Does adding exercise or physical activity to pharmacological osteoporosis therapy in patients with increased fracture risk improve bone mineral density and lower fracture risk? A systematic review and meta-analysis

Inclusion criteria followed the Participants, Interventions, Comparators, Outcomes, and Study (PICOS) design framework [29]. Studies that did not meet the following criteria were excluded.

An electronic database search of MEDLINE (via PubMed), EMBASE (via Ovid), CINAHL (via EBSCO Host), and CENTRAL (via the Cochrane Library) was conducted using keywords, Medical Subject Headings (MeSH) terms, as well as related text words. Searches were performed from the database inception to May 6, 2022. The full search strategy is contained in Supplement 4. Unpublished and ongoing trials were searched via International Standard Randomized Controlled Trial Number (https://​www.​isrctn.​com/​), US National Institutes of Health (https://​clinicaltrials.​gov/​), EU Clinical Trials Register (https://​www.​clinicaltrialsre​gister.​eu/​), German Clinical Trials Register (https://​www.​drks.​de/​), and Australian New Zealand Clinical Trials Register (https://​www.​anzctr.​org.​au/​). In addition, the reference lists of prior relevant systematic reviews were screened for additional studies. Forward and backward citation tracking of included studies was performed via Web of Science Core Collection. Hand searches were performed for older bisphosphonates (see Supplement 5).

Two independent assessors (from a pool of three screeners: AKS, NKA, EAC) screened each record against predefined inclusion criteria based on title and abstract and subsequently full text. Covidence (www.​covidence.​org) was used for the screening process. Duplicate records were detected and removed via the Covidence software. Disagreements were resolved through discussions between reviewers and, if necessary, an adjudicator (NLM) and a senior member of the research team (DLB, PJO, BB) were consulted.

Data was extracted independently by two extractors (from a pool of three extractors: AKS, NKA, EAC) using GoogleSheets. Differences and extraction errors for the measures of spread and/or impact of studies reporting were identified by a reviewer (AKS) and were resolved through discussions between reviewers and, if necessary, an adjudicator (NLM) and a senior member of the research team (DLB, PJO, BB) were consulted. Prior to commencing data extraction, this method was piloted on ten studies chosen at random and the results were collated by one team member (AKS). In the custom GoogleSheets for extraction, comment bubbles contained detailed information for the extractors as to what data to extract for each parameter and information on how to address unclear information in publications. The following study information were extracted: publication information (author, year, title, journal, funding), study design, study demographics (e.g., age, sex, number of participants), medication interventions (e.g., category, dose, and frequency), exercise interventions (e.g., types, session and overall intervention duration, intensity, and frequency), measurement time points, and outcomes (e.g., BMD, BTM). If multiple follow-ups existed within each timeframe, we extracted the follow-up data closest to 12 months for BMD outcomes and end of intervention if exercise period deviated. For BTM outcome, data closest to 3 months and end of intervention were extracted (see Supplement 3 for further information).

Data for the main outcomes were extracted as number of participants, mean, and standard deviation (SD) where possible. In case these were reported using other measures of center and spread (e.g., median and interquartile range reported instead), we used standard equations to convert the data [31]. All applied data handling methods are available for each study and outcome in Supplement 6. Where data was presented in a figure only, ImageJ (https://​imagej.​nih.​gov/​ij/​) was used to extract the values by measuring the length of the axes in pixels followed by the length of the relevant data of interest [32]. If it was not possible to extract the required data, information was requested from the corresponding author with a minimum of three times over a 4-week period.

The Cochrane Collaboration Risk of Bias Tool 2 (RoB2) was used to examine potential bias from the randomisation process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported results for randomized trials [33]. An overall risk of bias judgement was made for BMD and BTM outcomes.

The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach was used to assess the certainty of the evidence [34]. Risk of bias, imprecision, inconsistency, indirectness, and publication bias are considered in the GRADE approach to rate the certainty [34]. Further details on the criteria used for GRADE are available in Supplement 7. Due to the exclusive presence of RCTs, a high level of confidence in the evidence was assumed with a possibility of downgrade after the assessment. Two independent assessors (AKS, EAC) assessed risk of bias and GRADE. Conflicts were resolved by discussion and, if necessary, consultation of an adjudicator (NLM).

All statistical analyses and forest plots were completed in R version 4.2.1. (https://​www.​r-project.​org/​) and the R packages meta [35], metafor [36], and metadat [37].

As all outcomes of interest were continuous, but could be measured on different scales, standardized mean difference (SMD) with Hedges correction (Hedges’ g) was used as the effect estimate [38]. Random-effects meta-analysis was used. To estimate the 95% confidence intervals (95%CI), the Hartung-Knapp-Sidik-Jonkman adjustment with ad hoc correction was used [39, 41]. Statistical significance was set as an alpha of ≤ 0.05.

Heterogeneity was assessed by I² statistic via the reported 95% prediction intervals (95%PI) where possible. The heterogeneity parameter τ (tau) was estimated via restricted maximum likelihood (REML) estimation. Funnel plots and the Egger’s test modified by Pustejovsky [42] to assess the risk of publication bias and small study effects are only recommended if at least 10 studies were available [43]. Furthermore, this threshold also applies for meta-regression and subgroups analysis [43]. Accordingly, these statistical approaches would only be used if ≥ 10 studies were included. In order to test the robustness of the results, sensitivity analyses exploring the role of outliers and influential trials were included [44].

The details of the five included RCTs [45‐48] are shown in Table 1 and Supplement 10. The sample size of each record varied from 35 to 249 participants with a total sample of 530 participants with a mean age from 46.6 to 81 years old. The length of interventions in included studies ranged from 6 to 24 months.

A summary of the risk of bias assessment for each study is shown in Supplement 11. Summary risk of bias plots for BMD and BTM outcomes and a detailed risk of bias assessment for each included study and outcome were created (Supplement 11).

For the five trials assessing BMD outcome, 80% had a high risk of bias and 20% some concerns overall. There was low risk of bias for randomization process (40%), deviations from intended interventions (60%), measurement of outcome (40%), and selection of reported results (20%). Missing outcome data was rated as having high risk of bias in 80% of included studies.

For the four trials assessing BTM outcome, 75% had a high and 25% a low risk of bias overall. There was a low risk of bias for randomization process (50%), deviations from intended interventions (75%), missing outcome data (25%), measurement of outcome (50%), and selection of reported results (25%). The certainty of evidence was rated as very low for all study outcomes (Supplement 12). The main reasons for downgrading were risk of bias, inconsistency, and imprecision. Publication bias was assessed by funnel plots (Supplement 13) in addition to criteria used (Supplement 7).

Three trials had the mean and SD extracted directly [45, 46, 49]. For one study [47], the mean and standard error were extracted from a figure. Standard error was converted to SD. One study [48] had mean and SD as percentage change from baseline data. Due to missing data and non-responsiveness of the authors, no transformation of the data was possible. Two trials [45, 47] required pooling of groups due to multiple groups and various recruitment sites in both intervention and comparator groups.

The meta-analytic findings of all BMD related meta-analyses are presented in Table 2.

The meta-analytic findings of all BMT related meta-analyses are presented in Table 3.

We decided to add a search for unpublished and ongoing trials in five trial registries and to search for potentially relevant studies with older bisphosphonates by performing an additional hand search in all databases used. No other protocol deviations occurred.

Previous systematic reviews and meta-analyses have largely focused on specific populations [19, 53, 54], pharmacological therapies [53, 55, 56], and/or types of exercises [57, 58, 58, 59]. The meta-analysis by Yan et al. (2021) [60] has proceeded in a comparable manner, as they also compare EX + PT versus PT alone. They also included different types of exercises with focus on BMD, BTM, and pain. No further details were given regarding the pharmacological therapy used. Their meta-analysis showed significant improvements for BMD, but no improvements for BTMs. The reason of different results may occur because they included more studies (n = 20) and therefore involved more participants (n = 1824). Additionally, any activity to enhance and improve physical health with intervention periods less than 6 months were included. Furthermore, they searched in Chinese databases and included reports only published in Chinese except of one German article. In comparison to the existing literature, our current study focusses on the comparison of exercises instead of routine activities, with currently prescribed pharmacological therapies. In addition, we included only studies that could reach clinically based results due to their duration of intervention (see Supplement 3).

Recent clinical practice guidelines [61‐63] give very similar recommendations in recommending exercise training in combination to pharmacological therapy for treatment of osteoporosis. On the basis of the findings of the current study, we cannot yet recommend additions to clinical practice guidelines that exercise can have an additive effect to pharmacological therapies. While it seems logical to prescribe both exercise and pharmacological therapy to patients with osteoporosis, the certainty of evidence is very low and heterogeneity very high, precluding a possible recommendation.

High-quality RCTs are required to examine the additive effect of exercise to antiresorptive and/or osteoanabolic medication. Specifically, future RCTs should be adequately powered and provide sufficient and useful data quality for a meta-analysis. In terms of participants, future studies should focus primarily on primary osteopenia/osteoporosis. Also, it is possible a greater benefit of exercise may be seen in less active individuals; this should be considered in study design and recruitment. Sex, age, race, treatment history, and comorbidities may moderate treatment effects and should be considered in study design and/or statistical analysis. In terms of the exercise intervention, high-impact exercise and high-intensity progressive resistance training have been shown to have with osteoanabolic effects [64]. Future primary RCTs could and should focus on these interventions, but also consider that these forms of exercise may not suit all populations and other forms of exercise and physical activity (e.g., aerobic dance) should also be evaluated. The frequency of exercise sessions and duration of exercise should be considered, and we recommend adhering to current guideline recommendations [64]. In terms of the pharmacological intervention, it should be investigated if the efficacy of currently in several guidelines suggested sequential therapies using an osteoanabolic agent first followed by an antiresorptive for individuals with very high fracture risk can be enhanced by exercise. For outcome measures, BMD is often taken as a surrogate marker of bone fracture risk, and is a more feasible (e.g., total hip or femoral neck BMD) outcome to assess than long-term fracture rates, even if the latter is the ultimate goal of osteoporosis management. Furthermore, the method of measuring BMD (DXA, QCT, pQCT, and/or QUS) should be considered versus their advantages and limitations. Further, in terms of study design, 9 months of intervention is a conservatively set time period to measure detectable changes in BMD via, for example, DXA [65]. Finally, while it is difficult to blind participants to exercise, the assessors of the bone outcomes must be blinded to the group allocation of the participants.

This study has some limitations. Nearly all study outcomes had a high risk of bias. The lack of blinding participants to exercise treatment may overestimate the intervention effects. Due to the small number of included studies (n = 5), it was not reasonable to perform subgroup analysis or to assess the publication bias and meta-regression by using a statistical approach. A further limitation is the small population in the respective studies indicating further demand on studies [66]. Regarding the BMD outcome, we only included aBMD and vBMD to examine the effect of intervention on bone mineral density. This leads to limited findings [67]. Besides exercise and pharmacological therapies, BTM also depends on various other controllable factors such as nutrition, as well as non-controllable factors such as genetic predisposition [68]. The influenceable factors might also depend on social-economical aspects [69, 70]. These factors should be considered while interpreting the final results and deriving measurements according to the results.

The strength of the current study was a systematic review approach, which is the highest level of evidence. Additionally, the Hartung-Knapp-Sidik-Jonkman adjustment was used. This is important to implement in meta-analyses that include less than five trials [39‐41]. The role of outliers in sensitivity analysis was explored to examine the robustness of the results. The addition of searches of trial registries and backward/forward citation tracking strengthened the search process.