Evaluation of FRAX in patients with periprosthetic fractures following primary total hip and knee arthroplasty

The fracture risk assessment tool (FRAX) is a tool which calculates an individual 10-year fracture risk based on epidemiological data in patients with a risk of osteporosis. The aim of this study was to evaluate the value of FRAX to estimate the risk of postoperative periprosthetic fractures (PPF) in patients following with total hip and knee arthroplasty. 167 patients (137 periprosthetic fractures in total hip arthroplasty and 30 periprosthetic fractures in total knee arthroplasty) were included in this study. Patients’ data was retrieved retrospectively. In each patient the 10-year probability of a major osteoporotic fracture (MOF) and an osteoporotic hip fracture (HF) was calculated using FRAX. According to the NOGG guideline 57% of total hip arthroplasty (THA) patients and 43.3% of total knee arthroplasty (TKA) patients were in need of osteoporosis treatment, whereas only 8% and 7% received an adequate one respectively. 56% of the patients with PPF after THA and 57% of the patients with PPF after TKA reported about a previous fracture. Significant associations between the 10-year probability of a MOF and HF calculated by FRAX and PPF in THA and TKA were seen. The results of the present study show that FRAX might have the potential to estimate the PPF in patients following THA and TKA. FRAX should be calculated before and after THA or TKA in order to assess the risk and counsel patients. The data show a clear undertreatment of patients with PPF in respect to osteoporosis.


Results
167 patients (137 periprosthetic fractures in total hip arthroplasty and 30 periprosthetic fractures in total knee arthroplasty) were included in this study (see Table 1). The indication for primary THA was osteoarthritis in 112 (81%) %, hip fracture in 11 (8%), osteonecrosis of the hip in 9 (6.5%) and developmental dysplasia of the hip in 5 (3.6%) patients respectively. The indication for primary TKA was osteoarthritis in 30 patients (100%). Mean age of patients was 81.0 ± 10.7 years in PPF in THA and 76.5 ± 10.2 years in PPF in TKA respectively. Females represented 74% of the PPF in THA and 90% PPF in TKA. Mean height of patients was 168.0 ± 8.3 cm in PPF in THA and 165.0 ± 7.5 cm in PPF in TKA respectively. Mean weight of patients was 68.0 ± 13.2 kg in PPF in THA and 76.0 ± 14.1 kg in PPF in TKA respectively. Mean BMI of patients was 24.0 ± 3.6 kg/m 2 in PPF in THA and 27.7 ± 4.7 kg/m 2 in PPF in TKA respectively. In 106 patients (88 periprosthetic fractures in total hip arthroplasty and 18 periprosthetic fractures in total knee arthroplasty) the implant survival could be calculated. Results of the correlation analyses can be seen in Table 2 (Table 2).
Periprosthetic fractures in THA. The most frequent fracture was a Vancouver type B2 periprosthetic fracture of the femur in 57%, followed by a Vancouver B1 periprosthetic fracture of the femur in 20% and Vancouver C periprosthetic fracture of the femur in 11%. www.nature.com/scientificreports/ In 56% of the patients with periprosthetic fracture in THA there was a previous fracture in the patient's history, whereas secondary osteoporosis was present in 26% and rheumatoid arthritis in 10%.
According to the NOGG guideline 88 of 137 (57%) were in need of treatment according to the FRAX values in THA patients. In the group of patients considered for treatment (very high risk or high risk) according to the NOGG guideline just 8% (7 out of 88) received adequate treatment pharmacologic for osteoporosis.
At the time of initial surgery the 10-year probability of a MOF was at a mean of 19.31% (± 14.5) and of a HF 9,8% (± 10.0).
At Periprosthetic fractures in TKA. The most frequent fracture was a Su type II periprosthetic fracture of the distal femur in 83%, followed by Felix Type I fracture at the proximal tibia (7%) and a Su type II periprosthetic fracture in 4%. The majority of fractures occurred in female patients (27/30-90%). Therefore statistical analyses were just done in female patients. In 57% of the patients with periprosthetic fracture in TKA there was a previous fracture in the patient's history. The second most frequent cause/risk factor was secondary osteoporosis (27%) and rheumatoid arthritis (10%). According to the NOGG guideline 43.3% (13/30) of the patients would require medical treatment (very high risk or high risk), whereas in fact only 1 (7%) received it before fracture.
At the time of initial surgery the 10-year probability of a MOF was at a mean of 18.0% (± 13.4) and of a HF 7.8% (± 8.9).
At the time of postoperative periprosthetic fracture the 10-year probability of a MOF was at a mean of 22.9% (± 13.1) and of a HF 11.3% (± 9.0).
In Results of the correlation analyses can be seen in Table 1 (Table 1), whereas Fig. 1 shows distribution of patients in respect to 10-year probability of MOF according to the NOGG guidelines (Fig. 1).

Discussion
FRAX is a tool widely used to screen patients a risk of osteoporosis and osteoporotic fractures. In this study, FRAX has been evaluated among 167 patients with PPF following THA or TKA. Results show that the majority of patients who suffered PPF had an increased risk (10-year probability) of MOFs and HFs according to FRAX. In this respect in more than half of the patients, medical treatment would be indicated, which is in contrast to reality in which less then 10% received an adequate one. This is in accordance with a recent study of the incidence of osteoporosis among TKA patients in which osteoporosis was present in 36% of the cases 8 . Furthermore, the same group invested osteoporosis specific medical treatment in TJA patients who had osteoporosis and found that just about 20% received adequate treatment 10 . Similar results were seen by Ha and Park who found an osteoporosis incidence of 50% among TKA patients of which only 15% received medical treatment for osteoporosis 11 . In a study of 268 elderly THA patients osteoporosis was present in 18% and osteopenia in 41% 9 . This points towards the need of additional osteoporosis screening not just at the time of implantation, but also on a regular base in the clinical follow-up to potentially prevent complications such as PPF. Furthermore, if indicated, medical treatment needs to be applied and patients counselled in this respect. In addition, patients should get a multidimensional treatment including fall prevention programs. It has been shown that lower scores of the 36-Item Short Form Survey (SF-36) subdomains of physical functioning and vitality are predictors for PPFs after primary TKA 22 .
The results of these study show significant associations between the 10-year probability of a MOF and HF calculated by FRAX and PPF in THA and TKA. Age and previous fracture, both components of the FRAX tool, seem www.nature.com/scientificreports/ to be major factors attributed to PPF risk which was seen in a high percentage among the studied patients and in previous publications 1,4,7,23 . Various studies showed that increased patient age is associated with the incidence of PPF 6,7 Previous fractures are associated with an increased PPF risk 23 . As in osteoporosis where previous fractures are associated with subsequent fractures, a previous osteoporotic fracture seems to be an indicator of PPF 24 .
There are limitations to this study which includes the retrospective nature of data collection which causes missing data. As a such implant survival could not be calculated in all of the patients. Furthermore, important bone specific data is missing at all as no such data was available at the time of implantation. Furthermore, the study was limited to data that was retrieved from a documentation system, so additional information such as bone metabolic parameters, BMD or microCT are not available. The patient collective is quite heterogenous and includes to a majority THA patients, but only 30 TKA patients.
The results of this study can be seen as a preliminary report, however FRAX can easily be used in a clinical setting or even by patients themselves to estimate their current situation in respect to bone health. It was seen that the majority of patients that experienced PPFs are in need of osteoporosis specific treatment and have an increased likelihood of fractures. Further studies with an increased number of participants are needed to define the definitive value of FRAX to estimate PPF risk. At the moment however, it seems to be a good and easily accessible tool to be applied to define a status quo of bone health in patients before and following THA or TKA.

Conclusions
The results of the present study show that FRAX might have the potential to be used as a tool to estimate individual PPF risk in patients following THA and TKA. The data shows that there is a association of the 10-year probability of MOF and HF and PPF. Furthermore, a clear undertreatment of patients with PPF in respect to osteoporosis has been seen. Further research is necessary to support the results of the present study.

Data availability
The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.