Development of the Asia Pacific Consortium on Osteoporosis (APCO) Framework: clinical standards of care for the screening, diagnosis, and management of osteoporosis in the Asia-Pacific region

APCO members were invited to determine which aspects of care required clinical standards to be developed, based on a list informed by the findings of the 5IQ comparative analysis. The questionnaire included 32 questions under three domains (Appendix 2):

Members were asked to express their agreement (or not) with the wording of 16 draft clinical standards, of which several included proposed wording on levels of attainment (5-point scale from “strongly agree” to “strongly disagree”). For each draft standard, members were also invited to propose alternative wording or make comments. Consensus was defined as a ranking of “strongly agree” or “agree” by at least 75% of respondents.

The wording of clinical standards and levels of attainment was amended, based on the results of Round 2. APCO members were invited to approve amendments (yes/no options). Consensus was defined as a “yes” response to a proposed rewording by at least 75% of respondents.

A fourth round was conducted after some of the standards and levels of attainment were again reworded for clarity and precision without changing their intent. The minor amendments to the wording of the standards, with commentary itemizing each change, were sent to the APCO members who had participated in one or more of the previous Delphi rounds. Members were instructed to reply to the Chairperson by a stipulated date if they had any objection to the rewording of the standards.

Fifteen of the 18 guidelines cited risk factors separately for men and women, two guidelines for women only, and another did not state the sex to which cited risk factors applied.

The most commonly cited risk factors for osteoporosis (cited in at least half the guidelines) were excessive alcohol consumption (17 guidelines), a family history of osteoporosis and/or fractures (17 guidelines), smoking (17 guidelines), low body mass index (BMI)/weight (16 guidelines), height loss (13 guidelines), age over 70 years (13 guidelines), and early menopause (12 guidelines).

A range of other age thresholds for increased risk was also identified, with several citing one or more of the following: adults (four guidelines), postmenopausal women (three guidelines), 50 years and above (two guidelines), 60 years and above (one guideline), and 65 years and above in men (one guideline).

All 18 guidelines cited a history of fragility fracture as a risk factor for subsequent fracture. Thirteen of 18 guidelines specified the type of fragility fractures, most commonly spine, hip, proximal femur, wrist, forearm, distal radius, humerus, pelvis, and ribs. In the context of case identification, one guideline from China mentioned the importance of FLSs to promote multidisciplinary joint diagnosis and treatment.

While all the guidelines identified glucocorticoid treatment as being associated with bone loss and/or increased fracture risk, the lists of other medicines associated with bone loss varied between guidelines. More than half of the guidelines also identified users of androgen deprivation therapy (11/18 guidelines), anticoagulants (9/18 guidelines), anticonvulsants (10/18 guidelines), and aromatase inhibitors (11/18 guidelines) as at-risk groups. Other medications that were identified included proton pump inhibitors, thiazolidinediones, and selective serotonin reuptake inhibitors.

The medical conditions most frequently cited as risk factors were rheumatoid arthritis, malabsorption, hyperthyroidism, and multiple myeloma. Diabetes was noted as a risk factor in 10 of 14 (71%) of the newer guidelines, compared with none of the older guidelines.

The extent of commentary on assays for bone turnover markers and other biochemical parameters varied considerably between the guidelines. Some guidelines only briefly mentioned biochemical tests in the evaluation of osteoporosis, while others provided detailed guidance. Total procollagen type 1 N-terminal propeptide (P1NP) was the most frequently recommended bone turnover marker (16/18 guidelines). Other laboratory tests that were commonly recommended in the guidelines included calcium, alkaline phosphatase, 25-hydroxyvitamin D, creatinine, urinary N-telopeptide (uNTX), parathyroid hormone (PTH), and osteocalcin.

The FRAX® fracture risk assessment tool [30] was by far the most commonly recommended risk assessment tool (17/18 guidelines), followed by the Osteoporosis Self-Assessment Tool for Asians (OSTA) (11/18 guidelines) [31]. Some newer guidelines also recommended a range of other tools: the Garvan nomogram [32] (three guidelines), IOF Risk Check (“one-minute osteoporosis risk test”)[33] (three guidelines), the Khon Kaen Osteoporosis Study (KKOS) scoring system [34] (one guideline), the Male Osteoporosis Risk Estimation Score (MORES) [35] (1 guideline), QFracture algorithm [36, 37] (one guideline), and Simple Calculated Osteoporosis Risk Estimation (SCORE) [38] (one guideline).

All 18 guidelines mentioned the use of radiography to identify vertebral fractures, with some directly recommending x-ray of the thoracic and lumbar spine, some advising when it should be considered, and one simply noting that bone loss and healing fractures may be evident on radiographs. Dual-energy x-ray absorptiometry (DXA)–based vertebral fracture assessment (VFA) was mentioned in eight guidelines, with guidance ranging from just statements of its existence or simple advice to consider DXA for identifying fractures, to the inclusion of DXA-identified fractures in treatment indications.

Thirteen of 18 guidelines referred to the importance of falls risk assessment. The degree of detail varied widely with some amount of detail specified in the guidelines from Australia, China, Hong Kong SAR, Chinese Taipei, India, Malaysia, Japan, Thailand, and New Zealand. Some guidelines emphasized that falls risk assessment should be conducted routinely in all older adults (Australia), while others discussed falls risk assessment for patients known to be at risk of fragility fractures or with a history of fracture. One of the Chinese guidelines recommended that attention should be paid to the evaluation of fall-related risk factors in elderly patients with osteoporosis. Some guidelines recommended certain aspects of assessment, such as muscle strength and balance, while the Hong Kong SAR and one of the other Chinese guidelines provided comprehensive guidance on specific medical and environmental risk factors to consider when performing the assessment. Some mentioned strategies or structured tools for performing falls risk assessment (Thailand).

Ten of 18 guidelines recommended specialist referral for assessment in certain clinical scenarios. Four recommended that any patient with fragility fractures or osteoporosis should be referred for specialist assessment or care. Other specified scenarios included fracture or persistent loss of BMD despite treatment, complex conditions, problems considered “difficult” or beyond the scope of primary care, secondary osteoporosis, severe or unusual presentation, when first-line treatments were contraindicated or not tolerated, when assessment for specialized treatments is needed (e.g., teriparatide or long-term estrogen therapy), and when bone densitometry is not available.

All 18 guidelines recommended that information should be provided on calcium intake, and 17 recommended providing information on exercise.

Ten out of 18 guidelines (including one older guideline) recommended that patients should be provided with information on sun exposure to maintain healthy vitamin D levels.

Only seven out of 18 guidelines (all newer guidelines) recommended that patients should be given information about the link between osteoporosis and fracture risk.

All 18 guidelines cited BMD T-score of ≤ − 2.5 standard deviations (SD) as an indication for treatment. The reference database to be used to delineate T-scores was specified in only a minority of the guidelines. A broad range of other indications for treatment were cited in the various guidelines, with each guideline typically featuring three or four of a total of 14 indications identified.

Most guidelines cited hip and vertebral fractures as an indication for treatment. Fifteen of the 17 guidelines that included hip fracture as an indication for treatment, and all 17 guidelines that included vertebral fracture as an indication for treatment, stated that BMD testing was not required to initiate treatment. Of the 15 guidelines that included non-hip, non-vertebral fracture as an indication for treatment initiation, nine suggested that BMD testing was not necessary prior.

Thirteen guidelines mentioned or recommended treatment thresholds based on FRAX® probabilities. Six guidelines cited the threshold recommended by the US National Osteoporosis Foundation guidelines [39] (BMD T-Score in the osteopenic range, in combination with a FRAX® 10-year probability of ≥ 3% for hip fracture risk or ≥ 20% for major osteoporotic fracture risk). Of these, only two (the Malaysian guideline and one of the guidelines from China) explicitly recommended that this threshold should be adopted for their local target population. The Hong Kong SAR guideline stated that the US criteria were provided for clinical guidance only and that treatment decisions should consider individual patient factors. Other guidelines cited the US treatment threshold as an example or in commentary, without further interpretation.

Some guidelines provided guidance on the logistics and limitations of applying FRAX® to local populations. One of the guidelines from China [15] stated that the current prediction results may underestimate fracture risk in the Chinese population. Several others recommended the use of population specific FRAX® algorithms (e.g., Hong Kong SAR, India, Singapore, Thailand, Vietnam).

Osteopenia in combination with other criteria (e.g., presence of risk factors, risk scores, ≥ 10 years post menopause) was included in several guidelines.

Less than one-third of guidelines classified treatment options as first line, second line, and third line. All guidelines recommended bisphosphonates (oral and intravenous) and raloxifene as treatment options for osteoporosis. Most guidelines also recommended denosumab, teriparatide, and hormone replacement therapy (HRT). Predictably, the recently introduced treatments, abaloparatide and romosozumab, were recommended in only two newer guidelines (India and Korea).

Calcium and vitamin D were recommended as adjunctive therapies more often than as active treatments. That calcium was not a treatment option was explicitly stated in less than half of the guidelines.

The volume of commentary on potential adverse effects of osteoporosis treatments varied considerably between guidelines. Eleven of 18 guidelines provided information on adverse effects associated with all or several classes of anti-osteoporosis agents.

Almost all the guidelines provided information on the role of repeated measurements of BMD and biochemical markers of bone turnover in monitoring. Some specified intervals for repeating tests (Table 3). Overall, there was limited guidance on making long-term management plans and providing them to the patient or primary care provider. Only two guidelines explicitly stated the need for a long-term plan, while 8 others implied the need for such planning.

Duration of treatment was specified in 13 guidelines and discussed very generally in two others. Most guidelines advised using bisphosphonates for 3–5 years before reassessing to determine whether to continue or discontinue with observation (“drug holiday”), based on assessment of treatment response and fracture risk. Another guideline recommended continuing bisphosphonate therapy for 5–10 years before reconsideration. Several guidelines gave detailed guidance based on risk assessment. Three guidelines recommended that teriparatide treatment should not exceed 2 years, while one guideline recommended 3 years of treatment with teriparatide. Of the four guidelines that discussed HRT and mentioned initiation of, and duration of therapy with it, one (Australia) recommended that its long-term use is not recommended. The Hong Kong SAR guideline recommended that patients should be adequately counselled on the risks and benefits of long-term use (beyond 3-5 years for combined estrogen-progesterone therapy and beyond 7 years for Estrogen alone therapy). The Indian Menopause Society guideline recommended it as first line therapy for women with menopausal symptoms for up to 10 years after menopause and specifically stated that HRT should not be started solely for bone protection after 10 years of menopause. The guidance from Singapore specified that menopausal hormone therapy can be considered for prevention of osteoporosis and fragility fractures in women who experience early menopause and that it can be continued until the normal age of menopause unless contraindicated. Of the two guidelines that mentioned duration of denosumab therapy, one recommended long-term use and the other recommended reevaluation of fracture risk after 5–10 years.

Nine guidelines made specific recommendations on assessing and encouraging adherence.

Five guidelines made specific recommendations for referral to falls prevention programs. A further three guidelines provided detailed commentary on falls risk assessment.

Consensus was reached on the inclusion of clinical standards on the following components (i.e., ≥ 75% rated its inclusion as important or extremely important (Table 4)):

Consensus was reached (i.e., ≥ 75% of respondents strongly agreed or agreed) on the wording of all 16 standards (Fig. 1). Agreement was unanimous or nearly so (> 90% agreement) for the wording of draft standards on:

Diversity of opinion was greatest for the wording of the draft standard on proactive case identification based on common risk factors, although consensus was reached. The volume of comments (excluding minor editing suggestions) was highest for the standards on, medicines associated with risk, assessment of vertebral fractures, intervention thresholds for osteoporosis-specific therapies (Table 5), and assessment of adherence to pharmacological therapies.

Consensus was achieved on all the proposed amendments, involving six standards, and levels of attainment for two standards. At the end of this round, consensus had been reached on the final Framework of minimum clinical standards.

There were no objections to the minor amendments proposed to improve clarity and precision of the wording of some of the standards.

Up to 50% of patients with a hip fracture have already sustained fractures at other skeletal sites during the previous months or years [47, 48]. A prior fracture at any site is associated with a doubling of future fracture [49, 50] and mortality [51] risks. Accordingly, these “signal” or “sentinel” fractures should alert healthcare providers to the opportunity for secondary fracture prevention.

Over the past decade and a half, coordinated post-fracture models of care have been designed to ensure that healthcare providers consistently respond to the first fracture to prevent the second and subsequent fractures [48]. These are generally hospital-based, but involvement of primary care physicians is also important to ensure continuity of care [52, 53]. Components of effective FLSs include multidisciplinary involvement, dedicated case managers and clinician champions, regular assessment and follow-up, multifaceted interventions, and patient education [54].

Fracture liaison service programs significantly increase rates of BMD testing, initiation of osteoporosis treatment, adherence to treatment, and reduce refracture incidence and mortality rates, compared with usual care [54, 55]. They are cost-effective in comparison with usual care or no treatment, regardless of the program intensity or the country [55]. The IOF strongly promotes the concept of FLSs through its global Capture the Fracture® initiative [40]. This initiative has set an international benchmark for FLSs by defining 13 globally endorsed standards for service delivery.

The number of FLSs in the Asia-Pacific region is rapidly increasing. Analyses of indigenous fracture liaison programs have been conducted in several countries and regions in the Asia-Pacific region including Australia [56], China [57, 58], Chinese Taipei [59], Hong Kong SAR [60], Singapore [61], Japan [62‐64], New Zealand [65], South Korea [66‐69], and Thailand [70]. In 2017, the fracture liaison service consensus meeting held in Taipei concluded that the 13 Best Practice Framework Standards of the IOF’s Capture the Fracture® campaign [40] were generally applicable in the Asia-Pacific region (albeit with minor modifications). As of 17 September 2020, 111 FLSs from the Asia-Pacific region are included on the IOF Global Map of Best Practice (https://​www.​capturethefractu​re.​org/​map-of-best-practice). Eighty-four have been fully evaluated, of which 19 have been awarded a Gold star. These include FLSs from Australia, Japan, New Zealand, Singapore, Taiwan, and Thailand.

Despite this burgeoning interest, the 5IQ analysis revealed that only four guidelines in the Asia-Pacific region advocated the role of FLSs, suggesting that the crucial role of secondary fracture prevention is still sadly under-recognized among guideline developers and policy makers in the Asia-Pacific [52]. Two recent initiatives are, however, expected to enable more widespread adoption of FLSs in the region: the FLS tool box for Asia-Pacific developed by the Asia-Pacific Bone Academy Fracture Liaison Service Focus Group [71], and the Capture the Fracture® Partnership launched by the IOF (https://​www.​capturethefractu​re.​org/​capture-fracture-partnership), which includes several countries in the Asia Pacific region.

Upon the knowledge that “fractures beget fractures” [47, 49, 50] has subsequently come the understanding that the risk of a subsequent fragility fracture is particularly acute immediately after the sentinel or the index fracture and that this risk wanes progressively with time [72‐78]. This very high fracture risk in the 1 to 2 years after the index fracture has been termed “imminent risk” [79, 80] and appears to be age-dependent, with the transient effect of increased risk being more evident at older ages [76]. The ratio of 10-year fracture probability after a recent fragility fracture to that after a fracture, irrespective of its site or recency, has been termed the probability ratio or adjustment ratio. Probability ratios that provide adjustments to conventional FRAX® estimates for recency of sentinel fractures have now been derived, though how they will be included in the FRAX® algorithm is still being debated [81].

Other factors can also contribute to, or result in, a very high fracture risk. The simultaneous presence of multiple risk factors (e.g., family history of fracture, glucocorticoid use) can additively contribute to fracture probabilities and shift fracture risk categories to higher strata of risk [30]. Individuals have also been somewhat arbitrarily designated as very high risk if they have fractures while on approved osteoporosis therapy, have multiple prevalent fractures or fractures while on drugs causing skeletal harm such as glucocorticoids, have very low T-score (e.g., < − 3.0), have high risk for falls or a history of falls, or have a FRAX® probability for major osteoporosis fracture > 30%, or hip fracture > 4.5% over a 10-year period [82].

This concept of stratification of osteoporotic fracture risk that may enable or guide the choice of therapeutic agent is gaining traction globally. Those at higher risk may require initiation of treatment with more potent therapies such as an anabolic agent. For those at low risk, a decision will need to be made on whether therapy with an anti-osteoporosis agent is even indicated. Updated position papers and guidelines published by the European Society for Clinical and Economic Aspects of Osteoporosis (ESCEO) and the International Osteoporosis Foundation (IOF) [80] and by the American Association of Clinical Endocrinologists (AACE) [82] provide recommendations for this stratification of risk. The ESCEO/IOF position paper applies age-specific thresholds to define high versus very high risk, while the AACE guidelines suggest an arbitrarily decided fixed threshold as well as the presence of certain pre-defined risk factors as described previously [80, 82].

At a regional and national level, these recommendations will have to be reviewed to determine the relevance to local practices and to see whether they offer tangible benefits to patients. Some challenges to implementation of these guidelines in the Asia-Pacific would need to be considered. The FRAX® thresholds that are recommended in the AACE guidelines are likely only applicable to the USA. While FRAX® has been validated in many Asian countries, most of them have not implemented specific intervention thresholds for use yet. Recommendations also need to consider local reimbursement criteria, which often dictate eligibility criteria for pharmacological treatment options.

Osteoporosis is a chronic condition. Therefore, patients with osteoporosis need a long-term, personalized management plan, with many patients requiring multiple anti-osteoporosis medications during their lifetime. The specific and individualized treatment plan traditionally has involved initiation of therapy with an antiresorptive agent. Most of the extant guidelines in the Asia-Pacific region also favor this approach. However, there is evidence for a stronger anti-fracture efficacy of bone-forming treatments compared to antiresorptive ones from two clinical trials in patients with severe osteoporosis [83, 84] and for a blunted/delayed BMD response to bone-forming treatment, with prior antiresorptive use, though this effect somewhat depends on the potency of the antiresorptive agent [85‐87].

In certain scenarios, a deleterious effect on BMD is seen in patients previously treated with antiresorptive agents who are then switched to an anabolic agent [88]. In contrast, sequential therapy with an anabolic agent followed by an antiresorptive agent appears largely to be associated with maintenance or continued increase in BMD [88] and continued reduction in fractures [84, 89].

This has resulted in increasing interest in sequential treatment with an anabolic agent given as first-line treatment in the very-high-risk fracture patient, whose fracture risk needs to be addressed quickly and in whom BMD needs to be improved quickly and as much as possible. However, the fact that this may not always be possible, due to restrictions in reimbursement criteria and limitations imposed by guidelines, should be considered before advocating large-scale adoption of such recommendations.

Health economic analysis is playing an increasingly important role to inform the relative value of osteoporosis therapies and to help determine how best to allocate finite health care resources. The development of absolute fracture risk–based assessment and intervention thresholds has led also to the exploration of cost-effectiveness of interventions along the lines of fracture probability. A significant body of evidence now exists to show that anti-osteoporosis therapies are cost-effective in women at high risk of fracture [90].

However, the cost-effectiveness of various pharmacological therapies is determined in part by the costs and benefits of treatment. Intervention thresholds will also vary, since they depend critically on country- and region-specific factors such as reimbursement issues, health economic assessment, willingness to pay for healthcare, and access to DXA scanning. Due to this vast heterogeneity in epidemiologic and economic characteristics between countries, such thresholds should be country-specific. In the USA, a 10-year absolute hip fracture probability of 3% and its equivalent major osteoporotic fracture probability of 20% have been considered as cost-effective intervention thresholds [39]. Various other major osteoporotic fracture risk fixed intervention thresholds ranging from 7 to 15% have been deemed to be cost-effective in other countries [91, 92]. In the UK, age-dependent FRAX®-based intervention thresholds have been shown to provide clinically appropriate access to treatment as well as to be cost-effective [93].

Given the evidence supporting initial treatment of patients at very high risk of fragility fracture with an anabolic agent followed by consolidating the effect of these agents with antiresorptive agents, health economic appraisals exploring the cost-effectiveness of such sequential therapies are also slowly emerging [94].

Health economic evaluation studies in osteoporosis are few in the Asia-Pacific region. Robust studies have been conducted in a few countries including Singapore [95], China [96, 97], Japan [98, 99], and Australia [100]. These studies have employed different modeling strategies and have explored cost-effectiveness of fracture intervention thresholds as well as of various osteoporosis medications. The varying results obtained from these studies highlight the crucial differences in economic, epidemiological, and clinical practice factors between the countries in the vast region that is the Asia-Pacific.