC1/2 phosphorylation → Rheb-GTP → mTORC1 lysosomal activation is attenuated) → reduced S6K1 and 4E-BP1 phosphorylation → decreased cap-dependent translation efficiency for myofibrillar proteins (myosin heavy chain, actin, titin) → lower muscle protein synthesis (MPS) rate. Simultaneously, caloric deficit activates the HPA axis: reduced blood glucose → glucoprivic stress signal → CRH/AVP → ACTH → cortisol elevation. Cortisol binds glucocorticoid receptor (GR), and the GR/GRE complex drives transcription of two master atrogene regulators: FOXO1 (forkhead box O1) and KLF15 (Krüppel-like factor 15). FOXO1 drives expression of MAFbx/atrogin-1 (FBXO32) and MuRF1 (TRIM63) — the two principal E3 ubiquitin ligases of the skeletal muscle ubiquitin-proteasome system (UPS). MAFbx ubiquitinates MyoD and eIF3-f (an eukaryotic initiation factor component), suppressing translational capacity. MuRF1 polyubiquitinates myofibrillar proteins (myosin heavy chains MHC I/II, troponin, myosin binding protein C) → 26S proteasome degradation → net loss of myofibrillar mass. KLF15 amplifies the FOXO1/atrogin-1 arm by also inducing branched-chain amino acid (BCAA) catabolism through BCAT2 (branched-chain aminotransferase 2) → valine, leucine, and isoleucine are transaminated → the carbon skeletons enter the TCA cycle → muscle-derived BCAA provide anaplerotic substrate for hepatic gluconeogenesis during caloric deficit. The consequence is a shift in skeletal muscle net protein balance from positive (synthesis > degradation) to negative (degradation > synthesis), driven by the convergence of reduced anabolic (insulin/Akt/mTORC1) and increased catabolic (cortisol/FOXO1/MAFbx-MuRF1) signaling — not from any drug-specific mechanism. GLP-1-induced weight loss produces the same insulin reduction and cortisol-mediated UPS activation as dietary restriction. This explains why GLP-1 therapy’s ~25–40% lean mass loss rate falls within the expected range for diet-induced weight loss, and why resistance exercise (which generates mechanical mTORC1 activation via integrin/FAK/Akt independently of insulin) and adequate leucine (which reactivates mTORC1 via SESN2/GATOR2/RagA independently of insulin signaling) are the mechanistically correct interventions — they restore the anabolic arms of protein balance without requiring caloric excess. However:

Caloric restriction without exercise: Typically 25–35% of weight lost as lean mass
GLP-1 therapy without exercise: Approximately 25–40% lean mass loss
Bariatric surgery: Similar or higher lean mass loss in some analyses
GLP-1 + resistance exercise: Lean mass loss substantially reduced in preliminary data

The proportion of lean mass loss with GLP-1 drugs appears similar to diet-induced weight loss at the population level. However, the concern is that GLP-1 drugs enable faster, larger weight loss — meaning the absolute lean mass lost is greater even if the proportion is similar.

The regain compounding problem: When patients stop GLP-1 therapy and regain weight (as documented in SURMOUNT-4), weight regain occurs predominantly as fat mass, not as lean mass reconstitution — because the two compartments have fundamentally different reconstitution requirements. Fat reconstitutes passively: as described above, lipid-depleted adipocytes have restored Akt/AS160/Rab10/GLUT4 trafficking and disinhibited LPL via reduced ANGPTL4, making them highly efficient at capturing glucose and fatty acids from any caloric surplus without requiring contractile activity. Lean mass reconstitution, by contrast, requires the simultaneous satisfaction of three independent inputs to mTORC1: the RHEB arm (IGF-1 → Akt → TSC2 → RHEB-GTP), the leucine/Rag arm (Sestrin2 → GATOR2 → RagA/B-GTP → lysosomal mTORC1 recruitment), and critically the mechanical arm — titin stretch during eccentric loading → FAK/ILK autophosphorylation → MEK/ERK → mechano-growth factor (MGF) splice variant of IGF-1 → satellite cell HGF/c-Met/MyoD activation → myonuclear fusion into existing fibers. Passive caloric surplus generates the nutrient and hormonal inputs but does not activate the mechanical arm, which requires eccentric resistance loading; without titin-FAK activation and MGF-driven satellite cell recruitment, net myofibrillar hypertrophy beyond restoring resting protein turnover does not occur. Additionally, in the absence of mechanical mTORC1 activation, FOXO1 remains incompletely phosphorylated and partially nuclear → MuRF1/MAFbx-mediated protein degradation continues at its elevated basal rate — the muscle catabolic floor from the preceding GLP-1 caloric deficit phase does not spontaneously reset without exercise. The result is a metabolic partition where caloric surplus after GLP-1 cessation flows preferentially to hyper-efficient, exercise-independent adipocytes while muscle reconstitution requires active resistance training, adequate leucine per meal, and potentially restored IGF-1 — conditions that rarely co-occur without deliberate intervention. A patient who loses 25 kg (with ~8 kg lean mass loss) and then regains 12 kg (mostly fat) ends up with less muscle and more fat than before treatment — a metabolically worse body composition despite having lost weight on net.

What the Evidence Supports to Minimize Lean Mass Loss

Protein intake: Higher dietary protein during GLP-1-induced weight loss preserves lean mass. The rate-limiting amino acid for muscle protein synthesis is leucine, which activates a specific nutrient-sensing cascade: dietary leucine is sensed intracellularly by Sestrin-2 (SESN2), which in its unbound state physically associates with and inhibits the GATOR2 complex (GAP activity toward Rags 2, a component of the mTORC1 lysosomal activation pathway). Leucine binding to SESN2 induces a conformational change that dissociates SESN2 from GATOR2 → GATOR2 suppresses GATOR1’s GTPase-activating activity toward RagA/B → RagA/RagB GTPases remain GTP-loaded → Rag GTPase heterodimer recruits mTORC1 to the lysosomal surface via the Ragulator scaffold → lysosome-associated mTORC1 is activated by RHEB-GTP → mTORC1 phosphorylates S6K1 Thr389 (→ eIF4B + ribosomal protein S6 activation → translational elongation) and 4E-BP1 (→ release of eIF4E → eIF4F cap-dependent translation initiation) → net increase in muscle protein synthesis (MPS). Current evidence supports targeting 25–30g protein per meal (not just per day) to maximize muscle protein synthesis throughout the day. This threshold reflects leucine’s dose-response: MPS plateaus at approximately 2–3g leucine per meal (achieved by ~25–30g of complete protein), with negligible additional mTORC1 activation from larger single-meal doses — explaining why distributing protein across multiple meals drives substantially more MPS than equivalent total protein consumed in fewer, larger meals. GLP-1-induced appetite reduction can make this difficult — smaller meal volumes require higher protein density.

Resistance exercise: Resistance training during GLP-1-induced weight loss preserves lean mass more than aerobic exercise alone. Studies show 2–3 resistance training sessions per week can meaningfully reduce the proportion of weight lost as lean mass. For patients considering long-term GLP-1 therapy, initiating a resistance exercise habit early — rather than after weight loss is complete — is supported by the biology.

Tirzepatide vs. semaglutide body composition: Some analyses suggest tirzepatide may have a modestly more favorable lean mass preservation profile than semaglutide, possibly due to the GIP component’s effects on adipose tissue differentiation and energy partitioning. The proposed molecular basis is the GIP receptor’s “fat buffering” function in adipocytes. The GIP receptor (GIPR) is highly expressed on white adipocytes and preadipocytes — among the highest tissue expression levels for this receptor. GIP → GIPR → Gs → cAMP → PKA → in adipocytes, this cascade promotes GLUT4 translocation (increasing adipocyte glucose uptake post-meal) and activates CREB transcription → CREB drives expression of PPAR-γ (the master adipocyte differentiation transcription factor) and its co-activator C/EBPα → promotes differentiation of preadipocytes into small, insulin-sensitive adipocytes rather than hypertrophic, insulin-resistant ones. The net result is what metabolic researchers call “fat buffering”: dietary caloric excess is preferentially channeled into well-differentiated adipose tissue with high GLUT4 and lipid uptake capacity, rather than accumulating as ectopic lipid in the liver (hepatic steatosis) or skeletal muscle (intramyocellular lipid, IMCL). Ectopic lipid accumulation — particularly IMCL — impairs insulin signaling in muscle through IRS-1 serine phosphorylation by DAG-activated PKC-θ, activating a pathway that blocks PI3K/Akt and suppresses muscle protein synthesis. By reducing ectopic lipid deposition via GIPR-driven adipocyte fat buffering, tirzepatide may maintain a more insulin-sensitive muscle environment during caloric restriction — supporting mTORC1-driven muscle protein synthesis and attenuating the net protein catabolism that causes lean mass loss. This is a mechanistic hypothesis that aligns with the observed GIP biology; prospective trials specifically comparing body composition with DXA scanning between tirzepatide and semaglutide are needed to confirm the clinical magnitude. This comparison is not definitively established but is consistent with the mechanistic expectations from GIP receptor biology in adipose tissue.

Oral Semaglutide for Weight Loss: OASIS-1

The appeal of a daily or weekly pill rather than a weekly injection has driven significant interest in oral GLP-1 formulations. Rybelsus (oral semaglutide) has been FDA-approved for T2D since 2019 at doses up to 14 mg daily — but at doses too low for robust weight loss effects.

The OASIS-1 trial (PMID: 37742727; published Lancet 2023) evaluated a once-weekly oral semaglutide 50 mg tablet (a substantially higher dose than Rybelsus, enabled by a reformulated absorption-enhancing technology) in adults with obesity without type 2 diabetes over 68 weeks.

Results:

15.1% body weight loss with oral semaglutide 50 mg vs. 2.4% placebo
Comparable efficacy to injectable semaglutide 2.4 mg (the STEP-1 result was 14.9%)
Tolerability: similar GI profile to injectable semaglutide, with slightly higher early nausea

What this means: A once-weekly oral pill achieving the same weight loss as the weekly injection represents a meaningful access and adherence improvement for patients with needle aversion, injection site issues, or logistical barriers. Regulatory submissions for this higher-dose oral formulation are in progress in the US and other markets.

Oral absorption caveats: Oral semaglutide requires strict administration conditions (fasting, specific water volume, no food/drink/other medications for 30 minutes after). These conditions affect absorption substantially — and the reason is mechanistic. The tablet formulation includes SNAC (sodium N-[8-(2-hydroxybenzoyl)amino]caprylate) as an absorption enhancer, a molecule specifically engineered to solve the problem of oral peptide delivery. SNAC works through two concurrent mechanisms: first, it transiently chelates Ca²⁺ ions in the gastric mucosal surface layer → disruption of epithelial tight junction integrity through two parallel molecular mechanisms. Claudin strand disruption: claudin paracellular strands form by head-to-head dimerization of extracellular loop 1 (ECL1) between apposing gastric epithelial cell membranes; ECL1 of each claudin contains a conserved glutamate residue (e.g., Glu64 in claudin-3) whose carboxylate, together with adjacent backbone carbonyl oxygens, coordinates a Ca²⁺ ion in an octahedral coordination shell that bridges the intercellular space and locks ECL1 dimers in their paracellular-sealing orientation. SNAC’s Ca²⁺ chelation collapses these intercellular Ca²⁺ bridges → claudin ECL1 dimerization fails → strand continuity is disrupted → paracellular gaps open. Occludin MARVEL domain destabilization: occludin’s MARVEL domain (a four-transmembrane-helix bundle with an extracellular domain contacting the opposing membrane) requires Ca²⁺ coordination for stable folding of its loop regions; SNAC chelation → MARVEL domain destabilization → additional paracellular strand discontinuities at occludin-containing TJ regions. Both mechanisms are rapidly reversible: once SNAC is diluted by gastric fluid or clears the mucosal surface, free Ca²⁺ rebinds claudin ECL1 coordination sites and occludin MARVEL → TJ resealing within minutes — creating a narrow delivery window rather than sustained barrier compromise → paracellular transport of the semaglutide-SNAC complex across the gastric epithelium — a local, reversible permeability effect that normalizes rapidly; second, SNAC forms a non-covalent hydrophobic complex with semaglutide that raises the local pH at the gastric mucosa surface and shields the peptide’s amide bonds from pepsin active-site access and from acid denaturation, maintaining semaglutide in its active conformation through the hostile gastric environment. Both mechanisms require SNAC to reach an effective local concentration at the gastric mucosa, which only occurs when the stomach is empty and the tablet dissolves in a small volume of water (~120 mL specified in the protocol). Food in the stomach dilutes SNAC, reduces the local concentration below the threshold for tight-junction modulation, and accelerates gastric emptying, routing the tablet to the intestine where semaglutide faces far more aggressive peptidase activity without mucosal permeability enhancement. This explains why even partial protocol violations — eating shortly before, drinking excess water, or taking other medications that alter gastric pH or motility — meaningfully reduce bioavailability. The OASIS-1 protocol enforced these conditions rigorously; real-world adherence to the administration protocol may reduce efficacy below trial conditions.

Adolescent Data: STEP TEENS

Until recently, GLP-1 pharmacotherapy for obesity was exclusively studied and approved for adults. That changed with the publication of STEP TEENS.

The STEP TEENS trial (PMID: 37159090; published NEJM 2022) evaluated semaglutide 2.4 mg in adolescents aged 12–17 with obesity (BMI ≥95th percentile) over 68 weeks.

Results:

16.1% reduction in BMI with semaglutide vs. -0.6% with placebo
This corresponds to approximately 14.8% of body weight lost
Cardiometabolic markers (waist circumference, blood pressure, lipids) improved significantly
Tolerability similar to adult trials; no new safety signals specific to adolescents

FDA approval: Wegovy is now FDA-approved for weight management in adolescents aged 12 and older with obesity.

Tirzepatide pediatric data: A Phase 3 trial evaluating tirzepatide in adolescents with obesity is ongoing. Results are anticipated in 2026–2027.

Clinical context: Pediatric obesity pharmacotherapy remains an active and evolving area. The availability of GLP-1 drugs for adolescents represents a meaningful expansion of treatment options for a population in whom lifestyle intervention alone has historically modest efficacy and bariatric surgery has high barriers. The modest behavioral response to lifestyle restriction in adolescent obesity is not a motivational failure — it reflects three overlapping biological features that are most pronounced during this developmental stage.

Pubertal physiology compounds adiposity-driven insulin resistance. During active pubertal development (Tanner stages 2–4), GH pulse amplitude increases substantially under rising hypothalamic GHRH tone. The resulting hepatic IGF-1 surge drives bone and tissue growth, but also generates physiological insulin resistance in peripheral tissues — a Tanner-stage-specific phenomenon in which elevated GH activates the GH receptor/JAK2/STAT5b axis in skeletal muscle and adipose tissue, and JAK2 leads to inhibitory phosphorylation of IRS-1 at Ser307 and Ser612 (the same sites targeted by ceramide and TNF-α in adult insulin resistance) — but the mechanism is indirect, because JAK2 is a tyrosine kinase and cannot directly phosphorylate these serine residues; the bridge requires an intermediate serine kinase: GH/GHR → JAK2 trans-phosphorylation at Tyr1007/1008 → JAK2 activates PI3K via direct binding of the p85 regulatory subunit → PIP3 → PDK1/Akt → mTORC1 → S6K1 (p70 ribosomal S6 kinase 1), which is the serine kinase that directly phosphorylates IRS-1 at Ser307/Ser612/Ser636 as a negative-feedback brake on the same IRS-1/PI3K pathway it was activated through — S6K1 is the molecular bridge that converts GH’s tyrosine-kinase cascade into IRS-1 serine phosphorylation. This is the same short-loop negative feedback that chronic hyperinsulinemia establishes over years in adult obesity via insulin/IR → Akt → mTOR → S6K1 → IRS-1 Ser307, but which pubertal GH surges establish transiently over hours through the parallel GH/JAK2/PI3K/S6K1 route rather than through ceramide or TNF-α. A parallel SOCS3-dependent arm reinforces IRS-1 suppression: JAK2 → STAT5b → SOCS3 transcription → SOCS3 protein binds IRS-1 via its SH2 domain and recruits an Elongin-BC/Cullin-5/RBX2 E3 ligase → IRS-1 polyubiquitination → proteasomal degradation, further depleting the IRS-1 pool available for insulin-driven PI3K signaling. This phosphoserine IRS-1 cannot recruit PI3K, reducing Akt/GLUT4 signaling. In a lean adolescent, the effect is transient and metabolically tolerable. In an adolescent with obesity, this pubertal JAK2/IRS-1 Ser307 insulin resistance is layered on top of existing ceramide- and diacylglycerol-driven IRS-1 Ser307 phosphorylation from adipose tissue inflammation — the two mechanisms converge on the same serine residue, producing a degree of systemic insulin resistance that substantially exceeds either cause alone. Lifestyle caloric restriction that succeeds in non-obese adolescents — where only pubertal insulin resistance is present — frequently fails to achieve adequate metabolic improvement in obese adolescents because the inflammatory IRS-1 Ser307 signal is not reduced by moderate caloric restriction alone.

Reward circuit immaturity increases food cue salience. The prefrontal cortex (PFC) — particularly the ventromedial PFC and anterior cingulate cortex — provides the primary top-down inhibitory control over food reward signals generated in the nucleus accumbens (NAc). This PFC-to-NAc inhibitory projection, operating through GABAergic medium spiny neuron (MSN) D2 receptor pathways, suppresses impulsive, cue-driven eating behavior in proportion to its myelination and synaptic maturity. PFC axonal myelination and PFC-NAc functional connectivity are not completed until the early-to-mid twenties, with adolescence representing the developmental trough of top-down inhibitory capacity relative to the subcortical dopamine system. In adolescents, food cues — palatable odors, visual exposure, peer eating — trigger dopamine release in the NAc via the mesolimbic ventral tegmental area (VTA) projection that is proportionally larger than the adult response (adolescent NAc D1 MSNs are hypersensitive to dopamine, partly due to incomplete pruning of the dopamine reuptake transporter DAT). The result is that the NAc “wanting” signal generated by a food cue significantly exceeds the immature PFC’s capacity to suppress it through cognitive override — a structural neurobiological constraint, not a character deficit. Lifestyle counseling that relies on prefrontal deliberate inhibition (“choose the salad, not the burger”) is asking adolescents to consistently override a limbic dopamine signal with an executive control system that is biologically not yet fully online. GLP-1 receptor activation in the VTA and NAc — reducing dopamine neuron firing (via HCN channel modulation) and attenuating incentive salience (via DARPP-32/PP1 pathway) — bypasses this structural limitation by reducing the size of the NAc dopamine signal before the PFC needs to inhibit it, explaining why pharmacological appetite suppression outperforms behavioral counseling in this age group.

Hypothalamic leptin sensitivity is acutely vulnerable during adolescence. Animal models have established that the arcuate nucleus undergoes a postnatal critical window of organizational development in which excess dietary fat, gut-derived inflammatory signals, or hyperleptinemic states can produce epigenetic modifications at the POMC and AgRP gene loci (H3K4me3 and H3K27ac mark shifts driven by DNMT3a methyltransferase activity) that durably shift the arcuate hunger circuit setpoint toward a higher defended adiposity (Vogt et al., PMID: 24813520). These epigenetic marks, once established during the adolescent developmental window, resist reversal by caloric restriction in adulthood — the AgRP/NPY orexigenic output remains elevated relative to the body’s fat mass, continuously counteracting caloric deficit with compensatory hunger escalation that behavioral strategies cannot fully override without pharmacological support. GLP-1R activation in the arcuate POMC neurons and NTS suppresses this AgRP-mediated hunger drive through Gs/cAMP/PKA signaling that operates downstream of the epigenetically modified promoter state — addressing the functional consequence of the hypothalamic setpoint shift rather than requiring the epigenetic programming itself to be reversed.

Early Signals: GLP-1 Drugs and Addiction

Beyond metabolic disease, an emerging — and potentially transformative — signal involves GLP-1 drugs reducing addictive behaviors.

Observational data: Multiple large database studies have found that GLP-1 drug users have lower rates of alcohol use disorder diagnoses, smoking continuation, and opioid overdose events compared to matched controls on other medications. These are observational associations, subject to confounding, but they are consistent across multiple independent datasets.

Mechanistic rationale: GLP-1 receptors are expressed in the mesolimbic dopamine system — the brain’s reward circuit — including the ventral tegmental area and nucleus accumbens. This is the same circuit that mediates substance use disorders. GLP-1 receptor activation in these regions may reduce dopamine reward signals from substances, blunting the craving and reinforcing properties of alcohol, nicotine, and opioids.

At the circuit level, two parallel mechanisms converge to reduce substance-driven reward. First, GLP-1Rs on ventral tegmental area (VTA) dopamine neurons couple to Gαs → adenylyl cyclase → cAMP → PKA, which phosphorylates and modulates hyperpolarization-activated cyclic nucleotide-gated (HCN) channels on the dopamine neuron membrane. HCN channels normally generate the pacemaker current (Ih) that drives spontaneous dopamine neuron firing; PKA-mediated modulation shifts HCN channel gating, reducing the Ih depolarizing drive → lower spontaneous and substance-cue-evoked dopamine neuron firing rate → smaller dopamine release bursts in the nucleus accumbens in response to substance cues (Fortin & Bhatt, 2021, PMID: 34285093).

Second, GLP-1Rs on nucleus accumbens (NAc) medium spiny neurons (MSNs) — the primary targets of mesolimbic dopamine projections — activate cAMP/PKA → phosphorylation of DARPP-32 (dopamine and cAMP-regulated phosphoprotein, 32 kDa) at Thr34, converting DARPP-32 into a potent inhibitor of protein phosphatase-1 (PP1). PP1 normally dephosphorylates AMPA receptor subunit GluA1 at Ser845, a phosphorylation site that controls AMPA receptor synaptic insertion and conductance. With PP1 inhibited by pThr34-DARPP-32, GluA1 Ser845 phosphorylation is maintained — but crucially, when dopamine stimulation is attenuated (via the VTA mechanism above), the net synaptic strengthening response to substance cues is blunted: fewer dopamine-driven postsynaptic potentials reach the threshold required to drive long-term potentiation at corticoaccumbal glutamate synapses. The behavioral correlate is reduced incentive salience — the “wanting” signal that drives substance-seeking — without eliminating baseline hedonic capacity or the ability to experience pleasure from natural rewards (Dickson et al., 2012, PMID: 22326765).

Animal model support is substantial: rodents treated with GLP-1 receptor agonists drink less alcohol, consume less nicotine, and show reduced opioid-seeking behavior in standard addiction paradigms.

Current clinical trial status: Multiple registered RCTs are now actively enrolling:

Semaglutide for alcohol use disorder (NCT05421676)
Semaglutide for smoking cessation (multiple trials) — the nicotine-specific mechanism is worth unpacking because it differs from the opioid pathway above. Nicotine binds α4β2 nicotinic acetylcholine receptors (nAChRs) on VTA dopamine neurons and on presynaptic cholinergic projections terminating in the VTA. α4β2 nAChR activation depolarizes the neuron membrane and increases the baseline firing rate of VTA DA neurons, driving phasic dopamine release into the nucleus accumbens shell. Critically, this nicotinic excitation is additive with existing VTA excitatory tone — the more depressed or stressed the basal state, the larger the relative DA burst nicotine produces, which is why stress exposure reliably increases craving in smokers trying to quit. GLP-1R agonism intervenes at the GABAergic interneuron step: GLP-1R is expressed on VTA GABAergic interneurons, and Gs/cAMP/PKA signaling in those neurons increases inhibitory GABA output onto DA projection neurons. Elevated GABAergic tone activates both GABA-A (rapid Cl⁻ influx → hyperpolarization) and GABA-B (Gi-coupled → reduced Ca²⁺ current → suppressed DA neuron action potential generation) receptors on the same VTA DA neurons that nicotine tries to excite. The net result: GLP-1R agonism reduces the ceiling dopamine burst amplitude that nicotine can produce, even when nicotine successfully engages and activates nAChRs. This is mechanistically distinct from varenicline (which directly occupies α4β2 nAChR as a partial agonist, reducing the DA signal via competitive receptor occupancy) — GLP-1 drugs work downstream of receptor binding, at the level of DA neuron excitability itself. An additional feature relevant to chronic smokers: prolonged nicotine exposure causes α4β2 nAChR desensitization, reducing cholinergic drive over time and contributing to the “flat affect” withdrawal syndrome. GLP-1R’s inhibitory route (GABAergic → DA neuron hyperpolarization) is entirely independent of nAChR occupancy state, meaning it remains pharmacologically active even when the nAChR desensitization cycle has dampened the excitatory signal — a potential reason GLP-1 agonism may complement rather than substitute for nicotinic approaches
GLP-1 agonists for opioid use disorder (multiple trials) — the opioid-specific circuit mechanism differs from both the alcohol and nicotine mechanisms above, because opioids produce their reward via a disinhibition cascade rather than through direct DA neuron excitation or synaptic modulation. Mu-opioid receptors (MOR, OPRM1) are densely expressed on VTA GABAergic interneurons — the tonic inhibitory neurons that provide a continuous hyperpolarizing brake on VTA dopamine projection neurons. Opioid binding to MOR on these interneurons → Gi → inhibition of adenylyl cyclase (reduced cAMP) + Gβγ → GIRK channel (G-protein-coupled inwardly-rectifying K⁺ channel, Kir3.1/3.2 heteromer) activation → K⁺ efflux → sustained GABAergic interneuron hyperpolarization → interneurons fire less and release less GABA onto the VTA DA projection neurons they normally restrain. With their primary inhibitory input removed, VTA DA neurons undergo disinhibition — their firing rate and burst propensity increase substantially — producing the large-amplitude phasic dopamine release into the nucleus accumbens shell that underlies opioid euphoria and reinforcement. This disinhibition mechanism is mechanistically distinct from nicotine’s direct nAChR excitation of DA neurons: nicotine adds an excitatory current to DA neurons; opioids remove an inhibitory current. The behavioral result — increased NAc DA — is similar, but the circuit node is different. GLP-1R agonism intervenes at the level of the VTA DA neuron’s intrinsic excitability (via HCN channel gating, as described above) — reducing the ceiling amplitude of DA neuron burst firing regardless of whether the disinhibition is triggered by opioid MOR/GIRK interneuron silencing or by nicotinic direct excitation. When GLP-1R reduces HCN-driven pacemaker current on the VTA DA neurons, those neurons cannot fire as rapidly even when they are maximally disinhibited by opioid-silenced interneurons → NAc dopamine release is attenuated at a step downstream of MOR activation → reduced incentive salience from opioid cues. A secondary mechanism on NAc MSNs (DARPP-32/PP1 reducing corticoaccumbal LTP, as described for the general reward circuit above) further reduces the associative memory consolidation linking opioid administration to reward context. This dual-point intervention — reduced DA neuron ceiling excitability + reduced NAc synaptic plasticity from drug-paired cues — provides a mechanistically coherent rationale for GLP-1 agonism in opioid use disorder that is opioid-circuit-specific rather than simply a downstream consequence of appetite suppression

Results from these trials are expected in 2026–2028. Current evidence is insufficient to support using GLP-1 drugs as addiction treatments — but if the RCT data confirms the observational and preclinical signals, this would represent one of the most significant expansions of GLP-1 drug indications to date.

The New Safety Signal: Nonarteritic Anterior Ischemic Optic Neuropathy (NAION)

In 2025, the FDA issued a safety communication requiring label updates for semaglutide injectable products following reports of nonarteritic anterior ischemic optic neuropathy (NAION) — a rare but serious condition that can cause sudden, painless vision loss in one eye due to reduced blood flow to the optic nerve.

What Is NAION?

NAION affects approximately 10 per 100,000 people annually in the general population and is associated with risk factors including hypertension, diabetes, hyperlipidemia, sleep apnea, and small “crowded” optic disc anatomy.

The FDA’s safety communication was prompted by observational data — including a 2024 study in JAMA Ophthalmology (PMID: 38935394) — suggesting that semaglutide users had a higher rate of NAION compared to matched non-users. The association was more pronounced in patients with pre-existing diabetes and prior non-arteritic optic neuropathy.

The Proposed Mechanism: Why Semaglutide Might Cause Optic Nerve Ischemia

The optic nerve head receives its arterial supply exclusively from short posterior ciliary arteries (SPCAs) — small-caliber terminal branches of the ophthalmic artery that enter the sclera posterior to the globe and perfuse the pre-laminar and laminar portions of the optic nerve. Unlike cerebral arteries, which have well-characterized pressure autoregulation over a wide MAP range, the SPCA capillary bed has limited autoregulatory reserve — it cannot fully compensate for reductions in systemic perfusion pressure through arteriolar dilation. Two converging mechanisms have been proposed to explain how semaglutide might reduce effective optic nerve head perfusion in susceptible patients.

Mechanism 1 — Systemic blood pressure and perfusion pressure reduction: GLP-1 receptor agonists reduce systolic blood pressure by approximately 5–7 mmHg — through their natriuretic effects (GLP-1R/cAMP/PKA on proximal tubule NHE3) and direct vasodilatory effects (GLP-1R → eNOS/NO → cGMP/PKG → vascular smooth muscle relaxation). Weight loss independently reduces cardiac output and systemic vascular resistance over weeks. In patients with pre-existing microvascular disease (diabetic retinopathy, hypertensive arteriolosclerosis of the SPCAs), the functional perfusion reserve of the optic nerve head capillary bed is already reduced — these vessels cannot dilate further in response to falling systemic MAP. When MAP reduction from GLP-1 therapy approaches the lower limit of SPCA autoregulation, optic nerve head blood flow falls below the critical ischemia threshold → segmental infarction of the anterior optic nerve fiber layer → axonal swelling and vision loss characteristic of NAION. This explains the stronger signal in diabetic patients, whose SPCA microvascular disease is a known independent NAION risk factor.

Mechanism 2 — Trans-laminar pressure gradient shift: GLP-1-induced weight loss in obese patients reduces elevated intracranial pressure (ICP) — ICP is directly correlated with BMI (each 5 BMI unit increase raises ICP by approximately 1–2 cmH₂O in population studies). Elevated ICP provides a “back-pressure” at the lamina cribrosa — the fenestrated scleral plate through which all optic nerve axons pass — that counterbalances intraocular pressure (IOP). The trans-laminar pressure gradient = IOP − ICP; elevated ICP reduces this gradient, buffering the mechanical stress on axons crossing the lamina. When weight loss rapidly normalizes ICP, the trans-laminar gradient increases: the same IOP (unchanged) now acts across a lower ICP support → increased mechanical stress at the lamina cribrosa on axons passing through its tight fenestrations → in patients with small “crowded” optic disc anatomy where the disc-to-cup ratio is structurally unfavorable, this new mechanical stress can precipitate ischemia in already marginal vascular beds. This mechanism is independent of semaglutide’s direct receptor pharmacology — it is a consequence of the desired weight loss — and would theoretically apply to any weight loss intervention achieving comparable ICP reduction in susceptible patients.

Both mechanisms are supported by biological plausibility rather than direct experimental evidence in the semaglutide context. The causal relationship between semaglutide and NAION is not established — the observational associations require confirmation in prospective data with individual-level covariates accounting for disc anatomy, ICP trajectory, and perfusion reserve.

How to Interpret This

The absolute risk remains low. NAION is rare in the general population, and even with an elevated relative risk in semaglutide users, the absolute additional risk per patient per year is small.

However, certain patient groups deserve specific discussion:

Patients with prior NAION in one eye: Risk of NAION in the second eye may be elevated; discuss with an ophthalmologist before starting or continuing semaglutide
Patients with optic disc drusen or small “disc at risk” anatomy: Structural vulnerability may increase risk. “Disc at risk” anatomy refers to a small optic disc diameter with high axon-to-disc-area ratio — the same number of retinal ganglion cell axons (approximately 1.2 million) must pass through a narrower scleral canal, increasing axon packing density at the lamina cribrosa. This structural configuration amplifies NAION risk through two converging mechanisms: (1) Perfusion reserve: The optic disc rim is supplied by short posterior ciliary arteries (SPCAs) whose branches are proportional in number and caliber to disc surface area; a small disc has fewer SPCA branch points and less collateral perfusion reserve, so any reduction in SPCA perfusion pressure — from nocturnal hypotension, vasoconstriction, or ICP-driven trans-laminar pressure increase — crosses the ischemic threshold across a larger fraction of the available territory than it would in a large disc with redundant collateral channels. (2) Lamina cribrosa mechanics: The lamina cribrosa pore fenestrations scale with disc diameter; small-disc patients have narrower fenestrations, concentrating axon bundles under higher compressive stress per unit cross-sectional area when trans-laminar pressure gradients increase. The combination of reduced perfusion reserve and increased mechanical vulnerability per axon means that hemodynamic events that are subclinical in normal anatomy produce frank ischemia and axon loss in small-disc anatomy — explaining why “disc at risk” is an independent NAION risk factor that compounds the perfusion and ICP mechanisms described above
Patients with poorly controlled hypertension: The combination of hypertensive optic nerve stress and any semaglutide-associated effect warrants attention

What to do: Any new vision change — sudden blurring, dimming, or loss of vision in one eye — in a patient taking semaglutide should prompt urgent evaluation by an eye care professional. This is not a reason to avoid semaglutide for appropriate candidates, but it is a reason to know the symptom and act on it promptly.

Note: The current safety communication concerns semaglutide specifically. Tirzepatide has not yet accumulated equivalent epidemiological data; similar signals are possible but not currently established.

Updated Anesthesia Guidance

Updated guidance from the American Society of Anesthesiologists (ASA, 2023) — now being implemented with broader clinical experience — recommends:

Weekly GLP-1 medications (semaglutide, tirzepatide): hold for at least 7 days before elective procedures requiring general anesthesia
Daily GLP-1 medications (liraglutide): hold for at least 1 day before procedures
If the drug cannot be held (urgent surgery), treat as a full stomach with appropriate precautions

The mechanism: GLP-1 receptor agonists significantly slow gastric emptying through three converging pathways. First, enteric nervous system inhibition: GLP-1Rs on myenteric plexus motor neurons → Gs/cAMP/PKA → reduced acetylcholine release from excitatory motor neurons that drive antral peristalsis, while augmenting VIP and nitric oxide release from inhibitory motor neurons → antrum contracts less forcefully and less frequently. Second, vagal afferent ileal brake: GLP-1Rs on nodose ganglion vagal afferents → increased vagal sensory firing → NTS → dorsal motor nucleus of vagus (DMNV) → efferent cholinergic tone to the pyloric sphincter → maintained pyloric closure between antral contractions → increased gastric outflow resistance. Third, direct pyloric smooth muscle effect: GLP-1R/Gs/cAMP/PKA on pyloric smooth muscle → PKA phosphorylation of myosin light-chain kinase (MLCK) → reduced contractile responsiveness of antral drive relative to tonic pyloric resistance → net slowing of gastric outflow. At pharmacological doses, these mechanisms are sustained around the clock — unlike endogenous GLP-1, which is rapidly inactivated by DPP-IV. This sustained, 24-hour gastroparesis-like slowing is what creates aspiration risk during anesthesia even after standard overnight fasting; residual gastric contents can remain hours longer than expected. Even patients who have fasted according to standard NPO guidelines may have residual gastric contents under GLP-1 influence, increasing aspiration risk during airway management.

Practical action: If you are taking semaglutide or tirzepatide and have any planned surgical procedure — including dental procedures under sedation — inform both your prescribing physician and your surgical team. The conversation about when to hold your dose should happen before the procedure, not the morning of.

Updated Evidence Summary: 2026

Question	At Approval (~2021–2022)	Evidence in 2026
Which drug is more effective for weight loss?	Unknown (no head-to-head)	Tirzepatide superior (~20% vs ~14% body weight loss, SURMOUNT-5)
Is cardiovascular benefit durable?	3-year data only	Yes — 5-year SELECT extension confirms durability; partially weight-independent
Do these drugs protect kidneys?	Preliminary signals	Yes — FLOW trial: 24% reduction in CKD composite outcome; new FDA indication
Do these drugs help heart failure?	Unknown	Yes — STEP-HFpEF: symptom improvement + weight loss; new FDA indication
What happens when you stop?	Expected regain based on mechanism	Substantial regain: ~14 percentage points over 88 weeks (SURMOUNT-4)
How much lean mass is lost?	Uncertain	~25–40% of weight lost is lean mass; resistance exercise + protein mitigate this
Can an oral GLP-1 pill match injection?	Unlikely at approved doses	OASIS-1: once-weekly oral sema 50mg achieves ~15% weight loss — comparable to injection
Are these drugs safe for adolescents?	Adult-only data	STEP TEENS: 16.1% BMI reduction; Wegovy FDA-approved age 12+
Rare safety signals?	Initial GI profile	NAION signal added to semaglutide labeling (2025)
Can GLP-1 drugs reduce addiction?	Mechanistic hypothesis only	Compelling observational data; RCTs underway; not yet clinical recommendation

Practical Guidance for Current and Prospective GLP-1 Patients

If you are deciding between semaglutide and tirzepatide: SURMOUNT-5 data supports tirzepatide as the more effective weight-loss agent. For patients with established CVD, semaglutide has more mature cardiovascular outcome data. For CKD with albuminuria, semaglutide now has a specific FDA indication — tirzepatide does not. Discuss your full medical picture with your prescriber.

If you are achieving partial response on semaglutide: Switching to tirzepatide has clinical evidence supporting it as a reasonable next step before escalating other interventions. The SURMOUNT-5 data suggests the weight loss advantage is clinically meaningful, not trivial.

If you are concerned about muscle loss: Prioritize resistance training (2–3 sessions weekly) and high-protein intake (aim for 25–30g per meal) throughout the weight-loss phase. These are not optional lifestyle additions — they are the main evidence-based tools to preserve lean mass during GLP-1 therapy.

If you are considering stopping GLP-1 therapy: Have an honest conversation with your prescriber about what SURMOUNT-4 means for your situation. Stopping is sometimes the right decision, but it should be made with clear expectations about probable weight regain and a plan for managing it. The muscle loss/regain-as-fat compounding problem makes stopping without a maintenance plan riskier than the weight loss alone suggests.

If you have vision symptoms while on semaglutide: Any sudden or new vision change — particularly blurring or vision loss in one eye — warrants same-day or next-day evaluation by an eye care professional, not a “wait and see” approach.

If you have surgery planned: Inform your surgical team that you are on a GLP-1 medication and discuss ASA guidance on holding your dose before the procedure.

References

Jastreboff AM, et al. Tirzepatide vs. Semaglutide for Weight Loss in Adults with Obesity (SURMOUNT-5). Presented ObesityWeek 2025. ClinicalTrials.gov NCT05822154
Lincoff AM, Brown-Frandsen K, Colhoun HM, et al. Semaglutide and Cardiovascular Outcomes in Obesity without Diabetes (SELECT). N Engl J Med. 2023;389(24):2221-2232. PMID: 37952131
Aronne LJ, et al. Continued Treatment with Tirzepatide for Maintenance of Weight Reduction (SURMOUNT-4). JAMA. 2024;331(1):38-48. PMID: 38078870
Perkovic V, et al. Semaglutide and Kidney Outcomes in Type 2 Diabetes and Kidney Disease (FLOW). N Engl J Med. 2024;391(2):109-121. PMID: 38587991
Kosiborod MN, et al. Semaglutide in Patients with Heart Failure with Preserved Ejection Fraction and Obesity (STEP-HFpEF). N Engl J Med. 2023;389(12):1069-1084. PMID: 37350574
Knop FK, et al. Oral Semaglutide 50 mg for Weight Loss in Overweight or Obesity (OASIS-1). Lancet. 2023;402(10403):705-719. PMID: 37742727
Weghuber D, et al. Once-Weekly Semaglutide in Adolescents with Obesity (STEP TEENS). N Engl J Med. 2022;387(24):2245-2257. PMID: 37159090
Hathaway JT, Shah MP, Hathaway DB, et al. Risk of Nonarteritic Anterior Ischemic Optic Neuropathy in Patients Prescribed Semaglutide. JAMA Ophthalmol. 2024;142(8):732-739. PMID: 38935394

GLP-1 Medications in 2026: What New Clinical Data Reveals About Tirzepatide, Semaglutide, Weight Regain, and Long-Term Use

The Head-to-Head: SURMOUNT-5

What This Means in Practice

Long-Term Cardiovascular Benefit: SELECT 5-Year Extension

Kidney Protection: The FLOW Trial

Heart Failure: STEP-HFpEF

The Cardiac Mechanisms: How GLP-1 Drugs Help HFpEF

The Weight Regain Problem: SURMOUNT-4

Why This Matters — and Why Regain Is Biologically Inevitable Without Medication

Body Composition: The Lean Mass Concern

What the Evidence Supports to Minimize Lean Mass Loss

Oral Semaglutide for Weight Loss: OASIS-1

Adolescent Data: STEP TEENS

Early Signals: GLP-1 Drugs and Addiction

The New Safety Signal: Nonarteritic Anterior Ischemic Optic Neuropathy (NAION)

What Is NAION?

The Proposed Mechanism: Why Semaglutide Might Cause Optic Nerve Ischemia

How to Interpret This

Updated Anesthesia Guidance

Updated Evidence Summary: 2026

Practical Guidance for Current and Prospective GLP-1 Patients

References

Read Next

Anti-Aging Peptide Protocols: A Framework for Evidence-Based Longevity Approaches

Anti-Aging Peptides: What the Research Actually Shows

AOD-9604 and Fragment 176-191: What the Clinical Trial Record Actually Shows

Ready to Apply These Insights?

The Head-to-Head: SURMOUNT-5

What This Means in Practice

Long-Term Cardiovascular Benefit: SELECT 5-Year Extension

Kidney Protection: The FLOW Trial

Heart Failure: STEP-HFpEF

The Cardiac Mechanisms: How GLP-1 Drugs Help HFpEF

The Weight Regain Problem: SURMOUNT-4

Why This Matters — and Why Regain Is Biologically Inevitable Without Medication

Body Composition: The Lean Mass Concern

What the Evidence Supports to Minimize Lean Mass Loss

Oral Semaglutide for Weight Loss: OASIS-1

Adolescent Data: STEP TEENS

Early Signals: GLP-1 Drugs and Addiction

The New Safety Signal: Nonarteritic Anterior Ischemic Optic Neuropathy (NAION)

What Is NAION?

The Proposed Mechanism: Why Semaglutide Might Cause Optic Nerve Ischemia

How to Interpret This

Updated Anesthesia Guidance

Updated Evidence Summary: 2026

Practical Guidance for Current and Prospective GLP-1 Patients

References

Stay Informed on Peptide Research

Read Next

Anti-Aging Peptide Protocols: A Framework for Evidence-Based Longevity Approaches

Anti-Aging Peptides: What the Research Actually Shows

AOD-9604 and Fragment 176-191: What the Clinical Trial Record Actually Shows

Ready to Apply These Insights?