Osteoporosis Update: Screening and Treatment Recommendations

What we know

Osteoporosis is common and associated with significant morbidity, mortality, and cost. Over the last 20 years, there has been an increase in available pharmacologic management options, making treatment decisions more complex. Multiple medical societies have periodically published recommendations to provide a review of the evolving evidence and guidance to the practicing clinician. ^1,2 Differing guidelines may have key differences based on the prioritization of different study designs (prospective vs. observational), opinions of the experts on the committee (primary care vs. specialty), and emphasizing different values (cost, efficacy). The American College of Physicians (ACP), representing primary care physicians, regularly publishes rigorous systematic reviews and evidence-based guideline recommendations for common clinical conditions. Since publication of their 2017 osteoporosis recommendations, multiple studies have been published on new treatments and their comparative efficacy.

Study results

ACP commissioned an independent systematic review and network meta-analysis of the evidence from which the recommendations are based. Instead of focusing on specific risk assessment tools, the committee considered fracture risk based on the diagnosis of osteoporosis, history of fractures, treatment failure or intolerance, and for those who had multiple risk factors for fractures. The guideline committee prioritized (1) longer-term benefits versus harms that last at least 3 years and (2) prevention of hip and clinical vertebral fractures (based on their greater clinical impact). Recommendations for menopausal hormone therapy (HT), treatment duration, or monitoring of bone mineral density (BMD) were not included.

The guideline recommends bisphosphonates for initial pharmacologic treatment in those with osteoporosis, with the RANK ligand inhibitor (denosumab) as a second-line treatment. The preferred bisphosphonates include alendronate, risedronate, or zoledronate; ibandronate was not recommended given the lack of evidence for hip fracture reduction. In those at very high risk for fracture, a sclerostin inhibitor (romosozumab), or recombinant parathyroid hormone (teriparatide) is suggested, followed by a bisphosphonate. Very high risk was based on older age, recent fracture (i.e., within the last year), history of multiple fractures, or multiple risk factors. The committee felt that evidence was too limited for abaloparatide or selective estrogen receptor modulators (SERMs; bazedoxifene, raloxifene) to make definitive recommendations. For those who have low bone mass (osteopenia as opposed to osteoporosis), the guideline recommends an individualized approach to treatment in women over age 65, given more limited data in this group. The recommendations include a helpful table of the average Medicare spending for each medication. Because of the number of ongoing studies, ACP plans to create a “living guideline” moving forward, which will involve a quarterly review of evidence and periodic updating of the review and recommendations.

What this changes

Perceived differences within guidelines can be a source of confusion for clinicians and patients. Understanding the perspectives, strengths, and limitations of each recommendation helps to provide comprehensive, individualized care. This guideline focuses on management of established osteoporosis (typically in those 65 and older), and longer-term outcomes. Concerns raised about this recommendation include not considering the high risk for vertebral fractures after denosumab discontinuation, use of HT, and the implications of decreasing cost of anabolic agents in the future. ³ In women between the ages of 40 and 60, a randomized trial with 4 years follow-up showed that HT improvements in BMD were similar to or better than alendronate (depending on the HT formulation). ⁴ Recently updated recommendations from the Menopause Society (formerly known as the North American Menopause Society) ⁵ and Endocrine Society ⁶ provide updated evidence and support for the use of HT (typically in individuals age 60 years or younger) or SERMs for bone health in those who are appropriate candidates. ¹

What we know

Because most osteoporosis treatment trials are placebo-controlled, there are few head-to-head studies to guide clinical care. The last Cochrane-type reviews were published in 2008, before the introduction of new treatments. ^7,8 There is substantial variation between randomized controlled trials (RCTs) with respect to studied patient populations, fracture diagnoses, and reported risk reductions, making comparisons of treatment outcomes difficult. An updated meta-analyses to compare relative effectiveness of treatment options was needed.

Study results

The authors conducted a systematic review, network meta-analysis, and meta-regression analyses to understand the benefits of available treatments for fracture risk reduction, and to stratify benefits based on risk of fracture. The analyses (which are separate from the ACP systematic review noted above) were conducted consistent with the methodology recommended by the Cochrane Collaboration as well as the Preferred Reporting Items for Systematic review and Meta-analysis (PRISMA) for Network Meta-Analysis. ⁹ All RCTs of postmenopausal women between 1996 and 2021 (without age restrictions) were included, except for RCTs from Asian countries given their treatment doses are different than in other regions of the world.

The primary outcome was all clinical fractures (excluding fingers and toes), and secondary outcomes included vertebral fractures, hip fractures, and serious adverse events (including all-cause mortality and cardiovascular safety). Interventions included bisphosphonates, denosumab, SERMs (raloxifene, bazedoxifene, and bazedoxifene with conjugated estrogen), parathyroid hormone receptor agonists, and romosozumab. Of the 6,244 initial entries identified, the results of 69 RCTs met criteria for analysis, representing >80,000 patients.

For the outcome of clinical fractures, all treatments showed protective effects except for denosumab and SERMs. Of note, the authors did not include a pivotal study for denosumab, the Fracture Reduction Evaluation of Denosumab in Osteoporosis Every 6 Months (FREEDOM) study, because of lack of aggregated data for clinical fracture in these publications. Bisphosphonates were less effective in reducing clinical fractures compared to parathyroid hormone receptor agonists, with odds ratios (OR) 1.49, 95% confidence interval (CI) 1.12–2.00.

All treatments were effective in vertebral fracture reduction compared to placebo, with denosumab, romosozumab, and parathyroid hormone receptor agonists being more effective compared to oral bisphosphonates. Overall, baseline fracture risk did not impact treatment effects except in the case of antiresorptive agents, where greater clinical fracture risk reduction was seen with increasing age. For hip fracture, all treatments showed a protective effect except for SERMs; romosozumab was more effective than oral bisphosphonates during active treatment. There was no significant increase in mortality, cardiovascular, or other adverse safety events. Data for nonvertebral fractures could not be analyzed. It is important to note that because of limitations of the reporting, the certainty in the effect estimates was moderate to low for all outcomes, which may introduce risk of bias.

What this changes

Studies show that BMD is improved by most osteoporosis treatments, but fracture prevention is the most relevant clinical outcome of interest. This updated meta-analysis shows that all available postmenopausal osteoporosis medications will help reduce vertebral fracture risk, but there may not be consistent fracture risk reduction of all clinical fractures or at the hip. Head-to-head RCTs suggest that anabolic agents are probably more effective in preventing clinical and vertebral fractures than antiresorptive medications.

Though more head-to-head trials would be helpful, the number of individuals with fractures are lower in these types of studies (compared to placebo-controlled -studies), leading to low statistical power and greater financial and operational hurdles. Thus, more consistent reporting on baseline fracture risks and fracture diagnoses would be helpful to compare future studies. Another important outcome from the study is a reminder that antiresorptive treatments such as bisphosphonates, denosumab, and SERMs seemed to prevent fracture better with increasing age. This is an important reminder that patients should not be considered “too old” to consider treatment.

What we know

The US Preventive Services Task Force (USPSTF) guidelines endorse screening for osteoporosis in all women over the age of 65, but recommend the use of risk assessment tools to determine who should be screened below this age cutoff. ¹⁰ The guidelines recommend a variety of risk assessment tools, including the commonly used Fracture Risk Assessment Tool (FRAX) and Osteoporosis Self-Assessment Tool (OST). FRAX has four options for race and ethnicity (Caucasian, Black, Hispanic, Asian), but it's ability to discern differential fracture risk based on these categories within the US population is not clear. Additionally, the overall utility of these tools to screen for osteoporosis needs to be better clarified.

Study results

From the 161,888 participants in the Women's Health Initiative (WHI) observational study, the authors identified a cohort of 67,169 women who were treatment naïve, had complete data regarding risk variables, and provided 10 years of follow-up. The primary goal for this study was to assess the ability of FRAX (without BMD information) to discriminate fracture risk based on racial/ethnic grouping, as compared to OST. Questionnaires were utilized to assess incident fractures throughout the follow-up duration, as well as to ascertain relevant risk factors (age, race/ethnicity, smoking, alcohol intake, parental hip fracture, fall frequency, medical comorbidities, prior history of fracture after age 55, early menopause, medication use, as well as physical function). Absolute 10-year risks for major osteoporosis fracture were calculated using FRAX and OST, the latter that does not include race/ethnicity variables. Among the 67,169 participants, 5,594 women experienced a major osteoporosis fracture. Both FRAX and OST performed poorly in women who did versus did not suffer a fracture among all specified race/ethnicity categories.

A secondary goal of the study was to assess overall ability of FRAX and OST to identify those between the ages of 50 and 64 who are likely to have a BMD T-score within the osteoporosis range. Of this original cohort, 4,607 who had complete data with BMD screening were included for the secondary analysis. In contrast to the FRAX tool, the OST had excellent ability to identify femoral neck osteoporosis in each of the four racial and ethnic groups.

What this changes

The conclusion of the authors is that the US version of FRAX does not reliably predict major osteoporosis fracture risk based on race/ethnicity, or the chances of a femoral neck BMD reading within the osteoporosis diagnostic range. If race/ethnicity information is to be included in risk calculations (a growing area of controversy), more validated tools may be needed. In contrast to FRAX, OST was considered an excellent tool for identifying young postmenopausal women who may be potential candidates for osteoporosis drug therapy, without the need to incorporate race/ethnicity information.

What we know

Fragility fractures related to osteoporosis are an important indicator to initiate medical management, however, osteoporosis remains undertreated, even in the setting of fracture history. Because of their mechanism of action, there are theoretical concerns that early initiation of bisphosphonates may contribute to nonunion in the setting of acute fracture, especially after surgery. Although SERMs and HT have shown important physiologic benefits in bone and fracture healing, ¹¹ clinicians may be concerned about early initiation due to minimal fracture safety data. This study aimed to understand whether treatment with bisphosphonates or SERM/HT was associated with fracture nonunion risks, and whether timing of treatment had any impact on this risk.

Study results

This was a retrospective analysis of Medicare claims using Current Procedural Terminology (CPT) and International Classification of Diseases (ICD-10) codes to identify individuals who had a surgically treated long bone fracture between the years of 2016 and 2019. The analysis was limited to those who were 65 years of age or older; the database was subsequently searched for any claims made regarding fracture union status within the following year. Multivariable logistic regression models were conducted to assess association between medication treatment (broken into two categories of bisphosphonates or SERM/HT) and fracture union status, while also controlling for medical comorbidities, fracture type, age, sex, and race.

There were 111,343 long bone fractures identified, of which 10,452 (9.4%) were associated with a diagnosis of nonunion within the next year. Bisphosphonate use was higher in the nonunion group (12.2% nonunion vs. 11.4% union), while there was no difference in the SERM/HT users (1.1% vs. 1.2%). However, after controlling for race, age, sex, and comorbidity index, neither bisphosphonate nor SERM/HT use was associated with nonunion. When stratified based on timing of treatment, even those who initiated treatment within 90 days postfracture did not have an increased risk of nonunion (OR, 0.94; 95% CI 0.86–1.03; p = 0.175).

What this changes

The American Society for Bone and Mineral Research urges the initiation of treatment after acute fracture to minimize future fracture risk; they highlight clinician hesitancy and misinformation as key factors leading to undertreatment. ¹² This study provides reassurance that early treatment, including treatment within 3 months, does not significantly increase delayed healing complications after surgery. The authors acknowledge a potential theoretical possibility of increased nonunion risks, but highlight that clinicians should not be concerned about this risk when there is a 15 times greater odds of benefiting the patient (in terms of future risk reduction) than harm. If HT or SERM treatments are initiated, it is important to always consider thrombosis risks based on a careful clinical reassessment.