Introduction: Radial Impaction Grafting (RIG) has been proposed as a method for improving durability and reliability of cementing a collared femoral component in radial compressed bonegraft. In a prospective, baseline-control, multicenter clinical study the clinical and quality of life scores, survival rate, complications and radiographic findings were evaluated. The main objective of this study is to establish and indentify factors that contribute to the fracture rate. We tried to establish a model that represents the simplest combination of factors that predict femur fracture following the RIG technique.

Patients and Methods: Eighty-eight patients enrolled the study: 59 women and 29 men with average age of 74 (38–90) years. The average follow-up was 4 years (2–7.2). The results of 81 patients were evaluated. The Endo-classification (EC) was used for the determination of the pre-operative femoral bone loss; type 1 was noticed in 25 cases, type 2 in 44, type 3 in 17 and type 4 in 1 case.

At 70 patients the femur was augmented (mesh, cerclage(s) and/or plate). A short stem (115–135mm) was used in 48% and a long stem (165–225mm) in 52% of the cases.

Results: The average total HHS increased from a pre-operative score of 44 (median 44) to 91 (median 94) at > 5 years fu and the HHS-pain score from 17 (median 20) respectively to 43 (median 44). The average VAS for pain at > 5 years fu was 2.5 (0–100 scale). Ninety-three percent of the patients were very-extremely satisfied with the result.

Postoperative the femoral stem had a varus position (6 –10°) in 4 cases. In 6 cases the stem was in posterior and 3 cases in anterior position.

Progressive subsidence was demonstrated in 1 patient increasing to 5 mm at 1 year fu. Three patients showed subsidence of 3 mm at 3 months, but were stable afterwards. Slightly increasing radiolucent lines (> 2 mm) were noticed in 2 patients in subcortical areas, but never any signs of osteolysis.

Intra- and postoperative femoral fractures occurred in 12 patients (13,6%). There were 3 dislocations and 1 loosening of a trochanteric osteotomy. Despite the fractures there was no loosening of a stem during follow-up period.

After bivariate multilevel analysis, conducted to determine a relationship between the fracture complication and all other variables, a multivariate model was developed of the most significant variables to determine the predictor factors for femoral fracture. Only the Endo-classification and age are predictors for a fracture following a RIG procedure (p-value 0.003).

Each higher EC type increases the probability of a fracture after RIG 2.01 times and each added year of life 1.07 times. So the risk of getting a femoral fracture increases exponentially with age from 1.7% at 55 years to 15.5% at 90 years in EC type 1. The fracture risk in EC type 4 increases from 4.7% at 40 years to almost 60% at 90 years.