The clinical and biomechanical results of a triple osteotomy of the innominate bone performed in twenty-two patients at The Children's Hospital and the University of Colorado Health Sciences Center, Denver, between 1980 and 1991, were reviewed. Three patients were lost to follow-up and one died. An additional patient had an incongruous hip joint and a flattened femoral head, which, in retrospect, were not considered indications for a triple osteotomy. This left a study group consisting of seventeen patients (twenty-two hips). There were sixteen female patients (twenty hips) and one male patient (two hips). The mean age of the patients at the time of the operation was twenty-four years (range, twelve to forty-one years). Five patients had a concomitant femoral varus-derotation osteotomy. Previous operations on the involved hip included a femoral osteotomy in one patient, a Salter osteotomy in one, and a shelf procedure in one. The etiology of the dysplasia was congenital in nine hips; in the other thirteen hips, the diagnosis was primary acetabular dysplasia (Table I). The indications for the operation were acetabular dysplasia, defined radiographically by a center-edge angle19 of less than 20 degrees and an acetabular index13 of greater than 30 degrees, and pain that interfered with normal activities. Preoperatively, all hips demonstrated a sufficient degree of congruity, such that the contour of the acetabulum made it possible to redirect the acetabulum over the femoral head, and all had a well preserved articular cartilage space and a functional range of motion. It was believed that the biomechanics of the hip would be improved by redirection of the acetabulum. No patient had an incongruous or osteoarthrotic hip joint for which a salvage osteotomy or a joint replacement would have been indicated. The mean preoperative center-edge angle was 9 degrees (range, -4 to 23 degrees) (Table I). The five patients who had a concomitant femoral osteotomy all had had a femoral neck-shaft angle that was greater than 155 degrees preoperatively. All of the operations were performed by the senior two of us (R. E. E. and J. D. W.) with use of Steel's technique15,16.
Radiographic and clinical data were obtained preoperatively, in the immediate postoperative period (within six months), and at the time of the latest follow-up. The mean duration of follow-up was 6.8 years (range, 2.2 to 13.8 years). Patients were located for long-term follow-up with the technique described by Chung.
A two-dimensional biomechanical analysis of each hip was done with use of anteroposterior radiographs of the pelvis that were made preoperatively, postoperatively, and at the time of the most recent follow-up. All of the results were calculated with use of a spreadsheet program (Quattro Pro; Borland International, Scotts Valley, California) on an IBM-compatible personal computer. The model for determining the joint pressure, or stress, across the hip joint was adapted from Legal's model. All calculations were made on the basis of the conditions during the single-limb-stance phase of gait, during which the greatest load is exerted on the hip. Hip load, or the resultant force acting across the hip joint, is calculated in relation to the partial body weight, which is the weight of the subject minus the weight of the support limb, or five-sixths of the total body weight9, and is thus normalized for the actual body weight of the subject so that comparisons can be made among different subjects. Hip load per partial body weight (relative load) and stress per partial body weight (relative stress) were determined from geometric parameters and reference points on an anteroposterior radiograph of the pelvis (Fig. 1). The stress on a joint depends on the magnitude of the total force acting on the joint, the area of the weight-bearing surface, and the distribution of forces on that surface. The area of the weight-bearing surface was determined with use of measurements made on the anteroposterior radiograph, approximating the mean acetabular inlet plane. The theory and derivation of the calculations of the biomechanics of the hip are based on the work of Legal. Because the direction of the muscular force (angle s) (Fig. 1) is variable, the stress is calculated as a function of s. Legal showed, however, that s ranges from 60 to 75 degrees in both normal and abnormal hips; he called this the physiological range. Therefore, for each hip, the stress before the operation, immediately (within six months) after the operation, and at the time of the most recent follow-up was determined as a function of s between 60 and 75 degrees. Because the equation is linear, the mean relative stress and relative load within the physiological range could be used for comparisons among subjects and between preoperative and postoperative values. The absolute stress was also independently evaluated for each hip on the basis of individual partial body weight. The Harris hip-rating system was used for the clinical evaluation of each hip, and the scores were correlated with the biomechanical results.
A control group was composed of twenty-one patients, matched for age and gender, who had asymptomatic hips. Hip load, stress, and the area of the weight-bearing surface were calculated from measurements on the normal anteroposterior radiographs of the pelvis. The significance was analyzed with use of repeated-measures analysis of variance and the Student t test. The level of significance was p < 0.001.
The iliac osteotomy successfully united in all of the patients. Two patients had a non-union of the osteotomy of the superior pubic ramus; one patient was asymptomatic, and the other was managed with subsequent bone-grafting and application of a plate. Another patient had bone-grafting for a non-union of the femoral osteotomy.
Biomechanical Analysis of the Control Hips
In the twenty-one control hips, the mean values (and standard deviation) were 1.38 ± 0.263 megapascals for absolute stress, 0.21 ± 0.043 centimeter-2 for relative stress, 3.14 ± 0.222 for relative load, 15.72 ± 2.647 square centimeters for the area of the weight-bearing surface, and 36 ± 5.1 degrees for the center-edge angle.
Biomechanical Analysis of the Dysplastic Hips
The mean relative stress (and standard deviation) postoperatively (0.38 ± 0.137 centimeter-2) was significantly less (p < 0.001) than that preoperatively (0.85 ± 0.375 centimeter-2). After a mean of 6.8 years of follow-up, the relative stress remained essentially unchanged (0.37 ± 0.119 centimeter-2) from the postoperative value (Table I and Fig. 2-A), but it was not as low as the control level. The osteotomy increased the center-edge angle by a mean of 19 degrees (range, 5 to 37 degrees); at the time of the latest follow-up, the center-edge angle remained increased by a mean of 17 degrees (p < 0.001) (Table I).
The postoperative relative load (mean, 4.28) was essentially unchanged compared with the preoperative value (mean, 4.35) (p > 0.05) (Fig. 2-B). However, the mean area of the weight-bearing surface increased significantly postoperatively (12.56 compared with 6.39 square centimeters preoperatively) (p < 0.001) (Fig. 2-C). The five hips in which a concomitant femoral osteotomy had been done were compared with the other seventeen hips. When a femoral osteotomy had been done, there was a significantly greater decrease in the mean relative stress (from 0.99 centimeter-2 preoperatively to 0.35 centimeter-2 at the time of the latest follow-up compared with corresponding values of 0.81 and 0.38 centimeter-2) and in the mean relative load (4.64 to 3.84 compared with 4.27 to 4.40) (p < 0.001 for both comparisons). However, there was no significant change in the mean area of the weight-bearing surface (from 5.67 square centimeters preoperatively to 12.13 square centimeters at the time of the latest follow-up compared with corresponding values of 6.61 and 12.71 square centimeters) (p > 0.05).
Harris Hip Score
All of the patients had relief of pain as well as clinical improvement, and all were satisfied with the results of the procedure. At the time of the latest follow-up, nineteen hips (86 per cent) were rated as excellent or good (80 to 100 points) on the Harris hip scale; one, as fair (70 to 79 points); and two, as poor (60 to 69 points). There were no failures (less than 60 points). The mean Harris hip score was 70 points (range, 39 to 93 points) preoperatively and 93 points (range, 62 to 100 points) at the latest follow-up evaluation (p < 0.001). At the time of the most recent follow-up, the mean relative stresses were 0.42 and 0.45 centimeter-2 for the two hips with a poor score and 0.40 centimeter-2 for the hip with a fair score. For the hips with a good or excellent score, the relative stress averaged 0.36 centimeter-2.
CASE 13. A twenty-five-year-old active woman was first seen because of progressive pain in the right hip. She had no history of hip disease, but the pain had become moderate and a limp had developed, necessitating the occasional use of one crutch. The preoperative radiographs of the pelvis demonstrated moderate dysplasia of the hip with a shallow acetabulum and a center-edge angle of 9 degrees (Fig. 3-A). After a triple osteotomy of the innominate bone, the center-edge angle increased to 28 degrees and the coverage of the femoral head was much better (Fig. 3-B). At nine years of follow-up, adequate coverage had been maintained and the center-edge angle was 24 degrees (Fig. 3-C and Table I). Clinical improvement was demonstrated by a hip score of 96 points (compared with a score of 53 points preoperatively). She reported that she was very active, enjoyed downhill skiing, and had had an uneventful pregnancy and delivery.
At the time of the latest follow-up, the biomechanical analysis of the hip demonstrated no significant change in the relative load, which was 4.16 (compared with 3.61 preoperatively). The area of the weight-bearing surface had increased to 11.65 square centimeters, from 6.93 square centimeters preoperatively, and the relative stress had decreased to 0.36 centimeter-2, from 0.54 centimeter-2 preoperatively. The body weight was 52.3 kilograms (partial body weight, 43.6 kilograms), while it had been 52.4 kilograms (partial body weight, 43.7 kilograms) preoperatively. The absolute stress during single-limb stance was 1.55 megapascals, compared with 2.31 megapascals preoperatively.
It is well recognized that acetabular dysplasia predisposes a patient to early osteoarthrosis3,7,10,12,14,17,18. The goals of pelvic osteotomy for acetabular dysplasia are to create a painless, functional joint and to prevent osteoarthrosis10,12. Whether the operation notably improves the natural history of these hips is unclear. However, a triple osteotomy of the innominate bone appears to accomplish these goals by restoring the normal joint biomechanics and reducing the stress on the hip.
Our results show that a triple osteotomy decreases the stress on the hip specifically by increasing the area of the weight-bearing surface. We found that an increase in the center-edge angle correlated well with a decrease in the relative stress (r = 0.77). The operative technique was designed to maximize the center-edge angle, so as to provide the greatest amount of coverage of the femoral head. Redirection of the acetabulum not only increases the weight-bearing area of the joint toward the superior acetabular margin but also increases it by a corresponding amount in the direction of the acetabular floor. An increase in the center-edge angle effectively increases the weight-bearing surface. In contrast, the over-all load on the hip was not significantly changed after the osteotomy. We know that the hip load can be affected by a change in the abductor moment arm or in the partial body-weight moment arm9. However, because the femoral head is not substantially medialized or lateralized by the triple osteotomy, it follows that the lever arm of the partial body weight, and therefore load, is not altered.
When a femoral osteotomy was performed with the pelvic osteotomy, the abductor moment arm increased. This, coupled with a decrease in the femoral neck-shaft angle, resulted in a decrease in the hip load, and subsequently a greater decrease in the stress, compared with the values after a pelvic osteotomy only (Fig. 2-A). Most of the difference in the decrease in stress is the result of a significant reduction in load (p < 0.001), without a significant alteration in the area of the weight-bearing surface, when a femoral osteotomy is also done (Fig. 2-C). Our indication for a femoral osteotomy is severe dysplasia of the proximal part of the femur (a neck-shaft angle of greater than 155 degrees or abnormal femoral anteversion, or both). It appears that, with dysplasia of the proximal part of the femur, a combination of triple osteotomy and femoral varus-derotation osteotomy is more successful in reducing the stress on the hip than pelvic osteotomy only.
The mean values for relative load, relative stress, and absolute stress in our control group were similar to those found by other investigators. Legal calculated a mean relative hip load of 3.08 and a mean relative stress of 0.20 centimeter-2. For a person who weighs seventy kilograms, this corresponds to an absolute hip load of 1763 newtons (3.08 x 70 kilograms x 5/6 x 9.81 meters per square second) and an absolute stress of 1.15 megapascals. Brinckmann et al. calculated the mean relative and absolute stresses as 0.22 centimeter-2 and 1.45 megapascals, respectively.
Although the data demonstrated that the decrease in the relative stress was associated with a significant improvement in the Harris hip score (p < 0.001), the hip score correlated poorly with the relative stress (r = 0.55). This variance can be explained by the fact that the hips that had a high stress preoperatively did not necessarily have a poor preoperative score; similarly, the hips that had a relatively low stress did not always have a high score. All of the patients were symptomatic before the operation, but six hips had a preoperative hip score that was good or excellent (range, 80 to 93 points). These hips still had acetabular dysplasia that was severe enough to warrant operative intervention. Furthermore, a pelvic osteotomy is sometimes performed for acetabular dysplasia alone, regardless of symptoms, in order to prevent the development of osteoarthrosis (although most dysplastic hips are not seen until after the patient has become symptomatic). Nevertheless, the data demonstrated that the functional outcome was better when the biomechanical goals had been accomplished. An improved hip score was associated with a lower value for stress.
In this study, we had hoped to use biomechanical analysis of relative stress to determine an objective criterion for operative treatment, so that a surgeon could calculate the stress on a dysplastic hip and then determine if the patient would benefit from a pelvic or a femoral osteotomy, or both. Although there was no absolute value to indicate that the stress was too high and an osteotomy was indicated, we are able to make several assumptions on the basis of the data. First, a preoperative relative hip stress of 0.85 centimeter-2 or more correlated with a fair or poor hip score and an operation was necessary (r = 0.86). Second, the hips on which the relative stress was 0.37 centimeter-2 or less at the time of the most recent follow-up had a good or excellent functional outcome, as assessed with the Harris hip score. Preoperatively, the relative stress on all of the hips was at least 0.37 centimeter-2. A relative stress between 0.37 and 0.85 centimeter-2 constituted the so-called gray zone, where it was difficult to make functional correlations. Some hips on which the preoperative relative stress was less than 0.85 centimeter-2 were symptomatic and an operation was necessary; other hips on which the postoperative relative stress was more than 0.37 centimeter-2 still had a good or excellent score. In fact, one patient (Case 4, left hip) had an excellent clinical result (hip score, 100 points), after more than thirteen years of follow-up, despite a relative stress of 0.61 centimeter-2 (absolute stress, 2.80 megapascals). Although the result was excellent, the long-term effects of this particularly high stress on the hip are unknown. We can only speculate that the triple osteotomy prolonged the longevity of the hip by decreasing the stress. A longer duration of follow-up should demonstrate whether a relative stress of 0.61 centimeter-2 or more should be left untreated or if it predisposes the hip to osteoarthrosis.
The decision to treat dysplasia of the hip in adolescents or adults can be a difficult one12,16. Sometimes, a pelvic osteotomy is warranted as prophylaxis against osteoarthrosis when there are no notable symptoms. However, the indications for an osteotomy, the type of osteotomy, and the timing of the operation are yet to be delineated. The present indications for an osteotomy are symptoms, primarily pain, and the radiographic appearance of the hip. The center-edge angle, the acetabular index, and the appearance of the sourcil have been used as objective means to help in the decision-making process3,17. The congruity of the joint, the shape of the femoral head, and the neck-shaft angle are all considered. However, none of these usual radiographic parameters for evaluation of the necessity for a pelvic osteotomy yields an accurate assessment of biomechanical function. Understandably, Legal's model of biomechanical analysis can only approximate the actual stress on the hip. The method relies on a two-dimensional analysis of a three-dimensional problem. The model assumes that the stress is evenly distributed throughout a presumed spherical surface, but osteoarthrosis arises from eccentric loading of the joint and, in a dysplastic hip, the stress is not evenly distributed. Therefore, it is only an assumption that our demonstration of increased stress in the dysplastic hips correlates with, rather than represents, the actual mechanics responsible for osteoarthrosis. Nonetheless, we believe that biomechanical analysis of the hip can be a helpful adjunct in the decision-making process and the preoperative planning for treatment of dysplasia of the hip.
Although the biomechanical and clinical efficacy of the triple osteotomy was demonstrated in this study, follow-up of longer duration is needed to delineate completely the natural history of the hip after a triple osteotomy. However, we observed that triple osteotomy, performed for the indications described here, produces better biomechanics of the hip. Whether the osteotomy extends the longevity of the hip by avoiding osteoarthrosis remains to be seen.