A hospital-affiliated endocrinology clinic in Shanghai ran its first formulary review of 2026 with a simple question on the table: which GLP-1 receptor agonist belonged on the preferred list for patients who had plateaued on semaglutide. The name that kept surfacing in the trial registries wasn't one most Western clinicians had heard of. Ecnoglutide XW003 Phase 3 data, published across two pivotal China-based trials, describe a molecule engineered around a different pharmacological premise than the GLP-1 agonists already on the market: biasing the GLP-1 receptor toward cyclic AMP (cAMP) signaling while minimizing the beta-arrestin recruitment that drives receptor internalization. Sciwind Biosciences, the Hangzhou-based developer, now holds China's National Medical Products Administration (NMPA) approval for ecnoglutide in type 2 diabetes, a Phase 3 obesity dataset published in The Lancet Diabetes & Endocrinology, and a commercialization deal with Pfizer China worth up to $495 million in milestones. None of that has translated into FDA review. What follows is a look at what the trial data actually show, what the biased-signaling mechanism is hypothesized to change, and what remains unverified outside a Chinese regulatory and patient population.
What Is Ecnoglutide (XW003)?
Ecnoglutide, internally coded XW003 during development, is a long-acting GLP-1 receptor agonist administered by once-weekly subcutaneous injection. It is structurally a modified GLP-1 analog rather than a small molecule, placing it in the same peptide-engineering lineage as semaglutide (Ozempic/Wegovy) and liraglutide (Victoza/Saxenda), but with amino-acid substitutions and a fatty-acid conjugation strategy designed to extend half-life and to shift which downstream signaling pathway the activated receptor favors.
In January 2026, ecnoglutide injection became the first cAMP-biased GLP-1 receptor agonist to receive NMPA approval, specifically for glycemic control in adults with type 2 diabetes. Sciwind has separately filed for a weight-management indication in China, and Phase 3 obesity data are already in peer-reviewed print. Shortly after the diabetes approval, Pfizer entered a licensing agreement to commercialize ecnoglutide in mainland China — Sciwind retains the marketing authorization and continues to lead R&D, registration, and manufacturing.
An oral formulation, XW004, has completed a Phase 1 trial with reported positive results, suggesting Sciwind is pursuing the same injectable-to-oral trajectory that Novo Nordisk followed with semaglutide. As of this writing, ecnoglutide has no FDA submission on record and is not approved, prescribed, or legally marketed in the United States, the EU, or the UK.
The cAMP-Biased Mechanism: Why Signal Bias Matters
GLP-1 receptor activation triggers at least two distinguishable downstream processes: G-protein-coupled cAMP production, which drives insulin secretion and satiety signaling, and beta-arrestin recruitment, which drives receptor internalization and subsequent desensitization. Most approved GLP-1 agonists, including semaglutide, activate both pathways without strong preference. Ecnoglutide was engineered to separate them.
In vitro pharmacology from the compound's discovery paper, published in Acta Pharmaceutica Sinica B (Zhang et al., 2023), reports a cAMP EC50 of 0.018 nM against an internalization EC50 exceeding 10,000 nM — a greater than 5.5 × 10^5-fold separation between the two signaling arms. Head-to-head cell-based assays in the same body of work found lower maximal beta-arrestin recruitment and markedly lower receptor internalization for ecnoglutide relative to semaglutide at equivalent concentrations.
The hypothesis driving the clinical program is that reduced receptor internalization preserves a larger pool of surface-available receptors over the weekly dosing interval, which could translate into greater or more sustained efficacy per unit dose. That is a mechanistic hypothesis, not a settled clinical fact — bias ratios measured in HEK293 reporter assays do not automatically predict human outcomes, and comparative efficacy still needs to be demonstrated in blinded, adequately powered trials rather than inferred from binding kinetics. For context on how an earlier generation of agonists without engineered bias performed against each other in confirmatory trials, see the head-to-head data from semaglutide vs. liraglutide in the STEP 8 trial.
EECOH-1: Phase 3 Glycemic Data in Type 2 Diabetes
EECOH-1, the registrational trial behind China's diabetes approval, was a multicenter, randomized, double-blind, placebo-controlled Phase 3 study (Nature Communications, PMID 41501026; registered as a type 2 diabetes monotherapy study, NCT05680155). It randomized 211 adults with type 2 diabetes across four arms: ecnoglutide 0.6 mg (n=69), ecnoglutide 1.2 mg (n=71), and two matched placebo cohorts (n=36 and n=35), followed by a 24-week double-blind period and a 28-week open-label extension.
At week 24, HbA1c fell by 1.96% with 0.6 mg and 2.43% with 1.2 mg, against a 0.87% reduction with placebo (p=0.0003 and p<0.0001, respectively). HbA1c below 7.0% was achieved by 68.1% (0.6 mg) and 80.3% (1.2 mg) of participants, versus 21.1% on placebo. Fasting plasma glucose fell by 2.89 mmol/L and 3.32 mmol/L in the two active arms, versus 1.21 mmol/L with placebo.
Body weight, a secondary endpoint in this diabetes population, decreased by 4.51% (0.6 mg) and 4.74% (1.2 mg) against 2.02% with placebo. Treatment-emergent adverse events occurred in 78.3% (0.6 mg) and 77.5% (1.2 mg) of ecnoglutide-treated patients versus 63.4% on placebo, driven by diarrhea (24.6% in the 0.6 mg arm) and nausea (12.7% in the 1.2 mg arm). Three participants discontinued for adverse events across the entire study, one per cohort, and no severe hypoglycemia was reported.
SLIMMER: Phase 3 Weight-Loss Data in Adults With Obesity
The obesity program's primary readout, the SLIMMER trial (ClinicalTrials.gov NCT05813795), published in The Lancet Diabetes & Endocrinology and presented at the American Diabetes Association's 85th Scientific Sessions, enrolled 664 adults across 36 centers in China. Eligibility required a BMI of 24-28 kg/m2 with at least one weight-related comorbidity, or higher — a threshold considerably lower than the ≥30 kg/m2 (or ≥27 with comorbidity) bar used in most Western obesity trials, reflecting regional BMI-risk thresholds rather than a different disease definition.
Participants were randomized to placebo (n=165) or ecnoglutide 1.2 mg (n=166), 1.8 mg (n=166), or 2.4 mg (n=167), once weekly for 48 weeks. At the top dose, mean weight loss reached 13.2% by week 24 and 15.4% by week 48, against an essentially flat placebo trajectory — yielding an approximate 15-percentage-point placebo-adjusted effect.
Response-rate data at week 48 for the 2.4 mg arm: 92.8% of participants lost ≥5% of body weight (versus 13.9% on placebo), 79.6% lost ≥10% (versus 4.8%), 63.5% lost ≥15% (versus 1.2%), and 27.5% lost ≥20% (versus 0%). Secondary endpoints included a 12.8 cm reduction in waist circumference, a 53.1% relative reduction in liver fat content among those with baseline hepatic fat ≥8%, and a uric acid reduction of up to 54.3 µmol/L. The most frequent adverse events were mild-to-moderate gastrointestinal disorders and decreased appetite, concentrated in the dose-escalation period; the trial reported no cases of pancreatitis, medullary thyroid carcinoma, or severe gallbladder disease.
Head-to-Head: Ecnoglutide vs. Semaglutide at Week 20
A smaller, open-label Phase 2 trial (NCT07073417) directly compared ecnoglutide against semaglutide in 163 adults with obesity (BMI ≥30 kg/m2) across 17 Chinese centers. This is the only published randomized comparison of the two molecules head-to-head, and its evidence tier — Phase 2, open-label, 20-week — is meaningfully lower than the blinded, 48-week SLIMMER dataset above.
At week 20, the least-squares mean percent change in body weight was -12.8% for ecnoglutide versus -9.5% for semaglutide (p<0.0001), a relative difference the sponsor's press materials characterized as roughly 35% greater weight loss. Response rates followed the same pattern: 99% of ecnoglutide-treated participants reached ≥5% weight loss versus 86% on semaglutide (p<0.01), and 74% reached ≥10% versus 40% on semaglutide (p<0.001). Waist circumference fell 10.5 cm with ecnoglutide versus 8.7 cm with semaglutide.
These numbers are notable, but open-label design removes blinding as a safeguard against assessment bias, and the trial was not powered or designed as a definitive superiority study in the way STEP or SURMOUNT-class trials were for semaglutide and tirzepatide. A double-blind, adequately powered confirmatory trial against a licensed semaglutide formulation — ideally outside the Chinese regulatory and dosing-titration context — has not yet been published. Until then, the 35% figure should be read as a Phase 2 signal worth tracking, not a settled comparative-efficacy claim.
Adverse Event Profile and Safety Signals Across Trials
Pooling the safety signal across EECOH-1 and SLIMMER, the adverse-event pattern looks consistent with the established GLP-1 class: gastrointestinal events (nausea, diarrhea, decreased appetite, and vomiting) dominate, cluster in the dose-escalation window, and are described as mild to moderate in severity. Treatment-emergent AE rates ran roughly 15-30 percentage points above placebo depending on dose and trial, which is in line with what pivotal trials for semaglutide and tirzepatide have reported.
No pancreatitis, medullary thyroid carcinoma, or severe gallbladder events were reported in either Phase 3 trial, though both trials' sample sizes (n=211 and n=664) are far too small to rule out rare events — pancreatitis incidence in pooled GLP-1 trial data runs in the range of a few cases per thousand patient-years, which requires much larger or longer-running datasets to characterize with confidence.
Neither EECOH-1 nor SLIMMER reported DEXA or other body-composition substudy data, so it remains unknown what proportion of ecnoglutide-induced weight loss is fat mass versus lean mass. That question has been directly studied for other GLP-1 agents; see the DEXA substudy findings on muscle loss on GLP-1 therapy for the kind of data that is currently missing from the ecnoglutide program. Cardiovascular outcomes, which took years of dedicated outcome trials to establish for semaglutide and liraglutide, have no published ecnoglutide equivalent.
Regulatory Status: NMPA Approval, the Pfizer Deal, and the US Gap
Ecnoglutide's regulatory position is almost entirely China-specific. NMPA approval for type 2 diabetes was granted in January 2026, and a weight-management filing is pending with the same agency. Pfizer's licensing deal, announced in February 2026 and worth up to $495 million in milestone payments, grants Pfizer commercialization rights in mainland China while Sciwind retains the marketing authorization and continues to run manufacturing, registration, and ongoing R&D.
There is no FDA Investigational New Drug filing, no EMA submission, and no ClinicalTrials.gov record of a US-based trial site for ecnoglutide as of this writing [CITATION NEEDED: current FDA/EMA filing status for ecnoglutide, subject to change]. That absence matters for anyone encountering the compound outside China: it is not a prescribable option in US clinical practice, and any product claiming otherwise is operating outside the approved regulatory pathway that governs drugs like semaglutide and tirzepatide domestically.
That regulatory gap is distinct from, but conceptually related to, the compounding and gray-market dynamics that have shaped US access to other GLP-1 molecules after supply disruptions. For a detailed look at how one such channel was resolved domestically, see the current legal status of compounded semaglutide after the FDA shortage resolution. No equivalent compounding pathway exists for ecnoglutide in the US, because the molecule has never been FDA-approved in the first place — a materially different legal position than a shortage-driven compounding exception for an approved drug.
Supplier Landscape and Research-Use Realities
Because ecnoglutide is unapproved outside China, any vial reaching a US or EU address through a peptide-research vendor arrives under research-use-only (RUO) labeling rather than pharmaceutical-grade chain of custody. That labeling is a legal distinction, not a quality guarantee: RUO suppliers are not required to meet the same GMP manufacturing, batch-testing, or cold-chain documentation standards that NMPA-approved clinical material from Sciwind must satisfy for the Chinese market.
Lyophilized peptide products in this category are also sensitive to reconstitution and storage practice in ways that directly affect potency. Bacteriostatic water ratios, mixing technique, and refrigeration windows all influence how much intact peptide remains active by the time a vial is used — variables that are well documented for other lyophilized peptides but have no published stability data specific to ecnoglutide [CITATION NEEDED: peer-reviewed reconstitution stability data for ecnoglutide]. General protocols for handling this class of compound are covered in the breakdown of reconstituting lyophilized peptides, including bacteriostatic water ratios and storage life.
The absence of independent third-party purity testing across RUO vendors selling ecnoglutide-labeled product is a documented pattern across the broader research-peptide market, not specific to this compound, but it is a material consideration for anyone evaluating sourced material against the dosing used in the published Phase 3 trials.
How Ecnoglutide Compares With Other Next-Generation GLP-1 Agonists
Ecnoglutide enters a crowded field of molecules trying to outperform first-generation GLP-1 monotherapy through different mechanisms rather than simple dose escalation. Tirzepatide pairs GLP-1 and GIP receptor agonism; retatrutide adds glucagon receptor activity to that dual mechanism; survodutide and mazdutide pursue GLP-1/glucagon dual agonism along separate development paths; cagrilintide pairs an amylin analog with semaglutide rather than altering GLP-1 receptor signaling itself.
Ecnoglutide's differentiator is narrower and more mechanistic: it keeps single-receptor GLP-1 agonism but changes which intracellular pathway that single receptor favors. That is a meaningfully different strategy than combination or multi-receptor approaches, and it has not yet been tested against them in a shared trial — there is no published trial comparing ecnoglutide against tirzepatide, retatrutide, or a cagrilintide/semaglutide combination. The amylin-pathway strategy — a leading example of a non-GLP-1-receptor approach to the same obesity-pharmacology problem — is covered in the mechanism breakdown of why cagrilintide is being paired with GLP-1 agonists.
Cross-trial comparison across any of these programs is inherently limited — different populations, different BMI eligibility criteria, different trial durations, and different placebo-adjustment methods make percentage-point comparisons between press releases unreliable. The 15.4% week-48 weight loss reported for ecnoglutide's top dose in SLIMMER sits in the same general range as tirzepatide's SURMOUNT-1 results and below retatrutide's reported Phase 2 ceiling, but none of these figures come from a shared comparator arm.
What's Still Unclear: Gaps in the Evidence Base
Several open questions limit how far the current ecnoglutide dataset can be extrapolated. First, every published Phase 2 and Phase 3 trial to date has enrolled exclusively Chinese participants under China-specific BMI eligibility criteria; pharmacokinetic and efficacy differences across ethnic and regional populations are well documented for other drug classes and have not been tested for ecnoglutide specifically.
Second, no dedicated cardiovascular outcomes trial has been registered or published — a category of study that took Novo Nordisk and Eli Lilly years to complete for their respective GLP-1 portfolios and that regulators increasingly expect before broad long-term use is considered established practice. Third, durability beyond 48 weeks, discontinuation effects on weight regain, and combination use with amylin analogs or SGLT2 inhibitors remain unpublished.
Fourth, the head-to-head semaglutide comparison (NCT07073417) is open-label and Phase 2; a blinded confirmatory trial would carry substantially more weight in any comparative-efficacy claim. Researchers and clinicians tracking this compound have a concrete next step: monitor the NCT05813795 and NCT07073417 ClinicalTrials.gov records directly for extension-phase results, long-term safety updates, and any move toward FDA pre-IND engagement, rather than relying on sponsor press releases as a stand-in for peer-reviewed, blinded, adequately powered confirmatory data.
This article summarizes research and does not constitute medical advice. Consult a licensed clinician for diagnosis, treatment, or any decisions about medications or supplements.