Abstract: TH-PO783
FRAX-HD: A Two-Year Incident Hip Fracture Risk Assessment Tool for Japanese Hemodialysis Patients
Session Information
- Dialysis: Epidemiology, Outcomes, Clinical Trials - Non-Cardiovascular - I
November 02, 2017 | Location: Hall H, Morial Convention Center
Abstract Time: 10:00 AM - 10:00 AM
Category: Dialysis
- 607 Dialysis: Epidemiology, Outcomes, Clinical Trials - Non-Cardiovascular
Authors
- Fujii, Naohiko, Hyogo Prefectural Nishinomiya Hospital, Nishinomiya, Japan
- Hamano, Takayuki, Osaka University Graduate School of Medicine, Suita, OSAKA-FU, Japan
- Masakane, Ikuto, Honcho-Yabuki Clinic, Yamagata, Japan
Group or Team Name
- Committee of Japanese Renal Data Registry, Japanese Society for Dialysis Therapy
Background
Dialysis patients are at 5-to-6-fold higher risk of hip fracture than general population. FRAX® is a useful tool for assessing future fracture risk; however, it is not optimized for hemodialysis (HD) patients. Our aim was to create a 2-year risk assessment tool for hip fracture among Japanese HD patients (FRAX_HD).
Methods
We extracted patients on facility-based HD, aged 20-100, with ≥1 yr of vintage at the end of 2007 using the Japanese nation-wide dialysis registry (JRDR-08102). Hip fractures were detected annually up to 2 years by questionnaire. We excluded patients who later received cinacalcet during the follow-up period, and randomly-selected 60%, 20%, and 20% of the eligible subjects to generate TRAIN, VALIDATE, and TEST sets, respectively. Multivariable Poisson regression analyses with interaction terms were performed to create prediction models, and validation was done later.
Results
TRAIN, VALIDATE, and TEST sets included 73999, 24667, and 24665 patients and 1574, 525, and 524 hip fractures, respectively. The subjects, aged 65.2 ± 12.3, included 39% of female and 34% of diabetes, with median vintage of 5.6 yrs. In TRAIN set, there were significant interactions between sex and age, sex and prior PTX/PEIT, and diabetes and vintage. Incidence rate ratio (IRR) elevated with age more prominently in female than in male (IRR 5.62 [95%CI: 3.32, 7.91]; females aged ≥80 vs. males aged <50). Prior history of PTX/PEIT was associated with lower risk only in female (IRR 0.58 [0.36, 0.81]). The model including these interactions demonstrated best performance and was selected as final model, which also achieved similar results in TEST set (AUC 0.739).
Conclusion
We successfully generated a 2-year risk assessment tool for hip fracture among HD patients. Further refinement and sensitivity analyses are required before its clinical application.