|Determination of Radiographic Guidelines for Percutaneous Fixation of Proximal Humerus Fractures Using a Cadaveric Model|
cadaveric shoulders – humeral heads underwent fluoroscopic evaluation with the head divided into three zones on both AP and axillary views creating 9 zones. 5 AP and 3 axillary fluoroscopic images in different rotational positions were assessed for pin penetration. All images were evaluated for pin penetration and the AP view was evaluated for lesser tuberosity location. Pins placed appropriately below the subchondral bone did not appear to penetrate the joint on any fluoroscopic image. Pins placed 2 mm beyond the articular surface were appropriately viewed exiting the head on most views (64%) but falsely appeared within the head on several others (36%). Pins perforating the posterior head were problematic for accurate detection on AP views (missed in 87%), but this was avoided by externally rotating the humerus to 60°. Articular penetration cannot always be appreciated radiographically and special efforts are necessary to avoid this problem including the use of various rotational views as well as the use of appropriate landmarks for orientation such as the lesser tuberosity position.
Several studies have shown that closed reduction and percutaneous fixation of proximal humerus fractures can be an effective technique. Advantages of this technique include its minimal invasiveness, maintenance of fracture site vascularity, and high rate of reported successful outcome1-5 With these positive attributes, the indications for performing proximal humerus fractures have increased to include two-part surgical neck or greater tuberosity fractures, three-part fractures, and valgus-impacted four-part fractures.
The purpose of this study was to develop safe radiographic and surgical guidelines for avoiding articular penetration while accurately placing guide wires into the proximal humerus for fracture fixation.
Materials and Methods
One fixed whole-body cadaver was used for this study. The right shoulder was used with the proximal humerus divided into nine zones. Anteroposterior (AP) and axillary fluoroscopic images were obtained on each specimen. The head was divided into three zones on each of these images. On the AP view these zones were referred to as the inferior, middle, and superior zones, while on the axillary view these were referred to as the anterior, neutral, and posterior zones. Finally, the intersection point, defined as the point where the lateral humeral shaft and the lateral metaphyseal proximal humerus flare intersect, was determined. The intersection point was considered the distal extent of the surgical neck (Figure 1).
With the humeral head fluoroscopically sectioned, a pin was placed into the middle zone on the AP view. The pin was placed on the anterior third of the lateral humeral shaft by walking the pin anteriorly and posteriorly to determine the location of this axis. The guide wire was driven into the middle zone on the AP view and neutral zone on the axillary view. The distance below the intersection point was measured on the skin surface. Next, the proximal humerus was stripped of its soft tissue and the angulation between the shaft and the guide wire was measured (Figure 2) and varied to determine the arc by which the guide wire would remain in the middle zone.
With the C-arm redirected to allow axillary imaging of the humeral head, the location of the guide wires already placed on the AP view was evaluated. If the guide wire was placed into the neutral zone, the angle between guide wire and forearm was determined. If the pin was not within the neutral zone, the pin was redirected using the same starting point on the anterior third of the humeral shaft. Once positioned within the neutral zone, the angle between the forearm and guide pin were determined. As performed on the AP images, the guide wire was again taken through an arc of angulation to determine the parameters by which the guide wires could be maintained within the neutral zone of the humeral head.
With the starting point marked and the proximal humeral shaft stripped, the anterior capsule of the glenohumeral joint was exposed. A capsulotomy was performed, and the humeral head was dislocated. The articular surface of the humeral head was divided into nine zones (Figure 3A) with three equidistant lines drawn horizontally and three equidistant lines drawn vertically.
Using the common starting point and an anterior cruciate ligament (ACL) guide (Figure 3B), a threaded guide wire was placed under direct visualization into each of the nine zones. Using the same track that placed the guide wire into the center of the humeral head, the original pin was driven beyond the central portion of the articular surface by 2 mm with the head dislocated for direct visualization (Figure 3C). The humeral head was relocated and the pin position was evaluated by fluoroscopic imaging.
Five AP views were obtained, three by rotating the humerus and two by angling the C-arm. The three humerus positions were neutral, 30° of internal rotation, and 30° of external rotation. With the arm held in a neutral position, the C-arm was directed 15° caudad and cephalad completing the five AP views (Figures 4A-4C). Three different axillary views were obtained with the C-arm placed horizontal to the ground as well as directed 15° anterior and 15° posterior (Figures 4E-4F). Each fluoroscopic image was printed for later evaluation of guide wire position.
Once these eight different fluoroscopic images were obtained, the humeral head was re-dislocated. The guide wire was pulled back to a position 2 mm below the articular surface measured by placing the blunt end of a second wire into the hole created by the guide wire’s articular penetration. The head was relocated and the same eight fluoroscopic views were again obtained for evaluation. In the AP plane, the angle between the shaft and guide wire was measured (Figure 2). In the axillary plane, this angle was again measured between the guide wire and forearm to determine retroversion.
The first guide wire was completely removed from the humerus and the head was re-dislocated. Using the common starting point and the ACL guide, the guide wire was re-directed into one of the other nine zones 2 mm beyond the articular cartilage. With the head relocated, the eight fluoroscopic images were obtained. The head was dislocated and the guide wire was again backed out until it was 2 mm below the articular surface, and the eight fluoroscopic images were repeated. After the images were completed the coronal angle between the shaft and the guide wire as well as the version between the forearm and guide wire were measured. The process was repeated for the remaining seven zones of the humeral head.
Guide wires placed through the anterior portion of the humeral head (three anterior zones) were detected in 16 (67%) of 24 views. Guide wires placed 2 mm beyond the posterior articular surface, however, were appreciated in only 2 (9%) of 24 views (Figure 5). This difference was statistically significant (P<.05). Guide wires placed through the anterior articular surface were best visualized on AP views in internal rotation; and there were no unrecognized anterior penetrations using that view. However, pins placed posteriorly were missed on 66% of the views taken with the humerus 30° externally rotated. Humeral shaft external rotation to 60° (Figure 5D) resulted in 100% posterior pin penetration detection.
The position of the lesser tuberosity was noted to track from medial (internally rotated view) to lateral (externally rotated view) within the humeral head, which also aided in ensuring that adequate rotation was being performed (Figure 5). Guide wires placed in the middle zones (detection failure=9/15) were less accurately visualized than superior (detection failure=5/15) or inferior guide wires (detection failure=5/15) on AP views although this difference was not statistically significant. This trend did not seem to be affected by rotating the fluoroscopic beam either caudad or cephalad (Figure 4A-4C).
Although the axillary images failed to recognize guide wire penetration in 7 (30%) of 24 views, a recognized pattern was not appreciated as seen on the AP views. However, rotating the beam 15° anteriorly or posteriorly (30° arc) (Figures 4D-4F) allowed 100% guide wire penetration detection. Anteriorly directed axillary views were more accurate at detecting superior than inferior guide wire tip penetration (100% versus 33%), while posteriorly directed axillary views were more accurate in evaluating inferior than superior guide wire penetration (66% versus 33%) (Figure 6).
In assessing for guide wires that were not detected as penetrating the articular cartilage on at least one AP view and one axillary view (as may be the case clinically), three pin positions were missed. These guide wire positions were anterosuperior, posterosuperior, and posteroinferior. However, all guide wire positions were detected on at least one AP view or one axillary view when considering all rotational views.
All pins placed 2 mm below the articular surface did not penetrate the humeral head on any radiographic view. However, a significant proportion of the views obtained of guide wires placed 2 mm beyond the articular surface falsely did not demonstrate joint penetration (26 of 72).
The site of skin entry averaged 6 cm (range: 5.1-7.2 cm) below the intersection point with an average angle of 30° (range: 24°-37°) for placing the pin into the middle zone of the humeral head. The actual entry point into the humeral shaft averaged 2.9 cm (range: 1.7-4.1 cm) distal to the intersection point with the remaining distance necessary for directing the pin through the soft tissue. A 30° arc maintained the guide wire within the middle zone. The average retroversion measured between the forearm and guide wire when placed in the neutral zone on the axillary view was 29° (range: 19°-36°). An arc of 26° maintained the guide wire within the neutral zone on the axillary views.
Although both hip fracture and proximal humerus fracture literature report significant advantage in terms of initial stability in placing the fixation device within the subchondral bone and spreading the fixation as far apart as possible (thus using all zones),9,10,12-18 these both increase the risk of inadvertent articular penetration. Since a significant risk of articular penetration has been reported when placing fixation for either proximal femur or humerus fractures,6,13,14 the main goal of this study is to provide specific radiographic and surgical guidelines to avoid this problem.
Posterior guide wire penetration is especially difficult to appreciate on the AP view. Therefore, posterior guide wires should be evaluated in >60° of external rotation (Figure 5D), as all penetrating posterior pins were appreciated on this view. This occurs due to the relative retroversion of the humeral head, and this factor can be eliminated by externally rotating the humeral shaft. The same phenomena has been reported in the hip literature in which pins placed in the anterior cortex are especially prone to inadvertent pin penetration.13 We believe this is due to the relative anteversion of the femoral neck.
Because in the true clinical situation, rotation of the humeral head relative to the shaft, and therefore forearm, may be difficult to quantify, the position of the lesser tuberosity was assessed in various positions of rotation (Figure 5). The lesser tuberosity was noted to move laterally with external rotation and was noted to lie on the lateral cortex on the 60° external rotation views (Figure 5D). Therefore, clinically the lesser tuberosity should be seen along the lateral cortex to ensure adequate imaging has been performed and posterior guide wire penetration has not occurred.
A limitation of this study is that nonfractured humeri were used. During the reduction of a true fracture, an unknown degree of angulation (retroversion or anteversion) of the surgical neck may make precise radiographic views (eg, AP in 60° external rotation) difficult. Appreciation of the lesser tuberosity position may be helpful.
Several recent studies have suggested placing an additional guide wire from the anterior cortex to the posterior aspect of the humeral head to aid in stability.6 Although not evaluated in our study, these guide wires, which intentionally place the tip in the posterior aspect of the humeral head, would seem especially prone to penetrate the articular cartilage without radiographic detection. Guide wires placed from proximal to distal through the greater tuberosity also have been suggested for improving stability that would eliminate the risk of inadvertent articular penetration,9 but would increase the risk of axillary nerve injury6 and may make rehabilitation more difficult with a guide wire lying just under the acromion.
When evaluating guide wire placement using the axillary views, none of the specific guide wire locations were noted to be problematic. However, a high incidence (35%) of articular penetration was not appreciated on the various views. Our purpose in evaluating axillary views in different positions of rotation is that this view can be technically difficult to obtain in the operating room especially in large patients. Therefore, the surgeon often accommodates by rotating the C-arm into various configurations until an axillary image is obtained. We sought to determine the effect of rotating the C-arm (Figure 4) and did not find that any one orientation was most accurate. Importantly, when the C-arm was rotated through a 30° arc, pin penetration detection was 100%.
Because the C-arm oftentimes is inadvertently rotated cephalad and caudad relative to the patient sitting in a slight beach-chair position, views also were obtained by rotating the C-arm slightly cephalad or caudad (Figure 4) to determine whether the accuracy was affected. No one view was found to be less accurate than the others; however, guide wire positions were found that appeared within the humeral head on one view and out on another (Figure 7). Guide wires placed in the middle of the head on the AP view were at increased risk of penetrating the articular cartilage without fluoroscopic detection, which seems rational, as pins placed at the poles are less likely to have enough overlapping cortex to avoid accurate detection.
Although not a primary goal of this study, we found the neck-shaft angle and humeral head-shaft retroversion to be similar to accepted current values. By using these angles as well as the arcs of acceptable angulation reported, ideally multiple guide wire passage and pin penetration may be avoided.
Percutaneous fixation for proximal humeral shaft fractures has become a commonly performed procedure with positive clinical results. Developing appropriate guidelines for safely and effectively placing these guide wires is an appropriate next step as the use of this technique is broadened. The following study has provided guidelines to avoid articular surface pin penetration while using a percutaneous pinning method.