Effects of Statins on Bone Mineral Density and Fracture Risk: A PRISMA-compliant Systematic Review and Meta-Analysis.

Although observational studies have identified the protective effect of statins on bone health, the effects remain controversial in randomized controlled trials (RCTs). We conducted a meta-analysis of RCTs to evaluate the effects of statins on bone mineral density (BMD) and fracture risk among adults.We searched electronic databases of Medline, Embase, and the Cochrane Central Register of Controlled Trials (CENTRAL) and conducted a bibliography review to identify articles published until May, 2015.Studies included in this meta-analysis should be randomized controlled trials conducted in adults, using statins in the intervention group. Information on changes in BMD or odds ratio, relative risk or hazard ratio (HR) for fracture risk with the corresponding 95% confidence interval (CI) was provided.Two investigators independently reviewed the title or abstract, further reviewed the full-texts and extracted information on study characteristics and study outcomes. Net change estimates of BMD and pooled HR of fracture risk comparing the intervention group with the control group were estimated across trials using random-effects models.Of the relevant 334 citations, 7 trials (including 27,900 randomized participants in total) meeting the eligibility criteria were included. Of the 7 trials, 5 were conducted to assess the association of statins use with BMD change and 2 with fracture risk. Compared with the control group, statins use was associated with significant increase in BMD of 0.03 g/cm (95% CI: 0.006, 0.053; I = 99.2%; P < 0.001), but null association with fracture risk, with the pooled HR of 1.00 (95% CI: 0.87, 1.15; I = 0; P = 0.396). Sensitivity analyses revealed that the associations were consistent and robust.The effect of statins use on bone health among subpopulation could not be identified due to limited number of trials.These findings provide evidence that statins could be used to increase BMD other than decreasing fracture risk in participant with dyslipidemia. In addition, further trials with the primary outcome of bone health-related measurements in subpopulation are warranted to ensure the effect of statins use.

INTRODUCTION

Osteoporosis, which is defined based on bone mineral density (BMD), is a skeletal disorder characterized by compromised bone strength. It is induced by an imbalance between osteoblastic bone formation and osteoclastic bone resorption.[1] Osteoporosis, which is a process operative in almost all individuals past middle age, will greatly increase the risk of fractures in both men and women.[2,3] Osteoporosis or osteoporotic fracture is also the great disease burden in an aging population due to their association with increased mortality and substantial long-term loss of independence.[4] It has been demonstrated that osteoporosis causes more disability-adjusted life years loss than any type of cancer other than lung cancer.[5] Therefore, a cost-effectiveness of treatment on osteoporosis should be considered.

Cardiovascular diseases are also age-related disease and several epidemiologic studies have identified that they may share common biological pathways.[6,7] The 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) are widely used for primary prevention of cardiovascular disease.[2,8,9] Previous studies have revealed the positive biologic effects of statins on bone, including simulating bone formation[10] and sharing the same pathway as nitrogen-containing bisphosphonate drugs.[11] In addition, the pleiotropic effect has attracted particular attention of statins on bone metabolism. Therefore, statins might be clinically significant in the prevention and treatment of osteoporosis. Furthermore, some in vitro and in vivo experiments have investigated the mechanism of statins influencing bone metabolism.[12,13] The positive effects of statins on osteoblast differentiation and bone formation[14-16] have been identified to be related with the inhibition of the isoprenoid biosynthetic pathway. Therefore, the depletion of GGPP, especially FPP, may be necessary for statin-induced bone formation.[17] Moreover, simvastatin was proved to be involved in the inhibition of receptor activator of nuclear factor-κB ligand (RANKL)-induced osteoclast differentiation by preventing the production of reactive oxygen species (ROS).[18]

Several observational studies have found the association of statins use with improved BMD,[19-21] as well as reduced risk of fractures.[22-26] However, some other observational studies and post hoc analysis of randomized clinical trials (RCTs) did not find consistent results.[27-30] Due to controversial results and cumulative reports of RCTs on the association of statins use with osteoporosis-related measurement, we performed the meta-analysis to explore the association of statins use with BMD and fracture risk and provide evidence for the treatment of osteoporosis or improvement of bone health.

METHODS

We conducted the literature search, study selection, data extraction, and results synthesis following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.[31] The PRISMA checklist is shown in the appendix.

Search Strategy and Study Selection

A literature search was conducted in electronic databases, including PubMed (1966 to May, 2015), Embase (1947 to May, 2015), and the Cochrane Central Register of Controlled Trials (CENTRAL) (issue April, 2015) for articles examining the association of statins use with BMD and bone fracture without language restriction. Detailed search strategies for 3 databases are shown in the Supplementary materials. Briefly, the following search terms were included in our literature search strategy: “statin,” “bone mineral density,” “bone fracture,” and “osteoporosis.” In addition, reference lists from all eligible articles, reviews, systematic reviews, and meta-analyses were also searched to identify relevant articles.

After removing duplicates, 2 investigators reviewed the articles independently and discrepancies regarding study eligibility were discussed with another investigator. The inclusion criteria of eligible articles are as follows: adults participants aged 18 years or older; statins were used as the intervention or at least part of the intervention; changes in BMD and corresponding variance or confidence interval (CI) or information which could be used to calculate above indicator were provided, or odds ratio, relative risk, or hazard ratio (HR) with their corresponding 95% CI for fracture risk was provided; and randomization was used to conduct group allocation. Articles with latest information were included, if several articles were generated from the same study. Ethical approval was not necessary for the current meta-analysis.

Data Extraction and Quality Assessment

Two investigators also conducted data extraction and quality assessment independently and further discussed with another investigator for discrepancies. The following data were extracted: title of articles, authors, year of publication, name of the trial, study design (primary outcome of study, randomization, and blinding), participants’ characteristics, intervention drug and corresponding dose, information on BMD or bone fracture and outcome measurement, and statistical analysis methods. The Jadad score was used to assess the quality of included studies. The scoring system included randomization, blinding, description of drop-out and withdrawal, and evaluation of randomization and blinding.[32]

Data Synthesis and Statistical Analysis

For each RCT, if net effect size of BMD was not provided, it was calculated as the change in BMD-related measures (from the baseline to the end of intervention) in the intervention group minus the change in BMD in the control group: . For studies without variance data, we calculated variance from CIs or test statistics. If the variance for change between baseline and end of intervention was not reported, it was calculated from the following equation[33]: where corresponds to the variance at baseline, corresponds to the variance at the end of intervention, and an imputed of 0.5 is the correlation coefficient between measurements at baseline and the end of intervention.[34]

We used random-effect models to estimate BMD net change or pooled HR of fracture risk across trials. Heterogeneity across studies was assessed by the Cochrane Q and the I2 statistics.[35] We conducted influence analysis by removing each trial sequentially to determine its influence magnitude on the overall estimates. To further assess the robustness of our results, we performed several sensitivity analyses by only including trials with Jadad score ≥ 3, using BMD or bone fracture as the primary outcome, and trials using intention-to-treat (ITT) analysis.

Funnel plots were used to inspect publication bias visually and the Egger test was used to assess the asymmetry of the funnel plot.[36] In addition, we used “trim-and-fill” method to examine the influence of publication bias on the overall findings.[37] A two-sided P value less than 0.05 was considered statistically significant and all the analyses were performed with Stata 12.0 (StataCrop LP, College Station, TX).

RESULTS

Of the retrieved 334 relevant citations, 7 trials of 27,900 randomized participants were included in the current meta-analysis (Figure 1). Characteristics of the 7 trials are shown in Table 1. The trials, published between 2001 and 2014, varied from 64 to 17,802 participants. Study durations ranged from 12 months to 6 years. The studies were conducted in the US, Denmark, Australia, and countries from East Asia, as well as multiple centers. Of the 7 trials, 5 were conducted to assess the association of statins use with BMD change[38-42] and 2 with fracture risk.[43,44] A total of 4 trials included participants with osteoporosis or osteopenia. Five trials had the primary outcome of BMD change and the other 2 trials assessed fracture risk as the secondary outcome. Four trials applied ITT analysis and 5 were categorized as high quality (Jadad score ≥ 3).[32] Participants in intervention groups received statins treatment with various dosages daily, including atorvastatin, simvastatin, and rosuvastatin; and participants in control groups received placebo, diet or lifestyle guidance, or nonstatin treatment.

FIGURE 1

Selection of eligible randomized controlled trials examining the association of statins use on bone health.

TABLE 1

Characteristics of 7 Trials Examining the Effect of Statin Use on Net Change in Bone Related Measures Among Adults

Baseline characteristics of participants in the intervention group and control group were shown Table 2. In both groups, the average age ranged from 58.6 to 80.8 years old with the proportion of males participants from 0% to 100%. Among intervention groups, average BMD ranged from 0.51 to 0.93, with those from 0.58 to 0.91 in their corresponding control groups.

TABLE 2

Baseline Information of Study Participants

Among 5 trials with the outcome of BMD change, 4 reported comparisons of absolute BMD change and 1 reported percentage of BMD change[38] at various time point. In the current analysis, only information at the end of the study was used (Table 3). In addition, most of the studies reported BMD change of lumbar spine, and Chuengsamarn et al[40] reported that of distal radius. Among the 4 trials reported absolute BMD change, net change ranged from −0.002 to 0.045 g/cm2 in intervention groups and from −0.02 to 0.006 g/cm2 in control groups. As shown in Table 4, 2 trials assessed the association of statins use and fracture risk.

TABLE 3

Average Change in BMD

TABLE 4

Overview of Multivariable-Adjusted Associations of Statin Use With Fracture

Pooled estimate of the net change of BMD is presented in Figure 2 and pooled HR of fracture risk is presented in Figure 3, respectively. On average, compared with the control group, statins use resulted in significant increases in BMD, with net BMD change of 0.030 g/cm2 (95% CI: 0.006, 0.053; I2 = 99.2%; P < 0.001= but null association with fracture risk, with the pooled HR of 1.00 (95% CI: 0.87, 1.15; I = 0; P = 0.396). In order to examine the robustness of our findings, we also conducted sensitivity analyses based on restricting BMD location and study population. For example, when we excluded the study that only included males,[39] the pooled net change of BMD was 0.040 (−0.006, 0.085) g/cm2 and when we further pooled the results of studies conducted only in females,[41,42] the net change was 0.030 (−0.027, 0.088) g/cm2. The results did not substantially differ from the overall findings. In addition, the influence analysis did not identify any trials’ removal would significantly alter the findings. Although the somewhat asymmetrical funnel plot was shown regarding to net BMD change estimates (Figure 4), the Begg test did not indicate significant publication bias (P = 0.174).

FIGURE 2

Average net change in bone mineral density in randomized controlled trials comparing statins use with control group. The size of each square is proportional to the percent weight that each study contributed in the pooled estimate. The pooled effect size is indicated by the diamond. Bars, 95% confidence interval (CI).

FIGURE 3

The association of statins use with bone fracture in randomized controlled trials comparing statin use with control group. The size of each square is proportional to the percent weight that each study contributed in the pooled estimate. The pooled effect size is indicated by the diamond. Bars, 95% confidence interval (CI).

FIGURE 4

Funnel plot of the meta-analysis on the association of statins use with bone mineral density.

DISCUSSION

The current meta-analysis pooled results from 7 RCTs with almost 30,000 participants. We have identified that statins use significantly increased BMD by approximately 0.030 g/cm2 and was not associated with higher fracture risk, with robust findings across sensitivity analyses. Our findings indicate that statins use could be a potential prevention or treatment for bone health.

Osteoporosis is responsible for 2 million broken bones and $19 billion in related costs every year.[45] It is estimated that osteoporosis will be responsible for approximately 3 million fractures and $25.3 billion in costs each year by 2025 (http://www.nof.org/article/7). As the most important predictor of osteoporotic fractures, the decrease in BMD significantly increased the fracture risk.[46] A previous systematic review suggested that statins use is effective for increasing bone turnover.[47] The current meta-analysis provides important information on quantitative benefits of statins use on BMD from the accumulation evidence of RCTs. Additional study strengths include the inclusion of only RCTs, thereby reducing the likelihood that the observed association statins use with BMD and fracture risk related traits could be explained entirely by bias and confounding. In addition, only 2 of 4 trials of BMD were individually statistically significant, highlighting the benefits of meta-analysis to identify important effect sizes with increased statistical power. In addition, sensitivity analysis did not substantially change the findings.

The “statin for osteoporosis” hypothesis has drawn great attention and many studies have revealed the mechanism on the protective effect of statins use on the prevention of osteoporosis.[48] A very complex and still incomplete picture showed that statins could increase osteogenesis or suppress osteoblast apoptosis.[49] In addition, other pathways, including reduction of oxidative stress and restoration of NO formation, and antiinflammatory effects of statins also contribute to the protection against osteoporosis. Although BMD significantly increased after statins use, the fracture risk was not reduced otherwise. The potential explanation includes that small changes in BMD might not translate to changes on bone surfaces, which is critical to protect against fracture.[50] Although it might be more cost-effective when treating dyslipidemia and osteoporosis together, the current study did not identify the significant association between statins use and osteoporosis.

Still, certain limitations should be addressed and some of these limitations provided hints for further investigations. First, the number of RCTs regarding to the association of statins use with BMD and fracture risk is very small, which has limited subgroup analysis and further limited to identify subgroup population who were more susceptible to statins therapy on both dyslipidemia and osteoporosis. Therefore, more research is needed to determine whether statins intervention can present its benefits among participants with various lipid levels or different disease status, etc. Second, most of the included studies were conducted in females, which have limited the generality of the results to male patients. The prevalence of osteoporosis was more prevalent in females, about 40% of females in developed countries will experience an osteoporosis-related fracture through their lifetime, while males experiencing approximately one-third to one-half the risk of females.[51,52] In despite of this, the effect of statins use on bone health in males should not be ignored. Third, although we searched for “gray literature,” none of them was in accordance with our inclusion criteria. Therefore, there was some indication of possible publication bias for the BMD trait. In addition, most of trials in this meta-analysis did not use BMD as its primary outcome, which highlighted the need for relevant RCTs.

In conclusion, this meta-analysis provides evidence that statins are an effective strategy for bone health. Although these findings are encouraging, further trials to better understand the effect of statins use on BMD in certain subgroups are warranted. Research will also be needed to assess the cost-effectiveness of statins use on bone health. In aggregate, results of the current meta-analysis suggested that statins use could contribute to meaningful increments in BMD at the population level.