Clinical efficacy and safety of drug interventions for primary and secondary prevention of osteoporotic fractures in postmenopausal women: Network meta-analysis followed by factor and cluster analysis

Fei Wen; Hongheng Du; Liangliang Ding; Jinxi Hu; Zifeng Huang; Hua Huang; Kaikai Li; Yuxia Mo; Anyin Kuang

doi:10.1371/journal.pone.0234123

. 2020 Jun 3;15(6):e0234123. doi: 10.1371/journal.pone.0234123

Clinical efficacy and safety of drug interventions for primary and secondary prevention of osteoporotic fractures in postmenopausal women: Network meta-analysis followed by factor and cluster analysis

Fei Wen ^1,^#, Hongheng Du ^2,^#, Liangliang Ding ^1,^#, Jinxi Hu ³, Zifeng Huang ⁴, Hua Huang ⁵, Kaikai Li ⁶, Yuxia Mo ^7,^*, Anyin Kuang ^8,^*

Editor: Carlos M Isales⁹

PMCID: PMC7269244 PMID: 32492050

Abstract

We aimed to evaluate the comparative efficacy and safety of drugs respectively for primary prevention and secondary prevention of osteoporotic fractures in postmenopausal women (PMW), and to further identify the optimal intervention(s) respectively for the two groups when efficacy and safety both considered. We searched three databases. Bayesian network meta-analyses were conducted for two efficacy outcomes (vertebral fractures and nonvertebral fractures) and two safety outcomes (tolerability and acceptability) respectively in primary prevention group and secondary prevention group. We synthesized hazard ratios (HRs) and 95% confidence intervals (CIs) for nonvertebral fractures, and risk ratios (RRs) for three others. Factor and cluster analyses on surface under the cumulative ranking curve (SUCRA) values were conducted to identify the best intervention(s) with efficacy and safety both considered. The study protocol has been registered in PROSPERO. We included 57 randomized trials involving fifteen anti-osteoporotic interventions and 106320 PMW. For primary prevention, only zoledronate (once per 18 months) reduced both vertebral (RR 0.46, 95% CI 0.28–0.74) and nonvertebral (HR 0.66, 95% CI 0.51–0.85) fractures. For secondary prevention, abaloparatide, alendronate, denosumab, lasofoxifene, risedronate, romosozumab, teriparatide, and zoledronate (once per 12 months) reduced both vertebral (RRs: from 0.17 to 0.62) and nonvertebral (HRs: from 0.54 to 0.81) fractures. PTH (1–84) and abaloparatide increased withdrawal risk. Romosozumab, teriparatide, denosumab and risedronate, with the greatest composite scores, constituted the optimal cluster having both superior efficacy and superior safety. Zoledronate used at 5 mg per 18 months, with the similar safety as placebo, is the only drug intervention which has been shown to significantly reduce both vertebral and nonvertebral fractures for primary prevention of osteoporotic fractures in PMW; while romosozumab, teriparatide, denosumab, and risedronate are the optimal treatments for secondary prevention when efficacy and safety both considered. A limitation is that safety outcomes failed to consider the severity of adverse effects.

Introduction

Osteoporosis, as the most common skeletal disease, leads to 1.5 million osteoporotic fractures per year in the USA, which results in a total annual cost of $20 billion [1,2]. On the other hand, osteoporotic vertebral and nonvertebral fractures can also cause kyphosis, chronic pain, a loss of self-esteem, an increased risk of death, a dependent living situation, and poor quality of life [3]. Postmenopausal osteoporosis (PMO), as the most common form of osteoporosis, accompanies a substantial proportion of postmenopausal women (PMW) and advances with aging [4]. The number of women at risk for osteoporotic fractures and the relevant medical costs in the next decades are expected to continue to rise, due to the fact that life expectancy of women in Western countries is presently over 80 years and continually grows [4]. Thus, selecting efficacious and safe anti-osteoporotic agents to prevent osteoporotic fractures in PMW is more than important.

Although osteoporosis or osteopenia (low bone density) is defined by bone mineral density (BMD) T score measured via dual energy X-ray absorptiometry (DXA) (namely, a BMD T score at or below -2.5 indicates the former, and T score between -1 and -2.5 indicates the latter) [5], BMD is just one of various important risk factors for fragility fractures, of which the vast majority occur in osteopenic individuals (i.e., those with a BMD T-score between -1 and -2.5) [6–9]. It reveals the importance of primary prevention of osteoporotic fractures in PMW. Meanwhile, women who have developed a fragility fracture are at higher risk for secondary fractures [10–12], and prevalent vertebral fracture (PVF) is a strong factor to predict future fractures [3], which reveals the importance of secondary prevention of osteoporotic fractures in PMW. Therefore, it is meaningful for clinicians and susceptible individuals to know the optimal drug intervention(s) respectively for primary prevention and for secondary prevention of osteoporotic fractures in PMW when efficacy and safety both considered. However, the 2017 American College of Physicians (ACP) guideline [13] shows that the comparative effectiveness of different osteoporosis drugs for PMW with osteopenia or osteoporosis is unknown. Meanwhile, evidence on comparative safety in the guideline [13] is also lacking.

Due to the lack of a randomized controlled trial (RCT) comparing all drug interventions for PMO, the best way of assessing their comparative effectiveness and safety is to perform network meta-analysis using the data deriving from RCTs in which an active drug intervention was compared with placebo or active interventions were compared among them. Although there are several network meta-analyses [14–21] having reported the comparative anti-fracture efficacy of some drug treatments for PMO, none of these studies provided evidence on comparative efficacy respectively for primary prevention and for secondary prevention of PMO. Meanwhile, none of them evaluated a novel drug intervention of zoledronate intravenously used at a dosage of 5 mg per 18 months [22], while none but two [14,15] evaluated two newly approved agents of romosozumab [23] and abaloparatide [24]. Furthermore, none of them performed relevant analyses to identify the optimal drug treatment(s) when both efficacy and safety considered, while none but two [16,18] performed analysis on safety outcome.

Hence, we carried out this network meta-analysis to assess the comparative efficacy and safety of fifteen anti-osteoporotic drug interventions for PMO respectively in the primary prevention group and in the secondary prevention group using vertebral fractures and nonvertebral fractures as efficacy outcomes and using tolerability and acceptability as safety outcomes, and when possible conducted factor and cluster analysis after network meta-analysis to identify the optimal treatment(s) with both efficacy and safety considered respectively for primary prevention and for secondary prevention of osteoporotic fractures in PMW.

Methods

This network meta-analysis was carried out according to the PRISMA extension statement for network meta-analysis [25], and its protocol has been registered in Research Registry (registration number: reviewregistry749) and in PROSPERO (registration number: CRD42019139165), and is available in protocols.io (http://dx.doi.org/10.17504/protocols.io.bf9zjr76). The PRISMA checklist for this study is provided in S1 Appendix.

Search strategy

We searched through 1 February 2019 for English articles reporting RCTs which assessed the anti-fracture efficacy and/or safety of drug interventions for primary osteoporosis in PMW via the Cochrane Library, PubMed and Embase without sample size restrictions. The second search was done on 17 April 2019. The search strategy for this study was developed using text words (e.g., "postmenopausal osteoporosis", "postmenopausal osteopenia", "postmenopausal women", "osteopenia", "osteoporosis", "romosozumab", "zoledronate", "Parathyroid Hormone (1–84)", "Denosumab", "Teriparatide", "abaloparatide", "Bone Loss*", "fracture*", "controlled clinical trial", and "randomized controlled trial"), and medical subject headings (MeSHs; e.g., "Osteoporosis, Postmenopausal", "Osteoporosis", "teriparatide", "Alendronate", "Risedronic Acid", and "Raloxifene Hydrochloride"), and medical supplementary concepts (e.g., "AMG 785", "Abaloparatide", "bazedoxifene", "Lasofoxifene", and "strontium ranelate") used as substitutes for MeSHs when appropriate. S2 Appendix provides the full search strategy. Moreover, the relevant articles referred in meta-analyses previously published were assessed for eligibility and Google Scholar was also searched for related RCTs.

Inclusion and exclusion criteria

Four outcomes of interest for this study were efficacy endpoints (vertebral fractures, and nonvertebral fractures), and safety endpoints (tolerability, namely, withdrawals due to adverse events; and acceptability, namely, study discontinuation due to any cause). Fifteen active drug interventions of interest were abaloparatide (subcutaneous, 80 μg/day), teriparatide (subcutaneous, 20 μg/day), romosozumab (subcutaneous, 210 mg/month), denosumab (subcutaneous, 60 mg/6 months), parathyroid hormone (1–84) [subcutaneous, 100 μg/day; referred to as PTH (1–84)], zoledronate (intravenous, 5 mg, once per 24 months), zoledronate (intravenous, 5 mg, once per 18 months), zoledronate (intravenous, 5 mg, once per 12 months), strontium ranelate (oral, 2 g/day), lasofoxifene (oral, 0.5 mg/day), bazedoxifene (oral, 20 mg/day), raloxifene (oral, 60 mg/day), risedronate (oral, 35 mg/week or 5 mg/day), alendronate (oral, 35 or 70 mg/week, or 5 or 10 mg/day), and ibandronate (oral, 150 mg/month).

Eligible studies included for this study were RCTs: ① which lasted one or more than one year; ② in which participants were PMW with osteoporosis or osteopenia; ③ which compared at least one active intervention with placebo, or compared active interventions of interest among them; ④ in which women in each study arm were given an adjuvant therapy of vitamin D and/or calcium; and ⑤ which assessed one or more than one of the four outcomes of interest. Studies, in which participants contained PMW with secondary osteoporosis, most participants were Asians, the Caucasian proportion of participants was below 50%, or the identical data analyzed in previously-published articles were reanalyzed, were excluded.

We defined those included trials in which each participant had a BMD T score of between -1 and -2.5 and didn’t experience an osteoporotic fracture as primary prevention trials, and meanwhile defined those trials in which each participant had a BMD T score of below -2.5 and/or experienced at least an osteoporotic fracture as secondary prevention trials. If those trials which enrolled both women with prevalent osteoporotic fractures and women without prevalent osteoporotic fractures, we would try to search for and include the corresponding papers in which subgroup analyses were performed stratified by women with or without prevalent fractures. If those papers with corresponding subgroup analyses done were available, in them the data deriving from women without prevalent fractures would be used as the primary prevention data and the data deriving from women with prevalent fractures would be used as the secondary prevention data; if not available, we would defined relevant trials in accordance with baseline statistics as follows. Those trials in which PVF proportion of participants was equal to or more than 20% [26,27] or mean BMD T score was below -2.5 [28] when PVF proportion was available, or in which mean BMD T score was below -2.5 or mean age was above 62 years [26,27] when PVF proportion wasn't available would be defined as secondary prevention trials. On the contrary, other trials would be defined as primary prevention trials. BMD T score at the femoral neck (FN) measured by DXA was preferred [29]. BMD measured at the total hip or lumbar spine would be considered when BMD measured at the FN was unknown.

Study selection

Two authors independently deleted duplicated records via the Endnote X9 software, and then scanned the titles and abstracts of the remaining records for the assessment of potential eligibility, and finally reviewed full-text articles for final eligibility for inclusion. Possible disagreements on study selection would be addressed through discussion between them or by another author involved if necessary.

Data extraction and quality assessment

Two authors independently extracted the data as follows from included studies: first author name, publication year, quality assessment results, duration of drug used and RCT performed, mean age, Caucasian proportion, PVF proportion, BMD, drug name, usage and dosage of drug, sample size of each arm, and outcome data (including study-level survival data of hazard ratios (HRs) and their 95% confidence intervals (CIs) for the outcome of nonvertebral fractures, and study-level dichotomous data of total participant numbers and the numbers of participants having experienced the event of interest for the outcomes of vertebral fractures, tolerability, and acceptability).

Two independent reviewers assessed the quality of included studies based on the Jadad scale [30]. Possible total score according to this scale ranges from 0 point to 5 points. The score derives from three aspects (namely, randomization, double blinding, and withdrawals). As for randomization, a study in which an accurate randomization method was used, or in which randomization method was simply mentioned, or in which an inaccurate randomization method was used or randomization method was not mentioned, gains 2 points, or 1 point, or 0 point, respectively. Similarly, the score as for double blinding was given. As for withdrawals, a study in which the information of withdrawals and dropouts was detailedly provided, or in which not, respectively gains 1 or 0 point. Possible disagreements on data extraction or quality assessment would be addressed through discussion between them or by another author involved if necessary.

Statistical analysis

The Bayesian random-effects model [31] was used to perform pairwise meta-analysis and network meta-analysis stratified by the primary prevention data and the secondary prevention data to synthesize HRs and their 95% CIs for the outcome of nonvertebral fractures and to synthesize risk ratios (RRs) and their 95% CIs for the three other outcomes. 95% CI of effect size not including 1.0 means difference of statistical significance. We computed I² statistic for heterogeneity assessment, and built node-splitting model for inconsistency assessment. I² greater than 50% denotes substantial heterogeneity [32], and p value from test of inconsistency less than 0.05 denotes significant inconsistency [33]. We plotted comparison-adjusted funnel plots to assess publication bias [34].

We calculated the surface under the cumulative ranking curve (SUCRA) to help identify the best drug intervention [35]. However, the relative ranking for each of the four outcomes in this study might be different, which would make the choice of the optimal intervention challenging. Therefore, we tried to use the factor analysis method to reduce the four variables consisting of SUCRA values for four outcomes into two possible common factors. Then the Stata command of "clusterank"[34] was applied to hierarchical cluster analysis [36] using the data of the two common factors, and meanwhile a two-dimensional scatter plot was plotted to show the best treatment cluster which might contain one or greater than one drug intervention. The optimal number of clusters for this Stata command is chosen according to an internal cluster validation measure, named clustering gain [37]. This measure with a maximum value represents maximum intracluster similarity and minimum intercluster similarity. If Kaiser-Meyer-Olkin Measure of Sampling Adequacy was above 0.5, p value form Bartlett's Test of Sphericity was below 0.05, and cumulative proportion of variance explained by first two common factors was above 80% in the results of factor analysis, which suggested the appropriateness of factor analysis in this condition, the subsequent cluster analysis and the analysis of computing composite score based on common factors would be performed. Otherwise, the subsequent analysis wouldn't be performed.

We grouped included trials into the primary prevention group vs. the secondary prevention group to conduct all analyses for each outcome. Meanwhile, we only kept double-blind RCTs to conduct relevant analyses for the safety outcome of acceptability. Network meta-analyses were done using JAGS (version 4.3.0) and R (version 3.6.0), factor analyses were done using SPSS (version 24.0), and network plots, clustered ranking plots and comparison-adjusted funnel plots were plotted using Stata (version 15.1).

Results

Summary of included studies

A total of 367 full-text articles were reviewed for eligibility, of which 57 were included for quantitative synthesis (Fig 1). The included 57 RCTs are listed in S3 Appendix. Baseline characteristics, outcome data and quality assessment results of these studies are given in S4 Appendix. Included studies had an average Jadad score of 4.2 points (standard deviation: 1.1), and the score of most included studies (54/57) was equal to or greater than 3 points, which indicated that primary studies included in this study had high quality in general. The average duration of drug interventions used for participants was 27.5 months (standard deviation: 16.3), and the duration was in the range of 12 to 72 months. Of included studies, 19 provided the primary prevention data, 42 provided the secondary prevention data, and 4 provided both of the two kinds of data.

In included RCTs, mean age had an average value of 67.5 years (standard deviation: 5.5) in the range of 53.5 to 85.5 years; and the proportion of Caucasians had an average value of 90.9% (standard deviation: 9.9%) in the range of 60.4% to 100%. The four outcomes of interest in this study totally involved 16 drug interventions (i.e., 15 active interventions and an inactive intervention of placebo) and 106320 PMW. The primary prevention trials in total involved 29642 PMW and 10 interventions (i.e., alendronate, bazedoxifene, denosumab, ibandronate, PTH (1–84), raloxifene, risedronate, zoledronate (once per 18 months), zoledronate (once per 24 months), and placebo), while the secondary prevention trials in total involved 76678 PMW and 15 interventions (i.e., abaloparatide, alendronate, bazedoxifene, denosumab, ibandronate, lasofoxifene, PTH (1–84), raloxifene, risedronate, romosozumab, strontium ranelate, teriparatide, zoledronate (once per 12 months), zoledronate (once per 24 months), and placebo).

Fig 2A–2H are network plots of vertebral fractures (A), nonvertebral fractures (B), tolerability (C) and acceptability (D) in the primary prevention group, and vertebral fractures (E), nonvertebral fractures (F), tolerability (G) and acceptability (H) in the secondary prevention group. There is no closed loop in Fig 2D, whereas there are at least two closed loops in any other network plot. It meant that test of inconsistency was necessarily conducted for all outcomes in the two subgroups except for acceptability in the primary prevention group. In addition, placebo was most commonly compared to other interventions in each of these network plots. On the contrary, the lines among active drugs were thin, and the number of those lines was limited, which suggested the lack of head-to-head trials comparing active drugs.

Subgroup network meta-analyses

Fig 3 (active drugs compared to placebo) and S5–S8 Appendixes (active drugs compared among them) show the results of network meta-analysis based on the primary prevention data. In terms of vertebral fractures (Fig 3A and S5 Appendix), alendronate (RR 0.60, 95% CI 0.42–0.84), PTH (1–84) (RR 0.27, 95% CI 0.11–0.61), raloxifene (RR 0.57, 95% CI 0.47–0.73), and zoledronate (once per 18 months) (RR 0.46, 95% CI 0.28–0.74) reduced fracture risk compared with placebo. Meanwhile, no significant difference was found in comparisons among active drugs. In terms of nonvertebral fractures (Fig 3B and S6 Appendix), zoledronate (once per 18 months) reduced fracture risk compared with placebo (HR 0.66, 95% CI 0.51–0.85) or raloxifene (HR 0.71, 95% CI 0.54–0.94). In terms of tolerability (Fig 3C and S7 Appendix) and acceptability (Fig 3D and S8 Appendix), no significant difference was observed in all comparisons.

Fig 4 (active drugs compared to placebo) and S9–S12 Appendixes(active drugs compared among them) show the results of network meta-analysis based on the secondary prevention data. In terms of vertebral fractures (Fig 4A and S9 Appendix), abaloparatide (RR 0.17, 95% CI 0.049–0.40), alendronate (RR 0.62, 95% CI 0.51–0.78), bazedoxifene (RR 0.62, 95% CI 0.46–0.83), denosumab (RR 0.32, 95% CI 0.23–0.43), lasofoxifene (RR 0.60, 95% CI 0.45–0.79), PTH (1–84) (RR 0.44, 95% CI 0.20–0.95), raloxifene (RR 0.69, 95% CI 0.53–0.90), risedronate (RR 0.59, 95% CI 0.47–0.74), romosozumab (RR 0.30, 95% CI 0.23–0.41), strontium ranelate (RR 0.72, 95% CI 0.63–0.81), teriparatide (RR 0.26, 95% CI 0.19–0.35), and zoledronate (once per 12 months) (RR 0.31, 95% CI 0.24–0.44) reduced fracture risk compared with placebo. Meanwhile, abaloparatide, denosumab, romosozumab, teriparatide, and zoledronate (once per 12 months) were superior to alendronate, bazedoxifene, lasofoxifene, raloxifene, risedronate, and strontium ranelate; and zoledronate (once per 24 months) was inferior to abaloparatide. In terms of nonvertebral fractures (Fig 4B and S10 Appendix), abaloparatide (HR 0.54, 95% CI 0.31–0.96), alendronate (HR 0.81, 95% CI 0.70–0.94), denosumab (HR 0.80, 95% CI 0.64–0.99), lasofoxifene (HR 0.76, 95% CI 0.60–0.96), risedronate (HR 0.70, 95% CI 0.57–0.86), romosozumab (HR 0.69, 95% CI 0.56–0.84), teriparatide (HR 0.63, 95% CI 0.48–0.81), and zoledronate (once per 12 months) (HR 0.76, 95% CI 0.63–0.94) reduced fracture risk compared with placebo. Meanwhile, risedronate, romosozumab and teriparatide were superior to bazedoxifene and raloxifene; and strontium ranelate was inferior to teriparatide. In terms of tolerability (Fig 4C and S11 Appendix), abaloparatide (RR 1.7, 95% CI 1.2–2.3) and PTH (1–84) (RR 1.9, 95% CI 1.4–2.7) increased the risk of withdrawals due to adverse events compared with placebo. Meanwhile, abaloparatide and PTH (1–84) were less tolerated than alendronate, bazedoxifene, lasofoxifene, risedronate, romosozumab, and teriparatide; and PTH (1–84) was less tolerated than denosumab, ibandronate, raloxifene, strontium ranelate, and zoledronate (once per 12 months) as well. In terms of acceptability (Fig 4D and S12 Appendix), PTH (1–84) (RR 1.2, 95% CI 1.1–1.4) increased the risk of withdrawals due to any cause compared with placebo, whereas risedronate (RR 0.87, 95% CI 0.79–0.95) reduced this risk. Meanwhile, abaloparatide and PTH (1–84) were less acceptable than alendronate, bazedoxifene, raloxifene, risedronate, strontium ranelate, and teriparatide; and risedronate was more acceptable than bazedoxifene and lasofoxifene.

Assessment of inconsistency and heterogeneity

S13–S15 Appendixes provide the results from test of inconsistency for vertebral fractures (S13 Appendix), nonvertebral fractures (S14 Appendix), and tolerability (S15 Appendix) in the primary prevention group. The p values from test of inconsistency in this subgroup ranged from 0.244 to 0.993. Test of inconsistency for acceptability in this subgroup wasn’t performed due to the lack of evidence loop in this outcome network. S16–S19 Appendixes provide the results from test of inconsistency for vertebral fractures (S16 Appendix), nonvertebral fractures (S17 Appendix), tolerability (S18 Appendix), and acceptability (S19 Appendix) in the secondary prevention group. The p values from test of inconsistency in this subgroup ranged from 0.051 to 0.999. In summary, all p values from test for inconsistency in this study were greater than 0.05, which suggested there was no significant inconsistency in network meta-analysis for any outcome in the two different subgroups.

S20–S27 Appendixes provide the results of heterogeneity assessment in this network meta-analysis. No substantial heterogeneity was observed in network meta-analysis for vertebral fractures (S20 Appendix), nonvertebral fractures (S21 Appendix) and acceptability (S23 Appendix) in the primary prevention group, and for nonvertebral fractures (S25 Appendix) and tolerability (S26 Appendix) in the secondary prevention group according to the I² statistic <50%.

Although there was substantial heterogeneity (I² = 81.1%) for risedronate vs. placebo in pairwise meta-analysis of tolerability in the primary prevention group (S22 Appendix), the estimate value of RR in any primary study was not statistically significant, which was essentially consistent with the result of pooled direct estimate or network estimate as for this comparison. Although there was substantial heterogeneity (I² = 87.4%) for zoledronate (once per 12 months) vs. placebo in network meta-analysis of vertebral fractures in the secondary prevention group (S24 Appendix), the indirect estimate with wide 95% CI (representing the lack of statistical power) was invalid and the results of pooled direct estimate and network estimate were similar as for this comparison. Although there was substantial heterogeneity (I² = 55.8%) for teriparatide vs. raloxifene in network meta-analysis of acceptability in the secondary prevention group (S27 Appendix), the estimate value of RR resulting from either direct evidence or indirect evidence was not statistically significant, which was essentially consistent with the result of pooled network estimate as for this comparison. Therefore, these results of network estimate deriving from the three different comparisons were still valid despite the existence of substantial heterogeneity in these comparisons.

Factor analyses and cluster analysis

S28 Appendix presents SUCRA values of drug interventions for 4 different outcomes estimated based on the primary prevention data. The results (S29 Appendix) of factor analysis on these SUCRA values show that Kaiser-Meyer-Olkin Measure of Sampling Adequacy was equal to 0.470 (<0.5) and p value form Bartlett's Test of Sphericity was equal to 0.797 (>0.05), which suggested the data were not suitable for factor analysis. Thus, common factors extracted in this factor analysis were not used to perform cluster analysis.

S30 Appendix presents SUCRA values of drug interventions for 4 different outcomes estimated based on the secondary prevention data. The results (S31 Appendix) of factor analysis on these SUCRA values show that Kaiser-Meyer-Olkin Measure of Sampling Adequacy was equal to 0.509 (>0.5) and p value form Bartlett's Test of Sphericity was equal to 0.002 (<0.05), which suggested the data were suitable for factor analysis. Meanwhile, Factor 1 and Factor 2 totally explained 84.67% of the total variance deriving from 4 primary variables, which implied the appropriateness of the two common factors, as substitutes for 4 primary variables, used for cluster analysis. The rotated component matrix in S31 Appendix shows the correlation coefficient between Factor 1 and SUCRA for vertebral fractures or for nonvertebral fractures approximated to 1, which indicated that Factor 1 represented the anti-fracture efficacy of drug interventions. Similarly, Factor 2 represented safety. The data of Factor 1 (efficacy score) and Factor 2 (safety score) are provided in S30 Appendix. The safety score of risedronate was greater than its efficacy score, whereas the efficacy scores of romosozumab, teriparatide, denosumab, abaloparatide, and zoledronate (once per 12 months) were greater than their safety scores. The result of cluster analysis based on the two common factors is shown in Fig 5. All drug interventions were optimally clustered into 6 groups since in this condition the statistic of clustering gain had the maximum value of 14.19. The Cluster 1 consisting of romosozumab, teriparatide, denosumab, and risedronate simultaneously had superior efficacy and safety. The Cluster 2 consisting of abaloparatide and zoledronate (once per 12 months) had the similar efficacy and the inferior safety compared to Cluster 1. The four other clusters had the inferior effectiveness compared to Cluster 1 or Cluster 2, and had the non-superior safety compared to Cluster 1 in general. Additionally, composite scores computed based on Factor 1 and Factor 2 are listed in S30 Appendix, and the four drugs in Cluster 1 had the greatest composite scores.

Fig 5 — PTH (1–84), parathyroid hormone (1–84); romo, romosozumab; abal, abaloparatide; teri, teriparatide; zole24, zoledronate (5 mg per 24 months, intravenous); zole12, zoledronate (5 mg per 12 months, intravenous); deno, denosumab; stro, strontium ranelate; laso, lasofoxifene; baze, bazedoxifene; ralo, raloxifene; plac, placebo; alen, alendronate; iban, ibandronate; rise, risedronate.

Detection of publication bias

Comparison-adjusted funnel plots are provided in S32 Appendix, and all of them did not suggest dominant publication bias.

Discussion

Summary of evidence

This study is the first one in which evidence of the comparative efficacy and safety of 16 drug interventions (15 active interventions and placebo) was respectively provided for primary and secondary prevention of PMO via network meta-analysis, and this study is also the first one in which the optimal drug interventions in the condition of efficacy and safety simultaneously taken into account were identified for secondary prevention of PMO via factor analysis and cluster analysis. In summary, there are the following main findings in this study.

For secondary prevention of PMO, the cluster analysis results showed that romosozumab, teriparatide, denosumab and risedronate constituted Cluster 1 which simultaneously had superior efficacy and safety, and meanwhile the same four drug interventions had the greatest composite scores that were computed based on the results of factor analysis. It means that these four drug interventions are the optimal options for secondary prevention of osteoporotic fractures among PMW in the context of both efficacy and safety considered. It is worth mentioning that in this study the safety outcomes consisted of tolerability and acceptability, and failed to consider the severity of adverse effects. Furthermore, the factor analysis results revealed that the safety score of risedronate was greater than its efficacy score and there was a contrary case with the three other treatments. Hence, risedronate was able to enter into the optimal cluster with superior efficacy and safety, which mainly depended on its advantage in safety; whereas the other three interventions mainly depended on their advantage in efficacy. This point is also supported by the following results found by network meta-analysis in our study. Compared with placebo, romosozumab, teriparatide, and denosumab had the similar tolerability and acceptability, and risedronate had the similar tolerability and superior acceptability, which supports risedronate with advantage in safety. Although risedronate reduced the risk of vertebral fractures (RR 0.59) and nonvertebral fractures (HR 0.70) compared with placebo, romosozumab (HR 0.69) and teriparatide (HR 0.63) had greater risk reduction as for nonvertebral fractures while romosozumab (RR 0.30), teriparatide (RR 0.26) and denosumab (RR 0.32) had greater risk reduction as for vertebral fractures, which supports these three drug treatments with advantage in efficacy. These effect sizes of the four treatments for reducing vertebral and nonvertebral fractures are similar with those estimated in two prior studies [14,15] in which network meta-analyses were performed based on the combined data for primary and secondary prevention of osteoporotic fractures in PMW.

For secondary prevention of PMO, oral risedronate was observed to have superiority over the other oral bisphosphonates (i.e., alendronate, and ibandronate) in efficacy and safety in the clustered ranking plot, which was mainly because compared with alendronate or ibandronate risedronate had greater risk reduction in terms of nonvertebral fractures (vs. placebo, HRs for risedronate, alendronate and ibandronate: 0.70, 0.81 and 0.94) and withdrawals resulting from adverse events (vs. placebo, RRs for risedronate, alendronate and ibandronate: 0.97, 1.0 and 1.0) or any cause (vs. placebo, RRs for risedronate, alendronate and ibandronate: 0.87, 0.95 and 1.0) despite these differences without statistical significance. Similarly, risedronate with statistically insignificant advantage over alendronate and ibandronate in reducing nonvertebral fractures was demonstrated in several previous network meta-analyses which enrolled patients with PMO [14,19,38] or with osteoporosis of any cause [39,40], and in reducing adverse events was demonstrated in two other studies [18,41] which enrolled patients with primary osteoporosis.

For secondary prevention of PMO, abaloparatide and zoledronate (once per 12 months) constituted Cluster 2 which had the similar efficacy and the inferior safety compared to Cluster 1. Their superior efficacy derived from the ability of significantly reducing vertebral fractures (vs. placebo, RRs for abaloparatide and zoledronate: 0.17 and 0.31) and nonvertebral fractures (vs. placebo, HRs for abaloparatide and zoledronate: 0.54 and 0.76). These effect size estimates of abaloparatide and zoledronate approximate to those calculated in three previous studies [14,15,42]. Meanwhile, they significantly or insignificantly increased withdrawals deriving from adverse events (vs. placebo, abaloparatide: RR 1.7, 95% CI 1.2–2.3); zoledronate: RR 1.2, 95% CI 0.83–1.6)), which led to their inferior safety. Consistent with the results, abaloparatide was found to increase this withdrawal risk (vs. placebo: RR 1.62, 95% CI 1.15–2.27; vs. teriparatide: RR 1.44, 95% CI 1.04–2.00) in a primary study [43], while zoledronate was found to, with statistical significance, increase the risk of adverse events in two secondary studies [16,18].

For primary prevention of PMO, oral bisphosphonates (i.e., alendronate, risedronate and ibandronate) were not observed to reduce vertebral and nonvertebral fractures compared to placebo except that alendronate reduced vertebral fractures (RR 0.60, 95% CI 0.42–0.84) in our study. Similar results were found in two Cochrane systematic reviews [26,27] in which in terms of primary prevention of osteoporotic fractures in PMW alendronate showed a significant reduction in vertebral fractures (RR 0.55, 95% CI 0.38–0.80) but not in nonvertebral fractures (RR 0.89, 95% CI 0.76–1.04) [26] while risedronate didn't show a significant reduction in vertebral (RR 0.97, 95% CI 0.42–2.25) and nonvertebral fractures (RR 0.81, 95% CI 0.25–2.58) [27]. In the primary prevention group of this study, zoledronate (once per 18 months) was the only drug intervention which reduced the risk of both vertebral fractures (RR 0.46, 95% CI 0.28–0.74) and nonvertebral fractures (HR 0.66, 95% CI 0.51–0.85), and meanwhile didn't increase withdrawal risk. This evidence on zoledronate used at a dose of 5 mg at 18-month intervals derives from a recently-published well-designed RCT [22] which mainly focused on PMW with osteopenia and without PVF, and in which compared to placebo zoledronate led to an increase in BMD and a decrease in markers of bone turnover, apart from having lowered fracture risk.

Clinical implications

For secondary prevention of osteoporotic fractures in PMW, when both anti-fracture efficacy and safety is taken into consideration our study supports that the best drug treatments are romosozumab, teriparatide, denosumab and risedronate, of which risedronate has greater safety superiority, and the others have greater efficacy superiority. It implies that risedronate is the optimal drug intervention for PMW with mild to moderate osteoporosis, while the three others are the optimal drug interventions for PMW with severe osteoporosis or at high risk for developing fragility fractures. In clinical reality, anabolic agents are preferred in severe osteoporosis patients due to their capability of reducing fracture risk more quickly and to a greater extent than antiresorptive agents [44]. However, fractures and bone loss may rapidly occur after bone-building drugs such as teriparatide are discontinued [45,46]. For preventing this reversal effect, a sequential treatment strategy, starting with a bone-forming agent, followed by an antiresorptive, is recommended for those patients [1,47]. Nowadays, there are five sequential treatments by which antiresorptives are able to maintain or further increase BMD gains obtained by anabolic drugs, and they are respectively 1 year of alendronate switching from 1 year of PTH (1–84) [48], 2 years of denosumab switching from 2 years of teriparatide [49], 2 years of alendronate switching from 1 year of romosozumab [50], 1 year of denosumab switching from 1 year of romosozumab [51], and 24 months of alendronate switching from18 months of abaloparatide [52]. Which sequential treatment is the best remains unclear, and the findings in our study can help address this issue. The two treatment strategies starting with PTH (1–84) or abaloparatide shouldn’t be considered at first owing to the poor safety and anti-fracture efficacy with PTH (1–84) and the poor safety with abaloparatide found in our study. On the contrary, the treatment strategies starting with teriparatide or romosozumab should be preferred owing to their superior anti-fracture efficacy and safety found in our study. Based on RCTs, evidence about teriparatide followed by denosumab with significant risk reduction in fractures is lacking [49], whereas there is clear evidence that the sequential treatment starting with romosozumab then switching to alendronate or denosumab has significantly reduced fracture risk [50,51]. Besides, the combination of romosozumab and bisphosphonates (e.g., alendronate) might reduce the risk of cardiovascular adverse events [53]. Accordingly, the sequential treatment starting with romosozumab followed by alendronate could be the best one for PMW with severe osteoporosis or at high risk for osteoporotic fractures.

For secondary prevention of mild to moderate osteoporosis, oral bisphosphonates are often recommended as first-line agents. There are some osteoporotic patients for whom oral bisphosphonates are not enough effective, such as, fragility fractures occurring during the treatment, and BMD having an unsatisfactory response. Among this population denosumab was demonstrated to lead to greater increase in BMD than bisphosphonates did in Lyu et al.’s study [54], in which, however, evidence on the comparative anti-fracture effectiveness was limited. Our study demonstrated that a significantly greater risk reduction in vertebral fractures derived from denosumab than derived from oral bisphosphonates (vs. alendronate, RR 0.51; vs. risedronate, RR 0.53), which, to some degree, further supports denosumab used in osteoporotic PMW on which a bisphosphonate hasn’t resulted in a satisfactory effect. Similar with teriparatide, there are reverse effects on bone markers, BMD and fracture risk after denosumab discontinued [46,55]. Thus, other agents, mainly bisphosphonates, should be immediately given in this case.

For primary prevention of osteoporotic fractures in PMW, oral bisphosphonates are widely used in clinical practice. However, there hasn’t been evidence deriving from RCTs that a bisphosphonate could significantly reduce both vertebral and nonvertebral fractures until now. Fortunately, we have identified the newest evidence on zoledronate intravenously used at a dose of 5 mg per 18 months, which is the only drug intervention which has been shown to significantly reduce both vertebral (RR 0.46) and nonvertebral fracture risk (HR 0.66) for primary prevention of PMO, and which will help improve the management of PMO in terms of primary prevention due to its superior anti-fracture efficacy and good safety profile.

For primary or secondary prevention of osteoporotic fractures in PMW, although the evidence of vitamin D and calcium used for the treatment of fractures remains limited their supplements should be considered, given that we included those trials in which all women received calcium and/or vitamin D, and that the supplements are recommended for adults with no known osteoporosis or vitamin D deficiency to prevent fractures [56]. Meanwhile, other important points (e.g., drug compliance, administration route, cost effectiveness, patient comorbidities, and a clinician’s familiarity with use of a particular treatment) should be also considered in a particular clinical setting, apart from anti-fracture efficacy and safety assessed in our study. Moreover, we included those RCTs which mainly enrolled Caucasian women, which implies evidence on the comparative efficacy and safety of osteoporosis drugs identified in this study is most applicable to white PMW.

Strengths and limitations

The greatest strength of our study is that we assessed the comparative efficacy and safety of different agent interventions for PMO respectively in the primary prevention group and in the secondary prevention group, and identified the optimal treatments for secondary prevention in consideration of both effectiveness and safety via factor and cluster analysis on the data of SUCRA values deriving from network meta-analysis. Meanwhile, we synthesized the up-to-date evidence deriving from newly-published RCTs [22,43,50,51,57] mainly involving two newly approved drugs (namely, romosozumab [23] and abaloparatide [24]) and a new drug intervention (namely, zoledronate 5 mg per 18 months), which could become a strong basis of updating the 2017 ACP guideline [13]. Second, no significant inconsistency existed in all network meta-analyses in this study, and no substantial heterogeneity existed in the majority of network meta-analyses while in the minority of them existed the heterogeneity that didn’t affect the validity of network estimates. Meanwhile, RCTs included in our study generally had high quality, and no dominant publication bias was found. These facts resulted in the validity of this network meta-analysis. Third, we improved the accuracy of network meta-analysis results by doing the following things: ① an identical drug used at different intervals was taken for different interventions, such as zoledronate (once per 12 months), zoledronate (once per 18 months) and zoledronate (once per 24 months) were considered as three interventions in this study, although alendronate used at a dose of 5 or 10 mg was taken for one intervention because the anti-fracture efficacy of these two dosages was combined in included trials [58,59]; ② we only kept double-blind RCTs for the assessment of comparative acceptability; ③ we used the fixed-effects model to perform network meta-analyses for vertebral fractures and nonvertebral fractures in the primary prevention group, given the presence of only low or no inconsistency and heterogeneity in those analyses; and ④ we synthesized HRs (instead of RRs) for nonvertebral fractures since primary studies had provided study-level survival data for this outcome and survival data contains more information (i.e., the time when an event of interest occurs) than dichotomous data does. Fourth, we further improved the network transitivity in this study compared to prior network meta-analyses [14,15,42] by only considering those trials which mainly enrolled white people given that the majority of participants were Caucasians in those trials of new drug interventions (e.g., zoledronate used once per 18 months [22], romosozumab [50,51], and abaloparatide [43]), by only considering those trials in which all participants were women with primary osteoporosis since those new drug trials [22,43,50,51] completely enrolled this type of patients, and by grouping all trials into the primary prevention subgroup and the secondary prevention subgroup to increase the comparability of baseline characteristics inside each subgroup.

This study has some limitations. First, for primary prevention of osteoporotic fractures, we failed to find evidence about the efficacy of three bone-building drugs (i.e., abaloparatide, teriparatide and romosozumab), and failed to observe denosumab with statistically significant efficacy since its wide 95% CIs across 1.0 suggested the lack of statistical power. Given that these four drugs were observed to have superior effectiveness for secondary prevention of osteoporotic fractures, future studies with adequate statistical power are needed to clarify whether it is possible to use them for primary prevention by lowering their dosages or by prolonging their dose intervals, such as zoledronate used once per 12 months is effective for secondary prevention while this drug used once per 18 months is effective for primary prevention. Second, we failed to perform factor and cluster analysis to synthesize evidence of efficacy and safety to help identify the optimal interventions for primary prevention owing to the data limited. Instead, we identified that the only drug intervention of zoledronate (once per 18 months) reduced the risk of both vertebral and nonvertebral fractures and had the similar safety as placebo. Third, calcitonin and hormone replacement therapy were not considered in this study owing to the fact that the former is not widely used for treatment of osteoporosis any longer [13] and the latter has a limited indication for osteoporosis treatment [60,61]. Although strontium ranelate is not any more in clinical use in most countries, we still assessed it in this study since this drug was also considered in at least one of recently-published network meta-analyses [14,15,42]. Other drugs [i.e., lasofoxifene and PTH (1–84)] not approved by the U.S. Food and Drug Administration (FDA) were assessed in this study for the same reason. Fourth, although the cluster analysis in the paper ensured maximum intracluster similarity and minimum intercluster similarity according to the maximum of clustering gain, it failed to tell us whether the difference among various clusters was statistically significant or not. Thus, the statistical significance of difference among active drugs needs to be identified by more head-to-head trials. Fifth, we used tolerability and acceptability as safety endpoints, but failed to perform analysis on any of specific adverse events which we should be cautious of in clinical reality, such as romosozumab causing cerebrovascular and cardiovascular events [62], abaloparatide and teriparatide causing hypercalcemia [24], denosumab or bisphosphonates causing atypical femoral fractures (AFF) and osteonecrosis of the jaw (ONJ) [55], and bisphosphonates causing gastrointestinal side effects [55]. The two safety outcomes failed to consider the severity of adverse effects. Thus, it is meaningful that future studies assess specific adverse events and serious adverse effects. Sixth, this study failed to compare different sequential treatments owing to the absence of head-to-head trials assessing their anti-fracture efficacy. Given that an anabolic or an antiresorptive agent has the limited ability to reduce fracture risk and has the limited usage time (e.g., anabolics restricted to 2 years, and bisphosphonates restricted to 3 to 5 years) [1], and that the discontinuation of some osteoporosis drugs (e.g., teriparatide, and denosumab) may immediately cause bone loss and fracture occurrence [45,46,55], appropriate sequential treatments are considered to be essential for the long-term management of PMO [1,44,47]. Therefore, further studies are needed on which one is the optimal sequential treatment and how is the optimal timing and dosing of anti-osteoporotic agents used as sequential therapy to maintain the maximum decrease in fracture risk. Seventh, since there was a limited number of those papers with subgroup analyses done stratified by women with or without prevalent fractures, a minority of PMW grouped into the primary prevention group were women with prevalent fractures and a minority of PMW grouped into the secondary prevention group were women without prevalent fractures. Thus, the relative efficacy of drugs severally for primary prevention and secondary prevention identified in this study only approximates to the real situation, and accordingly, it requires to be validated by analyzing individual patient data. Eighth, further studies should use evidence at different time to stratify to compare different anti-osteoporotic interventions in order to provide more accurate estimates for differences between short-term interventions (e.g., romosozumab, abaloparatide and teriparatide) and chronic interventions (e.g., oral bisphosphonates and denosumab). Ninth, we failed to consider including hip fractures as an endpoint of interest; thus, it is meaningful that further studies compare anti-osteoporosis drugs in preventing hip fractures. Last, SUCRA does not consider the magnitude of differences in effects between treatments, and fails to tell us whether the differences in the rankings based on SUCRA are with statistical significance.

Conclusions

Based on current evidences, zoledronate intravenously used at a dose of 5 mg per 18 months, with the similar safety as placebo, is the only drug intervention which has been shown to significantly reduce both vertebral and nonvertebral fracture risk for primary prevention of osteoporotic fractures in PMW; while romosozumab, teriparatide, denosumab, and risedronate are the optimal treatments for secondary prevention when anti-fracture efficacy and safety both considered. Romosozumab followed by alendronate could be the best sequential treatment for PMW with severe osteoporosis or at high risk for osteoporotic fractures. Future studies are needed to clarify whether it is possible to use bone-building drugs for primary prevention by lowering their dosages or prolonging their dose intervals, and to identify the optimal sequential treatment and the optimal timing and dosing of sequential treatments.

Supporting information

S1 Appendix. PRISMA checklist.

(DOC)

Click here for additional data file.^{(64.5KB, doc)}

S2 Appendix. Full search strategy.

(PDF)

Click here for additional data file.^{(117.7KB, pdf)}

S3 Appendix. Checklist of studies included for network meta-analysis.

(PDF)

Click here for additional data file.^{(67.5KB, pdf)}

S4 Appendix. Study characteristics, outcome data and quality assessment.

(XLSX)

Click here for additional data file.^{(24.5KB, xlsx)}

S5 Appendix. Network meta-analysis for vertebral fractures in the primary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(29.7KB, pdf)}

S6 Appendix. Network meta-analysis for nonvertebral fractures in the primary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(55.3KB, pdf)}

S7 Appendix. Network meta-analysis for tolerability in the primary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(57.6KB, pdf)}

S8 Appendix. Network meta-analysis for acceptability in the primary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(56.2KB, pdf)}

S9 Appendix. Network meta-analysis for vertebral fractures in the secondary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(164.6KB, pdf)}

S10 Appendix. Network meta-analysis for nonvertebral fractures in the secondary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(138.1KB, pdf)}

S11 Appendix. Network meta-analysis for tolerability in the secondary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(111.9KB, pdf)}

S12 Appendix. Network meta-analysis for acceptability in the secondary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(139.1KB, pdf)}

S13 Appendix. Inconsistency test for vertebral fractures in the primary prevention group.

(PDF)

Click here for additional data file.^{(31.1KB, pdf)}

S14 Appendix. Inconsistency test for nonvertebral fractures in the primary prevention group.

(PDF)

Click here for additional data file.^{(57.3KB, pdf)}

S15 Appendix. Inconsistency test for tolerability in the primary prevention group.

(PDF)

Click here for additional data file.^{(31KB, pdf)}

S16 Appendix. Inconsistency test for vertebral fractures in the secondary prevention group.

(PDF)

Click here for additional data file.^{(87.2KB, pdf)}

S17 Appendix. Inconsistency test for nonvertebral fractures in the secondary prevention group.

(PDF)

Click here for additional data file.^{(88.3KB, pdf)}

S18 Appendix. Inconsistency test for tolerability in the secondary prevention group.

(PDF)

Click here for additional data file.^{(89.2KB, pdf)}

S19 Appendix. Inconsistency test for acceptability in the secondary prevention group.

(PDF)

Click here for additional data file.^{(89.3KB, pdf)}

S20 Appendix. Heterogeneity assessment for vertebral fractures in the primary prevention group.

(PDF)

Click here for additional data file.^{(63.2KB, pdf)}

S21 Appendix. Heterogeneity assessment for nonvertebral fractures in the primary prevention group.

(PDF)

Click here for additional data file.^{(93.8KB, pdf)}

S22 Appendix. Heterogeneity assessment for tolerability in the primary prevention group.

(PDF)

Click here for additional data file.^{(93.4KB, pdf)}

S23 Appendix. Heterogeneity assessment for acceptability in the primary prevention group.

(PDF)

Click here for additional data file.^{(61.6KB, pdf)}

S24 Appendix. Heterogeneity assessment for vertebral fractures in the secondary prevention group.

(PDF)

Click here for additional data file.^{(179.5KB, pdf)}

S25 Appendix. Heterogeneity assessment for nonvertebral fractures in the secondary prevention group.

(PDF)

Click here for additional data file.^{(186.2KB, pdf)}

S26 Appendix. Heterogeneity assessment for tolerability in the secondary prevention group.

(PDF)

Click here for additional data file.^{(180.4KB, pdf)}

S27 Appendix. Heterogeneity assessment for acceptability in the secondary prevention group.

(PDF)

Click here for additional data file.^{(181.1KB, pdf)}

S28 Appendix. SUCRA values of drug interventions for primary prevention of osteoporotic fractures.

(XLSX)

Click here for additional data file.^{(9.1KB, xlsx)}

S29 Appendix. Factor analysis of SUCRA values estimated based on the primary prevention data.

(PDF)

Click here for additional data file.^{(58.4KB, pdf)}

S30 Appendix. SUCRA values of drug interventions for secondary prevention of osteoporotic fractures.

(XLSX)

Click here for additional data file.^{(10.6KB, xlsx)}

S31 Appendix. Factor analysis of SUCRA values estimated based on the secondary prevention data.

(PDF)

Click here for additional data file.^{(58.4KB, pdf)}

S32 Appendix. Comparison-adjusted funnel plots.

(PDF)

Click here for additional data file.^{(262.9KB, pdf)}

Abbreviations

PMW: postmenopausal women
PMO: postmenopausal osteoporosis
PVF: prevalent vertebral fracture
BMD: bone mineral density
PTH (1–84): parathyroid hormone (1–84)
RCT: randomized controlled trial
SUCRA: surface under the cumulative ranking curve
RR: risk ratio
HR: hazard ratio
CI: confidence interval
FDA: Food and Drug Administration
ACP: American College of Physicians

Data Availability

All relevant data are within the manuscript and its Supporting Information files.

Funding Statement

No specific funding was received for this study

References

1.Lorentzon M. Treating osteoporosis to prevent fractures: current concepts and future developments. J INTERN MED. 2019; 285(4):381–94. 10.1111/joim.12873 [DOI] [PubMed] [Google Scholar]
2.Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. LANCET. 2002; 359(9319):1761–7. 10.1016/S0140-6736(02)08657-9 [DOI] [PubMed] [Google Scholar]
3.Black DM, Rosen CJ. Clinical Practice. Postmenopausal Osteoporosis. N Engl J Med. 2016; 374(3):254–62. 10.1056/NEJMcp1513724 [DOI] [PubMed] [Google Scholar]
4.Hernlund E, Svedbom A, Ivergard M, Compston J, Cooper C, Stenmark J, et al. Osteoporosis in the European Union: medical management, epidemiology and economic burden. A report prepared in collaboration with the International Osteoporosis Foundation (IOF) and the European Federation of Pharmaceutical Industry Associations (EFPIA). ARCH OSTEOPOROS. 2013; 8(136. 10.1007/s11657-013-0136-1 [DOI] [PMC free article] [PubMed]
5.Cosman F, de Beur SJ, LeBoff MS, Lewiecki EM, Tanner B, Randall S, et al. Clinician's Guide to Prevention and Treatment of Osteoporosis. Osteoporos Int. 2014; 25(10):2359–81. 10.1007/s00198-014-2794-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Trajanoska K, Schoufour JD, de Jonge E, Kieboom B, Mulder M, Stricker BH, et al. Fracture incidence and secular trends between 1989 and 2013 in a population based cohort: The Rotterdam Study. BONE. 2018; 114(116–24. 10.1016/j.bone.2018.06.004 [DOI] [PubMed] [Google Scholar]
7.Siris ES, Brenneman SK, Miller PD, Barrett-Connor E, Chen YT, Sherwood LM, et al. Predictive value of low BMD for 1-year fracture outcomes is similar for postmenopausal women ages 50–64 and 65 and Older: results from the National Osteoporosis Risk Assessment (NORA). J BONE MINER RES. 2004; 19(8):1215–20. 10.1359/JBMR.040508 [DOI] [PubMed] [Google Scholar]
8.Schuit SC, van der Klift M, Weel AE, de Laet CE, Burger H, Seeman E, et al. Fracture incidence and association with bone mineral density in elderly men and women: the Rotterdam Study. BONE. 2004; 34(1):195–202. 10.1016/j.bone.2003.10.001 [DOI] [PubMed] [Google Scholar]
9.Siris ES, Chen YT, Abbott TA, Barrett-Connor E, Miller PD, Wehren LE, et al. Bone mineral density thresholds for pharmacological intervention to prevent fractures. Arch Intern Med. 2004; 164(10):1108–12. 10.1001/archinte.164.10.1108 [DOI] [PubMed] [Google Scholar]
10.Saito T, Sterbenz JM, Malay S, Zhong L, MacEachern MP, Chung KC. Effectiveness of anti-osteoporotic drugs to prevent secondary fragility fractures: systematic review and meta-analysis. Osteoporos Int. 2017; 28(12):3289–300. 10.1007/s00198-017-4175-0 [DOI] [PubMed] [Google Scholar]
11.Center JR, Bliuc D, Nguyen TV, Eisman JA. Risk of subsequent fracture after low-trauma fracture in men and women. JAMA. 2007; 297(4):387–94. 10.1001/jama.297.4.387 [DOI] [PubMed] [Google Scholar]
12.Johnell O, Kanis JA, Oden A, Sernbo I, Redlund-Johnell I, Petterson C, et al. Fracture risk following an osteoporotic fracture. Osteoporos Int. 2004; 15(3):175–9. 10.1007/s00198-003-1514-0 [DOI] [PubMed] [Google Scholar]
13.Qaseem A, Forciea MA, McLean RM, Denberg TD. Treatment of Low Bone Density or Osteoporosis to Prevent Fractures in Men and Women: A Clinical Practice Guideline Update From the American College of Physicians. ANN INTERN MED. 2017; 166(11):818–39. 10.7326/M15-1361 [DOI] [PubMed] [Google Scholar]
14.Barrionuevo P, Kapoor E, Asi N, Alahdab F, Mohammed K, Benkhadra K, et al. Efficacy of Pharmacological Therapies for the Prevention of Fractures in Postmenopausal Women: A Network Meta-Analysis. J Clin Endocrinol Metab. 2019; 104(5):1623–30. 10.1210/jc.2019-00192 [DOI] [PubMed] [Google Scholar]
15.Reginster J, Bianic F, Campbell R, Martin M, Williams SA, Fitzpatrick LA. Abaloparatide for risk reduction of nonvertebral and vertebral fractures in postmenopausal women with osteoporosis: a network meta-analysis. Osteoporos Int. 2019; 30(7):1465–73. 10.1007/s00198-019-04947-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Yang XC, Deng ZH, Wen T, Luo W, Xiao WF, Zhao RB, et al. Network Meta-Analysis of Pharmacological Agents for Osteoporosis Treatment and Fracture Prevention. CELL PHYSIOL BIOCHEM. 2016; 40(3–4):781–95. 10.1159/000453138 [DOI] [PubMed] [Google Scholar]
17.Ellis AG, Reginster JY, Luo X, Cappelleri JC, Chines A, Sutradhar S, et al. Bazedoxifene versus oral bisphosphonates for the prevention of nonvertebral fractures in postmenopausal women with osteoporosis at higher risk of fracture: a network meta-analysis. VALUE HEALTH. 2014; 17(4):424–32. 10.1016/j.jval.2014.01.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Tan X, Wen F, Yang W, Xie JY, Ding LL, Mo YX. Comparative efficacy and safety of pharmacological interventions for osteoporosis in postmenopausal women: a network meta-analysis (Chongqing, China). MENOPAUSE. 2019; 26(8):929–39. 10.1097/GME.0000000000001321 [DOI] [PubMed] [Google Scholar]
19.Freemantle N, Cooper C, Diez-Perez A, Gitlin M, Radcliffe H, Shepherd S, et al. Results of indirect and mixed treatment comparison of fracture efficacy for osteoporosis treatments: a meta-analysis. Osteoporos Int. 2013; 24(1):209–17. 10.1007/s00198-012-2068-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Zhang L, Pang Y, Shi Y, Xu M, Xu X, Zhang J, et al. Indirect comparison of teriparatide, denosumab, and oral bisphosphonates for the prevention of vertebral and nonvertebral fractures in postmenopausal women with osteoporosis. MENOPAUSE. 2015; 22(9):1021–5. 10.1097/GME.0000000000000466 [DOI] [PubMed] [Google Scholar]
21.Migliore A, Broccoli S, Massafra U, Cassol M, Frediani B. Ranking antireabsorptive agents to prevent vertebral fractures in postmenopausal osteoporosis by mixed treatment comparison meta-analysis. Eur Rev Med Pharmacol Sci. 2013; 17(5):658–67. [PubMed] [Google Scholar]
22.Reid IR, Horne AM, Mihov B, Stewart A, Garratt E, Wong S, et al. Fracture Prevention with Zoledronate in Older Women with Osteopenia. N Engl J Med. 2018; 379(25):2407–16. 10.1056/NEJMoa1808082 [DOI] [PubMed] [Google Scholar]
23.Markham A. Romosozumab: First Global Approval. DRUGS. 2019; 79(4):471–6. 10.1007/s40265-019-01072-6 [DOI] [PubMed] [Google Scholar]
24.Rachner TD, Hofbauer LC, Gobel A, Tsourdi E. Novel therapies in osteoporosis: PTH-related peptide analogs and inhibitors of sclerostin. J MOL ENDOCRINOL. 2019; 62(2):R145–54. 10.1530/JME-18-0173 [DOI] [PubMed] [Google Scholar]
25.Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. ANN INTERN MED. 2015; 162(11):777–84. 10.7326/M14-2385 [DOI] [PubMed] [Google Scholar]
26.Wells GA, Cranney A, Peterson J, Boucher M, Shea B, Robinson V, et al. Alendronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database Syst Rev. 2008;1):D1155 10.1002/14651858.CD001155.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Wells G, Cranney A, Peterson J, Boucher M, Shea B, Robinson V, et al. Risedronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database Syst Rev. 2008;1):D4523 10.1002/14651858.CD004523.pub3 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Cosman F, de Beur SJ, LeBoff MS, Lewiecki EM, Tanner B, Randall S, et al. Clinician's Guide to Prevention and Treatment of Osteoporosis. Osteoporos Int. 2014; 25(10):2359–81. 10.1007/s00198-014-2794-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Siris ES, Simon JA, Barton IP, McClung MR, Grauer A. Effects of risedronate on fracture risk in postmenopausal women with osteopenia. Osteoporos Int. 2008; 19(5):681–6. 10.1007/s00198-007-0493-y [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials. 1996; 17(1):1–12. 10.1016/0197-2456(95)00134-4 [DOI] [PubMed] [Google Scholar]
31.Dias S, Sutton AJ, Ades AE, Welton NJ. Evidence synthesis for decision making 2: a generalized linear modeling framework for pairwise and network meta-analysis of randomized controlled trials. MED DECIS MAKING. 2013; 33(5):607–17. 10.1177/0272989X12458724 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003; 327(7414):557–60. 10.1136/bmj.327.7414.557 [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Dias S, Welton NJ, Caldwell DM, Ades AE. Checking consistency in mixed treatment comparison meta-analysis. STAT MED. 2010; 29(7–8):932–44. 10.1002/sim.3767 [DOI] [PubMed] [Google Scholar]
34.Chaimani A, Higgins JP, Mavridis D, Spyridonos P, Salanti G. Graphical tools for network meta-analysis in STATA. PLOS ONE. 2013; 8(10):e76654 10.1371/journal.pone.0076654 [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial. J CLIN EPIDEMIOL. 2011; 64(2):163–71. 10.1016/j.jclinepi.2010.03.016 [DOI] [PubMed] [Google Scholar]
36.Handl J, Knowles J, Kell DB. Computational cluster validation in post-genomic data analysis. BIOINFORMATICS. 2005; 21(15):3201–12. 10.1093/bioinformatics/bti517 [DOI] [PubMed] [Google Scholar]
37.Jung Y, Park H, Du D, Drake B. A Decision Criterion for the Optimal Number of Clusters in Hierarchical Clustering. J GLOBAL OPTIM. 2003; 25(1):91–111. [Google Scholar]
38.Zhou J, Ma X, Wang T, Zhai S. Comparative efficacy of bisphosphonates in short-term fracture prevention for primary osteoporosis: a systematic review with network meta-analyses. Osteoporos Int. 2016; 27(11):3289–300. 10.1007/s00198-016-3654-z [DOI] [PubMed] [Google Scholar]
39.Sanderson J, Martyn-St JM, Stevens J, Goka E, Wong R, Campbell F, et al. Clinical effectiveness of bisphosphonates for the prevention of fragility fractures: A systematic review and network meta-analysis. BONE. 2016; 89(52–8. 10.1016/j.bone.2016.05.013 [DOI] [PubMed] [Google Scholar]
40.Murad MH, Drake MT, Mullan RJ, Mauck KF, Stuart LM, Lane MA, et al. Clinical review. Comparative effectiveness of drug treatments to prevent fragility fractures: a systematic review and network meta-analysis. J Clin Endocrinol Metab. 2012; 97(6):1871–80. 10.1210/jc.2011-3060 [DOI] [PubMed] [Google Scholar]
41.Tadrous M, Wong L, Mamdani MM, Juurlink DN, Krahn MD, Levesque LE, et al. Comparative gastrointestinal safety of bisphosphonates in primary osteoporosis: a network meta-analysis. Osteoporos Int. 2014; 25(4):1225–35. 10.1007/s00198-013-2576-2 [DOI] [PubMed] [Google Scholar]
42.Yang L, Kang N, Yang JC, Su QJ, Liu YZ, Guan L, et al. Drug efficacies on bone mineral density and fracture rate for the treatment of postmenopausal osteoporosis: a network meta-analysis. Eur Rev Med Pharmacol Sci. 2019; 23(6):2640–68. 10.26355/eurrev_201903_17414 [DOI] [PubMed] [Google Scholar]
43.Miller PD, Hattersley G, Riis BJ, Williams GC, Lau E, Russo LA, et al. Effect of Abaloparatide vs Placebo on New Vertebral Fractures in Postmenopausal Women With Osteoporosis: A Randomized Clinical Trial. JAMA. 2016; 316(7):722–33. 10.1001/jama.2016.11136 [DOI] [PubMed] [Google Scholar]
44.Cosman F. Long-term treatment strategies for postmenopausal osteoporosis. CURR OPIN RHEUMATOL. 2018; 30(4):420–6. 10.1097/BOR.0000000000000509 [DOI] [PubMed] [Google Scholar]
45.Leder BZ, Neer RM, Wyland JJ, Lee HW, Burnett-Bowie SM, Finkelstein JS. Effects of teriparatide treatment and discontinuation in postmenopausal women and eugonadal men with osteoporosis. J Clin Endocrinol Metab. 2009; 94(8):2915–21. 10.1210/jc.2008-2630 [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Buckley L, Humphrey MB. Glucocorticoid-Induced Osteoporosis. N Engl J Med. 2018; 379(26):2547–56. 10.1056/NEJMcp1800214 [DOI] [PubMed] [Google Scholar]
47.Cosman F, Nieves JW, Dempster DW. Treatment Sequence Matters: Anabolic and Antiresorptive Therapy for Osteoporosis. J BONE MINER RES. 2017; 32(2):198–202. 10.1002/jbmr.3051 [DOI] [PubMed] [Google Scholar]
48.Black DM, Bilezikian JP, Ensrud KE, Greenspan SL, Palermo L, Hue T, et al. One year of alendronate after one year of parathyroid hormone (1–84) for osteoporosis. N Engl J Med. 2005; 353(6):555–65. 10.1056/NEJMoa050336 [DOI] [PubMed] [Google Scholar]
49.Leder BZ, Tsai JN, Uihlein AV, Wallace PM, Lee H, Neer RM, et al. Denosumab and teriparatide transitions in postmenopausal osteoporosis (the DATA-Switch study): extension of a randomised controlled trial. LANCET. 2015; 386(9999):1147–55. 10.1016/S0140-6736(15)61120-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Saag KG, Petersen J, Brandi ML, Karaplis AC, Lorentzon M, Thomas T, et al. Romosozumab or Alendronate for Fracture Prevention in Women with Osteoporosis. N Engl J Med. 2017; 377(15):1417–27. 10.1056/NEJMoa1708322 [DOI] [PubMed] [Google Scholar]
51.Cosman F, Crittenden DB, Adachi JD, Binkley N, Czerwinski E, Ferrari S, et al. Romosozumab Treatment in Postmenopausal Women with Osteoporosis. N Engl J Med. 2016; 375(16):1532–43. 10.1056/NEJMoa1607948 [DOI] [PubMed] [Google Scholar]
52.Bone HG, Cosman F, Miller PD, Williams GC, Hattersley G, Hu MY, et al. ACTIVExtend: 24 Months of Alendronate After 18 Months of Abaloparatide or Placebo for Postmenopausal Osteoporosis. J Clin Endocrinol Metab. 2018; 103(8):2949–57. 10.1210/jc.2018-00163 [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Asadipooya K, Weinstock A. Cardiovascular Outcomes of Romosozumab and Protective Role of Alendronate. Arterioscler Thromb Vasc Biol. 2019; 39(7):1343–50. 10.1161/ATVBAHA.119.312371 [DOI] [PubMed] [Google Scholar]
54.Lyu H, Jundi B, Xu C, Tedeschi SK, Yoshida K, Zhao S, et al. Comparison of Denosumab and Bisphosphonates in Patients With Osteoporosis: A Meta-Analysis of Randomized Controlled Trials. J Clin Endocrinol Metab. 2019; 104(5):1753–65. 10.1210/jc.2018-02236 [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Anastasilakis AD, Polyzos SA, Makras P. THERAPY OF ENDOCRINE DISEASE: Denosumab vs bisphosphonates for the treatment of postmenopausal osteoporosis. EUR J ENDOCRINOL. 2018; 179(1):R31–45. 10.1530/EJE-18-0056 [DOI] [PubMed] [Google Scholar]
56.Bischoff-Ferrari HA, Bhasin S, Manson JE. Preventing Fractures and Falls: A Limited Role for Calcium and Vitamin D Supplements? JAMA. 2018; 319(15):1552–3. 10.1001/jama.2018.4023 [DOI] [PubMed] [Google Scholar]
57.Kendler DL, Marin F, Zerbini C, Russo LA, Greenspan SL, Zikan V, et al. Effects of teriparatide and risedronate on new fractures in post-menopausal women with severe osteoporosis (VERO): a multicentre, double-blind, double-dummy, randomised controlled trial. LANCET. 2018; 391(10117):230–40. 10.1016/S0140-6736(17)32137-2 [DOI] [PubMed] [Google Scholar]
58.Liberman UA, Weiss SR, Broll J, Minne HW, Quan H, Bell NH, et al. Effect of oral alendronate on bone mineral density and the incidence of fractures in postmenopausal osteoporosis. The Alendronate Phase III Osteoporosis Treatment Study Group. N Engl J Med. 1995; 333(22):1437–43. 10.1056/NEJM199511303332201 [DOI] [PubMed] [Google Scholar]
59.Black DM, Schwartz AV, Ensrud KE, Cauley JA, Levis S, Quandt SA, et al. Effects of continuing or stopping alendronate after 5 years of treatment: the Fracture Intervention Trial Long-term Extension (FLEX): a randomized trial. JAMA. 2006; 296(24):2927–38. 10.1001/jama.296.24.2927 [DOI] [PubMed] [Google Scholar]
60.Compston JE, McClung MR, Leslie WD. Osteoporosis. LANCET. 2019; 393(10169):364–76. 10.1016/S0140-6736(18)32112-3 [DOI] [PubMed] [Google Scholar]
61.Jiang XD. Hormone therapy for the treatment of postmenopausal osteoporosis: will it soon become a lost art in medicine? MENOPAUSE. 2018; 25(7):723–7. 10.1097/GME.0000000000001124 [DOI] [PubMed] [Google Scholar]
62.Sancar F. Caution With New Osteoporosis Drug. JAMA. 2019; 321(19):1862 10.1001/jama.2019.6189 [DOI] [PubMed] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0234123.r001

Decision Letter 0

David Fyhrie

27 Jan 2020

PONE-D-19-32748

Clinical efficacy and safety of drug interventions for primary and secondary prevention of osteoporotic fractures in postmenopausal women: network meta-analysis followed by factor and cluster analysis

PLOS ONE

Dear Dr. Ding,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Mar 12 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

David Fyhrie

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements:

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf and http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In the Methods, please specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies). Please ensure that the specific method of assessment (funnel plot, Egger's test, Begg's test, etc) is mentioned.

3. Thank you for stating the following on the title page of your manuscript:

"Funding: No specific funding was received for this study."

We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows: "NO"

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have made a substantial job in summarizing evidence from RCTs on drug intervention for the prevention of osteoporotic fractures.

The manuscript is rather technical and some aspects of the analysis of data can be difficult to follow. The English of this article is rather good but could be further improved by English editing.

I have some comments:

When judging the evidence (e.g. figure 5), is uncertainty taken properly into account? At the bottom of page 8, Cluster 1 and 2 are described, but these seem determined based on the estimates presented in the figure. Are the differences statistically significant?

Tolerability is defined as withdrawals due to adverse events. However, it does not take into account the severity of the adverse effect. This is acknowledged by the authors in the discussion, but should it also be taken into account when summarizing the main findings concerning safety?

Figs 2A-2H are very nice, but it is unclear how they should be judged. ‘evidence loop in the network plot’ could be defined more clearly. Is it purely the visual inspection saying that there was an evidence loop or no.

Did you consider including hip fracture as an additional endpoint. If not, this should be noted as a limitation of the study.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2020 Jun 3;15(6):e0234123. doi: 10.1371/journal.pone.0234123.r002

Author response to Decision Letter 0

1 Feb 2020

Dear Editor,

Thank you very much for sending us the valuable comments of the reviewers on our manuscript (No.: PONE-D-19-32748; Title: Clinical efficacy and safety of drug interventions for primary and secondary prevention of osteoporotic fractures in postmenopausal women: network meta-analysis followed by factor and cluster analysis), which have helped us to improve the quality of our paper greatly. We have thoroughly revised our manuscript according to those comments. All comments have been addressed and itemized as follows.

Responses to Journal Requirements

Question 1:

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf and http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf.

Answer

We have revised title page, main body and references of our manuscript as well as file naming of supplementary materials according to PLOS ONE's style requirements.

Question 2:

In the Methods, please specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies). Please ensure that the specific method of assessment (funnel plot, Egger's test, Begg's test, etc) is mentioned.

Answer

We have performed the detection of publication bias, and have added corresponding sentences into the methods, results, and discussion section in the revised manuscript.

Question 3:

Thank you for stating the following on the title page of your manuscript:

"Funding: No specific funding was received for this study."

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows: "NO".

Answer

No specific funding was received for this study, and accordingly, we have not provided funding information in the Funding Statement or in any section of the manuscript.

Responses to Comments to the Author

Question 1:

Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Partly

Answer

Although the process of data analyses in this paper was complex, we believe that every analysis step was logical and rigorous, which ensured the conclusions supported by the data.

Question 2:

Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Answer

Question 3:

Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #1: Yes.

Answer

None.

Question 4:

Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #1: No.

Answer

We have made every effort to revise the manuscript in terms of English grammar, sentence structure, and spelling. If something wrong still exists in these aspects, please point out them and we will try our best to revise them again.

Question 5:

Review Comments to the Author

Reviewer #1: The authors have made a substantial job in summarizing evidence from RCTs on drug intervention for the prevention of osteoporotic fractures.

The manuscript is rather technical and some aspects of the analysis of data can be difficult to follow. The English of this article is rather good but could be further improved by English editing.

Answer:

Thank you very much for the positive comments.

The statistical analyses in this paper are complicated indeed. We performed network meta-analysis and calculated the SUCRA value for each treatment respectively in the primary prevention group and in the secondary prevention group at first. Then we did factor analysis based on the SUCRA values to obtain factor scores. At last, we conducted cluster analysis based on the factor scores to rank various interventions with both efficacy and safety taken into account. Although the statistical analyses are complex, we believe that they are logical and rigorous. Meanwhile, we have provided detailed methods and results in the manuscript for all data analyses to make them clear to readers. On the other hand, we have made every effort to revise the manuscript in terms of English grammar, sentence structure, and spelling. If something wrong still exists in these aspects, please point out them and we will try our best to revise them again.

Reviewer #1: I have some comments:

Answer:

This is a good point.

Although the cluster analysis in the paper ensured maximum intracluster similarity and minimum intercluster similarity according to the maximum of clustering gain [1], it failed to tell us whether the difference among various clusters was statistically significant or not. This point has been added into the fourth point in the limitations section of the revised manuscript.

References

1. Jung Y, Park H, Du D, Drake B. A Decision Criterion for the Optimal Number of Clusters in Hierarchical Clustering. J GLOBAL OPTIM. 2003; 25(1):91-111.

Reviewer #1: Tolerability is defined as withdrawals due to adverse events. However, it does not take into account the severity of the adverse effect. This is acknowledged by the authors in the discussion, but should it also be taken into account when summarizing the main findings concerning safety?

Answer:

This is a good point.

Given that the evidence summary section in the paper mainly stated key findings and comparisons with previous studies, we have added this point into the fifth point in the limitations section.

Reviewer #1: Figs 2A-2H are very nice, but it is unclear how they should be judged. ‘evidence loop in the network plot’ could be defined more clearly. Is it purely the visual inspection saying that there was an evidence loop or no.

Answer:

“Evidence loop in the network plot” means “closed loop in the network plot”, and that it exists in the network plot means the necessity of test for inconsistency. There is no closed loop in Fig 2D, which meant that it was not necessary to conduct test of inconsistency for the outcome of acceptability in the primary prevention group. In network plots, the size of every circle is proportional to the number of randomly assigned participants and indicates the sample size; and the width of the lines corresponds to the number of trials. Thus, those network plots also tell us: placebo was most commonly compared to other interventions in each network plot; on the contrary, the lines among active drugs were thin, and the number of those lines was limited, which suggested the lack of head-to-head trials comparing active drugs. Accordingly, we have revised the third paragraph in the results section and the legend of Fig 2. Meanwhile, the fact of the lack of head-to-head trials was clarified in the fourth point in the limitations section.

Reviewer #1: Did you consider including hip fracture as an additional endpoint. If not, this should be noted as a limitation of the study.

Answer:

This limitation of the paper has been added into the last point in the limitations section.

Question 6:

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Answer

None.

Hopefully, we have made an appropriate revision based on these valuable comments and suggestions. Please let me know if there is anything else we need to do with the revision, and we’ll complete it as soon as possible.

We thank you wholeheartedly for your excellent work. Your kind assistance is greatly appreciated. We look forward to any future correspondence.

Yours sincerely,

Liangliang Ding

E-mail: 15887847798m0@sina.cn

Attachment

Submitted filename: Response to Reviewers.doc

Click here for additional data file.^{(149.5KB, doc)}

PLoS One. doi: 10.1371/journal.pone.0234123.r003

Decision Letter 1

Carlos M Isales

11 May 2020

PONE-D-19-32748R1

PLOS ONE

Dear Dr. Ding,

We would appreciate receiving your revised manuscript by Jun 25 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

We look forward to receiving your revised manuscript.

Kind regards,

Carlos M. Isales, M.D.

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Reviewer #1: Thank you for the revised version of the manuscript.

I appreciate that you have included in the limitation section that:

'Meanwhile, the two safety outcomes failed to consider the severity of adverse effects.'

(I suggest omitting 'Meanwhile')

This limitation should also be indicated in the abstract and other places where the main findings are summarized. For example, when it is stated (abstract) that: 'Romosozumab, teriparatide, denosumab and risedronate, with the greatest composite scores, constituted the optimal cluster having both superior efficacy and superior safety', the words 'superior safety is a very strong statement. Many reads do not read the limitation section.

You have also mention some other very relevant limitation. They should also influence on the rest of the discussion.

It is stated on page 11: 'Fortunately, we have identified the newest evidence on zoledronate intravenously used at a dose of 5 mg per 18 months, which nowadays is the only one drug intervention that has the capability to significantly reduce both vertebral (RR 0.46) and nonvertebral fracture risk (HR 0.66) for primary prevention of PMO'

COMMENT: As you indicate in the limitation section, this could be due to lack of studies/statistical power. I therefor suggest that you reformulate to 'which is the only drug which has been shown to significantly reduce....'

In the abstract, SUCRA should be defined

Reviewer #2: This study provided a network meta-analysis (NMA) that enabled the authors to pool data from 57 individual studies that included evidence from both direct and indirect comparisons of the 15 interventions for two efficacy outcomes and two safety outcomes.

The surface under the cumulative ranking curve (SUCRA) is a numeric presentation of the overall ranking of an intervention and presents a single number associated with each treatment. SUCRA values range from 0 to 1. Higher SUCRA values indicate that the intervention was more likely to be in the top rank while lower SUCRA values indicate that the intervention likely to be in the lower rank.

Since the SUCRA values were used to rank the interventions as well as used in the subsequent factor and cluster analyses it is important to consider their quality. There are a few problem associated with SUCRA values (1) some of which the authors have addressed.

First, the evidence on which the SUCRA rankings are based may be of very low quality and therefore untrustworthy. The authors have provided the Jadad scores along with the data corresponding to all of the selected studies. They have also provided visual representation of NMA point estimates and certainty intervals comparing estimates of each treatment against placebo in the main text as well as the comparisons among the treatments in the Supplemental files.

Second, there are typically several relevant outcomes. A treatment that is best in one outcome (say, an efficacy outcome) may be the worst in another outcome (for example, a safety outcome). The authors were able to combine efficacy and safety outcomes for the secondary prevention group via factor/cluster analyses to attempt to address this and thereby provide rankings for the interventions based on all outcomes.

There are a few problems that were not addressed and might be good to include in the Discussion or the Limitations section. Issues such as cost and a clinician’s familiarity with use of a particular treatment may also bear consideration when ranking interventions. Also, in the process of calculation, SUCRA does not consider the magnitude of differences in effects between treatments (e.g. the first ranked treatment may be only slightly, or a great deal better than the second ranked treatment). And finally, chance may explain any apparent difference between treatments, and SUCRA does not capture that possibility.

All in all it appears that all of the methods associated with this NMA are appropriate well described. The presentations of the assessments of heterogeneity and publication bias are important aids for the interpretation of the results.

(1) Mbuagbaw, L., Rochwerg, B., Jaeschke, R. et al. Approaches to interpreting and choosing the best treatments in network meta-analyses. Syst Rev 6, 79 (2017). https://doi.org/10.1186/s13643-017-0473-z

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

PLoS One. 2020 Jun 3;15(6):e0234123. doi: 10.1371/journal.pone.0234123.r004

Author response to Decision Letter 1

13 May 2020

Dear Editor,

Thank you very much for sending us the valuable comments of the reviewers on our manuscript (No.: PONE-D-19-32748R1; Title: Clinical efficacy and safety of drug interventions for primary and secondary prevention of osteoporotic fractures in postmenopausal women: network meta-analysis followed by factor and cluster analysis), which have helped us to improve the quality of our paper greatly. We have thoroughly revised our manuscript according to those comments. All comments have been addressed and itemized as follows.

Responses to Journal Requirements

Answer

We have deposited the study protocol for this network meta-analysis in protocols.io, and the DOI of the study protocol has been presented in the first paragraph of the METHODS section.

Responses to Comments to the Author

Question 1:

If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: (No Response)

Answer

None.

Question 2:

Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

Reviewer #2: Yes

Answer

Thanks very much for these positive responses.

Question 3:

Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: Yes.

Answer

This point has been detailed in the first response letter, and we believe the statistical analyses in this paper were logical and rigorous although they were complicated. Thanks very much for the positive response of Reviewer #2.

Question 4:

Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #1: Yes

Reviewer #2: Yes.

Answer

Thanks very much for these positive responses.

Question 5:

Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #1: Yes

Reviewer #2: Yes.

Answer

Thanks very much for these positive responses.

Question 6:

Review Comments to the Author

Reviewer #1: Thank you for the revised version of the manuscript.

I appreciate that you have included in the limitation section that:

'Meanwhile, the two safety outcomes failed to consider the severity of adverse effects.'

(I suggest omitting 'Meanwhile')

Answer:

We have omitted 'Meanwhile' in the sentence of 'Meanwhile, the two safety outcomes failed to consider the severity of adverse effects'.

We have added this limitation into the ABSTRACT section and into the second paragraph of the DISCUSSION section.

Reviewer #1: You have also mention some other very relevant limitation. They should also influence on the rest of the discussion.

Answer:

We have revised this sentence. Meanwhile, we have revised the sentences relevant with this point in the ABSTRACT and CONCLUSIONS sections.

Reviewer #1: In the abstract, SUCRA should be defined.

Answer:

SUCRA has been defined in the ABSTRACT section.

Answer:

Thanks very much for these positive comments.

Reviewer #2: There are a few problems that were not addressed and might be good to include in the Discussion or the Limitations section. Issues such as cost and a clinician’s familiarity with use of a particular treatment may also bear consideration when ranking interventions. Also, in the process of calculation, SUCRA does not consider the magnitude of differences in effects between treatments (e.g. the first ranked treatment may be only slightly, or a great deal better than the second ranked treatment). And finally, chance may explain any apparent difference between treatments, and SUCRA does not capture that possibility.

Answer:

The first point has been added into the last paragraph of the Clinical Implications section in the DISCUSSION section. The second and third points have been added into the last point of the Limitations section.

Question 7:

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Answer

None.

We thank you wholeheartedly for your excellent work. Your kind assistance is greatly appreciated. We look forward to any future correspondence.

Yours sincerely,

Liangliang Ding

E-mail: 15887847798m0@sina.cn

Attachment

Submitted filename: Response to Reviewers.doc

Click here for additional data file.^{(147KB, doc)}

PLoS One. doi: 10.1371/journal.pone.0234123.r005

Decision Letter 2

Carlos M Isales

20 May 2020

PONE-D-19-32748R2

Dear Dr. Ding,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Carlos M. Isales, M.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

PLoS One. doi: 10.1371/journal.pone.0234123.r006

Acceptance letter

Carlos M Isales

22 May 2020

PONE-D-19-32748R2

Dear Dr. Ding:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor Carlos M. Isales

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Appendix. PRISMA checklist.

(DOC)

Click here for additional data file.^{(64.5KB, doc)}

S2 Appendix. Full search strategy.

(PDF)

Click here for additional data file.^{(117.7KB, pdf)}

S3 Appendix. Checklist of studies included for network meta-analysis.

(PDF)

Click here for additional data file.^{(67.5KB, pdf)}

S4 Appendix. Study characteristics, outcome data and quality assessment.

(XLSX)

Click here for additional data file.^{(24.5KB, xlsx)}

S5 Appendix. Network meta-analysis for vertebral fractures in the primary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(29.7KB, pdf)}

S6 Appendix. Network meta-analysis for nonvertebral fractures in the primary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(55.3KB, pdf)}

S7 Appendix. Network meta-analysis for tolerability in the primary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(57.6KB, pdf)}

S8 Appendix. Network meta-analysis for acceptability in the primary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(56.2KB, pdf)}

S9 Appendix. Network meta-analysis for vertebral fractures in the secondary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(164.6KB, pdf)}

S10 Appendix. Network meta-analysis for nonvertebral fractures in the secondary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(138.1KB, pdf)}

S11 Appendix. Network meta-analysis for tolerability in the secondary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(111.9KB, pdf)}

S12 Appendix. Network meta-analysis for acceptability in the secondary prevention group with active drugs compared among them.

(PDF)

Click here for additional data file.^{(139.1KB, pdf)}

S13 Appendix. Inconsistency test for vertebral fractures in the primary prevention group.

(PDF)

Click here for additional data file.^{(31.1KB, pdf)}

S14 Appendix. Inconsistency test for nonvertebral fractures in the primary prevention group.

(PDF)

Click here for additional data file.^{(57.3KB, pdf)}

S15 Appendix. Inconsistency test for tolerability in the primary prevention group.

(PDF)

Click here for additional data file.^{(31KB, pdf)}

S16 Appendix. Inconsistency test for vertebral fractures in the secondary prevention group.

(PDF)

Click here for additional data file.^{(87.2KB, pdf)}

S17 Appendix. Inconsistency test for nonvertebral fractures in the secondary prevention group.

(PDF)

Click here for additional data file.^{(88.3KB, pdf)}

S18 Appendix. Inconsistency test for tolerability in the secondary prevention group.

(PDF)

Click here for additional data file.^{(89.2KB, pdf)}

S19 Appendix. Inconsistency test for acceptability in the secondary prevention group.

(PDF)

Click here for additional data file.^{(89.3KB, pdf)}

S20 Appendix. Heterogeneity assessment for vertebral fractures in the primary prevention group.

(PDF)

Click here for additional data file.^{(63.2KB, pdf)}

S21 Appendix. Heterogeneity assessment for nonvertebral fractures in the primary prevention group.

(PDF)

Click here for additional data file.^{(93.8KB, pdf)}

S22 Appendix. Heterogeneity assessment for tolerability in the primary prevention group.

(PDF)

Click here for additional data file.^{(93.4KB, pdf)}

S23 Appendix. Heterogeneity assessment for acceptability in the primary prevention group.

(PDF)

Click here for additional data file.^{(61.6KB, pdf)}

S24 Appendix. Heterogeneity assessment for vertebral fractures in the secondary prevention group.

(PDF)

Click here for additional data file.^{(179.5KB, pdf)}

S25 Appendix. Heterogeneity assessment for nonvertebral fractures in the secondary prevention group.

(PDF)

Click here for additional data file.^{(186.2KB, pdf)}

S26 Appendix. Heterogeneity assessment for tolerability in the secondary prevention group.

(PDF)

Click here for additional data file.^{(180.4KB, pdf)}

S27 Appendix. Heterogeneity assessment for acceptability in the secondary prevention group.

(PDF)

Click here for additional data file.^{(181.1KB, pdf)}

S28 Appendix. SUCRA values of drug interventions for primary prevention of osteoporotic fractures.

(XLSX)

Click here for additional data file.^{(9.1KB, xlsx)}

S29 Appendix. Factor analysis of SUCRA values estimated based on the primary prevention data.

(PDF)

Click here for additional data file.^{(58.4KB, pdf)}

S30 Appendix. SUCRA values of drug interventions for secondary prevention of osteoporotic fractures.

(XLSX)

Click here for additional data file.^{(10.6KB, xlsx)}

S31 Appendix. Factor analysis of SUCRA values estimated based on the secondary prevention data.

(PDF)

Click here for additional data file.^{(58.4KB, pdf)}

S32 Appendix. Comparison-adjusted funnel plots.

(PDF)

Click here for additional data file.^{(262.9KB, pdf)}

Attachment

Submitted filename: Response to Reviewers.doc

Click here for additional data file.^{(149.5KB, doc)}

Attachment

Submitted filename: Response to Reviewers.doc

Click here for additional data file.^{(147KB, doc)}

Data Availability Statement

All relevant data are within the manuscript and its Supporting Information files.

[pone.0234123.ref001] 1.Lorentzon M. Treating osteoporosis to prevent fractures: current concepts and future developments. J INTERN MED. 2019; 285(4):381–94. 10.1111/joim.12873 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref002] 2.Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. LANCET. 2002; 359(9319):1761–7. 10.1016/S0140-6736(02)08657-9 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref003] 3.Black DM, Rosen CJ. Clinical Practice. Postmenopausal Osteoporosis. N Engl J Med. 2016; 374(3):254–62. 10.1056/NEJMcp1513724 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref004] 4.Hernlund E, Svedbom A, Ivergard M, Compston J, Cooper C, Stenmark J, et al. Osteoporosis in the European Union: medical management, epidemiology and economic burden. A report prepared in collaboration with the International Osteoporosis Foundation (IOF) and the European Federation of Pharmaceutical Industry Associations (EFPIA). ARCH OSTEOPOROS. 2013; 8(136. 10.1007/s11657-013-0136-1 [DOI] [PMC free article] [PubMed]

[pone.0234123.ref005] 5.Cosman F, de Beur SJ, LeBoff MS, Lewiecki EM, Tanner B, Randall S, et al. Clinician's Guide to Prevention and Treatment of Osteoporosis. Osteoporos Int. 2014; 25(10):2359–81. 10.1007/s00198-014-2794-2 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref006] 6.Trajanoska K, Schoufour JD, de Jonge E, Kieboom B, Mulder M, Stricker BH, et al. Fracture incidence and secular trends between 1989 and 2013 in a population based cohort: The Rotterdam Study. BONE. 2018; 114(116–24. 10.1016/j.bone.2018.06.004 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref007] 7.Siris ES, Brenneman SK, Miller PD, Barrett-Connor E, Chen YT, Sherwood LM, et al. Predictive value of low BMD for 1-year fracture outcomes is similar for postmenopausal women ages 50–64 and 65 and Older: results from the National Osteoporosis Risk Assessment (NORA). J BONE MINER RES. 2004; 19(8):1215–20. 10.1359/JBMR.040508 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref008] 8.Schuit SC, van der Klift M, Weel AE, de Laet CE, Burger H, Seeman E, et al. Fracture incidence and association with bone mineral density in elderly men and women: the Rotterdam Study. BONE. 2004; 34(1):195–202. 10.1016/j.bone.2003.10.001 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref009] 9.Siris ES, Chen YT, Abbott TA, Barrett-Connor E, Miller PD, Wehren LE, et al. Bone mineral density thresholds for pharmacological intervention to prevent fractures. Arch Intern Med. 2004; 164(10):1108–12. 10.1001/archinte.164.10.1108 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref010] 10.Saito T, Sterbenz JM, Malay S, Zhong L, MacEachern MP, Chung KC. Effectiveness of anti-osteoporotic drugs to prevent secondary fragility fractures: systematic review and meta-analysis. Osteoporos Int. 2017; 28(12):3289–300. 10.1007/s00198-017-4175-0 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref011] 11.Center JR, Bliuc D, Nguyen TV, Eisman JA. Risk of subsequent fracture after low-trauma fracture in men and women. JAMA. 2007; 297(4):387–94. 10.1001/jama.297.4.387 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref012] 12.Johnell O, Kanis JA, Oden A, Sernbo I, Redlund-Johnell I, Petterson C, et al. Fracture risk following an osteoporotic fracture. Osteoporos Int. 2004; 15(3):175–9. 10.1007/s00198-003-1514-0 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref013] 13.Qaseem A, Forciea MA, McLean RM, Denberg TD. Treatment of Low Bone Density or Osteoporosis to Prevent Fractures in Men and Women: A Clinical Practice Guideline Update From the American College of Physicians. ANN INTERN MED. 2017; 166(11):818–39. 10.7326/M15-1361 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref014] 14.Barrionuevo P, Kapoor E, Asi N, Alahdab F, Mohammed K, Benkhadra K, et al. Efficacy of Pharmacological Therapies for the Prevention of Fractures in Postmenopausal Women: A Network Meta-Analysis. J Clin Endocrinol Metab. 2019; 104(5):1623–30. 10.1210/jc.2019-00192 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref015] 15.Reginster J, Bianic F, Campbell R, Martin M, Williams SA, Fitzpatrick LA. Abaloparatide for risk reduction of nonvertebral and vertebral fractures in postmenopausal women with osteoporosis: a network meta-analysis. Osteoporos Int. 2019; 30(7):1465–73. 10.1007/s00198-019-04947-2 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref016] 16.Yang XC, Deng ZH, Wen T, Luo W, Xiao WF, Zhao RB, et al. Network Meta-Analysis of Pharmacological Agents for Osteoporosis Treatment and Fracture Prevention. CELL PHYSIOL BIOCHEM. 2016; 40(3–4):781–95. 10.1159/000453138 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref017] 17.Ellis AG, Reginster JY, Luo X, Cappelleri JC, Chines A, Sutradhar S, et al. Bazedoxifene versus oral bisphosphonates for the prevention of nonvertebral fractures in postmenopausal women with osteoporosis at higher risk of fracture: a network meta-analysis. VALUE HEALTH. 2014; 17(4):424–32. 10.1016/j.jval.2014.01.008 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref018] 18.Tan X, Wen F, Yang W, Xie JY, Ding LL, Mo YX. Comparative efficacy and safety of pharmacological interventions for osteoporosis in postmenopausal women: a network meta-analysis (Chongqing, China). MENOPAUSE. 2019; 26(8):929–39. 10.1097/GME.0000000000001321 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref019] 19.Freemantle N, Cooper C, Diez-Perez A, Gitlin M, Radcliffe H, Shepherd S, et al. Results of indirect and mixed treatment comparison of fracture efficacy for osteoporosis treatments: a meta-analysis. Osteoporos Int. 2013; 24(1):209–17. 10.1007/s00198-012-2068-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref020] 20.Zhang L, Pang Y, Shi Y, Xu M, Xu X, Zhang J, et al. Indirect comparison of teriparatide, denosumab, and oral bisphosphonates for the prevention of vertebral and nonvertebral fractures in postmenopausal women with osteoporosis. MENOPAUSE. 2015; 22(9):1021–5. 10.1097/GME.0000000000000466 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref021] 21.Migliore A, Broccoli S, Massafra U, Cassol M, Frediani B. Ranking antireabsorptive agents to prevent vertebral fractures in postmenopausal osteoporosis by mixed treatment comparison meta-analysis. Eur Rev Med Pharmacol Sci. 2013; 17(5):658–67. [PubMed] [Google Scholar]

[pone.0234123.ref022] 22.Reid IR, Horne AM, Mihov B, Stewart A, Garratt E, Wong S, et al. Fracture Prevention with Zoledronate in Older Women with Osteopenia. N Engl J Med. 2018; 379(25):2407–16. 10.1056/NEJMoa1808082 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref023] 23.Markham A. Romosozumab: First Global Approval. DRUGS. 2019; 79(4):471–6. 10.1007/s40265-019-01072-6 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref024] 24.Rachner TD, Hofbauer LC, Gobel A, Tsourdi E. Novel therapies in osteoporosis: PTH-related peptide analogs and inhibitors of sclerostin. J MOL ENDOCRINOL. 2019; 62(2):R145–54. 10.1530/JME-18-0173 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref025] 25.Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. ANN INTERN MED. 2015; 162(11):777–84. 10.7326/M14-2385 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref026] 26.Wells GA, Cranney A, Peterson J, Boucher M, Shea B, Robinson V, et al. Alendronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database Syst Rev. 2008;1):D1155 10.1002/14651858.CD001155.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref027] 27.Wells G, Cranney A, Peterson J, Boucher M, Shea B, Robinson V, et al. Risedronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database Syst Rev. 2008;1):D4523 10.1002/14651858.CD004523.pub3 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref028] 28.Cosman F, de Beur SJ, LeBoff MS, Lewiecki EM, Tanner B, Randall S, et al. Clinician's Guide to Prevention and Treatment of Osteoporosis. Osteoporos Int. 2014; 25(10):2359–81. 10.1007/s00198-014-2794-2 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref029] 29.Siris ES, Simon JA, Barton IP, McClung MR, Grauer A. Effects of risedronate on fracture risk in postmenopausal women with osteopenia. Osteoporos Int. 2008; 19(5):681–6. 10.1007/s00198-007-0493-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref030] 30.Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials. 1996; 17(1):1–12. 10.1016/0197-2456(95)00134-4 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref031] 31.Dias S, Sutton AJ, Ades AE, Welton NJ. Evidence synthesis for decision making 2: a generalized linear modeling framework for pairwise and network meta-analysis of randomized controlled trials. MED DECIS MAKING. 2013; 33(5):607–17. 10.1177/0272989X12458724 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref032] 32.Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003; 327(7414):557–60. 10.1136/bmj.327.7414.557 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref033] 33.Dias S, Welton NJ, Caldwell DM, Ades AE. Checking consistency in mixed treatment comparison meta-analysis. STAT MED. 2010; 29(7–8):932–44. 10.1002/sim.3767 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref034] 34.Chaimani A, Higgins JP, Mavridis D, Spyridonos P, Salanti G. Graphical tools for network meta-analysis in STATA. PLOS ONE. 2013; 8(10):e76654 10.1371/journal.pone.0076654 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref035] 35.Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial. J CLIN EPIDEMIOL. 2011; 64(2):163–71. 10.1016/j.jclinepi.2010.03.016 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref036] 36.Handl J, Knowles J, Kell DB. Computational cluster validation in post-genomic data analysis. BIOINFORMATICS. 2005; 21(15):3201–12. 10.1093/bioinformatics/bti517 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref037] 37.Jung Y, Park H, Du D, Drake B. A Decision Criterion for the Optimal Number of Clusters in Hierarchical Clustering. J GLOBAL OPTIM. 2003; 25(1):91–111. [Google Scholar]

[pone.0234123.ref038] 38.Zhou J, Ma X, Wang T, Zhai S. Comparative efficacy of bisphosphonates in short-term fracture prevention for primary osteoporosis: a systematic review with network meta-analyses. Osteoporos Int. 2016; 27(11):3289–300. 10.1007/s00198-016-3654-z [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref039] 39.Sanderson J, Martyn-St JM, Stevens J, Goka E, Wong R, Campbell F, et al. Clinical effectiveness of bisphosphonates for the prevention of fragility fractures: A systematic review and network meta-analysis. BONE. 2016; 89(52–8. 10.1016/j.bone.2016.05.013 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref040] 40.Murad MH, Drake MT, Mullan RJ, Mauck KF, Stuart LM, Lane MA, et al. Clinical review. Comparative effectiveness of drug treatments to prevent fragility fractures: a systematic review and network meta-analysis. J Clin Endocrinol Metab. 2012; 97(6):1871–80. 10.1210/jc.2011-3060 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref041] 41.Tadrous M, Wong L, Mamdani MM, Juurlink DN, Krahn MD, Levesque LE, et al. Comparative gastrointestinal safety of bisphosphonates in primary osteoporosis: a network meta-analysis. Osteoporos Int. 2014; 25(4):1225–35. 10.1007/s00198-013-2576-2 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref042] 42.Yang L, Kang N, Yang JC, Su QJ, Liu YZ, Guan L, et al. Drug efficacies on bone mineral density and fracture rate for the treatment of postmenopausal osteoporosis: a network meta-analysis. Eur Rev Med Pharmacol Sci. 2019; 23(6):2640–68. 10.26355/eurrev_201903_17414 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref043] 43.Miller PD, Hattersley G, Riis BJ, Williams GC, Lau E, Russo LA, et al. Effect of Abaloparatide vs Placebo on New Vertebral Fractures in Postmenopausal Women With Osteoporosis: A Randomized Clinical Trial. JAMA. 2016; 316(7):722–33. 10.1001/jama.2016.11136 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref044] 44.Cosman F. Long-term treatment strategies for postmenopausal osteoporosis. CURR OPIN RHEUMATOL. 2018; 30(4):420–6. 10.1097/BOR.0000000000000509 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref045] 45.Leder BZ, Neer RM, Wyland JJ, Lee HW, Burnett-Bowie SM, Finkelstein JS. Effects of teriparatide treatment and discontinuation in postmenopausal women and eugonadal men with osteoporosis. J Clin Endocrinol Metab. 2009; 94(8):2915–21. 10.1210/jc.2008-2630 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref046] 46.Buckley L, Humphrey MB. Glucocorticoid-Induced Osteoporosis. N Engl J Med. 2018; 379(26):2547–56. 10.1056/NEJMcp1800214 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref047] 47.Cosman F, Nieves JW, Dempster DW. Treatment Sequence Matters: Anabolic and Antiresorptive Therapy for Osteoporosis. J BONE MINER RES. 2017; 32(2):198–202. 10.1002/jbmr.3051 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref048] 48.Black DM, Bilezikian JP, Ensrud KE, Greenspan SL, Palermo L, Hue T, et al. One year of alendronate after one year of parathyroid hormone (1–84) for osteoporosis. N Engl J Med. 2005; 353(6):555–65. 10.1056/NEJMoa050336 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref049] 49.Leder BZ, Tsai JN, Uihlein AV, Wallace PM, Lee H, Neer RM, et al. Denosumab and teriparatide transitions in postmenopausal osteoporosis (the DATA-Switch study): extension of a randomised controlled trial. LANCET. 2015; 386(9999):1147–55. 10.1016/S0140-6736(15)61120-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref050] 50.Saag KG, Petersen J, Brandi ML, Karaplis AC, Lorentzon M, Thomas T, et al. Romosozumab or Alendronate for Fracture Prevention in Women with Osteoporosis. N Engl J Med. 2017; 377(15):1417–27. 10.1056/NEJMoa1708322 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref051] 51.Cosman F, Crittenden DB, Adachi JD, Binkley N, Czerwinski E, Ferrari S, et al. Romosozumab Treatment in Postmenopausal Women with Osteoporosis. N Engl J Med. 2016; 375(16):1532–43. 10.1056/NEJMoa1607948 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref052] 52.Bone HG, Cosman F, Miller PD, Williams GC, Hattersley G, Hu MY, et al. ACTIVExtend: 24 Months of Alendronate After 18 Months of Abaloparatide or Placebo for Postmenopausal Osteoporosis. J Clin Endocrinol Metab. 2018; 103(8):2949–57. 10.1210/jc.2018-00163 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref053] 53.Asadipooya K, Weinstock A. Cardiovascular Outcomes of Romosozumab and Protective Role of Alendronate. Arterioscler Thromb Vasc Biol. 2019; 39(7):1343–50. 10.1161/ATVBAHA.119.312371 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref054] 54.Lyu H, Jundi B, Xu C, Tedeschi SK, Yoshida K, Zhao S, et al. Comparison of Denosumab and Bisphosphonates in Patients With Osteoporosis: A Meta-Analysis of Randomized Controlled Trials. J Clin Endocrinol Metab. 2019; 104(5):1753–65. 10.1210/jc.2018-02236 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0234123.ref055] 55.Anastasilakis AD, Polyzos SA, Makras P. THERAPY OF ENDOCRINE DISEASE: Denosumab vs bisphosphonates for the treatment of postmenopausal osteoporosis. EUR J ENDOCRINOL. 2018; 179(1):R31–45. 10.1530/EJE-18-0056 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref056] 56.Bischoff-Ferrari HA, Bhasin S, Manson JE. Preventing Fractures and Falls: A Limited Role for Calcium and Vitamin D Supplements? JAMA. 2018; 319(15):1552–3. 10.1001/jama.2018.4023 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref057] 57.Kendler DL, Marin F, Zerbini C, Russo LA, Greenspan SL, Zikan V, et al. Effects of teriparatide and risedronate on new fractures in post-menopausal women with severe osteoporosis (VERO): a multicentre, double-blind, double-dummy, randomised controlled trial. LANCET. 2018; 391(10117):230–40. 10.1016/S0140-6736(17)32137-2 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref058] 58.Liberman UA, Weiss SR, Broll J, Minne HW, Quan H, Bell NH, et al. Effect of oral alendronate on bone mineral density and the incidence of fractures in postmenopausal osteoporosis. The Alendronate Phase III Osteoporosis Treatment Study Group. N Engl J Med. 1995; 333(22):1437–43. 10.1056/NEJM199511303332201 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref059] 59.Black DM, Schwartz AV, Ensrud KE, Cauley JA, Levis S, Quandt SA, et al. Effects of continuing or stopping alendronate after 5 years of treatment: the Fracture Intervention Trial Long-term Extension (FLEX): a randomized trial. JAMA. 2006; 296(24):2927–38. 10.1001/jama.296.24.2927 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref060] 60.Compston JE, McClung MR, Leslie WD. Osteoporosis. LANCET. 2019; 393(10169):364–76. 10.1016/S0140-6736(18)32112-3 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref061] 61.Jiang XD. Hormone therapy for the treatment of postmenopausal osteoporosis: will it soon become a lost art in medicine? MENOPAUSE. 2018; 25(7):723–7. 10.1097/GME.0000000000001124 [DOI] [PubMed] [Google Scholar]

[pone.0234123.ref062] 62.Sancar F. Caution With New Osteoporosis Drug. JAMA. 2019; 321(19):1862 10.1001/jama.2019.6189 [DOI] [PubMed] [Google Scholar]

PERMALINK

Clinical efficacy and safety of drug interventions for primary and secondary prevention of osteoporotic fractures in postmenopausal women: Network meta-analysis followed by factor and cluster analysis

Fei Wen

Hongheng Du

Liangliang Ding

Jinxi Hu

Zifeng Huang

Hua Huang

Kaikai Li

Yuxia Mo

Anyin Kuang

Roles

Abstract

Introduction

Methods

Search strategy

Inclusion and exclusion criteria

Study selection

Data extraction and quality assessment

Statistical analysis

Results

Summary of included studies

Fig 1. Flow chart of study selection.

Fig 2.

Subgroup network meta-analyses

Fig 3.

Fig 4.

Assessment of inconsistency and heterogeneity

Factor analyses and cluster analysis

Fig 5. Clustered ranking plot of drug interventions for secondary prevention of osteoporotic fractures.

Detection of publication bias

Discussion

Summary of evidence

Clinical implications

Strengths and limitations

Conclusions

Supporting information

Abbreviations

Data Availability

Funding Statement

References

Decision Letter 0

David Fyhrie

Roles

Author response to Decision Letter 0

Decision Letter 1

Carlos M Isales

Roles

Author response to Decision Letter 1

Decision Letter 2

Carlos M Isales

Roles

Acceptance letter

Carlos M Isales

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases