Introduction

Short-stem hip implants have been introduced as a bone preserving option for total hip arthroplasty (THA) in the younger and more active population [33]. However, there are debates whether the long-term survival rates of these implants meet current standards and if short-stem implants are indicated for all pathologies, e.g. osteonecrosis of the femoral head [8, 27]. Implant longevity in THA relies upon good initial fixation and long-term axial and rotational stability [21, 31]. In modern, uncemented THA, primary stability is frequently achieved by creating a ‘press fit’ between the implant and the surrounding bone [31], which is crucial to the rapid osseointegration of the implant [1]. With a firm press fit, the subsidence is expected to be minimal in the first postoperative months. However, migration patterns differ depending on the fixation philosophy and design of the implant [20], and few publications have thus far reported on the stability of short-stem implants [23, 33].

Radiostereometric analysis (RSA) is the gold-standard measure of migration [15], and has evolved from the marking of implants to a ‘model-based’ technique [32]. The importance of RSA lies not only in its accuracy, which allows for precise results in only a small number of patients [25], but also in its predictive value [15]. Several THA studies, involving different cemented component designs have confirmed that early migration, as measured by RSA, predicts clinically relevant loosening [6, 18, 29]. For cementless femoral stems, the situation is somewhat more complex, as well-functioning stems may show initial postoperative migration [6], which can be characterized as settling.

We hypothesized that early migration of the Fitmore Hip Stem would be modest, with no substantial migration after initial settling of the prosthesis. The aim of this 2 year, prospective, single-center study was therefore to examine migration of the stem in THA using model-based RSA, and thus predict the implant’s long-term survival. In addition, we evaluated the stem’s clinical performance using standard measures.

Patients and methods

Study settings and participants

34 patients with radiological signs of osteoarthritis and severe hip pain and disability requiring THA were consecutively enrolled between March 2011 and September 2011. Informed consent was obtained from all patients prior to any procedures, and this prospective study was approved by the cantonal ethic committee of Zurich (KEK-ZH-Nr. 2010-0272/4). Baseline characteristics were collected for all patients.

The inclusion criteria were patients who were able to understand the follow-up program, patients who have given written consent to take part in the study, a pre-operative level of pain and function requiring a conventional joint replacement, between 18 and 70 years of age, good general health for age (e.g. no ongoing cancer, no chronic diseases limiting life expectancy), male and female. Exclusion criteria included patients unwilling to sign consent or to comply with the follow-up program, pregnancy, active infection, Parkinson’s disease and revision arthroplasty.

Implant and surgery

All patients received a Fitmore Hip Stem alongside an uncemented Fitmore Shell acetabular component, which combined a polyethylene liner with a ceramic or metal head. The Fitmore Hip Stem (Zimmer, Winterthur, Switzerland) is a curved, uncemented triple-tapered stem made from titanium alloy. Proximally, it has a titanium vacuum plasma spray coating, while the distal part is rough-blasted. The anchorage concept is based on apposition to the calcar supported by the medial curvature, which is available in three different options, and contact with the lateral cortex via axial loads. Its trapezoidal cross-section and support on the medial curvature underpins the primary stability, which is enhanced by a slightly oversized titanium coating. The triple-tapered stem design should achieve rotational stability.

Three surgeons performed the operations using either a minimally invasive anterolateral approach (N = 30), as described by Röttinger [28], or a standard lateral approach (N = 4), as set out by Hardinge [2] depending with which they felt more comfortable.

Follow-up and outcomes

RSA assessments were performed postoperatively after mobilization with partial weight bearing, and at 3 months, 6 months, 1 year and 2 years follow-ups. Clinical evaluation, which included Harris Hip Score (HHS) [14] and Oxford Hip Score (OHS) [5], was conducted preoperatively and at all follow-up examinations.

Radiostereometric analysis (RSA)

During surgery, 7–9 tantalum markers (Ø 0.8 mm, Industrial Techtonics, Ann Arbour, MI, USA) were inserted into the proximal femur using the UmRSA Injector (RSA Biomedical, Umeå, Sweden). Following surgery, postoperative RSA radiographs were taken after mobilization with partial weight bearing. This first examination represented the baseline position and reference point for all subsequent measurements. Follow-up RSA radiographs were taken at 3 months, 6 months, 1 year and 2 years postoperatively (Fig. 1).

Fig. 1
figure 1

RSA Fitmore. Fitmore shaft with tantalum markers

Full size image

Model-based RSA analysis (RSAcore v3.34; LMUC, Leiden, The Netherlands) was performed using combined three-dimensional images of the stem and head construct. RSA precision was determined after 19 double examinations at 2 years follow-up by calculating the 95 % prediction interval of the accuracy of zero motion [26, 32]. In addition, the condition number and the average of the mean error of rigid body fitting were determined.

Analysis of translation and rotation was performed by RSAcore at the Department of Orthopaedics, Leiden University Medical Center, in line with recommendations by Valstar et al. [32]. Translations and rotations of the stem were calculated using a coordinate system, with the origin in the center of the stem model.

Patient positioning allowed description of hip stem migration using anatomical directions.

Statistical analysis

Analysis of longitudinal RSA and clinical data were based on linear mixed models, with inclusion of time included as a linear spline. Random effects were included for the time term.

Statistical analyses were performed using SPSS 22 (Chicago, Illinois, USA). RSA outcome values are reported as mean, 95 % CI and range. Wald tests were used to assess the significance of migration over time. Clinical outcome data are reported as mean and range. Differences were considered significant at a p value <0.05.

Results

The baseline characteristics of the patients are shown in Table 1. 28 Patients were available at the 3- and 6-months clinical and RSA examinations; 27 patients were available at the 1 year follow-up and 24 at the 2-year follow-up. One patient died after 6 months from causes unrelated to the surgery. A further three patients underwent revision THA between 1 and 2 years post-surgery: one underwent two-stage revision surgery due to a deep hip infection; one had a one-step revision due to aseptic loosening; and one was revised at an external institution due to non-specific thigh pain (Fig. 2).

Table 1 Patient characteristics
Full size table
Fig. 2
figure 2

Patient flowchart

Full size image

RSA outcomes

The radiological baseline examination was performed after mobilization with partial weight bearing at a median of 3 (range 2–10) days postoperative. RSA precision after the 19 double examinations at 2-year follow-up showed a 95 % prediction interval of the accuracy for zero motion for translation along the x-, y- and z-axis of 0.06, 0.08 and 0.18 mm, respectively, and the precision of rotation about the x-, y- and z-axis was 0.35, 0.61 and 0.08°, respectively. The average (±SD) condition number was 61 (±38) and the average mean error of rigid body fitting was 0.19 (±0.10).

Migration and rotation occurred during the first 3 months post-surgery but subsequently stabilized (Figs. 3, 4). 3 months analysis revealed a mean translation along the y-axis, indicating axial subsidence, of −0.39 mm (95 % CI −0.60 to −0.18; range −1.82 to +0.67) (p < 0.001). This was unchanged at the 2-year follow-up, when the translation along the y-axis was again −0.39 mm (95 % CI −0.64 to −0.14; range −2.3 to +0.5). There was no significant difference in translation along the y-axis compared with the 3-month assessment at 6 months (−0.36 mm), 1 year (−0.40 mm), and 2 years, at p values of 0.516, 0.670 and 0.958, respectively.

Fig. 3
figure 3

Mean RSA translation. Error bars represent 95 % confidence interval. First examination (0) postoperative after mobilization with partial weight bearing; further examinations after 3, 6, 12 and 24 months

Full size image
Fig. 4
figure 4

Mean RSA rotation. Error bars represent 95 % confidence interval. First examination (0) immediately postoperative after mobilization with partial weight bearing; further examinations after 3, 6, 12 and 24 months

Full size image

Rotation along the y-axis, indicating internal rotation of the stem, was, on average, 0.85° (95 % CI 0.44–1.26; range −1.18 to +3.28; p < 0.001) after 3 months. At 2 years, the internal rotation of the stem was, on average, 1.09° (95 % CI 0.52–1.66; range −1.0 to 4.1°). The stem was rotationally stable between 3 months and 2 years post-surgery, with no significant differences between the results at 3 months and those at 6 months (0.81°, p = 0.747), 1 year (0.92, p = 0.654) and 2 years (p = 0.265).

The surgical approach, i.e. lateral and antero-lateral approach did not influence stem migration at 3, 6, 12 or 24 months. There was no significant difference of the maximum total point motion between the two approaches.

Migration was recorded up to the 6-month follow-up in one patient, at 2 mm axial subsidence and 3° of internal rotation of the femoral stem. This subsequently stabilized, and the patient had a good clinical outcome, with no signs of loosening.

Clinical assessment

Mean HHS increased markedly, from 60 (range 30–80) preoperatively to 92 (range 61–100) at 3 months post-surgery (p < 0.01). At 6 months and 1 and 2 years, the respective mean HHS was 96 (range 54–100), 95 (range 69–100) and 99 (range 83–100) points. The improvement in HHS after 3 months was not statistically significant, at p values of 0.167, 0.240 and 0.084, respectively.

Mean OHS also showed significant improvements, increasing from 23 points (range 10–37) preoperatively to 44 points (range 30–48) at 3 months (p < 0.01). At 6 months and 1 and 2 years, the mean OHS was 46 (range 17–48), 45 (range 35–48) and 46 (range 42–48) points. Again, the improvement in OHS after 3 months was not statistically significant, at p values of 0.351, 0.495 and 0.134, respectively. Nevertheless, 95.8 % of the analyzed patients were pain-free at the 2-year follow-up.

Revision cases

The first patient to be revised had very good 1 year functional and radiological outcomes, with an HHS of 100 and an OHS of 48 (Fig. 5). Unfortunately, the patient developed a septic arthritis due to unknown reasons and had to be revised. For final treatment, a two-stage procedure was necessary.

Fig. 5
figure 5

Mean subsidence (migration along y-axis), with standard and single subsidence results, in the three revision cases. Case 1 showed excellent clinical results, albeit with septic arthritis after 1 year. Case 2 reported persisting hip pain. Case 3 had aseptic loosening with persistent pain, requiring revision surgery

Full size image

The second revision patient reported non-specified thigh pain after an initial straight lateral approach. An initially very good HHS of 100 at 3 months deteriorated to 69 at 1 year (Fig. 5). Similar results were seen with an initial good OHS of 48 at 3 months deteriorating to 35 at 1 year. He was revised at another institution after 12 months. Although revision surgery revealed a stable stem, he had a tear of the hip abductors that were responsible for the pain subsequent to a lateral approach. Unfortunately, the patient was not available for further clinical follow-up.

The third revision patient complained of persistent hip pain while walking. RSA analysis indicated progressive migration over the first postoperative 12 months (Fig. 5). Single-photon emission computed tomography at 1 year revealed loosening/insufficient osseointegration of the femoral shaft component. Revision surgery revealed only a very small amount of osseointegration on the stem in zone 7, based on the Gruen classification [12]. No bacterial contamination was found on implant sonication. The reason for aseptic loosening remained unclear.

Discussion

In this study, we analyzed the clinical and RSA results of 28 consecutive THA patients who received a short stem for symptomatic osteoarthritis. The stem showed good short-term clinical results and a small degree of initial migration. In the vast majority of cases, stem stabilization occurred in the first 3 months postoperatively, with no substantial implant migration seen subsequently up to the 2-year follow-up.

In uncemented THA, primary stability is a prerequisite for osseointegration and long-term success [18, 25, 29]. This includes a tight proximal fit with small reversible axial micromotion between the implant and the bone, as well as stem migration with axial and rotational stability. These fit and fill criteria are highly influenced by femur anatomy and prosthesis design [16]. Micromotion is also linked to fibrous ingrowth and resorption of bone at the implant-bone interface [30]. Freitag et al. compared the implant-specific periprosthetic bone mineral density (BMD) reduction between the Fitmore short-stem and the CLS long stem [9]. The Fitmore short-stem not only showed similar BMD results after 1 year in most zones, but even significant less BMD reduction in zone 6. In experimental and clinical studies, a critical threshold for cyclic micromotion appears to be 150 μm [7, 17]. Above this limit, bony ingrowth will not occur [17, 30]. With axial migrations greater than 1.5 mm, Krismer et al. experienced stem loosening at an accuracy of 79 % within 2 years [19]. However, a recent review found that excessive average migration is not always a predictor of implant longevity [6]. It is therefore unclear at which location the critical threshold must be avoided, and whether reversible micromotion or irreversible migration is more detrimental.

Early axial migration has been investigated in different designs of uncemented tapered stems with fixation in the proximal or slight metaphyseal area. In a biomechanical study, Bieger et al. [3] compared the primary stability and axial migration of the short Fitmore Hip Stem and the well-established long CLS Spotorno stem (Zimmer). After 100,000 load cycles, the interface motion for the Fitmore stem was below the critical threshold of 150 μm, and not significantly different from that for the longer CLS stem. In terms of migration, our clinical results with the Fitmore stem were comparable to those for other proximal fixed stems, at a mean migration in the y-axis after 2 years of 0.39 mm [11].

Migration patterns for the proximal plasma-sprayed hydroxyapatite-coated Taperloc stem (Biomet Inc, Warsaw, IN, USA) indicated that the majority of migration occurred within the first 3 months, after which the stems stabilized, at a mean subsidence of 0.28 at 3 months and 0.25 mm after 2 years [4]. For the CLS Spotorno stem (Zimmer), mean subsidence after 5 years was reported to be 1.7 mm [34]. Both stems designs had excellent clinical results [4, 34].

Rotational stability is also regarded as a further fundamental indicator for osseointegration [22]. In a comparative study, custom-made cementless stems showed a higher stability compared to conventional cementless stems [11]. Pepke et al. compared the primary stability of the Fitmore stem and the long CLS Spotorno stem, analyzing rotational stability and bending characteristics [24]. They found no significant difference in rotational stability. However, the mediolateral bending characteristics revealed that the CLS stem had significantly higher flexibility. It is not yet know whether the increased rigidity of the Fitmore stem influences load transfer and, potentially, bone remodeling. However, Rohrl et al. presented RSA analysis of the CFP stem, a femoral neck preserving short stem, that showed an increase in initial internal rotation that stabilized after 1 year [27]. These findings correspond to our results with increased internal rotation of the Fitmore stem in the first 3 months.

Another factor affecting initial stem stability is bone mineral density (BMD) [1]. Short-stem implants have the advantage of avoiding proximal–distal mismatch, particularly in osteoporotic patients with disproportionately wide intramedullary canals. In women, there is a negative correlation between a low BMD, the intra-osseous dimension of the proximal femur and initial stability [1]. Gustke suggests that additional undersizing of the femoral stem can lead to subsidence [13]. Also Jahnke et al. emphasized the importance of these fit and fill criteria [16].

In general, short stems appear to yield good clinical results [33], which was reflected in the preliminary clinical results achieved with the current stem [13]. Short stems therefore appear to provide initial stability and long-term fixation in patients undergoing THA, and address the challenges of proximal–distal mismatch, less-invasive surgical exposures and bone preservation for potential revision surgery [23]. In our patients, the short-term results are also promising.

We note several limitations. The study had a relatively small sample size that was not randomized or controlled. Furthermore, the small sample size did not allow us to correlate BMI and implant migration although a tendency was seen with the Fitmore stem in previously published studies [10]. Six patients had to be excluded from the RSA analysis because of insufficient marker presentation. The follow-up time with RSA was limited to 2 years. While these limited numbers of implants appeared to stabilize after 3 months we cannot predict future migration. However, other RSA studies suggest stabilization in this time frame is typical for those implants that do well.

In conclusion, the current stem for primary THA showed early minimal migration and rotation in the first postoperative 3 months, after which the stems stabilized. There were significant improvements in clinical performance in the 2-year postoperative period.