Influence of Design Variations on Early Migration of a Cemented Stem in THA

Abstract

The clinical results of using different cemented stems in total hip arthroplasty may vary because of shape, surface finish, and choice of metal alloy. Less is known about the influence of size and offset using one and the same design. Data from 140 patients (140 hips) of a cohort constituting 197 patients (231 hips) implanted with cemented Spectron Primary stems were extracted from patients studied in four randomized studies primarily designed for other purposes. Stem size (1, 2, 3, or larger), normal/extra offset, neck length and true offset (horizontal length between the stem and head center of the inserted modular implant) were recorded in millimeters. The radiographic appearance of the cementing was graded on the immediate postoperative radiograph according to Barrack. The patients were examined with radiostereometric analysis of stem migration for a 2-year followup period. Stem Size 1 (the smallest) showed an insignificant increase in subsidence compared with the bigger sizes. Using stepwise linear regression analysis, the quality of the cementing (Barrack C2) was the only parameter that had any certain influence of the subsidence at 2 years. Stem size 2 tended to show minimum valgus alignment and sizes 1 and 3 and bigger showed minimum varus tilting. Our findings suggest that placement of the stem with the tip against the posterior cortex (C2) and stem size influence the primary fixation in different ways. This effect is small, however, indicating that further confirmation with longer followup is necessary.

Level of Evidence: Therapeutic Level II-1 (prospective cohort study). See the Guidelines for Authors for a complete description of levels of evidence.

The influence of stem material selection, cementing technique,¹⁸ surgical experience¹¹ implant shape, and surface structure on the clinical results of total hip arthroplasty (THA) has been evaluated in numerous studies.^7-9,21 During the last decade many manufacturers have expanded their inventories of some implant designs by introducing options with different offsets and sometimes with a larger variation of sizes. Such changes should facilitate restoration of offset, optimize the width of the cement mantle, and reduce the need for change of implant type or customized devices in patients with small femurs. These changes have been made under the presumption that they would improve intraoperative utility, but their effectiveness in improving clinical performance has not been documented.

Restoration of normal offset during THA theoretically has advantages.^6,15,19,20 If offset decreases after THA, the tension in the abductor muscles may decrease and risk of instability. With less offset higher muscular forces are theoretically required, which would result in higher hip joint loads and stresses at the cement-implant and cement-bone interfaces and possibly an increased risk of loosening. Higher forces may result in greater wear.²⁴

In practice, there are several ways to change offset including change in neck length, stem-neck angle, and medial or lateral shift of the neck. These measures can be combined in various ways, resulting in additional options. In addition, choice of implant size may influence stem size because the length of the femoral neck often increases with increasing stem size. During the operation offset not only will be determined by implant parameters but also may vary depending on stem positioning.

Because there is no information available on the influence of stem geometry and offset on the fixation of cemented stems, we extracted subgroups from a number of randomized trials using the same implant design and the same type of bone cement. All patients had been followed up with repeated radiostereometric analysis (RSA) recordings of migration for up to 2 years. We hypothesized that higher migration in terms of more subsidence and more stem rotations would be recorded with decreasing implant size and higher offset.

MATERIALS AND METHODS

Data from 140 patients (140 hips) implanted with THA prostheses were extracted from four randomized studies including 197 patients (231 hips) designed to study new and undocumented bone cements or polyethylene (PE). The selection of patients was based on use of the same prosthesis design (Spectron Primary, Smith and Nephew, Memphis, TN) and type of bone cement (Palacos cum gentamicin, Schering Plough, Heraus Kultzer, Wehrhrim Germany). In patients who had bilateral THAs, the first hip operated on was used for data collection and RSA examination with up to 2 years of followup. There were 50 men and 90 women (140 hips) with a median age of 60 years (range, 29-81 years) and weight of 75 kg (range, 38-120 kg).

The Spectron primary prosthesis is a straight stem is made of cobalt-chromium (Co-Cr) alloy and has a 12/14 global taper. Its proximal 1/3 is grit blasted (surface roughness [Ra], 2.8 μm). The distal part is smoother (0.7 μm) and a centralizer is attached to the tip of the stem. The Spectron stem is available in five sizes. Each size is available with standard or increased offset (Fig 1). In the high offset design the neck is longer and is medially shifted. All stems have the same neck angle (131°). Neck length varies between −3 to +16 mm. Twenty-eight-mm femoral heads made of Co-Cr alloy were used in all hips. The stems had been supplied with three titanium towers, each with a tantalum marker attached to its tip (Fig 1). During the operation tantalum markers (ø = 0.8 mm) were inserted into the proximal femur.

FIG 1: The Spectron primary stem with ordinary and extra offset (right) is shown. The manufacturer has supplied these stems with tantalum markers inserted into each titanium tower (white arrows).

Preoperative planning was done in all patients to facilitate correct choice of prosthesis size, neck length, and offset. The anterolateral approach (modified Hardinge) was used in all operations. Third-generation cementing technique (brush, high-pressure lavage, distal plugging of the femoral canal, tamponades soaked in adrenalin solution, retrograde injection of cement into femur, and cement pressurization) was used. All patients were operated on at the same hospital (Sahlgrenska University Hospital) by 15 surgeons.

Implant parameters (stem size, standard/high offset, and neck length) were recorded. The true implant offset (horizontal distance between a central stem line and the modular head center) was recorded using tables supplied by the manufacturer. This offset could vary between 32 mm and 48 mm depending on the parts chosen during the operation. Because stem positioning also influences offset, we also measured offset on radiographs (radiographic offset). This offset was defined as the shortest horizontal distance between a central line in the femoral canal and the head center. The length of the resected femoral neck (the shortest distance between the prosthetic collar and the most medial point on the lesser trochanter) and proximal or distal stem positioning (the vertical distance between the tip of the greater trochanter and the head center) were recorded. These measurements were made on digitized radiographs using specific software (M-desk, RSA Biomedical, Umeå, Sweden). Radiographs from the early part of this study were scanned using a flat-bed scanner with a resolution of 300 dpi (16 bits/pixel). Later in the study, radiograph plates with a resolution of 254 dpi (grey scale depth 10 bits/pixel) were used. The 28-mm femoral head was used for calibration. In 134 hips these measurements could be made on the postoperative examination. In one patient, all radiographs were unavailable. In five patients, radiographs at the 1-year followup were used.

The radiographic appearance of the cement mantle was graded on anteroposterior (AP) and lateral radiographs taken immediately postoperatively. The Barrack classification¹ was used (A = white out; B = slight radiolucency at the cement-bone interface; C = radiolucency involving 50-90% of the cement-bone interface or defective or incomplete mantle; C1 = cement mantle thickness < 2 mm; C2 = presence of stem-cortex contact (Fig 2); and D = radiolucency at the cement-bone interface of 100% in any projection or a failure to fill the canal with cement such that the tip of the stem was not covered). On examination, 134 patients' radiographs were found to be of sufficient quality for this evaluation; in five of these patients, the 1-year followup radiographs were used. Eleven patients were classified as A, 87 were classified as B, 25 were classified as C1, 11 were classified as C2 and none were classified as D. None of the primarily included 197 patients (231 stems) had undergone revision operation at the 2-year followup.

FIG 2: Proximal (+)/distal (−) migration of the gravitational stem center related to stem size is shown. Stem size 3 includes data for one size 4 stem and one size 5 stem. Data are shown mean ± SE.

Radiostereometric examinations were made with the patient supine between 3 and 7 days postoperatively, after 6 months, and at 1 and 2 years postoperatively. The uniplanar technique with the calibration cage positioned under the examination table was used (15). Rotations of the stems about the three cardinal axes and proximal and distal translation were measured.

To maintain the precision of the measurements this analysis was only made if at least three well-defined markers could be identified. The precision of the measurements was evaluated from 63 repeated examinations. The precision expressed as 2.7 standard deviations (SD) of the differences between two such subsequent radiostereometric examinations (time interval about 15 minutes) were calculated. Proximal or distal stem migration of 0.11 mm or more and rotation of 0.20° to 0.71° or more could be established with a certainty of 99% (99% confidence interval [CI] limits).The mean error of rigid body fitting for all segments studied was 0.15 (range, 0.03-0.35) and the corresponding mean value of the condition numbers representing degree of marker scatter was 47 (range, 19-125).

At the 2-year followup 140 hips were examined with RSA. There were 13 missing 6-month observations and nine missing 1-year examinations because the patients did not present for followup or because the stereoradiographs did not qualify for adequate evaluation (insufficient numbers or only poorly scattered tantalum markers visible).

There were only three patients with the longest neck (+12) and those were classified as having the second longest necks (+8) to simplify the statistical calculation. Accordingly two patients with stem size 4 or 5 were grouped as size 3 stems. The Kruskall-Wallis test was used to evaluate any difference of proximal-distal stem translations and stem rotations between sizes 1 to 3 using data from the 2-year followup. The influence of age, gender, weight, geometric stem parameters (size, standard/high offset, neck length, and true offset), stem position according to conventional radiographs, and cement mantle quality on stem migration (proximal-distal translations, rotations) were evaluated using a stepwise linear regression (Tables 1, 2). Probability values less than 0.05 were regarded as significant difference.

TABLE 1: Patient and Implant Data

TABLE 2: Variables and Codes Used in the Regression Analysis

RESULTS

The proximal and distal stem migration did not vary between the three classes of stem sizes studies (sizes 1, 2, and 3-5). At 2 years, there was no difference in subsidence between the three groups of stem size (size 1, mean 0.18 mm, SD = 0.20; size 2, mean 0.09 mm, SD = 0.24; sizes 3-5, mean 0.11 mm, SD = 0.15; Kruskall-Wallis test based on the three classes: p = 0.1) (Fig 2).

Size 2 stems rotated into varus (0.02º, SD = 0.24), whereas size 1 and sizes 3 to 5 rotated into valgus (size 1, mean −0.07º, SD = 0.17; sizes 3-5, mean −0.06º, SD = 0.19; p = 0.005) (Fig 3). All stems showed mean posterior tilt between 0.18 and 0.42° (p = 0.75) and retroversion (0.03 - 0.14°, p = 0.09) without any difference between the three size groups.

Fig 3: Valgus (+)/varus (−) tilt of the three classes of stem sizes are shown. Stem size 3 includes data for one size 4 stem and one size 5 stem. See also text to Fig 2. Data are shown mean ± SE.

The radiographic offset was on average 4.7 mm (SD = 6.3) smaller than those values obtained by the manufacturer based on the final assembly of the modular components in each individual (Fig 4).

Fig 4: Difference between the offset of each individual stem and the corresponding offset as measured on AP radiographs (n = 136).

A stepwise linear regression analysis revealed that only cement mantle quality had any influence on the proximal or distal migration at 2 years. This influence was weak (r-square = 0.06). Hips classified as Barrack C2 subsided more (p = 0.04) (Fig 5) than those with other levels of Barrack classifications. None of the demographic variables or those related to stem geometry (Fig 6) or radiographic position had any certain influence.

Fig 5: Proximal (+)/distal (−) stem migration related to cement mantle quality is shown. Data are shown mean ± SE for Barrack groups A to C2.

Fig 6: The plot of proximal (+)/distal (−) stem migration related to prosthetic offset (the combined effect of stem geometry and neck length) is shown. Note the absence of correlation (r² = 0.002, Spearman's rho = 0.025, p = 0.8) and presence of outliers in the interval with the shortest prosthetic offset.

DISCUSSION

The degree of femoral stem offset determines the lever arm of the hip joint and may thereby influence component fixation. Restoring offset also may improve THA function and may minimize the risk of complications such as dislocation,^1,3,20 impingement, and limp.¹⁷ Sakalkale et al²⁴ observed decreased PE wear in THAs in which a lateralized femoral component was used compared to those in which a standard femoral component was used. Kalairajah¹² et al reported the clinical and radiographic outcomes of Taperloc arthroplasties implanted with Boneloc bone cement. The patients who had smaller stems (7.5 mm or 10 mm) had a 27% failure rate whereas those patients who were implanted with stems equal to or greater than 12.5 mm had a 12% failure rate. Bourne et al² reported that bone cement pressure during stem insertion will increase when progressing from a small to a large stem, which could be expected to influence the fixation. Sivananthan et al²⁵ did, however, not observe any failures in a group of 46 small Exeter stems inserted in an Asian study population, but the mean followup was only 4 years.

We tried to limit the number of confounders in our study, but still it has limitations. Our patients were operated on at the same hospital; each patient had the same surgical approach used and all had the same stem implanted (Spectron stem). The variables studied were, however, not independent and not chosen by random. One could in fact question whether one could definitively ascertain the influence of stem geometry on fixation in a clinical study. In clinical practice, choice of stem geometry will be more or less biased by factors related to the patient and the surgeon. The most obviously biased data in our study were that all patients who were implanted with size 1 stems were women except for one. Weight and total offset showed an uneven distribution whereas the other demographic variables seemed to be distributed more symmetrically. None of the demographic factors or those related to stem geometry seemed to influence fixation. Authors of previous RSA studies on stems have reported no influence of patient related factors on stem fixation,^14,22 which our findings confirm. It is possible that despite being a comparatively large patient cohort with RSA followup, the number of observations is still too small to detect any true difference. It should be emphasized that we studied only one specific stem design and therefore the results likely cannot be generalized for describing other stem geometries or surface treatments.

Considering the wide selection of implant designs on the market, there is little information on the influence of stem size on the revision rate. In our study, we found no certain influence of stem size on component subsidence. The observation of different rotations around the AP axis between the three groups of stem sizes is difficult to explain. In the regression analysis we found no difference, suggesting interdependence between two or several variables.

The quality of the cement mantel is important for stem fixation.^{4,7,10,18,23,26} Ramaniraka et al²³ evaluated micromovements at the cement-bone and cement-stem interfaces. Movements were minimal if the cement mantel had a thickness of 3 to 4 mm but increased if it became wider. Abnormally high micromovements occurred when the cement was thinner than 2 mm.²³ As early as 1983 Carlsson et al⁵ observed scalloping around stems with broken cement mantles. They suggested use of a centering device to avoid this complication. In our study, patients who were classified as C2 had more subsidence than the patients who had other classification levels. This observation suggests that patients classified as C2 will have stems that will debond more easily from the mantle, which will facilitate transport of joint fluid and debris from the joint to the interface. In the Swedish National Register, contact between the stem and the cortex corresponding to a C2 position has been pointed out as the second main reason for revision surgery (6%) within 6 years.¹⁶

In the Spectron primary design the minimal cross-sectional diameter (range, 6-12 mm) and the length (range, 115-135 mm) vary with stem size. The smallest stem sizes have a more rounded shape than the larger one, which are more tapered. These size and shape variations mean that load distribution will vary. Provided that the surgeon chooses sizes according to femoral anatomy and that the femoral size increases with increasing demands dictated by patient-related variables, clinical loosening might be avoided. The femoral anatomy may show large individual variations. Highly active and heavy patients may have thick cortical bone and a narrow canal and other factors such as cement mantle thickness may vary considerably depending on the internal shape of the femoral canal. This means that any stem must have safety margins and be able to share loads higher than predicted in the standard case. This is especially important when the stem is rough and not aimed to migrate inside the mantle.

Our findings suggest that in the majority of cases, customized modular stems based on careful preoperative planning result in similar stability in the short term. This finding does not, however, exclude the possibility that some combinations may be less successful in the long term, especially when they are used in heavy and/or active patients. We emphasize our results only concern one specific stem design and probably cannot be generalized for results using other stem geometries or surface treatments.