Osteoporosis International (2021) 32:1961–1971 https://doi.org/10.1007/s00198-021-05890-x ORIGINAL ARTICLE The treatment gap after major osteoporotic fractures in Denmark 2005-2014: a combined analysis including both prescription-based and hospital-administered anti-osteoporosis medications M. K. Skjødt 1,2 & M. T. Ernst 2,3 & S. Khalid 4 & C. Libanati 5 & C. Cooper 4,6 & A. Delmestri 4 & K. H. Rubin 2 & M. K. Javaid 4 & D. Martinez-Laguna 7,8,9 & E. Toth 5 & D. Prieto-Alhambra 4 & B. Abrahamsen 1,2,4 Received: 12 October 2020 / Accepted: 11 February 2021 / Published online: 15 March 2021 # The Author(s) 2021 Abstract Summary This study demonstrates a substantial and persistent anti-osteoporosis treatment gap in men and women ≥50 years old who sustained major osteoporotic fracture(s) between 2005 and 2014 in Denmark. This was not substantially reduced by including hospital-administered anti-osteoporosis treatments. Strengthened post-fracture organization of care and secondary fracture prevention is highly needed. Introduction The purpose of this study was to evaluate the Danish anti-osteoporosis treatment gap from 2005 to 2014 in patients sustaining a major osteoporotic fracture (MOF), and to assess the impact of including hospital-administered anti-osteoporosis medications (AOM) on the treatment gap among these patients. Methods In this retrospective, registry-based study, we included men and women aged 50 years or older and living in Denmark, who sustained at least one MOF between 2005 and 2014. We applied a repeated cross-sectional design to generate cohorts of patients sustaining a first MOF, hip, vertebral, humerus, or forearm fracture, respectively, within each calendar year. We evaluated the treatment gap as the proportion of patients within each cohort not receiving treatment with AOM within 1 year of the fracture. Hospital-administered AOM was identified by SKS code. Results The treatment gap among MOF patients decreased from 85% in 2005 to 79% in 2014. The gap was smaller among hip and vertebral fracture patients as compared to humerus and forearm fracture patients, and it was smaller in women than in men. The use of hospital-administered AOM was relatively uncommon, with a maximum of 0.9% of MOF patients initiating hospital-administered AOM (in 2012). We observed substantial variations in this proportion between fracture types and gender. Hospital-administered AOM was most commonly used among vertebral fracture patients. * M. K. Skjødt 5 UCB Pharma, Anderlecht, Belgium miksk@regionsjaelland.dk 6 MRC Lifecourse Epidemiology Unit, University of Southampton, 1 Department of Medicine, Hospital of Holbaek, Smedelundsgade 60, Southampton General Hospital, Southampton, England, UK DK-4300 Holbaek, Denmark 7 2 OPEN, Open Patient data Explorative Network, Department of GREMPAL Research Group, Fundació Institut Universitari per a la Clinical Research, University of Southern Denmark and Odense recerca a l’Atenció Primària de Salut Jordi Gol i Gurina University Hospital, Odense, Denmark (IDIAPJGol), CIBERFES, Barcelona, Spain 3 Department of Public Health, Clinical Pharmacology, Pharmacy and 8 Environmental Medicine, University of Southern Denmark, Universitat Autònoma de Barcelona, Bellaterra (Cerdanyola del Odense, Denmark Vallès), Spain 4 NDORMS, Nuffield Department of Orthopaedics, Rheumatology 9 and Musculoskeletal Sciences, Oxford University Hospitals, Atenció Primària Barcelona Ciutat, Institut Català de la Salut, Oxford, England, UK Barcelona, Spain 1962 Osteoporos Int (2021) 32:1961–1971 Conclusion A significant treatment gap among patients sustaining a major osteoporotic fracture was present throughout our analysis, and including hospital-administered AOM did not significantly improve the treatment gap assessment. Improved secondary fracture prevention is urgently needed. Keywords Osteoporosis . Fracture . Fracture prevention . Osteoporosis treatment . Bisphosphonate Introduction systems. It is therefore unclear if substantial reductions in the treatment gap may in fact already have been achieved Osteoporosis is a highly prevalent disease, particularly among through increased use of hospital-administered AOM. the older population, and is characterized by low bone mineral In this paper we therefore aim to evaluate the Danish anti- density (BMD), impaired bone microarchitecture, and an in- osteoporosis treatment gap from 2005 to 2014 in annualized creased risk of fractures [1, 2]. In Denmark, the incidence rates cohorts of patients sustaining a MOF, and to assess the impact of fractures in men and women aged 50 years or older have of including hospital-administered AOM on the treatment gap been found to be 170 and 320 per 10,000 patient-years for all among these patients. fractures, and 78 and 200 per 10,000 patient-years for major osteoporotic fractures (MOF, a composite of hip, vertebral, humerus, and forearm fractures), respectively [3]. Methods Patients with a previous osteoporotic fracture face an in- creased risk of subsequent fractures [4], most excessively el- This is a post hoc analysis of Danish data from the evated in the first years after the index fracture [5, 6]. Non- Multinational Observational Database Study on Imminent surprisingly, it has been reported that up to 45% of all hip Osteoporotic Fracture Risk (the IFRISK study), which is a fracture patients have had a prior fracture, and a study in register-based study evaluating short-term fracture risk in pa- Danish hip fracture patients demonstrated that 28% had tients at high risk of osteoporotic fractures [30]. In Denmark, sustained a MOF within the past 10 years [7–9]. The potential the study was approved by the Danish Medicines Agency, the of anti-osteoporosis medications (AOM) to disrupt this frac- Danish Data Protection Agency, and Statistics Denmark (ref. ture cascade by reducing subsequent fracture risk has been number 706638). For this type of studies, ethical committee demonstrated [10]. Hence, Danish guidelines—generally in approval is not required. accordance with international guidelines—recommend as- In this retrospective analysis, we applied a repeated cross- sessment for osteoporosis in patients with any fragility frac- sectional design to evaluate the post-fracture anti-osteoporosis ture, while recognizing vertebral and proximal hip fractures as treatment gap in consecutive annualized cohorts of patients highly suggestive of osteoporosis and thus mandating AOM sustaining an incident MOF from 2005 to 2014. therapy unless disqualifying circumstances prevail [11–16]. Furthermore, we assessed—as part of this analysis—the im- Despite these recommendations, the majority of fragility pact of hospital-administered AOM on the treatment gap. The fracture patients are not initiated on AOM, as has been dem- treatment gap was defined as the proportion of fracture pa- onstrated consistently across settings and healthcare systems tients not treated with AOM within 1 year of the index frac- [17–26]. This post-fracture treatment gap seems to be more ture. We performed the analyses separately for MOF and for pronounced in men [24–26]. We have recently published a each of the components of the composite MOF outcome, study of the changing treatment gap from 2005 to 2015 across respectively. the United Kingdom, Catalonia, and Denmark, which con- To better understand the AOM treatment initiation and firmed the continuing existence of this treatment gap—with treatment gap, we also evaluated if any baseline covariates a persistent treatment gap of approximately 88–90% in where associated with an increased likelihood of initiating Denmark—among patients with any first incident osteoporot- AOM treatment. In addition, we also evaluated treatment per- ic fracture [27]. We have only identified one treatment gap sistence, defined as the proportion of patients who remained study attempting to identify the use of hospital-administered on AOM treatment one year after the first AOM treatment AOM [28], and only few studies reporting the use of AOM administration following the index fracture. which in Denmark would be administered at a hospital— zoledronic acid and potentially denosumab—in their treat- ment gap assessment [19, 29]. Not including hospital- Data sources administered AOM could inflate treatment gap estimates but capturing such infusions and injections at the individual pa- We utilized data from the Danish Health Registries. This in- tient level is surprisingly challenging in most healthcare cluded the National Patient Register, which contains informa- tion on all admissions to public Danish hospitals since 1977 Osteoporos Int (2021) 32:1961–1971 1963 and out-patient visits since 1995, including diagnosis codes administered AOM group. The ATC and SKS codes used to and treatments; the National Cause of Death Register, giving identify AOM are listed in supplementary table 1. date and cause of death from 1970 and onwards; and the Treatment with glucocorticoids (only oral glucocorticoids National Prescription Database, which contains information were included) and hormone replacement therapy was defined on all filled prescriptions from 1995. These registries cover by the filling of at least one prescription at a pharmacy. the entire Danish population until death or migration. Statistical analysis Study population Each annual fracture cohort was analyzed separately. As part of the repeated cross-sectional design, patients entered each Eligible for inclusion were men and women with incident hip, cohort at the time of their first eligible incident fracture in the vertebral, humerus, and/or forearm fractures between 2005 respective calendar year. Patients were censored at the time of and 2014, aged 50 years or older at the time of the fracture. death, migration, or 1 year after the index fracture, whichever Any first incident hip, vertebral, humerus, and forearm frac- occurred first. ture in a given calendar year (= index fracture) qualified for Treatment persistence was operationalized as either the fill- inclusion in the respective annual cohort for that fracture type, ing of a receipt for an AOM within 120 days of the 1-year while only the first of these fractures within a given calendar anniversary for the first AOM treatment administration fol- year would qualify for inclusion in the respective annual MOF lowing the index fracture (as a pack of alendronic acid lasts cohort. If a fracture had been coded for the same anatomical up to 84 days, and then adding a grace period of 25%) or as location within the past 6 months, the new fracture code filled prescriptions (in Defined Daily Doses) accounting for at would be considered to be either a complication to or least one full year of AOM use (including a 25% grace peri- follow-up for the index fracture and thus not be a cause for od). Patients receiving zoledronic acid infusion were consid- inclusion in the respective fracture cohort. Hence, patients can ered treatment persistent at 1-year. contribute to the treatment gap analysis in several fracture To identify factors associated with AOM treatment initia- cohorts, yet only once per fracture group (MOF, hip, vertebral, tion, we performed a logistic regression analysis for the 2014 humerus, forearm) per calendar year, with the rationale being MOF cohort, with all baseline covariates used as input into the that the same patient can present a missed opportunity for anti- model. osteoporosis treatment in more than one calendar year and for Data were analyzed using Stata version 16.1. Categorical more than one type of fracture. Patients were excluded if they data were summarized by number and proportion of patients, had a history of breast or prostate cancer, bone metastasis, while continuous variables were summarized by mean. All and/or Paget’s disease. analyses were stratified according to the year and type of The diagnosis codes used to identify fractures are listed in fracture. Outcomes analyses were performed on an overall supplementary table 1. level, and also stratified according to gender and age at index fracture (<75 vs ≥75 years). Data extraction For each participant in each cohort, the date of the index frac- Results ture was defined as baseline. The baseline period to identify demographic and clinical characteristics was defined as the Demographics of the fracture populations year prior to the index fracture for use of pharmaceutical prod- ucts, the last 5 years prior to the index fracture for Charlson For the MOF cohorts, a total of 249,897 fracture events met Comorbidity index, and any time prior to baseline for other the criteria for evaluation for the study. Following exclusion baseline covariates including medical history. of patients with a diagnosis code for breast cancer (n = Follow-up data included treatment with AOM from baseline 10,133), prostate cancer (n = 3192), or bone metastasis or until censored, defined by treatment with bisphosphonates (in- Paget’s disease (n = 392), a total of 236,180 fracture events cluding zoledronic acid), raloxifene, teriparatide, denosumab, were included in the MOF cohorts (Table 1). For the hip, or strontium ranelate. Anti-osteoporosis medication therapy vertebral, humerus, and forearm fracture cohorts, a total of was identified as the filling of at least one prescription at a 81,498, 18,343, 46,303, and 98,602 fracture events were in- pharmacy, and/or by at least one procedure code (SKS code, cluded, respectively (supplementary tables 2 to 5). The num- which is a Danish healthcare classification system) for admin- ber of hip fractures decreased over the years, while the number istration of zoledronic acid or denosumab. Users of AOM who of vertebral fractures increased. filled prescriptions but also received hospital-administered The mean age of the MOF cohorts were 72.6–73.9 years AOM (by “SKS” code) were counted only in the hospital- and 24.3–26.6% were men (Table 2). A major osteoporotic 1964 Osteoporos Int (2021) 32:1961–1971 Table 1 Flow table for patients sustaining a first incident major Year Evaluated Excluded Included osteoporotic fracture in a given year. History of bone metastasis History of breast cancer History of prostate cancer History of bone and history of Paget’s disease has metastasis or been pooled due to small n Paget’s disease 2005 25,279 800 276 34 24,169 2006 25,184 892 244 25 24,023 2007 24,158 896 259 34 22,969 2008 24,163 872 282 60 22,949 2009 24,196 921 292 41 22,942 2010 27,499 1132 362 42 25,963 2011 26,154 1106 376 40 24,632 2012 24,110 1153 357 42 22,558 2013 24,674 1190 364 35 23,085 2014 24,480 1171 380 39 22,890 Total 249,897 10,133 3192 392 236,180 fracture had previously been sustained in 24–29% of the MOF around 80 years), and more were men (28–31%). A larger patients, while 33–40% had previously sustained any fracture. proportion had previously sustained a major osteoporotic frac- Less than 10% had been treated with corticosteroids in the ture (30–35%) or any fracture (39–45%), and 12–13% had previous year. previously sustained a hip fracture. The vertebral, humerus, Compared to the MOF cohorts, the hip fracture cohorts and forearm fracture cohorts (supplementary tables 7 to 9) (supplementary table 6) were older at baseline (mean age were overall comparable to the MOF cohorts, except that the Table 2 Baseline characteristics of patients sustaining a first incident major osteoporotic fracture in a given year. AOM, anti-osteoporosis medication; HRT, hormone replacement therapy; MOF, major osteoporotic fracture 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 N 24,169 24,023 22,969 22,949 22,942 25,963 24,632 22,558 23,085 22,890 Age (years (mean)) 73.8 73.5 73.9 73.8 73.5 72.6 72.9 73.3 73.3 73.4 Sex (male) 24.4% 24.6% 25.1% 24.9% 25.2% 24.3% 25.8% 26.1% 25.9% 26.6% Index fracture Vertebral 6.5% 6.6% 6.7% 7.1% 7.1% 6.7% 7.6% 8.8% 8.9% 9.9% Humerus 18.2% 18.0% 18.5% 19.0% 19.3% 18.0% 18.3% 19.0% 19.1% 18.8% Hip 36.2% 35.1% 36.5% 36.0% 34.7% 31.3% 31.2% 32.7% 31.8% 31.9% Forearm 39.0% 40.3% 38.3% 37.9% 38.9% 44.0% 42.8% 39.5% 40.2% 39.4% Previous fracture MOF 24.0% 24.8% 25.6% 26.4% 26.4% 25.6% 26.2% 28.2% 28.0% 28.8% Any 32.7% 33.8% 35.1% 36.0% 36.4% 35.7% 36.6% 39.3% 39.2% 40.2% Medical therapy1 Corticosteroids 9.4% 8.9% 9.4% 8.9% 8.9% 8.5% 8.9% 8.6% 8.7% 8.7% HRT 8.7% 9.4% 9.5% 9.5% 9.5% 10.2% 9.6% 10.4% 10.6% 11.0% AOM 8.0% 9.2% 9.6% 10.4% 10.7% 11.1% 11.8% 12.6% 12.2% 12.0% Charlson comorbidity index score2 =0 83.2% 83.2% 82.8% 83.1% 82.3% 84.1% 83.9% 83.2% 83.2% 83.2% =1 10.0% 9.9% 9.9% 9.5% 10.0% 9.2% 9.0% 9.5% 9.3% 9.2% =2 4.1% 4.1% 4.6% 4.4% 4.6% 4.0% 4.3% 4.3% 4.3% 4.4% ≥3 2.8% 2.9% 2.8% 3.0% 3.0% 2.7% 2.8% 2.9% 3.2% 3.3% 1 Within 1 year prior to the index fracture 2 Within 5 years prior to the index fracture Osteoporos Int (2021) 32:1961–1971 1965 vertebral fracture cohorts had higher proportions of men (40– over the time span of our analysis, from a starting point in 44%), while the forearm fracture cohorts were younger (mean 2005 at 82% and 88%, respectively, to 74% and 82% in age around 70 years) and had lower proportions of men (18– 2014. This was driven by an increase in the proportion of 19%). patients already on AOM therapy at the time of the fracture and remaining on therapy post-fracture, and—primarily—for Treatment gap hip fracture patients also by an increase in the proportion of patients being initiated on AOM therapy after the fracture. The post-fracture treatment gap in the 2005 MOF cohort was In comparison, the treatment gaps for patients with verte- approximately 85%, decreasing to 79% in 2014 (Fig. 1). The bral (panel 2B) and humerus (panel 2C) fractures were rela- narrowing treatment gap was mainly due to more patients tively stable from 2005 to 2014. Patients with vertebral frac- already on anti-osteoporosis therapy at the time of fracture, tures were significantly more likely to receive AOM treatment increasing from 8% in 2005 to 12% in 2014, although dem- within 1 year of the index fracture (treatment gap of 63 to onstrating a decreasing trend from 2012 to 2014. The propor- 66%), driven both by more patients already on treatment at tion of fracture patients initiating treatment within 1 year of the time of the fracture and by more patients initiating treat- their fracture increased from 7% to 9%. While most treatment ment after the fracture. The treatment gap in humerus fracture initiations were prescription-based, an increasing proportion patients varied between 83 and 88%. of fracture patients were initiated on hospital-administered Initiation of hospital-administered AOM increased for all AOM. In 2014, however, this was still as little as 0.6% of all fracture types from 2005 to 2014, yet were more frequently patients in the MOF cohort. initiated in patients with vertebral (0.8 to 2.2%) and hip (0.1 to Treatment persistence, defined as patients receiving AOM 1.3%) fractures, as compared to patients with humerus (0 to within 1 year of the index fracture and still receiving AOM 0.6%) and forearm (0 to 0.4%) fractures. For all fracture types, after an additional year, was found to be 81% across the an- initiation of hospital-administered AOM seemed to stagnate nualized fracture cohorts. or even decrease from 2012 and onwards. Treatment gap according to fracture location Treatment gap according to gender Stratifying the treatment gap analysis according to fracture The treatment gap was significantly smaller in women than in location (Fig. 2), we found that the treatment gaps for hip men, irrespective of index fracture cohort and index year (Fig. (panel 2A) and forearm (panel 2D) fracture patients decreased 3 and supplementary figures 1–4). In men, the treatment gap Fig. 1 Treatment rate and treatment gap for patients sustaining a first incident major osteoporotic fracture within any given year (2005–2014). The numbers above the bars indicate the proportion of patients initiating hospital-administered treatment (zoledronic acid or denosumab) 1966 Osteoporos Int (2021) 32:1961–1971 Fig. 2 Treatment rate and treatment gap for patients sustaining a first indicate the proportion of patients initiating hospital-administered treat- incident hip (panel 2A), vertebral (2B), humerus (2C), or forearm (2D) ment (zoledronic acid or denosumab); no number listed if too few fracture within any given year (2005–2014). The numbers above the bars observations in the MOF and hip fracture cohorts decreased from 2005 to forearm fracture cohorts. In women, the treatment gap de- 2014, while it was stable in the vertebral, humerus, and creased from 2005 to 2014, irrespective of fracture cohort. Fig. 3 Treatment rate and treatment gap for patients sustaining a first indicate the proportion of patients initiating hospital-administered treat- incident major osteoporotic fracture within any given year (2005– ment (zoledronic acid or denosumab). Treatment gap lines are trend lines 2014), stratified according to gender. The numbers above the bars Osteoporos Int (2021) 32:1961–1971 1967 Fig. 4 Use of hospital- administered AOM in patients sustaining a first incident major osteoporotic fracture in any given year (2005–2014). The figure shows the proportion of all AOM- treated MOF patients treated with hospital-administered AOM (zo- ledronic acid or denosumab), stratified according to gender. Index year has been grouped for smoothening. AOM, anti- osteoporosis medication; MOF, major osteoporotic fracture Treatment gap according to age at index fracture Factors associated with AOM treatment The treatment gap was consistently smaller in older (age ≥ 75 In the 2014 MOF cohort, age (odds ratio [OR] 1.01; 95% years at index fracture) than in younger (<75 years) patients, confidence interval [CI] 1.00–1.01), treatment with corticoste- except for hip fracture patients in whom the treatment gap was roids (OR 2.32; 95% CI 2.10–2.59), and hormone replace- similar regardless of the age stratification (supplementary fig- ment therapy (OR 1.22; 95% CI 1.11–1.35) were associated ures 5–9). In the MOF, humerus, and forearm fracture cohorts with an increased likelihood of receiving AOM treatment. the differences between the younger vs the older age group Male gender was associated with a reduced likelihood of treat- was generally below 10%, whereas in the vertebral fracture ment (OR 0.37; 95% CI 0.34–0.40). Patients with a prior cohort the differences between the groups fluctuated around MOF (OR 1.38; 95% CI 1.22–1.54) or any prior fracture 15%. During the time span of our analysis, the treatment gap (OR 1.32; 95% CI 1.19–1.48) were more likely to receive demonstrated a decreasing trend in both age groups across the treatment. fracture cohorts, except for vertebral fracture patients in whom As compared to patients with an index forearm fracture, it was relatively stable. patients with an index vertebral fracture (OR 3.00; 95% CI 2.69–3.35) or an index hip fracture (OR 1.60; 95% CI 1.48– 1.73) were more likely to receive AOM treatment, whereas Proportion of AOM not captured in prescription index humerus fractures where associated with a reduced like- registers lihood of treatment (OR 0.84; 95% CI 0.76–0.94). Using a CCI score of 0 as the reference, a CCI of 1 was The amount of hospital-administered AOM out of the total associated with an increased likelihood of receiving AOM uptake of osteoporosis medications was relatively small over treatment (OR 1.22; 95% CI 1.09–1.36), a CCI of 2 was not the period studied. Specifically, the use of hospital- different from the reference at a statistically significant level administered AOM as a proportion of total AOM use within (OR 1.11; 95% CI 0.94–1.30), and a CCI of 3 or above was 1 year after a major osteoporotic fracture increased initially, associated with a reduced likelihood of receiving AOM treat- yet appears to decrease towards the end of our analysis (Fig. ment (OR 0.79; 95% CI 0.64–0.96). 4). During our analysis, an average of 3.9% of the AOM- treated MOF patients where using hospital-administered AOM. The general trend was that hospital-administered Discussion AOM were relatively more often used in men, and use of hospital-administered AOM varied significantly between frac- Our findings demonstrate a critical gap in post-fracture care ture cohorts and index fracture year, ranging from 0 to 18% with large and persistent anti-osteoporosis treatment gaps in (supplementary table 10). men and women sustaining a MOF. For the hip and forearm 1968 Osteoporos Int (2021) 32:1961–1971 fracture cohorts, the treatment gap—based on the filling of at gaps in Denmark [27]. This current paper was conceptualized least one prescription or receiving at least one parenteral oste- to evaluate if inclusion of hospital-administered AOM would oporosis drug dose—decreased marginally from 2005 to reduce this treatment gap. Further, we here evaluate the treat- 2014, yet in the 2014 cohorts, 3 out of 4 hip fracture patients ment gap after any MOF allowing each individual to contrib- and 4 out of 5 forearm fracture patients were still not treated ute to several fracture cohorts representing additional missed with AOM within 1 year of the index fracture. The treatment opportunities for AOM treatment initiation, whereas the first gaps for vertebral and humerus fracture patients were gener- paper evaluated the treatment gap after any first incident fra- ally stable at around 65% and 85%, respectively, across the gility fracture. Reassuringly, this current paper demonstrates 2005 to 2014 time period. The treatment gap was more pro- smaller—yet still very large—treatment gaps in Denmark in nounced in men than in women, and in younger than in older patients with a major osteoporotic fracture as compared to any patients (<75 vs ≥75 years). It should be noted that not all osteoporotic fracture as was reported previously. For hip, ver- fragility fracture patients are candidates for AOM treatment tebral, and forearm fractures (MOF and humerus fractures [31], as contraindications, patient preferences, and other indi- were not independently evaluated in the former paper) the vidual or local factors may speak against treatment initiation. treatment gaps are generally smaller in this current analysis, Furthermore, Danish treatment guidelines recommend that probably owing to the differences in study design and the AOM treatment decisions in patients with non-hip non-verte- inclusion of hospital-administered AOM [27]. bral fragility fractures should be based also on further evalu- The cause of the treatment gap could plausibly, at least in ations to estimate the risk of subsequent fractures [11, 12], part, be attributed to insufficient coordination of care. Several which may be reflected in the treatment gap among humerus structural models to improve post-fracture care have been de- and forearm fracture patients identified in this study. veloped, and a review evaluating 10 such interventions report- However, in order to benchmark the performance of the ed all of them demonstrating improvements in post-fracture Danish hospital services against international metrics we use initiation of AOM therapy [34]. Among these care models, the the term treatment gap in the internationally accepted sense. fracture liaison service (FLS) has, in particular, been demon- Also, it is worth noting that FRAX based guidelines largely strated to effectively improve treatment initiation rates and set the intervention threshold at the equivalence point of a treatment compliance in a cost-effective manner [35, 36]. postmenopausal woman with a prior fragility fracture [32]. However, while the use of FLS units is being employed across Our findings demonstrate that the use of AOM in secondary the globe and a best practice framework has been defined, no fracture prevention is currently insufficient, which is in line accredited FLS units were in place in Denmark during the with what has been demonstrated in previous studies. In MOF time span of our analysis, and today only four Danish hospi- patients not treated with AOM at the time of the fracture, tals run an internationally accredited FLS (www. Leslie et al. (2012) found a persistent treatment gap above capturethefracture.org) [36]. The lack of a coordinated effort 85% in Canada between 1996 and 2008, with the gap being to initiate evaluation and treatment in fracture patients almost larger in men [25]. Also including patients treated with AOM certainly contributed to the high rates of untreated MOF at the time of the fracture, Wilk et al. (2014) demonstrated a 1 patients seen in our analysis. Another reason for the year post-fracture treatment gap around 75% in women with treatment gap may include a lack of knowledge in patients hip fractures in the USA [21], consistent with a treatment gap and healthcare providers about the risks and benefits of around 73% in men and women with hip fractures in Hawaii AOM, as suggested elsewhere [29, 37]. as demonstrated by Nguyen et al. (2018) [26]. Other studies In this analysis, we also found that inclusion of hospital- have demonstrated similar findings [17, 19], indicating that administered AOM in the treatment gap assessment did not the lack of post-fracture anti-osteoporosis medical therapy materially reduce the treatment gap after a MOF in Denmark. seems to be a universal problem. Barton and colleagues Hence, an average of only 3.9% of AOM-treated patients in (2019) demonstrated decreasing treatment rates in AOM the MOF cohorts were using hospital-administered AOM (zo- naïve patients sustaining an incident vertebral fracture, sug- ledronic acid or denosumab) within 1 year of their index frac- gesting that this may be due to fear of rare bisphosphonate side ture. This demonstrated substantial variation between genders effects such as osteonecrosis of the jaw and atypical femur and index fracture types, with proportionately more men than fractures [33]. However, while some time trend analyses indi- women using hospital-administered AOM, and with hospital- cate that the problem may be getting worse in recent years [19, administered AOM being used more often in vertebral fracture 25, 29], our analysis suggests a stable or slightly improving patients (supplementary table 10). Hospital-administered treatment gap from 2005 to 2014 in Denmark. AOM was initiated more often in vertebral and hip fracture The post-fracture treatment gap in Denmark was evaluated patients than in humerus and forearm fracture patients (Fig. 2). recently—conjointly with the United Kingdom and In a recent study, Axelsson et al. (2020) attempted to iden- Catalonia—using comparable data sources linked to the tify the use of hospital-administered AOM by combining a IFRISK study, demonstrating large and persistent treatment diagnosis code for osteoporosis with codes for intravenous Osteoporos Int (2021) 32:1961–1971 1969 or subcutaneous drug administration, using Swedish health indicating that patients who would seem to be at higher risk registers. While the validity of this approach is uncertain, they of subsequent fractures are indeed more likely to receive found an increasing proportion of fracture patients meeting AOM treatment. Treatment was less likely in subjects with a these criteria for parenteral AOM-administration following low comorbidity score (Charlson index of zero) but also in implementation of an FLS unit (going from 11.8 to 18.9%) patients with substantial comorbidity (Charlson index of 3 or [28]. Cheung et al. (2018) reported in an analysis of the Hong higher), the latter likely results from a combination of contra- Kong AOM treatment gap, that of all patients on AOM, 4% indications against bisphosphonates or a clinical decision where on zoledronic acid and 1% on denosumab [19]. based on expected short remaining lifespan. We also con- Solomon et al. (2014) reported that 0.3% of AOM-treated firmed that male gender was associated with a markedly re- patients were using denosumab within 1 year of the index duced propensity to receive AOM treatment, indicating that fracture, yet did not provide distinct figures on the use of male osteoporosis is still neglected in clinical practice. These zoledronic acid [29]. Including hospital-administered AOM findings could be used to tailor future secondary fracture pre- in the treatment gap assessment does not seem to materially vention strategies. alter the treatment rate in Denmark, yet this may differ signif- The strengths of this study include the use of validated icantly between countries. In this regard, it is important to note health registries and prescription databases, and the large that zoledronic acid was first granted marketing authorization number of patients and fractures accrued over a substantial for use in osteoporosis by the European Medicines Agency in time period without data breaks. Further, to our knowledge 2005, and denosumab in 2010. Thus, the years covered in this this is the first study using Danish data to include hospital- study reflects the early use of both drugs, which may partially administered AOM therapy in the evaluation of the post- explain our findings. Hence, future treatment gap evaluations fracture AOM treatment gap. should preferably include hospital-administered AOM. Limitations do exist, including the small risk of diagnosis The reason for vertebral and hip fractures more commonly miscoding which is an inherent part of registry-based research. being a cause for initiating hospital-administered AOM than Further, prescription data originates at pharmacy level, thus humerus and forearm fractures, could be that Danish osteopo- reflecting filled prescriptions. Hence, we are not able to ascer- rosis guidelines recognize the former as diagnostic for osteo- tain the number of patients actually prescribed an AOM. porosis [11, 12] in line with these fractures carrying a very Further, we cannot assess from our data if patients filling their high risk of subsequent fractures [38]. This could potentially prescriptions actually ingest their AOM. This, coupled with instill a greater willingness among the treating physicians to the fact that our study did not require patients to remain on use hospital resources to ensure AOM treatment insofar con- AOM throughout the full year post-fracture, likely resulted in traindications or side effects to oral therapies exist. The sig- an underestimation of the true treatment gap. To this end, we nificant reduction in mortality in hip fracture patients treated found a 1-year treatment persistence around 81%. with zoledronic acid, as demonstrated in the HORIZON An additional limitation pertains to the use of zoledronic Recurrent Fracture Trial, may also have contributed to this acid. From the data, we cannot ascertain if it has been admin- group more frequently being prescribed a hospital- istered due to osteoporosis or in the oncology-setting (e.g. administered AOM [39]. tumor-induced hypercalcaemia). Hence, in this study the indi- The notion of a higher proportion of men than women cation for using zoledronic acid following an index fracture being treated with hospital-administered AOM may reflect may be biased, yet even so, fracture patients receiving zole- the possibility that side effects and/or contraindications to oral dronic acid—regardless of indication—does not contribute to AOM could occur more frequently in men. Further, hospital the treatment gap, why this would merely serve to deflate the administration of AOM may be less prone to primary non- treatment gap and thus does not affect the conclusions of our compliance; hence, the finding may reflect a higher proportion study. of men than women being prescribed a hospital-administered Another aspect is the completeness of our data in terms of therapy due to a perceived risk of primary non-compliance at hospital administration of zoledronic acid and denosumab by the time of AOM-prescription. the use of procedure codes. While this has not been formally Based on the logistic regression analysis for patients sus- validated, there is a sizeable yet redeemable cost to the depart- taining a MOF (in 2014), we reassuringly found that patients ment for such hospital administration; hence, we are confident with an index vertebral or hip fracture were more likely to that the vast majority of administrations were indeed captured. receive AOM treatment, which is in line with the Danish treat- In conclusion, this analysis has demonstrated that a major ment guidelines recognizing these fractures as diagnostic for gap in post-fracture AOM treatment was evident in patients osteoporosis [11, 12]. We also found an apparent association sustaining a major osteoporotic fracture in Denmark from between common fracture risk factors (including age, prior 2005 to 2014. Regrettably, the magnitude of this gap was fractures and corticosteroid treatment) and AOM treatment, not materially altered by the inclusion of hospital- 1970 Osteoporos Int (2021) 32:1961–1971 administered AOM. While our results echo the call for References strengthening post-fracture care, this study also highlights the continued need for a systematic, proactive, nationwide 1. Brandi ML (2009) Microarchitecture, the key to bone quality. Rheumatol (United Kingdom) 48:iv3–iv8. https://doi.org/10.1093/ approach to the organization of secondary fracture prevention. rheumatology/kep273 2. Hernlund E, Svedbom A, Ivergård M et al (2013) Osteoporosis in Supplementary Information The online version contains supplementary the European Union: Medical management, epidemiology and eco- material available at https://doi.org/10.1007/s00198-021-05890-x. nomic burden: a report prepared in collaboration with the International Osteoporosis Foundation (IOF) and the European Funding UCB funded this study. All analyses were conducted indepen- Federation of Pharmaceutical Industry Associations (EFPIA). dently by the academic researchers involved. Arch Osteoporos 8:1–115. https://doi.org/10.1007/s11657-013- 0136-1 Data availability Not available due to legislative restrictions on data 3. Driessen JHM, Hansen L, Eriksen SA et al (2016) The epidemiol- sharing. ogy of fractures in Denmark in 2011. Osteoporos Int 27:2017– 2025. https://doi.org/10.1007/s00198-016-3488-8 4. Klotzbuecher CM, Ross PD, Landsman PB et al (2000) Patients Code availability Not available. with prior fractures have an increased risk of future fractures: a summary of the literature and statistical synthesis. J Bone Miner Declarations Res 15:721–739. https://doi.org/10.1359/jbmr.2000.15.4.721 5. Van Geel T, Van Helden S, Geusens P et al (2009) Clinical subse- Ethics approval The study was approved by the Danish Medicines quent fractures cluster in time after first fractures. Ann Rheum Dis Agency, the Danish Data Protection Agency, and Statistics Denmark 68:99–102. https://doi.org/10.1136/ard.2008.092775 (ref. number 706638). For this type of studies, ethical committee approval 6. Johansson H, Siggeirsdóttir K, Harvey NC et al (2017) Imminent is not required. For this type of study formal consent is not required. risk of fracture after fracture. Osteoporos Int 28:775–780. https:// doi.org/10.1007/s00198-016-3868-0 Consent to participate Not applicable. 7. Port L, Center J, Briffa NK et al (2003) Osteoporotic fracture: missed opportunity for intervention. Osteoporos Int 14:780–784. https://doi.org/10.1007/s00198-003-1452-x Consent for publication Given by all authors. 8. Edwards BJ, Bunta AD, Simonelli C et al (2007) Prior fractures are common in patients with subsequent hip fractures. Clin Orthop Conflict of interest MKS: Educational grant and institutional research Relat Res:226–230. https://doi.org/10.1097/BLO. grant, UCB. 0b013e3180534269 MTE: Institutional grant, UCB. 9. Frederiksen A, Abrahamsen B, Johansen PB, Sorensen HA (2018) SK: No conflict of interest. Danish, national cross-sectional observational study on the preva- CL and ET: Employees of UCB pharma. lence of prior major osteoporotic fractures in adults presenting with CC: Personal fees from Alliance for Better Bone Health, Amgen, Eli hip fracture-limitations and scope for fracture liaison services in Lilly, GSK, Medtronic, Merck, Novartis, Pfizer, Roche, Servier, Takeda prevention of hip fracture. Osteoporos Int 29:109–114. https://doi. and UCB. org/10.1007/s00198-017-4247-1 AD: No conflict of interest. KHR: No conflict of interest. 10. Geusens P, Lems WF, Bours S, vd Bergh JP (2019) Secondary MKJ: Personal fees from Amgen and UCB, unrestricted grants from fracture prevention: drug treatment, fall prevention and nutrition Amgen. requirements. Best Pract Res Clin Rheumatol 33:290–300. https:// DML: Consulting or speakers fees from Amgen, Lilly, Novartis, doi.org/10.1016/j.berh.2019.04.005 Italfarmaco, Ferrer and Rubió, outside the submitted work. 11. Hitz M, Harsløf T, Ejersted C, et al (2019) NBV: postmenopausal DPA: Research grants from Amgen, Johnson & Johnson, and UCB. osteoporose Consultation or speaker fees paid to my department by Amgen, Astellas, 12. Hermann P, Frost M, Abrahamsen B, et al (2020) NBV: and UCB. Behandling af mandlig osteoporose BA: Institutional research grants UCB, Novartis. Consulting or 13. Kanis J, Cooper C, Rizzoli R et al (2019) European guidance for the speakers fees UCB, Kyowa-Kirin, Amgen, Eli-Lily, Pharmacosmos. diagnosis and management of osteoporosis in postmenopausal women. Osteoporos Int 30:3–44. https://doi.org/10.1007/s00198- 018-4704-5 Open Access This article is licensed under a Creative Commons 14. Cooper C, Ferrari S, IOF Board and Executive Committee (2017) Attribution-NonCommercial 4.0 International License, which permits IOF Compendium of Osteoporosis any non-commercial use, sharing, adaptation, distribution and reproduc- 15. Compston J, Cooper A, Cooper C et al (2017) UK clinical guideline tion in any medium or format, as long as you give appropriate credit to the for the prevention and treatment of osteoporosis. Arch Osteoporos original author(s) and the source, provide a link to the Creative Commons 12:43. https://doi.org/10.1007/s11657-017-0324-5 licence, and indicate if changes were made. The images or other third 16. Conley RB, Adib G, Adler RA et al (2020) Secondary fracture party material in this article are included in the article's Creative prevention: consensus clinical recommendations from a Commons licence, unless indicated otherwise in a credit line to the ma- multistakeholder coalition. J Bone Miner Res 35:36–52. https:// terial. If material is not included in the article's Creative Commons licence doi.org/10.1002/jbmr.3877 and your intended use is not permitted by statutory regulation or exceeds 17. Elliot-Gibson V, Bogoch ER, Jamal SA, Beaton DE (2004) Practice the permitted use, you will need to obtain permission directly from the patterns in the diagnosis and treatment of osteoporosis after a fra- copyright holder. To view a copy of this licence, visit http:// gility fracture: a systematic review. Osteoporos Int 15:767–778. creativecommons.org/licenses/by-nc/4.0/. https://doi.org/10.1007/s00198-004-1675-5 Osteoporos Int (2021) 32:1961–1971 1971 18. Papaioannou A, Giangregorio L, Kvern B et al (2004) The osteo- 30. (2018) Multinational Observational Database Study on Imminent porosis care gap in Canada. BMC Musculoskelet Disord 5:1–6. Osteoporotic Fracture Risk: Stage 1. In: Eur. Union Electron. https://doi.org/10.1186/1471-2474-5-11 Regist. Post-Authorisation Stud. (EU PAS Regist. http://www. 19. Cheung M, Ho AW, Wong S (2018) Post-fracture care gap: a ret- encepp.eu/encepp/viewResource.htm?id=24100. Accessed 23 rospective population-based analysis of Hong Kong from 2009 to Jun 2020 2012. Hong Kong Med J 24:579–583. https://doi.org/10.12809/ 31. Javaid MK, Sami A, Lems W et al (2020) A patient-level key hkmj187227 performance indicator set to measure the effectiveness of fracture 20. Fraser LA, Ioannidis G, Adachi JD et al (2011) Fragility fractures liaison services and guide quality improvement: a position paper of and the osteoporosis care gap in women: the Canadian Multicentre the IOF Capture the Fracture Working Group, National Osteoporosis Study. Osteoporos Int 22:789–796. https://doi.org/10. Osteoporosis Foundation and Fragility Fracture. Osteoporos Int 1007/s00198-010-1359-2 31:1193–1204. https://doi.org/10.1007/s00198-020-05377-1 21. Wilk A, Sajjan S, Modi A et al (2014) Post-fracture pharmacother- 32. Kanis JA, Harvey NC, Johansson H et al (2017) FRAX update. J apy for women with osteoporotic fracture: analysis of a managed Clin Densitom 20:360–367. https://doi.org/10.1016/j.jocd.2017.06. care population in the USA. Osteoporos Int 25:2777–2786. https:// 022 doi.org/10.1007/s00198-014-2827-x 33. Barton DW, Behrend CJ, Carmouche JJ (2019) Rates of osteopo- 22. Shibli-Rahhal A, Vaughan-Sarrazin MS, Richardson K, Cram P rosis screening and treatment following vertebral fracture. Spine J (2011) Testing and treatment for osteoporosis following hip frac- 19:411–417. https://doi.org/10.1016/j.spinee.2018.08.004 ture in an integrated U. S. healthcare delivery system. Osteoporos 34. Little EA, Eccles MP (2010) A systematic review of the effective- Int 22:2973–2980. https://doi.org/10.1007/s00198-011-1536-y ness of interventions to improve post-fracture investigation and 23. Papaioannou A, Kennedy CC, Ioannidis G et al (2008) The osteo- management of patients at risk of osteoporosis. Implement Sci 5. porosis care gap in men with fragility fractures: the Canadian https://doi.org/10.1186/1748-5908-5-80 Multicentre Osteoporosis Study. Osteoporos Int 19:581–587. 35. Ganda K, Puech M, Chen J et al (2013) Models of care for the https://doi.org/10.1007/s00198-007-0483-0 secondary prevention of osteoporotic fractures: a systematic review 24. Lüthje P, Nurmi-Lüthje I, Kaukonen JP et al (2009) Undertreatment and meta-analysis. Osteoporos Int 24:393–406. https://doi.org/10. of osteoporosis following hip fracture in the elderly. Arch Gerontol 1007/s00198-012-2090-y Geriatr 49:153–157. https://doi.org/10.1016/j.archger.2008.06.007 25. Leslie WD, Giangregorio LM, Yogendran M et al (2012) A 36. Mitchell P, Åkesson K, Chandran M et al (2016) Implementation of population-based analysis of the post-fracture care gap 1996- Models of Care for secondary osteoporotic fracture prevention and 2008: the situation is not improving. Osteoporos Int 23:1623– orthogeriatric Models of Care for osteoporotic hip fracture. Best 1629. https://doi.org/10.1007/s00198-011-1630-1 Pract Res Clin Rheumatol 30:536–558. https://doi.org/10.1016/j. 26. Nguyen ET, Posas-Mendoza T, Siu AM et al (2018) Low rates of berh.2016.09.008 osteoporosis treatment after hospitalization for hip fracture in 37. Jaglal SB, McIsaac WJ, Hawker G et al (2003) Information needs in Hawaii. Osteoporos Int 29:1827–1832. https://doi.org/10.1007/ the management of osteoporosis in family practice: an illustration of s00198-018-4553-2 the failure of the current guideline implementation process. 27. Skjødt MK, Khalid S, Ernst M et al (2020) Secular trends in the Osteoporos Int 14:672–676. https://doi.org/10.1007/s00198-003- initiation of therapy in secondary fracture prevention in Europe: a 1421-4 multi-national cohort study including data from Denmark, 38. Center JR, Bliuc D, Nguyen TV, Eisman JA (2007) Risk of subse- Catalonia, and the United Kingdom. Osteoporos Int 31:1535– quent fracture after low-trauma fracture in men and women. JAMA 1544. https://doi.org/10.1007/s00198-020-05358-4 297:387–394. https://doi.org/10.1001/jama.297.4.387 28. Axelsson KF, Johansson H, Lundh D et al (2020) Association 39. Lyles KW, Colón-Emeric CS, Magaziner JS et al (2007) Zoledronic Between recurrent fracture risk and implementation of fracture liai- acid and clinical fractures and mortality after hip fracture. N Engl J son services in four swedish hospitals: a cohort study. J Bone Miner Med 357:1799–1809. https://doi.org/10.1056/NEJMoa074941 Res 35:1216–1223. https://doi.org/10.1002/jbmr.3990 29. Solomon DH, Johnston SS, Boytsov NN et al (2014) Osteoporosis Publisher’s note Springer Nature remains neutral with regard to jurisdic- medication use after hip fracture in U.S. patients between 2002 and tional claims in published maps and institutional affiliations. 2011. J Bone Miner Res 29:1929–1937. https://doi.org/10.1002/ jbmr.2202