Dyslipidemia is characteristic of many metabolic disorders, including type 2 diabetes mellitus (T2DM) and obesity. Derangement of serum lipid biomarkers, such as low levels of high-density lipoprotein cholesterol (HDL-C) and high levels of low-density lipoprotein cholesterol (LDL-C) or triglycerides (TG/TAG), are important risk factors for morbid cardio- and cerebrovascular complications such as ischemic heart disease and stroke.¹ They are also implicated in the pathophysiology of diseases such as T2DM, in which inflammation, high adiposity, dyslipidemia, and the formation of harmful lipid-derived metabolites contribute to insulin resistance.² Consequently, the maintenance of patient lipid levels within acceptable bounds is a critical component of managing these conditions. This is typically achieved using lifestyle modifications to diet and exercise and the administration of pharmacological agents such as statins, anti-diabetic drugs, and insulin.³

Glucose-dependent insulinotropic peptide (GIP) and glucagon-like peptide 1 (GLP-1) are incretin hormones with a prominent physiological role in promoting the release of post-prandial insulin.^4,5 This enables an appropriate metabolic response to nutrition and the maintenance of normal blood glucose after food intake, accounting for the difference between insulin levels following IV administration of glucose and those after oral ingestion.⁶ Research has established numerous functions for both hormones that broaden their pharmacological potential well beyond the use of their incretin effects. For example, GIP inhibits the release of gastrin, show glucose-dependent glucagonotropic activity, and regulate signaling pathways in the brain and adipose tissues involved in energy homeostasis.⁷ Similarly, GLP-1 has been shown to decrease gastric emptying, suppress appetite (and thus support behavioral changes to decrease body weight), and improve lipid metabolism, and it might have additional cardio and neuroprotective properties.^4,5,8

GLP-1 agonists (GLP-1A) and dipeptidylpeptidase 4 (DPP-4) inhibitors are well-established pharmacologic agents for leveraging the physiologic effects of GLP-1 in patients with diabetes and related conditions. Tirzepatide is a novel, subcutaneously administered, dual agonist of both GIP and GLP-1 receptors that is being evaluated for use in managing diabetes and obesity.^7-9 In patients with diabetes, the potential utility of tirzepatide has been considered with immense enthusiasm, as increased adherence to lifestyle changes (i.e., by suppressing appetite) and adequate improvement in lipid metabolism and glycemic-control indicators indicate the potential to target multiple points in the disease process, including its behavioral determinants, pathophysiology, and symptomatology, via a single drug. Similarly, considerable benefit has been hypothesized in patients with obesity, which is now understood as a multifaceted metabolic disorder with neurophysiological, behavioral, and endocrine components.⁷ A major challenge in the management of obesity is the prevention of a return to pathologically high levels of weight and adiposity after initial success in weight loss and body re-composition.^10,11 Encouraging evidence suggests that the dual incretin and hormonal effects of tirzepatide could produce substantial weight loss in obese patients, further spurring optimism that tirzepatide could greatly improve the health and long-term clinical course of these patients.

Previous reviews have focused on the drug’s glycemic control, weight loss, and safety, but the efficacy of tirzepatide in improving lipid metabolism across various doses and patient populations is unclear.^12,13 Our aim in this systematic review and meta-analysis was to pool the available randomized data on the effects of tirzepatide on serum lipid levels.

Pre-registration and protocol

This study was preregistered with the International Prospective Register of Systematic Reviews (PROSPERO) under ID CRD42022380144 and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (Supplementary Table 1).

Search strategy

Tirzepatide is a newer drug for which only limited published evidence is available. Consequently, when searches of the PubMed and ClinicalTrials.gov databases with the keyword ‘tirzepatide’ were performed on January 6, 2024, no limits or filters, such as by time, language, or study design, were applied.

Study selection criteria

The inclusion criteria were (1) publications reporting randomized controlled trials (RCTs); (2) studies that assessed any of the following parameters: total cholesterol, HDL-C, LDL-C, and TG; and (3) studies in which at least one group was treated with any dose of tirzepatide.

Article screening process

The titles and abstracts of all records retrieved by the search strategy were screened by two independent reviewers (Muhammad Umar Mahar [MUM] and Omar Mahmud [OM]). Conflicts were resolved by consensus. Subsequently, the full texts of all articles identified as relevant were screened independently by five authors (MUM, OM, Salaar Ahmed [SA], Saleha Ahmed Qureshi [SAQ], and Wasila Gul Kakar [WGK]) to assess their eligibility for inclusion. Any conflicts were resolved by consensus. If any data necessary for statistical analysis were not reported, graphical extraction was performed if possible, or the corresponding authors were contacted via e-mail and asked to share the missing information.

Data extraction

Data from each article were extracted independently by at least two authors (MUM, OM, SAQ, and WGK), and any discrepancies were resolved by consensus. The extracted variables were patient demographic data, number of patients in each study arm, and percentage or absolute change from baseline for HDL-C, LDL-C, total cholesterol, and TG.

The clinical trial reported by Frias et al.¹⁴ (NCT03131687) used multiple comparator groups (placebo and dulaglutide). To include their data in the analysis without double counting the patients treated with tirzepatide, we divided the tirzepatide group by 2 when pooling the results in the corresponding placebo and GLP-1A subgroups. This method allowed us to incorporate an additional data point in our results and preserve the distinction between different control groups without double counting the patients in the intervention group. This method is approved for use at the authors’ discretion by the Cochrane Handbook for Systematic Reviews.¹⁵ No other data included in the analysis required this adjustment.

For studies that did not report numerical data, WebplotDigitizer¹⁶ (Automeris LLC) was used to extract graphical data from the available figures. WebplotDigitizer, an online tool used for extracting graphical data, has been used previously in systematic reviews and has a high level of accuracy.^17-20 Two authors (MUM and OM) independently extracted the graphical data, and then the mean and inter-rater correlations were calculated for each variable. The mean values were then used in the meta-analysis. More detail regarding the data obtained through graphical extraction is reported in Supplementary Methods and Supplementary Tables 2-6.

Statistical analysis

The meta-analysis was performed using Review Manager version 5.4 (Cochrane Collaboration). Separate comparisons were performed for each dose (5, 10, and 15 mg) of tirzepatide. Pooled mean differences in percentage change from baseline were calculated with a 95% confidence interval (CI) using an inverse variance random-effects model.

For each biomarker, the percentage change from baseline resulting from each dose of tirzepatide was compared with the control. The control groups were treated with placebo, insulin, or GLP-1A, and subgroup analyses by control group were performed. Detailed forest plots are reported in Supplementary Figs. 1-12. Heterogeneity was quantified using the τ² and I² statistics. All analyses were two-sided, and a P-value of <0.05 was considered significant.

Publication bias, study quality, and certainty of evidence assessment

Publication bias was modeled using funnel plots (Supplementary Figs. 13-24). Study quality assessment used the Cochrane risk of bias tool. Several included studies were previously included in systematic reviews exploring different outcomes, such as the efficacy of tirzepatide in improving glycemic control.^12,13 We adhered to the previously published risk of bias assessment results for those studies. For newer studies, two authors (SAQ and SA) independently evaluated the study quality, and disagreements were resolved via consensus among all authors.^21-25 All risk of bias assessments are provided in Supplementary Table 7. The certainty of evidence assessment was performed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach by two independent authors (SAQ and OM), with conflicts resolved by a third author (MUM) (Supplementary Tables 8-11).

Literature search

In total, 433 articles were retrieved from the databases and subjected to the initial title and abstract screening, which identified 18 articles for full-text screening. After that, 14 articles (13 RCTs) met the overall inclusion criteria: a 2018 phase 2 RCT by Frias et al.,¹⁴ the SURPASS 1, 2, 3, 4, 5, and 6 RCTs, SURPASS J-MONO, SURPASS AP-Combo, and the SURMOUNT 1, 2, 3, and 4 RCTS.^7,8,14,21-30 Details of the article selection process are depicted in our PRISMA flowchart (Fig. 1).

Study and participant characteristics

The 13 studies analyzed comprised 12 phase-3 clinical trials and one phase 2b clinical trial. All trials were multicenter studies, and 12 were international. Seven trials were double-blinded, and six were open label. Seven trials included active controls (two with dulaglutide, one with semaglutide, two with insulin glargine, one with insulin degludec, and one with insulin lispro). Study durations ranged from 27 to 72 weeks. Further details about the studies and patients included are reported in Tables 1 and 2, respectively. A choropleth world map showing the global distribution of trial centers, particularly centers in Asia, is provided in Fig. 2.

Study quality, publication bias, and certainty of evidence assessment

Nine studies were considered to have a low risk of bias; two studies raised concerns about a non-trivial risk of bias, and two studies were deemed to have a high risk of bias. The complete risk of bias assessment results are provided in Supplementary Tables 8-11.

Results of meta-analysis

Total cholesterol

When dosed at 5 mg, tirzepatide lowered total cholesterol levels more than controls, with a statistically significant mean difference of –4.77% (95% CI, –6.43 to –3.10; K [number of pooled cohorts]=9, N [total patients pooled]=6,421). This advantage increased to –5.39% (95% CI, –6.87 to –3.91; K=10, N=7,041) for 10 mg and –6.55% (95% CI, –8.26 to –4.84; K=12, N=8,291) for 15 mg doses. No statistically significant difference was seen between the placebo, insulin, and GLP-1 agonist subgroups across all three doses of tirzepatide. However, across doses, no significant differences were seen between tirzepatide and the GLP-1A subgroup. Moderate to high heterogeneity was seen across the 5 mg (τ²=4.07, I²=65%), 10 mg (τ²=3.35, I²=61%), and 15 mg (τ²=7.51, I²=77%) doses.

HDL-C

When dosed at 5 mg, tirzepatide raised HDL-C levels more than controls, with a statistically significant mean difference of 3.78% (95% CI, 1.83 to 5.74; K=9, N=6,466). This advantage increased to 5.75% (95% CI, 4.11 to 7.40; K=10, N=7,044) for 10 mg and 6.94% (95% CI, 5.15 to 8.74; K=12, N=8,296) for 15 mg doses. The test for subgroup differences was insignificant across all three doses of tirzepatide. However, tirzepatide was significantly superior to the controls in all three subgroups across all three doses. Moderate to high heterogeneity was seen across the 5 mg (τ²=6.35, I²=75%), 10 mg (τ²=4.70, I²=70%), and 15 mg (τ²=7.63, I²=78%) doses.

LDC-C

When dosed at 5 mg, tirzepatide lowered LDL-C levels more than controls, with a statistically significant mean difference of –5.60% (95% CI, –7.52 to –3.68; K=9, N=5,507). This advantage increased to –5.84% (95% CI, –8.34 to –3.34; K=10, N=6,129) for 10 mg and –7.82% (95% CI, –10.62 to –5.02; K=12, N=7,373) for 15 mg doses. The test for subgroup differences was insignificant across tirzepatide doses. Moreover, no significant difference between tirzepatide and the control was seen in the GLP-1A subgroup regardless of the tirzepatide dose. Moderate to high heterogeneity was seen across the 5 mg (τ²=2.41, I²=29%), 10 mg (τ²=9.42, I²=61%), and 15 mg (τ²=17.96, I²=77%) doses.

TG

When dosed at 5 mg, tirzepatide lowered TG levels more than controls with a statistically significant mean difference of –13.18% (95% CI, –17.08 to –9.28; K=11, N=6,581). This advantage increased to –17.59% (95% CI, –21.46 to –13.73; K=12, N=7,199) for 10 mg and –22.08% (95% CI, –25.38 to –18.78; K=14, N=8,453) for 15 mg doses. The test for subgroup differences was significant in the 15 mg dose analysis, where a greater advantage with tirzepatide was seen versus GLP-1 agonists or placebo, as opposed to insulin. Moderate to high heterogeneity was seen across the 5 mg (τ²=27.47, I²=73%), 10 mg (τ²=31.48, I²=76%), and 15 mg (τ²=26.04, I²=74%) doses.

All meta-analysis results, stratified by dose and lipid profile parameter, are depicted in Fig. 3. A detailed breakdown of the results from the subgroup analyses is reported in Supplementary Figs. 1-12. Funnel plots depicting publication bias are reported in Supplementary Figs. 13-24. Certainty of evidence was assessed using GRADEPro software to rate the level of evidence provided by the meta-analysis results and the references (Supplementary Tables 8-11).

Tirzepatide is a novel dual GIP/GLP-1 receptor agonist (GLP-1 RA) under investigation as a treatment for various metabolic conditions, including T2DM and obesity. Previous reviews have explored its efficacy and safety, but a current and methodologically robust synthesis of the evidence on improvements in patient lipid profiles has been lacking.^31,32 Our meta-analysis is the first to pool all currently available randomized data. Using an appropriate random-effects meta-analysis, we have demonstrated that tirzepatide shows consistent and dose-dependent efficacy in improving all dimensions of the lipid profile across different patient populations and comparators.

Across 5, 10, and 15 mg doses, our data indicate increasingly pronounced improvements in total cholesterol, LDL-C, HDL-C, and TG levels. These results highlight the efficacy of tirzepatide in improving lipid metabolism, in addition to the glycemic control and weight loss benefits emphasized by previous reviews.^31,32 Correcting dyslipidemia in patients with diabetes, obesity, and metabolic syndrome can mitigate the risk of atherosclerotic cardio- and cerebrovascular complications.^33-35 This efficacy might be complemented by tirzepatide’s ability to produce dose-dependent decreases in ApoC-3 and Apo-B, which are implicated in an increased risk of cardiovascular events.³⁶ Current evidence also suggests that tirzepatide might optimize levels of circulating adipokines and biomarkers such as adiponectin and proinsulin, which are associated with inflammation, insulin resistance, and dyslipidemia.^8,14,37,38 Thus, tirzepatide appears to produce a uniquely pronounced and multidimensional improvement in patients’ metabolic status that is superior to alternative therapies, which supports its proposed role as a novel singular cardioprotective agent.^39,40

Our analysis shows that tirzepatide can raise HDL-C levels more than insulin or GLP-1 agonists. This finding is important, given the historical difficulty in finding drugs that effectively improve HDL-C levels. Other conventional agents that raise HDL-C include niacin and fibrates.⁴¹ The significant increases in HDL-C observed with tirzepatide are comparable to those with niacin therapy and show potential superiority to fibrates and statins.⁴² Moreover, niacin produces insulin resistance and the risk of diabetes in obese patients.³⁴ Thus, the ability of tirzepatide to raise HDL-C levels might be superior to that of conventional alternatives and improve patient outcomes with fewer drawbacks.

Unlike with HDL-C and TG levels, the SURPASS-2 trial did not show a significant difference between tirzepatide and semaglutide in the analyses of LDL-C and total cholesterol levels. This null result partially accounts for the heterogeneity in our results for these markers. Conversely, the SURPASS J-MONO trial found that tirzepatide was significantly superior to dulaglutide in improving those parameters. These results reflect the now well-established superiority of semaglutide to dulaglutide in glycemic control, weight loss, and other metabolic parameters.⁴³

The literature also includes studies that report strong evidence of tirzepatide’s utility in conditions such as fatty liver disease. For instance, Ludvik et al.’s analysis²⁹ of a subpopulation of the SURPASS-3 trial showed that tirzepatide was superior to insulin degludec at reducing liver fat content (insulin degludec produced a 15.71% change from baseline, whereas tirzepatide dosed at 15 mg achieved a 47.11% decrease).

Although an increasing volume of evidence supports the clinical use of tirzepatide to treat various conditions, caution is warranted because the drug is relatively new to the market, and many trials remain ongoing. As with other injectable anti-diabetic agents and insulin, the subcutaneous administration of tirzepatide can produce significant adverse effects at the injection site, which could lead to hesitancy and discontinuation.^44,45 This might be especially consequential in patients receiving the drug to reduce body weight (a more elective usage). Tirzepatide also often presents with gastrointestinal disturbances, including abdominal discomfort, nausea, and diarrhea. This relationship is dose-dependent and can affect nearly half of all patients treated with the 15 mg dose, leading to discontinuation rates of up to 10% in some studies.⁴⁵ Finally, although the drug has been shown to aid in weight loss, no data indicate whether this effect is maintained long term in patients using tirzepatide to treat obesity.^44,46,47

It is also important to consider tirzepatide’s efficacy across various patient populations. One established aspect is the differing disease profiles of metabolic conditions in Asians and non-Asians and corresponding differences in how these populations respond to therapeutics. The literature shows that Asian populations tend to respond better to DPP-4 and sodium glucose cotransporter 2 inhibitors than Europeans, whereas GLP-1 RAs might not be as effective.⁴⁸ A recent meta-analysis studying tirzepatide’s efficacy in both Asian and non-Asian populations revealed that tirzepatide was more effective in reducing fasting blood glucose and glycosylated hemoglobin levels in non-Asians than in Asians.⁴⁹ The difference was significant enough to warrant dose adjustments based on race. The observed variation in efficacy could be attributed to the higher prevalence of visceral obesity.⁵⁰ This could in turn extend to a difference in the improvement of lipid levels in Asians. Our results are likely to generalize to Asian populations because the clinical trials included here have adequate representation of Asian centers, as depicted in our choropleth diagram. However, given the rising prevalence of diabetes and obesity in Asia, there is a pressing need for localized evidence tailored to individual patients and local population requirements. This is especially true when prescribing drugs such as tirzepatide, which can potentially affect the metabolic profile of each race and individual differently. Therefore, further investigation into tirzepatide’s efficacy across different populations is warranted.

The retail price of tirzepatide, at more than $12,000, is relatively expensive compared with other agents and ‘cost-effective’ usage depends on patient willingness to spend large sums for improved glycemic control and weight loss.^51,52 Thus, ultimately, economic considerations are a major facet in the clinical utility of tirzepatide, particularly in the lower middle income countries (LMICs) that bear an increasingly disproportionate burden of T2DM and obesity.^53-55

Limitations

Given that tirzepatide is a newer drug, relatively few trials were available for pooling per subgroup. Moreover, study durations and operational definitions, such as that of obesity, varied across studies. Also, due to a lack of data in the available articles, important markers of lipid metabolism such as adipokines and markers of insulin resistance could not be analyzed. However, our study is the only review to include data from all 13 major clinical trials that have assessed the effects of tirzepatide on lipid parameters. In addition, our use of graphical extraction methods maximized the amount of data available for our analysis. Finally, our review is the first to use a random-effects meta-analysis to assess the efficacy of tirzepatide in improving lipid levels. This is the correct statistical approach and incorporates heterogeneity across the included studies into the effect estimates to limit bias. Thus, we provide the first robust and complete analysis of tirzepatide’s effects on the lipid profile that has used an appropriate methodology.

Conclusion

Our review is the first to use a methodologically sound random-effects meta-analysis and show that tirzepatide can optimize a patient’s LDL-C, HDL-C, and TG. The ability of a single drug to improve a broad array of biomarkers while also supporting behavioral components of disease management (through appetite suppression) is encouraging because it might offer a means to both improve and simplify the medical management of conditions such as T2DM, obesity, and metabolic syndrome in certain patient populations. However, tirzepatide is a relatively new drug, and long-term studies are needed to fully characterize its safety and efficacy with prolonged use.

Supplementary materials can be found online at https://doi.org/10.7570/jomes24008.

The authors declare no conflict of interest.

The authors would like to acknowledge Mr. Muhammad Abdullah for providing valuable feedback and assistance in improving the initial manuscript draft.

Study concept and design: SA; acquisition of data: MUM, OM, and SA; analysis and interpretation of data: MUM, OM, SAQ, and WGK; drafting of the manuscript: MUM, OM, SA, SAQ, WGK, and SSF; critical revision of the manuscript: MUM, OM, SA, SAQ, WGK, and SSF; statistical analysis: MUM and OM; administrative, technical, or material support: SSF; and study supervision: SSF.