• Background: Tirzepatide (brand name Mounjaro) is a once-weekly dual GIP and GLP-1 receptor agonist, approved in 2022 for type 2 diabetes and since developed for obesity and cardiometabolic disease. This review summarises its efficacy and safety.

• Methods: Scopus and PubMed were searched on 2026-09-25. The search returned 804 records; 749 remained after deduplication, 669 after a journal-quality filter, and 254 met an automated eligibility rule. The 50 most-cited eligible records (2021 to 2025) were synthesised narratively by theme, using data extracted automatically from abstracts.

• Results: In SURPASS trials in type 2 diabetes, HbA1c fell by 1.87 to 2.8 percentage points across dose arms. Tirzepatide beat semaglutide 1 mg (HbA1c differences of 0.15 to 0.45 points; 1.9 to 5.5 kg more weight loss) and insulin glargine (2.43 and 2.58 versus 1.44 points), with less hypoglycaemia than insulin (6 to 9% versus 19% with glargine). In SURMOUNT-3, adults without diabetes lost a further 18.4% of body weight, versus a 2.5% gain with placebo. SURPASS-CVOT showed non-inferiority to dulaglutide for major cardiovascular events (hazard ratio 0.92; 95.3% CI 0.83 to 1.01; superiority P = 0.09). One trial in obesity-related heart failure with preserved ejection fraction reported hazard ratios of 0.41 to 0.67 for heart failure composites. Kidney benefit (hazard ratio 0.58) came from a post hoc analysis; hepatic evidence was limited to ALT changes. Gastrointestinal events were the most common adverse events, mostly mild to moderate. Gallbladder or biliary disease was more frequent (relative risk 1.97; 95% CI 1.14 to 3.42); pancreatitis was not significantly increased (relative risk 1.46; 95% CI 0.59 to 3.61).

• Conclusion: Within the retrieved literature, tirzepatide lowered HbA1c and body weight more than placebo, semaglutide, dulaglutide and insulin. Limitations are substantial: only two databases were searched, data came from abstracts alone, no risk-of-bias or GRADE assessment was made, no meta-analysis was done, and citation-ranked selection favoured highly cited older papers. Findings need confirmation against full trial reports.

Obesity and type 2 diabetes mellitus (T2DM) are frequently described as twin epidemics, with a prevalence that is rising worldwide and is especially high in developed countries ^[1]. In the United States, obesity affects approximately 42% of adults and is associated with higher rates of T2DM, hypertension, cardiovascular disease, sleep disorders, osteoarthritis and premature death ^[2]. Among men with a body mass index (BMI) of 30 to 39, cardiovascular event rates are 20.21 per 1000 person-years, compared with 13.72 in men with a normal BMI ^[2]. Obesity is also a common comorbidity of T2DM, and even modest weight loss can improve glucose homeostasis and reduce cardiometabolic risk factors in these patients ^[3]. The kidney belongs to the same risk cluster: blood pressure, lipids, glycated haemoglobin and albuminuria are now considered together as cardiorenal risk factors ^[4].

Lifestyle intervention remains the base of obesity care, but its durability is limited. Multicomponent behavioural programmes often produce 5% to 10% weight loss, yet weight regain occurs in 25% or more of participants at two years ^[2]. Weight loss of this size may lower haemoglobin A1c by 0.6% to 1% in people with T2DM ^[2]. The gap between what lifestyle change delivers and what patients need has driven the development of pharmacotherapy. Of the six medications approved by the US Food and Drug Administration (FDA) for long-term weight management, two are glucagon-like peptide-1 (GLP-1) receptor agonists (semaglutide and liraglutide), and tirzepatide is reported to have the greatest effect, with mean weight loss of 21% at 72 weeks, within a range of approximately 8% to 21% reported for GLP-1 receptor agonists and dual agonists and 25% to 30% for bariatric surgery ^[2].

GLP-1 receptor agonists were first developed for T2DM, where their glycaemic effects and, with long-acting agents, their weight loss were evident ^[5]. Cardiovascular safety trials in patients with T2DM, most of whom had cardiovascular disease and excess weight, then showed that the class can lower cardiovascular risk ^[5]. Liraglutide and semaglutide subsequently led to obesity-specific programmes, the SCALE programme at 3 mg daily for liraglutide and the STEP programme at 2.4 mg weekly for semaglutide ^[5]. GLP-1 based therapy is now recommended early in the T2DM treatment algorithm because of its glycaemic efficacy, weight reduction and favourable cardiovascular outcomes ^[6]. Use has grown accordingly: in one academic health system, dispensing of GLP-1 receptor agonists rose rapidly between 2014 and 2022 ^[7].

The rationale for adding glucose-dependent insulinotropic polypeptide (GIP) followed a reversal in thinking. GIP is the main incretin hormone in healthy people, but its insulin response is strongly reduced in T2DM, and it was long regarded as an unappealing therapeutic target ^[8]. Resistance to GIP was later reported to be reversible with improved glycaemic control ^[8], and co-infusion of GLP-1 and GIP produced a synergistic increase in insulin and glucagonostatic responses compared with either hormone alone ^[6]. These observations led to a dual GIP/GLP-1 receptor agonist, or twincretin ^[6].

Tirzepatide (brand name Mounjaro) is that molecule. It is a synthetic peptide of 39 amino acids based on the native GIP sequence, given as a once-weekly subcutaneous injection ^[6,8]. Its half-life is about five days ^[1]. It was the first dual GIP/GLP-1 receptor co-agonist approved for T2DM, initially in the United States, Europe and the United Arab Emirates ^[9]. The first approval came in the United States in May 2022, as an adjunct to diet and exercise to improve glycaemic control in adults with T2DM ^[10,11]. Development extended to obesity, cardiovascular disorders in T2DM, heart failure, non-alcoholic steatohepatitis and obstructive sleep apnoea ^[10], and tirzepatide is now counted among the FDA-approved agents for long-term weight management ^[2]. The SURMOUNT programme of four phase 3 trials against placebo, in adults with a BMI of 27 kg/m² or more, was designed to define its role in chronic weight management ^[12].

The comparative question is what distinguishes tirzepatide from selective GLP-1 receptor agonists and from the insulin regimens that follow them in treatment algorithms. The SURPASS phase 3 programme addressed it in T2DM against semaglutide 1 mg ^[13], dulaglutide ^[14], insulin glargine ^[15], titrated insulin degludec ^[16] and prandial insulin lispro added to basal insulin ^[17]. In the head-to-head comparison with semaglutide (SURPASS-2, 1879 patients, 40 weeks), HbA1c fell by 2.01, 2.24 and 2.30 percentage points with tirzepatide 5, 10 and 15 mg and by 1.86 with semaglutide; body weight differences favoured tirzepatide by 1.9, 3.6 and 5.5 kg ^[13]. Against insulin glargine at 52 weeks, HbA1c fell by 2.43% (10 mg) and 2.58% (15 mg) versus 1.44%, with less hypoglycaemia ^[15]. Across SURPASS-1 to -5, a review reported HbA1c reductions of 1.24 to 2.58% and weight reductions of 5.4 to 11.7 kg ^[9]. In obesity, a clinical review reported placebo-corrected weight loss of about 5% with liraglutide, 12% with semaglutide and 18% with tirzepatide ^[18]. A network meta-analysis of weight-loss trials placed tirzepatide 15 mg (a 16.53% reduction in body weight) among the three most efficacious treatments, although it was less effective in patients with T2DM than in those without ^[19].

Efficacy is only one side of the assessment. In a meta-analysis, lean mass accounted for about a quarter of the weight lost with GLP-1 receptor agonists, and tirzepatide 15 mg and semaglutide 2.4 mg were among the least effective agents at preserving it ^[20]. In a meta-analysis of nine trials, tirzepatide was not significantly associated with pancreatitis (relative risk 1.46; 95% CI 0.59 to 3.61) but was associated with a higher risk of a composite of gallbladder or biliary disease (relative risk 1.97; 95% CI 1.14 to 3.42) ^[21]. Cardiovascular outcome evidence was slower to arrive. In SURPASS-4, adjudicated major adverse cardiovascular events were not increased against glargine (hazard ratio 0.74; 95% CI 0.51 to 1.08) ^[15], and in SURPASS-CVOT, which analysed 13,165 patients with T2DM and atherosclerotic cardiovascular disease, tirzepatide was noninferior to dulaglutide (hazard ratio 0.92; 95.3% CI 0.83 to 1.01) ^[22]. Kidney and cardiac endpoints widen the scope further: a SURPASS-4 post hoc analysis reported a lower composite kidney outcome rate than with glargine (hazard ratio 0.58) ^[23], and in patients with heart failure with preserved ejection fraction and obesity, tirzepatide improved health status and functional capacity ^[24] and reduced left ventricular mass and paracardiac adipose tissue ^[25].

The evidence on tirzepatide has grown quickly and spans diabetes, obesity, cardiovascular and kidney research, so a systematic, quality-filtered synthesis is needed. This review asks what the efficacy and safety of tirzepatide are across glycaemic, weight, cardiovascular, hepatic and adverse-event outcomes. It summarises glycaemic efficacy in T2DM against placebo, selective GLP-1 receptor agonists and insulin, characterises weight and body-composition outcomes, evaluates cardiovascular, cardiorenal and hepatic outcomes, and assesses the adverse-event profile.

The review is organised as follows. The Methods section describes the multi-database search, journal-quality filtering, DOI validation and selection process. The results are then presented by domain: the mechanism of action and pharmacology of tirzepatide, glycaemic efficacy and its place in T2DM management, obesity and weight management, cardiovascular, hepatic and other organ outcomes, and safety and tolerability. The Discussion places these findings alongside existing GLP-1 receptor agonist therapy, sets out the limitations of the evidence and identifies priorities for future research.

The overall aim of this review is to provide a narrative synthesis of the tirzepatide literature identified through a Scopus and PubMed search. The specific objectives are to address five review questions:

1. What is the mechanism of action and pharmacology of tirzepatide as a dual GIP/GLP-1 receptor agonist?
2. What is its glycaemic efficacy in adults with type 2 diabetes?
3. What is its efficacy for weight loss in people with obesity?
4. What are its effects on cardiometabolic, renal and hepatic outcomes?
5. What is its safety profile, including adverse events?

This review was produced with an automated literature pipeline rather than by a team of human reviewers. It follows the reporting structure of PRISMA 2020, but several PRISMA items were not carried out, and each is stated as such below. No protocol was written or registered (for example, in PROSPERO).

Two bibliographic databases were searched on 2026-09-25: Scopus, through the Elsevier Search API, and PubMed, through the NCBI E-utilities. Embase was not searched. There was no hand search of reference lists, no search of trial registries or conference abstracts, and no grey literature search.

The same query was used in both databases: "tirzepatide" OR "Mounjaro" OR "Zepbound". The search is based on title, abstract and keyword fields. In Scopus this was TITLE-ABS-KEY; in PubMed the default automatic term mapping was used. The search strategy consisted of free-text keyword search terms combined with the Boolean operator OR. No controlled-vocabulary subject headings (MeSH or Emtree terms) were hand-built, and no validated search filters were applied; the only limit was the publication year described below. No additional synonyms for GIP/GLP-1 dual agonists were included. No language restriction was applied in the search.

A single query with a record cap returns only the newest papers. To avoid this, the query was run separately for each publication year from 2021 to 2026, retrieving up to 70 records per database per year. The year 2026 is a partial year. The per-year cap means that a year with more than 70 matching records in one database was not retrieved completely.

Selection was performed by software in five sequential steps.

1. Deduplication. Records were merged when their DOI or normalised title matched.
2. Journal-quality filter. A record was kept if its journal had a CiteScore of at least 3.0 (Scopus serial-title API) or an SJR value above the set threshold. Records whose journal had no available metric (about 164 of the 749 deduplicated records, mostly in PubMed-only journals) were retained rather than excluded. These metrics describe the journal, not the quality of the individual study.
3. DOI validation. Each DOI was checked against the doi.org handle API; all 669 records resolved (100%).
4. Eligibility rule. A record was eligible if its abstract was at least 100 characters long and it either named tirzepatide, Mounjaro, Zepbound, SURPASS or SURMOUNT in the title or mentioned these terms at least twice in the abstract. This tests topical relevance only; it does not classify study design, population or outcome.
5. Final selection of 50 records. The eligible records were sorted into nine topic categories, which were generated for this topic from a sample of the retrieved abstracts. Within each category, the two most-cited eligible records were guaranteed inclusion, and the remaining slots were filled by citation count (Scopus "cited by" counts).

Because the last step ranks by citations, the selection favours older, highly cited papers. The included records span 2021 to 2025, and the newest 2026 literature is under-represented. The included set should not be read as a random or representative sample of the eligible records.

The flow of records was as follows.

1. Records identified: 804 (Scopus and PubMed combined, all years).
2. Duplicates removed: 55. Records remaining: 749.
3. Removed by journal-quality filtering: 80. Records remaining: 669.
4. DOIs validated via doi.org: 669 of 669 resolved.
5. Not meeting the eligibility rule: 415. Eligible records: 254.
6. Eligible records not chosen by the citation-ranked selection: 204.
7. Records included in the review: 50, drawn from 34 journals.

The exclusion counts in lines 2, 3, 5 and 6 are differences between stage totals reported by the pipeline; the reasons are the rules described above.

Data collection was automated and done in a single pass, from abstracts only. Each article was extracted into a standardized, structured record. No manual data extraction form was used, and study authors were not contacted to obtain or confirm data. Regular expressions were used to pull sample sizes, p-values, thresholds and doses from the abstract text. Full texts were not read. There was no second extractor and no arbitration process. The data items were therefore limited to what abstracts state and regular expressions can capture: sample size, p-values, thresholds and doses, together with the outcomes and effects as worded in the abstract. Extraction errors are possible and were not systematically checked.

Outcomes sought (data items, PRISMA item 10a) were the change in HbA1c, the change in body weight, cardiovascular, renal and hepatic outcomes, and adverse events and treatment discontinuation. Other variables sought (item 10b) were study design, sample size, dose, comparator, follow-up duration and funder, each as stated in the abstract. Any of these items that an abstract did not report could not be captured.

No study-level risk-of-bias tool was applied. RoB 2, ROBINS-I, the Newcastle-Ottawa Scale and AMSTAR-2 were not used. The only quality control was the journal-level filter described above. The certainty of the evidence was not graded, and no GRADE assessment was made. Reporting (publication) bias was not assessed: there was no funnel plot or Egger test. In PRISMA terms, the risk of bias due to missing results in the synthesis (reporting bias) was not assessed. Citation-ranked selection is a possible source of such bias. Because selection favours highly cited papers, publication and citation bias may be present and would tend to favour positive, high-profile findings.

To prepare the data for synthesis, the abstract text was standardized: units and symbols were normalized, and inline markup in PubMed abstracts (for example, subscripts in HbA1c) was restored. Results are displayed in the text, in one comparison table of the SURPASS trials, and in a PRISMA flow figure. The synthesis is a narrative, thematic one, organised by the nine topic categories (domains). No meta-analysis was performed, because study designs, comparators, populations and outcome definitions differ across the included records. Heterogeneity was handled by grouping records into thematic sections (subgroups by domain) and not statistically, and no other method for exploring heterogeneity was used; no heterogeneity statistics were calculated. Effect measures are reported as stated in the source abstracts, with no recalculation or conversion to a common scale.

• Only two databases were searched. Without Embase, hand searching and grey literature, relevant studies may be missing.
• Per-year caps of 70 records per database may have truncated the retrieved set in the busiest years.
• Records with no journal metric were retained, so the journal filter is only a partial screen of quality, and it says nothing about individual study quality.
• Eligibility was decided by a term-frequency rule and citation ranking. Relevant studies with few mentions of the drug name, or recent studies with few citations, could be missed.
• Data were taken from abstracts only, in one automated pass, without verification by a second person.
• No protocol was registered, and no risk-of-bias or certainty assessment was done.
• The search was not restricted by language, but the included records all had English abstracts, so studies published only in other languages are not reflected.

These limitations mean the review is best read as a structured, transparent overview of highly cited tirzepatide literature, and not as a full systematic review with formal appraisal of the evidence.

Tirzepatide is a single synthetic peptide of 39 amino acids, built on the native GIP sequence, that activates both the glucose-dependent insulinotropic polypeptide (GIP) receptor and the glucagon-like peptide-1 (GLP-1) receptor ^[6,26]. GIP and GLP-1 are the two principal incretin hormones, released from the intestine in response to nutrients; both stimulate insulin secretion by pancreatic beta cells, and GLP-1 additionally suppresses glucagon, delays gastric emptying and reduces food intake ^[27,10,8]. GIP was long considered an unattractive target because its insulinotropic effect is strongly reduced in type 2 diabetes, but that resistance has been reported to be reversible with improved glycaemic control, and co-infusion of GLP-1 and GIP produces a greater insulin and glucagonostatic response than either hormone alone ^[8,6], hence the term "twincretin" ^[1].

Cryo-electron microscopy structures of tirzepatide bound to GIPR and GLP-1R describe the interactions underlying its dual agonism ^[28]. In cell-based assays, tirzepatide preserved maximal cAMP production despite partial Gαs recruitment and reduced ligand-induced receptor internalisation, which the authors interpreted as biased agonism that alters receptor trafficking relative to native peptides and semaglutide ^[29]. These structural and in vitro observations show how the molecule engages its targets; they do not by themselves establish the clinical mechanism.

Tirzepatide lowered HbA1c more than the selective GLP-1 receptor agonist semaglutide 1 mg in SURPASS-2 (-2.01, -2.24 and -2.30 percentage points with 5, 10 and 15 mg versus -1.86) ^[13], and across SURPASS-1 to -5 it reduced body weight by 5.4 to 11.7 kg ^[9]. Which receptor produces which part of this advantage is only partly resolved. Human data show that tirzepatide improved insulin sensitivity and insulin secretory responses more than semaglutide, with lower prandial insulin and glucagon concentrations, and that the two drugs reduced appetite to a similar degree even though tirzepatide produced greater weight loss ^[9]. The same review noted that the food-intake and weight effects of GIP seen in rodents have not been demonstrated in humans ^[9].

A post hoc analysis in 316 participants with type 2 diabetes (tirzepatide 1, 5, 10 or 15 mg, dulaglutide 1.5 mg, or placebo) explored these mechanisms in humans. HOMA2-B increased with tirzepatide, proinsulin/insulin and proinsulin/C-peptide ratios fell with 10 and 15 mg relative to placebo and dulaglutide, and the insulin-sensitivity markers adiponectin, IGFBP-1 and IGFBP-2 rose ^[30]. Weight loss explained only 13% and 21% of the HOMA2-IR improvement with the 10 and 15 mg doses, which the authors took to indicate that dual receptor agonism provides mechanisms of glycaemic control distinct from weight reduction ^[30].

Preclinical work points to GIPR as one contributor. In obese mice, tirzepatide improved insulin sensitivity more than GLP-1R agonism, and in Glp-1r-null mice, where GLP-1R-driven weight loss is absent, it still enhanced glucose disposal in white adipose tissue; a long-acting GIPR agonist did the same ^[31]. In rodent food-choice experiments, tirzepatide suppressed total caloric intake and selectively suppressed lipid intake, while GIPR agonism alone did not alter food choice and the shift was absent in GLP-1R knockout mice ^[32]. These are rodent findings and may not transfer to people.

Tirzepatide carries a C20 fatty-diacid moiety that binds albumin, which supports once-weekly subcutaneous dosing with a half-life of about five days ^[1]. A phase 1 multiple-ascending-dose study in Japanese participants with type 2 diabetes confirmed a plasma half-life of approximately 5 days; after 8 weeks, fasting plasma glucose fell by 52.7, 69.1 and 68.9 mg/dL versus placebo with 5, 10 and 15 mg ^[33]. In a single 5 mg dose study of 45 participants, exposure was similar across renal function groups apart from a 25 to 29% AUC increase in moderate impairment, and the authors concluded that dose adjustment may not be required ^[34].

A review of oral contraceptive interactions describes a greater effect of tirzepatide on gastric emptying than typical GLP-1 receptor agonists, most marked after the first dose and followed by tachyphylaxis; it identified one tirzepatide study showing significantly reduced AUC, Cmax and time to maximum concentration of an oral hormonal contraceptive, whereas five studies of GLP-1 receptor agonists showed no significant effect ^[35].

The corpus also addresses incretin actions beyond glucose and weight. In a phase 2 post hoc analysis, tirzepatide 10 and 15 mg lowered YKL-40 and leptin relative to placebo and dulaglutide, and 15 mg lowered hsCRP relative to placebo, while IL-6 did not change significantly ^[36]. A review of cardiovascular mechanisms also proposes anti-inflammatory effects, reduced cell death and promotion of autophagy ^[37]. Both sources are hypothesis-generating.

For neurodegeneration, the evidence is preclinical. In a neuroblastoma cell line exposed to high glucose, tirzepatide activated the pAkt/CREB/BDNF pathway and counteracted hyperglycaemia-related neuronal insulin resistance and neurodegeneration ^[38]. In mice, tirzepatide did not cross the blood-brain barrier within 1 hour after intravenous injection and entered slowly over 6 hours, whereas albiglutide, dulaglutide and DA5-CH entered faster ^[39]. The only human evidence of possible central effects is self-reported: 153 current drinkers with obesity taking semaglutide or tirzepatide reported lower alcohol intake than before treatment and than controls ^[40].

Metabolic effects on amino acids offer another mechanistic thread. In obese mice and in humans with type 2 diabetes, tirzepatide improved insulin sensitivity alongside lower circulating branched-chain amino acids (BCAAs) and branched-chain keto acids, metabolites associated with insulin resistance and type 2 diabetes ^[31,41]. Stable-isotope tracer studies in diet-induced obese mice showed that tirzepatide stimulated BCAA/BCKA catabolism in brown adipose tissue, with increased breakdown products (glutamate, alanine, 3-hydroxyisobutyric acid) and TCA cycle intermediates, a profile resembling cold exposure ^[41]. This pathway was demonstrated in mice; the human finding is the reduction in circulating BCAAs.

Weight loss with potent incretin-based agents is not confined to fat. In a network meta-analysis of 22 randomised trials, lean mass loss accounted for roughly 25% of total weight loss with GLP-1 receptor agonists, and tirzepatide 15 mg was among the most effective for weight and fat mass reduction but among the least effective at preserving lean mass ^[20]. Dual and triple receptor agonists are being developed to act on carbohydrate, lipid and protein metabolism simultaneously ^[11], including the GIPR/GLP-1R/GCGR triagonist peptide 20 ^[28].

Tirzepatide is a once-weekly dual GIP and GLP-1 receptor agonist that received its first approval in the USA in May 2022 to improve glycaemic control in adults with type 2 diabetes, as an adjunct to diet and exercise ^[10,11]. GLP-1-based therapy was already recommended early in the T2DM treatment algorithm because of its glycaemic efficacy, weight reduction and favourable cardiovascular outcomes ^[6]. Two arguments place tirzepatide in that setting. Weight loss is itself a glycaemic lever: 5% to 10% weight loss may lower haemoglobin A1c (HbA1c) by 0.6% to 1% in people with type 2 diabetes ^[2]. In a systematic analysis of non-insulin glucose-lowering drugs, tirzepatide, liraglutide and semaglutide fell in the highest weight-loss stratum (more than 5% of initial weight), above dulaglutide and the SGLT2 inhibitors ^[3]. Early uptake reflected this interest: Mounjaro reached 1508 users in a University of California health-system cohort in 2022, with a monthly growth rate estimated at 254.3% ^[7].

The phase 3 SURPASS programme tested tirzepatide 5, 10 and 15 mg across a broad range of people with type 2 diabetes, as monotherapy or combination therapy ^[42]. Across SURPASS-1 to -5, HbA1c fell by 1.9% to 2.6%, body weight by 6.6% to 13.9%, and systolic blood pressure by 2.8 to 12.6 mmHg ^[43]. A review of the same five trials reported HbA1c reductions of 1.24% to 2.58% and weight reductions of 5.4 to 11.7 kg, with 23.0% to 62.4% of participants reaching an HbA1c below 5.7% ^[9]. Ranges differ between reviews because they pool different arms and time points, so the individual trial reports are compared below.

In SURPASS-1, adults with inadequately controlled type 2 diabetes on diet and exercise alone (mean baseline HbA1c 7.9%) received tirzepatide or placebo for 40 weeks. HbA1c decreased by 1.87%, 1.89% and 2.07% with 5, 10 and 15 mg versus +0.04% with placebo, giving estimated treatment differences of -1.91%, -1.93% and -2.11% (all p<0.0001) ^[44]. An HbA1c below 7.0% was reached by 87% to 92% of tirzepatide-treated participants versus 20% on placebo, and weight loss was dose-dependent at 7.0 to 9.5 kg. No clinically significant (<54 mg/dL) or severe hypoglycaemia was reported with tirzepatide ^[44].

SURPASS-2 randomised 1879 patients to tirzepatide or semaglutide 1 mg for 40 weeks, from a mean baseline HbA1c of 8.28%. HbA1c fell by 2.01%, 2.24% and 2.30% with tirzepatide 5, 10 and 15 mg and by 1.86% with semaglutide. The differences were -0.15 percentage points (95% CI -0.28 to -0.03; P = 0.02), -0.39 (95% CI -0.51 to -0.26; P<0.001) and -0.45 (95% CI -0.57 to -0.32; P<0.001), and tirzepatide was noninferior and superior at every dose. Weight reductions exceeded those with semaglutide by 1.9, 3.6 and 5.5 kg (P<0.001 for all) ^[13]. In the Japanese SURPASS J-mono trial, HbA1c decreased by 2.4%, 2.6% and 2.8% with tirzepatide 5, 10 and 15 mg versus 1.3% with dulaglutide 0.75 mg (differences -1.1, -1.3 and -1.5; all p<0.0001), with weight reductions of 5.8, 8.5 and 10.7 kg versus 0.5 kg ^[14].

Against basal or prandial insulin, tirzepatide lowered HbA1c more and caused less hypoglycaemia. In SURPASS-3, participants inadequately controlled on metformin with or without an SGLT2 inhibitor started from a mean HbA1c of 8.17%. At week 52 HbA1c fell by 1.93%, 2.20% and 2.37% with tirzepatide versus 1.34% with titrated insulin degludec. The estimated treatment differences ranged from -0.59% to -1.04% (p<0.0001 for all doses). An HbA1c below 7.0% was reached by 82% to 93% of tirzepatide-treated participants versus 61% on degludec. Body weight fell by 7.5 to 12.9 kg with tirzepatide and rose by 2.3 kg with degludec, and hypoglycaemia occurred in 1% to 2% of tirzepatide-treated participants versus 7% on degludec ^[16]. SURPASS-4 randomised 2002 participants (1995 received at least one dose of tirzepatide) with established cardiovascular disease or high cardiovascular risk. At 52 weeks HbA1c changed by -2.43% with 10 mg and -2.58% with 15 mg versus -1.44% with insulin glargine. The estimated differences were -0.99% (97.5% CI -1.13 to -0.86) and -1.14% (-1.28 to -1.00), and the 0.3% noninferiority margin was met for both doses. Hypoglycaemia occurred in 6% to 9% of tirzepatide-treated participants versus 19% with glargine, and adjudicated MACE-4 events were not increased (hazard ratio 0.74, 95% CI 0.51-1.08) ^[15]. In SURPASS-6, tirzepatide added to insulin glargine lowered HbA1c by 2.1% (pooled) versus 1.1% with prandial insulin lispro, a difference of -0.98% (95% CI -1.17% to -0.79%; P<.001), with mean weight change of -9.0 kg versus +3.2 kg. The HbA1c target of less than 7.0% was reached by 68% versus 36%, and hypoglycaemia occurred at 0.4 versus 4.4 events per patient-year ^[17]. The SURPASS J-combo trial, an add-on to a single oral agent in Japanese patients, was designed with safety as its primary endpoint. In that trial, mean HbA1c fell from 8.5% to 6.0% with 5 mg and from 8.6% to 5.6% with both 10 and 15 mg, with weight reductions of 3.8, 7.5 and 10.2 kg, irrespective of background oral medication ^[45].

Mechanistic work supports these effect sizes. In a post hoc analysis, tirzepatide improved markers of beta-cell function and insulin sensitivity more than dulaglutide, and weight loss explained only 13% and 21% of the improvement in insulin resistance at 10 and 15 mg ^[30]. This suggests that the glycaemic effect is not only secondary to weight loss. Preclinical data show weight-independent insulin sensitisation through GIP receptor agonism in obese mice ^[31]. Reviews also note that GIP responsiveness, which is reduced in type 2 diabetes, may be restored by better glycaemic control ^[8]. In SURPASS 1-4 participants adherent to treatment (N = 3188), higher tirzepatide doses, metformin background therapy and lower baseline HbA1c were associated with a weight reduction of at least 15%, and greater weight loss accompanied larger HbA1c improvements ^[46].

Glycaemic treatment in type 2 diabetes overlaps with cardiovascular and renal risk, and the SURPASS programme provides some data on this. In a post hoc analysis of SURPASS-4, the annual eGFR decline was -1.4 mL/min/1.73 m² with tirzepatide versus -3.6 with insulin glargine, and the composite kidney outcome was less frequent (hazard ratio 0.58, 95% CI 0.43 to 0.80) ^[23]. Systolic blood pressure reductions were mainly mediated by weight loss but had weight-independent components, notably in SURPASS-4 ^[43]. In SURPASS-CVOT, 13,299 patients with type 2 diabetes and atherosclerotic cardiovascular disease (mean HbA1c 8.4%) were randomised. Tirzepatide was noninferior to dulaglutide 1.5 mg for cardiovascular death, myocardial infarction or stroke (12.2% versus 13.1%; hazard ratio 0.92, 95.3% CI 0.83 to 1.01; P = 0.003 for noninferiority; P = 0.09 for superiority) ^[22].

Several limitations qualify these conclusions. The SURPASS trials whose abstracts state a funder were funded by Eli Lilly and Company. Primary endpoints were assessed at 40 to 52 weeks, although the SURPASS-4 abstract reports treatment for up to 104 weeks ^[15] and its kidney analysis a median treatment duration of 85 weeks ^[23]; the SURPASS-4 abstract gives HbA1c results only for the 10 and 15 mg doses. SURPASS-2, -3, -4 and -6 were open-label, and SURPASS J-mono and J-combo enrolled Japanese participants only, with J-combo having no comparator arm. HbA1c reductions generally increased with dose across the trials that reported all three doses, although in SURPASS J-combo the 10 mg and 15 mg arms ended at the same HbA1c ^[45]. The advantage over semaglutide in SURPASS-2 was small at 5 mg (-0.15 percentage points) and larger at 15 mg (-0.45), and tirzepatide's advantage over semaglutide was seen for both weight and HbA1c at every dose. Comparisons with insulin showed both a larger HbA1c reduction and fewer hypoglycaemic events (SURPASS-3, -4 and -6).

Obesity affects approximately 42% of adults in the United States, and multicomponent behavioral interventions typically produce 5% to 10% weight loss, with weight regain in 25% or more of participants at 2-year follow-up ^[2]. In the same review, glucagon-like peptide-1 (GLP-1) agonists and the dual glucose-dependent insulinotropic polypeptide (GIP)/GLP-1 agonist tirzepatide were credited with roughly 8% to 21% weight loss, and tirzepatide was described as having the greatest effect among approved long-term antiobesity medications, with a mean loss of 21% at 72 weeks ^[2]. The SURMOUNT programme evaluates once-weekly tirzepatide against placebo, as an adjunct to lifestyle intervention, in adults with a body mass index (BMI) of 27 kg/m² or higher with or without type 2 diabetes, with percentage change in body weight as the primary end point in every trial ^[12].

The most direct trial evidence here is SURMOUNT-3, a double-blind, placebo-controlled trial of 579 adults without diabetes who had already lost at least 5.0% of body weight during a 12-week intensive lifestyle intervention ^[47]. Participants received the maximum tolerated tirzepatide dose (10 or 15 mg) or placebo for 72 weeks. Additional mean weight change from randomization was -18.4% with tirzepatide versus +2.5% with placebo (estimated difference -20.8 percentage points; 95% CI -23.2 to -18.5; P < 0.001), and 87.5% versus 16.5% achieved a further reduction of at least 5%; gastrointestinal events were the most common adverse events and mostly mild to moderate ^[47]. Substantial additional loss is therefore possible even after a successful behavioral run-in.

Pooled and comparative estimates are consistent. A clinical review reported placebo-corrected weight loss of around 5% for liraglutide, 12% for semaglutide and 18% for tirzepatide ^[18]. A network meta-analysis of 27 randomized trials (15,584 patients) placed tirzepatide 15 mg at -16.53% body weight and -13.23 cm waist circumference against placebo, among the three most efficacious regimens examined, with greater loss in patients without type 2 diabetes, with higher BMI and with longer treatment ^[19]. In type 2 diabetes, tirzepatide, liraglutide and semaglutide fell in the highest weight-loss stratum (more than 5% of initial weight) ^[3]; across SURPASS-1 to -5, tirzepatide 5, 10 and 15 mg lowered body weight by 6.6% to 13.9% ^[43], and 20.7% to 68.4% of participants lost more than 10% of baseline weight ^[9].

Weight reduction rises with dose. In the placebo-controlled SURPASS-1 trial (n = 478), tirzepatide produced dose-dependent loss of 7.0 to 9.5 kg over 40 weeks ^[44]. SURPASS J-mono (n = 636) recorded 5.8 kg (7.8%), 8.5 kg (11.0%) and 10.7 kg (13.9%) with 5, 10 and 15 mg at 52 weeks, versus 0.5 kg (0.7%) with dulaglutide 0.75 mg ^[14], and the add-on trial SURPASS J-combo (n = 443) reported 5.1%, 10.1% and 13.2% reductions at 52 weeks ^[45]. Among treatment-adherent participants in SURPASS 1-4 (N = 3,188), a reduction of at least 15% was associated with higher tirzepatide dose, female sex, White or Asian race, younger age, background metformin, and lower baseline HbA1c, fasting glucose and non-HDL cholesterol; greater weight reduction accompanied larger improvements in glycaemia, triglycerides, ALT, waist circumference and blood pressure ^[46].

Against semaglutide 1 mg, the open-label SURPASS-2 trial (n = 1,879) showed greater weight reduction with tirzepatide 5, 10 and 15 mg at 40 weeks, with least-squares mean differences of -1.9, -3.6 and -5.5 kg (P < 0.001 for all) ^[13]. A review of the programme noted similar reductions in appetite with both drugs but greater weight loss with tirzepatide ^[9]. Liraglutide is compared only indirectly: a body-composition network meta-analysis of 22 trials (2,258 participants) found tirzepatide 15 mg and semaglutide 2.4 mg the most effective for weight and fat-mass reduction, whereas liraglutide (3.0 mg weekly or 1.8 mg daily) was the only GLP-1 agonist that reduced weight significantly without significantly reducing lean mass ^[20].

In a SURMOUNT-1 post hoc analysis of 2,461 participants without prior atherosclerotic cardiovascular disease, the baseline median 10-year predicted risk was low (1.5% to 1.6%); relative change to week 72 was -23.5% to -16.4% with tirzepatide versus +12.7% with placebo (P < 0.001), with greater absolute reductions in participants at intermediate-to-high baseline risk ^[48]. These are risk-engine predictions, not observed cardiovascular events.

Lean mass is a recognised concern. GLP-1 receptor agonists overall reduced total body weight by 3.55 kg, fat mass by 2.95 kg and lean mass by 0.86 kg, so lean mass accounted for approximately 25% of weight lost, although relative lean mass (percentage change from baseline) was unaffected; tirzepatide and semaglutide were among the least effective at preserving lean mass ^[20]. Gastrointestinal effects dominate tolerability: nausea, diarrhoea, constipation and vomiting predominate, gallbladder disorders and acute pancreatitis are rare serious events, and individualized dosing with careful patient assessment is advised ^[18]. Adverse events were dose-related: in SURPASS J-combo, treatment-emergent events occurred in 84% at 15 mg versus 74% at 5 and 10 mg ^[45], and in J-mono nausea rose from 12% at 5 mg to 20% at 10 and 15 mg ^[14]. Rodent work suggests that tirzepatide selectively lowers preference for fat-rich food, which may contribute to weight loss ^[32]. Obesity guidance places pharmacotherapy alongside behavioral, nutritional and physical-activity components of comprehensive care ^[2].

Two problems remain open. The first is durability. Lifestyle-based weight loss is not long-term effective in many patients ^[3], maintaining loss is difficult, and guidelines support long-term antiobesity medication when lifestyle measures alone fail to maintain weight ^[2]. The longest tirzepatide obesity follow-up summarized here is 72 weeks ^[47]. The second is adherence, which shapes how effect sizes are read: SURMOUNT-3 reports a treatment-regimen estimand regardless of adherence ^[47], whereas the SURPASS predictor analysis was restricted to participants who took at least 75% of doses and remained on treatment at week 40 or 42 ^[46]. Uptake is also changing quickly: in one University of California health system, Mounjaro reached 1,508 users in 2022 with a monthly growth rate of 254.3%, the highest of any product examined, and the authors called for longitudinal studies and evidence-concordant prescribing ^[7]. Narrative reviews present tirzepatide as one of a new generation of antiobesity agents whose place will depend on further data ^[49,5].

The cardiovascular evidence for tirzepatide has to be read against the class-level record for glucagon-like peptide-1 receptor agonists (GLP-1 RAs). Reviews of the class state that cardiovascular safety trials in people with type 2 diabetes, most of whom had established cardiovascular disease and excess weight, showed that GLP-1 RAs lowered cardiovascular risk ^[5], and that these agents have been shown to reduce major adverse cardiovascular events ^[37]. That class-level evidence, however, does not automatically transfer to a dual GIP/GLP-1 agonist, and the same reviews frame tirzepatide as a compound whose own outcome data were still pending: liraglutide and semaglutide programmes were the reference points, while tirzepatide was awaiting the SURPASS-CVOT trial ^[5]. Reviews of the compound's cardiovascular biology describe possible benefits beyond glycaemic control through anti-inflammatory effects, reduced cell death, promotion of autophagy and indirect effects on blood pressure, obesity and lipid metabolism, but these are mechanistic proposals rather than outcome data ^[37,26].

Tirzepatide-specific outcome evidence now comes from randomised trials. In SURPASS-4, an open-label trial of 2002 adults with type 2 diabetes and established cardiovascular disease or high cardiovascular risk, adjudicated four-component major adverse cardiovascular events (MACE-4) were not increased on tirzepatide compared with insulin glargine (hazard ratio 0.74, 95% CI 0.51-1.08), with 109 events in total, so the trial supports safety rather than benefit ^[15]. A meta-analysis across the SURPASS programme likewise found no hazard ratio above 1.0 for MACE-4 or its components against pooled comparators, with upper confidence bounds below 1.3, although event numbers were low ^[9]. The main cardiovascular comparison in the set is SURPASS-CVOT, a double-blind, active-comparator noninferiority trial in 13,299 patients with type 2 diabetes and atherosclerotic cardiovascular disease. The primary composite of cardiovascular death, myocardial infarction or stroke occurred in 801 patients (12.2%) receiving tirzepatide and 862 (13.1%) receiving dulaglutide 1.5 mg (hazard ratio 0.92; 95.3% CI 0.83 to 1.01). Noninferiority was met (P = 0.003), but superiority was not (P = 0.09) ^[22]. Because the comparator, dulaglutide, is itself a GLP-1 RA with demonstrated cardiovascular benefit, this result shows tirzepatide was noninferior to an active class member; it does not show a reduction in events beyond that class effect.

Surrogate and risk-factor data point in the same direction. Across SURPASS-1 to -5, systolic blood pressure fell by 2.8 to 12.6 mmHg, and mediation analyses attributed the reduction primarily to weight loss, with weight-loss-independent effects explaining 33% to 57% of the difference from insulin glargine in SURPASS-4 ^[43]. In the SURMOUNT-1 post hoc analysis of 2461 participants with obesity or overweight, without diabetes or a history of atherosclerotic cardiovascular disease, the predicted 10-year ASCVD risk score, whose baseline median was low (1.5% to 1.6%), changed by -23.5% to -16.4% with tirzepatide against +12.7% with placebo at week 72 (P < 0.001) ^[48]. This is a modelled risk estimate driven by changes in risk factors, not observed events, and it should not be read as an outcome result. Narrative reviews also describe improvements in insulin sensitivity, weight and lipid profile, while noting that long-term cardiovascular effects still required verification ^[8,27].

The SUMMIT trial provides randomised evidence for tirzepatide in obesity-related heart failure with preserved ejection fraction. In the expanded analysis of 731 patients (BMI 38.2 kg/m² on average) followed for a median of 104 weeks, tirzepatide gave consistently favourable hazard ratios of 0.41 to 0.67 across composites of death and worsening heart failure, and at 52 weeks improved the Kansas City Cardiomyopathy Questionnaire Clinical Summary Score by 6.9 points and 6-minute walk distance by 18.3 m compared with placebo ^[24]. The cardiac magnetic resonance substudy (106 evaluable patients) found that left ventricular mass fell by 11 g and paracardiac adipose tissue by 45 mL relative to placebo, with the change in left ventricular mass tracking weight loss; the authors suggest this may contribute to the reduction in heart failure events ^[25]. These findings are specific to tirzepatide in a population with obesity, and one review notes that the effect of the GLP-1 RA class on heart failure remained uncertain ^[37].

Kidney data derive mainly from SURPASS-4. In a post hoc analysis, the annual eGFR decline was -1.4 mL/min/1.73 m² with tirzepatide versus -3.6 with insulin glargine (between-group difference 2.2; 95% CI 1.6 to 2.8), the urine albumin-creatinine ratio rose 36.9% with glargine but not with tirzepatide, and the composite kidney endpoint was less frequent with tirzepatide (hazard ratio 0.58; 95% CI 0.43 to 0.80) ^[23]. Because this was a post hoc analysis of an open-label trial whose comparator was insulin, and the kidney composite included new-onset macroalbuminuria, the finding is hypothesis-generating; the same limits apply to reviews that repeat it ^[4]. A multilevel meta-analysis of eight randomised trials (9533 participants) reported a 26.9% greater reduction in urine albumin-to-creatinine ratio with tirzepatide than controls, significant against placebo and insulin but not against semaglutide, alongside a neutral effect on eGFR (mean difference 0.39 mL/min/1.73 m²; P = 0.46) ^[50]. Tirzepatide therefore appears to lower albuminuria, while evidence that it preserves filtration remains limited to the SURPASS-4 post hoc slope analysis. A pharmacokinetic study in 45 participants found no clinically relevant effect of renal impairment on exposure, so the authors concluded that dose adjustment may not be required ^[34].

Liver disease appears in the tirzepatide literature as a development indication rather than a reported outcome. Tirzepatide is described as under investigation for non-cirrhotic non-alcoholic steatohepatitis ^[42], and the renal pharmacokinetic study lists nonalcoholic steatohepatitis among the conditions for which the drug was being developed ^[34]. Both statements describe programme scope only; efficacy and safety on liver histology or fibrosis are not reported in these sources.

Evidence for neurological effects is preclinical. In a neuroblastoma cell line exposed to normal and high glucose, tirzepatide activated the pAkt/CREB/BDNF pathway and mitigated high-glucose-induced neurodegeneration and neuronal insulin resistance; the authors describe this as insight into potential benefit in diabetes-related neuropathy, and it was an in vitro study ^[38]. A mouse pharmacokinetic study found that tirzepatide did not appear to cross the blood-brain barrier within 1 hour of intravenous injection but entered slowly over 6 hours, presumably via extracellular pathways, whereas albiglutide and dulaglutide entered the brain faster; the authors therefore prioritised the faster-penetrating agents as candidates for Alzheimer's and Parkinson's disease ^[39]. Neither study involved human participants, and neither provides evidence on dementia or Parkinson's disease outcomes with tirzepatide.

Class-level GLP-1 RA evidence supports cardiovascular benefit in type 2 diabetes, whereas tirzepatide's own record consists of cardiovascular safety in SURPASS-4 and noninferiority to dulaglutide in SURPASS-CVOT, together with favourable surrogate effects and heart-failure outcomes in SUMMIT ^[15,22,24]. Kidney findings for tirzepatide rest on post hoc and pooled albuminuria analyses, hepatic use remains a development indication, and neurological hypotheses remain confined to laboratory and animal work.

The safety evidence for tirzepatide comes from three kinds of source that carry different weight: randomized trials, which give comparative event rates under defined doses and populations; pooled analyses and reviews, which add rarer outcomes; and reports of real-world behaviour, which are hypothesis-generating only. The findings below are organized by organ system, with the design behind each stated.

Gastrointestinal events were the most consistent finding in the phase 3 programme. In SURPASS-2, which randomized 1879 adults with type 2 diabetes to tirzepatide 5, 10 or 15 mg or semaglutide 1 mg for 40 weeks, nausea occurred in 17 to 22% of tirzepatide recipients and 18% of semaglutide recipients, diarrhoea in 13 to 16% and 12%, and vomiting in 6 to 10% and 8%; most events were mild to moderate ^[13]. Against placebo in SURPASS-1 (monotherapy, 40 weeks), nausea occurred in 12 to 18% versus 6%, diarrhoea in 12 to 14% versus 8%, and vomiting in 2 to 6% versus 2% ^[44]. Gaps were larger against basal insulin: in SURPASS-4, nausea was 12 to 23% and diarrhoea 13 to 22% with tirzepatide versus 2% and 4% with glargine, and events clustered in the dose-escalation phase ^[15]. In SURPASS-3, treatment discontinuation due to an adverse event was more common with tirzepatide than with insulin degludec ^[16].

Dose dependence is described across sources. A review of the SURPASS programme reports that nausea, vomiting, diarrhoea and constipation were more common at higher doses ^[9]. In the Japanese SURPASS J-mono trial, nausea rose from 12% at 5 mg to 20% at 10 and 15 mg (8% with dulaglutide) ^[14], and in SURPASS J-combo overall treatment-emergent adverse events were more frequent at 15 mg (84%) than at 5 or 10 mg (74% each) ^[45]. In the largest comparison, tirzepatide versus dulaglutide 1.5 mg in 13,165 analysed patients with type 2 diabetes and atherosclerotic disease, overall adverse-event incidence appeared similar but gastrointestinal events were more frequent with tirzepatide ^[22]. In SURMOUNT-3 (adults without diabetes, 72 weeks), gastrointestinal events were again the most common, mostly mild to moderate ^[47]. A network meta-analysis of 27 trials in obesity or overweight found a higher adverse-event incidence in participants without type 2 diabetes than in those with it ^[19].

A meta-analysis of nine randomized trials (9871 participants; 6828 on tirzepatide) examined pancreatitis and gallbladder disease ^[21]. Pancreatitis was not significantly associated with tirzepatide (relative risk 1.46, 95% CI 0.59 to 3.61), but the wide interval leaves the estimate imprecise. The composite of gallbladder or biliary disease was associated with tirzepatide versus placebo or basal insulin (relative risk 1.97, 95% CI 1.14 to 3.42), although cholelithiasis, cholecystitis and biliary diseases were not significant when analysed separately. A clinical review lists gallbladder disorders and acute pancreatitis among rare serious adverse events of GLP-1 receptor agonists, including tirzepatide ^[18].

Hypoglycaemia depended on the comparator and background therapy. No clinically significant (below 54 mg/dL) or severe hypoglycaemia was reported with tirzepatide monotherapy in SURPASS-1 ^[44]; in SURPASS-2 the rate was 0.2 to 1.7% across doses, versus 0.4% with semaglutide ^[13]. Against insulin the direction reversed: in SURPASS-4, hypoglycaemia (below 54 mg/dL or severe) occurred in 6 to 9% with tirzepatide versus 19% with glargine ^[15], and in SURPASS-6, with basal insulin in both arms, at 0.4 versus 4.4 events per patient-year for pooled tirzepatide versus prandial lispro ^[17]. The obesity network meta-analysis found no increase in serious adverse events or in hypoglycaemia below 54 mg/dL for any agent studied ^[19].

For cardiovascular safety, SURPASS-4 found adjudicated major adverse cardiovascular events not increased versus glargine (hazard ratio 0.74, 95% CI 0.51 to 1.08), with 25 deaths (3%) on tirzepatide and 35 (4%) on glargine ^[15]. A programme-wide meta-analysis found no hazard ratio above 1.0 for these events, with upper confidence bounds below 1.3, though event numbers were low ^[9]. SURPASS-CVOT randomized 13,299 patients to tirzepatide or dulaglutide; the primary composite occurred in 12.2% versus 13.1% (hazard ratio 0.92, 95.3% CI 0.83 to 1.01), meeting non-inferiority but not superiority (P = 0.09) ^[22]. Blood pressure reductions occurred irrespective of antihypertensive use, which the authors interpret as easing theoretical concern about hypotension ^[43].

For kidney outcomes, a post hoc SURPASS-4 analysis found slower eGFR decline with tirzepatide than glargine (-1.4 versus -3.6 mL/min/1.73 m2 per year) and a lower risk of a composite kidney endpoint (hazard ratio 0.58, 95% CI 0.43 to 0.80) ^[23]. A review notes that an early eGFR dip can occur, as with other kidney-protective drugs ^[4], and a meta-analysis of eight trials found a neutral pooled effect on eGFR ^[50]. After a single 5 mg dose in 45 subjects, exposure was similar across renal impairment groups apart from a 25 to 29% higher AUC in moderate impairment, and dose adjustment may not be required ^[34].

A literature review of six trials, only one of them on tirzepatide, reported reduced AUC and maximum concentration of an oral hormonal contraceptive given with tirzepatide, whereas five studies of other GLP-1 receptor agonists showed no significant effect; the authors attribute this to tirzepatide's greater delay of gastric emptying and call for further study ^[35]. A clinical review of the drug class also reports increased aspiration risk during anaesthesia and difficult bowel preparation for colonoscopy, findings drawn from class-level rather than tirzepatide-specific evidence ^[18].

A network meta-analysis of 22 randomized trials (2258 participants) found that GLP-1 receptor agonists reduced lean mass (mean difference -0.86 kg), about 25% of total weight loss, while relative lean mass was unchanged. Tirzepatide 15 mg and semaglutide 2.4 mg gave the greatest weight and fat loss but were among the least effective at preserving lean mass ^[20]. This is a body-composition finding, not a demonstrated functional harm. Behavioural evidence is weaker: a study combining analysis of about 68,250 Reddit posts with a survey of 153 drinkers with a BMI of 30 or higher reported lower self-reported alcohol intake after starting semaglutide or tirzepatide, but self-report, the comparison design and social-media sampling limit inference ^[40].

Gastrointestinal effects, the lower hypoglycaemia rate against insulin, and the composite gallbladder or biliary excess are supported by randomized data. Pancreatitis remains unresolved because the interval is wide. The contraceptive interaction, perioperative aspiration and lean-mass findings rest on small studies, narrative reviews or class-level evidence and are signals needing tirzepatide-specific confirmation.

The included evidence points to several practical considerations without amounting to clinical guidance. Gastrointestinal events concentrate during dose escalation and rise with dose. Hypoglycaemia risk depends on concomitant insulin rather than on tirzepatide alone. The pancreatitis and biliary estimates are pooled trial-level figures and do not identify which individuals are at higher risk. Reviewers recommend individualized dosing, thorough patient assessment and continued vigilance, particularly regarding delayed gastric emptying and anaesthesia ^[18], and the contraceptive interaction has prompted calls for provider and patient education ^[35].

The 50 selected articles support different conclusions in different domains, and the strength of the evidence should be read domain by domain. Glycaemic efficacy in type 2 diabetes rests on the strongest base: randomised phase 3 trials showed larger HbA1c reductions with tirzepatide than with placebo ^[44], semaglutide 1 mg ^[13], insulin glargine ^[15], insulin degludec ^[16], prandial insulin lispro ^[17] and dulaglutide in a Japanese population ^[14]. Against basal or prandial insulin, tirzepatide also lowered hypoglycaemia (6 to 9% versus 19% with glargine, and 0.4 versus 4.4 events per patient-year with lispro) ^[15,17]. Weight loss followed the same pattern, with body weight falling by 7.5 to 12.9 kg while insulin degludec added 2.3 kg ^[16]. In adults with overweight or obesity without diabetes, the one primary trial in the set reported an additional 18.4% weight loss against a 2.5% gain with placebo after a lifestyle programme ^[47]. For cardiovascular outcomes, SURPASS-CVOT found tirzepatide non-inferior to dulaglutide (hazard ratio 0.92; 95.3% CI 0.83 to 1.01; P = 0.003 for non-inferiority, P = 0.09 for superiority) ^[22]. Heart failure with preserved ejection fraction and obesity is a setting where the set shows placebo-controlled cardiovascular clinical benefit, from one trial and its imaging substudy ^[24,25]. Kidney findings come from a post hoc analysis and a meta-analysis of albuminuria ^[23,50]. Hepatic evidence is close to absent, and the safety data are dominated by gastrointestinal events plus a small number of pooled analyses.

The clinical superiority of tirzepatide over selective GLP-1 receptor agonists comes mainly from a single open-label 40-week comparison with semaglutide, in which the HbA1c advantage was small at the lowest dose (−0.15 points) and larger at 15 mg (−0.45 points), while the weight difference grew from 1.9 to 5.5 kg ^[13]. One review in the set places tirzepatide with liraglutide and semaglutide in the strongest weight-loss group among glucose-lowering drugs ^[3], and another quotes placebo-corrected losses of about 5%, 12% and 18% for liraglutide, semaglutide and tirzepatide ^[18]. These are cross-trial figures. A network meta-analysis placed tirzepatide 15 mg (−16.53%) behind retatrutide 12 mg (−22.10%) and reported smaller losses in people with type 2 diabetes ^[19], so the class is still moving beyond dual agonism.

Why the dual mechanism works remains unsettled. Human and mouse data suggest that tirzepatide improves insulin sensitivity partly independently of weight loss: weight loss explained only 13% and 21% of the HOMA2-IR improvement at 10 and 15 mg ^[30], and the effect persisted in obese mice lacking the GLP-1 receptor ^[31]. On food intake the evidence is less favourable to GIP: in rodents, GIPR agonism alone did not change food choice, and the shift produced by tirzepatide disappeared in GLP-1 receptor knockout mice ^[32]. One review notes that GIP effects on food intake have not been shown in humans ^[9]. Structural work on the receptor complexes ^[28] and signalling work describing biased agonism ^[29] exist, but none of this links receptor pharmacology to a clinical outcome.

For cardiovascular outcomes, the class already carried outcome-trial evidence ^[5]. SURPASS-CVOT adds that tirzepatide was not inferior to an active comparator that itself reduces events ^[22]. It therefore rules out excess risk against dulaglutide but cannot show an absolute benefit over placebo.

Patient selection follows the trial populations. In type 2 diabetes, the SURPASS data cover metformin-treated, insulin-treated and high cardiovascular risk patients, with the clearest advantages over insulin and lower hypoglycaemia ^[16,15,17]. A single-dose pharmacokinetic study in 45 participants found no clinically relevant effect of renal impairment on exposure, although the moderate-impairment group had 25 to 29% higher AUC ^[34]. For obesity, tirzepatide worked after an intensive lifestyle programme ^[47], which fits a review recommending that pharmacotherapy be combined with behavioural, nutritional and physical-activity interventions ^[2]. Among treatment-adherent participants in SURPASS, a weight reduction of 15% or more was more likely with higher doses, female sex, White or Asian race, younger age, metformin background and better glycaemic status ^[46]. This is descriptive and should not be used to withhold treatment.

Monitoring follows the safety signals. Gastrointestinal adverse events were the most common events in trials and mostly mild to moderate ^[13,15]. Gallbladder or biliary disease was more frequent than with placebo or basal insulin in a pooled analysis (risk ratio 1.97), while pancreatitis was not ^[21]. A literature review reports that tirzepatide reduced exposure to an oral contraceptive, unlike GLP-1 receptor agonists ^[35]. Another review lists aspiration risk during anaesthesia and difficult bowel preparation ^[18]. Lean mass made up about a quarter of total weight loss with GLP-1-based agents, and tirzepatide 15 mg was among the least effective at preserving it ^[20]. Blood pressure fell without dependence on antihypertensive treatment, and there was no further decrease when baseline systolic pressure was below 122 mmHg ^[43]. Use of tirzepatide grew rapidly in one health system, and its authors call for evidence-concordant prescribing and stable supply ^[7].

Most efficacy data come from sponsor-funded trials of 40 to 104 weeks ^{[13,15,16,22]}. Several key comparisons were open-label ^{[13,15,16,17]}, which matters for patient-reported and adverse-event outcomes. Cardiovascular and renal statements often rest on post hoc, surrogate or modelled outcomes: kidney outcomes ^[23], biomarkers described by the authors as hypothesis-generating ^[36], mediation of blood pressure change ^[43] and predicted ASCVD risk ^[48]. Alcohol findings come from social-media posts and a self-reported survey of 153 people ^[40]. Mechanistic claims about neuroprotection rest on a neuroblastoma cell line ^[38]. Narrative reviews in the set largely summarise the same trials and preclinical work ^[1,8,26,37].

This was an automated review and not a formal systematic review. Only Scopus and PubMed were searched, on 2026-09-25, retrieving up to 70 records per database per year for 2021 to 2026, so busy years may be truncated. The citation-ranked selection of 50 papers favoured highly cited older work: the included articles span 2021 to 2025, with none from 2026, and the newest literature is under-represented. Data were extracted from abstracts in a single automated pass with no second reader, and no risk-of-bias tool or GRADE assessment was applied. Publication bias was not assessed, no meta-analysis was performed, and the review was not registered.

The selected set also leaves large evidence areas uncovered. It contains no abstract addressing eye or skin safety, eating disorders or other psychiatric outcomes, thyroid or cancer safety, paediatric or older populations, or cost and access beyond utilisation counts. Hepatic outcomes are represented only by ALT changes ^[46] and by mentions of steatohepatitis as a development indication ^[10,42], with no histology or liver-outcome trial. Of the obesity programme, the set has the design paper ^[12], SURMOUNT-3 ^[47] and a SURMOUNT-1 post hoc analysis ^[48], but no primary SURMOUNT-1 or SURMOUNT-2 report, and no abstract reports weight regain after stopping tirzepatide. These are gaps in the retrieved set. They should not be read as gaps in the underlying literature.

Long-term data after discontinuation are the most immediate need, given that weight regain occurs in 25% or more of people after behavioural interventions ^[2] and that no tirzepatide equivalent appears in the set. Head-to-head trials of tirzepatide against semaglutide in obesity are needed, since the only direct comparison here was in type 2 diabetes ^[13]. Placebo-controlled outcome trials in people without diabetes would show whether the predicted cardiovascular risk reduction ^[48] translates into events. Body-composition trials should test whether lean-mass loss ^[20] has functional consequences and can be prevented. Surveillance of biliary events ^[21] and of drug interactions and peri-procedural risks ^[35,18] needs larger and longer data. Finally, next-generation multi-agonists such as retatrutide, mazdutide and orforglipron are already compared indirectly with tirzepatide ^[19], and should be tested against it directly.

Within the 50 articles retrieved, tirzepatide reduced HbA1c and body weight more than placebo, semaglutide, dulaglutide and insulin, with less hypoglycaemia than insulin, and it was not inferior to dulaglutide for major cardiovascular events. Heart failure with preserved ejection fraction is supported by one trial, kidney findings come from post hoc or pooled analyses, and hepatic evidence is limited to ALT changes in this set. Gastrointestinal events are the main tolerability issue, with gallbladder disease, drug absorption and lean-mass loss requiring attention. The mechanism of the added GIP effect is unresolved, and no abstract addresses durability of weight loss after treatment stops. Because the review used abstracts only, one search date, two databases and no formal appraisal, these conclusions describe the retrieved literature and should be confirmed against full trial reports before they guide practice.

This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement. The following checklist documents compliance with each of the 27 items.

• Checklist adapted from: Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.