Prediction of Fragility Fractures and Mortality in a Cohort of Geriatric Patients

Background

Risk factors of refracture after fragility fractures include osteoporosis, female gender and advanced age among others. We hypothesized that the assessment of functionality, muscle health and nutrition status contribute to the risk prediction for further fractures and death.

Methods

We assessed 334 patients admitted to the department of acute geriatrics for sociodemographic data, bone fragility, selected laboratory tests, body composition and data on functionality using the comprehensive geriatric assessment. Patients had follow-ups until the occurrence of further fractures or death. Dual-energy X-ray absorptiometry and pulse echo measurements were performed to assess bone mineral density. Fracture risk was assessed using the FRAX score and muscle strength according to published guidelines on sarcopenia.

Results

The mean age was 81 years (70–95), and 82.3% (275/334) were women. An incidence of 10.4% (35/334) new fragility fractures was observed within 24 months, and the mortality rate was 12.2% (41/334). A significantly higher rate of further fractures was associated with lower BMI (body mass index) (HR 0.925, CI 0.872–0.98; p = 0.009), lower parathyroid hormone levels (HR 0.986, CI 0.973–0.998; p = 0.026) and with the diagnosis of osteoporosis (HR 2.546, CI 1.192–5.438; p = 0.016). No significant associations were present in patients with previous fractures, with higher age, higher FRAX scores, sarcopenia, in women, sarcopenic obesity, frail patients, lower grip strength, lower walking speed, lower Barthel index or lower DI (density index) values. The predictive power for further fractures was 10.7% higher adding osteosarcopenia, BMI and parathyroid hormone levels to standard assessment parameters osteoporosis, age and the status of previous fractures. Mortality was significantly higher with advanced age (HR 1.101, CI 1.052–1.151; p < 0.001), in men (HR 6.464, CI 3.141–13.305; p < 0.001), in smokers (p = 0.002), higher FRAX score (HR 1.039, CI 1.009–1.070; p = 0.010), lower renal function (HR 0.987, CI 0.976–0.997; p = 0.010), lower Tinetti test scores (HR 0.943, CI 0.900–0.987; p = 0.012), lower walking speed (HR 0.084, CI 0.018–0.382; p = 0.001), lower hand grip (HR 0.876, CI 0.836–0.919; p < 0.001) and lower Barthel index scores (HR 0.984, CI 0.971–0.997; p = 0.015).

Conclusions

In a cohort of geriatric patients, the addition of BMI, low parathyroid hormone levels and osteosarcopenia increases the predictive power for further fractures by 10.7%. These parameters are a valuable addition to the standard assessment parameters age and history of sustained fractures. Mortality is partly associated with potentially treatable functional parameters.