Background

The age-related loss of skeletal muscle and bone, i.e., sarcopenia and osteoporosis, both represent a substantial and growing public health issue as both conditions individually contribute to frailty, poor balance, fall and fractures [1,2,3,4]. Back in 2009, Binkley et al. [5], classified individuals with both conditions as “Sarco-osteopenic” and suggested that the combination of these two conditions could be expected to identify individuals at a higher fracture risk than currently thought.

Definitions

Sarcopenia was first named by Rosenberg in 1989 and referred to the putative age-dependent decline in muscle mass [6]. In 1998, Baumgartner et al. [7] defined the first cut-off values for sarcopenia based on whole body dual-energy X-ray absorptiometry (DXA) of relative appendicular lean mass (RaLM) as 2 standard deviations (SD) below the mean of a young reference group (women: 5.45 kg/m2; men: 7.26 kg/m2). Although the cut-off values are still widely used, the current recommendations suggest parameters of muscle strength and physical performance to be incorporated in the definition of sarcopenia [7,8,9,10,11]. Consequently, The European Working Group on Sarcopenia in Older People (EWGSOP) suggested a conceptual staging of sarcopenia as pre-sarcopenia (low muscle mass only), sarcopenia (low muscle mass plus either low muscle strength or low physical performance) or severe sarcopenia (low muscle mass, strength and physical performance) [8]. In parallel, The International Working Group on Sarcopenia (IWG/IWGS) have suggested a definition based on appendicular muscle mass (adjusted for squared height) and gait speed (GS), whereas The Foundation of the National Institute of Health and Sarcopenia Project (FNIH) defines sarcopenia based on appendicular muscle mass (adjusted for BMI) and hand grip strength (HGS) [9, 10]. Despite an overall agreement that sarcopenia should be defined from a combination of low muscle mass, strength and/or function, the introduction of a wide number of different cut-off values makes it difficult to compare results from different populations and studies as the prevalence varies significantly depending on the cut-off values used [12, 13].

In contrast, there has been consensus regarding the definition of osteoporosis since 1994, where WHO defined a T-score (i.e., the SD of bone mineral density (BMD) in reference to the mean of healthy young adults of the same gender) between − 1.0 and − 2.5 and a T-score ≤ − 2.5 as osteopenia and asymptomatic osteoporosis, respectively. Low-energy fracture of the hip or spine defined symptomatic osteoporosis even with a normal BMD T-score [14].

As a more recent approach, Binkley et al. [15] introduced an even more comprehensive syndrome, the “Dysmobility Syndrome” to reflect a geriatric condition, identifying patients at high risk of adverse musculoskeletal conditions, in line with the metabolic syndrome in patients with increased risk of cardiovascular diseases. The syndrome includes osteoporosis, low muscle mass, history of fall, slow GS, low HGS and high fat mass—all components included in an arbitrary equally weighted scale—and has been demonstrated to better identify individuals with a previous fall or fracture than the sarcopenia definitions alone [15]. This finding is supported in a new review by Hill et al. [16], who found the dysmobility syndrome to be associated with a functional decline, increased number of falls and fractures, as well as increased mortality.

Prevalence

Both sarcopenia and osteoporosis are common conditions in the older population. The International Sarcopenia Initiative, based on EWGSOP criteria, found sarcopenia to be present in 1–29% of the home-dwelling populations, in 14–33% of the long-term care population and in 10% of acute hospitalized geriatric patients [17]. In a more recent paper, though, the prevalence of sarcopenia in patients hospitalized at a geriatric ward was 6.6 and 18.7% sarcopenia and severe sarcopenia, respectively [18]. In line, the “EU27-report” from 2013 stated that the prevalence of osteoporosis in Europe in 2010 was approximately 22.0 million (22.1%) and 5.5 million (6.6%) in 50- to 84-year-old women and men, respectively [19]. Furthermore, the annual fracture incidence related to osteoporosis is approximately 8.9 million worldwide, with a higher incidence in Scandinavia [20].

The linkage of bones and muscle disorder

The genesis of both osteoporosis and sarcopenia are multifactorial, and it is believed that some of the known factors causing osteoporosis also contribute to the development of sarcopenia [21, 22]. Notably, bone and muscle tissue mass are tightly correlated throughout life and share environmental, endocrine and paracrine influences [23]. Moreover, parallel changes in muscle and bone mass are known to be brought about by exercise, disuse and aging [21]. “The mechanostat theory”, e.g., states that a decline in mechanical loading reduces the bone formation leading to a fragile bone status [24,25,26,27]. Thus, muscle has long been recognized as the primary source of anabolic mechanical stimuli for bone tissue [23]. However, the precise mechanisms responsible for orchestrating bone and skeletal mass are still not well-characterized and little is known about potential mechanisms other than loading [10]. Secreted factors such as myostatin, activins and pro-inflammatory cytokines represent potential underlying common mechanisms linking bone and muscle, yet little is known regarding how these factors work and affect muscle and bone mass. However, emerging data supports the concept that muscle secretes factors that target other tissues and are involved in glucose metabolism while simultaneously similar data has been emerging for bone [28]. An integrated physiology is therefore hypothesized to exist between muscle and bone tissue.

Another key player in the muscle bone cross-talk is thought to be the Wnt/β-catenin signaling pathway, a major regulator of bone mass and of muscle development and growth. The regulation of this pathway by osteocytes may play a central role in regulating bone mass [22]. In particular, sclerostin (encoded by the SOST gene), is highly express in osteocytes and inhibits Wnt signaling [29].

The consequences of sarcopenia and osteoporosis

Falls and/or fractures leads to immobilization, hospitalization, impaired level of function and a markedly increased risk of mortality. Consequently, sarcopenia [30,31,32,33,34] and osteoporosis [19, 35,36,37] creates significant economic, societal, and social burden for individuals. This burden is expected to increase with the increasing life expectancy.

The future therefore brings challenges: the increasing aging population and their increased demands for healthy old age. In this context, clinicians, prevention, early detection and better treatment of both sarcopenia and osteoporosis may be necessary to meet these demands.

Aim and research question

The aim of the present systematic review was to provide an overview of the current knowledge of sarcopenia and osteoporosis in older Caucasians (aged 65 years or older) as well as a meta-analysis of studies comparing sarcopenic with non-sarcopenic patients with regard to BMD/T-score and risk of low-energy fractures.

Current knowledge of the following topics is sought:

  1. 1.

    The prevalence of sarcopenia among older people suffering from osteopenia, osteoporosis and/or low-energy fractures.

  2. 2.

    Fracture risk assessment among older people suffering from sarcopenia and symptomatic osteoporosis.

  3. 3.

    Markers of bone turnover as assessed in osteosarcopenic older people compared to non-sarcopenic osteoporotic or osteopenic.

Method

Search strategy and study eligibility

The following search queries in PubMed until the 10th of March 2018 were used with the filters “abstract and full text available” added.

Search strings: “Osteoporosis OR Osteopenia AND Sarcopenia AND older”; “Fracture AND Sarcopenia AND Older”; “Osteoporosis OR Osteopenia AND Sarcopenia AND Bone Markers”; “Fracture AND Sarcopenia AND Bone Markers”; and “BMD AND Sarcopenia AND Bone Markers”. The results were exported to EndNote X8 (Clarivate Analytics 2017; Philadelphia; PA 19130; USA) to remove duplicates. Title and abstracts were screened for eligibility with regard to the following inclusion criteria: “original papers only”; “human studies only”; “mean age ≥ 65 years”; “Caucasian ancestry only”; and as well as “primary focus on osteosarcopenia”. Full-text papers were screened according to the same inclusion criteria. Afterwards, additional records were screened and identified through other sources as described above.

Data extraction for meta-analysis

Eligible studies were screened to provide data comparing sarcopenic vs non-sarcopenic patients with regard to osteopenia/osteoporosis from baseline BMD or T-score as well as with regard to fractures. Studies on fractures were required to provide numbers of fracture events in both groups. Studies not providing such data were excluded.

All demographic data and prevalence of sarcopenia were extracted from each study. In fracture studies wherein all patients had fractures, the number (prevalence) of sarcopenic participants were calculated as well. Percentages were recalculated back to absolute numbers and medians with ranges were recalculated to their means with standard deviations wherever required. In case of missing information the author of the study was contacted. The first author (BRN) conducted the primary search and data extraction, which were subsequently confirmed by an experienced author (JA). Disagreements were solved by consensus.

Statistics

Weighted mean difference method was used to compare the baseline BMD and T-score in sarcopenic vs non-sarcopenic participants, by pooling their means and standard deviations. The relative risk (RR) ratio with 95% confidence interval (CI) of fractures was estimated by pooling the number of events and total populations together in both groups. The prevalence of sarcopenia in the included studies was pooled as proportions. Heterogeneity among studies was tested using the Chi squared method and the I2 statistic. We used fixed effect model in all analyses to restore the real sizes of the influential studies. All analyses were performed with 95% CI and all p values < 0.05 were considered statistically significant. All analyses and plots were performed using the meta-analysis package of the statistic software program STATA version 15 (STATA Corporation, Lakeway Drive, College Station, TX, USA).

Results

Twenty-seven original papers were included after assessment of eligibility. Of those, 17 were included in the meta-analysis. The results of the search strategy are presented in the PRISMA flow chart [38] in Fig. 1 and selected data from the papers included are presented in Table 1.

Fig. 1
figure 1

Flow chart of data search, screening and evaluation of eligibility

Full size image
Table 1 A list of the 27 original articles included in the systematic review and the reference numbers in the first column corresponds to the reference numbers in the paper
Full size table

Overall, study settings were very diverse due to heterogeneous populations, designs and methodology. The populations comprised healthy volunteers; volunteers with increased risk of fall; participants referred to outpatient clinics treating osteoporosis and preventing fall; as well as patients with verified low-energy fractures. The designs were mostly cross-sectional; however, few prospective observational designs were present (Table 1). Furthermore, the methods and classification used to determine sarcopenia were diverse, even though a greater consensus appeared after the introduction of the EWGSOP recommendations [8]. The main difference though was whether sarcopenia was classified with regard to muscle mass only or whether strength and/or physical performance were included (Table 1).

The associations of sarcopenia with low bone mass and fracture in older people

Prevalence

The overall prevalence of osteosarcopenia (from T-score and diverse sarcopenic classifications) in the cohorts including both gender varied between 5.0 and 37.0% [39,40,41,42,43]. In studies including women only, osteosarcopenia varied between 0.5 and 45.0% [2, 44, 45]. However, the prevalence of osteosarcopenia seemed to vary depending on whether participants were classified primarily with sarcopenia or osteopenia/osteoporosis. Notably, there was a tendency, that in subjects classified as sarcopenic, osteopenia was more frequent than osteoporosis. In a population of 171 women and 118 men, it was observed that sarcopenia was more frequently associated with osteopenia than osteoporosis, 81.3 vs 30.2% [39]. This finding was supported by Kirchengast et al. [46] who found an increased prevalence of osteopenia (women = 58.8%; men = 50.0%) compared to osteoporosis (women = 25.5%; men = 16.7%), in sarcopenic women and men. On the other hand, in participants suffering from osteoporosis Locquet et al. [39] and Genaro et al. [47] found 36.1 and 21.4%, respectively, to be sarcopenic. In line with these results, Huo et al. [41] found sarcopenia to be more frequent in osteoporotic (62.7%) vs osteopenic (47.7%) participants, as well as Gillette-Guyonnett et al. [48] who found sarcopenia to be present in 32.2% of the osteoporotic participants vs 26.2% in non-osteoporotic participants.

In general, the relative percentage of participants with osteosarcopenia is greater in women (25.5–82.6%) compared to men (16.4–32.0%) [40, 41, 46].

Another general finding is a marked variance in the prevalence of sarcopenia in low-energy fracture participants depending on the studied population. In participants with low-energy osteoporotic fractures, sarcopenia was present in 7.8–58% and 1.3–96.3% of the cases, women and men, respectively [2, 32, 49,50,51,52,53]. Noteworthy, the very high prevalence in men were from small-size studies based on the initial definition of sarcopenia [52]. When both genders were included, sarcopenia was present in 17.1–58% [39, 54,55,56,57] of participants with fractures. Interestingly, it has been demonstrated that participants with dysmobility syndrome, i.e., the suggested geriatric syndrome, are more likely to have incurred a prior facture (66.2%), whereas participants with sarcopenia (FNIH classification) and severe sarcopenia (EWGSOP classification) had suffered a prior fracture in only 7.8 and 10.4% of the cases, respectively (Table 1) [49]. This finding may point towards the dysmobility syndrome as a more sensitive marker of frailty.

Meta-analysis of the pooled mean prevalence of sarcopenia in patient populations with fractures is 46% (95% CI 44, 48; p < 0.001) (Fig. 2), although, the heterogeneity of the included studies was significant.

Fig. 2
figure 2

Meta-analysis of the prevalence of sarcopenic patients in populations presented with fractures. CI confidence interval, ES estimated size

Full size image

The included studies demonstrated a substantial difference in observed fracture incidences among the old patients and the really old patients. Patients with a mean age of 67.2 years had a history of fractures in 7.8–10.4% of sarcopenic cases [49], whereas sarcopenic participants with mean age between 80 and 85 years had a fracture incidence that ranged between 17 and 58% [2, 54, 55]. This difference may solely be ascribed to the aging bone, though.

Risk of low bone mineral density/T-score and fracture with regard to sarcopenia

Of the 27 original studies included in this review, 17 studies addressed BMD/T-score (8 eligible for meta-analysis) and 16 studies addressed low-energy fractures (4 were eligible for meta-analysis) (Table 1).

In our meta-analysis, we found statistically significant lower weighted mean differences of BMD (0.07 g/cm2; p < 0.001) and T-score (0.34; p < 0.001) in sarcopenic compared to non-sarcopenic individuals (Fig. 3).

Fig. 3
figure 3

Meta-analysis of the weighted mean differences of baseline BMD and T-score of femoral neck comparing sarcopenic with non-sarcopenic patients. BMD bone mineral density, CI confidence interval, WMD weighted mean difference

Full size image

Moreover, there was an increased risk of osteoporotic fractures in sarcopenic compared to non-sarcopenic participants with a relative risk of 1.37 (95% CI 1.18, 1.59; p < 0.001) (Fig. 4). However, the heterogeneity between studies was significant (p < 0.001).

Fig. 4
figure 4

Meta-analysis of the risk of fracture in sarcopenic versus non-sarcopenic. N number, F fracture, CI confidence interval, RR relative risk

Full size image

Studies on fracture, not eligible for meta-analysis, estimated the odds ratio (OR) of osteoporotic fractures in sarcopenic vs non-sarcopenic at 2.43 (95% CI 1.12, 5.27; p < 0.05) and 2.7 (95% CI 1.4, 5.5; p = 0.005), respectively [58, 59]. Considering osteosarcopenic vs normal BMD/normal muscle mass—participants, the OR of fracture was estimated by Hars et al. [58] and Huo et al. [41] at 3.39 (95% CI 1.54, 7.45; p = 0.002) and 2.71 (95% CI 1.7, 4.4; p < 0.001), respectively. One study, however, estimated the hazard ratio of fracture in sarcopenic and non-sarcopenic participants and did not find a significant difference (Hazard ratio 1.53, 95% CI 0.70, 3.31) [60].

Osteosarcopenia and markers of bone turnover

Markers of bone turnover was assessed in two of the included 27 studies within the given search criteria. In the work by Gonelli et al. [61] markers of bone turnover were included in the regression models to predict BMD at different skeletal sites from different body composition measures, however, it was not explored whether markers of bone turnover were altered in osteosarcopenic compared to osteopenic or sarcopenic participants. In the study by Drey et al. [40] osteosarcopenic participants had increased markers of bone turnover compared to controls, in contrast to osteopenic and sarcopenic participants. Consequently, it was hypothesized, that bone loss was faster in osteosarcopenic compared to participants with only one condition (sarcopenia or osteopenia).

Discussion

In summary, the presented systematic review and meta-analysis demonstrated a high prevalence of osteosarcopenia in various geriatric populations. Somewhat surprisingly, the prevalence of sarcopenia combined with osteopenia was in some studies more frequent, compared to sarcopenia with osteoporosis [39, 46], whereas others has found sarcopenia to more frequent in osteoporotic participants, as expected [41]. This finding could partly be explained by the different classifications of sarcopenia used. Despite the various definitions of sarcopenia, the present meta-analysis showed a high prevalence of sarcopenia in patients with osteoporotic fractures [2, 49,50,51,52, 54,55,56,57]. Importantly, the risk of incurring a low-energy fracture when being sarcopenic was much higher compared to non-sarcopenic controls [32, 39, 41, 58]. Moreover, BMD and T-scores were lower in sarcopenic subjects [32, 39, 41, 47, 58, 62, 63], although there was a high heterogeneity among the included studies.

Future studies on the prevalence of osteopenia and osteoporosis with sarcopenia within different study populations (healthy volunteers, outpatient clinics and admitted patients) are, however, needed to understand the true relationship.

In a clinical setting, the association of markers of bone turnover and importance in muscle metabolism is somewhat unclear and the hypothesis of an association may be contradictory. To our knowledge, only one study in humans have reported increased plasma bone turnover markers in osteosarcopenic participants as compared with either sarcopenic or osteopenic participants, indicating that markers of bone turnover might supplement other clinical and para-clinical biomarkers in identifying participants at high risk of bone fractures [40]. This observation is supported by an in vitro study investigating cultured osteoblasts with serum from participants classified as normal, obese, obese osteopenic, obese sarcopenic and obese osteopenic sarcopenic [64]. In this study the authors found that the RunX2 (an essential transcription factor in osteoblast maturation) level was decreased in all pathological groups compared to healthy controls. However, the osteoblast marker osteocalcin was only altered in obese participants and not in the obese osteopenic sarcopenic participants indicating that different amounts of adipose tissue and muscle mass may alter the bone biology [64]. This could be of interest in future studies on markers of bone turnover, osteosarcopenia and whole body composition with increased age.

In general, it was difficult to compare the evaluated studies in this review due to the lack of consensus regarding the definition of sarcopenia. Since 2010 though, more and more studies rely on the sarcopenia classification including muscle mass, strength and/or physical performance, as suggested by the EWGSOP [8]. However, even by this simple classification a lot of different tests and cut-offs are suggested making the conclusions tricky, as discussed by Bijlsma et al. [65]. In a new review by Reijnierse et al. [12], including patients from a geriatric outpatient clinic, three widely used definitions of sarcopenia did not capture the same patients. With regard to fractures, this was supported by Cawthon et al. [32] who found that participants classified from the Baumgartner and Newman criteria did not have an increased risk of osteoporotic fractures, but participants classified from EWGSOP, IWG and NHIH criteria were more exposed. As a consequence of the lack of a well-defined definition of sarcopenia, the bone field is far more advanced than the muscle field in clinical terms. In contrast to the diagnostics and treatment of sarcopenia, the bone field has been successful in developing therapeutics for prevention and treatment of osteoporosis on the basis of well-defined clear parameters, while the definition of sarcopenia still remains somewhat unsettled [9, 66].

Strengths and limitations

In the present systematic review, both cross-sectional data and prospective studies were included, well aware of the higher evidence level (and the opportunity to consider causality) from the latter design which in this context expresses the relevant outcome “fracture” due to osteosarcopenia. However, regarding participants of interest, most studies were population surveys including home-dwelling older individuals with a high functional level, but also more frail patients groups (patients referred to outpatient clinics and hip-fracture patients) were represented. A great strength of the present study was that we were able to perform 3 meta-analyses, which enabled a comparison.

What could be of interest in future studies?

Studies to gain consensus on the classification of sarcopenia in relevant clinical settings is highly warranted. To gain more knowledge about the underlying mechanisms leading to sarcopenia and osteosarcopenia, biomarkers of bone and muscle should be included in future studies.

Of special interest in the clinical setting may be the dysmobility syndrome, which has been shown to be associated with a functional decline, increased number of falls and fractures, as well as increased mortality [16]. This syndrome may pinpoint participants at high risk of bone fracture and functional disability better than participants with osteosarcopenia, sarcopenia or osteopenia alone. No intervention studies, however, are currently present, as well as prospective studies are limited.

The high prevalence of osteosarcopenia and its consequences may very well alter future assessment of osteoporotic patient to include assessment of sarcopenia to focus on sarcopenic prevention strategies beside the well-known medical treatment of osteoporosis.

Conclusion

The concept “osteosarcopenia” was very frequent in older people and might be a better predictor of physical decline, falls and fractures than osteopenia or sarcopenia alone. From meta-analysis, the relative risk of fracture was higher in sarcopenic subjects making this field interesting in future studies. However, the lack of consensus regarding the classification of sarcopenia is a large barrier for clarification and future treatment targets.