The efficacy of statically locked intramedullary nails inserted after reaming for the treatment of fractures of the femoral shaft has been well established2,21,22. In comminuted fractures, the locking screws bear the transmitted load until the fracture has consolidated. Several authors have reported hardware failure, either through the distal screw-holes in the nail or of the distal locking screws, when fracture union was delayed or when full weight-bearing was initiated before callus formation3,4,13,17,20. The standard protocol for postoperative rehabilitation at our institution calls for protected weight-bearing until evidence of early healing is noted on radiographs. These restrictions on weight-bearing are particularly confining to patients who have bilateral injuries and to those who have a fracture of the contralateral extremity or of an upper extremity that would preclude weight-bearing or walking with the use of crutches. Such restrictions would necessitate bed-to-chair transfers for a minimum of four to six weeks while the patient waited for early consolidation of the fracture, making self-care difficult and delaying rehabilitation and discharge from the hospital.
Currently, the level of loading that leads to failure of the hardware in a fractured femur is unknown. Improvements in the design of the nails, including the elimination of welds and an increase in the amount of material in cross section, may have greatly improved the fatigue characteristics of these locking devices, distinguishing them from their predecessors. Failure of locking screws or nails used to stabilize fractures of the femoral shaft has been rare in our experience. Our clinical suspicion was that intramedullary nail systems of modern design could safely withstand immediate physiological loading without failure if the fracture united in a timely manner. Therefore, this two-part study was designed to determine the fatigue strength of eleven different configurations of locking nails and to use the data to form the basis of a prospective clinical study of immediate weight-bearing after treatment of fractures of the femoral shaft with a statically locked intramedullary nail.
Part I: Biomechanical Testing
Materials and Methods
We simulated segmentally comminuted midisthmal fractures using two sections of polyvinyl chloride pipe that were sixty-four millimeters long and thirty-two millimeters in diameter. Because different manufacturers use locking screws of different diameters, there is no standard size of drill for insertion of the screws. We tested devices of different sizes from multiple manufacturers, inserting each locking screw according to the technique suggested by the manufacturer. A segmental gap (average, twenty-seven centimeters) between the sections of the pipe was used to mimic a comminuted fracture in the middle of the femoral shaft. Plastic, washer-like spacers were inserted in the sections of the pipe to keep the nail centered.
We stabilized the simulated fractures with eleven different statically locked femoral nail constructs. These constructs included (1) RT-1, a 12.0-millimeter-diameter Russell-Taylor nail (Smith and Nephew Richards, Memphis, Tennessee) with one oblique 6.4-millimeter proximal locking screw (core diameter, 4.8 millimeters) and one 6.4-millimeter distal locking screw; (2) RT-2, a 12.0-millimeter-diameter Russell-Taylor nail with one oblique 6.4-millimeter proximal locking screw and two 6.4-millimeter distal locking screws; (3) RT-10, a 10.0-millimeter-diameter Russell-Taylor Delta nail with one oblique 5.0-millimeter proximal locking screw (core diameter, 4.0 millimeters) and two 5.0-millimeter distal locking screws; (4) S-O, a 12.0-millimeter-diameter Synthes nail (Paoli, Pennsylvania) with one oblique 4.9-millimeter proximal locking screw (core diameter, 4.3 millimeters) through a nail cap and two 4.9-millimeter distal locking screws; (5) S-T, a 12.0-millimeter-diameter Synthes nail with one transverse 4.9-millimeter proximal locking screw and two 4.9-millimeter distal locking screws; (6) S-10, a 10.0-millimeter-diameter Synthes nail with one transverse 4.9-millimeter proximal locking screw and two 4.9-millimeter distal locking screws; (7) Z-2, a 12.0-millimeter-diameter Zimmer nail (Warsaw, Indiana) with one oblique 5.5-millimeter proximal locking screw (core diameter, 5.0 millimeters) and two 5.5-millimeter distal locking screws; (8) Z-1, a 12.0-millimeter-diameter Zimmer nail with one oblique 5.5-millimeter proximal locking screw and one 5.5-millimeter distal locking screw; (9) Z-10, a 10.0-millimeter-diameter Zimmer nail with one oblique 5.5-millimeter proximal locking screw and two 4.2-millimeter distal locking screws (core diameter, 3.7 millimeters); (10) H-12, a 12.0-millimeter-diameter Alta nail (Howmedica, Rutherford, New Jersey) with two 5.0-millimeter proximal locking screws (core diameter, 4.3 millimeters) and two 5.0-millimeter distal locking screws; and (11) H-10, a 10.0-millimeter-diameter Alta nail with two 5.0-millimeter proximal locking screws and two 5.0-millimeter distal locking screws. Fifteen nails were tested for each of the eleven different configurations. All nails were forty centimeters long, and all screws were fully threaded. No nail or screw was used for more than one test.
Each construct was mounted in a specially designed fixture (Fig. 1) on a materials testing machine (Instron, Canton, Massachusetts). The fixture accommodated the curvature of the nails to permit only axial loading of the construct. Each construct was cyclically loaded in compression at eight hertz for 500,000 cycles. If a construct experienced 500,000 cycles without failure, it was said to have run out; that is, it had exceeded its anticipated service life. The fatigue-test protocol and data analyses were conducted according to the so-called staircase method with a step size of 222 newtons5. With this method, all tests resulting in a run-out are followed by a test of a new specimen at a higher load (increased by one step size) and all failures are followed by a test at a lower load (decreased by one step size); this process effectively brackets the mean fatigue strength. Therefore, approximately one-half of the tests resulted in failure of the hardware, whereas the other half ran out with the construct still intact. In the current study, fatigue strength refers to the mean load at which fatigue occurred in the construct at 500,000 cycles. We considered a construct to have failed if any of the locking screws or nails fractured before the end of the test. The Student t test was used to determine significant differences in mean fatigue loads among the eleven nail configurations. Significance was set at p < 0.05.
The 500,000-cycle run-out limit was based on unproved estimates of normal walking. Unpublished data18, as well as our own pedometer readings, indicate that each lower limb of an active adult who is physiologically normal is loaded approximately 5000 to 7000 times per day. This approximates 50,000 cycles per week and 500,000 cycles over ten weeks, which is the time required for callus to form and the fracture site to begin sharing a substantial portion of the load16.
All three Russell-Taylor constructs (RT-1, RT-2, and RT-10) and two of the Zimmer constructs (Z-1 and Z-10) failed as a result of fatigue fracture of the distal locking screw or screws at the junction with the intramedullary nail. All but one of the H-12 constructs failed through one or both distal locking screws. In contrast, all failures of the Z-2 constructs occurred through the proximal locking screw. Similarly, we noted a preponderance of failures through the proximal screw of the three Synthes constructs and the H-10 construct. All failures of the S-T constructs occurred through the proximal screw. Of the eight failures of S-O constructs, seven were through the proximal screw; five of the seven also exhibited failure of the distal screw or screws. We were unable to determine whether the failure occurred first in the proximal screw or the distal screw or screws. All failures of the S-10 constructs occurred through the transverse proximal locking screw. During testing, a distal screw on each of two Z-10 constructs backed out. Data from these nails were discarded, and two replacement specimens were tested.
The mean fatigue strengths (and standard deviation) of the RT-2 and Z-2 constructs were 2171 ± 107 newtons and 2113 ± 58 newtons, respectively (Fig. 2). None of these constructs failed at loads of less than 2002 newtons. These two constructs demonstrated significantly higher fatigue strength than all other constructs tested (p < 0.001), but they were not significantly different from each other. The mean fatigue strengths of the H-12 and Z-1 constructs were 1721 ± 44 newtons and 1699 ± 71 newtons, respectively, and were significantly greater (p < 0.001) than that of the S-O construct (1446 ± 53 newtons) or that of the H-10 construct (1370 ± 53 newtons). The S-O and H-10 constructs were significantly stronger (p < 0.05) than the Z-10, RT-1, RT-10, S-T, and S-10 constructs. There was no significant difference (p > 0.20) in fatigue strength between the Z-10 constructs (1254 ± 71 newtons) and the RT-1 constructs (1196 ± 116 newtons) or between the RT-1 constructs and the RT-10 constructs (1130 ± 49 newtons). The fatigue strengths of the S-T and the S-10 constructs (1001 ± 53 newtons and 970 ± 53 newtons, respectively) also did not differ significantly (p > 0.3).
Part II: Clinical Investigation
Materials and Methods
After the results of the biomechanical testing were analyzed, a clinical study of patients was begun at The R Adams Cowley Shock Trauma Center, Baltimore, Maryland, in January 1996, with approval of the institutional review board. The purpose of this investigation was to evaluate the safety and efficacy of immediate weight-bearing after treatment of a comminuted fracture of the femoral shaft with a statically locked intramedullary nail. The study consisted of a consecutive series of fractures treated with a 12.0-millimeter-diameter Russell-Taylor intramedullary nail that was inserted after reaming and locked with two distal locking screws (the RT-2 construct). All patients with a fracture of the femoral shaft were considered for the study. The inclusion criteria included (1) a type-III or type-IV comminuted fracture21; (2) a fracture with the proximal fragment circumferentially intact to a level distal to the lesser trochanter; (3) a fracture that did not extend to within 12.5 centimeters of the knee joint, as measured on an anteroposterior radiograph of the femur; (4) an endosteal canal that, in the estimation of the attending surgeon, could accommodate reaming and the insertion of a 12.0-millimeter-diameter intramedullary nail; (5) a patient who was able to provide informed consent within one week after the injury; and (6) a fracture of the femoral shaft treated within twenty-four hours after the injury at our institution. The exclusion criteria included (1) a fracture that did not meet the inclusion criteria, such as a fracture of the femoral shaft with ipsilateral injury of the femoral neck, ipsilateral fractures of the femoral shaft and the distal aspect of the femur, a fracture with type-I or type-II comminution21, and a fracture in a femur in which the diameter of the endosteal canal was not large enough for a 12.0-millimeter intramedullary nail; (2) a head injury or another injury that precluded walking with full weight-bearing on the injured limb within one week after the injury; (3) an open fracture of the femoral shaft (fractures secondary to low-velocity gunshots were not excluded from the study group); (4) a patient who, according to the judgment of the attending surgeon, should not be managed with reaming and intramedullary nailing; (5) a fracture stabilized with a second-generation nail; (6) a fracture stabilized before transfer to our institution; (7) an interlocking nailing performed for delayed union, malunion, or nonunion; (8) a patient who was skeletally immature; and (9) a patient who was seventy years of age or older.
All patients were permitted to bear weight immediately after the operation, and all began physical therapy within a week after the injury. The patient progressively increased weight-bearing as much as he or she could tolerate throughout the first few postoperative weeks. Demographic data, including the weight of the patient, as well as data on initial injuries (including other orthopaedic injuries) were collected. Routine follow-up was performed at two, six, ten, sixteen, and twenty-four weeks after the operation. Weight-bearing at the time of each follow-up visit was recorded.
At the time of each follow-up visit, the healing of the fracture of the femoral shaft and the integrity of the nail and screws were assessed with use of anteroposterior and lateral radiographs. The presence or absence of back-out or breakage of a nail or screw was recorded. The final follow-up examination was considered to be the one performed twenty-four weeks after fracture of the femoral shaft. The need for any additional operations on the femur was also recorded.
Since January 1996, thirty-five patients (thirty-six fractures) met the criteria for the study. Twenty-eight patients (twenty-nine fractures) had complete follow-up (twenty-four weeks) through the time of union of the fracture. Three patients never returned for follow-up, and four had incomplete follow-up.
There were twenty-one male patients and seven female patients. The average age of the patients was thirty-three years (range, sixteen to sixty-six years); the average weight of the patients was eighty-one kilograms (range, fifty to 127 kilograms). There were sixteen type-III and thirteen type-IV comminuted fractures21. Four fractures were caused by gunshots, and the rest occurred after blunt trauma. Eight of the patients had no other orthopaedic injury or general injury necessitating an operation. Of the other twenty patients, eight had an injury of the contralateral lower extremity or an injury of the upper extremity that may have precluded walking with or without crutches if it had not been possible to bear weight on the limb with the fracture of the femoral shaft (for example, a patient had a concomitant brachial plexus palsy, fracture of the olecranon, and fracture of the ulna; a patient had a comminuted fracture of the ipsilateral humeral shaft; a patient had a bimalleolar fracture of the contralateral ankle; and a patient had bilateral fracture of the femoral shaft).
Of the twenty-eight patients who had complete follow-up, two patients were recorded as not having progressed to full weight-bearing at six weeks after the injury. It was believed that this failure to bear full weight was due to continued restricted motion of the knee and discomfort. The fractures in both of these patients healed, and satisfactory function of the knee was regained by twelve weeks. Eight patients were walking without any assistive device, and the rest were bearing full weight with use of a cane or a single crutch.
All fractures united. No nails broke or became deformed, and no screw fractured or backed out. The screws were removed (the fracture was converted to a dynamic construct) from one patient to accelerate fracture-healing five months after nailing, and the fracture healed within three months after that procedure. Three other patients, in whom the fracture healed uneventfully, had the screws removed because the screw heads were causing irritation of the soft tissues.
In this two-part study, we determined the fatigue strength of eleven intramedullary nail constructs that had been used to stabilize simulated fractures of the midpart of the femoral shaft. Our results indicate that the RT-2 and Z-2 constructs had significantly higher fatigue strengths than the other constructs that were tested; the RT-2 and Z-2 constructs were still intact after 500,000 cycles of loads equivalent to three times the body weight of a seventy-kilogram patient23.
Forces acting across the hip joint are reported to range from one to four times body weight during normal walking1,7. Because of muscle activity spanning the hip, the joint forces are likely higher than the loads on the diaphysis of the femur. There is little information regarding the axial loads transmitted through the femur. Michel et al.15 used a custom intramedullary nail that was instrumented with strain gauges to measure the force transmitted through the nail in a patient who had a comminuted diaphyseal fracture of the femur. Initially, the load on the intramedullary nail was 23 percent greater than the applied weight-bearing load. However, this load-share decreased to half of the weight-bearing load at the time of the removal of the nail. Results from a recent three-dimensional model of the femur suggest that the midpart of the femoral shaft is subjected to a peak axial load of approximately two times body weight during gait8.
Run-out was set at 500,000 cycles because that number of cycles represents the activity level of a patient walking, on average, five miles (eight kilometers) each day, seven days per week, commencing immediately postoperatively and continuing for ten weeks, which is the amount of time that we assumed was necessary for substantial osseous consolidation. In comparison with the clinical situation, these mechanical stresses represent an extreme demand on the interlocking construct because no osseous consolidation was considered in our model. Furthermore, patients rarely have this level of activity immediately after a fracture of the femoral shaft. Thus, the fatigue strengths determined in the present study represent an estimated worst-case scenario for a typical patient who would be walking during the weeks after the operation. The results suggest that immediate walking with full weight-bearing may be possible despite substantial osseous comminution and the lack of additional support from fracture-healing.
The biomechanical portion of this study did not address certain clinical issues. No torsional loads were applied to these constructs, which could have affected the fatigue failure results. Also, the application of bending moments was intentionally avoided. All specimens were loaded in compression along the axis of the nail. In vivo, the femur is loaded along the mechanical axis of the lower extremity, creating a bending moment in the diaphysis. Currently, the effect of this bending moment on the fatigue strength of a locking nail and screws is unknown. Furthermore, investigators have found that breakage is not the only mode of failure of interlocking screws or a locking nail. In the study by Hajek et al.11, acute failure occurred at the interface between the bone and the screw at loads six to seven times body weight. These excessive loads to failure may help to explain why loss of fixation after interlocking intramedullary nailing is rarely seen in the clinical setting.
Although we initially considered using bone from cadavera as the specimens for our biomechanical test, we chose to avoid introducing another uncontrolled variable into this investigation19. The use of polyvinyl chloride pipe as a simulated femur permitted the testing of the nail-and-screw construct alone, without the added variable conditions of the interface between the bone and the screw.
We found no study in the literature in which the fatigue strength of locked intramedullary nails was determined experimentally. Thus, we were unable to compare the results of our study directly with those of other studies. Previous biomechanical studies of intramedullary nails have focused on torsional and bending stiffness or on axial load to failure6,11,12,14 and documented axial failure loads in ranges far greater than the loads associated with normal gait. Bucholz et al.4 addressed the fatigue issue in a clinical and finite element analysis of nailing for the treatment of fractures of the distal aspect of the femur; they concluded that fatigue of the nails could occur through distal locking holes, but failure of the screws was not documented or tested.
Recent clinical and biomechanical studies have raised questions concerning the necessity of using both distal screws to statically lock the nail for fractures of the femoral shaft9-11. Measurement of torsional and axial failure loads on the distal aspect of cadaveric femora in which an intramedullary nail had been implanted demonstrated that a single screw provides adequate fixation11. In the biomechanical testing of this construct in bone from cadavera, the most common mode of failure was cut-out of the distal screws11. Clinically, when the fracture was located in the middle-to-proximal region of the shaft, no difference in the rate of failure of the hardware or in the final result was noted between constructs with one distal locking screw and those with two distal locking screws11. Part I of our study, however, demonstrated that the use of two distal locking screws, compared with one, significantly increased the fatigue strength of a construct (p < 0.05).
We realize that the intramedullary nailing systems that we tested are not similar, which makes comparison difficult. The Russell-Taylor and Zimmer nails are made of stainless steel, whereas the Synthes and Howmedica systems are manufactured from titanium. No apparent advantage with regard to fatigue strength was seen with the titanium system. The lower fatigue strengths observed with the Synthes system may be related more to differences in design and size among the nails and screws than to the metallurgy used in the manufacture.
The results of our clinical study strongly suggest that early weight-bearing after stabilization of a comminuted femoral fracture with an RT-2 construct (a 12.0-millimeter-diameter Russell-Taylor nail and two 6.4-millimeter distal locking screws) or a Z-2 construct (a 12.0-millimeter-diameter Zimmer nail and two 5.5-millimeter distal locking screws) may be safe for a typical patient (for example, a seventy-five-kilogram man). Early weight-bearing may also be safe after treatment with other nail systems, depending on the body weight of the patient, the level of postoperative walking, and the design of the implants. As failure of the interface between the bone and the screw may be related to the degree of osteoporosis, if present, our biomechanical model mimics fixation of a fracture in the relatively young person. The results of this study may not apply to fractures occurring in osteopenic bone.
We chose to include in the clinical portion of this study only comminuted fractures of the femoral shaft because stable fracture patterns with contact between the proximal and distal fracture fragments can absorb a substantial portion of the load placed on the limb, and this can protect the locking system. Our biomechanical model and our criteria for inclusion in the clinical study created a worst-case scenario for the fatigue testing of the locking nails and screws. It is certainly possible, however, to stabilize a type-I comminuted fracture in slight distraction, recreating our laboratory and clinical model for the study of a stable fracture pattern. It is our opinion that the slight distraction at the fracture site after nailing is a common clinical reality.
Early weight-bearing could substantially increase the potential for walking for patients who have multiple injuries of extremities. Patients who have fractures involving more than one limb often must wait for early consolidation of the fracture before they can begin weight-bearing. This delay may slow rehabilitation, prolong hospitalization, and increase the cost of care. Although we made no attempt to measure those parameters, eight (29 percent) of the twenty-eight patients in our investigation would not have been permitted to walk postoperatively if they had not been allowed to bear weight immediately on the limb with the fractured femoral shaft.
However, we do not suggest continuance of full weight-bearing on the affected limb if there is no evidence of consolidation of the fracture or progression toward consolidation by twenty-four weeks after the injury. All systems have a finite fatigue strength, and fracture-healing must take place to prevent eventual failure of the nail or screws. While this study did not address this matter, we believe that the surgeon should consider some stimulatory procedure (for example, converting the fracture to a dynamic construct, exchange nailing with reaming, or bone-grafting) if progressive fracture-healing is not demonstrated on serial radiographs by that time.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Funds (or hardware) were received in total or partial support of the research or clinical study presented in this article. The funding (or hardware) sources were Zimmer (Warsaw, Indiana), Synthes (Paoli, Pennsylvania), Howmedica (Rutherford, New Jersey), and Smith and Nephew Richards (Memphis, Tennessee).
Investigation performed at the Section of Orthopaedics, Program in Trauma, The R Adams Cowley Shock Trauma Center, The University of Maryland Medical System, and the Orthopaedic Biomechanics Laboratory, University of Maryland at Baltimore, Baltimore
1. Bergmann, G.; Graichen, F.; and Rohlmann, A.: Hip joint loading during walking and running, measured in two patients. J. Biomech.
, 26: 969-990, 1993.
2. Brumback, R. J.; Uwagie-Ero, S.; Lakatos, R. P.; Poka, A.; Bathon, G. H.; and Burgess, A.: Intramedullary nailing of femoral shaft fractures. Part II: Fracture-healing with static interlocking fixation. J. Bone and Joint Surg.
, 70-A: 1453-1462, Dec. 1988.
3. Brumback, R. J.; Ellison, P. S., Jr.; Poka, A.; Lakatos, R.; Bathon, G. H.; and Burgess, A. R.: Intramedullary nailing of open fractures of the femoral shaft. J. Bone and Joint Surg.
, 71-A: 1324-1331, Oct. 1989.
4. Bucholz, R. W.; Ross, S. E.; and Lawrence, K. L.: Fatigue fracture of the interlocking nail in the treatment of fractures of the distal part of the femoral shaft. J. Bone and Joint Surg.
, 69-A: 1391-1399, Dec. 1987.
5. Collins, J. A.: Failure of Materials in Mechanical Design: Analysis, Prediction, Prevention, pp. 369-374. New York, John Wiley and Sons, 1981.
6. Covey, D. C.; Saha, S.; Lipka, J. M.; and Albright, J. A.: Biomechanical comparison of slotted and nonslotted interlocking nails in distal femoral shaft fractures. Clin. Orthop.
, 252: 246-251, 1990.
7. Davy, D. T.; Kotzar, G. M.; Brown, R. H.; Heiple, K. G.; Goldberg, V. M.; Heiple, K. G., Jr.; Berilla, J.; and Burstein, A. H.: Telemetric force measurements across the hip after total arthroplasty. J. Bone and Joint Surg.
, 70-A: 45-50, Jan. 1988.
8. Duda, G. N.; Schneider, E.; and Chao, E. Y.: Internal forces and moments in the femur during walking. J. Biomech.
, 30: 933-941, 1997.
9. George, C. J.; Lindsey, R. W.; Noble, P. C.; Alexander, J. W.; and Kamaric, E.: Optimal location of a single distal interlocking screw in intramedullary nailing of distal third femoral shaft fractures. J. Orthop. Trauma
, 12: 267-272, 1998.
10. Grover, J., and Wiss, D. A.: A prospective study of fractures of the femoral shaft treated with a static, intramedullary, interlocking nail comparing one versus two distal screws. Orthop. Clin. North America, 26: 139-146, 1995.
11. Hajek, P. D.; Bicknell, H. R., Jr.; Bronson, W. E.; Albright, J. A.; and Saha, S.: The use of one compared with two distal screws in the treatment of femoral shaft fractures with interlocking intramedullary nailing. A clinical and biomechanical analysis. J. Bone and Joint Surg.
, 75-A: 519-525, April 1993.
12. Johnson, K. D.; Tencer, A. F.; Blumenthal, S.; August, A.; and Johnston, D. W. C.: Biomechanical performance of locked intramedullary nail systems in comminuted femoral shaft fractures. Clin. Orthop.
, 206: 151-161, 1986.
13. Kempf, I.; Grosse, A.; and Beck, G.: Closed locked intramedullary nailing. Its application to comminuted fractures of the femur. J. Bone and Joint Surg.
, 67-A: 709-720, June 1985.
14. Kinast, C.; Frigg, R.; and Perren, S. M.: Biomechanics of the interlocking nail. A study of the proximal interlock. Arch. Orthop. and Trauma Surg.
, 109: 197-204, 1990.
15. Michel, M. C.; Schneider, E.; Genge, M.; and Perren, S. M.: Loading history of an interlocking femoral nail subsequent to fracture treatment. Orthop. Trans.
, 15: 412-413, 1991.
16. Perren, S. M.: The biomechanics and biology of internal fixation using plates and nails. Orthopedics
, 12: 21-34, 1989.
17. Seiler, J. G., III, and Swiontkowski, M. F.: A prospective evaluation of the AO/ASIF universal femoral nail in the treatment of traumatic and reconstructive problems of the femur. J. Trauma
, 31: 121-126, 1991.
18. Swiontkowski, M.: Personal communication, 1996.
19. Tencer, A. F.; Johnson, K. D.; Johnston, D. W.; and Gill, K.: A biomechanical comparison of various methods of stabilization of subtrochanteric fractures of the femur. J. Orthop. Res.
, 2: 297-305, 1984.
20. Webb, L. X.; Gristina, A. G.; and Fowler, H. L.: Unstable femoral shaft fractures: a comparison of interlocking nailing versus traction and casting methods. J. Orthop. Trauma
, 2: 10-12, 1988.
21. Winquist, R. A.; Hansen, S. T., Jr.; and Clawson, D. K.: Closed intramedullary nailing of femoral fractures. A report of five hundred and twenty cases. J. Bone and Joint Surg.
, 66-A: 529-539, April 1984.
22. Wiss, D. A.; Fleming, C. H.; Matta, J. M.; and Clark, D.: Comminuted and rotationally unstable fractures of the femur treated with an interlocking nail. Clin. Orthop.
, 212: 35-47, 1986.
23. Woodson, W. E.; Tillman, B.; and Tillman, P.: The body. In Human Factors Design Handbook: Information and Guidelines for the Design of Systems, Facilities, Equipment, and Products for Human Use. Ed. 2, pp. 544-609. New York, McGraw-Hill, 1992.