Research Article: Stable fixation of an osseointegated implant system for above-the-knee amputees

Date Published: April 24, 2012

Publisher: Informa Healthcare

Author(s): Audrey Nebergall, Charles Bragdon, Anne Antonellis, Johan Kärrholm, Rickard Brånemark, Henrik Malchau.


Rehabilitation of patients with transfemoral amputations is particularly difficult due to problems in using standard socket prostheses. We wanted to assess long-term fixation of the osseointegrated implant system (OPRA) using radiostereometric analysis (RSA) and periprosthetic bone remodeling.

51 patients with transfemoral amputations (55 implants) were enrolled in an RSA study. RSA and plain radiographs were scheduled at 6 months and at 1, 2, 5, 7, and 10 years after surgery. RSA films were analyzed using UmRSA software. Plain radiographs were graded for bone resorption, cancellization, cortical thinning, and trabecular streaming or buttressing in specifically defined zones around the implant.

At 5 years, the median (SE) migration of the implant was –0.02 (0.06) mm distally. The rotational movement was 0.42 (0.32) degrees around the longitudinal axis. There was no statistically significant difference in median rotation or migration at any follow-up time. Cancellization of the cortex (plain radiographic grading) appeared in at least 1 zone in over half of the patients at 2 years. However, the prevalence of cancellization had decreased by the 5-year follow-up.

The RSA analysis for the OPRA system indicated stable fixation of the implant. The periprosthetic bone remodeling showed similarities with changes seen around uncemented hip stems. The OPRA system is a new and promising approach for addressing the challenges faced by patients with transfemoral amputations.

Partial Text

51 patients with transfemoral amputations (mean age at implant surgery: 45 (21–65) years; 28 males) were enrolled in the prospective OPRA (Osseointegrated Prostheses for the Rehabilitation of Amputees) study. There were 45 unilateral patients and 6 bilateral patients. 2 bilateral patients were only treated on one side in this study; thus, 55 prosthetic systems were enrolled and followed in this study. Most patients had had a traumatic injury or a tumor.

All patients in the OPRA study passed the 2-year follow-up after S2 and clinical details will be presented in a separate paper. Of the 55 implants in the study, the following numbers of implants were analyzed with RSA at each follow-up interval: 47 implants at 6 months, 42 implants at 1 year, 40 implants at 2 years, 15 implants at 5 years, 12 implants at 7 years, and 3 implants at 10 years. Due to the low number of implants with a 10-year follow-up, the 10-year data are not reported at this time. 8 patients were excluded from the study and 4 had their implants removed; 3 implants were removed due to loosening and 1 was removed due to infection. 1 patient did not have his implant removed but was excluded due to complications that were not from surgery, 1 patient was lost to follow-up, and 2 patients died. Only 2 of the 4 failed implants had RSA films analyzed at 1 and 2 years of follow-up. 1 patient did not receive any tantalum beads and did not therefore participate in the RSA part of the study (Table 1).

This is the first prospective RSA study to evaluate osseointegrated prostheses for use in transfemoral amputee patients. It is widely accepted that loosening of an orthopedic implant is one of the most common reasons for failure of the device. However, the RSA analysis for the OPRA system indicated that there was no statistically significant migration of the osseointegrated implant to this point. Graphically, some implants appeared to be rotating more than in the rest of the cohort. The 2 implants that showed the greatest degree of rotation (almost 3 degrees) had very high condition numbers in the RSA analysis. This “motion” could most likely be attributed to unreliable RSA analyses, since the median condition number for these patients was much higher (226) than the median condition number of the rest of the cohort (53). These high rotations occurred around the longitudinal axis. For geometric reasons, these have poorer resolution than rotations about the transverse and sagittal axes with this type of implant and radiographic set-up.