The Evolution of Femoral Shaft Plating Technique

CLINICAL ORTHOPAEDICS AND RELATED RESEARCH Number 354, pp 195-208 0 1998 LippincottWilliams 8 Wilkins

The Evolution of Femoral Shaft Plating Technique S. Robert Rozbruch, MD*; Urs Miiller, MD**; Emanuel Gautier, MD**; and Reinhold Ganz, MD**

There has been an evolution in the AOIAssociation for the Study of Internal Fixation plating technique during the past 3 decades that includes the use of longer plates and fewer plate screws, fewer lag screws outside the plate, fewer unicortical screws at the plate periphery, and greater use of the 95' blade plate to achieve balanced fixation of proximal and distal shaft fractures. These changes reflect an evolving technique of plate osteosynthesis that emphasizes indirect reduction techniques, biologic internal fixation, and improved biomechanics. Outcome data suggest that there has been an improvement with time that is reflected by shorter time to union, a decrease in the frequency of implant failures, delayed unions, nonunions, malunions, From the *Department of Orthopaedic Surgery, Beth Israel Medical Center North Division, New York, NY; *Department of Orthopaedic Surgery, The New York Hospital, Cornell Medical Center, Hospital for Special Surgery, New York, NY; Orthopaedic Surgery Specialists PC, New York, NY; and **Department of Orthopaedic Surgery, Inselspital, University of Bern, Bern, Switzerland. Work performed at Inselspital, University of Bern, CH3010 Bern Switzerland Reprint request to S. Robert Rozbruch, MD, Orthopaedic Surgery Specialists PC, 420 East East Seventy-Second Street, New York, NY 10021. Received: October 3,1996 Revised: February 28, 1997; February 11, 1998; April 9, 1998 Accepted: April 2 1,1998

195

number of reoperations, and in overall rate of failure. The best predictor of success was the length of plate by logistic regression analysis. With the evolution of plating techniques and a greater emphasis on biology of fracture healing, the incidence of complications and failures has decreased after femoral shaft plating. Plate osteosynthesis of the femoral shaft is particularly advantageous in many situations and can be quite successful (87% success rate in Group 111).

There has been an evolution in the techniques used for plating of long bone fractures. Despite the changes in practice, particularly during the last decade, there is little reflecting this in the English literature. The principles and guidelines for plate fixation discussed in many of the current orthopaedic textbooks are not what would be considered current state of the art,AO/Association for the Study of Internal Fixation (AO/ASIF) technique. Plate osteosynthesis is an important technique in the treatment of femoral shaft fractures. It is particularly advantageous in certain situations where an intramedullary nail may not be ideal. These may include incidences of adult and pediatric polytrauma14J5.32 especially with head trauma,23 pulmonary compromise,8,9,17,27,2*,39,40 ipsilateral femoral neck and shaft fractures,32fracture location in the proxi-

196

Clinical Orthopaedics and Related Research

Rozbruch et al

mal or distal shaft,4J*open fracture with a vascular injury,15 and an excessively narrow intramedullary canal. Additionally, in many underdeveloped areas of the world where the equipment necessary for intramedullary nailing is not available, plating can be performed. Femoral shaft plating is an important technique. The clinic that has generated the cases in this study has significant experience with femoral shaft fractures and their treatment (Table 1). The techniques of plate osteosynthesis have evolved as a result of experience with previous failures and an improved understanding of the biology of fracture healing and plate mechanics leading to the current state of biologic internal fixation. The concept of biologic internal fixation entails preserving the biologic reactivity of the tissue as much as possible. This includes careful tissue dissection, epiperiosteal exposure of bone, and indirect reduction of the fracture to avoid stripping and the devascularization of bone fragments. Reduction of the fracture to achieve anatomic alignment of extraarticular fractures and optimal, rather than maximal, stability is the goal. The biologic status of the injury site is not compromised to achieve a mechanical goal. The internal fixation includes implants that minimize bone necrosis, the judicious use of screws to avoid unnecessary additional trauma, and longer plates for better leverage and mechanical stability.4,5,24,30 To document and describe the evolution of AO/ASTF plating techniques three groups of patients who underwent femoral shaft plating, representing 3 decades of practice at one major AO/ASIF center, were studied. The purpose of this study is to document the evolution of plating technique objectively through a radiographic analysis and to evaluate the clinical results.

MATERIALS AND METHODS Eighty-one femoral shaft fractures, as classified by the AO/ASIF comprehensive classification of fractures,26 consecutively treated with open reduction

TABLE 1. Distribution of Plating and lntramedullary Nailing of Femoral Shaft Fractures During the Study Periods Period

Plating

1972-1 973 1982 1993-1994

25 (49%) 30 (66%) 28 (47%)

lntramedullary Nailing Total 26 (51%) 15 (33%) 31 (53%)

51 45 59

and internal fixation with a plate at Tnselspital, University of Bern, during the years 1972 to 1973, 1982, and 1993 to 1994 were reviewed retrospectively. This included fractures of the diaphyseal segment of the femur, designated by the number 32 in the AO/ASIF comprehensive classification of fractures. The proximal limit of the diaphyseal segment is determined by a transverse line passing through the inferior edge of the lesser trochanter and the distal limit of the diaphyseal segment is determined by the square method where the side of the square is the same length as the widest part of the epiphysis.26 Fractures beyond this zone were excluded. Pathologic fractures and periprosthetic fractures were excluded also. Group I included all of the femoral shaft fractures treated with plating during 1972 to 1973, a total of 25 fractures in 23 patients. Group I1 included all of the femoral shaft fractures treated with plating during 1982 (30 fractures in 30 patients). Group 111 included all the femoral shaft fractures treated with plating during 1993 to 1994 with the exception of three patients, and totaled 25 fractures in 23 patients. Three patients had moved to different geographic areas so their records and radiographs were unable to be located. During these same times, other femoral shaft fractures had been treated with intramedullary nailing (Table 1). Outcome data of at least 1 year followup were available for 73 fractures (91%). The patient lists of Groups I and IT were retrieved from the AO/ASIF documentation cards at Inselspital. The patient list of Group I11 was retrieved from a computerized file of operative reports compiled during the contemporary years. Clinical information was retrieved from hospital records and the parameters recorded are listed in Tables 2 and 3. Radiographic information was retrieved from direct analysis of the radiographs and these parameters are listed in Table 4.

Number 354 September, 1998

Femoral Shaft Plating Technique

197

TABLE 2. Clinical Data of 80 Fractures of the Femoral Shaft ~~

Variable Years treated Number of fractures Number of patients Age (average) (years) Age (range) (years) Open fractures Grade 1 Grade 2 Grade 3 Femoral shaft location Proximal third Middle third Distal third AO/ASIF Classification A

B C .Implant used Round hole plate Dynamic compression plate Limited contact dynamic compression plate 95" condylar blade plate Primary bone graft Satisfactory reduction

Group I

~

Group II

Group ill

1972-1973 25 23 32.8 12.1-64.4 8 (32%) 4 3 1

1982 30 30 26.7 14.0-57.6 10 (33%) 5 4 0

1993-1 994 25 23 49.5 19.8-86.3 3 (12%) 1 1 1

6 (24%) 15 (60%) 4 (16%)

6 (20%) 18 (60%) 6 (20%)

12 (48%) 6 (24%) 7 (28%)

8 (32%) 11 (44%) 6 (24%)

12 (40%) 14 (35%) 4 (13%)

14 (56%) 3 (12%) 8 (32%)

0 25 0

0 0 6 19 1 (4%) 25 (100%)

3 21 (1 narrow) 0

1

5

4 (16%) 23 (92%)

11 (30%) 30 (100%)

AO/ASIF = Association for the Study of Internal Fixation

The grading of open fractures was done according to the classification system of Gustilo and Anderson.' The fracture classification used was the AO/ASIF comprehensive classification.26 The implants used were the AO/ASIF 4.5 mm round hole plate, narrow and wide plates, the AO/ASIF 4.5 mm wide dynamic compression plate, the AO/ASIF 4.5 mm wide limited contact dynamic compression plate, and the AO/ASIF 95" condylar blade plate (Synthes, Paoli, PA). Perioperative antibiotics were administered to all of the patients of Groups I1 and 111. Patients of Group I routinely did not receive perioperative antibiotics. Time to union was determined from the clinical records and was considered the time at which the treating physician allowed the patient to begin full weightbearing and stated in the record that there was bony union. Implant failure was defined as breakage or bending of the plate or screws resulting in a loss of original postoperative position. Implant loosening was defined as loosening of the screws resulting in a loss of orig-

inal postoperative position. Delayed union was defined as time to healing more than 6 months. Nonunion was defined as a failure of progression toward union for more than 6 months. Nonunion was classified additionally as atrophic or hypertrophic based on the radiographic appearance. Malunion was defined as bony union in an abnormal position (>5' varus or valgus or abnormal rotation beyond 15'). Reoperations did not include routine removal of hardware after bony union. Success was defined as bony union in anatomic axial alignment and length and the absence of reoperation requiring a revision of the original implant. Failure was defined as the presence of malunion, nonunion, or reoperation requiring a revision of the original implant. The radiographic data (Table 4) were retrieved from direct analysis of the injury and postoperative radiographs. The standard full length anteroposterior (AP) view of the femur was used for the analysis, and the measurements were made with a standard centimeter ruler. The magnification was not corrected. The length of the fracture was con-

198


Rozbruch et al

TABLE 3. Outcome Data of 73 Fractures of the Femoral Shaft Variable Time to union (months)* Callus grade Implant failures Type of failures Average time (months) Implant loosening Average time (months) Delayed unions Nonunions Type Malunions Infections Reoperations* (excluding removal of hardware) Successes* AO/ASIF Type A (n = 31) AO/ASIF Type B (n = 26) AO/ASIF Type C (n = 16) Failures* Open fractures (n = 20)

Group I (n = 21)

Group II (n = 29)

Group 111 (n = 23)

4.9 3.13 4 (19%) AH plates 5.3 2 (9.5%) 1 3 (14.3%) 2 (9.5%) Hypertrophic 2 (9.5%) 1 (4.8%) deep 9 (43%)

4.98 2.96 3 (10.3%) All plates 4.7 0

3.38 3.2 1(4.3%) Screws 5 2 (8.7%) 3 0 1 (4.3%) Hypertrophic 0 1 (4.3%) deep 3 (13%)

13 (62%) 4/8 (50%) 5/9 (55%) 4/4 (100%) 8 (38%) 3/7 (43%)

2 (6.9%) 1 (3.4%) Hypertrophic 2 (6.9%) 1 (3.4%) deep 9 (31%) 24 (83%) 9/11 (82%) 12/14 (86%) 3/4 (75%) 5 (17%) 2/10 (20%)

20 (87%) 9/12 (75%) 3/3 (100%) 8/8 (100%) 3 (13%) 2/3 (66%)

Total

(n = 73) 4.45 8(11%)

4 (5%)

5 (7%) 4 (5%) 4 (5%) 3 (4%) 21 (29%) 57 (78%) 22/31 (71%) 20/26 (77%) 15/16 (94%) 16 (22%) 7 (35%)

AO/ASIF = Association for the Study of Internal Fixation. 'Statistically significant.

sidered the proximal to distal span of the fracture after reduction. The plate span ratio was defined as the ratio of the plate length to the fracture length. The lag screw index was defined as the number of lag screws outside the plate per centimeter of fracture length multiplied by 100. The plate screw density was defined as the number of plate screws per total plate holes multiplied by 100 to yield a percentage. The uniscrew density was defined as the number of peripheral unicortical plate screws per centimeter of the plate. The screw to fracture density was defined as the number of plate screws per centimeter of fracture length. Because most of these data are ratios or fractions, the magnification of the radiographs was not thought to be a confounding factor. Callus of the united fracture was classified by its greatest extension from the bone on the AP or lateral radiograph. A number value then was assigned: (1) no callus; ( 2 ) less than 5 mm of callus; (3) between 5 mm and 10 mm of callus; and (4) greater than 10 mm of callus. Data were recorded and calculations were performed on Microsoft Access and Excel software (Microsoft, Cambridge, MA). Statistical analysis

was performed using the most appropriate of various tests that included the Mantel-Haenszel linear trend test, Kruskal-Wallis test, test for linearity, analysis of variance (ANOVA), Student's t test, and logistic regression analysis.

RESULTS The overall patient profiles of the three groups were similar regarding the number of fractures, patient age, fracture location within the femoral shaft, AO/ASIF fracture classification, and satisfactory reduction of the fracture at surgery having been achieved. The proportion of open fractures was lower in Group I11 (12%) than in Groups I and 11 (32%, 33%) (Table 2). The implants used and the tendency to choose a certain implant varied among the groups. The AO/ASIF round hole plate and the dynamic compression plate were used most often in Group I. The 4.5 mm narrow

Number 354 SeDtember, 1998

Femoral Shaft Platina Techniaue

199

TABLE 4. Radiographic Data and Calculations of 80 Fractures of the Femoral Shaft Variable Length of fracture (cm) Number of plate holes* Length of plate (cmy Number of plate screws* Plate span ratio (length of plate/length of fracture) Lag screw index’ (lag screws outside of plate/ length of fracture) x 100 Plate screw density* (plate screwslplate holes) x 100 Uniscrew index* (peripheral unicortal screws/ length of plate) x 100 Screw fracture density (screws per cm) (plate screwdlength of fracture)

Group I (n = 25)

Group II

(n = 30)

Group 111 (n = 25)

8.7 11.8 21.7 11 3.9

10.1 13.9 25.5 11.4 5.3

10.3 14.6 27.8 6.5 6

10 95%

9 84%

0.75 45%

6

6

0

2.02

2.52

1.47

“Statistically significant.

round hole plate was used only for two fractures of Group I. Double plating (lateral and anterior plates) was used for internal fixation of femoral fractures in two patients of Group I. The use of the 95” condylar blade plate increased with time. In Group I, it was used for only one of 10 proximal and distal shaft fractures and, in Group 111, for 19 of 19 proximal and distal shaft fractures. The limited contact dynamic compression plate was used only in Group 111because of its unavailability before that time. Bone grafting was performed most frequently in the patients of Group I1 and least frequently in those of Group I11 (Table 2). The overall length of the fractures did not vary considerably among the groups. There was a tendency over time to use longer plates for a given length of fracture as reflected by the plate span ratio values. Despite the use of longer plates, the proportion of plate holes filled with screws decreased significantly with time as reflected by the plate screw density values (ANOVA, p < 0.0001). The tendency to use lag screws outside the plate to fix fragments internally decreased significantly with time as reflected by the lag screw index values (Kruskal-Wallis test, p = 0.0004). The present use of unicortical screws at the periphery of the plates virtually disappeared as

reflected by the uniscrew index values (Kruskal-Wallis test, p c 0.0001). There was a significant trend toward the use of fewer plate screws for a given length fracture as reflected by the screw fracture density values (KruskalWallis test, p = 0.01) (Table 4). Fracture healing outcome data (Table 3) of at least 1 year followup were available for 73 of the 80 total fractures (91%). Outcome information was available for 21 of the 25 fractures (19 of 23 patients) of Group I. Two patients received their postoperative care in different geographic areas and followup information was not available. Two patients died, one as a result of multiple organ failure at 1 week after the polytrauma, and one patient at l month after from meningitis resulting from an open skull injury during the trauma. Outcome data were available for 29 of the 30 fractures (29 of 30 patients) of Group 11. One patient received his postoperative care in a different geographic area so followup information was not available. Outcome data were available for 23 of the 25 fractures (23 of 25 patients) of Group 111. Two patients died, one as a result of a myocardial infarction 1 day after surgery, and one from congestive heart failure 3 months after surgery.

Clinical Orthopaedics

200

and Related Research

Rozbruch et al

The overall infection rate was 4% and did not differ significantly among the three groups. The overall success rate was 78%, and the A 0 Type C fractures had the best overall success rate (94%). Longer fractures had a better success rate than shorter fractures. Using Student's t test to evaluate predictors of success, the length of fracture was found to be a significant predictor of success (p = 0.03). The overall failure rate was higher for the open fractures (35%) than for the entire group (22%). A significantly shorter time to fracture union was seen in Group TI1 compared with Groups I and I1 (Kruskal-Wallis test, p = 0.004). There was no significant difference among the groups in the amount of callus measured on the radiographs of the healed fracture. The frequency of implant failures decreased with time. Implant loosening was similar in Groups I and 111and was lower in Group 11. The frequency of delayed union decreased with time. The nonunion rate of Groups I1 and III was lower than those of Group I. All of the nonunions seen were of a hypertrophic nature. The rate of malunion decreased with time. The rate of infection was low and similar in all of the groups. The frequency of reoperation (excluding routine removal of hardware) significantly decreased with time (Mantel-Haenszel linear trend test, p = 0.03). The success rate of the index procedure to achieve union without malunion significantly improved with time (Mantel-Haenszel linear trend test, p = 0.049) (Table 3). The best predictor of success using logistic regression analysis was an increased length of plate. There were three failures in Group I11 (13%). One was a nonunion in a Gustilo and Anderson Grade 1 open fracture with periosteal stripping. The second was an infection in a Gustilo and Anderson Grade 3B open fracture leading to bone resorption and eventual implant failure at 5 months. The third was a closed transverse fracture internally fixed with an 8-hole plate that loosened within 5 months after implantation.

DISCUSSION An evolution has occurred in the techniques used for fracture fixation that has been based on an improved understanding of the biology of fracture healing, of the biomechanics of fracture healing and fracture fixation, and on experience analyzing previous failures. This evolution has involved the implants and the techniques of application. In the current study, the evolution of plate osteosynthesis is documented and described through the study of femoral shaft fractures as a model for long bone fractures. This new approach, biologic internal fixation, has led to improved clinical outcomes. At the Swiss A 0 clinic, there is significant experience with plating of femoral shaft fractures. As can be seen in Table 1, plate fixation was used approximately 50% of the time. Particular clinical situations and fracture patterns lend themselves toward the technique to be used. Although intramedullary nailing is the technique of choice for the standard femoral shaft fracture,l6 there are clinical situations where plating is particularly advantageous. Femoral shaft plating is an important technique and should placed in the armamentarium of orthopaedic surgeons so that they can choose the best technique for the individual patient. Plate osteosynthesis is technically demanding, however, and the clinical success is related to the surgical technique used. To describe and quantify the technique, radiographs were chosen for study. A critical analysis of the radiographs reveals the technique used during surgery. For example, multiple lag screws placed outside the plate in various planes along the entire length of a comminuted fracture indicates soft tissue stripping of the bony fragments during surgery (Fig 1). Various indices and ratios were created to quantify the techniques used. The previous AO/ASIF guidelines for a specific number of screws and cortices in each fragment25 no longer are used in the clinic of the current authors. Implants, particularly screws, are used optimally and judiciously to

Number 354 SetXember, 1998

Femoral Shaft PlatinnTechnique

201

Fig 1A-E. (A) Anteroposterior injury radiograph showing an A 0 Type C femoral shaft fracture from 1973. (B) Anteroposterior and (C) lateral radiographs 6 months postoperatively showing reduction of individual fracture fragments and internal fixation with a 14 hole 4.5 mm wide round hole plate with multiple anterior lag screws, low plate span ratio, high plate screw density, and peripheral unicortical plate screws. Fracture lines are visible despite clinical union. Also, varus deformity is present. (D) Anteroposterior and (E) lateral radiographs 1 year postoperatively showing obliteration of fracture lines and the presence of callus.

avoid unnecessary surgical trauma to the bone. Use of double plating has been abandoned in the femur because of the obvious need for greater soft tissue stripping and devascularization of the bone. Similarly, the use of mul-

tiple lag screws outside the plate mandates stripping and devascularization of the bone so the use lag screws has decreased dramatically (Fig 1). The vascularity of fracture fragments is not compromised for absolute stability. Several clinical studies and basic

202

Rozbruch et al

science studies have shown that in comminuted fractures, the bony healing is improved by spanning the fracture zone and leaving it undisturbed rather than enhancing the primary postoperative stability and compromising the bony vascularity with dissection of each fracture fragment.l.l0-14.34.36,3* In shaft fractures, the exact reduction of each bony fragment is no longer a goal in itself. Rather, the overall restoration of length, axial alignment, and rotation are the goals (Figs 2,3). In comminuted fractures, absolute stability can be achieved only at the expense of the remaining viability of the fracture fragments. Such reduction and fixation techniques lead to a delay in fracture healing because bone healing can start only in areas of bone viability (Fig 1). As time elapses for the needed revascularization of the surgically devitalized bone tissue, the plate used for fixation will undergo high strain because of imposed external loading and is at risk of fatigue failure. The most critical stress of the bone-implant construct in a femoral fracture is the bending moment that results from the physiologic eccentric axial loading of the bone. There are two methods for plate stabilization of a femoral fracture. The first is a compression plate and the second is a bridge plate or splint. The method should be chosen based on the fracture type and personality. Compression plating can be used only in cases where at least partial contact of the main fracture fragments can be achieved, such as in transverse, oblique, or simple butterfly fragment fracture patterns. Tensioning of the plate during surgery leads to compression of the fracture fragments resulting in a loadsharing situation between the bone and the plate. This leads to a very stable situation with reduced strain or elastic plate deformation. The high degree of stability leads to at least partial pri-


mary bone healing with little callus. The specific healing pattern or biologic response is not the goal, but is the consequence of the mechanical stability. If a transverse fracture were plated without compression, only a relatively stable situation would be achieved. This would result in high strain in the small bony gap and in the short segment of plate between the two innermost screws. This would delay fracture healing and increase the possibility of nonunion.4,29,30 In comminuted fractures, interfragmentary compression and absolute stability is unnecessary, difficult to achieve, and only could be achieved at the expense of tissue viability. In these cases, the plate is used as an extramedullary splint without compression of the fracture. Although the bending moment on the plate is high because of its eccentric position, failure does not occur if the plate spans a long segment of bone. Under a constant external bending moment, the strain of the longer plate is reduced. The longer plate also reduces the force on the screws for a given external bending moment. In the comminuted fracture, the plate initially is the only load carrier of the construct. Little or no load transmission is possible through the comminuted area until bridging callus forms. The plate and the fracture undergo elastic deformation. The strain of the plate is lessened because of its increased length and leverage. This increases the number of loading cycles that the implant will tolerate before fatigue failure. The comminuted fracture also experiences low strain because of the relatively large surface area of the tissue, and tissue differentiation is not inhibited. Underneath the periosteal envelope of the cortex, bone formation can take place as a result of the limited strain. The internal structure of this woven bone can be com-

Fig 2A-F. (A) Anteroposterior and (B)lateral injury radiographs showing an A 0 Type C proximal femoral shaft fracture from 1993. (C) Anteroposterior and (D) lateral postoperative radiographs showing the product of indirect reduction and balanced fixation with an 18 hole 95" condylar blade plate, high plate span ratio, and low plate screw density. (E) Anteroposterior and (F) lateral radiographs 6 months postoperatively showing union with a massive callus response.

I)



203

204

Rozbruch et al

pared with a spring where the single elements of the material are less deformed than a linear wire connecting the two fragment ends. As the callus forms, loadsharing between the plate and the bone progresses (Fig 2). The highest success rate was found in the most comminuted AO/ASIF Type C fractures, and a greater length of fracture was found to be a statistically significant predictor of success. If the tissues are preserved meticulously and the surgeon resists the temptation to expose the entire fracture, bone grafting is unnecessary. Primary bone grafting is contraindicated if soft tissue dissection is necessary to allow placement of the graft.5 This approach is reflected by a primary bone grafting rate of 4% in the 1993 to 1994 group compared with 30% in the 1982 group. This practice is not consistent with the recommendations for routine bone grafting found in older studies.21,22.33,34 In those series, greater initial soft tissue dissection likely compromised the biologic healing potential, then mandating a biologic stimulus for healing, namely the bone graft. The design and biomaterial of newer implants is a part of the evolution of plating. For example, the limited contact dynamic compression plate which is composed of pure titanium and has undersurface grooves was used for all the midshaft fractures in Group 111. The work of Gautier and Perren2.3 and Perren et a131 suggests that titanium is advantageous because of its improved tissue compatibility relative to stainless steel, and the undersurface grooves are thought to improve the blood supply to the plated bone segment by minimizing the zone of ischemia under the plate. This is speculative and probably of limited relevance to the principal observations of this study. Balanced fixation more recently has been achieved better with increased use of the 95" condylar blade plate in proximal and distal fractures of the shaft. Despite a 40% incidence of proximal and distal shaft fractures in Group I, only one 95" condylar blade plate


was used in contrast to it being used for all 19 proximal and distal shaft fractures in Group 111. Because only a short lever arm is possible for the plate on one side of a proximal or distal shaft fracture, it is compensated by enhanced fixation at the plate's end (Fig 2).4J* The proportion of proximal and distal shaft fractures in Group I11 was 76% versus 40% in Groups I and 11. This probably reflects the current tendency to plate rather than nail proximal and distal shaft fractures and to reserve the intramedullary nail for the midshaft fracture. Also, the use of the narrow dynamic compression plate has been abandoned for use on the femoral shaft because the implant does not have adequate fatigue strength. Longer plates relative to the fracture length are thought to be particularly important and have been used more recently. The longer plate improves the construct by increasing the lever arm of the plate.6J9.30 With indirect reduction techniques, axial alignment of the femur is achieved without anatomic reduction of individual fragments.24 The longer plate requires fewer screws to achieve optimal fixation of the plate onto the bone surface when the screws are placed in the positions closest to the fracture in both major fragments and at the plate's periphery (Figs 2,3). Gotzen et aI6 have shown that a 10-hole plate fixed with two screws placed as mentioned on each side of a cadaver bone osteotomy has similar stability to an 8-hole plate fixed with four screws on each side of the osteotomy. They concluded that the former is a more efficient construct. Laurence et a l l 9 concluded that the use of more than four screws for one plate is redundant mechanically at the moment of implantation. They found that the highest tensile load applied to any screw by any plate during bending stress was never more than half the load needed to pull a screw out of even the weakest bone. Their data showed that the majority of tensile stresses are borne by the screw closest to the fracture and that as the length of a plate increased, the tensile

Fig 3A-E. (A) Anteroposterior injury radiograph showing an A 0 Type C segmental femoral shaft fracture from 1994. (B)Anteroposterior and (C) lateral postoperative radiographs showing the product of indirect reduction and balanced fixation with an 18 hole, 4.5 mm wide, limited contact dynamic compression plate, high plate span ratio, and low plate screw density. (D) Anteroposterior and (E) lateral radiographs 4 months postoperatively showing union with a moderate callus response.

206

Rozbruch et al

stress on the innermost screw decreased for a given bending moment. Furthermore, additional biologic compromise from the plate is minimized with improved plate designs. In the current study, an increasing length of plate was found by logistic regression analysis to be the best predictor of success. Despite the findings of Gotzen et a16 and Laurence et al,19 the common clinical recommendations have been to use four plate screws on either side of a femur fracture. This was first advocated by Lindahl20 in 1967. Seinsheimer37 agreed with these recommendations and the AO/ASIF Manual of Internal Fixation25 illustrates such a plate construct for a femoral fracture. The modern trend toward using fewer screws is reflected by the decreasing plate screw density and the screw fracture density in this study. By optimally rather than maximally using screws, damage to the bone is minimized.30 Bone in close apposition to a screw initially undergoes necrosis. If the screw does not move, then the dead bone gradually is replaced by living bone. If, however, there is micromotion at the bone screw interface, there is bone resorption that may be accompanied by fibrous tissue, synoviallike cells, cartilaginous tissue, and osteoclastic activity.29.35 After removal of screws, the remaining, temporary defects weaken the bone considerably in bending and torsion.19 The use of unicortical screws at the plate's periphery has been abandoned (Fig 1). Although originally thought to smooth the abrupt transition in stiffness at the plate's end, this currently is not thought to be beneficial. Having the best possible holding power at the end of the plate enhances its mechanical leverage,6.19 and the holding power of a unicortical screw is only approximately ?hof that of a bicortical screw.20 The use of a unicortical screw to hold a comminuted fracture fragment to the middle of the plate may be a way to minimize surgical trauma, while situating a major fragment in proper alignment. Although in conventional AO/ASIF technique plates were used with lag screws, in


the newer techniques of internal fixation, plates often are used without lag screws. The conventional lag screw technique is not forgiving. Even with a neutralization or protection plate, one single excessive load applied to the lag screw can pull the lag screw out of its tight connection to the bone and its holding power is lost. One overload permanently will negate the function of the screw. A tensioned plate used without a lag screw acts as an elastic but rigid spring. The load applied to the plate will deform the plate somewhat within the elastic zone and the fracture may undergo micromovement. In these cases callus formation is not an unwanted side effect but rather a consequence of the mechanical stability.4.30 The expectation of seeing bony union with more callus on recent radiographs compared with that seen on older radiographs was not supported by the data. Many of the older cases that were thought to have absolutely rigid fixation still underwent secondary bone healing with callus (Fig 1). Perhaps, the older constructs were not as rigid as originally thought. Also, the stability of the construct may have decreased with time as a result of bony resorption during healing. With newer techniques and implants, the success rate in the 1993 to 1994 group reached 87%. Again, the definition of success was stringent in that bony union without deformity resulting from only the index procedure was considered a success. The time to union and the complications including malunion, nonunion, implant failure, delayed union, and need for reoperation decreased with time. One confounding factor was the smaller number of open fractures in Group I11 compared with the numbers in Groups I and I1 because the overall failure rate for open fractures was higher than that of the entire group (35% versus 22%). The number of open fractures, however, was small and the difference in failure rates was not statistically significant. With the evolution of plating techniques, improved clinical outcomes have been observed, as other European authors have reported. Kinast et al'* showed improved


results in the treatment of subtrochanteric fractures fixed with the 95" condylar blade plate after the onset of indirect reduction techniques and discontinuation of bone grafting and dissection of the medial cortex. Heitemeyer and associatesl2-14 studied severely comminuted femur fractures that either were stabilized after anatomic reduction of the fractured area or stabilized with a bridging plate that spanned the undisturbed fracture area. They found quicker union and fewer complications in the latter group. Similarly, Baumgaertel and Gotzen,' Schoots et al,36 and Thielemann et aP8 reported favorable results after using long bridging plates and biologic techniques to stabilize comminuted femoral shaft fractures. Although use of the intramedullary nail for standard femoral shaft fractures is recommended, plate osteosynthesis is advantageous in special situations that were outlined. The success of plating is technique dependent, and the use of modem plating techniques and biologic internal fixation is recommended. This includes careful tissue dissection, extraperiosteal bone exposure, indirect fracture reduction techniques, bridge plating of comminuted shaft fractures, focus on achieving proper length, alignment, and rotation rather than anatomic reduction of extraarticular fracture fragments, judicious use of lag screws outside of the plate because it minimizes additional dissection and devascularization of fragments, balanced fixation with 95" condylar blade plates for proximal and distal shaft fractures, wide, 4.5 mm, limited contact dynamic compression plates for midshaft fractures, longer plates for better mechanical leverage, optimal rather than maximal use of plate screws, and the unusual need for primary bone grafting. Concrete recommendationsregarding plate length and number of screws needed cannot be made based on the current data. Additional study is needed for this. Preliminary guidelines, however, can be suggested. The plate span ratio of six and the plate screw density of approximately 50% seen in Group I11 can be used as a guide for successful treatment. Two


207

or three screws in each major fragment generally is recommended if the screw purchase into bone is excellent. A simple transverse fracture generally can be stabilized in compression with a 10- or 12-hole plate with two or three screws in each major fragment. The positions closest and farthest from the fracture should be used and an additional screw can be placed between these.

Acknowledgments The authors thank Kaj Klaue, MD, and Ralph Hertel, MD, for their intellectual contributions about biologic internal fixation, and David Helfet, MD, for his help with the manuscript.

References 1. Baumgaertel F, Gotzen L: Biological plate fixation of comminuted fractures of the shaft of the femur. A prospective study. Unfallchimrg 97:78-84,1994. 2. Gautier E, Perren SM: Limited contact dynamic compression plate (LC-DCP)-Biomechanical research as basis to new plate design. Orthopade 21:11-23,1991. 3. Gautier E. Perren S M The limited contact dvnamic compression plate (LC-DCP)-Biomechaniial research as a basis for the modified plate design. Orthopade 21:11-23,1992. 4. Gautier E, Perren SM, Ganz R: Principals of internal fixation. Current Orthopaedics 6:22&232, 1992. 5 . Gerber C, Mast JW, Ganz R: Biological internal fixation of fractures. Arch Orthop Trauma Surg 109: 295-303,1990. 6 . Gotzen L, Haas N, Riefenstahl L: Biomechanische untersuchungen zur plattenfixation an die hauptfragmente. Heftezur unfallheilkunde 165: 10-12,1983. 7. Gustilo RB, Anderson JT: Prevention of infection in the treatment of one thousand and twenty-five open fractures of long bones: Retrospective and prospective analysis. J Bone Joint Surg 58A:453458,1976. 8. Heim D: Experimental results of intraosseous pressure in various surgical procedures of the long bones. Orthopade 24:104-113,1995. 9. Heim D, Schlegel U, Perren SM: Intramedullary pressure in reamed and unreamed nailing of the femur and tibia: An in vitro study on intact human bones. Injury 24(Suppl):56-63, 1993. 10. Heitemeyer U, Claes L, Hierholzer G The significance of postoperative mechanical stability for bony repair of a comminuted fracture. An animal experiment. Unfallchirurg 93:49-55, 1990. 11. Heitemeyer U, Claes L, Hierholzer G, Korber M: Significance of postoperative stability for bony reparation of comminuted fractures. An experimental study. Arch OrthopTrauma Surg 109:144149,1990. 12. Heitemeyer U, Hierholzer G: Indications for a bridging plate osteosynthesis of compound femoral

208

Rozbruch et al

shaft fractures. Aktuelle Traumatologie 21:173-181, 1991. 13. Heitemeyer U, Hierholzer G: Bridging plate osteosynthesis in closed comminuted fractures of the femoral shaft. Aktuelle Traumatologie 15:205-209, 1994. 14. Heitemeyer U, Kemper F, Hierholzer G, Haines J: Severely comminuted femoral shaft fractures: Treating by bridging plate osteosynthesis. Arch Orthop Trauma Surg 106:327-330, 1987. 15. Jekic IM, Jekic ML: The status of plate osteosynthesis in the treatment of femur shaft fractures in polytrauma patients. Helvetica Chirurgie Acta 60:6 11-6 13, 1994. 16. Johnson KD: Femoral Shaft Fractures. In Browner BW, Jupiter JB, Levine AM, Trafton PG (eds). Skeletal Trauma. Vol 1. Philadelphia, WB Saunders 1525-1641,1992. 17. Khalfa K, Metais P: Assay of free plasma lipids after femoral and tibial fractures and centromedullary nailing. Ann Biol Clin 44:397-402, 1986. 18. Kinast C, Bolhofner BR, Mast JW, Ganz R: Subtrochanteric fractures of the femur. Results of treatment with the 95" condylar blade plate. Clin Orthop 238:122-130,1989. 19. Laurence M, Freeman MAR, Swanson SAV Engineering considerations in the internal fixation of fractures of the tibial shaft. J Bone Joint Surg 51B:754-768,1969. 20. Lindahl L The rigidity of fracture immobilization with plates. Acta Orthop Scand 38:101-114, 1967. 21. Loomer RL, Meek R, De Sommer F: Plating of femoral shaft fractures: The Vancouver experience. J Trauma 20:1038-1042, 1980. 22. Magerl F, Wyss A, Brunner C, Binder W: Plate osteosynthesis of femoral shaft fractures in adults. A followup study. Clin Orthop 138:62-73, 1979. 23. Marks PH, Paley D, Kellam JF: Heterotopic ossification around the hip with intramedullary nailing of the femur. J Trauma 28:1207-1213,1988. 24. Mast J, Jakob R, Ganz R: Planning and Reduction Technique in Fracture Surgery. Berlin, SpringerVerlag 1989. 25. Miiller ME, Allgower M, Schneider R, Willenegger H: Manudl of Internal Fixation. ed. 3. Berlin, Springer-Verlag 546-547, 1992. 26. Miiller ME, Nazarian S, Koch P, Schatzker J: The Comprehensive Classification of Fractures of Long Bones. Berlin, Springer-Verlag 1990. 27. Neudeck F, Obertacke U, Wozasek G, et al: Pathophysiologic consequences of various osteosynthesis


procedures in polytrauma patients. Part 1: Experimental studies of intramedullary pressure development in reamed and unreamed intramedullary nailing and plate osteosynthesis of the femur. Aktuelle Traumatologie 24: 114-120, 1994. 28. Pel1 AC, Christie J, Keating JF, Sutherland GR: The detection of fat embolism by transesophageal echocardiography during reamed intramedullary nailing. A study of 24 patients with femoral and tibial fractures. J Bone Joint Surg 75B:921-925, 1993. 29. Perren SM: Physical and biological aspects of fracture healing with special reference to internal fixation. Clin Orthop 138:175-196, 1979. 30. Perren SM, Buchanan J, Hertel R, Colton C: Basic AO/ASIF Technique: Aims and Principles. In Texhammer R, Colton C (eds). AO/ASIF Instruments and Implants-A Technical Manual. Berlin, Springer-Verlag 9-23, 1994. 31. Perren SM, Klaue K, Pohler 0, et al: The limited contact dynamic compression plate. Arch Orthop Trauma Surg 109:304-310, 1990. 32. Reimer B, Foglesong ME, Miranda MA: Femoral plating. Orthop Clin North Am 25:625-633, 1994. 33. Rinaldi E, Marenghi P: Plate and screw fixation by autogenous corticocancellous grafts in comminuted fractures of the femur. Italian Journal of Orthopaedics and Traumatology 10:349-354, 1984. 34. Riiedi TP, Liischer N: Results after internal fixation of comminuted fractures of the femoral shaft with DC plates. Clin Orthop 138:74-76, 1979. 35. Schatzker J, Home JG, Sumner-Smith G: The reaction of cortical bone to compression by screw threads. Clin Orthop 111:263-265,1975. 36. Schoots FJ, Van den Wildengerg FA, Van der Sluis RF, Goris RJ: Internal fixation of femoral fractures with extra long plates. Unfallchirurg 92:373-378, 1989. 37. Seinsheimer F: Concerning the proper length of femoral side plates. J Trauma 2 1:42-45, 1981. 38. Thielemann FW,Blersch E, Holz U: Biological plating of the shaft of the femur in polytraumatized patients. Unfallchirurg 91:389-394, 1988. 39. Wenda K, Runkel M, Degrief J, Ritter G: Pathogenesis and clinical relevance of bone marrow embolism in medullary nailing demonstrated by intraoperative echocardiography. Injury 24(Suppl): 73-81, 1993. 40. Wozasek G, Simon P, Red1 H, Schlag G: Intramedullary pressure changes and fat intravasation during intramedullary nailing: An experimental study in sheep. J Trauma 36:202-207, 1994.