Effects of Glucagon-Like Peptide-1 Receptor Agonists on Bone Metabolism in Type 2 Diabetes Mellitus: A Systematic Review and Meta-Analysis

Background Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are an intriguing class of antihyperglycemic drugs for type 2 diabetes mellitus (T2DM). Such drugs not only play a primary role in regulating blood glucose levels but also exhibit additional pleiotropic effects, including potential impacts on bone metabolism and fracture risk. However, the mechanism of such drugs is unclear. The purpose of this study was to evaluate the effect of GLP-1 RAs on bone metabolism in T2DM. Methods From database inception to May 1, 2023, the searches were conducted on multiple databases such as Web of Science, Embase, PubMed, CNKI, the Cochrane Library, Wanfang, and VIP. We systematically collected all randomized controlled trials of bone metabolism in patients with T2DM treated with GLP-1 RAs. The quality evaluation was performed according to the Cochrane Handbook for Systematic Reviews of Interventions. Data extraction was analyzed using Review Manager 5.4 software, and funnel plots were drawn to evaluate publication bias. Results Twenty-six randomized controlled trials that met the inclusion criteria were included, involving a total of 2268 participants. In this study, compared to other antidiabetic drugs or placebo, GLP-1 RAs were found to significantly increase serum calcium (mean difference (MD) = 0.05, 95% confidence interval (CI) (0.01, 0.09), P = 0.002], bone alkaline phosphatase [standardized MD (SMD) = 0.76, 95% CI (0.29, 1.24), and P = 0.001), and osteocalcin (SMD = 2.04, 95% CI (0.99, 3.08), and P = 0.0001) in T2DM. Specifically, liraglutide increased procollagen type 1 N-terminal propeptide (SMD = 0.45, 95% CI (0.01, 0.89), and P = 0.04). GLP-1 RAs were also associated with a reduction in cross-linked C-terminal telopeptides of type I collagen (SMD = −0.36, 95% CI (−0.70, −0.03), and P = 0.03). In additionally, GLP-1 RAs increased lumbar spine bone mineral density (BMD) (SMD = 1.04, 95% CI (0.60, 1.48), and P < 0.00001) and femoral neck BMD (SMD = 1.29, 95% CI (0.36, 2.23), and P = 0.007). Conclusions GLP-1 RAs can not only improve BMD in the lumbar spine and femoral neck of patients with T2DM but also protect bone health by inhibiting bone resorption and promoting bone formation. Systematic Review Registration. PROSPERO, identifier CRD42023418166.


Background
Type 2 diabetes mellitus (T2DM) is the most common endocrine disorder.Due to long-term exposure to high blood glucose, patients can develop a series of complications, mainly afecting the major blood vessels of the heart, microvessels in the kidneys and retina, as well as the nervous and skeletal systems [1].Such conditions exert a substantial impact on the quality of life for patients and their families.
In recent times, T2DM is increasingly recognized as a signifcant contributor to secondary osteoporosis and fragility fractures in the skeletal system [2].In clinical practice, several commonly prescribed antidiabetic medications not only contribute to controlling the blood glucose level of patients but also have diverse efects on their bone health.For instance, a study has indicated that metformin plays a benefcial role in promoting bone formation and improving bone metabolism [3].However, as shown in a study, metformin does not have a substantial impact on enhancing bone mineral density (BMD) in patients with T2DM [4].Conversely, thiazolidinedione drugs have been found to induce osteoblast apoptosis, leading to reduced bone formation and an increased risk of fractures [5].
Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are a novel class of antidiabetic medications favored by T2DM patients due to their dual benefts of lowering blood glucose levels and promoting weight loss [6].According to a study, GLP-1 RAs can reduce bone breakdown by affecting osteoclasts, thereby inhibiting bone resorption.Besides, GLP-1 RAs can enhance osteoblast activity and promote bone formation [7].Furthermore, GLP-1 RAs can control blood glucose levels positively, thus infuencing bone health.Such processes exert an antiosteoporotic role [8].Although GLP-1 RAs can inhibit bone resorption, stimulate bone formation, enhance BMD, and improve overall bone quality [8], the impact of GLP-1 RAs on fracture risk remains highly controversial.A meta-analysis demonstrated that compared to other antidiabetic medications, the use of GLP-1 RAs does not result in a reduction in fracture risk among patients with T2DM [9].Nevertheless, a separate network meta-analysis published in 2018 indicated that GLP-1 RAs signifcantly decrease the risk of fractures in patients with T2DM compared with placebo or other antidiabetic medications [10].
As a result, the impact of GLP-1 RAs on fracture risk and their infuence on sensitive bone metabolism markers remain uncertain.In addition, diferences in the structure and duration of action among various GLP-1 RAs may contribute to variations in their efects.Currently, comprehensive meta-analyses focusing on the relationship between GLP-1 RAs and fracture risk are very limited, and the efects of these medications on bone metabolism markers and BMD have not been extensively studied in meta-analyses.Terefore, the purpose of this study was to systematically evaluate and analyze the efects of GLP-1 RAs on selected bone metabolism markers and BMD in T2DM.

Protocol and Registration.
Te protocol for this systematic review and meta-analysis has been registered with PROSPERO (registration number: CRD42023418166).Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [11], we present our methods and fndings (Supplemental Table 1).

Eligibility Criteria. Randomized controlled trials (RCTs)
were included to compare the efcacy of GLP-1 RAs with other antidiabetic medications or placebo.Each trial included participants who were aged 18 years or older and diagnosed with T2DM according to the diagnostic criteria of the World Health Organization or the American Diabetes Association if glycated hemoglobin (A1C) reaches or exceeds 6.5%, or if fasting plasma glucose levels reach or exceed 126 mg/dL (7.0 mmol/L), or if the 2-hour plasma glucose during an oral glucose tolerance test (OGTT) reaches or exceeds 200 mg/dL (11.1 mmol/L), or if random plasma glucose levels at any time reach or exceed 200 mg/dL (11.1 mmol/L) accompanied by typical symptoms of diabetes, the diagnosis of T2DM can be made based on any of these criteria [12].Tere were no restrictions based on race, gender, or duration of the disease, and exclusion criteria did not consider the presence of osteoporosis or a history of fractures.Te primary outcomes of interest were BMD or bone metabolism markers.In the case of multiple publications from the same study, we selected the data that provided the most comprehensive and longest follow-up period.

Search Strategy.
We searched PubMed, Cochrane, Embase, Web of Science, CNKI, Wanfang, and VIP for relevant literature from the database inception to May 1, 2023.Detailed information regarding our search strategy is presented in the electronic supplementary material (Supplemental Table 2).To ensure inclusiveness, we included search terms related to "GLP-1 RAs" as well as specifc terminologies of diferent types of GLP-1 RAs such as semaglutide, liraglutide, exenatide, dulaglutide, and benaglutide to capture all potentially eligible studies and avoid any omissions.

Selection Process. All search results were imported into
EndNote (version X9, Tomson Reuters, Philadelphia, PA, USA) to remove any duplicate records.Two reviewers independently conducted an initial screening based on the title and abstract of each article.Subsequently, the remaining articles underwent a full-text assessment to determine their eligibility for inclusion, with reasons for exclusion carefully documented.In case of disagreements, a third reviewer was consulted to reach a consensus.Articles that did not provide the necessary data were excluded, as well as those for which the required data could not be obtained even after contacting the corresponding authors.

Data Collection and Risk of Bias
Assessment.Two researchers independently extracted the relevant data.Te included data consisted of the frst author's name, publication year, country where the study was conducted, sample sizes and relevant information for the treatment and control groups, names of GLP-1 RAs, dosage, duration of the study, and results of bone metabolism-related markers before and after treatment.In cases of diferences of opinion during the extraction process, the resolution was achieved through discussion or involvement of a third researcher.Te Review Manager 5.4 software was used by the two researchers to assess the risk of bias in the included RCTs.Tis assessment was conducted using the bias risk assessment tool provided in the Cochrane Handbook for Systematic Reviews of Interventions, version 5.3 [13].
2.6.Statistical Analysis.Te relevant outcome markers refecting bone metabolism and BMD collected in this study were all continuous variables.Terefore, the mean diference (MD) or standardized mean diference (SMD) with standard 2 International Journal of Endocrinology deviation (SD) was chosen as efect markers.Heterogeneity between included studies was evaluated using the Q test and I 2 statistic.Te Q test primarily was used to assess the p value, and if the result of the heterogeneity test was p ≥ 0.1 and I 2 < 50%, it indicated that there was no statistically signifcant heterogeneity among the studies, and a fxedefect model was used for meta-analysis.Otherwise, the subgroup analysis could be performed to identify the source of heterogeneity or a random-efects model could be used to pool the efect sizes for meta-analysis.Sensitivity analysis was conducted to assess the stability of the results by sequentially excluding individual studies and reanalyzing the data.If the exclusion of a particular study led to signifcant changes in the pooled efect size or its heterogeneity, further reading and evaluation of that study were necessary.Publication bias was assessed by a visual funnel plot of the main outcome markers.Statistical analysis of all predetermined outcome markers was performed using RevMan 5.4 provided by the Cochrane Collaboration.P < 0.05 was considered statistically signifcant.

Search Results
. According to the established retrieval strategy, a total of 6081 studies were screened from 7 databases.Besides, EndNote X9 was used to remove 1389 duplicate records and 1986 records were automatically marked as ineligible.Ten, we read the titles and abstracts of the remaining articles based on the inclusion and exclusion criteria, ultimately excluding 2353 articles.Among the remaining 353 articles, 8 were inaccessible in full text.After thoroughly reading the full texts, we found that 302 articles did not contain the required outcome markers, 10 were not RCTs, and 7 were study protocols.Terefore, 26 studies were ultimately included , involving a total of 2268 participants (Figure 1).[34][35][36], 1 on the combination of exenatide and dulaglutide [37], and 1 study on benaglutide [38].For a comprehensive understanding of the included studies, the detailed characteristics were shown in Table 1.Summary of fndings is provided in Supplemental Table 3.

Study
Of the 25 studies included, 16 studies provided detailed descriptions of the methods used for random sequence generation, and the remaining 9 studies mentioned randomization without specifying the exact methods used.Tere were 10 studies explicitly stating the allocation concealment method, and 10 studies in blinding for participants and personnel, as well as 24 studies on outcome assessors.Te risk of bias assessment for each study was presented in Figure 2.
International Journal of Endocrinology Such a result indicated a statistically signifcant diference (Figure 8).According to a subgroup analysis based on the type of medication, liraglutide markedly improved the osteocalcin level (SMD � 2.35 and 95% CI (1.11, 3.60)) (Figure 9).
In addition, a sensitivity analysis was performed by systematically excluding each study, and no signifcant change was found in the overall efect size and heterogeneity.
Funnel plots were created to assess the impact of GLP-1 RAs on CTX and lumbar spine BMD (Supplemental Figures 4  and 5).Te plots displayed a symmetrical distribution of studies on both sides of the axis, indicating a relatively low risk of publication bias.

Discussion
Ultimately, GLP-1 RAs played a benefcial efect on BMD and bone metabolism in patients with T2DM.Compared to other antidiabetic drugs or placebo, GLP-1 RAs may demonstrate greater potential benefts for bone health in the treatment of T2DM.Specifcally, dipeptidyl peptidase-4 (DPP-4) inhibitors and GLP-1 RAs share similar mechanisms of action.However, some studies suggest that DPP-4 inhibitors have a neutral or mildly positive efect on bone health [40].In contrast, clinical trials have demonstrated more pronounced efects of GLP-1 RAs in reducing glycated hemoglobin and body weight, making them a more favorable option for most patients [41].Furthermore, insulin   GLP-1 Ras, glucagon-like peptide-1 receptor agonists; BMI, bone mineral density; HbA1c, hemoglobin A1c; Lira, liraglutide; Exen, exenatide; Bena, benalutide; NR, not reported; QD, quaque die; BID, bis in die; QW, quaque week; TID, ter in die.
International Journal of Endocrinology therapy for diabetes may alter levels of certain biomarkers associated with bone metabolism, such as advanced glycation end products and other indicators related to bone density and fracture risk, thereby infuencing the processes of bone formation and resorption [42].Our study also concluded that the longer treatment duration was associated with the more signifcant improvements in lumbar spine BMD.Furthermore, CTX could be reduced and BALP, osteocalcin, as well as P1NP, could be increased.Tese fndings indicated that GLP-1 RAs could suppress bone resorption and promote bone formation in T2DM.However, notably, the studies included in this analysis primarily focused on liraglutide and no statistically signifcant changes were revealed in subgroup analyses for other GLP-1 RAs.Tis could be attributed to the limited number of studies and sample sizes for other GLP-1 RAs, as well as potential diferences in molecular structure between exenatide and dulaglutide in contrast to liraglutide.
In recent years, GLP-1 RAs have attracted much attention as an antidiabetic medication.As shown in the study, these drugs not only reduced glucose sugar levels and had a cardiovascular protective efect but also may have a certain protective efect on bone health.According to the animal experiments, as opposed to the wild-type control group, GLP-1 RAs knockout mice exhibit an increase in osteoclast numbers as well as bone resorption levels, and a decrease in BMD [43].In a mouse model induced by lipopolysaccharide, the combined treatment with GLP-1 RAs signifcantly reduces osteoclast numbers and CTX in comparison with mice receiving lipopolysaccharide alone [44].As reported by Sedky, GLP-1 RAs can enhance osteocalcin in diabetic rats, thereby increasing bone mass and strength [45].Tese animal studies indicated that GLP-1 RAs can inhibit bone resorption, promote bone formation, and improve BMD.Such results were consistent with the fndings of this study.
However, there was still no defnitive and consistent conclusion from existing clinical studies on the impact of GLP-1 RAs on bone metabolism markers and fracture risk in patients with T2DM.In a clinical trial published by Gilbert in 2016 [18], the trial lasting for 104 weeks demonstrated that liraglutide as monotherapy does not afect total BMD in patients.However, the dropout rate of this trial is as high as 52%, which may greatly afect the fnal results and may be the reason for the inconsistency with our study fndings.Zhang [28] pointed out that liraglutide had no signifcant efect on whole-body BMD and bone formation markers in obese and overweight patients with T2DM after 26 weeks of treatment.However, it does reduce the level of the bone resorption marker CTX, which was consistent with the results of our study.Te diference in results might be only obese and overweight patients with T2DM were included in the clinical study, and the low-grade infammatory state in patients with obesity also afected bone metabolism [46,47].In addition, some studies have suggested that exenatide may increase fracture risk.However, this meta-analysis only assessed fractures as adverse events without including detailed bone metabolism markers or bone quality indicators, such as bone resorption and formation markers, BMD, and calcium and phosphate levels [48].In contrast, our study provided a detailed analysis of these markers, enabling a more accurate evaluation of the efects of GLP-1 RAs on bone health.Overall, although the analyses suggest that liraglutide may reduce fracture risk in patients with T2DM, patients using GLP-1 RAs do not exhibit a signifcant increase in fracture risk; in some cases, GLP-1 RAs may be associated with a lower fracture risk.Compared to other antidiabetic medications, GLP-1 RAs generally show a safer profle regarding fracture risk [49].6 International Journal of Endocrinology Tis meta-analysis has several limitations.First, most of the included RCTs primarily used liraglutide, while the number of RCTs for other GLP-1 RAs was limited, with smaller sample sizes, potentially afecting the comprehensiveness of the results.Second, the efects of the medication may take a longer time to manifest, with treatment durations  International Journal of Endocrinology in the studies ranging from 4 weeks to 104 weeks.Shorter treatment periods could impact the strength of the evidence.Tird, we only included published clinical trials and did not include unpublished trials, which may introduce publication bias and omit relevant data.Furthermore, although the heterogeneity of results was high, sensitivity analyses were performed by systematically excluding each study to assess their impact on heterogeneity and overall efect size, with fndings indicating that these exclusions did not signifcantly alter the overall efect size or heterogeneity.Finally, it is noteworthy that despite using multiple international databases to ensure comprehensive data retrieval, regional bias may still be present.For instance, studies from Western countries generally suggest that GLP-1 receptor agonists have no signifcant impact on BMD or bone turnover markers, while studies from China have shown potential Heterogeneity: tau 2 = 0.37; chi 2 = 6.07, df = 1 (P = 0.01); I 2 = 84% Test for overall effect: Z = 0.27 (P = 0.79) Total (95% CI) Heterogeneity: tau 2 = 0.32; chi 2 = 82.62,df = 12 (P < 0.00001); I 2 = 85% Test for overall effect: Z = 2.13 (P = 0.03) Test for subgroup differences: chi 2 = 0.31, df = 1 (P = 0.58), I 2 = 0%      International Journal of Endocrinology as opposed to exenatide and dulaglutide.Patients with T2DM are already at high risk of osteoporosis or fractures, so it is important to choose antidiabetic medications that not only lower blood glucose but also minimize the risk of osteoporosis or fractures.

Figure 4 :
Figure 4: Comparison of total hip BMD (a) and femoral neck BMD (b) in the GLP-1 RAs group compared with the control group.

Figure 6 :
Figure 6: Comparison of CTX in the GLP-1 RAs group compared with the control group.

Table 1 :
Characteristics of included studies.
Figure 5: Comparison of serum calcium in the GLP-1 RAs group compared with the control group.
RAs can not only improve BMD at the lumbar spine and femoral neck but also enhance bone quality.Such results can delay the occurrence and progression of osteoporosis and reduce the risk of fractures in patients with T2DM.Among the GLP-1 RAs, liraglutide seems to have more efective efects in reducing CTX and increasing osteocalcin Figure 10: Comparison of P1NP in the liraglutide group compared with the control group.