Anti-osteoporosis treatments changed body composition in postmenopausal women: A systematic review and meta-analysis

Background:

Whether the widespread anti-osteoporosis treatments in postmenopausal women also benefit the change of body composition (lean body mass [LBM] and body fat mass [FM]) remains controversial. In order to solve this issue and find out the most effective treatment, we conducted this meta-analysis.

Methods:

We searched the literature, via PubMed, Embase, Scopus, Web of Science, and Cochrane to screen citations from inception to March 26, 2022, for inclusion in this study. Only clinical trials that used anti-osteoporosis treatments in postmenopausal women and displayed the alteration of body composition were included. Stata 14.0 was used for the meta-analysis.

Results:

Our meta-analysis results presented that: compared with placebo, hormone replacement therapy (HRT) was associated with increased LBM (standardized mean differences [SMD] = 0.32, 95% confidence interval [CI] = 0.02–0.61) and reduced FM (SMD = −0.30, 95% CI = −0.51 to −0.09) in postmenopausal women. Compared with placebo, physical exercise training showed an effect of decreasing FM (SMD = −0.66, 95% CI = −0.94 to −0.38) but not significant influence LBM (SMD = 1.31, 95% CI = −0.29 to 2.91). The network meta-analysis of our study showed that oral estrogen and progestogen plus exercise (OEPE) treatment might be the most effective anti-osteoporosis treatment (surface under the cumulative ranking curve 99.9) to reduce FM in postmenopausal women.

Conclusions:

anti-osteoporosis treatments, especially HRT, affect body composition. Furthermore, the combination of HRT and exercise training are the most effective treatment to reduce FM while maintaining LBM.

1. Introduction

Menopause is characterized by the disturbance of estrogen and results in metabolic disorders in postmenopausal women, including adiposity, postmenopausal osteoporosis, and other metabolic diseases.^[1,2] Adiposity is presented by the alteration in body composition, including the reduction of fat-free mass or lean body mass (LBM), and the increase of body fat mass (FM) as women age, and results in many metabolic and cardiovascular diseases.^[3,4]

Postmenopausal osteoporosis, which contributes to approximately 65% of white women aged 50 years and older suffering from at least one osteoporotic fracture,^[5] has been associated with an increased risk of adiposity due to osteoporotic pain, reduced mobility and function, fear of falling and bone fracture.^[6] Fortunately, widespread anti-osteoporosis treatments such as hormone replacement therapy (HRT), selective estrogen receptor modulator (SERM), or diphosphonate could improve bone mineral density and lower the fracture risk. Some reports also showed that HRT has beneficial effects on the regulation of fat distribution and possible mediation roles in fat accumulation.^[7–9] However, it remains unclear whether the anti-osteoporosis treatments could have positive effects on the change of body composition (LBM and FM) in postmenopausal women.

Therefore, we reviewed the current studies on the changes in FM and LBM in postmenopausal women receiving anti-osteoporosis therapy, and analyzed the effects of them on the changes in FM and LBM. We also performed a network analysis to further compare the influence on FM among different anti- osteoporosis treatments.

2. Methods

2.1. Data sources and search strategy

This study focused on the changed body composition (FM and LBM) in postmenopausal women receiving anti-osteoporosis therapy. This meta-analysis followed the preferred reporting items for systematic reviews and meta-analyses statement.^[10] Briefly, we performed a systematic electronic search in 5 databases including PubMed, Embase, Scopus, Web of Science, and Cochrane from database inception to March 26, 2022. The search strategy was related to population, diseases, intervention and outcomes, and constructed as following: (((((((((((((((Senile Osteoporoses) OR (Senile Osteoporosis)) OR (Age-Related Bone Loss)) OR (Age-Related Bone Losses)) OR (Age-Related Osteoporosis)) OR (Age Related Osteoporosis)) OR (Age-Related Osteoporoses)) OR (Perimenopausal Bone Loss)) OR (Postmenopausal Bone Losses)) OR (Post-Menopausal Osteoporoses)) OR (Post-Menopausal Osteoporosis)) OR (Postmenopausal Osteoporosis)) OR (Perimenopausal Bone Losses)) OR (Postmenopausal Bone Loss)) AND (((((((Diphosphonates) OR (Bisphosphonates)) OR (Bisphosphonate)) OR ((Estrogens) OR (Estrogen))) OR (RANK ligand inhibitor)) OR (Parathyroid hormone receptor agonist)) OR ((((SERMs) OR (Selective Estrogen Receptor Modulator)) OR (SERM)) OR (Selective Estrogen Receptor Modulators)))) AND ((((Body Fat Distribution) OR (Fat Distribution)) OR (Body Fat Patterning)) OR (((((Adipose Tissue) OR (Fatty Tissue)) OR (Fat Pad)) OR (Fat Pads)) OR (Body Fat))). Ethical approval was not required for this study since the study was conducted based on the data retrieved from previously published research.

2.2. Selection criteria

Two independent reviewers (SL and XX) first screened the title and abstract of the studies by the inclusion criteria. On the second screen, the reviewers screened the full text according to the inclusion and exclusion criteria. We also screened the references from relevant research and meta-analysis to identify additional eligible studies. The inclusion criteria showed as follows: Clinical trials were conducted in postmenopausal women to prevent or treat postmenopausal osteoporosis. The anti-osteoporosis interventions had to be recommended by the UK National Osteoporosis Guidelines Group^[11] or US National Osteoporosis Foundation.^[12] Studies should include comparisons between experimental and control groups, or between multiple groups of anti-osteoporosis treatments. Outcomes should include FM or LBM before and after the intervention, or their changes, in each group. The exclusion criteria were as follow: The study was performed in mixed-gender or pre-and postmenopausal women without separate analysis on FM or LBM. The postmenopausal women had other diseases that influence bone metabolism and body composition. The data was replicated from another study. Review articles, case reports, editorials, conference abstracts, and letters were not considered. Studies were not written in English.

2.3. Data extraction

Data extraction was performed by 2 independent reviewers (SL and XX), and a third reviewer (DS) re-evaluated the studies if there exists disagreement. We used a standardized Excel file to collect the relevant data including the following: study cohort, year, study region, sample size, treatments, follow-up, age, study type, and FM or LBM values at baseline and after the treatments. The primary outcome of the study was the alteration of FM or LBM between baseline and the end of the follow-up.

2.4. Quality assessment and risk of bias assessment

Two independent reviewers (SL and XX) assessed the quality of the included studies in duplicate using Cochrane Collaboration’s tool for randomized trials^[13] or Newcastle-Ottawa Scale for cohort study.^[14] Any discrepancies were solved by discussion including a third assessor (DS) until a consensus was reached. The risk of bias was classified as low, unclear, and high, which was summarized using Review Manager (version 5.3). Trials with a high risk of bias for >1 key domain were considered to be at high risk of bias whereas trials with a low risk of bias for all key domains were considered to be at low risk of bias; otherwise, they were considered to be at unclear risk of bias.

2.5. Data synthesis and analysis

According to the Cochrane Collaboration in the Cochrane Handbook for Systematic Reviews of Interventions 4.2.5 (Section 8.5.2.10),^[15] we calculated the missing standard deviation (SD) using pre-and-post-SDs and correlation coefficients. We calculated the correlation coefficients in both groups intervening by HRT and placebo from Wu J’s study.^[16] We chose it because the study has the most complete data and the interventions are similar to those in the other included studies.^[17–21] Next, we used the correlation coefficients and the available data to estimate the missing SDs in those studies.^[17–21] The effect size value was presented as the standardized mean differences (SMDs) combined with the 95% confidence interval (CI). The I² statistic was used to calculate the heterogeneity of the data, and I² > 50% indicated significant heterogeneity. Thus, a random-effects or fixed-effects model meta-analysis was performed according to clinical heterogeneity. Statistical significance was determined as a P value <0.05 (2-tailed). Publication bias was assessed by visually inspecting using Begg test. For those studies with 2 and more different intervention groups, the control group was proportionally split into groups for comparison against each intervention group. All statistical analyses were performed using Stata 14.0 (Stata Corp., College Station, TX).

3. Results

3.1. Study characteristics and quality assessments

The process of literature screening was presented by the preferred reporting items for systematic reviews and meta-analyses statement flowchart (Fig. 1). A total of 16 studies^[16–31] including 13 randomized controlled trials (RCTs) and 3 cohort studies^[19,22,28] were identified as inclusion studies and the main characteristics of them were displayed in Table 1. These trials were published from 1991 to 2020, including 4532 participants. Among the 16 anti-osteoporosis treatment studies, 12 studies^{[16–19,22–24,26,28–31]}used HRT alone or HRT combined with other anti-osteoporosis treatments, one used bisphosphate alone,^[27] one used SERM alone,^[25] one used calcium combined with vitamin D,^[21] and one used physical training alone.^[20] Finally, 9 studies^{[16–18,22,23,26,28,30,31]} were included in the meta-analysis of HRT on FM (3 lacked placebo group^[19,24,29]), and 8 studies^{[16–18,22,23,26,28,30]} were included in the meta-analysis of HRT on LBM (3 lacked placebo group^[19,24,29] and one lacked LBM data^[31]). All 16 studies^[16–31] were included in the network mate-analysis to compare the influences of anti-osteoporosis treatments on FM.

Figure 1.

The flow diagram of study selection.

Table 1.

Baseline characteristics of participants in included studies (n = 16).

Study cohort	Year	Study region	No.	Length (mo)	Treatment*	Age (years), Mean (SD) or range	Baseline of LBM (kg), Mean (SD)	Posttreatment change of LBM (kg), Mean (SD)	Baseline of FM (kg), Mean (SD)	Posttreatment change of FM (kg), Mean (SD)	Study characteristics
Wu J, et al	2006	Japan	33	6	Placebo	54.9 (2.9)	36.9 (3.7)	0.28 (0.09)	15.1 (4.5)	0.03 (1.02)	RCT
			31	6	Exercise	55.2 (2.8)	37.9 (4.2)	0.09 (0.86)	16.8 (4.3)	−0.57 (1.07)
			33	6	OE	53.8 (2.9)	37.1 (3.2)	0.30 (0.65)	15.0 (4.1)	−0.44 (1.32)
			31	6	OEE	54.4 (2.9)	37.5 (3.2)	0.30 (0.69)	16.1 (3.5)	−0.70 (0.97)
Figueroa A, et al	2003	US	28	12	Placebo	55.2 (0.7)	37.7 (0.8)	−0.4 (0.2)	26.3 (1.6)	1.0 (0.7)	Cohort studies
			20	12	OEE	53.7 (1.1)	39.8 (1.2)	0.6 (0.3)	27.5 (2.1)	−0.9 (0.7)
			22	12	OE	57.4 (0.9)	38.8 (0.9)	0.2 (0.4)	26.6 (1.9)	0.1 (0.8)
			24	12	Exercise	57.8 (0.9)	39.7 (1.0)	0.7 (0.2)	27.7 (1.7)	−0.6 (0.5)
Jensen L B, et al	2003	Denmark	449	60	Placebo	50.0 (2.8)	NA	0.12 (2.25)	NA	2.63 (3.81)	RCT
			448	60	OEP	49.5 (2.7)	NA	0.19 (1.82)	NA	1.93 (4.06)
Villareal D T, et al	2003	US	14	9	OEP	81 (3)	37.6 (2.0)	0.12 (0.33)	23.1 (8.2)	0.14 (0.55)	RCT
			14	9	OEPE	81 (3)	40.7 (6.4)	0.59 (0.61)	23.2 (7.4)	−2.70 (0.40)
Jacobsen D E, et al	2010	Netherlands	73	12	Placebo	73.4 (3.2)	43.2 (3.9)	0.03 (1.50)	26.1 (6.6)	−0.15 (2.00)	RCT
			70	12	Raloxifene	73.6 (3.2)	43.1 (4.2)	0.83 (2.40)	25.5 (8.0)	−0.61 (2.20)
Bea J W, et al	2011	US	337	72	Placebo	63.4 (7.1)	38.23 (5.42)	−0.40 (2.15)	32.51 (11.19)	0.43 (5.33)	RCT
			392	72	OEP	63.2 (7.2)	37.73 (5.17)	−0.29 (1.99)	32.52 (11.00)	0.64 (5.43)
Reid I R, et al	2020	Europe	878	72	Placebo	72 (5.1)	38.5 (4.2)	−1.01 (0.06)	27.6 (9.0)	−0.50 (0.16)	RCT
			916	72	Zoledronate	71 (5.0)	38.5 (4.3)	−1.19 (0.05)	27.2 (8.7)	−0.06 (0.15)
Sørensen M B, et al	2001	Europe	14	3	Placebo	55.5 (2.6)	39.0 (4.10)	−1.00 (1.58)	31.7 (9.20)	0.84 (1.34)	Cohort studies
			14	3	OEP	55.5 (2.6)	39.0 (4.10)	0.35 (0.86)	31.7 (9.20)	−0.40 (1.98)
O’Sullivan A J, et al	1998	Australia	8	24	OE	57 (1)	NA	−0.8 (0.3)	NA	1.2 (0.4)	RCT
			10	24	TEP	57 (1)	NA	0.4 (0.2)	NA	0.1 (0.5)
Kohrt W M, et al	1998	US	10	12	Calcium	68 (3)	38.9 (2.7)	0.53 (1.61)	NA	−0.11 (0.60)	RCT
			18	12	EC	66 (3)	39.4 (4.4)	0.62 (1.31)	NA	−1.29 (0.81)
			10	12	OEPC	65 (3)	38.6 (5.2)	1.22 (2.10)	NA	−0.16 (1.02)
			16	12	OEPCE	66 (4)	37.8 (4.0)	1.42 (1.41)	NA	−1.80 (1.93)
Haarbo J, et al	1991	Denmark	24	24	Placebo	45–55	46.6 (3.6)	−0.7 (0.10)†	19.5 (7.5)	0.9 (1.87) †	RCT
			19	24	OEP	45–55	43.3 (3.1)	0.6 (1.17) †	20.4 (6.5)	0.5 (2.18) †
			19	24	OEP	45–55	45.5 (5.2)	−0.2 (0.95) †	23.3 (9.4)	0.2 (3.37) †
Atkinson C, et al	2004	UK	88	12	Placebo	55.2 (4.9)	38.00 (4.12)	−0.22 (0.03) †	24.19 (6.88)	−0.67 (1.63) †	RCT
			83	12	OE	55.1 (4.7)	37.98 (4.14)	−0.18 (0.71) †	24.36 (7.89)	−0.46 (2.55) †
Milliken L A, et al	2009	US	115	48	OEPC	55.3	38.3 (4.4)	−0.5 (0.76) †	25.8 (8.7)	−0.7 (2.80) †	Cohort studies
			52	48	Calcium	57.5	38.0 (4.3)	−0.2 (0.34) †	25.4 (7.3)	0.2 (1.78) †
Silverman N E, et al	2009	US	40	6	Placebo	58 (5)	41.8 (5.0)	−0.6 (0.56) †	41.5 (10.0)	−5.4 (2.80) †	RCT
			46	6	Exercise	60 (5)	38.9 (8.3)	−0.2 (1.20) †	41.2 (5.2)	−5.1 (2.06) †
Sukumar D, et al	2011	US	21	12	Placebo	57.4 (4.7)	41.7 (5.4)	1.4 (0.86) †	36.9 (8.4)	4.5 (1.94) †	RCT
			26	12	Calcium	58.5 (4.1)	45.2 (6.4)	1.2 (1.20) †	39.6 (9.5)	4.2 (3.44) †
Paoletti A M, et al	2015	Italy	20	12	Placebo	51.6 (2.8)	NA	NA	NA	1.06 (2.13)	RCT
			40	12	OEP	50.9 (4.2)	NA	NA	NA	−0.80 (2.37)
			40	12	OEP	50.6 (4.6)	NA	NA	NA	−0.87 (2.52)

^*Abbreviations: EC, exercise plus calcium; FM = body fat mass, LBM = lean body mass, OE, oral estrogen; OEE, oral estrogen plus exercise; OEP, oral estrogen plus progestogen; OEPC, oral estrogen plus progestogen plus calcium; OEPCE, oral estrogen plus progestogen plus calcium plus exercise, OEPE, oral estrogen plus progestogen plus exercise; TEP, transdermal estrogen patches.

^†This data are estimated by calculating the correlation coefficient, which is described in the Data Synthesis section.

3.2. Risk of bias assessment

In the 9 studies^{[16–18,22,23,26,28,30,31]} included in the meta-analysis of HRT on FM, 3 studies were at high risk of bias, 4 were unclear, and 2 were at low risk of bias. Of the 8 studies on LBM,^{[16–18,22,23,26,28,30]} 3 studies were at high risk of bias, 4 were unclear, and 1 was at low risk of bias. Overall, 5 of 16 studies were at high risk of bias, 7 were unclear, and 4 were at low risk of bias (Fig. 2).

Figure 2.

Risk of bias assessments. The top shows the risk of bias graph for cohort studies (left) and risk of bias summary for included cohort studies (right). The bottom shows the risk of bias graph for RCTs (left) and risk of bias summary for RCTs (right). Review authors’ judgments about each risk of bias item presented as percentages across include studies in each bias graph. In the bias summaries: +, no bias; -, bias; ?, bias unknown. RCTs = randomized controlled trials.

3.3. Effect of anti-osteoporosis therapies on LBM

Among all the included studies, one lacked LBM data,^[31] 3 lacked a placebo group^[19,24,29] and 4 were non-HRT treatments.^{[20,21,25,27]} Raloxifene for 12 months was used as the treatment in one study,^[25] and there was no significant effect on LBM compared to the placebo group (P = .05). One was calcium and vitamin D supplementation for 12 months^[21] and showed no influence on LBM (P = .504). Four studies^{[16,20,22,30]} used physical exercise as an intervention^[20] and also present no statistical difference (SMD 1.31, 95% CI −0.29 to 2.91) (Figure S1, Supplemental Digital Content 1, http://links.lww.com/MD/H276). While, the study using zoledronate (5 mg every 18 months) for 72 months^[27] displayed a slight decrease of LBM (−1.19 ± 0.05 kg vs −1.01 ± 0.06 kg, P = .019). Across other 9 HRT studies,^{[16–19,22,23,26,28,30]} the overall estimate of random-effect analysis was SMD 0.32, 95% CI 0.02 to 0.61, with a substantial level of heterogeneity (I² = 84.3%, Q = 51.05), and was no evidence of publication bias according to the inspection of Begg test (P = .251) (Fig. 3).

Figure 3.

Forest plot for comparing physical exercise versus placebo in terms of LBM change (Top), and the Begg test for public bias of the included data (bottom). LBM = lean body mass.

3.4. Effect of anti-osteoporosis therapies on FM

Thirteen studies^{[16–18,20–23,25–28,30,31]} were included in this section (3 lacked placebo group^[19,24,29]), and 4 of which are non-HRT studies.^{[20,21,25,27]} Among them, raloxifene (P = .13)^[25] and calcium and vitamin D supplementation (P = .834)^[21] showed no statistical difference in FM changes between the intervention and placebo group. Physical exercise^{[16,20,22,30]} treatment showed an ability to reduce FM (SMD −0.66, 95% CI −0.94 to −0.38) (Figure S1, Supplemental Digital Content 1, http://links.lww.com/MD/H276). The zoledronate treatment (5 mg every 18 months)^[27] resulted in a reduced FM loss compared with placebo (−0.06 ± 0.15 kg vs −0.50 ± 0.16 kg, P = .046). Nine studies used HRT as an intervention, and the overall estimate of SMD was −0.30, 95% CI −0.51 to −0.09 (I² = 67.0%, Q = 30.34), with no evidence of publication bias according to the Begg test (P = .755). (Fig. 4).

Figure 4.

Forest plot for comparing physical exercise versus placebo in terms of FM change (Top), and the Begg test for public bias of the included data (bottom). FM = body fat mass.

3.5. Subgroup analysis

We found that HRT might have beneficial effects on LBM and FM when used as an anti-osteoporosis therapy in postmenopausal women. However, the meta-analysis on both LBM and FM showed a substantial level of heterogeneity (I² = 84.3% and I² = 67.0%, respectively). Therefore, we planned to perform subgroup analyses on them. Given that one study^[22] did not show the detailed methods of HRT, and other studies used different HRTs, it is hard to subgroup the studies by specific type and dosage of HRT. Thus, the subgroup analyses were performed in terms of study region, age, duration of follow-up, and study type. The results in LBM analyses were all presented as obvious heterogeneity or no significance (Figure S2, Supplemental Digital Content 2, http://links.lww.com/MD/H277). Meanwhile, the study region, age, and duration of follow-up subgroup analysis on FM also showed no significance (Figure S3, Supplemental Digital Content 3, http://links.lww.com/MD/H278). while, the subgroup analysis used study type on FM showed that HRT could slightly but significantly reduce FM in postmenopausal women both in cohort studies (SMD: −1.02, 95% CI −1.50 to −0.55) and RCTs (SMD: −0.10, 95% CI −0.19 to −0.01) with acceptable heterogeneity (I² = 0.0%, Q = 0.90 and I² = 48.4%, Q = 15.50, respectively) (Fig. 5).

Figure 5.

The subgroup analyses used study type to compare HRT versus placebo in terms of FM change, presented by using forest plot. HRT = hormone replacement therapy.

3.6. Network meta-analysis

Our data suggested that anti-osteoporosis treatments could have an influence on FM in postmenopausal women. Of note, there existed different anti-osteoporosis treatments, and the HRT methods also included various patterns. It prompted us to conduct a network meta-analysis to compare the effect of FM among different anti-osteoporosis treatments. A total of 16 studies^[16–31] include in the network meta-analysis. Of the 78 possible pair-wise comparisons between the 13 treatments, 18 have been present directly in one or more studies (Fig. 6). The global inconsistency (P = .205) was not significant (Figure S4, Supplemental Digital Content 4, http://links.lww.com/MD/H279). Oral estrogen and progestogen plus exercise (OEPE) (SMD −6.07, 95% CI −8.34 to −3.79), transdermal estrogen patches (TEP) (SMD −2.62, 95% CI −4.56 to −0.69), oral estrogen plus exercise (SMD −1.26, 95% CI −2.17 to −0.35) and exercise (SMD −0.77, 95% CI −1.53 to −0.01) treatments were significantly in reducing FM, and zoledronate significantly reduced the loss of FM compared with placebo (SMD 2.84, 95% CI 1.58–4.10) (Fig. 7). According to surface under the cumulative ranking curve (SUCRA), the top 5 treatments for reducing FM were OEPE (SUCRA 99.9), TEP (SUCRA 86.6), oral estrogen and progestogen plus calcium plus exercise (SUCRA 72.7), oral estrogen plus exercise (SUCRA 70.6) and exercise plus calcium (EC) (SUCRA 62.9) (Fig. 8). The funnel chart also showed that the studies are uniformly distributed in the middle and upper part of the funnel, which can be considered that the risk of small sample effect or publication bias is very small (Figure S5, Supplemental Digital Content 5, http://links.lww.com/MD/H280).

Figure 6.

Network of eligible comparisons for the multiple-treatment meta-analysis for efficacy (SMD). The width of the lines is proportional to the number of trials comparing each pair of treatments, and the size of each node is proportional to the number of randomized participants (sample size). SMDs = standardized mean differences.

Figure 7.

Efficacy of the 13 anti-osteoporosis treatments. Treatments are displayed in the blue diagonal tables. The gray tables below the blue diagonal tables shows the estimated SMD (95% CI) compared between the column-defining treatment and the row-defining treatment in terms of reducing FM. SMD higher than 0 favor the column-defining treatment. The data displayed in the gray tables above the blue diagonal tables are the opposite SMD to that in the below gray tables. CI = confidence interval, FM = body fat mass, SMDs = standardized mean differences.

Figure 8.

Cumulative rankograms: plots of the surface under the cumulative ranking curves (SUCRAs) for the efficacy of decreasing FM with various treatments in postmenopausal women. A larger SUCRA score indicates a stronger effect on decreasing FM with the intervention. FM = body fat mass, SUCRA = surface under the cumulative ranking curve.

4. Discussion

Our current analysis showed that: compared with placebo, HRT might increase LBM in postmenopausal women, while zoledronate could result in the loss of LBM. Compared with placebo, exercise and HRT could reduce FM, while zoledronate hinders the reduction of FM. OEPE might be the most effective anti-osteoporosis treatment to reduce FM in postmenopausal women.

Increased FM is associated with obesity, which resulted in most metabolic dysfunctions. The Global Burden of Disease group has also estimated that elevated BMI values were responsible for 4 million deaths in 2015, with two-thirds of this number attributed to cardiovascular disease.^[32] Of note, Osteoporosis can lead to reduced mobility,^[33,34] which contributes to the increased FM in postmenopausal women. Therefore, we conducted the meta-analyses to answer the question that whether the widespread anti-osteoporosis treatments in postmenopausal women could also benefit to the alteration of body composition (FM and LBM).

According to the UK National Osteoporosis Guidelines Group^[11] and US National Osteoporosis Foundation,^[12] bisphosphonate, RANK ligand inhibitor, HRT, SERM, and parathyroid hormone receptor agonists are the recommended anti-osteoporosis drugs. Meanwhile, vitamin D and calcium supplementation, and physical exercise are also recommended strategies for the prevention and treatment of osteoporosis. We reviewed and analyzed current studies and found that, for LBM, zoledronate might accelerate it loss while HRT might prevent it loss. However, with the obvious heterogeneity (I² = 84.3%, Q = 51.05), the effects of HRT on LBM is less confident. We also noticed an unexpected result that physical exercise failed to present a significant protective role in LBM (SMD 1.31, 95% CI −0.29 to 2.91). The possible reasons for the result might be that the training program used in these studies for postmenopausal women was low- or moderate-intensity exercise training, and the training mainly was weight-bearing aerobic exercises rather than strength and resistance exercises. For FM, the meta-analysis on HRT presented a positive effect in reducing FM, which is also consistent with our previous findings that ovariectomy-induced bone loss in mice tends to display a markedly increased body weight.^[35] Of note, the meta-analysis still showed substantial heterogeneity (I² = 67.0%, Q = 30.34). Therefore, we conducted the subgroup analyses on study region, age, duration of follow-up, and study type to find out the possible explanations. Finally, after we classified the studies into RCTs and cohort studies, the results showed positive effects in reducing FM with acceptable heterogeneity in both subgroups.

There are multiple anti-osteoporosis treatments in the studies, and the type or dosage of HRT varies. This prompted us to further compare the effect on LBM or FM between different anti-osteoporosis treatments. Considering that the results of the meta-analysis on LBM are less confident, we performed a network analysis to compare that on FM. As expected, estrogen and progesterone combined with physical training is the most effective manner of reducing FM in postmenopausal women. We also found an interesting result that TEP treatment was shown to be the second most effective intervention among the thirteen anti-osteoporosis treatments in our network analysis. However, the comparisons between TEP and other treatments were mostly indirect. Therefore, the conclusion needs more strong evidence.

5. Conclusion

In conclusion, we found that: compared with placebo, HRT might increase LBM in postmenopausal women, while zoledronate could result in the loss of LBM. Compared with placebo, exercise and HRT could reduce FM, while zoledronate inhibits the reduction of FM. OEPE might be the most effective anti-osteoporosis treatment to reduce FM in postmenopausal women. The results indicated that anti-osteoporosis treatments could affect FM and also have a possibility to change LBM, resulting in the alteration of body composition, and thus, have an influence on the metabolic health in postmenopausal women.

Author contributions

Conceptualization: Lingfeng Shi, Xiaoli Xu, Guangda Xiang, Shanshan Duan.

Data curation: Lingfeng Shi, Xiaoli Xu.

Formal analysis: Lingfeng Shi.

Funding acquisition: Shanshan Duan.

Methodology: Lingfeng Shi, Xiaoli Xu, Guangda Xiang.

Project administration: Shanshan Duan.

Software: Lingfeng Shi.

Supervision: Lingfeng Shi, Shanshan Duan.

Writing – original draft: Lingfeng Shi, Xiaoli Xu.

Writing – review & editing: Lingfeng Shi, Xiaoli Xu.