Influence of bone morphology and femur preparation method on the primary stability of hip revision stems

Aseptic loosening is one of the major reasons for re‐revisions of cementless revision stems. Insufficient primary stability is associated with bone characteristics and the surgical process. This study aimed to investigate how femur morphology and preparation methods influence the primary stability of revision stems. The Femur morphology was described by the upper femoral curvature (UFC) and an individualized Dorr type classification based on the ratio between the canal‐to‐calcar ratio (CCR*) and the cortical index (CI*) introduced as the cortical‐canal shape (CCS). Manual and powered reaming in combination with helical and straight reamers were used to prepare the bone cavity of 10 cadaveric human femur pairs. Forces during stem impaction were recorded (Reclaim, Depuy Synthes). Micromotion at the bone–implant interface during cyclic axial loading and torsional load to failure was determined. The CCS and impaction forces (R2 = 0.817, p < 0.001) or torsional strength (R2 = 0.577, p < 0.001) are inversely related. CCS did not correlate with micromotion during axial loading (R2 = 0.001, p > 0.999), but proximal femoral curvature did (R2 = 0.462, p = 0.015). Powered reaming and straight reamers led to an improved torsional strength (both: p = 0.043). The Individualized Dorr classification CCS and UFC allows a good estimation of the primary stability of revision stems. For severely curved Dorr type‐C femurs, an alternative anchorage method should be considered clinically.

with cementless revision stems. 18,19 Postoperative stem subsidence less than 2 mm 20 and micromotions of less than 30 µm are considered harmless, 21,22 while micromotions >150 µm are considered an unfavorable condition for osseointegration. 23 Patient-related factors such as femur shape or bone quality but also upper femoral curvature (UFC) may have an impact on the primary stability of cementless revision stems. Bone quality and decreased femoral strength are often age-related 24 and associated with comorbidities. 25 Macroscopically, the femoral morphology is usually described by the Dorr type. 26 Femurs of young, active patients often have a funnel-shaped medullary cavity with a thick cortex categorized as Dorr type A. Femurs with a stove-piped shaped canal and a thin cortexcommon in the elderly 27 -are assigned to Dorr type C. Since revision stems are anchored in the diaphysis, the cortical morphology of the femur influences the stability. Dorr type C bones with reduced bone strength often require stems with larger diameters. 27 Preoperative determining of the bone morphology according to the Dorr type is important for planning but usually is only performed qualitatively. Dorr's classification either refers to the canal-to-calcar ratio (CCR) or the cortical index (CI), determined at fixed positions without considering body-specific factors such as patient height, 26,28 and an ideal AP or ML view orientation cannot be guaranteed. Furthermore, using the CCR or the CI does not necessarily yield the same Dorr type for a specific femur, which makes a unique classification impossible in some cases. Combined, this can lead to ambiguous results even for the most experienced surgeons. 29,30 A unique classification to evaluate the morphology of the proximal femur based on the original work of Dorr is desirable to support surgeons' implant choices and to elucidate differences in clinical outcomes.
In addition to proximal femur morphology, especially important for elderly, 2 the cavity preparation method bears the potential to influence the outcome of the procedure. Likewise, in primary THA, where preparation affects stem seating for low bone mineral density (BMD), 31 broach design influenced the shape of the cavity, 31 this could also apply to the reamer design in revision THA. Furthermore, manual reaming could cause more variable cavities compared to powered reaming. However, the effect of these factors on the primary stability of revision stems has not been fully investigated.
This study aimed to investigate how femur morphology and preparation methods influence the primary stability of revision stems.
The BMD of the cortical bone was determined at a 10 mm wide ring of the femur shaft, 100 mm below the trochanter minor (threshold: [500, 1500 mgHA/cm 3 ]; AVIZOLite 9.7.0, Thermo Fisher Scientific). The cortical bone and medullary canals were segmented based on thresholds (cortical bone: 500-1500 mgHA/cm 3 ; medullary canal: −250 to 500 mgHA/cm 3 ). 31 CT data were pre-aligned based on the homogeneously distributed point cloud of the cortical shell using its three main principal components (Matlab 2020b, The MathWorks).
Percentages of donors' body height (BH) were used as bodyspecific measures for femur classification. The anatomical femur axis was defined by a best-fit line through the diaphysis between 2% BH (z 2% ) and 8% BH (z 8% ) distally to the trochanter minor (TM). This femur section was further divided into two equal sub-sections (z 2% to z 5% and z 5% to z 8% ) and used to determine the UFC (β) as the angle between the respective canal axes of the sub-sections ( Figure 1A).
Individual body-height-specific femur types CCR* and CI* (* stands for the individualized adapted method) on the bases of Dorr et al. 26 were determined, by replacing the fixed positions of 30 mm (z 30mm ) and 100 mm (z 100mm ) below the TM with bodyspecific measures of 1.7% BH (z 1.7% ) and 5.8% BH (z 5.8% ).
Positions are the same for an average German body height of 172 cm 34 ( Figure 1B, Equation 1). (1) AP and ML (Lauenstein) radiographs were replaced by cross-sectional areas of the cortical shell (A cort ) and the intramedullary canal (A im ) determined from CT data. Equation (1) refers to the diameter that is available in the x-ray projection. In CT images, however, the cross-sectional area is available, which is roughly proportional to the square of the diameter. Hence, the individualized CI* was determined from the medullary cross-sectional area A im,5.8% and the cortical cross-sectional area A cort,5.8% ( Figure 1B; Equation 2).
Inline with the former, the individual CCR* was determined with cross-sectional areas instead of diameters ( Figure 1B, Equation 4).
The ratio between CCR* and CI* (CCR*/CI*) is further referred to as the cortical-canal shape (CCS) and is used to evaluate the in-vitro experiments. These changes had no effect on the bone-implant interface. Ultimate torque to failure (angular velocity: 0.5°/s) was determined in combination with a static axial preload (500 N).

F I G U R E 4
The canal-to-calcar ratio (CCR*) as a function of the cortical index (CI*) exhibits a negative correlation (f(x) = 1−0.5x; p = 0. 003). The cutoff limits for the Dorr classification according to the CCR and the CI are indicated in black. 26 The cutoff limits for the modified classification (CCS) are indicated in orange (values are given in Table 1).
Straight reamers appeared to be 11% superior compared to helical reamers in terms of torsional strength (p = 0.043, 1−β = 0.361; Figure 7B). With a decreasing CCS higher torsional strength was observed (R 2 = 0.579, p < 0.001; Figure 7A). F I G U R E 5 Statistically sig. negative correlation between maximum impaction force and the CCS. CCS, cortical-canal shape.

| DISCUSSION
F I G U R E 6 Statistically sig. positive correlation between the micromotion and the UFC for specimens exhibiting more than 5 μm of micromotion. UFC, upper femoral curvature.
torsional strength. 35 Since the tested stems had only a small difference in stem size (15 ± 1), the contact area was assumed to be similar since the stem-length, the number of vertical fins and their shape was identical. The observed torsional strength within the bone-implant interface was slightly above the loads occurring during high-load activities like jogging. 36 The use of body-height-specific measures to determine CCR* and CI* is justified by the correlation of BH and femur length found in this study alike to previous ones. 37 Standard powered reaming appears to be superior compared to the manual approach since it is time-efficient, involves less physical work and yields to higher torsional strength, presumably due to a more uniform cavity. It is uncertain whether straight reamers further improve the junction due to their changed geometry or merely due to the sharper cutting edges caused by the less previous operating time of the reamer.
One-third (35%) of stems exhibited micromotions of less than 5 μm, in the range of the signal noise, and therefore, no correlations between micromotion and CCS or BMD were observed. The lower performance observed in curved femurs in terms of increase in micromotion may be related to a smaller contact area within the less congruent interface, which is known to reduce primary stability. 35 Mechanically, a low CCS with a high BMD (R 2 = 0.47, p = 0.002), as found in Dorr type A bones, increases system stiffness leading to higher impaction forces and energies which improves primary stability. In contrast, the thin cortex of Dorr type C bones with low BMD, often found in elderly patients most commonly affected by revision THA, 2,27 does not provide comparable radial pressures at the implant-bone interface to achieve similar torsional strength. For these specimens, in particular, the maximum torsional moments are close to loads projected for activities like jogging. 36 Overall, femur morphology appears worthwhile to be considered during cementless revision THA, because of the exhibited correlation between impaction force, impaction energy, and bone-related parameters (BMD, CCS, and UFC) with primary stability.

| CONCLUSION
With the CCS a unique Dorr type* can be determined for any femur.
Including body-height-specific measures and a semi-automatic computation of the CCR* and the CI* from three-dimensional CT data improves the reproducibility and reduces inter-/intraobserver variations. The CCS appears to be able to predict the mechanical relationships between bone and implant, and such may support adequate prosthesis selection. Apart from the clinical situation, this allows objective body-specific femoral morphological parameters to be used for in-vitro analysis. Since micromotion is affected by UFC and torsional stability by CCS, it may be worthwhile to consider both parameters in combination. In contrast to bone morphology, the preparation method plays a minor role for primary stability. Slight advantages of powered compared to manual reaming and the reamer geometry will most likely not decide the revision's success. A combination of several adverse factors may lead to early implant failure and re-revision. In severely curved type C* femurs that have been manually prepared, primary stability could be critically compromised, so an alternative anchorage method may be beneficial for these patients. Study design; project coordination; writingreview and editing.

AUTHOR CONTRIBUTIONS
All authors have read and approved the submitted manuscript.