Effect of incremental increase in radial neck height on coronoid and capitellar contact pressures

Abstract

Background

Over-lengthening of the radial neck has been shown to affect ulnohumeral kinematics and has been proposed to affect radiocapitellar pressures. We hypothesized that an incremental increase in radial neck height increases the capitellar contact pressure and reduces the coronoid contact pressure. Knowledge of the effects of over-lengthening is clinically important in preventing pain and degenerative changes due to overstuffing.

Methods

Six human cadaveric elbows were prepared on a custom-designed apparatus simulating muscle loads and passive flexion from 0° to 90° under gravity valgus torque while measuring joint contact pressures in this biomechanical study. Each elbow was tested sequentially starting with the intact specimen followed by insertion of a radial head prosthesis with 0, +2, and +4 mm of radial neck height, respectively.

Results

Capitellar mean contact pressures significantly increased after insertion of +2 and +4 mm radial head prostheses (p < 0.03). The capitellar mean contact pressure with a 0 mm radial head prosthesis was 97 KPa. Insertion of +2 mm and +4 mm radial heads increased mean contact pressures to 391 KPa (p = 0.001) and 619 KPa (p = 0.001), respectively, with 90° of elbow flexion.

Discussion

Increasing radial prosthesis height by 2 mm significantly increases capitellar contact pressures and reduces coronoid contact pressures.

Introduction

Radial head fractures account for 2–5% of all fractures and for 33% of all elbow fractures.^1–3 Many displaced radial head fractures require some form of surgical management. Although open reduction and internal fixation is a successful treatment option for simple fractures, it may be less reliable for comminuted or osteoporotic fractures due to a higher rate of complications.⁴

Radial head arthroplasty is a reliable choice to preserve axial and valgus stability and restore elbow kinematics to near normal levels.^5,6 However, when the radial head prosthesis is overstuffed, ulnohumeral joint kinematics are altered causing malalignment of the radiocapitellar joint. This malalignment ultimately leads to hyaline cartilage damage and osteoarthritis.^7–9

This study was conducted to test the hypothesis that incremental lengthening of radial neck height increases capitellar pressures and reduces coronoid pressures by unloading the coronoid. Knowledge of these effects is clinically important, as overstuffing increases the risk of pain and early degenerative changes.

Materials and methods

This study was performed with the approval of the institutional biospecimen subcommittee. All data are presented as the mean ± standard error of the mean.

Specimen preparation

Six male fresh-frozen cadaveric upper limbs were used for this study. There were three right and three left arms with a mean age of 80 ± 8 years. All specimens were thawed overnight for 12 h at room temperature prior to the experiment. Each specimen was then confirmed to have a normal range of motion (ROM). No specimen demonstrated a flexion contracture of more than 10°, a pronation-supination rotation arc less than 140°, or any radiologic evidence of arthritis or deformity under C-arm fluoroscopy.¹⁰ Each specimen was carefully dissected to remove the skin and subcutaneous fat from the mid-humerus to 5 cm distal of the elbow joint. The biceps, brachialis, and triceps muscle bellies were removed, while their tendon insertions were preserved and prepared with locking Krackow stitches using 36-kg-test braided polyester fishing line.¹⁰ The humeral origins of the flexor-pronator and the extensor-supinator muscles were preserved. To permit placement of the pressure transducer, the anterior capsule was excised, taking care not to injure the collateral or annular ligaments. Any specimen with cartilage erosion to the subchondral bone was excluded, but we did not discard specimens exhibiting shallow erosion with fibrillation and fissuring with normal joint contact.¹⁰ Any specimen with ligament insufficiency was excluded using the posterolateral rotatory drawer test or the direct observation of ligaments. The proximal humeral end of the specimen was then potted into a cylindrical metal sleeve in parallel to its long-axis using a polyurethane resin (Smooth-Cast 65D, Techno-Industrial Products, Inc., Hartland, WI 53029-8302) to fix and load the specimen onto the testing machine. A lateral column humeral osteotomy was made and a 2 mm groove was created in the bare area of olecranon. A Tekscan 5051 thin-film pressure transducer (Tekscan, South Boston, MA) was inserted, from anterior to posterior after anterior capsule excision, until the end of the sensor reached the olecranon bare area groove and covered both the ulnohumeral and the radio-capitellar articulations.

The lateral column osteotomy was fixed with three screws and washers (1 Arbeitsgemeinschaft für Osteosynthesefragen (AO) 3.5 mm cancellous for intercondylar fixation and two 3.5 mm AO cortical screws for distal humerus fixation).

Pressure transducer

The thin-film Tekscan sensor has been validated for measuring pressure in rounded contact areas¹¹ and has been used in earlier reports of joint contact pressures,^12,13 specifically including use within the elbow.^10,14–17 Each 5051 sensor has one 56 × 56 mm matrix (3136 mm²), comprising of 1936 sensels (individual detection units of pressure) located on conductive ink grids. The Tekscan 5051 thin-film pressure transducers (Tekscan, South Boston, MA) with a saturation pressure of 8.3 MPa (84.4 kg/cm²) were prepared, preconditioned, and calibrated according to the manufacturer’s recommendations. The calibration was performed with the Tekscan I-Scan software using a customized pressurized air-piston/load cell system to apply eight sequential loads to the sensor, while it was sandwiched between two layers of 1.6 mm rubber membrane, which was in turn sandwiched between two polished aluminum plates. The calibration loads ranged from 690 to 5520 KPa (7 to 56 kg/cm²) and were applied in 690 KPa (7 kg/cm²) increments. Since it is recommended that sensors are calibrated under conditions that mimic those encountered during testing,¹⁸ a rubber membrane-aluminum block calibration construct was chosen to mimic the cartilage-subchondral bone conditions of the elbow joint. The Tekscan sensors were inserted between the joint surfaces from anterior to posterior as previously reported.^10,15–17 The osteotomized lateral humeral condyle was fixed with three screws as stated above and the sensor was secured in place by fixing to two proximal screws in the proximal posterior aspect of ulna. The Tekscan contact pressure data were captured at a frequency of 100 Hz.

Specimen mounting and testing

The specimen was mounted on a custom-made machine designed to test the elbow while it was passively flexed from 0° to 90° at 90° degree of humeral external rotation.² This places the medial epicondyle upward and the transcondylar axis perpendicular to the floor for gravity valgus loading (Figure 1). The biceps, brachialis, and triceps were connected to Airpel™ pneumatic pistons (Airpot, Corp. Norwalk, CT 06851) to simulate muscle loads intended to provide dynamic joint stability. As with previous studies using this biomechanical testing model, force was applied in a 1:1:2 ratio with the brachialis, biceps, and triceps receiving 25 N, 25 N, and 50 N, respectively.^17,19 The distance of each pulley from the joint line and humeral axis was set to simulate the physiologic position of the tendons²⁰ 5.5 cm proximal to the joint line. The brachialis, biceps, and triceps pulleys were set at 2, 3.5, and 2 cm away from the humeral axis, respectively. The elbow was passively flexed by pulling a braided polyester line perpendicular to the forearm over the ROM to avoid application of stabilizing or distracting forces created by manual manipulation of the forearm directly.²¹ The angle was measured and the data were collected using a pulley (white pulley in Figure 1) connected to a potentiometer. The voltages measured by the potentiometer were converted into angle values and captured at 100 Hz using a custom LabView Virtual Instrument program (VI) (National Instruments Corporation, Austin, TX 78759-3504).²²

Figure 1.

Gravity valgus testing system. Schematic illustration of the custom-made machine used to simulate muscle loads and to control humeral rotation. Dynamic joint stability was achieved by connecting the biceps, brachialis, and triceps to pneumatic pistons to simulate muscle loads. The forearm was passively flexed from 0° to 90° under valgus load, while contact pressure data was collected with a thin film pressure transducer within the elbow joint.

Testing protocol

To reduce the friction between the joint surface and the sensor, 2 ml of mineral oil (Sigma-Aldrich, St. Louis, MO 63103) were applied to the joint after the sensor insertion. The articular surfaces and the tendons were frequently moistened during testing with normal saline. The elbow was tested from 0° to 90° of flexion at 90° of humeral external rotation (gravity valgus). A radial neck cut was made at a distance of 9 mm relative to the radial head dish. Then, a radial head prosthesis (Acumed) of 0, 2, and 4 mm neck length with the head size as per the recommendations of the manufacturer were inserted for each step of the experiment.

Data acquisition

Data were recorded using the Tekscan software (I-Scan). At the beginning of each set of tests (INTACT, 0 mm RADIAL HEAD, +2 mm RADIAL HEAD and +4 mm RADIAL HEAD), the positions of the lateral margin of the radial head, the anterior margin of the radial head, the radial notch, and the outline of the coronoid were registered using the Tekscan sensor and a blunt probe. Based on landmarks recorded at the beginning of each set of tests, the coronoid surface was theoretically divided into quadrants: anterior and posterior (i.e. “base” of coronoid) with respect to a line perpendicular to the tangent of the medial edge of the radial head passing through the radio-ulnar notch, and medial and lateral with respect to the central trochlear ridge of the coronoid.¹⁰

Mean contact pressure data between the coronoid and the trochlea which we referred to as “coronoid pressure” as well as radiocapitellar pressure which we have referred to as “capitellar pressure” were collected. All data were filtered using a fourth order, low-pass Butterworth filter with a 50 Hz cut-off frequency. Data were then down-sampled using spline interpolation.¹⁰ This involved generating a curve from the thousands of data points in each test. From that curve, discrete angle values from 0° to 90° in one degree integer increments were determined, which was necessary to perform discrete data analysis. To obtain an average curve for each group, the contact pressures corresponding to each discrete angle from all of the specimens in that group were similarly analyzed.¹⁰

To ensure that the down-sampling process did not significantly affect the outcome measures, the R² correlation of the filtered and down-sampled data was confirmed to be ≥ .98. The data used for the comparisons were chosen at eight different points from 15° to 85° of flexion in 10° increments.

Statistical analysis

All data are presented as the mean ± standard error. To determine if radial neck status (i.e. INTACT, 0 mm RADIAL HEAD, + 2 mm RADIAL HEAD and + 4 mm RADIAL HEAD) had a significant effect on mean contact pressures in the whole coronoid, lateral coronoid, and radiocapitellar articulation, these mean contact data were analyzed using Kruskal-Wallis testing where appropriate. Pairwise statistical comparisons of the mean contact pressures between the aforementioned experimental groups were modelled using Wilcoxon Signed Rank tests (p-values < 0.05 were considered to be significant). Bonferroni p-value adjustments were made where appropriate. Power analysis using the data from a previous published study¹⁹ revealed that using the aforementioned Wilcoxon Ranked Sign test comparisons, with six samples compared at eight flexion angles, we had an 80% chance of detecting a significant difference (p < 0.05) in mean contact pressures of at least 43 kPa in the whole coronoid, 74 kPa in the lateral coronoid, and 56 kPa in the radiocapitellar joint.

Results

In the gravity valgus position, lengthening of the radial neck resulted in a significant increase in mean capitellar contact pressures (p < 0.0001). It also had a significant effect on the mean contact pressures seen in the coronoid (p < 0.0001).

All the capitellar mean contact pressures measured with the radial head prostheses were significantly different than those seen with the intact radial head (p = 0.0006). Initially, replacement of the radial head with a + 0 mm neck reduced capitellar mean contact pressures by 63 kPa ± 13 kPa (23% ± 5%) compared to those in the Intact group (p < 0.0006). Thereafter, radiocapitellar contact pressures increased with incremental increases in radial neck length. The +2 mm radial neck group had significantly higher capitellar mean contact pressures compared to the 0 mm radial neck group (p = 0.0006). The mean increase in the capitellar contact pressures following insertion of the +2 mm radial neck was 235 kPa ± 41 kPa (222% ±38%) compared to the 0 mm radial neck group. The +4 mm mean contact pressures were significantly higher than those seen in the +2 mm (p < 0.002) and 0 mm radial neck groups (p = 0.0006) (Figure 2(a)). The mean increase in the capitellar contact pressures following insertion of the +4 mm radial neck was 474 kPa ± 29 kPa (323% ± 26%) and 239 kPa ± 55 kPa (95% ± 18%) compared to the 0 mm and + 2 mm radial neck group, respectively. The +2 mm and +4 mm groups had mean capitellar contact pressures that were, on average, 173 kPa ± 35 kPa (95% ± 18%) and 412 kPa ± 31 kPa (196% ± 19%) higher, respectively, than in the Intact group.

Figure 2.

Mean contact pressures. (a) Capitellar, (b) whole coronoid: Under gravity valgus, capitellar contact pressures were significantly higher in the +4 mm radial neck group compared to the + 2 mm (p = 0.002) and 0 mm (p = 0.0006) groups. The mean decrease in the whole coronoid contact pressures following insertion of the+ 2 mm radial neck was statistically significant versus the 0 mm group (p < 0.002). Whole coronoid contact pressures reduced even further by insertion of the +4 mm radial neck (p = 0.0006). There was no significant difference in the mean whole coronoid pressures seen between the 0 mm and the Intact groups (p ≈ 1).

Lengthening the radial neck also reduced contact pressures on the coronoid (p < 0.0001, Figure 2(b)). The mean decrease in the coronoid contact pressures following insertion of the + 2 mm radial neck was 57 kPa ± 13 kPa (18% ± 5%) versus the 0 mm group (p < 0.002). Mean coronoid contact pressures reduced even further by insertion of the +4 mm radial neck (p < 0.0006). The decrease in the mean coronoid contact pressures in +4 mm radial neck group compared to 0 mm averaged 155 kPa ± 16 kPa (49% ± 4%). Insertion of the +4 mm radial neck also reduced contact pressures by 98 kPa ± 11 kPa (36% ± 4%) compared to the +2 mm radial neck group. There was no significant difference in the mean coronoid pressures seen in between the 0 mm and the Intact groups (p ≈ 1).

Discussion

This study showed that under valgus loading, an incremental increase in radial neck height significantly increases mean capitellar contact pressures and reduces mean coronoid contact pressures by unloading the coronoid. We were able to detect a significant difference in capitellar contact pressure with 2 mm increments of radial neck height, which shifted the load from the ulnohumeral joint to the radiocapitellar joint.

These findings may explain the effects of lengthening (overstuffing) and its influence on ulnohumeral kinematics as described by Van Glabbeek et al.⁹ They showed over-lengthening the radial head height by 2 mm or more forced the ulna into varus and an externally rotated position even with the application of valgus force. Radial neck lengthening causes the ulna motion to track in a significantly different pathway compared to the normal radial neck length.

Lanting et al.²³ also explored capitellar contact pressures with different radial head implant heights. Their study used absolute values of contact pressures due to the confined space of the radiocapitellar joint deforming the thin flexible transducers at the periphery. Therefore, they used intraspecimen standardization. As such, their study did not find a significant increase in capitellar contact pressure when the length of the radial head increased or decreased by 2 mm from optimal radial length. This was attributed to the effect of transducer positioning or sensitivity variability on the measurement of contact pressure. This is different from our study in which the methodology was sensitive enough to find significant differences in contact pressures following a 2 mm increase in radial neck height.

Another study by van Glabbeek et al.⁸ measured capitellar contact pressure after transection of the medial collateral ligament with a custom-made radial head prosthesis. Their study showed that over-lengthening by 2.5 mm affected ulnohumeral kinematics and radiocapitellar pressure. They also reported the irreversible deformation of the sensor that made additional pressure measurements impossible by 5 mm of lengthening of radial neck. As with the study by Lanting et al. above, the Van Glabbeek et al.⁸ study had some difficulties with deformation of the sensors and subsequent data collection. Since we used a distal humeral osteotomy, we were able to insert the Tekscan sensor inside the joint, thus avoiding significant wrinkling and deformation which allowed us to collect consistent data.

The present study has some limitations. Due to the passive motion of the elbow joint during this experiment, it was not possible to know the effects of active muscle contraction and dynamic loading. Simulated active elbow motion using dynamically controlled actuators has been used by other investigators, but they did not show significant differences in repeatability between passive and active flexion.²⁴ The anterior capsule was excised; however, it has been shown to not have a significant role in varus/valgus stability.^25,26 Due to the inherent limitations of the Tekscan sensor, the absolute values of contact pressure should be taken with caution. It is recommended to calibrate sensors frequently and in a way as similar to the final testing condition as possible.¹⁸ Although sensor calibration was done using our standardized machine calibration method,^10,15,16,19 we were not able to perform curved surface calibrations to perfectly match the contour of each elbow joint. However, we observed significant differences with relatively low variability making us feel confident in the reliability of our data. Although we did not observe any macroscopic movements across the lateral column osteotomy, we were not able to detect microscopic movements, which might have had an effect on contact area and pressure data. Since all specimens were osteotomized and prepared for testing in the same manner, we believe this limitation did not significantly affect our results. Any clinical applications of this in vitro biomechanical study should be considered with caution.

Increasing radial head implant length as little as 2 mm significantly increases the capitellar and reduces coronoid contact pressures. Therefore, restoration of anatomic radial head length is critical when performing radial head arthroplasty.