Pharmacotherapy change patterns after fragility fracture in patients receiving bone-active medication: a fracture liaison service quality improvement cohort study

Abstract

Purpose

We examined pharmacotherapy change patterns in fragility fracture patients who fractured while taking bone active medication, and were screened by Ontario’s Fracture Screening and Prevention Program (FSPP). A secondary objective was to identify baseline characteristics associated with receiving a pharmacotherapy change recommendation.

Methods

The cohort included fragility fracture patients (50 + years), who fractured while taking bone active medication, were screened by the FSPP during 2017–2024, and subsequently evaluated by bone health specialists. Descriptive statistics were used to summarize baseline factors and pharmacotherapy patterns. Factors associated with receiving a recommendation to change medication were investigated using multivariable regression.

Results

Of 828 patients meeting eligibility criteria, 60.4% (n = 500) were advised to change medication, 33.9% (n = 281) to remain on the same medication, and 2.7% (n = 22) to discontinue the medication. Data were unavailable for 3.0% (n = 25). Of those advised to change medication, 83.6% (n = 418) were changed to denosumab, 5.2% (n = 26) to teriparatide, and 2.4% (n = 12) to risedronate. Older age was associated with a higher likelihood of receiving a recommendation to change medication [1.84 (95% CI: 1.15–2.95)], while being on denosumab was associated with a lower likelihood [0.07 (95% CI: 0.05–0.11)].

Conclusion

Findings suggest that a fracture on treatment prompted re-evaluation and treatment escalation in many cases, while younger age and prior denosumab use were associated with lower likelihood of switching, likely reflecting a multitude of clinical considerations and jurisdiction-specific reimbursement constraints. This study provides real-world insights on pharmacotherapy change patterns in a high-risk, understudied patient cohort from a jurisdiction-wide FLS, addressing a group with therapeutic uncertainty.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-026-09548-9.

Introduction

Fracture Liaison Services (FLS) seek to identify and assess patients who sustain fragility fractures in order to optimize bone health management and reduce future fracture risk [1–3]. Patients deemed to be at high risk for future fractures have an indication to initiate bone active medication [4, 5], which can reduce the risk of recurrent fractures by up to 60% [6-8]. Despite their effectiveness, these medications do not fully eliminate fracture risk. Studies suggest on-treatment fracture rates of approximately between 1 and 4 per 100 person-years [9], underscoring the importance of ongoing monitoring and reassessment. When fragility fractures occur while taking bone active medications, physicians (whether or not affiliated with an FLS) are faced with complex clinical decisions regarding continuation, modification or escalation of treatment. Clinical practice guidelines are frequently clear in recommending bisphosphonates and other antiresorptive agents, or more recently, anabolic agents, as first-line treatment to lower the risk of future fractures among high-risk individuals [5, 10–14]. However, there is no formal consensus and limited explicit guidance on pharmacotherapy management for patients who sustain one or more new fragility fracture(s) while taking first-line treatment. Guidelines emphasize that fractures or bone mineral density decline during therapy do not necessarily indicate inadequate treatment response. Best practice recommendations include referral to specialist care, reassessing fracture risk, evaluating secondary causes, confirming adherence, and considering extending or switching therapy after more than one fracture or bone density decline; however, guidance is less clear on the specific criteria for medication switching [4, 14–16]. Evidence demonstrating the potential benefits of switching therapeutic agents remains limited [17, 18]. Moreover, the absence of risk-assessment tools that adequately measure the risk of future fractures in patients who are receiving treatment confers further challenges in evidence-based goal-directed therapeutic management of this patient population [19]. Patients who fracture while taking bone active medication may, correctly or otherwise, perceive their fracture event as a signal of treatment failure, and therefore constitute a critical sub-cohort for monitoring given the risk that they may independently discontinue prescribed pharmacotherapy or seek treatment with alternative agents.

Outcome evaluations of post-fracture interventions typically exclude patients who are taking bone active medication at the time of screening [20–23], contributing to a notable gap in monitoring how post-fracture initiatives manage this subgroup of patients. A subset of fragility fracture patients are already taking bone active medication when they sustain a fragility fracture and present to an FLS [24]; for example, in one community hospital-based fracture screening clinic, 23% of participants were noted to be taking first-line pharmacotherapy at the time of admission to the program [24]. Furthermore, while some FLS evaluations have reported rates of treatment continuation or maintenance as discrete outcomes for this subgroup of patients [25], we found no studies documenting real-world practices in pharmacotherapy change patterns among patients who fracture while taking bone active medication. One study examined treatment failure – defined as two or more fractures – in women receiving osteoporosis therapy, but it did not assess subsequent changes in treatment [26].

As part of a quality improvement initiative in 2017, the Fracture Screening and Prevention Program (FSPP) of Ontario began to collect follow-up data on patients who reported taking bone active medication at the time of screening for a new fragility fracture. Using data collected by this large jurisdiction-wide FLS, our primary study objective was to examine pharmacotherapy change or maintenance patterns among fragility fracture patients who were on bone active medication at the time of fracture and enrolled in the FSPP between 2017 and 2024. A secondary objective was to understand which baseline characteristics were associated with receiving a pharmacotherapy change recommendation, and to add to the limited available knowledge regarding the current management of this subpopulation of patients.

Methods

Data source and setting

The FSPP has been implemented, incrementally, at 39 medium- and high-volume outpatient fracture clinics since 2007 in the province of Ontario, Canada. The Fracture Prevention Coordinators identify patients aged 50 years and above who sustain a fragility fracture and initiate further investigation and referral for their bone health. Comprehensive descriptions of the FSPP have been previously published [27–29]. Patients are surveyed in person or by telephone, with data recorded electronically in a centralized database. This secure database serves as a repository for de-identified patient responses, and continues to be used for quality assurance and improvement of the FSPP. Approval for usage of these data for quality improvement purposes was obtained from the data custodians, Osteoporosis Canada.

Study cohort selection

Our study cohort included fragility fracture patients (aged ≥ 50 years) enrolled in the FSPP between August 2017 and March 2024, who had sustained a fragility fracture while taking bone active medication, and were subsequently referred to and assessed by a bone health specialist. The prevailing Canadian clinical practice guidelines at the time of this study stated that patients who fracture on treatment may benefit from a consultation with bone specialists [5]. The FSPP fast-tracks the referral process for patients who are screened for a new fragility fracture while taking bone active medication. The bone health specialists referred to may include endocrinologists, rheumatologists, geriatricians, or internal medicine specialists practicing in Ontario but are not embedded within the FSPP.

Pharmacotherapy data collection process

FSPP patients who reported taking bone active medication at the time of screening were asked to indicate which medication they were taking. The list of medication choices included etidronate, risedronate, alendronate, zoledronic acid, denosumab, romosozumab, teriparatide, raloxifene, hormone therapy and pamidronate, with the corresponding trade names. Patients who confirmed taking one of the listed bone active medications were then referred to a bone health or osteoporosis specialist for further assessment. The Fracture Prevention Coordinator (through the baseline screening form, or notes from the specialist appointment for patients who did not know their index medication) provided confirmation that the patient was taking bone active medication at the time of screening. Subsequent to the specialist visit, the Fracture Prevention Coordinator followed up with the specialist’s office to collect data on the patient’s pharmacotherapy status (categorized as: remained on the same medication; discontinued medication; changed to a new medication; or refused medication). Any changes to the dosage of the medication would have beeen categorized as “remained on the same medication”.

Covariate measurement

Baseline characteristics examined were age, sex (male/female), fracture site (wrist, elbow, shoulder, spine, pelvis, hip, femur (not hip), tibia/fibula, clavicle), history of falls (yes/no), number of comorbid conditions reported by the patient (heart disease, diabetes, respiratory disease, cancer, arthritis, high blood pressure and cholesterol), index bone active medication and duration of its use. Age was examined as a continuous variable, as well as dichotomous in multivariable analyses (comparing those 50–65 years versus 66 + years of age as this classification is relevant to eligibility for drug coverage benefits in Ontario). History of falls captured falls in the previous year not including any fall that may have led to the current fracture. The index bone active medications were analyzed and presented using their generic names, and further categorized as being antiresorptive, (e.g. risedronate, alendronate, zoledronic acid, etidronate, denosumab), anabolic (e.g. teriparatide) or hormonal therapy (e.g. raloxifene). The duration of index bone active medication use was categorized as ‘<1 year (started before index fracture)’, ‘1–5 years’, and ‘5 + years’.

Analysis

Using descriptive analytical techniques, we examined the baseline characteristics of the study cohort and reported on their pharmacotherapy change recommendations. Appropriate measures of central tendency and proportions for categorical variables were used to generate the descriptive summaries. Using multivariable logistic regression, we aimed to understand if any baseline characteristics were significantly associated with patients who received a recommendation to change their index pharmacotherapy versus those who did not. Covariates considered for adjustment were age, sex, index fracture site, history of falls in the previous year, number of comorbid conditions, and the type and duration of use for the index pharmacotherapy. Statistical significance was assessed at an alpha level of 0.05. All analyses were performed using SAS Version 9.4.

Results

Our study cohort comprised 828 patients, ≥ 50 years of age, who were confirmed to have fractured while taking bone active medication, and who then saw a bone health specialist for a bone health assessment.

Baseline descriptive characteristics

(see Table 1) The average age in the sample was 74.6 years, with 93.5% (n = 774) being female. The most common index fracture site was the fracture of the wrist (43.0%, n = 356), followed by the hip (19.4%, n = 161). 26% (n = 215) of patients reported a history of falls within the previous year and the majority (66.3%, n = 549) reported having one or more comorbid chronic conditions. Among the 828 patients in the cohort, 52.3% (n = 433) reported taking risedronate, 27.2% (n = 225) reported taking denosumab and 19.0% (n = 157) reported taking alendronate. There were six or fewer patients taking each of zoledronic acid, etidronate, and raloxifene. With respect to duration of the pharmacotherapy prior to screening, 8.2% (n = 68) of patients had been taking it for less than one year, 56.8% (n = 470) reported being on treatment for 1–5 years, and 35.0% (n = 290) had been taking it for more than 5 years.

Table 1.

Baseline characteristics of patients who fractured while taking bone active medication

	Full Study Cohort (N = 828)
Age (mean, SD)	74.6 (9.0)
Age 50–65 years (n, %)	113 (13.6)
Age > = 66 years (n, %)	715 (86.4)
Sex - female (n, %)	774 (93.5)
Index fracture site (n, %)
Only wrist	356 (43.0)
Only spine	46 (5.6)
Only hip	161 (19.4)
Other	227 (27.4)
Multiple simultaneous fractures	38 (4.6)
History of falls (n, %)	215 (26.0)
Number of comorbid chronic conditions (n, %)
0	279 (33.7)
1	206 (24.9)
2	180 (21.7)
>= 3	163 (19.7)
Index bone active medication (n, %)
Antiresorptive:
Risedronate	433 (52.3)
Denosumab	225 (27.2)
Alendronate	157 (19.0)
Zoledronic acid	*1-5
Etidronate	*1-5
Hormonal therapy - raloxifene	6 (0.7)
Duration of index pharmacotherapy
5 + years before index fracture	290 (35.0)
1–5 years before index fracture	470 (56.8)
< 1 year before index fracture	68 (8.2)

*Values suppressed due to small cells

Pharmacotherapy change patterns

(Fig. 1) At the time of assessment by a bone health specialist, 60.4% (n = 500) of 828 patients were advised by their specialist to change medication, 33.9% (n = 281) were advised to remain on the same medication, and 2.7% (n = 22) were advised to discontinue medication. Data were not available on the remaining 3.0% (n = 25). The majority of patients who received a recommendation to change were switched to denosumab (n = 418, 83.6%), followed by teriparatide (n = 26, 5.2%) or risedronate (n = 12, 2.4%). There were 21 patients (4.2%) who refused pharmacotherapy.

Fig. 1.

Pharmacotherapy change recommendations among fragility fracture patients who fractured while taking bone active medication, and screened by the fracture screening and prevention program, Ontario

We also examined pharmacotherapy change recommendations stratified by age group, sex and by duration of administration (See Supplementary Table 1). By age group, we found that patients who were ≥ 65 years old were more likely (62.2% vs. 48.7%) to have been recommended a change to their medication compared to those < 65 years. The majority of patients, in both age groups, were switched to denosumab. The results of these bivariable analyses also showed that a longer duration on medication was associated with a higher likelihood of receiving a recommendation to change medication (73.8% of patients on index medication for more than 5 years vs. 48.5% of patients on index medication for less than 1 year).

Baseline characteristics associated with receipt of pharmacotherapy change recommendation

(see Table 2) Multivariable logistic regression analyses (adjusted for sex, age as a categorical variable, index fracture site, falls in the previous year, number of comorbid conditions, type and duration of index medication) demonstrated that older age (66 + vs. 50–65 years) was associated with a higher likelihood of receiving a recommendation to change medication [1.84 (95% CI 1.15–2.95)], while being on denosumab was associated with a lower likelihood of receiving a recommendation to change medication [0.07 (95% CI: 0.05–0.11)].

Table 2.

Results from multivariable logistic regression examining the association between baseline characteristics and pharmacotherapy change at follow-up

Model covariate	Odds ratio [95% CI)
Intercept	1.3097 [0.5537, 3.0982]
Age: 66+ (vs. 50–65)	1.8445 [1.1524, 2.9524]
Gender: female (vs. male)	1.7097 [0.8756, 3.3385]
Index fracture: spine (vs. wrist)	0.9021 [0.4171, 1.9509]
Index fracture: hip (vs. wrist)	1.5904 [0.9715, 2.6036]
Index fracture: other (vs. wrist)	1.0105 [0.6714, 1.5207]
Fallen in the past year: yes (vs. no)	1.2301 [0.8197, 1.8459]
Number of comorbid conditions: 1 (vs. 0)	0.8803 [0.5594, 1.3851]
Number of comorbid conditions: 2 (vs. 0)	1.0166 [0.6307, 1.6388]
Number of comorbid conditions: 3+ (vs. 0)	0.8951 [0.5425, 1.4768]
Baseline pharmacotherapy: denosumab (vs. risedronate)	0.0727 [0.0469, 0.1128]
Baseline pharmacotherapy: alendronate (vs. risedronate)	0.7853 [0.5033, 1.2253]
Duration of index pharmacotherapy: < 1 year (started before index fracture) (vs. 5 + years)	0.5503 [0.2759, 1.0977]
Duration of index pharmacotherapy: 1–5 years (vs. 5 + years)	0.7914 [0.5377, 1.1649]

Discussion

We examined pharmacotherapy change patterns among fragility fracture patients who sustained a fracture while taking bone active medication, and were screened by the FSPP. We also investigated factors that were associated with receiving a recommendation to change the pharmacotherapeutic agent at the time of assessment with a bone health specialist. The majority of our cohort consisted of female patients, over 65 years of age, who were taking antiresorptive bone active medication prior to screening, with first-line bisphosphonates being most commonly reported. This was expected, as studies have shown that oral bisphosphonates are most commonly prescribed for fracture prevention in Ontario [30]. Most patients reported taking their previously prescribed medication for more than one year prior to their recent fracture, and 8.2% of patients had been taking their index medication for less than one year prior to admission. Our multivariable analyses showed that younger patients and those taking denosumab were less likely to receive a recommendation to change their medication after adjusting for measured covariates.

More than half of the cohort was advised to change their previously prescribed pharmacotherapy at the time of their assessment with the bone health specialist. The majority of the patients advised to change medication were switched to denosumab, while a smaller proportion were switched to anabolic therapy. This suggests that in many cases the fracture event prompted re-evaluation and escalation of treatment. Prescribing patterns observed in our study likely reflect multiple considerations, including drug availability, adherence, tolerability, patient preferences, and jurisdiction-specific drug reimbursement criteria, which affect patient access to medications. In Ontario, the cost of first-line oral bisphosphonates is reimbursed through the publicly funded Ontario Drug Benefit (ODB) plan for patients 65 years or older and those receiving various forms of social assistance [31]. Patients younger than 65 years of age pay out of pocket for bone active medications, or through employee or other private health insurance coverage. Denosumab, which is more expensive than first-line oral bisphosphonates, is reimbursed through the ODB plan only under limited-use criteria, which includes fracture while on oral bisphosphonates or contraindications [32, 33]. Zoledronic acid, the other ODB-reimbursed alternative, requires intravenous administration at infusion centres, access to which is challenging for some patients. The small proportion of patients in this study switched to an anabolic agent, which is not typically reimbursed under the ODB plan, likely reflects those with means to pay out of pocket or with private insurance coverage. Most patients already on denosumab were not switched, likely reflecting a combination of clinical judgement and these jurisdiction-specific reimbursement constraints. Prior research shows that Ontario residents with private drug insurance are more likely to use prescription drugs than those without such insurance [34].

Strengths

Our study addresses a notable gap in the literature by examining pharmacotherapy change recommendations by bone health specialists for patients presenting with a fragility fracture to an FLS while taking bone active medication - a clinical scenario for which there is sparse evidence and formal guidance. Another strength is that our study population is also representative of younger patients (50–65 years of age), who tend to be excluded from studies leveraging routinely collected health administrative data in Ontario that typically capture individuals aged 65 or over [33]. The study provided real-world evidence on the prescribing practices of bone health specialists in a large jurisdiction when faced with the scenario of a patient who fractures while taking bone active medication.

Limitations

We acknowledge several unmeasured factors that may influence clinical decisions regarding bone active medication change. Following a new fracture, patients are likely to undergo a comprehensive clinical reassessment that includes evaluation of medication adherence, bone mineral density test results, and investigation for secondary causes of fracture. A substantial proportion of our study cohort had comorbid conditions; however, without further details on the specific diagnoses and corresponding medications, we were unable to comment on the extent to which these factors may have impacted the specialist recommendations. Our findings are limited to patients who are screened by, and subsequently enrol in the FLS; sicker individuals or those with cognitive or physical impairments may be underrepresented. Further evaluation is needed to determine whether the prescribing patterns differ in subgroups that are not adequately represented in the program. Furthermore, bone densitometry results were not available in the FLS data sources, however this information may have been available to the clinical team. As patients were already receiving bone active medication at the time of FLS screening, the assumption was that they had been appropriately diagnosed and treated prior to referral and were likely at high risk for future fractures. In addition, we lacked data on socio-economic status or private drug insurance coverage for patients under 65 years of age, both of which may have influenced treatment recommendations. However, by stratifying treatment change by age groups corresponding to Ontario Drug Benefit eligibility thresholds, we identified differences that are consistent with jurisdiction-specific reimbursement criteria.

Conclusion

Our study provides new insights into pharmacotherapy change patterns prescribed by bone health specialists for patients who sustained a fragility fracture while taking bone active medication and were admitted to a large FLS. The majority of patients received a specialist recommendation to change their index pharmacotherapy, while younger patients and those already receiving denosumab were less likely to be switched. The prescribing patterns are likely to reflect a multitude of considerations, including clinical judgment and jusridiction-specific reimbursement criteria, and provide real-world evidence on the current state of pharmacotherapy management of this high-risk patient population.

Authors’ contributions

HA, JEMS, ERB, RS, VEG: Substantial contributions to the conception and design HA, AY, RS: Acquisition, analysis HA, JEMS, ERB, RS, AY, RJ, JW: Interpretation of data. All authors (HA, AY, RS, RJ, JW, DL, ERB, JEMS): drafted the work or substantively revised it.

Funding

This study was supported by funding from the Ontario Ministry of Health and Long-Term Care (MOHLTC) through the Ontario Osteoporosis Strategy. The views expressed are those of the researchers and do not necessarily reflect those of the MOHLTC.

Data availability

The data that support the findings of this study are available from Osteoporosis Canada but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Osteoporosis Canada.

Declarations

Ethics approval and consent to participate

This study is based on data collected through a quality improvement initiative. Approval for reporting of quality improvement data is not required by the Research Ethics Board at Unity Health Toronto. Approval for usage of these data was obtained from the data custodians, Osteoporosis Canada. This study adhered to the Declaration of Helsinki. Risk assessment was performed on the data, and all values under 5 were suppressed to maintain confidentiality.

Consent for publication

Not applicable.

Competing interests

ERB reports an unrestricted research grant from Amgen Canada Inc.