The influence of cement thickness on stem subsidence and cement creep in a collarless polished tapered stem

Prosthesis

We used clinically available collarless polished tapered (CPT) stems (Zimmer Biomet, Warsaw, Indiana) with less than 0.1 µm in surface roughness. The stems were tapered from proximal to distal on both the frontal and lateral sides. This study used a total of six CPT stems: two each of small, medium and large (Fig. 1). The most proximal transverse diameter increased by 2.5 mm, and the offset diameter increased by 2 mm, between the small and medium stems and between the medium and large stems, respectively. The stems were made of cobalt-chromium (CoCr) alloy.

Fig. 1

CPT stems and No. 3 rasp. These are CPT stem sizes small, medium and large, respectively, from the left (right, the No. 3 large rasp).

Composite femurs

We used composite femurs #3403 (Pacific Research Laboratories, Inc., Vashon, Washington) that resembled human femurs in form and in mechanical and material characteristics.¹⁶ The composite femurs were cut at the femoral neck. All the femurs were reamed with a large No. 3 rasp (Zimmer Biomet), and CPT stems in different sizes were inserted to provide a variety of cement thicknesses. As a marker for 3D analysis by micro-CT, four tantalum balls 0.4 mm in diameter (Bal-tec, Micro Surface Engineering, Los Angeles, California) were embedded in the outer circumference of the composite femurs 1 cm proximal to just above the lesser trochanter (Fig. 2a and Fig. 2b). Similarly, tantalum balls 0.4 mm in diameter were embedded circumferentially near the tips of the stems (Fig. 2c) that were to be inserted into the femur models. Each ball-composite femur unit was then soaked with blended vegetable oil, made of rapeseed oil and soybean oil, (The Nissin Oillio group Ltd., Tokyo, Japan) for 24 hours (Fig. 2d) to mimic the wet conditions of the in vivo femoral environment.

Fig.

a) Figure showing small drilled holes for inserting tantalum balls; b) radiograph film after fixing stem with cement. Tantalum markers are visible on the outside of the composite femur (white circles) and in the cement (black circles); c) Radiograph showing placement of tantalum balls near the tips of the stems and d) Figure showing a ball-composite femur unit soaked in vegetable oil.

Loading test equipment

The composite femurs were attached to the testing instrument, which was composed of S45C structural carbon steel and epoxy resin (Devcon B; ITW Industry Co. Ltd, Osaka, Japan). The epoxy resin was shaped in the same external form as the composite femurs. Bone cement (Osteobond; Zimmer Biomet) was mixed in a vacuum cement mixing system, and 80 g of the mixed cement was inserted into the composite femur using a cement injection gun. Tantalum balls (1 mm in diameter) (Bal-tec, Micro Surface Engineering) were attached to the tips of the stems using glue (Fig. 2c). A centraliser was also attached before inserting the stem into the composite femur. A bone plug made of wax was fixed into the femoral canal. Before the cement solidified, two or three 0.6 mm tantalum balls were placed in the cement at approximately the same height as the markers embedded in the proximal position of the composite femur (Fig. 2b).

The hip prosthesis model was kept at 37°C to mimic the conditions of the in vivo environment. A temperature sensor (T-35, Takigen MFG Co., Ltd, Tokyo, Japan) was attached to the epoxy resin used to fix the composite femur, so that the temperature from a heater (G6A92 240V, 250W; Takigen) was maintained at 37°C.

Methods and duration of loading

Loading was applied at an internal angle of 15° on the coronal plane to the metal femoral head (Co-Cr, 26 mm in diameter), attached to the stem neck.¹⁷ Sine wave load (1 Hz, 3000 N) was applied 500 000 times. This was estimated to be equivalent to the load of a six-month walk.^18,19 The test load of 3000 N was considered to be equivalent to the load to the hip joint of a 70 kg person in a standing position on one leg.¹⁸ The loading tests were conducted using a fatigue strength testing system (EHF-UM 300KN-70L; Shimadzu Corp., Kyoto, Japan). In this experiment, an eight-hour unloading period was applied between the 16-hour loading periods to mimic an actual sleeping period in clinical settings. Unloading periods were considered to be important for stress relaxation of the cement.

Parameters and methods of testing: stem subsidence

Measurement data were obtained over time using a 5 mm digital displacement gauge (DTH-A5; Kyowa Electronic Instruments Co. Ltd, Tokyo, Japan) placed at the stem shoulder (Fig. 3). Data were sent automatically to a personal computer by means of data collection and analysis software (Sensor Interface PCD-300A; Kyowa). Each set of data, measurements collected for eight minutes every hour automatically, was sorted into a file to collect 10 000 data sets, and the data were entered into the computer. A total of 216 files were extracted per loading test. The measurement data were adjusted and converted to units of distance. Each of the 16-hour loading and eight-hour unloading periods of each day were divided into the early, middle and late phases, and the first two files from each phase (a total of 20 000 data sets) were used. Subsidence was defined as the mean of the maximum values of sine waves from the 20 000 data sets of each phase of the loading period. The mean values of the 20 000 data sets of the unloading period were also collected. Perpendicular migration of the tantalum balls at the end of the stems was measured by micro-CT, and this was used as the final amount of subsidence.

Fig. 3

The testing instrument. A composite femur with stem was inserted into the testing instrument. A metal head was attached to the stem neck and was placed in contact with a loading bar (2). A digital displacement gauge (1) was placed at the stem shoulder to measure stem motion continuously during loading and unloading periods.

Parameters and methods of testing: micro-CT analysis

Before and after loading, 3D images of the areas within a 2 cm radius of the markers were obtained using a micro-CT (Microfocus X-ray CT System) scanner (TOSCANER-32250 µhd; Toshiba IT & Control Systems Corp., Tokyo, Japan) to measure the amount of migration for the tantalum balls within the cement and for the balls that were attached to the stem tip. Images were taken with a slice spacing of 95 µm. The maximum spatial resolution of the micro-CT scanner was 5 µm. In the cross-sectional CT images in which the tantalum balls were detected, the shortest distance of the line segment that intersected a ball between the bone and the stem was defined as “cement width.” These images were also used to measure the shortest distance between the centre point of the anterior, posterior, medial and lateral surfaces of the stem and the canal bone. The mean value for these four measurements was defined as the “mean cement thickness” for each stem (Fig. 4).

Fig. 4

A micro-CT slice showing a tantalum ball as a black ball in the cement. The white arrow shows cement width. The mean value of the four black arrows was defined as “mean cement thickness ”.

The distance of the migration of a tantalum ball can represent the distance of cement movement in a specific location. With this measuring method, the embedded markers in the composite femur were used as a reference point. The position of a tantalum ball before applying load was defined as the original point, and the 3D coordinate values after applying load were calculated. The horizontal distance of the outward or inward migration related to the femoral medulla was expressed as “+” and “−”, respectively. For the amount of perpendicular migration, data were obtained on the distance parallel to the axis of the bone. Movement in the direction of subsidence was expressed as “+”.

Statistical analysis

The correlation between stem subsidence and mean cement thickness was investigated using Pearson’s correlation coefficient. The relationship between the behaviour of tantalum balls and cement thickness was investigated using Spearman’s rank-correlation coefficient. A comparison of the movement of balls in two different thicknesses of cement was investigated using the Mann-Whitney U test. A significance level of 0.05 was used for these tests. Statistical analysis was performed using StatView version 5.0 (SAS Institute Inc., Cary, North Carolina).

Stem subsidence

A digital displacement gauge, placed at the stem shoulder during testing, showed that all stems subsided after cyclic loading (Fig. 5). This subsidence was partially reversed during the unloading periods, but increased gradually over time. The smaller stems tended to be associated with greater subsidence; each of the small stems subsided further than either of the large stems. The mean values and standard deviations (sd) for final subsidence, as measured at the balls attached to the tips of the stems, were 0.179 mm (sd 0.084) for the large stems, 0.375 mm (sd 0.091) for the medium stems, and 0.495 mm (sd 0.126) for the small stems.

Fig. 5

Graph showing stem subsidence. All stems subsided rapidly on the first day, after which subsidence continued gradually. All stems subsided down during loading periods and rose up during unloading periods. This stem motion was observed repeatedly on a daily basis during the testing of all stems.

Cement thickness

Mean cement thickness decreased as stem size increased: 1.52 mm (sd 0.45) for the large stems, 2.19 mm (sd 0.96) for the medium stems, and 2.57 mm (sd 1.61) for the small stems. A strong positive correlation was observed between mean cement thickness and final stem subsidence measured by the ball at the end of stem, with less subsidence associated with a thinner cement mantle. (Fig. 6).

Fig. 6

Graph showing the correlation between stem subsidence and cement thickness. This correlation diagram shows mean cement thickness and stem subsidence for each of six femurs. Pearson’s correlation coefficient was 0.94 (Y = 0.2469 x − 0.167, p ≤ 0.05).

Investigation of the amount and direction of cement creep

A total of 13 tantalum balls (three for one of the two small stems, and two each for the other stems) were visualised in the micro-CT images. All of the visualised tantalum balls were found to have moved in the direction of stem subsidence. The mean distance of ball subsidence was 0.301 mm (0.015 to 0.607, sd 0.231). Table I shows measured and calculated values for different ball sizes.

The mean ratio of ball subsidence:stem subsidence was 102.9% (97.8% to 113.5%, sd 6.4%) for small stems, 68.4% (19.6% to 97.3%, sd 36.7%) for medium stems, and 38.2% (8.0% to 92.4%, sd 40.6%) for large stems. For small stems, the balls subsided downwards to almost the same extent as the stems.

The mean horizontal movement was 0.038 mm (−0.082 to 0.290, sd 0.103).

Eight tantalum balls moved outwards, and five moved inwards. Inward movement was noted in three of five tantalum balls for the small stems, two of four for the medium stems, and zero of four for the large stems. The mean horizontal:downward ratio of ball movement was 66.3% (−14.4% to 286.7%, sd 102.2%) in all balls. This ratio for the large stems was 202% (146% to 287%), indicating effective radial cement creep. The mean ratio for the small and medium stems was −1.1% (−14.4% to 23.0%), indicating that the thicker cement tended not to move in a horizontal direction by stem subsidence. Spearman’s rank-correlation coefficient showed a negative correlation between the horizontal:downward ratio of ball movement and cement width (Spearman’s rank correlation coefficient ρ = - 0.70, p < 0.05); thin cement mantle was associated with a higher percentage of outward movement. When compared to the movement of balls in two different thicknesses of cement (≥ 2.5 mm and < 2.5 mm), we found a significant difference in the horizontal:outward ratio of movement between the two groups (p = 0.03); outward movement of the balls was observed significantly more frequently for the cement width of < 2.5 mm (Fig. 7).

Fig. 7

A diagram of the horizontal:downward ratio of ball movement and cement width. This diagram demonstrated that effective horizontal cement movement in relation to cement subsidence was seen for cement widths of less than 2.5 mm (dotted line).