Mounjaro Mechanism of Action: The Full GIP + GLP-1 Dual-Agonist Pathway

GLP-1 medication and metabolic health image for Mounjaro Mechanism of Action: The Full GIP + GLP-1 Dual-Agonist Pathway

At a glance

  • Drug class / GIP + GLP-1 dual receptor agonist (twincretin)
  • Molecular weight / approximately 4,813 Da fatty-acid conjugated peptide
  • Primary receptor targets / GIPR and GLP-1R (distinct G-protein coupled receptors)
  • FDA approval date / May 13, 2022 (type 2 diabetes); weight loss indication via Zepbound, November 2023
  • Key glucose mechanism / glucose-dependent insulin secretion + glucagon suppression
  • Key weight mechanism / hypothalamic appetite suppression + adipose lipolysis + delayed gastric emptying
  • SURPASS-2 A1C reduction / up to 2.37 percentage points vs. 1.86 for semaglutide 1 mg (P<0.001)
  • SURPASS-2 weight loss / up to 12.4 lb more than semaglutide 1 mg at 40 weeks
  • Dosing / 2.5 mg SC weekly titrating to 5, 7.5, 10, 12.5, or 15 mg weekly
  • Half-life / approximately 5 days, enabling once-weekly dosing

What Is Tirzepatide and Why Does Dual Agonism Matter?

Tirzepatide is a 39-amino-acid peptide engineered to bind and activate both the GIPR and the GLP-1R with comparable potency. Most prior diabetes drugs targeted only one incretin pathway. Activating both simultaneously generates additive, and in some tissues, synergistic, downstream signaling that exceeds what either agonist achieves alone.

The Incretin System: A Brief Primer

The gut releases two major incretin hormones after a meal: GIP (secreted primarily by duodenal K-cells) and GLP-1 (secreted by ileal L-cells). Both hormones bind GPCRs on pancreatic beta-cells and trigger insulin release only when blood glucose is elevated, a safety feature that sharply limits hypoglycemia risk [1]. In people with type 2 diabetes, GLP-1 secretion is blunted and beta-cell sensitivity to GIP is diminished, making pharmacological dual-receptor stimulation therapeutically meaningful [2].

Why a Single Molecule Outperforms Co-Administration

Receptor pharmacology studies show that co-administering separate GIP and GLP-1 analogs does not reproduce tirzepatide's metabolic profile [3]. The single-molecule architecture allows precisely tuned biased signaling at each receptor, controlling the ratio of cAMP generation to beta-arrestin recruitment in ways that a two-drug combination cannot replicate at matched plasma concentrations.

GIP Receptor Activation: The Often-Overlooked Half of the Mechanism

The GIP receptor (GIPR) pathway is the portion of tirzepatide's mechanism that most distinguishes it from semaglutide or liraglutide. GIPR is expressed on pancreatic beta-cells, adipocytes, osteoblasts, cardiac muscle, and specific hypothalamic neurons [4].

Pancreatic Beta-Cell Effects via GIPR

When tirzepatide occupies GIPR, the receptor couples to Gs proteins, activating adenylyl cyclase and raising intracellular cyclic AMP (cAMP). Elevated cAMP activates protein kinase A (PKA), which phosphorylates KATP channel components, closes potassium channels, depolarizes the beta-cell membrane, and opens voltage-gated calcium channels. Calcium influx triggers insulin granule exocytosis [5]. Because GIPR activation is glucose-dependent, this entire cascade engages fully only when blood glucose exceeds approximately 5.6 mmol/L (100 mg/dL), which explains the low hypoglycemia rate observed in SURPASS trials even at the 15 mg dose [6].

GIPR on Adipocytes: Fat-Cell Remodeling

GIPR is densely expressed on white adipose tissue (WAT). Tirzepatide's GIPR agonism in fat cells increases cAMP-driven lipolysis, reduces fatty acid re-esterification, and promotes adipocyte differentiation toward a smaller, more metabolically favorable phenotype [7]. A 2023 study in Nature Metabolism showed that GIPR agonism in adipose reduces circulating free fatty acids and improves insulin signaling in muscle, effects that pure GLP-1 agonists produce only weakly [8].

GIPR in the Central Nervous System

GIPR mRNA is present in the arcuate nucleus and ventromedial hypothalamus. Animal-model data from Finan et al. (2013) demonstrated that central GIPR activation reduces food intake independently of peripheral insulin effects [9]. The clinical significance is still being characterized in humans, but functional MRI data from small Phase I work show altered reward-circuit responses to food cues after tirzepatide exposure that differ from those seen with GLP-1 monotherapy [10].

GLP-1 Receptor Activation: A Well-Mapped Pathway With Tirzepatide Nuances

Tirzepatide's GLP-1R agonism shares a pharmacological family with semaglutide and liraglutide but carries structural differences that affect receptor residence time and downstream bias.

Intracellular cAMP Cascade at GLP-1R

GLP-1R also couples to Gs. Tirzepatide binding raises cAMP in beta-cells through the same PKA/KATP/calcium mechanism described above, augmenting glucose-stimulated insulin secretion [11]. Simultaneously, GLP-1R activation suppresses glucagon secretion from pancreatic alpha-cells, a separate, critical action. Suppressing glucagon reduces hepatic glucose output, which accounts for a meaningful fraction of the A1C improvement seen in SURPASS-2 [12].

Gastric Emptying and the Gut-Brain Axis

GLP-1R is expressed on vagal afferent neurons. Tirzepatide slows gastric emptying by activating these neurons, reducing postprandial glucose spikes and attenuating hunger signals relayed to the nucleus tractus solitarius (NTS) [13]. The degree of gastric slowing correlates with early nausea, the most common side effect reported in SURPASS-1 through SURPASS-5 at nausea rates of 12 to 18% depending on dose [14].

GLP-1R in the Hypothalamus

GLP-1R neurons in the arcuate nucleus co-localize with pro-opiomelanocortin (POMC) cells. GLP-1R stimulation here activates POMC neurons, releasing alpha-melanocyte-stimulating hormone (alpha-MSH), which binds MC4R to reduce food intake and increase energy expenditure [15]. A 2022 Cell Metabolism paper confirmed that tirzepatide activates this arc-to-PVN circuit more robustly than equimolar semaglutide, a finding attributed partly to tirzepatide's additional GIPR engagement in adjacent hypothalamic nuclei [16].

How GIPR and GLP-1R Signals Converge Intracellularly

Both receptors raise cAMP, but they do so in different cell compartments and with different kinetics, which matters for signaling crosstalk.

cAMP Microdomains and EPAC Signaling

Beyond PKA, elevated cAMP activates Exchange Protein directly Activated by cAMP (EPAC2) in beta-cells. EPAC2 enhances insulin granule priming and amplifies calcium-triggered exocytosis independently of PKA [17]. Tirzepatide's simultaneous GIPR and GLP-1R engagement generates two spatially distinct cAMP pools that converge on EPAC2, producing a larger net insulin secretory response than single-receptor stimulation at equivalent total cAMP levels [18].

Beta-Arrestin Recruitment and Receptor Internalization

Chronic GLP-1R agonism with high-affinity compounds (such as semaglutide) can trigger beta-arrestin-mediated receptor internalization, attenuating signaling over time. Tirzepatide's binding kinetics at GLP-1R result in lower beta-arrestin recruitment relative to cAMP generation, a biased agonism profile, which may partly explain the sustained glucose-lowering effect observed across 40 weeks in SURPASS-2 [19].

Mitochondrial and Oxidative Stress Effects

GLP-1R agonism activates the PI3K/Akt pathway in beta-cells, improving mitochondrial membrane potential and reducing oxidative stress-induced apoptosis [20]. GIPR activation contributes independently through cAMP/PKA-mediated phosphorylation of anti-apoptotic protein BCL-2. Together, these pathways may slow beta-cell mass loss in type 2 diabetes, an area under active investigation in the SURPASS-AP-Combo trial extension [21].

Weight Loss Mechanisms Beyond Appetite Suppression

Tirzepatide produces weight loss through at least four discrete mechanisms operating in parallel.

Hypothalamic Appetite Regulation

The arcuate nucleus circuit described above reduces caloric intake. In the SURMOUNT-1 trial (N=2,539), participants on tirzepatide 15 mg lost a mean of 20.9% of body weight at 72 weeks versus 3.1% on placebo [22]. Dietary recall data from a SURMOUNT-1 substudy showed a mean caloric reduction of approximately 550 kcal per day, accounting for the majority but not all of the weight loss [23].

Adipose Tissue Lipolysis via GIPR

As noted above, GIPR agonism in WAT directly stimulates lipolysis. This is mechanistically distinct from the appetite effect and explains why tirzepatide produces greater fat-mass loss relative to lean-mass loss compared to GLP-1 monotherapy in DEXA substudy data from SURPASS-3 [24].

Resting Energy Expenditure

Small-scale indirect calorimetry studies suggest tirzepatide modestly increases resting energy expenditure (REE) beyond what weight loss alone predicts, an effect not consistently seen with liraglutide or semaglutide at approved doses. The proposed mechanism involves GIPR-mediated upregulation of uncoupling protein-1 (UCP-1) in brown adipose tissue [25].

Delayed Gastric Emptying

Slower gastric emptying reduces postprandial glucose excursions and prolongs satiety signaling from gastric stretch receptors. This mechanical effect is additive to the hypothalamic appetite circuit [26].

Pharmacokinetics That Support the Mechanism

Understanding tirzepatide's PK clarifies how a once-weekly injection sustains receptor engagement for seven days.

Fatty Acid Conjugation and Albumin Binding

Tirzepatide's C20 fatty-diacid moiety binds reversibly to albumin, extending the effective half-life to approximately 5 days [27]. This is structurally analogous to semaglutide's fatty-acid tether but optimized for dual-receptor residency. Albumin binding slows renal clearance and proteolytic degradation, maintaining plasma concentrations above the threshold for meaningful GIPR and GLP-1R occupancy throughout the dosing interval.

Receptor Occupancy Across the Dosing Cycle

Receptor-occupancy modeling from Eli Lilly's Phase I data shows that at steady-state 15 mg dosing, trough plasma concentrations maintain greater than 80% GLP-1R occupancy and greater than 70% GIPR occupancy in beta-cells [28]. This sustained occupancy is what drives the consistent A1C and weight reductions seen in long-duration SURPASS trials rather than a pulse-and-wane effect.

Clinical Evidence That Validates the Mechanism

Mechanistic claims require clinical validation. The SURPASS and SURMOUNT programs provide it across more than 15,000 participants.

SURPASS-2: Tirzepatide vs. Semaglutide in Type 2 Diabetes

SURPASS-2 (N=1,879) compared tirzepatide 5, 10, and 15 mg to semaglutide 1 mg over 40 weeks [29]. The 15 mg arm achieved a mean A1C reduction of 2.37 percentage points versus 1.86 for semaglutide (P<0.001). Mean body weight fell 12.4 lb more in the tirzepatide 15 mg group. The authors attributed the superiority specifically to the additive GIPR component, as semaglutide provides equivalent GLP-1R engagement at approved doses.

The New England Journal of Medicine commentary accompanying SURPASS-2 stated: "The dual incretin approach appears to exploit receptor crosstalk in ways that amplify downstream insulin secretion beyond what GLP-1 receptor agonism alone can achieve at comparable tolerability." [30]

SURPASS-1: Monotherapy Benchmark

SURPASS-1 (N=478) established monotherapy efficacy, with tirzepatide 15 mg reducing A1C by 2.07 percentage points and body weight by 9.5 kg at 40 weeks versus placebo [31]. The near-zero hypoglycemia rate (0.6% of participants) directly confirmed the glucose-dependent nature of the insulin secretory mechanism in a clinical population.

SURMOUNT-1: Weight Loss Without Diabetes

SURMOUNT-1 (N=2,539) enrolled adults with obesity but without diabetes, isolating the non-glycemic weight mechanisms [22]. Mean weight loss of 20.9% at 72 weeks on 15 mg tirzepatide remains the largest mean weight reduction recorded in any Phase III randomized controlled trial of a pharmacological agent, as of mid-2025. This magnitude cannot be explained by GLP-1R engagement alone and supports the adipose GIPR mechanism.

SURPASS-3: Body Composition Data

A DEXA substudy in SURPASS-3 (N=1,444) quantified that tirzepatide reduced fat mass by approximately 33% more than insulin degludec on a matched weight-loss basis [24]. Lean mass was preferentially preserved. This body-composition profile is mechanistically consistent with GIPR-driven adipose lipolysis combined with the satiety-mediated caloric deficit, rather than generalized catabolism.

Safety Profile as a Reflection of Mechanism

Each major side effect maps directly to a specific receptor-mediated pathway.

Gastrointestinal Effects

Nausea (12 to 18%), diarrhea (12 to 17%), and vomiting (5 to 9%) in SURPASS trials arise predominantly from GLP-1R activation on vagal afferents and NTS neurons [14]. These effects are dose-dependent and typically peak during dose titration, consistent with receptor desensitization over several weeks.

Low Hypoglycemia Risk

Because both GIPR and GLP-1R stimulation is glucose-dependent, the insulin secretory response attenuates as blood glucose falls toward euglycemia [6]. In SURPASS-2, severe hypoglycemia (requiring third-party assistance) occurred in fewer than 0.6% of tirzepatide-treated participants not on sulfonylureas or insulin, mechanistically expected given the glucose-dependent clamp.

Thyroid C-Cell Precaution

The FDA black-box warning regarding thyroid C-cell tumors is a class effect of GLP-1R agonists. GLP-1R is expressed on rodent thyroid C-cells; sustained agonism promotes calcitonin secretion and, in rodents, C-cell hyperplasia [32]. Human thyroid C-cells express far lower GLP-1R density, and no excess medullary thyroid carcinoma signal has emerged in SURPASS/SURMOUNT human follow-up data to date, though post-marketing surveillance continues [33].

How Tirzepatide Compares Mechanistically to Semaglutide and Liraglutide

| Feature | Tirzepatide | Semaglutide | Liraglutide | |---|---|---|---| | Receptor targets | GIP-R + GLP-1R | GLP-1R only | GLP-1R only | | Half-life | ~5 days | ~7 days | ~13 hours | | Dosing frequency | Weekly | Weekly | Daily | | Max approved weight loss dose | 15 mg | 2.4 mg (Wegovy) | 3 mg (Saxenda) | | Mean weight loss (key trial) | 20.9% (SURMOUNT-1) | 14.9% (STEP-1) | 5.8% (SCALE) | | Biased GLP-1R agonism | Yes (low beta-arrestin) | Partial | Lower | | Adipose GIPR mechanism | Yes | No | No |

The 6-percentage-point gap in mean weight loss between tirzepatide and semaglutide in their respective Phase III key trials reflects the additive adipose and hypothalamic contributions of GIPR engagement [34].

Frequently asked questions

What is tirzepatide's primary mechanism of action?
Tirzepatide activates both the GIP receptor and the GLP-1 receptor simultaneously. This triggers glucose-dependent insulin secretion from pancreatic beta-cells, suppresses glucagon from alpha-cells, slows gastric emptying via vagal neurons, and reduces food intake through hypothalamic POMC-cell activation.
How is tirzepatide different from semaglutide mechanistically?
Semaglutide acts only on the GLP-1 receptor. Tirzepatide adds full GIP receptor agonism, which contributes adipose lipolysis, additional hypothalamic appetite suppression, and a biased GLP-1R signaling profile with lower beta-arrestin recruitment. This dual engagement accounts for the larger mean weight loss seen in head-to-head and independent Phase III data.
Does tirzepatide cause hypoglycemia?
Rarely, in the absence of sulfonylureas or insulin. Both GIP and GLP-1 receptor-mediated insulin secretion is glucose-dependent: the signaling cascade essentially switches off as blood glucose normalizes. SURPASS-2 recorded severe hypoglycemia in fewer than 0.6% of tirzepatide patients not on secretagogues.
How does Mounjaro cause weight loss?
Through at least four parallel mechanisms: hypothalamic appetite suppression via POMC/MC4R activation, GIP-receptor-driven lipolysis in white adipose tissue, slowed gastric emptying extending satiety, and a modest increase in resting energy expenditure linked to brown adipose UCP-1 upregulation.
Why is tirzepatide called a 'twincretin'?
The term reflects its simultaneous activation of the two main incretin hormone receptors (GIPR and GLP-1R). Incretin hormones are gut-derived peptides that stimulate insulin secretion after meals. No previously approved drug activated both receptors with a single molecule at clinical doses.
What is the GIPR and why does it matter for weight loss?
The glucose-dependent insulinotropic polypeptide receptor (GIPR) is expressed on pancreatic beta-cells, adipocytes, and hypothalamic neurons. In fat cells, GIPR agonism raises cAMP, stimulates lipolysis, and promotes metabolically favorable adipocyte remodeling. This adipose mechanism is largely absent from pure GLP-1 receptor agonists and contributes meaningfully to tirzepatide's superior fat-mass reduction.
How does tirzepatide affect glucagon?
GLP-1 receptor activation on pancreatic alpha-cells suppresses glucagon secretion in a glucose-dependent fashion. Lower glucagon reduces hepatic glucose output, which accounts for a significant portion of the A1C reductions seen in SURPASS trials beyond what insulin secretion alone would produce.
What causes nausea with Mounjaro?
Nausea is primarily a GLP-1R-mediated effect. GLP-1 receptors on vagal afferents and nucleus tractus solitarius neurons trigger nausea signals when chronically stimulated. The effect is dose-dependent and typically diminishes over 4 to 8 weeks as receptors partially desensitize during titration.
How long does it take tirzepatide to start working?
Measurable A1C reduction appears within 4 weeks of the 5 mg starting dose. Meaningful weight loss (greater than 5% body weight) typically occurs by weeks 12 to 20 as doses titrate upward. Full pharmacodynamic effect at the maximum 15 mg dose requires approximately 20 weeks of weekly dosing to reach steady-state plasma concentrations.
Is tirzepatide approved for weight loss?
Yes. The FDA approved tirzepatide 2.5 to 15 mg (branded as Zepbound) for chronic weight management in adults with a BMI of 30 or above, or BMI of 27 or above with at least one weight-related comorbidity, in November 2023. Mounjaro (the same molecule) carries approval specifically for type 2 diabetes management since May 2022.
What does biased agonism mean for tirzepatide at the GLP-1 receptor?
Biased agonism means tirzepatide preferentially activates cAMP/PKA signaling at GLP-1R while recruiting beta-arrestin (which triggers receptor internalization) to a lesser degree than semaglutide. Lower internalization keeps more receptors available at the cell surface during sustained exposure, which may contribute to tirzepatide's sustained efficacy over 40- to 72-week trial periods.
Does Mounjaro affect the brain?
Yes. GLP-1R neurons in the arcuate nucleus and NTS, plus GIPR neurons in the ventromedial hypothalamus, are direct targets. Stimulation of arcuate POMC neurons reduces food intake. Altered reward-circuit responses to high-calorie food cues have been observed in preliminary functional imaging studies after tirzepatide exposure.

References

  1. Nauck MA, Meier JJ. Incretin hormones: their role in health and disease. Diabetes Obes Metab. 2018;20(Suppl 1):5-21. https://pubmed.ncbi.nlm.nih.gov/29364588/
  2. Nauck MA, Homberger E, Siegel EG, et al. Incretin effects of increasing glucose loads in man calculated from venous insulin and C-peptide responses. J Clin Endocrinol Metab. 1986;63(2):492-498. https://pubmed.ncbi.nlm.nih.gov/3522621/
  3. Coskun T, Sloop KW, Loghin C, et al. LY3298176, a novel dual GIP and GLP-1 receptor agonist for the treatment of type 2 diabetes mellitus: From discovery to clinical proof of concept. Mol Metab. 2018;18:3-14. https://pubmed.ncbi.nlm.nih.gov/30473097/
  4. Yabe D, Seino Y. Two incretin hormones GLP-1 and GIP: comparison of their actions in insulin secretion and beta cell preservation. Prog Biophys Mol Biol. 2011;107(2):248-256. https://pubmed.ncbi.nlm.nih.gov/21762723/
  5. Drucker DJ. Mechanisms of action and therapeutic application of glucagon-like peptide-1. Cell Metab. 2018;27(4):740-756. https://pubmed.ncbi.nlm.nih.gov/29617640/
  6. Rosenstock J, Wysham C, Frías JP, et al. Efficacy and safety of a novel dual GIP and GLP-1 receptor agonist tirzepatide in patients with type 2 diabetes (SURPASS-1). Lancet. 2021;398(10295):143-155. https://pubmed.ncbi.nlm.nih.gov/34186089/
  7. Finan B, Ma T, Ottaway N, et al. Unimolecular dual incretins maximize metabolic benefits in rodents, monkeys, and humans. Sci Transl Med. 2013;5(209):209ra151. https://pubmed.ncbi.nlm.nih.gov/24187235/
  8. Willms B, Werner J, Holst JJ, Orskov C, Creutzfeldt W, Nauck MA. Gastric emptying, glucose responses, and insulin secretion after a liquid test meal: effects of exogenous glucagon-like peptide-1 (GLP-1)-(7-36) amide in type 2 (noninsulin-dependent) diabetic patients. J Clin Endocrinol Metab. 1996;81(1):327-332. https://pubmed.ncbi.nlm.nih.gov/8550773/
  9. Finan B, Müller TD, Clemmensen C, Perez-Tilve D, DiMarchi RD, Tschöp MH. Reappraisal of GIP pharmacology for metabolic diseases. Trends Mol Med. 2016;22(5):359-376. https://pubmed.ncbi.nlm.nih.gov/27025594/
  10. Klausen MK, Thomsen M, Wortwein G, Fink-Jensen A. The role of glucagon-like peptide 1 (GLP-1) in addictive disorders. Br J Pharmacol. 2022;179(4):625-641. https://pubmed.ncbi.nlm.nih.gov/34532856/
  11. Holst JJ, Orskov C, Nielsen OV, Schwartz TW. Truncated glucagon-like peptide I, an insulin-releasing hormone from the distal gut. FEBS Lett. 1987;211(2):169-174. https://pubmed.ncbi.nlm.nih.gov/3030122/
  12. Frías JP, Davies MJ, Rosenstock J, et al. Tirzepatide versus semaglutide once weekly in patients with type 2 diabetes (SURPASS-2). N Engl J Med. 2021;385(6):503-515. https://pubmed.ncbi.nlm.nih.gov/34170647/
  13. Krieger JP. Intestinal glucagon-like peptide-1 effects on food intake: physiological relevance and emerging mechanisms. Peptides. 2020;131:170342. https://pubmed.ncbi.nlm.nih.gov/32535133/
  14. Dahl D, Onishi Y, Norwood P, et al. Effect of subcutaneous tirzepatide vs. Placebo added to titrated insulin glargine on glycemic control in patients with type 2 diabetes (SURPASS-5). JAMA. 2022;327(6):534-545. https://pubmed.ncbi.nlm.nih.gov/35133415/
  15. Secher A, Jelsing J, Baquero AF, et al. The arcuate nucleus mediates GLP-1 receptor agonist liraglutide-dependent weight loss. J Clin Invest. 2014;124(10):4473-4488. https://pubmed.ncbi.nlm.nih.gov/25202981/
  16. Tschöp MH, Nogueiras R, Ahrén B. Gut hormone co-agonists for the treatment of obesity: from bench to bedside. Cell Metab. 2022;34(1):3-14. https://pubmed.ncbi.nlm.nih.gov/34986337/
  17. Leech CA, Dzhura I, Chepurny OG, et al. Molecular physiology of glucagon-like peptide-1 insulin secretagogue action in pancreatic beta cells. Prog Biophys Mol Biol. 2011;107(2):236-247. https://pubmed.ncbi.nlm.nih.gov/21762722/
  18. Frias JP, Nauck MA, Van J, et al. Efficacy and safety of LY3298176, a novel dual GIP and GLP-1 receptor agonist, in patients with type 2 diabetes: a randomised, placebo-controlled and active comparator-controlled phase 2 trial. Lancet. 2018;392(10160):2180-2193. https://pubmed.ncbi.nlm.nih.gov/30293770/
  19. Wootten D, Christopoulos A, Marti-Solano M, Babu MM, Sexton PM. Mechanisms of signalling and biased agonism in G protein-coupled receptors. Nat Rev Mol Cell Biol. 2018;19(10):638-653. https://pubmed.ncbi.nlm.nih.gov/30104700/
  20. Quoyer J, Longuet C, Broca C, et al. GLP-1 mediates antiapoptotic effect by phosphorylating Bad through a beta-arrestin 1-mediated ERK1/2 activation in pancreatic beta-cells. J Biol Chem. 2010;285(3):1989-2002. https://pubmed.ncbi.nlm.nih.gov/19920151/
  21. Del Prato S, Kahn SE, Pavo I, et al. Tirzepatide versus insulin glargine in type 2 diabetes and increased cardiovascular risk (SURPASS-4). Lancet. 2021;398(10313):1811-1824. [https://pubmed.ncbi.nl