The current study demonstrated comparable subsidence patterns of the AMIStem to those of previously published short stems (15, 18, 21, 34) with the highest subsidence in the first 3 months. At the latest follow-up, the average stem subsidence was 1.98 ± 1.20 mm, with 48% of the implants demonstrating subsidence of > 2 mm. Periprosthetic radiolucency of > 2 mm was found in 26% of the implants in Zone 1, and in 9% in Zone 7, respectively. Furthermore, no patient-related or implant-related factors were found to have a statistically significant influence on stem subsidence.
The examined short stem demonstrated almost 50% of the overall subsidence within the first 3 months and then slowed down markedly. However, there was some subsidence up to the last follow-up, which is in accordance with recent findings of Schaer et al. (34), who studied the Optimys short stem (Mathys, Bettlach, Switzerland) and Thalmann et al. (14) who studied the Fitmore short stem (Zimmer Inc., Warsaw, Indiana, USA). In contrast, most previously published data on cementless short stems found stabilization of subsidence during a shorter follow-up of 3 months(21, 22), 6 months(20), 12 months(19), or 24 months(15). Applying radiostereometric analysis (RSA), Acklin et al. observed an average subsidence of 0.39 mm at 3 months after implantation of a Fitmore stem (Zimmer Inc., Warsaw, Indiana, USA) with no further distal subsidence until a 2-year follow-up(21). Freitag et al. using EBRA-FCA observed an average subsidence of 1.1 mm (range: -5 mm to 1.5 mm) at a 5-year follow-up with stabilization from the 2-year mark with the same implant (15). For the Optimys short stem (Mathys, Bettlach, Switzerland), subsidence of 0.96 +/- 0.76 mm at 3 months and 2.04 +/- 1.42 mm at 5 years has recently been reported(34). Kutzner et al. published a mean axial subsidence of 0.55 mm (SD 0.78 mm) at 6 weeks and 1.43 mm (SD 1.45 mm) at final follow-up at 2 years in the same stem design(19), while in another study, the same author reports subsidence of > 2 mm in 15.7% of implants, which subsequently stabilized (20). However, they did not use EBRA-FCA for measurements. Brinkmann et al. analyzed subsidence of the Nanos stem (Smith & Nephew plc, London, UK) and the Metha stem (Aesculap AG, Tuttlingen, Germany) during a 1-year follow-up and reported an average distal subsidence of 2.04 +/- 2.65 mm and 1.96 +/- 2.37 mm, respectively(18).
When compared to conventional stems, short stems are reported to subside slightly more. Clauss et al. reported mean subsidence of 0.66 mm at 5 year follow-up in the twinSys® stem (Mathys Ltd., Bettlach, Switzerland) with 9.8% of the implants showing subsidence > 2 mm using EBRA-FCA(12). Campbell et al. used RSA to evaluate stem subsidence of a corail stem (Corail; Depuy Orthopaedics Inc., Warsaw, Indiana, USA) to find an average subsidence of 0.58 mm (range − 0.23 to 3.71 mm) 2 years after surgery(35). Some authors compared the subsidence patterns of short stems to conventional stem designs. In a randomized controlled trial, Ferguson et al. showed substantially lower subsidence (0.36 +/- 0.38 mm) of the Meta Fix conventional stem (Corin Group, Cirencester Gloucestershire, UK) compared to the MiniHip (Corin Group, Cirencester Gloucestershire, UK) short stem (0.62 +/- 0.56 mm) at 2 years (16). McCalden et al. found a higher, yet not significant, subsidence of the SMF short stem (Smith & Nephew plc) compared to the Synergy conventional stem (Smith & Nephew plc) (0.94 +/- 1.74 mm versus 0.32 +/- 0.45 mm) at 2-years follow-up using RSA (17).
Several authors focused on defining a threshold value of early subsidence for the prediction of aseptic failure. Freeman et al. described a threshold subsidence of 1.2 mm per year during the first 2 years for the prediction of aseptic failure with a specificity of 86% and a sensitivity of 78%(26). Using RSA, Kärrholm et al. reported a risk of over 50% of aseptic loosening, if subsidence of over 1.2 mm occurred within the first 2 years after surgery. If a subsidence of more than 2.4 mm was reached, the risk increased to 95%(24). In a similar study, Krismer et al. investigated subsidence of the femoral stem using EBRA-FCA. Early aseptic loosening could be predicted with a sensitivity of 69% and a specificity of 80%, if subsidence exceeded 1.5 mm during the first two years(36). However, none of these studies examined cementless short femoral stems. Studying a proximally-fixed Vision 2000 stem (Depuy Orthopaedics Inc, Warsaw, Indiana, USA), Stihsen et al. described subsidence of > 2 mm in 19% out of 102 implants after two years and found a highly significant correlation of subsidence > 2 mm at two years and subsequent aseptic loosening(37). On the other hand, studying the metaphyseal-anchored Fitmore hip system (Fitmore®, Zimmer Inc., Warsaw, Indiana, USA), Gustke et al. described subsidence of more than 2 mm on plain radiographs in 34% of 100 examined THA after a mean follow-up of 1.3 years. However, none of these implants had to be revised during this short follow-up(38). In our study, 15% of the implants showed axial subsidence > 2 mm after the 1-year follow-up and 48% after a mean follow-up of 64 months with only one case of aseptic stem loosening during the observation period. Considering this, the above-mentioned threshold values might not be applicable for proximally-fixed short femoral stem designs. However, due to the limited sample size as well as the average follow-up time of 64 month, the current study might be underpowered to detect a potential correlation of subsidence with aseptic loosening, where rates as low as 0.4% at 10 years in the DAA have been reported(39).
The present study should be interpreted in light of its potential limitations. First, the EBRA-FCA method was used instead of RSA, which is currently considered the gold standard for analyzing distal stem subsidence. However, the widely established, computer-assisted EBRA-FCA system is able to detect subsidence of more than 1 mm with a specificity of 100% and a sensitivity of 78%(31) and is therefore considered a reproducible and accurate tool for evaluation of distal femoral stem subsidence. Second, we only evaluated axial subsidence of the stem while tilt and rotation were not evaluated. Third, the small size of our cohort limits the power of our study, especially regarding aseptic loosening, where rates as low as 0.4% at 10 years in DAA have been reported(39). Fourth, our study is prone to some attrition bias with a high rate of rejected X-rays by EBRA-FCA and incomplete radiographic datasets resulting in the inclusion of 62.3% of initially enrolled patients. Finally, only the AMIStem femoral component was investigated in our study. Although the EBRA-FCA software was used by various authors to measure subsidence of different femoral implants, our findings might not apply to other stem designs.