Arthritis & Rheumatism, Volume 60,
October 2009 Abstract Supplement

The 2009 ACR/ARHP Annual Scientific Meeting
Philadelphia October 16-21, 2009.

Association Between Medication Adherence and Risk of Fracture Among Women Initiating Osteoporosis Therapy

Halpern1,  Rachel, Becker1,  Laura K., Iqbal2,  Sheikh Usman, Badamgarav2,  Enkhe, Macarios2,  David

i3 Innovus, Eden Prairie, MN,
Amgen, Thousand Oaks, CA


Adherence to osteoporosis medication is frequently low and may result in increased risk of non-traumatic fracture. This study examined the relationship between adherence to osteoporosis medication and risk of fracture among women initiating osteoporosis therapy.


This retrospective analysis used medical and pharmacy claims from a large US health plan. Subjects were female commercial health plan enrollees >=45 years old who initiated treatment on alendronate, risedronate, teriparatide, ibandronate, or raloxifene from 6/30/2002 to 12/31/2006. The date of the first medication claim was the index date. Subjects were continuously enrolled for 1 year pre-index (baseline) and >=180 days post-index. Non-traumatic fractures were identified with diagnosis codes and ICD-9 or CPT procedure codes. "Outcome" fractures were those that occurred >=90 days after index date. Early fractures occurred within 90 days after index date. Prevalent factures occurred during baseline. Adherence was measured with medication possession ratio (MPR), the number of days with osteoporosis medication divided by number of days observed; adherence categories were "high" (MPR>=0.8), "medium" (0.5<=MPR<0.8), and "low" (MPR<0.5). Control variables included: age; prevalent fracture; early fracture; baseline steroid use; Charlson comorbidity index score; fracture risk-related comorbidities, and geographic region. Time to fracture <=720 days after index date was modeled with Cox regression with time-dependent adherence.


The study sample included 88,122 women with mean age of 58.6 (±8.3) years; 5,404 (6.13%) subjects had outcome fractures <=720 days after the index date. Bivariate comparison adjusted for time-dependent adherence showed that low adherence was associated with an 11.2% higher risk of fracture compared with high adherence (p=0.0006). When covariates were included, patients with low adherence had 8.3% higher risk of fracture relative to patients with high adherence (p=0.0110), as shown in Table 1. Other covariates associated with fracture included early fracture (hazard ratio (HR) =5.939, 95% confidence interval (CI) =5.453–6.467), prevalent fracture (HR=2.020 CI=1.865–2.188), and baseline steroid use (HR=1.088, CI=1.017–1.163).

Table 1. Role of Adherence in Fracture Risk: Regression Results

Regression (Reference: High Adherence)Hazard Ratio95% Confidence Intervalp-value
Bivariate comparison   
  Low adherence (MPR <0.5)1.112(1.047–1.182)0.0006
  Medium adherence (0.5 <0.8)1.073(0.999–1.153)0.0544
Adjusted for all covariates   
  Low adherence (MPR <0.5)1.083(1.019–1.152)0.0110
  Medium adherence (0.5 <0.8)1.052(0.979–1.131)0.1651


The results indicate that low adherence to osteoporosis therapy is associated with increased risk of fracture. Research to examine factors related to medication non-adherence could inform treatment interventions and may subsequently impact rates of fracture.

To cite this abstract, please use the following information:
Halpern, Rachel, Becker, Laura K., Iqbal, Sheikh Usman, Badamgarav, Enkhe, Macarios, David; Association Between Medication Adherence and Risk of Fracture Among Women Initiating Osteoporosis Therapy [abstract]. Arthritis Rheum 2009;60 Suppl 10 :860
DOI: 10.1002/art.25940

Abstract Supplement

Meeting Menu