Validating the FRAC-Stroke score: targeting those at highest risk of fracture

Fracture risk is increased after stroke, but there is a lack of specific, effective tools to screen and prevent minimal trauma fractures (MTF) post-stroke. While the FRAC-Stroke score, developed to predict fracture risk post-stroke, shows promising performance, it has not yet been widely adopted or validated in large cohorts outside of Canada.

We aimed to: 1) evaluate the reliability and validity of the FRAC-Stroke score in predicting MTF; and 2) determine the optimal score to classify patients at highest fracture risk following ischaemic stroke.

A retrospective cohort study was undertaken using data from the PRECISE study, which included person-level linked administrative data from the Australian Stroke Clinical Registry (AuSCR, 2014-2016). The FRAC-Stroke score at hospital discharge was determined for survivors of ischemic stroke, aged >50 years, using coded comorbidities derived in the prior 5-year period and modified Rankin Scale. Fine-Gray models were built to evaluate associations of the FRAC-Stroke score with MTF within 12 months post-discharge, accounting for death as the competing risk.

Among 4545 adults, the median FRAC-Stroke score was 8 (interquartile range 5-11) and 118 patients (2.6%) sustained a MTF within 12 months post-discharge. The optimal FRAC-Stroke score cut-point to stratify MTF risk was 11.5, with a c-statistic of 0.70. Participants with a FRAC-Stroke score >12 (vs <12) had a 5.5-fold increased risk of MTF (confidence interval 3.8-7.9, p<0.01). When stratified by quintiles of FRAC-Stroke score, those in quintile 5 (FRAC-Stroke score >14) had a 15.2-fold increased risk of MTF than those in quintile 1 (FRAC-Stroke score <3; confidence interval 6.6-35.2, p<0.01).

The FRAC-Stroke score reliably identifies increased fracture risk in the Australian ischemic stroke population, with a score of 12+ able to discriminate those at high risk of fracture.