OTA 2004 Posters
The Use of Locking Small-Fragment Plates for Treatment of Humeral Shaft Fractures
Purpose: In this retrospective study, we compared the healing and complication rates of humeral shaft fractures plated with 4.5-mm low-contact dynamic compression (LCDC) plates, 3.5-mm LCDC plates, and 3.5-mm locking compression plates (LCP).
Methods: Chart and radiographic review was performed of consecutive patients with humeral shaft fractures treated by two surgeons with open reduction internal fixation at two level I trauma centers from August 2000 to December 2003. A total of 123 humeral shaft fractures in 117 patients were identified and placed in one of three groups based upon type of implant used. Twenty-four patients were excluded because of lack of follow-up after discharge. Six poly trauma patients died during their hospital admission, and one patient had an amputation as a result of a failed vascular repair. An additional 10 patients were excluded because their internal fixation did not fit into the three groups described (locking 4.5-mm, dual 3.5-mm plates, or a 4.5-mm LCDC/recon custom hybrid plate). The remaining 76 patients (78 fractures) had clinical and radiographic follow-up to union.
Fractures were classified according to the OTA classification scheme. Preoperative data collected included mechanism of injury, associated injuries, time to surgery, open fracture type, and presence of preoperative radial nerve palsy. Postoperative data included type and length of plate used and surgical approach. Follow-up data included time to healing, presence of postoperative radial nerve palsy, weight-bearing status on the injured extremity, time to resolution of radial nerve palsy, and complications. Group 1 consisted of 37 fractures in 46 patients treated with 4.5-mm LCDC plates. Group 2 included 20 fractures in 20 patients treated with 3.5-mm LCDC plates. Group 3 includes 21 fractures in 21 patients treated with 3.5-mm locking plates. One patient with bilateral fractures was included in both groups 2 and 3 because one fracture was treated with a locked 3.5-mm plate and the other was treated with an unlocked plate.
Results: The weight-bearing status was determined on the basis of fracture configuration not the type of implant used. The fractures in group 1 united at an average of 14.8 weeks, those in group 2, at 13.2 weeks; and in group 3, at 10.6 weeks. There were four hardware failures overall, two in group 1 and two in group 2. There were three nonunions in group 1, two of which were infected.
There were a total of 32 cases of radial nerve palsy (26%); 29 (23.6%) were present preoperatively and 3 (2.4%) postoperatively. All radial nerve palsies occurring in the setting of an intact nerve resolved at an average of 19 weeks for those present preoperatively (with the exception of one patient, requiring tendon transfers at 1 year) and 6 weeks for those that occurred postoperatively.
Conclusion/Significance: Plating of humeral shaft fractures has traditionally been performed with 4.5-mm dynamic compression plates, broad or narrow, with good results, and this method is treated as our control group in this study. The smaller 3.5-mm plate offers the theoretical advantage of more holes per unit length, especially useful in complex comminuted fractures or periarticular fractures, particularly of the distal third of the humeral shaft. They are generally more easily contoured and are more easily positioned on the humerus. Decreased need for dissection resulting in fewer traumas to the radial nerve may also be an advantage. The advent of locked-plating technology makes these implants even more attractive for routine use in fixation of the humeral shaft. In our series, no failures occurred in the locked-plating group, even with weight-bearing. Although the larger plates remain the likely implant of choice for nonunion of the humerus, we believe that these results support the use of 3.5-mm plates, particularly the locking-compression plates, for the routine operative treatment of humeral shaft fractures.
The relatively high incidence of preoperative radial nerve palsy in this series reflects the high-energy fracture patterns seen at today's level I trauma centers. The longer time to recovery for the preoperative palsies compared with those that developed postoperatively is also indicative of this higher-energy fracture population. Although higher than reported in previous studies, we believe that our numbers more accurately reflect the true incidence of radial nerve palsy in patients who have sustained high-energy humeral shaft fractures.