You are on page 1of 14

Copyright 1996 by The Journal of Bone and Joint Surgery, Incorporated

Fractures of the Acetabulum: Accuracy of Reduction and Clinical Results in Patients Managed Operatively within Three Weeks after the Injury*
BY JOEL M. MATTA, M.D.t, LOS ANGELES, CALIFORNIA

Investigation performed at Los Angeles CountyUniversity of Southern California Medical Center, Good Samaritan Hospital, Los Angeles

ABSTRACT: The results were reviewed for 259 patients who had open reduction and internal fixation of 262 displaced acetabular fractures within twentyone days after the injury. Two hundred and fifty-five hips were followed for a mean of six years (range, two to fourteen years) after the injury; the remaining seven, which clearly had a poor result, were followed for less than two years. According to the classification of Letournel and Judet, associated fracture types accounted for 208 (79 per cent) of the fractures, with both-column fractures being the most common type (ninety-two hips; 35 per cent). Two hundred and fifty-eight hips were operated on with a single operative approach (Kocher-Langenbeck, ilioinguinal, or extended iliofemoral). The four remaining hips were operated on with a Kocher-Langenbeck as well as an ilioinguinal approach. The reduction was graded as anatomical in 185 hips (71 per cent). The rate of anatomical reduction decreased with increases in the complexity of the fracture, the age of the patient, and the interval between the injury and the reduction. The over-all clinical result was excellent for 104 hips (40 per cent), good for ninety-five (36 per cent), fair for twenty-one (8 per cent), and poor for fortytwo (16 per cent). The clinical result was related closely to the radiographic result. The clinical result was adversely affected by associated injuries of the femoral head, an older age of the patient, and operative complications. It was positively affected by an anatomical reduction and postoperative congruity between the femoral head and the acetabular roof. Osteonecrosis of the femoral head was noted in eight hips (3 per cent), and progressive wear of the femoral head was seen in thirteen (5 per cent). Subsequent operations included a total replacement of seventeen hips (6 per cent), an arthrodesis in four (2 per cent), and
*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 were received in total or partial support of the research or clinical study presented in this article. The funding source was the AO/ASIF Foundation, Davos, Switzerland. f637 South Lucas Avenue, Suite 605, Los Angeles, California 90017.

excision of ectopic bone in twelve (5 per cent). These findings indicate that in many patients who have a complex acetabular fracture the hip joint can be preserved and post-traumatic osteoarthrosis can be avoided if an anatomical reduction is achieved. An increase in the rate of anatomical reduction and a decrease in the rate of operative complications should be the goals of surgeons who treat these fractures. Displaced fractures of the acetabulum are best treated with anatomical reduction and rigid internal fixation. Residual displacement of more than one or two millimeters may lead to progressive post-traumatic osteoarthrosis and a poor functional result19. In 1980, I designed a treatment protocol for displaced acetabular fractures. The purpose of the current study was to evaluate the results of treatment with that protocol and to determine the relationship between the functional result at the latest follow-up evaluation and the quality of the reduction that had been achieved operatively. Materials and Methods From July 1980 through December 1991,1 treated 494 displaced fractures of the acetabulum in 489 patients (five of whom had a bilateral fracture). All of the fractures were displaced a minimum of five millimeters; thus, the indication for the operation was based on the configuration of the fracture. Fifty-seven hips were not operated on either because they met the criteria for non-operative treatment or because operative treatment was contraindicated. The criteria for non-operative treatment included apparent congruence in a hip with a both-column fracture or an acetabulum that was intact superiorly as documented by roof arc measurements and congruence of the femoral head with the roof on the three standard radiographs made with the patient out of traction14'1619. The hips that were treated nonoperatively as well as fifty-one that were operated on more than twenty-one days after the injury were excluded from the current study. Three hundred and eighty-six hips in 381 patients (five of whom had a bilateral fracture) were operated on within twenty-one days after the injury. Data were gathered prospectively according to a standardized
THE JOURNAL OF BONE AND JOINT SURGERY

1632

F R A C T U R E S OF T H E A C E T A B U L U M : ACCURACY OF R E D U C T I O N AND CLINICAL RESULTS

1633

protocol established in 1980; separate computer files were used for the initial and follow-up data. Complete initial data were available for all 381 patients (386 hips); however, 122 patients (124 hips) had had inadequate follow-up and I was unable to contact any of them for additional evaluation. Two hundred and sixty-two hips in 259 patients (three of whom had a bilateral fracture) had been followed adequately, and the initial and follow-up data for these patients form the primary basis for this report. Hips were judged to have had adequate follow-up if they had been followed for a minimum of two years (255 hips) or, if not, they had had a clearly poor clinical result before two years (seven hips). Demographic Data and Fracture Types The study included 184 male patients and seventyfive female patients. One hundred and six right acetabula (40 per cent) and 156 left acetabula (60 per cent) were fractured. The ages of the patients ranged from eleven to ninety years (mean, thirty-seven years). One hundred and forty-seven patients (56 per cent) were from twenty through thirty-nine years old; only twentynine patients (11 per cent) were sixty years old or more. The most common mechanisms of injury were a motor-vehicle accident (181 patients; 70 per cent), a motor vehicle-pedestrian accident (thirty-four patients; 13 per cent), and a fall (thirty-five patients; 14 per cent). (The nine remaining patients had other mechanisms of injury.) One hundred and forty-eight fractures (147 patients; 56 per cent) were associated with at least one additional injury. Fifty-one fractures (19 per cent) were associated with a head injury; twenty (8 per cent), an abdominal injury; forty-eight (18 per cent), a chest injury; seventeen (6 per cent), a genito-urinary injury; ninety-one (35 per cent), an injury involving an extremity; eleven (4 per cent), an injury of the spine; and thirty-three (13 per cent), a nerve palsy. Thirty-four fractures (13 per cent) were associated with more than two injuries; fifty (19 per cent), with two; and sixty-four (24 per cent), with one. The remaining 114 fractures (44 per cent) were not associated with other injuries. All patients were initially evaluated with use of three standard plain radiographs (one anteroposterior radiograph and two 45-degree oblique radiographs of the pelvis [Judet radiographs])4'812. Computerized tomography was performed on all patients who were managed after 1983. Although computerized tomographic scans and three-dimensional reconstructions of the scans were made for many patients, they were not used to determine the need for operative treatment or in the analysis of the results. The displacement of the fracture was measured separately. The maximum displacement seen on each of the three radiographs was recorded without adjustment for magnification. If it is assumed that the maximum displacement seen on any of the radiographs was the most accurate measure of displaceVOL. 78-A, NO. 11, N O V E M B E R 1996

ment, then displacement ranged from five to seventy millimeters and averaged twenty millimeters. With use of the initial radiographs, the fractures were classified according to the criteria of Letournel and Judet12. Of the 262 fractures, fifty-four (21 per cent) were simple fracture types and 208 (79 per cent) were associated fracture types. The simple fracture types included three anterior wall fractures (1 per cent), twelve anterior column fractures (5 per cent), twenty-two posterior wall fractures (8 per cent), eight posterior column fractures (3 per cent), and nine transverse fractures (3 per cent). The associated fracture types included ten posterior column-posterior wall fractures (4 per cent), sixty transverse-posterior wall fractures (23 per cent), thirty-one T-shaped fractures (12 per cent), fifteen anterior column-posterior hemitransverse fractures (6 per cent), and ninety-two both-column fractures (35 per cent). There was one open fracture. The distribution of the fractures, with associated patterns accounting for 79 per cent, reflects the fact that 246 patients (95 per cent) were transferred to my care from other hospitals because of the need for operative treatment and the complexity of the fracture. Eighty-three hips had a posterior dislocation and one had an anterior dislocation. Of the eighty-three posterior dislocations, twenty-one were reduced closed before the operation and sixty-two were reduced at the time of the operation. Preoperative traction was used for only a few patients for whom the operation was delayed after they were admitted to my institution; it was used to prevent posterior dislocation or wear of the femoral head against a transtectal T-shaped or transverse fracture. Operative Technique On the basis of the classification and the specific configuration of the fracture, and according to the recommendations of Letournel and Judet12, a single operative approach was selected; it was expected that the entire reduction could be performed with use of one approach. The Kocher-Langenbeck approach, in which the patient is positioned prone on the Judet table, was used in 112 hips (43 per cent); the ilioinguinal approach, in which the patient is positioned supine on the Judet table, was used in eighty-seven (33 per cent); and the extended iliofemoral approach, in which the patient is positioned laterally on the Judet table, was used in fifty-nine (23 per cent). The goal of the operation was to achieve an anatomical reduction of the innominate bone and the articular surface of the acetabulum. The Kocher-Langenbeck and ilioinguinal approaches were considered preferable to the extended iliofemoral approach because they involve less stripping of muscle from the innominate bone and the femur; the extended iliofemoral approach was chosen only when an anatomical reduction was judged to be probably impossible with use of one of the other two approaches alone. In

1634

J. M. MATTA

TABLE I
OPERATIVE APPROACH FOR EACH FRACTURE T Y P E * !

Fracture Type Anterior wall (n = 3) Anterior column (n = 12) Posterior wall (n = 22) Posterior column (n = 8) Transverse (n = 9) Posterior column-posterior wall (n = 10) Transverse-posterior wall (n = 60) T-shaped (n = 31) Anterior column-posterior hemitransverse (n = 15) Both-column (n = 92) Entire series (n = 262)

Kocher-Langenbeck

Operative Approach Extended Ilioinguinal Iliofemoral 3 11 1 2 (9%)

Kocher-Langenbeck and Ilioinguinal

20 (91 %) 8 8 10 46 (77%) 19(61%)

1 13 (22%) 6(19%) 1 36(39%) 59(22%) 1 (2%) 2(6%)

4(13%) 14 54(59%) 87(33%)

1(1%) 112(43%)

1(1%) 4(2%)

*As recommended by Letournel and Judet12. t All values are given as the numbers of hips.

four hips (2 per cent), an initial Kocher-Langenbeck or ilioinguinal approach was inadequate to complete the reduction and fixation and a combined ilioinguinal and Kocher-Langenbeck approach was used121417, during the same session of anesthesia. For six fracture types, the choice of the operative approach was consistent. The ilioinguinal approach was used for anterior wall, anterior column, and anterior column-posterior hemitransverse fractures. The KocherLangenbeck approach was used for posterior wall, posterior column, and posterior column-posterior wall fractures (Table I). The operative approach for the four remaining fracture types was chosen according to the specific fracture pattern as follows. Transverse fractures: The Kocher-Langenbeck approach is usually chosen. If the fracture line crosses the acetabulum from proximal anterior to distal posterior and the displacement is greatest anteriorly, the ilioinguinal approach is selected. Transverse-posterior wall fractures: In most patients, the Kocher-Langenbeck approach is employed. If the surgeon expects unusual difficulties with reduction, the extended iliofemoral approach may be selected. Transverse-posterior wall fractures are often difficult to reduce when they include a transtectal transverse component, when there is an extended posterior wall fracture (involving the posterior border of the bone), when they consist of T-shaped and posterior wall fractures, and when they are associated with dislocation of the symphysis pubis or fracture of the contralateral pubis ramus. T-shaped fractures: Usually, the Kocher-Langenbeck approach is chosen. If the surgeon expects unusual difficulties with reduction, the extended iliofemoral approach may be selected. T-shaped fractures that are often difficult with regard to reduction include those with a transtectal transverse component, those with a wide separation along the vertical stem of the T, and those

associated with dislocation of the symphysis pubis or fracture of the contralateral pubic ramus. Both-column fractures: These fractures are most frequently treated through the ilioinguinal approach. If the reduction is expected to be unusually difficult, the extended iliofemoral approach is chosen. Bothcolumn fractures that include a complex fracture of the posterior column, a displaced fracture line crossing the sacro-iliac joint, or a wide separation of the anterior and posterior columns at the rim of the acetabulum are often difficult to reduce. The Judet table was not available until 1985 but was used consistently after that time. Before 1985, AMSCO (American Sterilizer Company, Erie, Pennsylvania) and Kirschner fracture tables were used. The table is used to position the extremity securely, with distal and lateral traction applied to the femoral head in order to facilitate reduction and visualization of the acetabulum. The reduction was performed by direct manipulation of the bone with special reduction instruments; the goal was to achieve an anatomical reduction of the innominate bone and the acetabulum. The fixation was performed with plates and screws in 226 hips (86 per cent) and with screws alone in thirty-six (14 per cent)7,1213. Screws alone were used when the fracture divided the innominate bone into large segments, as in high anterior column fractures or simple-pattern both-column fractures. A fracture of the posterior wall was not considered an indication for fixation with screws alone. Fixation with interfragmentary screws, with the reduction held with reduction forceps or clamps, was usually performed before definitive fixation with a plate. Provisional fixation with Kirschner wires was not used. The loss of blood and the duration of the operation varied according to the operative approach. The Kocher-Langenbeck approach was associated with a mean blood loss of 900 milliliters (range, 100 to 3500 milliliters); the ilioinguinal approach, with a mean of
THE JOURNAL OF BONE AND JOINT SURGERY

F R A C T U R E S O F T H E ACETABULUM: ACCURACY OF R E D U C T I O N AND CLINICAL RESULTS

1635

1366 milliliters (range, 300 to 6000 milliliters); the ex- if the patient had osteoarthrosis or every two years if tended iliofemoral approach, with a mean of 1642 milli- no or only slight problems were identified. Only a few liters (range, 500 to 3500 milliliters); and the combined patients were examined at all of the follow-up interKocher-Langenbeck and ilioinguinal approach, with a vals, but all were seen at the final interval and the data mean of 2025 milliliters (range, 1400 to 2700 milliliters). from that examination were used for this report. At the The mean duration of the operation was 2.4 hours (range, latest follow-up examination, separate radiographic and one to eight hours) with use of the Kocher-Langenbeck clinical grades were assigned. According to the radioapproach, 3.4 hours (range, one to twelve hours) with use graphic criteria, a grade of excellent indicates a normal of the ilioinguinal approach, 4.1 hours (range, two to appearance of the hip; a grade of good, mild changes, seven hours) with use of the extended iliofemoral ap- small osteophytes, moderate (one-millimeter) narrowproach, and 6.3 hours (range, four to seven hours) with ing of the joint, and minimum sclerosis; a grade of fair, use of the combined Kocher-Langenbeck and ilioingui- intermediate changes, moderate osteophytes, moderate (less than 50 per cent) narrowing of the joint, and nal approaches. moderate sclerosis; and a grade of poor, advanced The operative findings included intra-articular free changes, large osteophytes, severe (more than 50 per fragments in seventy-five hips (29 per cent); associated cent) narrowing of the joint, collapse or wear of the injury to the cartilage or the bone of the femoral head, femoral head, and acetabular wear. The clinical grade or both, in fifty hips (19 per cent); and acetabular articwas based on a modification of the system of Merle ular impaction in thirty hips (11 per cent). d'Aubigne and Postel that I have used in previous Postoperatively, immobilization or traction was not 1518 studies (Table II) . The modification involves a more used. Walking with thirty pounds (13.6 kilograms) of strict grading of the score for the range of motion of the weight-bearing on the ipsilateral extremity was usually hip, which is determined by comparison of the total begun during the first week after the operation, dependscore for the injured side with that for the uninjured ing on the severity of pain and the associated injuries. side. Flexion, abduction, and adduction were evaluated Full weight-bearing was allowed eight weeks after the with the patient supine, and rotation was evaluated with operation. the patient prone. Limitations in extension were subThe mean interval between the injury and the opertracted from the measurement of flexion, to give a single ation was nine days (range, zero to twenty-one days). score for flexion-extension. The mean duration of hospitalization was nineteen days The score of Merle d'Aubigne and Postel includes (range, three to 137 days; median, eighteen days). A uniform protocol for prophylaxis against deep one subjective assessment for pain. The patient was invenous thrombosis was not established until 1990; at that terviewed by the examining surgeon, and the score was time, it included use of continuous external pneumatic- based on the responses to six questions: 1. Do you have any pain? compression devices for both lower limbs and adminis2. If you have pain, does it require medication? tration of Coumadin (warfarin) beginning forty-eight 3. Are there any activities that you can't do now to seventy-two hours after the operation and continuing but could do before the acetabular fracture? for forty days1. No prophylaxis against heterotopic ossi4. Can you sit for a prolonged period of time, or fication was used. does hip pain or stiffness require moving at Data Collection intervals? 5. Do you participate in sports or exercise activities, The telephone numbers and addresses of the paand are you limited in these? tients were recorded as part of the initial collection of data. Attempts were made to contact, by telephone 6. Do you work, and did you return to your preand by letter, patients who did not return for followinjury occupation? up. Patients were assured that the follow-up examina- Patients who reported no problems in response to these tion could be performed free of charge and, if necessary, questions were assigned a score of 6 points. Several they could be reimbursed for costs of transportation. patients who participated in very high-demand sports A few patients had moved to a distant location that activities (marathon running, triathlons, or university precluded examination in Los Angeles. Those patients, varsity sports) and had a decrease in performance after if they could be contacted, were interviewed by tele- the injury also received a score of 6 points. A score of phone, and a clinical examination was performed and 5 points was assigned to patients who had mild or inan anteroposterior radiograph of the pelvis was made termittent symptoms or limitations that is, pain or by another orthopaedic surgeon, who was reimbursed discomfort associated with rigorous activity but not through funds allocated for this study. I assessed the necessitating analgesics, stiffness after sitting for one follow-up radiographs for those patients. or two hours, and aching or stiffness associated with Clinical and radiographic examinations were per- changes in the weather. Patients who needed pain medformed and data were recorded at six months, one year, ication or had pain that limited walking or work activiand two years; thereafter, they were performed annually ties received a score of 4 points or less.
VOL. 78-A, NO. 11, N O V E M B E R 1996

1636
TABLE II
CLINICAL GRADING SYSTEM

J. M. MATTA

Points Pain None Slight or intermittent After walking but resolves Moderately severe but patient is able to walk Severe, prevents walking Walking Normal No cane but slight limp Long distance with cane or crutch Limited even with support Very limited Unable to walk Range of motion* 95-100% 80-94% 70-79% 60-69% 50-59% <50% Clinical scoref Excellent Good Fair Poor .6 5 4 3 2 6 5 4 3 2 1 6 5 4 3 2 1 18 15, 16, c 13 or 14 <13

sis was the statistical test used most often on observed contingency tables in order to find possible associations between variables. Other statistical analyses included means and ranges, frequency tables to find distributions, and t tests and analysis of variance to determine if there were significant differences between means. Results Operative Reduction The mean postoperative displacement was one millimeter (range, zero to fifteen millimeters). The postoperative reduction was graded as anatomical for 185 hips (71 per cent), imperfect for fifty-two (20 per cent), poor for eighteen (7 per cent), and surgical secondary congruence for seven (3 per cent) (Table III). The quality of the reduction was strongly associated with the fracture type (chi square = 44.859; p < 0.001). Fifty-two (96 per cent) of the fifty-four simple fracture types were reduced anatomically, compared with only 133 (64 per cent) of the 208 associated fracture types. All eighteen poor reductions were of associated fracture types. Treatment of the T-shaped posterior wall fracture type (fourteen hips; not shown in the tables because it is a subset of the transverse-posterior wall fracture type) resulted in eleven anatomical reductions, two imperfect reductions, and one poor reduction. The patient's age was a significant variable with regard to the accuracy of the reduction (chi square = 37.650; p = 0.006) (Table III). The rate of anatomically reduced fractures decreased after the age of forty years. Fifty-five patients (57 per cent) who were forty years old or more had an anatomical reduction, compared with 130 (78 per cent) who were less than forty years old. The initial displacement (mean, twenty millimeters; range, five to seventy millimeters) was not predictive of the accuracy of the reduction (Table III). Hips that were operated on earlier had a higher rate of anatomical reductions and those that were operated on later had a higher rate of poor reductions. Ninetythree (74 per cent) of the 126 operations that were performed between zero and seven days after the injury resulted in an anatomical reduction, as did seventytwo (71 per cent) of the 101 that were performed between eight and fourteen days and twenty (57 per cent) of the thirty-five that were performed between fifteen and twenty-one days. Comparison of these three subsets showed no significant difference (chi square = 3.703; p = 0.16). However, when hips that were operated on between zero and fourteen days after the injury were compared with those operated on between fifteen and twenty-one days, significance was approached (chi square = 3.531; p = 0.060). Relationship between Clinical Results and Associated Factors The mean interval between the injury and the latest follow-up evaluation was six years (range, two to fourT H E J O U R N A L OF BONE AND J O I N T S U R G E R Y

*The range of motion is expressed as the percentage of the value for the normal hip. This is calculated by obtaining a total of the ranges, in degrees, of flexion-extension, abduction, adduction, external rotation, and internal rotation for the injured hip and dividing it by the total for the normal hip. fThe clinical score is determined by adding the points for pain, walking, and range of motion.

After the operation, the reduction and fixation was evaluated with use of an anteroposterior radiograph of the pelvis in the operating suite. Three to seven days postoperatively, anteroposterior and 45-degree oblique (Judet) radiographs were made in the radiology department. The reduction of the fracture was evaluated by measuring the residual postoperative displacements on the three plain radiographs. For each of these radiographs, the maximum displacement seen at any of the normal radiographic lines of the acetabulum or the innominate bone was recorded in millimeters, and the highest of the three values was used to grade the reduction according to one of four categories: anatomical (zero to one millimeter of displacement), imperfect (two to three millimeters), poor (more than three millimeters), or surgical secondary congruence (the acetabulum is reduced anatomically but displacements in the innominate bone alter the position of the joint). Surgical secondary congruence is possible only with both-column fractures and was described by Letournal and Judet12. Statistical A nalysis Throughout this study, there were many occasions to count occurrences of one characteristic at each level of another characteristic. Therefore, a chi-square analy-

FRACTURES OF THE ACETABULUM: ACCURACY OF REDUCTION AND CLINICAL RESULTS TABLE III
DISTRIBUTION OF THE QUALITY OF THE REDUCTION ACCORDING TO FRACTURE TYPE, A G E OF THE PATIENT, AND INITIAL DISPLACEMENT OF THE FRACTURE*

1637

Quality iof Reduction Surgical Secondary Congruence

Anatomical Fracture type Simple (n = 54) Anterior wall (n = 3) Anterior column (n = 12) Posterior wall (n = 22) Posterior column (n = 8) Transverse (n = 9) Associated (n = 208) Posterior column-posterior wall (n = 10) Transverse-posterior wall (n = 60) T-shaped (n = 31) Anterior column-posterior hemitransverse (n = 15) Both-column (n = 92) Age 11-19 yrs.(n = 19) 20-29 yrs. (n = 76) 30-39 yrs. (n = 71) 40-49 yrs. (n = 39) 50-59 yrs. (n = 28) 60-69 yrs. (n = 20) 70-90 yrs. (n = 9) Initial displacement 5-10 mm (n = 97) 11-20 mm (n = 97) 21-30 mm (n = 46) 31-70 mm (n = 22) *A11 values are given as the numbers of hips.

Imperfect

Poor

3 12 22 (100%) 7 8 10 48 (80%) 16 (52%) 7 52 (57%) 18 60 (79%) 52 (73%) 25 (64%) 17(61%) 7 (35%) 6 74 64 34 13 (76%) (66%) (74%) (59%)

1 1

10 (17%) 10 (32%) 6 24 (26%) 1 12 13 9 6 8 3 19 19 9 5

2 (3%) 5 (16%) 2 9 (10%) 7 (8%)

(16%) (18%) (23%) (21%) (40%)

4 6 3 3 2

(5%) (8%) (8%) (11%) (10%)

2 (5%) 2 (7%) 3 (15%)

(20%) (20%) (20%) (23%)

12 (12%) 2 (4%) 4 (18%)

4 (4%) 2 (2%) 1 (2%)

teen years) (when the seven patients who had a poor result and less than two years of follow-up were excluded). The over-all clinical result for the 262 hips was excellent for 104 (40 per cent), good for ninety-five (36 per cent), fair for twenty-one (8 per cent), and poor for forty-two (16 per cent). Fracture Type Forty (74 per cent) of the fifty-four hips that had a simple fracture pattern had either an excellent or a good clinical result at the latest follow-up evaluation, compared with 159 (76 per cent) of the 208 that had an associated fracture pattern (Table IV). The type of fracture pattern (simple or associated) was independent of the clinical result (chi square = 3.522; p = 0.32), as was the specific fracture type (chi square = 35.792; p = 0.12). However, treatment of the T-shaped-posterior wall fracture type (fourteen hips) was associated with the highest prevalence of poor results (three excellent, five good, one fair, and five poor). Posterior Dislocation Fifty-nine (71 per cent) of the eighty-three hips that had a posterior dislocation of the femoral head had either an excellent or a good clinical result, compared with 140 (78 per cent) of those that did not have a
VOL. 78-A, NO. 11, NOVEMBER 1996

posterior dislocation. With the numbers available, this difference was not significant (p = 0.18). Initial Displacement The mean initial displacement of the fractures was eighteen millimeters in the hips that subsequently had an excellent result, twenty-one millimeters in those that had a good result, seventeen millimeters in those that had a fair result, and twenty-three millimeters in those that had a poor result. With the numbers available, these differences were not significant (p = 0.28). Associated Injuries The clinical results for the patients who had associated injuries were similar to those for the patients who did not. Of the 148 hips in the 147 patients who had associated injuries, fifty-nine (40 per cent) had an excellent result; fifty-four (36 per cent), a good result; fourteen (9 per cent), a fair result; and twenty-one (14 per cent), a poor result. Of the 114 hips in the 112 patients who did not have associated injuries, forty-four (39 per cent) had an excellent result, forty-one (36 per cent) had a good result, seven (6 per cent) had a fair result, and twenty-two (19 per cent) had a poor result. The presence of a head injury did not adversely affect the clinical score.

1638

J. M. MATTA

TABLE IV
DISTRIBUTION OF THE CLINICAL RESULTS ACCORDING TO FRACTURE TYPE, QUALITY OF THE REDUCTION, AND RADIOGRAPHIC G R A D E *

Clinical Result Excellent Fracture type Simple (n = 54) Anterior wall (n = 3) Anterior column (n = 12) Posterior wall (n = 22) Posterior column (n = 8) Transverse (n = 9) Associated (n = 208) Posterior column-posterior wall (n = 10) Transverse-posterior wall (n = 60) T-shaped (n = 31) Anterior column-posterior hemitransverse (n = 15) Both-column (n = 92) Entire series (n = 262) Quality of the reduction Anatomical (n = 185) Imperfect (n = 52) Poor (n = 18) Surgical secondary congruence Radiographic grade Excellent (n = 141) Good (n = 59) Fair (n = 29) Poor (n = 33) *A11 values are given as the numbers of hips. Good Fair Poor

1 9 9 (41%) 2 5 7 21 (35%) 6 (19%) 8 36 (39%) 104 (40%) 82 (46%) 17 (33%) 3 2 82 (58%) 21 (36%) 1 (3%)

1 1 6 (27%) 3 3 2 21 (35%) 18 (58%) 5 35 (38%) 95 (36%) 68 (37%) 18 (35%) 6 3 49 (35%) 32 (54%) 13 (45%) 1 (3%)

1 1 2 1

1 7 (32%) 1

5 (8%) 2 (6%) 1 9 (10%) 21 (8%) 10 (5%) 7 (14%) 2 2 6 (4%) 3 (5%) 7 (24%) 5 (15%)

1 13 (22%) 5 (16%) 1 12 (13%) 42 (16%) 25 (12%) 10 (18%) 7

4 (3%) 3 (5%) 8 (28%) 27 (82%)

Damage of the Femoral Head Of the fifty hips that had damage of the femoral head, thirty (60 per cent) had an excellent or a good clinical result, compared with 169 (80 per cent) of the 212 hips that did not have such damage. The mean clinical score2"518 for the patients who had damage of the femoral head was 14 points (range, 5 to 18 points), significantly less (p = 0.01) than the mean score of 16 points (range, 4 to 18 points) for those who did not have such damage. Intra-Articular Fragments and Acetabular Impaction The mean clinical score for the seventy-five hips that had intra-articular fragments was 15 points (range, 5 to 18 points), compared with 16 points (range, 4 to 18 points) for those that did not have such damage. With the numbers available, this difference was not significant (p = 0.15). The mean clinical score for the thirty hips that had acetabular articular impaction was 15 points (range, 7 to 18 points), compared with 16 points (range, 4 to 18 points) for those that did not have impaction. With the numbers available, this difference was not significant (p = 0.21). Age The patients were divided into two groups those who were forty years old or more and those who were

less than forty years old to determine whether there was a relationship between the clinical result and the age-group. Of the 166 hips in patients who were less than forty years old, 134 (81 per cent) had an excellent or good result, compared with only sixty-five (68 per cent) of the ninety-six hips in patients who were forty years old or more. This difference was significant (chi square = 5.641; p = 0.02). However, when the variable of quality of the reduction was controlled for, the agegroup was found to be independent of the clinical result. An excellent or good clinical result was associated with 110 (85 per cent) of the 130 anatomical reductions in the younger age-group, compared with forty (73 per cent) of the fifty-five such reductions in the older agegroup. An excellent or good clinical result was associated with eighteen (69 per cent) of the twenty-six imperfect reductions in the younger age-group, compared with seventeen (65 per cent) of the twenty-six such reductions in the older age-group. Such a result was associated with six of the ten poor reductions in the younger age-group, compared with three of the eight in the older age-group. Reduction When hips with a congruent reduction and those with an intra-articular infection were excluded, the quality of the reduction was strongly associated with the clinical result (chi square = 14.982; p = 0.02) (Table IV). There was an even stronger association between the
T H E J O U R N A L OF BONE A N D J O I N T S U R G E R Y

F R A C T U R E S OF T H E ACETABULUM: ACCURACY O F R E D U C T I O N A N D CLINICAL RESULTS

1639

clinical results associated with an anatomical reduction and those associated with imperfect and poor reductions combined. This is because, with the numbers available, the clinical results associated with the imperfect reductions did not differ significantly from those associated with the poor reductions. An anatomical reduction was strongly associated with an excellent or good result (chi square = 9.910; p = 0.002) (Table IV). Of the seven hips that had surgical secondary congruence, two had an excellent clinical result, three had a good result, and two had a fair result. Postoperatively, the femoral head was judged to be congruent with the acetabular roof in 250 (95 per cent) of the 262 hips; of these 250 hips, 102 (41 per cent) had an excellent clinical result, ninety-two (37 per cent) had a good result, nineteen (8 per cent) had a fair result, and thirty-seven (15 per cent) had a poor result. The operation did not result in congruence in the remaining twelve hips (5 per cent); two had an excellent clinical result, three had a good result, two had a fair result, and five had a poor result. It follows that postoperative congruence with the acetabular roof is a strong predictor of the clinical result (chi square = 8.450; p = 0.04). Postoperative incongruence was observed in twelve of the seventy-seven hips that had an imperfect or a poor reduction. Of the sixty-five hips in this group that had congruence, forty-four (68 per cent) had an excellent or good clinical result, compared with five of the twelve that did not have congruence. Radiographic Grade Of the 262 hips, 141 (54 per cent) had an excellent radiographic result, fifty-nine (23 per cent) had a good result, twenty-nine (11 per cent) had a fair result, and thirty-three (13 per cent) had a poor result (Table IV). There was a strong association between the radiographic grade and the clinical result (chi square = 176.890; p = 0.00001) (Table IV). One hundred and eighty-four (92 per cent) of the 199 hips that had an excellent or good clinical result also had an excellent or good radiographic result. There was a trend in the clinical scores relative to the radiographic grades. A radiographic grade of excellent was associated with a mean clinical score of 17 points (range, 8 to 18 points); a grade of good, with a mean clinical score of 17 points (range, 8 to 18 points); a grade of fair, with a mean clinical score of 14 points (range, 6 to 18 points); and a grade of poor, with a mean clinical score of 9 points (range, 4 to 15 points). The radiographic findings included narrowing of the joint in twenty hips (8 per cent), osteophytes in thirty (11 per cent), wear of the femoral head in thirteen (5 per cent), sclerosis in eight (3 per cent), and ectopic bone in 107 (41 per cent). Poor Clinical Results Forty-two (16 per cent) of the 262 hips had a poor clinical result. Of these forty-two hips, seventeen were
VOL. 78-A, NO. 11, N O V E M B E R 1996

treated with a total hip replacement and four, with an arthrodesis. Thirteen hips had not had an additional operation at the time of writing, but total hip replacement or arthrodesis was considered appropriate for these hips if requested by the patient. Eight hips had a poor result due to severe ectopic bone. Complications Neurological injury (two sciatic-nerve palsies, one femoral-nerve palsy, and six peroneal-nerve palsies) occurred during nine operations. The nerve palsies were associated with three Kocher-Langenbeck, two ilioinguinal, two extended iliofemoral, and two combined Kocher-Langenbeck and ilioinguinal approaches. In addition, a nerve palsy occurred secondary to the injury of thirty-two hips. The clinical results for these forty-one hips were similar to those for the patients who did not have palsy (thirty-one [76 per cent] excellent and good results and ten [24 per cent] fair and poor results, compared with 168 [76 per cent] excellent and good results and fifty-three [24 per cent] fair and poor results). A wound infection was a complication of the operation in thirteen hips (five of which had an extra-articular infection and eight, an intra-articular infection). All thirteen hips had had adequate follow-up and were included in the study group. When infection was suspected, the patient was returned to the operating room for the removal of specimens for culture, for irrigation and debridement, and for closure over suction drainage; antibiotics were subsequently administered. No patient had an active infection or persistent drainage at the latest follow-up examination. Four infections (all intra-articular) developed after a Kocher-Langenbeck operative approach; four (all extra-articular), after an ilioinguinal approach; and five (four intra-articular and one extra-articular), after an extended iliofemoral approach. Six of the eight hips that had an intra-articular infection had complete destruction of the joint space and a poor clinical result. The remaining two intra-articular infections were associated with one good result and one fair result. The five extra-articular infections were associated with three good and two fair results. The mean clinical score of the patients who had an infection was 12 points, compared with 16 points for those who did not have an infection. Of the thirteen hips (5 per cent) that had wear of the femoral head, one had a fair clinical result and twelve, a poor result. In ten of the thirteen hips, the wear was severe. Eight hips (3 per cent) had osteonecrosis of the femoral head; two of them had a good clinical result and six, a poor result. Eighty-three hips (32 per cent) had a posterior dislocation; five (6 per cent) of them had osteonecrosis. Wear of the femoral head was associated with five fracture types: transverse-posterior wall fractures (five of sixty), both-column fractures (four of ninety-two), posterior wall fractures (two of twentytwo), posterior column-posterior wall fractures (one of

1640

J. M. MATTA

%';
mt

'&';.

_ij

.1. .

FIG. 1-A Figs. 1-A, 1-B, and 1-C: Radiographs of a forty-eight-year-old woman who had a both-column fracture of the right acetabulum. Fig. 1-A: Anteroposterior and 45-degree oblique radiographs showing the fracture.

FIG. 1-B Anteroposterior and 45-degree oblique radiographs made after open reduction and internal fixation, performed with the extended iliofemoral approach seven days after the injury. This approach was chosen because of a proximal posterior column fracture that entered the sacro-iliac joint. The reduction was graded as imperfect; there was two millimeters of displacement of the articular surface, best seen on the oblique radiograph of the iliac crest.

ten), and T-shaped fractures (one of thirty-one). Ectopic bone was either absent or slight in 216 (82 per cent) of the 262 hips, and it was present with or without limiting motion in forty-six (18 per cent). Moderate or severe ectopic bone, associated with at least

a 20 per cent loss of motion, developed in twenty-three hips (9 per cent). As previously reported2912, there is a very strong association between the operative approach and the prevalence of this amount of ectopic bone. This amount of ectopic bone was noted after nine (8 per
THE JOURNAL OF BONE AND JOINT SURGERY

FRACTURES OF THE ACETABULUM: ACCURACY OF REDUCTION AND CLINICAL RESULTS


t

1641

)=> /'/

Ir
t

_ >

FIG.

1-C

Sequential follow-up radiographs. Left: Five months after the operation, the joint space was narrowed secondary to wear of the cartilage. Center: At eight months, the femoral head was smaller and had migrated in a proximal direction secondary to wear of both the femoral head and the acetabulum. Right: One year after the injury, a total hip replacement was performed with the Kocher-Langenbeck approach. A graft from the femoral head was used to fill a defect in the acetabular roof that had resulted from wear; the graft was fixed with two screws.

cent) of the 112 Kocher-Langenbeck approaches, twelve (20 per cent) of the fifty-nine extended iliofemoral approaches, and two (2 per cent) of the eighty-seven ilioinguinal approaches. Eight hips (3 per cent) had a partial loss of reduction of the fracture after the initial open reduction and internal fixation. The clinical result was excellent for one of these hips, good for two, and poor for five. Repeat open reduction in three of these hips led to an excellent result in one with a T-shaped fracture, a good result in one with a both-column fracture, and a poor result in one with a transverse-posterior wall fracture. There were no non-unions in this series. Comparison of the Hips That Had Adequate Follow-up with Those That Did Not In order to test the validity of the sample of the hips included in this study, the initial data for the 262 hips that had adequate follow-up was compared with such data for the 124 hips that did not have adequate followup. The eight factors chosen for comparison were age, fracture type, initial displacement, posterior dislocation, femoral head injury, operative approach, reduction, and wound infection. With the numbers available, there was no significant difference between the two groups with regard to four factors: distribution of the fracture types (chi square = 9.566; p = 0.39), reduction (chi
VOL. 78-A, NO. 11, NOVEMBER 1996

square = 4.845; p = 0.18), age (chi square = 10.539; p = 0.10), and operative approach (chi square = 2.560; p = 0.46). However, a significant difference between the two groups was found for the distributions of femoral head injury (chi square = 9.554; p = 0.002), posterior dislocation (chi square = 5.541; p = 0.02), initial displacement (chi square = 12.211; p = 0.007), and wound infection (chi square = 6.3689; p = 0.02). The significant differences occurred because 12 per cent more hips that had a femoral head injury were included in the study than were lost to follow-up (fifty [19 per cent] of 262, compared with nine [7 per cent] of 124). Also, 12 per cent more hips that had a posterior dislocation were included in the study than were lost to follow-up (eighty-six [33 per cent] of 262, compared with twenty-six [21 per cent] of 124). Furthermore, 18 per cent more hips that had between five and ten millimeters of initial displacement were lost to follow-up than were included in the study (sixty-eight [55 per cent] of 124, compared with ninetyseven [37 per cent] of 262). Finally, all hips that had a wound infection were included in the study and none were lost to follow-up. Discussion The primary complication following a fracture of the acetabulum is post-traumatic osteoarthrosis that may necessitate a total hip replacement or an arthro-

1642

X M. MATTA

FIG.

2-A

Figs. 2-A, 2-B, and 2-C: Radiographs of a fifty-year-old man who had a posterior column-posterior wall fracture of the right acetabulum. Fig. 2-A: Anteroposterior and 45-degree oblique radiographs showing the fracture.

FIG.

2-B

Anteroposterior and 45-degree oblique radiographs made after open reduction and internal fixation with the Kocher-Langenbeck approach, performed seven days after the injury. The reduction was graded as anatomical.

desis3'6'9-11-12-'5-IM"-22'23-25. Other possible sequelae include osteonecrosis of the femoral head; acetabular non-union and bone defects; and shortening of the ipsilateral ex-

tremity secondary to bone deformity, defects, or wear1226, The goal of operative treatment is to preserve a functional, mobile, painless hip joint that continues to funcTHE JOURNAL OF BONE AND JOINT SURGERY

FRACTURES OF THE ACETABULUM: ACCURACY OF REDUCTION AND CLINICAL RESULTS

1643

FIG. 2-C Anteroposterior radiograph made 8.5 years after the injury. The result was excellent clinically and radiographically, and motion of the hip was normal. Full strength of the abductors did not return for four years. The very proximal fracture into the greater sciatic notch was judged to have probably injured the superior gluteal nerve.

tion for the rest of the patient's life. Not all potentially deleterious effects of the initial injury can be completely countered in all patients. The success of the operation after high-energy trauma is contingent on the articular cartilage of the hip remaining viable. If posttraumatic osteoarthrosis develops in the presence of viable cartilage, it is primarily the result of altered pressure distribution of the femoral head articulating with an inaccurately reduced acetabulum. Specifically, the contact area between the head of the femur and the acetabulum is markedly reduced by a malreduction, and the force per unit area to the articular cartilage increases. This results in loss of the joint space and, sometimes, in wear of the femoral head. One goal of the current study was to determine which factors associated with the injury could be used to predict the prognosis. Fracture type was not found to be a significant predictor; in fact, the purpose of the classification system of Letournel and Judet12 was anatomical description and guidance of the operative approach rather than prognosis. Of the subset of fourteen hips with a T-shaped posterior wall fracture, only eight had an excellent or good result, but this finding was not significant because of the small number of these fractures. Posterior dislocation at the time of the injury, associated injuries, and the initial displacement of the fracture were not predictors of a worse prognosis, as previously reported26. Also, as already noted, patients who were forty years old or more had significantly worse clinical results (p = 0.02); however, the difference is accounted for in the intermediate-result (good and fair) category that is, the patients who were more than forty years old had fewer good and
VOL. 78-A, NO. 11, NOVEMBER 1996

more fair results, while the percentages of excellent and poor results were almost identical for patients forty years old or less and those more than forty years old. Furthermore, when the factor of reduction is controlled for, the differences between the two age-groups are no longer significant, indicating that the result is more affected by reduction than by age. Neither impaction of the acetabular articular surface nor intra-articular free fragments, which indicate some degree of comminution and additional problems in reconstruction, were predictive of the prognosis. Finally, the most clearly predictive initial factor was injury to the cartilage or bone, or both, of the femoral head; this factor was significantly predictive of a worse prognosis (p = 0.01). Osteonecrosis of the femoral head, which is known to result from the injury, can produce an unsatisfactory clinical result regardless of the method of treatment; however, osteonecrosis occurred after only eight (3 per cent) of the 262 fractures in the current study. Letournel and Judet12 reported osteonecrosis after nineteen (4 per cent) of 492 fractures. Subsequent wear of the femoral head against a malreduced acetabulum was a more frequent problem and was seen in thirteen (5 per cent) of the 262 hips in the current series. Higher rates of osteonecrosis (thirteen [23 per cent] of fiftysix hips) have been reported in other series26; however, the wear may have been incorrectly attributed to osteonecrosis (Figs. 1-A, 1-B, and 1-C). Another goal of the current study was to assess the relationship between the quality of the operative reduction (including any complications) and the clinical result. When the results of anatomical reduction were compared with those of imperfect and poor reductions,

1644

J. M. MATTA

the former was found to be a highly significant predictor of an excellent or good clinical result (chi square = 9.910; p = 0.002). In earlier reports of the results in the current series, with smaller numbers of hips and shorter durations of follow-up, anatomical and imperfect reductions were found to have similar outcomes1618. In those reports, an anatomical reduction was defined as one that resulted in zero to one millimeter of displacement; an imperfect (satisfactory) reduction, as one that resulted in two to three millimeters of displacement; and a poor (unsatisfactory) reduction, as one that resulted in more than three millimeters of displacement16'8. The largest difference in the results was found when those of anatomical and imperfect reductions were grouped together and compared with those of poor reduction. With the greater number of patients and the longer duration of follow-up in the current report, the findings have changed. The results of anatomical reduction clearly differed from those of imperfect and poor reductions, a finding that supports that of Letournel and Judet12 (Figs. 2-A, 2-B, and 2-C). These new findings have led me to abandon the terms used previously for reduction (anatomical, satisfactory, and unsatisfactory). The term satisfactory is particularly misleading because it implies that displacement of as much as three millimeters is acceptable; although this was supported, to a certain degree, by the data reported previously1618, it is not supported by the current, more complete data. Despite the appearance of an anatomical reduction on radiographs, there may still be imperfections on areas of the articular surface that are invisible on standard plain radiographs or are hidden by plates and screws. Even when the quality of the reduction is known precisely, the allowed one millimeter of displacement is not strictly anatomical nor are the inevitable displacements that are too small to be seen, felt, or measured. I believe that these tiny displacements are always present, even with an anatomical reduction. It follows that there are always alterations in the distribution of pressure within the joint after an acetabular fracture and its operative reconstruction. It must therefore be hypothesized that satisfactory clinical and radiographic results are due in part to the capability of the acetabulum in an adult to tolerate limited changes in the distribution of pressure and perhaps to reshape itself over time. However, if a satisfactory long-term result is to be achieved, the term anatomical reduction must be inherent in the mindset of the operating surgeon, who must strive for

a reduction with no residual displacement. One of the main difficulties in the analysis of these results is the loss of almost one-third (124 of 386) of the hips to follow-up. Despite my continued efforts and those of my research assistants, we were only partially successful because of the mobility (due to changes in residence and reasons related to occupation) of this young-adult American population. The recourse was comparison of the initial data for the 262 hips that had adequate follow-up with those for the 124 hips that had inadequate follow-up, in order to test the validity of the sample available for study. Several factors indicate that the over-all rates of excellent and good clinical results would have been slightly better if all patients had been available for follow-up. Patients who had less than two years of follow-up and a clearly poor result were included, while those who had an excellent or good result but less than two years of follow-up were excluded. Significantly more patients (p = 0.02) who had an associated injury of the femoral head were available for follow-up. All thirteen hips that had an infection were available for follow-up. In addition, the series was skewed toward the more complex associated types of fractures, as most of the patients were not initially admitted to my institution but were referred later because of expected technical difficulty with the reduction and fixation. Therefore, the current data probably represent an underestimation of the possibility of preserving the hip joint after a displaced acetabular fracture. The strong relationship between the quality of the reduction and the clinical result closely parallels the findings of Letournel and Judet12. It can therefore be concluded that the positive results reported by Letournel and Judet can be reproduced by the concentrated efforts of a surgeon who treats these fractures frequently and over an extended period of time. Other authors have reported poor results for as many as twenty-four (43 per cent) of fifty-six hips when multiple surgeons with less experience and inconsistent protocols participated526. Thus, the indications for operative treatment may not be the same in all clinical settings, as a patient who has had an operation that resulted in a poor reduction or a serious complication can be worse off than one who has had no operation at all. Management of patients at specialized centers where these injuries are treated frequently will probably yield the best clinical results.
NOTE: The author thanks Margaret del Valle. B.S., for her help in the tabulation, description, and statistical analysis of the results.

References
1. Fishmann, A. J.; Greeno, R. A.; Brooks, L. R.; and Matta, J. M.: Prevention of deep vein thrombosis and pulmonary embolism in acetabular and pelvic fracture surgery. Clin. Orthop., 305:133-137,1994. 2. Ghalambor, N.; Matta, J. ML; and Bernstein, L.: Heterotopic ossification following operative treatment of acetabular fracture. An analysis of risk factors. Clin. Orthop., 305: 96-105,1994. 3. Johnson, E. E.; Matta, J. M.; Mast, J. W.; and Letournel, E.: Delayed reconstruction of acetabular fractures 21-120 days following injury. Clin. Orthop., 305:20-30,1994. THE JOURNAL OF BONE AND JOINT SURGERY

FRACTURES OF THE ACETABULUM: ACCURACY OF REDUCTION AND CLINICAL RESULTS

1645

4. Judet, R.; Judet, J.; and Letournel, E.: Fractures of the acetabulum: classification and surgical approaches for open reduction. J. Bone and Joint Surg., 46-A: 1615-1646, Dec. 1964. 5. Kaempffe, F. A.; Bone, L. B.; and Border, J. R.: Open reduction and internal fixation of acetabular fractures: heterotopic ossification and other complications of treatment. J. Orthop. Trauma, 5: 439-445,1991. 6. Kebaish, A.; Roy, A.; and Rennie, W.: Displaced acetabular fractures: long-term follow-up../. Trauma, 31:1539-1542,1991. 7. Knight, R. A., and Smith, H.: Central fractures of the acetabulum. /. Bone and Joint Surg., 40-A: 1-16, Jan. 1958. 8. Letournel, E.: Les fractures du cotyle, etude d'une serie de 75 cas. /. chir, 82: 47-87,1961. 9. Letournel, E.: The results of acetabular fractures treated surgically. 21 years' experience. In The Hip. Proceedings of the Seventh Open Scientific Meeting of The Hip Society. St. Louis, C. V. Mosby, 1979. 10. Letournel, E.: Open reduction and internal fixation of acetabular fractures: long term results and analysis of 1040 cases [abstract]. In 1st International Symposium on Surgical Treatment of Acetabular Fractures, p. 2. Paris, Fondation de L'Avenir, 1993. 11. Letournel, E.: The treatment of acetabular fractures through the ilioinguinal approach. Clin. Orthop., 292:62-76,1993. 12. Letournel, E., and Judet, R.: Fractures of the Acetabulum, edited by R. A. Elson. New York, Springer, 1993. 13. Levine, M. A.: A treatment of central fractures of the acetabulum./ Bone and Joint Surg., 25:902-906, Oct. 1943. 14. Matta, J.: Operative indications and choice of surgical approach for fractures of the acetabulum. Tech. Orthop., 1:13-22,1986. 15. Matta, J. M.: Operative treatment of acetabular fractures through the ilioinguinal approach. A 10-year perspective. Clin. Orthop., 305: 10-19,1994. 16. Matta, J. M., and Merritt, P. O.: Displaced acetabular fractures. Clin. Orthop., 230: 83-97,1988. 17. Matta, J. M.; Letournel, E.; and Browner, B. D.: Surgical management of acetabular fractures. In Instructional Course Lectures, The American Academy of Orthopaedic Surgeons. Vol. 35, pp. 382-397. St. Louis, C. V. Mosby, 1986. 18. Matta, J. M.; Mehne, D. K.; and Roofl, R.: Fractures of the acetabulum. Early results of a prospective study. Clin. Orthop., 205: 241250,1986. 19. Matta, J. M.; Anderson, L. M.; Epstein, H. C; and Hendricks, P.: Fractures of the acetabulum. A retrospective analysis. Clin. Orthop., 205:230-240,1986. 20. Mayo, K. A.: Open reduction and internal fixation of fractures of the acetabulum. Results in 163 fractures. Clin. Orthop., 305: 3137,1994. 21. Merle d'Aubigne, R. M., and Postel, M.: Functional results of hip arthroplasty with acrylic prosthesis. J. Bone and Joint Surg., 36-A: 451-475, June 1954. 22. Oransky, M., and Sanguinetti, C: Surgical treatment of displaced acetabular fractures: results of 50 consecutive cases. / Orthop. Trauma, 7:28-32,1993. 23. Pantazopoulos, T., and Mousafiris, C: Surgical treatment of central acetabular fractures. Clin. Orthop., 246: 57-64,1989. 24. Romness, D. W., and Lewallen, D. G.: Total hip arthroplasty after fracture of the acetabulum. Long-term results. /. Bone and Joint Surg., 72-B(5): 761-764,1990. 25. Ruesch, P. D.; Holdener, H.; Ciaramitaro, M.; and Mast, J. W.: A prospective study of surgically treated acetabular fractures. Clin. Orthop., 305: 38-46,1994. 26. Wright, R.; Barrett, K.; Christie, M. J.; and Johnson, K. D.: Acetabular fractures: long-term follow-up of open reduction and internal ' fixation. J. Orthop. Trauma, 8: 397-403,1994.

VOL. 78-A, NO. 11, NOVEMBER 1996