Changes in femoral bone mineral density after total knee arthroplasty: a systematic review and meta-analysis

Abstract

Background

Bone loss after total knee arthroplasty (TKA) may lead to periprosthetic fractures that are associated with significant costs (morbidity, economic, etc.) and pose a challenge to operative fixation. This meta-analysis quantifies the change in bone mineral density (BMD) of the distal femur after primary TKA.

Methods

A systematic review of six databases was performed by two independent reviewers. Studies that reported bone density after knee arthroplasty were identified and inclusion/exclusion criteria was applied. Data were extracted and analyzed using the Comprehensive Meta-Analysis Software.

Results

Fourteen studies were included in the analysis. The average decrease in BMD was 0.09 [0.05, 0.13], 0.14 [0.08, 0.20], 0.16 [0.10, 0.23], and 0.16 [0.12, 0.20] g/cm² at 3, 6, 12, and 24 months, respectively, corresponding to a 9.3%, 13.2%, 15.8%, and 15.4% BMD loss. A high degree of heterogeneity existed between the studies (I² > 90% at most time points).

Conclusion

In summary, there is a rapid and significant 15% decrease in BMD in the first 6 months after TKA that is sustained to 24 months. Better understanding regarding how perioperative optimization of bone health may affect BMD loss and the incidence of periprosthetic fracture is essential.

Level of evidence

Therapeutic Level II.

Appendix

Summary of investigations

Fourteen studies were included in the meta-analysis. Two of the first studies prospectively measuring BMD after TKA were by Petersen et al. [10, 15] using DPA. Petersen et al. [15] 1995 measured BMD in three areas of the distal femur (anterior to the fixation lugs, proximally to the lugs and posteriorly to the lugs) 2 years after TKA in 8 patients. The 1996 Petersen et al. [10] study measured and compared the BMD at similar regions (behind the anterior flange of femoral component and above the fixation lugs) for 29 patients at 1 year after implantation with different femoral components in hopes to reduce the previously observed loss in BMD. Another earlier study published, Liu et al. [16], was a case-control study measuring mostly supracondylar BMD bilaterally in 48 females, comparing two different implants to age-matched controls.

Several studies focused on quantifying supracondylar BMD changes. Gazdzik et al. [17] is a prospective longitudinal study that reported the BMD of an area proximal to the superior border of the femoral components in 106 postmenopausal females. In a similarly designed study, Mau-Moeller et al. [18] measured BMD three months postoperatively at the same femoral ROIs in 23 patients (65% male). Windisch et al. [19] also measured supracondylar femur BMD prospectively in 50 patients at several time points in the first postoperative year. In the Minoda et al. [20] study, twenty-eight patients with fixed-bearing TKA were matched with 28 patients with mobile-bearing TKAs, and the BMD of the anterior, central and posterior distal femur ROIs (spanning metaphyseal and intracondylar regions) were measured for 24 months. Van Loon et al. [21] studied the 1-year postoperative changes in BMD of the femoral neck, lumbar spine, and distal femur (distal anterior area of the femur behind the anterior flange and supracondylar area just superior to the anterior flange of the femoral component).

Soininvaara et al. [22, 23] developed standardized ROIs of the distal femur (anterior metaphyseal, central metaphyseal, posterior metaphyseal, total metaphyseal and diaphyseal). One of their studies in 2004 measured the changes in BMD of the operative leg in 69 patients over the course of a year [22]. Karbowski et al. [24] conducted a smaller longitudinal study measuring BMD of the same ROIs in the operative knee of 12 patients. Soininvaara et al. [23] 2008 studied the changes in BMD detected by DXA along with attempting to correlate single photon emission computed tomography measurements in the standard ROIs prospectively 2 years after TKA in 16 patients. Järvenpää et al. [25] compared the BMD at these standard ROIs defined by Soininvaara between 61 obese and nonobese patients over a period of 7 years. Shibuki et al. [26] retrospectively studied 22 patients collecting BMD at Soininvaara’s recommended ROIs over a 2-year postoperative period. In 2006, Saari et al. [27] reported 83 patients randomized into four groups based on varus/valgus alignment: a flat or a concave tibial plateau with the posterior cruciate ligament (PCL) retained and a concave or a posterior-stabilized tibial component with the PCL resected. They measured three intracondylar ROIs over 5 years.