Effect of exercise on bone-related outcomes in patients with cancer affected by bone metastases or bone loss: a systematic review and meta-analysis

Bone health impairment is a significant concern in the oncological scenario, affecting patients in both early- and advanced-stage disease [1]. Several anticancer treatments including hormone therapies such as aromatase inhibitors and androgen deprivation therapy (ADT), as well as chemotherapy agents, contribute to accelerated bone loss by reducing circulating sex steroids [1]. This often leads to an increased risk of osteopenia and osteoporosis, which is further exacerbated by aging, glucocorticoid and steroid use, physical inactivity, and inadequate nutrition status, all frequently observed in this population [2, 3]. Overall, up to 15% of patients with cancer may develop osteoporosis [4], particularly those with breast, lung, genitourinary, or skin cancers [4]. A large epidemiological study also reported a significantly increased risk of bone and osteoporotic fractures among patients with cancer [5]. Aside from fracture-related pain, reduced mobility, and quality of life, fractures are associated with an increased mortality risk, especially in the first year, but remain significant even for the subsequent 10 years after the event [6]. In the advanced- disease setting, bone health can be further compromised by the presence of bone metastases, which affect approximately 70%, 85%, and 40% of patients with metastatic breast, prostate, and lung cancers, respectively [7]. Bone metastases compromise skeletal integrity and increases the risk of skeletal-related events (SREs), including pathological fractures, spinal cord compression, bone pain, and hypercalcemia [8]. The incidence of SREs is high, affecting approximately 43.6% of patients with breast cancer and 45.9% of those with prostate cancer and bone metastases [7]. These complications significantly impact patients’ autonomy, quality of life, and overall survival, reinforcing the urgent need for effective strategies to manage bone health management in this population [7, 9].

Along with pharmacological treatments, such as bone-targeted agents and systemic therapies, non-pharmacological interventions like physical exercise have emerged as promising strategies for preserving or improving bone health and reducing the risk of fractures. Exercise, particularly weight-bearing and strength training, reduces the risk of falls through improved muscle function, coordination, and balance [10]. These adaptations are especially relevant in patients with cancer, where fall risk is a major contributor to fragility fractures. In addition, recent evidence suggest that exercise exerts anti-inflammatory effects by modulating cytokine and myokine profiles [11]. In particular, it promotes the release of anti-inflammatory mediators, such as interleukin-10 (IL-10), interleukin-1 receptor antagonist (IL-1ra), and myokines like irisin, while downregulating pro-inflammatory cytokines such as interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-α) [12]. These systemic changes may help regulate bone remodeling by promoting bone formation and inhibiting resorption, thereby contributing to stronger and more fracture-resistant bones [13, 14]. [8].Moreover, exercise exerts osteogenic effects through mechanical loading, thereby stimulating bone remodeling [15, 16]. In non-cancer populations, including individuals with osteopenia or osteoporosis, resistance training has shown efficacy in increasing bone mineral density (BMD) at clinically relevant sites, such as the lumbar spine and hip [17]. For the oncological population, systematic reviews and meta-analyses have reported positive effects of exercise on BMD in patients with cancer [1, 18, 19]. For instance, Rose et al., encompassing 22 studies, observed significant benefits from exercise in enhancing hip and lumbar spine BMD compared to controls. However, these previous analyses have primarily focused on patients without existing bone impairments, thereby exploring the preventive role of exercise on skeletal health [1, 18‐20]. It remains unclear whether exercise can also be effective in counteracting or reversing bone loss in patients with already compromised bone health, such as those diagnosed with osteopenia, osteoporosis, or bone metastases. To address this knowledge gap, the present systematic review and meta-analysis aims to determine the effects of exercise interventions on bone outcomes, including BMD, bone mineral content (BMC), and bone turnover markers, in patients with cancer and existing bone health impairments (i.e., osteopenia, osteoporosis, and/or bone metastases). By focusing specifically on this high-risk subgroup, our study seeks to move beyond prevention and explore whether exercise can serve as a therapeutic strategy to mitigate bone deterioration or skeletal complications in oncology.

This review was conducted in accordance with ethical standards for systematic reviews. As the study relies solely on previously published data, no new data collection involving human subjects was performed, and ethical approval was not required. The review follows the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines to ensure transparency, reproducibility, and methodological rigor [21]. The protocol was registered in the International Prospective Register of Systematic Reviews a priori (PROSPERO: CRD42024625654).

A total of 12,649 studies were identified across the four databases. After duplicate removal, 10,278 were retained and screened based on title and abstract. Following the exclusion of 10,029 studies due to irrelevance to the research question, 113 full-text papers were assessed, of which 95 did not meet the inclusion criteria. Consequently, 18 studies were included in the systematic review, with 11 providing sufficient data to perform the meta-analysis (Fig. 1). All included studies were randomized controlled trials (RCTs). In two cases, the same exercise intervention was described in separate studies evaluating different parameters and subgroup populations, thus warranting their inclusion [23‐26]. The risk of bias assessment revealed that most studies had a low risk of bias, particularly in outcome measurement and handling of missing data. However, some concerns were noted in the randomization process and selective reporting, with one study showing a high risk of bias in these domains (Supplementary Material). The certainty of the evidence ranged from low to very low for most outcomes, primarily due to concerns about risk of bias, indirectness of evidence, and imprecision. The detailed assessment is shown in Table S4 within Supplementary Materials.

Bild vergrößern

Overall, 1,478 patients were included in this analysis (Table 1). Ten investigations were focused on breast cancer [27‐36], six on prostate cancer [23‐26, 37, 38], and two involved mixed cancer populations [39, 40]. Most patients had stage I-III cancer (n = 915) and were receiving hormone therapy (n = 810) at the time of the intervention. Twelve studies included patients with osteopenia or osteoporosis [27‐38], while six focused on patients with bone metastases [23‐26, 39, 40]. Most investigations reported the supplementation or drugs administered to manage the bone impairment, except for five that did not report this information [23‐26, 28]. BMD and BMC, evaluated in 18 and 3 studies, respectively, were assessed using DEXA, while two used CT to evaluate BMD at metastatic sites [39, 40]. Bone turnover markers were analysed in nine trials [25, 26, 28, 31, 33‐35, 38, 40].

A variety of exercise modalities were employed across the included studies, as detailed in Table 2. Interventions ranged in duration from two to 24 months and included aerobic training, resistance training, impact-loading exercises, and football-based programs. Aerobic protocols primarily involved moderate-intensity walking sessions, while resistance training used body weight, free weights, elastic bands, or machines, typically prescribed at moderate intensity (e.g., 8–12 RM or 60–70% 1RM). Some studies combined different modalities, such as resistance plus aerobic or impact-loading exercises. Impact-loading exercises were included in three studies [28, 35, 38], consisted mainly of jumps with weighted vests loaded at 0–10% body weight, performed three times per week or in a circuit-training format or a series of bounding, hopping, skipping, leaping and drop jumping exercises. Football-based interventions included running, skill drills, strength, and balance exercises performed in structured sessions. Control groups generally received usual care, and participants were instructed to maintain their habitual activity levels.

Twelve investigations had explored the impact of exercise on bone outcomes in patients with breast (n = 10 studies) or prostate (n = 2 study), who were affected by osteopenia/osteoporosis [27‐38].

Regarding BMD, most of the investigations found no improvements in the whole-body (n = 4 studies), lumbar spine (n = 7 studies), total hip (n = 6 studies), or femoral neck (n = 6 studies). For instance, 6 months of clinic-based and home-based aerobic, resistance and impact training 3–5 times per week at moderate intensity did not significantly increase BMD on the lumbar spine (EX: 1.193 ± 0.197 vs. 1.188 ± 0.194 g/cm²; CG: 1.154 ± 0.173 vs. 1.136 ± 0.175 g/cm²; p = 0.111), total hip (EX: 1.013 ± 0.145 vs. 1.002 ± 0.141 g/cm²; CG: 1.000 ± 0.122 vs. 0.989 ± 0.130 g/cm²; p = 0.848), or whole-body (EX: 1.189 ± 0.115 vs. 1.179 ± 0.116 g/cm²; CG: 1.161 ± 0.121 vs. 1.149 ± 0.118 g/cm²; p = 0.827) compared to the controls in 104 patients with prostate cancer [38]. Similarly, 6 months of 150 min per week of aerobic training did not increase whole-body BMD (EX: 1.152 ± 0.109 vs. 0.144 ± 0.119 g/cm²; CG: 1.120 ± 0.094 vs. 1.112 ± 0.101 g/cm²; p = 0.97), but a between-group difference in favor of exercise training was observed at the 12-month follow-up (EX: 1.159 ± 0.112 vs. 0.167 ± 0.126 g/cm²; CG: 1.122 ± 0.098 vs. 1.097 ± 0.105 g/cm²; p = 0.043) [30]. By contrast, gain in lumbar spine (n = 2 studies) and total hip (n = 2 studies) BMD was observed. In one study, 30 patients with osteopenia and breast cancer undertook 3 months of supervised weight bearing training performed thrice weekly in addition to vitamin D and calcium supplementation and exhibited significant improvement in lumbar spine BMD compared to nutritional supplementation alone (EX: 0.48 ± 0.09 vs. 0.66 ± 0.09 g/cm²; CG: 0.46 ± 0.08 vs. 0.51 ± 0.07 g/cm²; p = 0.001) [29].

BMC did not exhibit between-group changes in the one study that evaluated this parameter [30]. Regarding bone turnover markers, bone ALP and NTX-I showed an improvement in one study [33] and no change in the other three [28, 31, 34, 38], whereas no changes were observed for deoxypyridinoline, osteocalcin, and TRACP-5b [28, 35].

Six studies explored the impact of exercise on bone health outcomes in patients with bone metastases [23‐26, 39, 40]. Patients had mainly metastatic prostate cancer, and the most affected bone regions were the thoracic spine (46.7% to 82.8%) and the lumbar spine (16.0% to 43.3%).

Focusing on BMD evaluated with DEXA, the investigations did not find any significant differences for total body (n = 2 studies) [23, 25], total hip (n = 2 studies) [23, 24], lumbar spine (n = 3 studies) [23, 24, 26], or femoral neck (n = 2 studies) [23, 26], as a result of exercise intervention. The following trials evaluated the impact of football intervention, in patients with metastatic prostate cancer, and neither 3-, 6- or 8-months of training resulted in any significant increase in BMD at any site, except for one study in which a gain in total hip BMD was observed after 8-month of exercise [(right hip EX: 0.992 ± 0.135 vs. 0.999 ± 0.129 g/cm²; CG: 1.031 ± 0.149 vs. 1.011 ± 0.139 g/cm²; p = 0.015)(left hip EX: 0.993 ± 0.129 vs. 1.002 ± 0.126 g/cm²; CG: 1.024 ± 0.151 vs. 1.016 ± 0.140 g/cm²; p = 0.030)] [26].

Interestingly, two studies explored the impact of training on BMD at the vertebral metastatic site via CT scans, reporting opposite results [39, 40]. Both trials adopted a similar intervention protocol, consisting of 6-months isometric resistance training performed 3–5 times weekly. The first study on 60 patients with mixed cancer types found a significant increase in BMD in all metastatic sites (EX: + 80.3% vs. CG: + 17.3%; p < 0.01) in favour of the exercise group compared to the controls [40]. However, in another study in 56 patients with mixed cancer types, the same intervention did not result in improvement in the bone metastatic BMD (EX: + 67.8% vs. CG: + 49.8%; p = 0.964) [39].

Regarding BMC, one investigation of 57 patients with prostate cancer reported an increase in the total body BMC [25], but this finding was not confirmed by a larger trial involving 214 patients [23]. Bone turnover markers were assessed in two investigations, reporting conflicting results for CTX-I, osteocalcin, PINP, bone ALP, NTX-I, deoxypyridinoline, and pyridinoline [25, 26, 40].

This systematic review and meta-analysis aimed to evaluate the effects of exercise interventions on bone health in patients with cancer and existing bone impairments. Overall, our findings revealed no significant effect of exercise on BMD or BMC at key anatomical sites (i.e., whole-body, lumbar spine, hip, femoral neck), and no consistent impact on bone turnover markers. While some individual studies reported within-group improvements, these did not translate into statistically significant pooled effects when compared to controls. Previous meta-analyses, although mxed, reported more favorable outcomes for exercise interventions, but they mostly included patients with early-stage cancer and no pre-existing bone impairment.. For instance, Rose et al., examined 26 studies involving patients with early-stage breast cancer mainly without bone dysfunction, finding a significant improvement in whole-body BMD (SMD = 0.25, 95%CI: 0.05 to 0.46), hip (SMD = 0.20, 95%CI: 0.04 to 0.36), trochanter (SMD = 0.19, 95%CI: 0.06 to 0.33), and femoral neck (SMD = 0.39, 95%CI:0.07 to0.71), but not in lumbar spine BMD (SMD = 0.40, 95%CI: −0.11 to 0.92) and whole-body BMC (SMD = 0.06, 95%CI: −0.17 to 0.28) [20]. Similarly, the meta-analysis of Singh et al. on 22 trials, mainly including early-stage breast and prostate cancer having normal bone health, confirmed the positive impact on hip BMD but not on whole-body BMD, while they found an improvement in the lumbar spine (SMD = 0.269, 95%CI: 0.036 to 0.501) [18]. A third study, exploring the effect of at least 6-months exercise training in patients who had completed anticancer treatment, revealed no effects on lumbar spine, hip, or femoral neck BMD compared to usual care [19]. In terms of BMC, our finding is in line with prior research, in which no statistical change was produced exercise training [18].

In contrast, our review focused on a high-risk population where reversing or slowing established bone deterioration may be more challenging. This distinction may explain the differences in findings and highlights the importance of stratifying analyses by baseline bone status. On the other hand, it could be possible that the traditional exercise training protocol (i.e., combined aerobic and resistance training at moderate intensity) may be more effective as a prevention strategy rather than a management strategy for bone deterioration, thus necessitating more specific adaptations. Critically, low-to-moderate intensity aerobic training, which was commonly implemented in the studies included in our review, is likely insufficient to reverse bone loss in the target populations. Only two studies, implemented specific modes and dosages to improve bone mass compared to previous investigations which mainly explored a combined aerobic, resistance, and impact training at moderate-to-vigorous intensity [18]. Apart from the type and intensity, some interventions were of short duration (> 4 weeks), which may be insufficient to induce detectable changes in bone structure. Thus, intervention length should be carefully considered in both future trials and clinical recommendations. Because of the heterogeneity of the exercise interventions, benefits of certain exercise programs may have been masked by other programs that could not be expected to have any influence on bone.

The current recommendations for exercise in patients with bone metastases clearly state that the approaches tested so far are likely to be too conservative, suggesting movements even more restrictive compared to many daily living activities, e.g., descending stairs [41]. In fact, the available evidence suggests that exercise prescriptions should be systematically tailored based on lesion location, avoiding resistance exercises that impose mechanical load on the affected area and weight-bearing activities (e.g., walking) when metastases are present in the lumbar spine, proximal femur, or pelvis [42‐45]. This may reinforce our previous speculation, suggesting that, despite the higher safety profile of this approach, there is an insufficient exercise load that, instead, is needed to increase BMD. Indeed, the same recommendations highlight how these are only a starting point for a future research area directed to evaluate specific programs to more effectively manage bone dysfunction in cancer. There remains considerable caution in prescribing specific bone-stimulating exercise such as heavy resistance training and impact loading due to concerns of fracture and other adverse events. This conservative approach may be inhibiting the demonstration of benefit.

From these perspectives, and based on the current evidence, a tailored exercise prescription aiming at counteracting bone disorders in cancer should be based on the following principles: type, intensity, duration, frequency, and length of the intervention. Regarding the specific type of activities, those applying mechanical stimuli have been demonstrated to remodel bone turnover to a more osteogenic capacity [1, 46]. Therefore, antigravity exercises are considered the most appropriate to achieve an anabolic bone response [47]. Theoretically, resistance training, e.g., multiarticular weight-lifting or body-weight exercises, and impact training, e.g., plyometric activities involving high rates of force change such as foot-stamping, heel drops, hops and jumping, can produce the high mechanical strain necessary to stimulate the cells to remodel the bone [47]. Regarding the intensity of resistance training, a progressive increase in load starting at least from 50% 1RM has been proposed as a potential strategy to maximize the bone response. However, the safety and applicability of high-intensity protocols in higher-risk populations, such as older patients with cancer, previous fractures, or bone pain, remain uncertain. For instance, the LIFTMOR trial involving 101 participants with osteoporosis, implemented 8 months of high-intensity resistance and impact training (≥ 80–85% 1RM, including deadlifts, overhead presses, and squats) and found no fractures or other adverse events, compliance of 92%, and a significant increase in lumber spine (+ 2.9% vs. −1.2%, p < 0.001) and femoral neck (0.3% vs. −1.9%, p = 0.004) BMD compared to controls [48]. The highly exercise-specific adaptation of bone was highlighted in the study of Newton et al., which demonstrated that traditional high-intensity resistance training was insufficient to slow bone loss in patients receiving ADT for prostate cancer. The addition of the impact loading protocol was essential to decrease the rate of bone loss due to ADT and importantly the group undertaking aerobic exercise only exhibited the greatest decline [49]. The need for caution in interpreting such recommendations impact-loading training needs to be acknowledged. While encouraging, these findings may not be generalizable to frailer populations or those with cancer-related bone involvement. A case report by Rosenberger et al. documented a vertebral fracture during 1RM testing in a 69-year-old breast cancer survivor who had recently completed treatment and had no known history of osteoporosis, raising important concerns not only about the exercise prescription itself, but also about the potential risks associated with certain physical assessments used to guide such programs [50]. Current clinical guidelines recommend avoiding mechanical loading on sites affected by metastases [37‐40], and some patients may be at increased risk of fracture. However future studies are needed to explore the real contribution of resistance and high-intensity training as well as to determine the relation between adherence and the effect on bone health, since most investigations do not report these data. The duration of the sessions, especially when combined with intensity, should be sufficient to have an acceptable “dosage” to induce an initially bone reabsorption, which should be followed by an increased bone formation during the recovery phase [51]. This concept, called supercompensation, introduces the importance of an appropriate prescription in terms of frequency and the necessity of balancing training days with adequate periods of rest. Indeed, bone remodeling requires at least 24 h of recovery to respond optimally to the workload [51]. The last component is the length of the intervention. Owing to the subgroup analysis of Singh et al. it is evident that interventions longer than 12 weeks were superior to the shorter ones, it can be speculated that prolonging the length may be essential particularly when the bone impairment already exists. Increasing the duration of the training exposure may allow the bone remodeling cycle to shift more towards bone formation, which could thus be detected by radiological evaluations. This hypothetical prescription will have to be further refined considering the type of impairment, i.e., bone loss and/or metastases, and the disease and patients’ characteristics, e.g., age, sex, limitations, etc., but is an important starting point for future research.

Some limitations of the present work warrant discussion. The limited number of studies has restricted the possibility to perform subgroup analyses, especially in terms of type and stage of oncological disease, bone impairments, and exercise prescription. In this light, we have decided to perform the literature search based on peer-reviewed, published articles, potentially excluding studies included in the gray literature. In addition, the certainty of the evidence ranged from low to very low across all outcomes. These limitations suggest that current evidence is insufficiently robust to support strong conclusions and highlight the need for higher-quality studies in this area. Moreover, there was substantial variability in patient characteristics, as well as in the design and prescription of exercise interventions which limited the interpretability of the pooled results. Finally, this review focused exclusively on bone-related outcomes. However, it is well established that fracture risk is multifactorial, and many fractures occur in patients without a diagnosis of osteoporosis. Functional limitations, muscle weakness, balance impairment, and increased fall risk are all relevant contributors and by not including these outcomes the present review may underestimate the broader potential benefits of exercise in this population. The main strength of this work is the inclusion of patients with bone health issues, which, until now, represented a gap in the literature.

The findings of this review and meta-analysis are that when combined across all the exercise intervention studies to date, BMD in patients with cancer and osteopenia, osteoporosis, and bone metastases do not seem to benefit. As with all other adaptations to exercise, the effects are highly specific to the type, dosage, and duration. Exercise as a treatment for people with these conditions requires more nuanced prescription, and future research should focus on incorporating higher intensities of resistance training with the incorporation of impact loading exercises to optimize bone health in these populations.