(Note from PFFDvsg Webmaster: Due to limited disk space and scanning problems, the Figures referred to in this document are not available online. My apologies for the inconvenience - hope to get them online soon.)
This is Section 3 of 5 from PFFD: A Congenital Anomaly, National Academy of Sciences, 1969.
HARLAN C. AMSTUTZ, M.D.
Treatment of proximal femoral focal deficiencies (PFFD) depends largely on the anticipated growth and development of the affected limb. A classification scheme based on the morphological natural history of sixty-eight PFFD's in fifty seven patients (Table 1) from the Hospital for Special Surgery, New York; the Royal National Orthopaedic Hospital, London; and the Hospital for Sick Children, London, has been developed to enable the surgeon to prognosticate accurately for the individual case of PFFD and to plan treatment at an early age (4, 5). Thirty-four extremities, followed serially without operative intervention for an average of 7.5 growth years with a range of two to fifteen (with seventeen the arbitrary maximum if the child was seen at birth), form the basis of the natural history study. The total length of comprehensive clinical and radiographic follow-up averaged 10.3 years for fifty-three patients, with four being under treatment for less than one year.
Proximal femoral focal deficiency is defined as the absence of some quality or characteristic of completeness of the proximal femur, including stunting or shortening of the entire femur. The broad terminology has been based on the classifications of Frantz and O'Rahilly (12), Hall et al (13), and Burtch (9). A portion of the distal femur was always present even if only represented by a distal ossicle in the severest form of the anomaly. A coxa vara in addition to a shortening greater than that due to the varus alone is characteristic of even a mild degree of PFFD. Patients with developmental coxa vara (so-called congenital, infantile, cervical) are not included in the series (3).
The present PFFD classification scheme includes five morphologically distinct groups identifiable by roentgenograms at birth, with six subtypes nearly always defined by the fifth year. The diagrammatic representations illustrating the five major groups (Figure 1) typify the conditions for the approximate age range of one to two years; Figure 2 illustrates the subtypes at adolescence. At birth the differences between the groups are more subtle. Each group included a sufficient number of patients to establish a specific natural history. This classification plan includes congenital bowed femur with coxa vara (2), which has not always been included with the PFFD entities (1, 7, 8, 11, 15). The major PFFD groups are otherwise similar to Aitken's categories (1), but additional specificity for Types I and III has been added by subclassification (Figures 1 and 2).
Through the growth years a constant growth ratio of the normal to the abnormal limb was found in 87 percent of the PFFD extremities reviewed. An increase in inhibition of growth of three percent was noted in one patient, while a mean decrease of five percent with a range of three to ten was noted in seven. It is believed that technical mensuration errors in data, analyzed retrospectively, could have been responsible for these deviations. Proportionality of growth was demonstrated in every patient after five years of age. With careful attention to technical factors, precise prognostication will likely be possible as early as two years of age. By that age nearly all capital femoral epiphyses will have ossified, even though the ossification is later than normal.
Since the segments of the abnormal congenitally anomalous limb, as well as those of the normal limb, grow in length at a nearly proportionate rate for each PFFD type, precise length prognostication is possible (4). However, a methodology for accurate mensuration is essential, preferably using, scanogram techniques and correcting for errors in projected length due to contractures. The limb-segment lengths are then plotted on a growth graph against skeletal age, and final lengths estimated.
A progressive coxa vara complicates mensuration and growth prediction. However, a growth ratio based on the length of the femur measured from the lateral femoral condyle to the greater trochanter of the normal and abnormal limbs can be used for prognosticating the growth of the involved portion of the femur, since this segmental proportionality has also been verified.
Associated lower-leg anomalies accompanied 30 percent of Types I, II, and V and 50 percent of Types III and IV and added to or compensated for the shortening present.
TABLE 1 Proximal Femoral Focal Deficiency Patients Reviewed
| PFFD Type |
Total Number of Extremeties |
Bilateral Symmetrical |
Number of Extremeties with Serial Follow-up |
Mean Years Followed |
Number of Extremeties with Serial Follow-up for Length |
Mean Growth Years Followed for Length |
Age Range (Growth Years-17 Maximum) |
Range Inhibition of Growth (%) |
Mean Inhibition of Growth (%) |
| 1A | 13 | - | 13 | 8.3 | 09 | 7.6 | 2-15 | 26.8 | 16-38 |
| 1B | 13 | 2 | 13 | 13.1 | 06 | 8.2 | 3-14 | 29.1 | 20-41 |
| 2 | 04 | - | 04 | 9.8 | 02 | 12.5 | 4-15 | 53.4 | 37-65 |
| 3 | 15(a) | - | 14 | 13.8 | 10 | 5.9 | 3.3-15 | 50 | 35-60 |
| 4 | 09 | - | 09 | 9.3 | 05 | 6 | 2-12.6 | 64 | 36-80 |
| 5 | 14 | 5 | 10 | 7.2 | 02 | 12 | 10-14 | 87 | 79-95 |
| Total | 68 | 7(b) | 63 | 34 |
(a) 1 patient Subtype A, 5 Subtype B, 1 Subtype C, 6 Subtype D, 3 not definitely classified prior to operation (1 probably Subtype 3, 2 probably Subtype C).
(b) 4 patients had bilateral asymmetrical PFFD'S. Total patients included-57.
In PFFD Type V, dysgenesis is so severe that none of the normal precursor hip-joint components, i.e., neither the capital femoral epiphysis nor the acetabulum, is present. Anatomically this condition was confirmed at operation by Mr. Lloyd-Roberts in one nine-month-old patient in this series. The hallmarks of this type are an iliac projection just above the usual acetabular anatomical site, present in eleven of fourteen hip regions, a spherical obturator foramen, and a "box-like" or square pelvis when the defect is bilateral. The severe deficiency extended to the middle and distal shaft in five patients with severe stunting and only a small bone segment that ossified late and represented the distal femoral epiphysis or femoral shaft (Figures 3-A and 3-B). Seven patients, including three affected bilaterally, for a total of 10 extremities, were followed for an average of 7.2 years. In four of these seven patients, a short femoral shaft developed but with severe proximal attenuation. Even though the proximal and distal femoral epiphyses appeared to be absent, the ossified portion of the femur continued to grow, although at a markedly inhibited rate.
Bilateral anomalies were present in six patients, five symmetrically. On three of these we have no follow-up. One bilaterally involved patient also had vertebral anomalies and measured only 3 ft 3 in. in height at maturity. The ankle of the anomalous, unilateral, Type V extremity was positioned near the contralateral normal knee at maturity, permitting placement of a prosthetic functional "knee" center below the foot even if it were not amputated. All Type V patients had positive Trendelenberg and hip instability. The severity of the abductor lurch depended largely on the degree of external rotation and/or abduction contracture and the length of the thigh segment. The greater the external rotation and the longer the thigh segment, the more difficulty the patient has in shifting his center of gravity to provide stability, and the more marked the abductor lurch. One patient had a 90 deg external-rotation contracture so that knee extension and flexion appeared to represent hip "abduction-adduction." When the hip and knee joints are in close proximity, separation of "hip"- and knee-joint mobility and contractures is difficult. However, the true knee seemed to manifest excessive mobility and often developed crepitus in adolescence and adult life, especially when accompanied by significant external rotation and abduction contractures. Despite heavy activity, the patient on whom we have the longest follow-up remains symptom-free with some degree of flexion, abduction, and external-rotation contracture at the age of thirty seven years. The degree of persisting contracture, especially in flexion, appears to reflect, at least partially, the method of prosthetic fitting with the patient "sitting" on his thigh. Flexion contractures averaged 35 deg in the unilateral patients and were more severe when the anomalies were bilaterally symmetrical. Operative findings have revealed considerable abnormality of musculature in these severe dysgeneses. However, the hypermobility and hip instability tend to lessen with age although the positive Trendelenberg remains.
In Type IV the patient's dysgenesis is less severe with formed hip-joint components, although without an arthrogam this is sometimes difficult to ascertain with certainty at birth. An acetabulum and capital femoral epiphysis seemed to be present in all patients on whom roentgenograms were available during the first ten years of life and serve to differentiate Type IV from Type V (Figures 4-A and 4-B). Ossification of the capital femoral epiphysis is often delayed up to two and one-half years. Occasionally the greater trochanter will ossify separately, but ossification may be delayed up to four years. Characteristically the proximal end of the distal femoral shaft tapers sharply, almost to a point. This severe proximal attenuation differentiates Type IV from Type III and represents an unfavorable prognostic sign for naturally occurring ossification of the neck. Proximal migration of the tapered, often sclerotic, femoral shaft occurs because of weakness in the area of persisting pseudarthrosis. The acetabulum becomes dysplastic irrespective of the ossified capital femoral epiphysis. The iliac projection prominently noted in Type V was also present in two of nine Type IV patients. In one patient, a 90-deg external-rotation contracture was noted similar to that seen frequently in the unilateral Type V (Figure 5). Further passive external rotation of the hip was possible to 180 deg, although some of the rotation occurred at the knee.
The inhibition of growth in five unilateral patients averaged 64 percent and remained constant throughout growth for each individual patient but contractures and proximal migration increased the functional inhibition an additional 20 percent (3).
The infant with Type III PFFD appears clinically and roentgenographicauy similar to a mild Type IV in the first year, except that the ossified portion of the femoral shaft has a variable degree of proximal ossific bulbousness. The acetabulum is present, but ossification of the capital femoral epiphysis is often markedly delayed. The subsequent developmental course varied in the fifteen extremities included in this group, necessitating four subtypes (Figure 2), although the size and configuration of the proximal shaft are of diagnostic and prognostic importance.
Type III-A was characterized in one patient by eventual complete ossification and stability of the cervical and trochanteric regions (at the age of five years) without varus progression (Figures 6-A through 6-D). The stability exhibited by the initially wide deficient cervical area is unusual, and is associated with marked bulbousness without attenuation of the proximal femoral shaft.
The distinguishing features of PFFD Type III-B are eventual complete ossification of the capital cervical trochanteric region, with marked varus progression that averaged 7.5 deg per year in four serially followed patients. Complete ossification was achieved by the age of six years for all four of these patients (Figures 7-A, 7-B, and 7-C). In three of the four patients the acetabulum became dysplastic in association with residual varus of 45 deg in two and 40 deg in one.
Spotty cervical ossification occurred in the single Type III-C patient with 115 deg of varus progression when followed to the age of eight years (Figures 8-A, 8-B, and 8-C). It is not likely that sufficient ossification will occur naturally to bridge the defect securely. This special subclassification is warranted when some ossification of a wide neck defect occurs, but the disturbance is so great that eventual complete stable bridging does not occur. This differentiation emphasizes the variability of ossification between III-A, III-B, and III-C types and has important treatment implications.
A wide persisting cervical pseudarthrosis characterizes Type III-D, although the capital femoral epiphysis, and frequently the trochanteric regions, ossifies. In the five patients of this type, three had marked proximal migration of the femoral shaft and progressive dysplasia of the acetabulum (Figures 9-A and 9-B).
Subgroup differentiation in infancy depends both on the qualitative and quantitative bulbousness of the proximal femoral shaft, but the length of the shaft has no prognostic significance. A large, broad, bulbously contoured, ossified proximal femoral shaft present in the first six months of life is likely to develop and morphologically become either Subtype III-A or Subtype III-B. On the other hand, some proximal attenuation of the femoral shaft, particularly if it persists into the third or fourth year, is likely to represent Type III-C or III-D. Careful follow-up is essential for proper subclassification, because a large bulb does not necessarily guarantee that the neck will ossify in some form.,
Three other Type III PFFD patients had osteosynthesizing operations at an early age before precise subtyping was possible, although two seemed to be differentiating into Type III-C and the other into Type III-B.
The stability of the hip of the Type III PFFD varies and is dependent on such factors as neck-shaft continuity, varus, hip contractures, and femoral length. Considerable difficulty can be encountered in analyzing femoral growth in this group because of the varus progression and changeable hip flexion contractures. In the present retrospective study, a proportionate inhibition of growth was found to remain constant in all of the Type III patients serially followed after the age of five years.
Type II PFFD is characterized by subtrochanteric pseudarthrosis and progressive varus, although the capitocervical area also has defective development, as manifested by delayed and irregular ossification. A higher incidence of the persisting subtrochanteric pseudarthrosis characteristic of Type II has been reported in patients where thalidomide had been an etiological factor (8). The morphology of the four patients in this group was somewhat heterogeneous, but the number of patients was insufficient to warrant subclassification. The classification of two short PFFD patients whose capital femoral epiphyses were ossified at one year and three months could not at that time be differentiated between Types II and III. A marked varus progression ensued in one, but the cervical region subsequently ossified with a persisting subtrochanteric defect characteristic of Type II (Figures 10-A, 10-B, and 10-C). The varus progression was not severe in the other patient and was perhaps modified by a drilling procedure performed at the age of two years. The subtrochanteric pseudarthrosis persisted to the age of ten and one-half years.
The subtrochanteric defect in a third patient simulated a pathological fracture through an ossific femoral attenuation. The area appeared to be healing with callus spontaneously when an osteosynthesizing procedure was performed at another hospital utilizing homogenous bone at three months of age. The area healed so that the femur is indistinguishable from Type I with a stable coxa vara and shaft bowing as described below (Figures 11-A through 11-D). A patient with a similar condition has been reported by Badger and Lambert (6). The possibility of traumatic insult at birth is real and had he not fractured he may well have been morphologically similar to one Type I patient with a subtrochanteric defect.
In the fourth patient the proximal portion of the femoral shaft was flattened with a superior tuft of irregular ossified material. His subtrochanteric pseudarthrosis apparently developed through this defective area with subsequent progressive varus. In addition to varus of the neck and subtrochanteric sites, two patients also had varus bowing of the shaft. Functionally these patients varied, but in general the hips were more stable than the more severe dysgeneses, although the Trendelenberg was positive.
Proximal femoral focal deficiency Type I, previously described as congenital bowed femur with coxa vara, is readily distinguishable from Types II and III by a stable, nonprogressive coxa vara and considerable femoral shaft bowing Even in the first year, differentiation is possible because of the bowing of the femoral shaft and the concavity in the proximal-medial subtrochanteric area. There is medial femoral cortical sclerosis and often a laterally positioned beak.
Because of its early appearance, Type I PFFD has often been mistaken for congenital dislocation of the hip, since the shaft is laterally and proximally positioned (Figure 14-A). However. the shortening and bowing, which are often severe, point the way to the correct diagnosis (Figures 12-A, 12-B, and 12-C). In one patient the severe shaft bowing spontaneously diminished from 75 to 5 deg (Figures 12-B and 12-C), but characteristically the lesser trochanter failed to ossify.
The capital femoral epiphysis ossifies and is well centered in the acetabulum. However, ossification may be delayed. Follow-up films on all patients have revealed a stable coxa vara of between 90 and 115 deg except in two where neckshaft angles of 120 and 125 deg were substantially reduced from the normal of 135 and 150 deg. However, a 30-deg midshaft varus produced a composite varus of 90 deg in the patient with the neck-shaft angle of 120 deg. It is important to emphasize that, to verify age changes of varus and bowing (5), serial roentgenograms need to be taken in two planes in a corresponding degree of rotation.
Eleven patients, two with bilateral involvement, were distinguished by the onset of progressive hip dysplasia which altered the prognosis with an early onset of degenerative arthritis and these have been categorized as Type I-B (Figures 13-A, 13-B, and 13-C). For some cases this differentiation was not obvious in early life. In seven patients between the ages of two and six years, dysplasia was noted on the first clinic visit. In two the capital femoral epiphysis was spherical in shape at the age of one and two years, respectively, only to become dysplastic by the time of subsequent follow-up visits between the ages of five and ten years. In ten of the eleven patients in this subtype, the capital femoral epiphysis became conical or bullet-shaped and the hip dysplasia progressed. In only one did the conical shape tend to become more spherical as the patient approached maturity.
All of the Type I-B patients had some abnormality of the combined capito-cervico-trochanteric epiphysis with a persisting (fetal) greater trochanteric-capital-femoral bridge in four (2), and a small epiphysis that ossified on the medial portion of an apparently broadened femoral neck in three. In one patient, marked irregularity of the capital femoral epiphysis was present in the first year of life, reminiscent of epiphyseal dysplasia. Three had irregular vertical epiphyseal lines and two had early acetabular dysplasia.
The neck-shaft angle of 95 deg for PFFD Type I-B was slightly less than the 106 deg of Type I-A, but the overlapping range of varus would indicate that this difference alone is not responsible for the onset of dysplasia in Type I-B. The shaft varus of PFFD Type I-B averaged 11 deg, compared with 14 deg for PFFD Type I-A. The exact etiology of dysplasia is, therefore, not apparent.
Inhibition of growth remained constant after all epiphyses were ossified in thirteen serially followed Type I-A and Type I-B patients, with the mean inhibition being 26.8 and 29.1 percent, respectively.
The treatment of proximal femoral focal deficiencies must be based on a sound morphological analysis and an appreciation of the natural history according to classification type and subtype. The type and timing of operations to improve hip development and stability, and the management of leg length inequality, depend on prognostication based on morphology and anticipated limb-length discrepancy. The indications for valgus osteotomy or osteosynthesis for progressive varus and abductor insufficiency must be based on a realistic appraisal of the functional improvement that can be achieved; that is, whether, if successful, the surgery will advance the patient to a morphological type with less anticipated functional impairment. Surgery, especially about the hip and the proximally deficient femur, has all too often resulted in an undesirable functional result. In this regard, our own experience dictates more often what not to do and, unfortunately, only occasionally what should be done.
The indications and advisability of performing a valgus osteotomy in Type I-A are not clear because of our inability to predict whether or not dysplasia of Type I-B will develop and if and when the varus of the femoral neck alone or possibly in association with retroversion may lead to a conical femoral head development.
Three Type I-A patients who have been followed to maturity with neck-shaft angles of 95, 90, and 105 deg, respectively, including one patient who was 66 years of age, had spherical femoral heads, so dysplasia does not necessarily develop with age.
Three Type I-A patients underwent valgus osteotomy. A 5-deg correction only was achieved in one patient on whom the osteotomy was performed through a subtrochanteric medial defect, with some delay in healing. The pre-existing retroversion was uncorrected, and at the age of twenty-four she has begun to experience a vague ache in her knee although no significant symptoms are thought to emanate from the hip. Two patients who had osteotomies at the age of four and thirteen, respectively, have been lost to follow-up. We do not have sufficient data to recommend or advise against a valgus osteotomy, since it is not certain that surgical restoration of a normal neck-shaft angle and anteversion will prevent the dysplasia and osteoarthritis that accompany Type I-B.
Since osseous continuity of the capito-cervico-trochanteric area will be established at an early age, and the coxa vara is not progressive, osteosynthesizing procedures are not indicated. An autogenous graft was inserted into one hip.before precise morphological typing could be done, although, with the bowing and large bulbous proximal shaft, classification as Type I or III-A was logical (Figures 14-A through 14-D). In retrospect, further development would have established the type more precisely, and, if it had proven to be Type 1, the operation probably could have been avoided.
When the femoral head dysplasia of Type I-B became evident with accompanying
acetabular dysplasia, valgus osteotomies in five hips of four patients were
accompanied by subluxation and further dysplasia with early onset of
osteoarthritis. We would, therefore, advise against the use of valgus osteotomy
alone when dysplasia is present, and it is merely a conjecture that, at the
earliest sign of dysplasia, a Chiari pelvic osteotomy, or some other shelving
procedure in combination with osteotomy, would prevent these changes.
An osteosynthesizing procedure is indicated in the presence of an increasing varus that results from pseudarthrosis in the subtrochanteric area. The timing of the operation is undoubtedly of considerable importance, and the surgery should be performed before the deformity becomes marked and secondary acetabular and femoral dysplasias develop. However, a sufficient amount of time should pass to allow for possible spontaneous ossification and union, even though a corrective valgus osteotomy may be required later. In this respect, decision for osteotomy and bone grafting in one patient might better have been postponed (Figures 11-A through 11-D). The varus might have diminished with growth.
Morphological subclassification is essential in Type III because of marked variability in prognosis from a remarkably similar infantile roentgenographic appearance. The rare Type III-A does not require an osteosynthesizing procedure or an osteotomy to prevent further varus. The same discussion about the indications for valgus osteotomy would apply to this subtype as to Type I-A.
The progressive varus of Type III-B is best treated by a valgus osteotomy. Possibly the operation should be delayed to allow maximum development of the capital femoral epiphysis. The ability to achieve full correction would not be compromised, since early closure of the capital femoral epiphysis is likely to occur with subsequent stunting. The operation should not be delayed, however, if symptoms, function, or impending dysplasia are apparent (Figures 15-A, 15-B, and 15-C). Technically, the interlocking osteotomy provides sufficient mechanical stability to obviate the need for internal fixation and a second operative procedure to remove the fixation. The procedure, as initially described for the progressive varus of the condition which is now preferably called developmental coxa vara, is applicable to this type (2). A strong spike of lateral cortex is fashioned in the distal fragment and mortised into a slot cut in the proximal fragment and controlled by a Steirunann pin (Figure 16). For full correction or stability it may be necessary to transfix through the neck into the head. Mobilization of the fragments by subperiosteal stripping, abductor tenotomy, or detachment of abductor muscles may be necessary. If the mortise-and-tenon joint does not provide sufficient stability, percutaneous pin fixation of both fragments for six weeks is advisable. Anteroposterior and lateral roentgenograms to check the correction should be obtained prior to closure of the wound, and a minimum of ten weeks of immobilization in a double hip spica is absolutely essential.
Osteosynthesizing procedures for Type III-C present complex technical problems. The rudimentary disordered ossification in the femoral neck might be stimulated by a bridging graft combined with valgus osteotomy, although resection of all of the cartilaginous anlage would provide the most certain method of effecting union.
One osteosynthesizing procedure previously reported (14), when the subtyping differentiation between C and D was not yet clearly defined and involving a patient six months of age, led to a massive osseous bridge between the ilium and the greater trochanter which had required two subsequent operations for removal by the time of follow-up examination at five years and eight months of age. The cause of the new bone growth in this patient is not clear. Retrospectively, the degree of varus was not severe and perhaps the initial procedure should have been delayed to observe what further natural ossification would have occurred.
Contractures in this group are often severe, indicating considerable muscle abnormality. We have had no personal experience with casting or traction as advocated by Blauth(8). Howcver, it is technically quite difficult to apply effective traction in a growing child with a short thigh segment. Even if the stress could be effectively relieved, subsequent weight-bearing and the effect of the abnormal muscle forces may lead to recurrence unless the other component abnormalities were stabilized by growth and development or operation.
. A case involving excision of the head and neck and fibular transplant previously reported (2) subsequently sustained dislocation. This procedure is not recommended, since the fibular epiphysis does not continue to grow and thus would not contribute sufficient growth or morphology to be of value.
Osteosynthesizing procedures for Subtype D pose severe technical problems and will undoubtedly fail unless the cartilaginous neck anlage is excised and bone-to-bone contact is secured, as emphasized by Westin (17). This procedure necessitates a 180-deg valgus angle eliminating the femoral neck. These procedures will have to be carefully evaluated and the results compared to those obtained on a similar nonoperated group, since some patients often function surprisingly well without operation even though the hip is somewhat unstable and there is a positive Trendelenberg and abductor lurch.
The severe stunting and marked anomalous musculature about the hip suggest difficulty in achieving a functional improvement even if an osteosynthesizing procedure were a technical success. Severe external rotation and abduction contractures, when present, create a most difficult problem because of marked instability. The person must awkwardly shift his weight far to the side for stability. The short thigh segment has little power in the normal flexion plane and knee extension and flexion appear to represent "hip abduction-adduction." The longer the thigh segment, the farther the weight must be shifted and the greater the abductor lurch. One patient, whose extremity could be externally rotated almost 180 deg passively, was fitted with the foot turned around as popularized by Van Nes, but because of the rotation, undue strain on the knee has occurred. We have had no experience in actually performing derotation osteotomies for this specific purpose, but have had the opportunity of following two performed at other hospitals. The theoretical advantage of converting the extremity to a functional below knee amputation rather than one above the knee was not clinically apparent in these patients who required knee extension assists despite the fact that they could flex and extend the "knee"-by plantar and dorsiflexion of the ankle with good power. However, the hip was quite unstable in one and stiff in the other as a result of operative attempts to fuse the hip. Total evaluation was very difficult. For a girl the poor cosmetic effect of the Van Nes procedure may be unacceptable socially and psychologically. One male patient even developed severe psoriasis. The other patient required multiple operations, including two for complete derotation, two unsuccessful operations to fuse the hip, and a knee arthrodesis. The spontaneous tendency to derotate with growth, requiring further procedures, has been reported. Careful preoperative and postoperative functional evaluation of this procedure must be made to determine its ultimate value.
No hip surgery is indicated for proximal femoral focal deficiency Type V, since normal components are absent.
All patients with proximal femoral focal deficiency, irrespective of type, present significant leg-length discrepancies, with the exception of the rare bilateral Type I (4). The predicted limb-segment discrepancy is determined by multiplying the percent inhibition of growth (difference between normal and abnormal length divided by normal times 100) by the anticipated normal limb-segment length. Serial scanograms are strongly recommended. The limb lengths are plotted on a Green and Anderson graph according to skeletal age. The predicted normal length is determined by transcribing the same spatial relationship of the growth curve to the mean-growth curve line at maturity.
The heights of the patient, obtained serially with sufficient block height to level the pelvis, are also plotted on a graph in order to prognosticate the ultimate height and determine the magnitude of the leg inequality, and to relate this to the child's estimated ultimate height. Recommendations for amputation must be individualized and based not only on the factors discussed above but also on projected intellectual capacity and personal, physical, and emotional needs. We have been impressed by the fact that the development of the child and his adjustment to life with a deformity are strongly dependent on the parents' understanding and attitudes. It is of particular benefit to be able to predict ultimate morphology and leg-length discrepancy and to prepare the parents at an early age for the ultimate treatment. We do not hesitate to perform an early amputation, particularly ankle disarticulation with immediate postsurgical prosthetic fitting and early weight-bearing, when this procedure is clearly the best form of treatment-as when the foot is abnormal with an associated paraxial fibular hemimelia and severe leg-length discrepancy. The amputations have been performed primarily to facilitate prosthetic fitting, since functional improvement in the active, growing child is not always apparent. In adolescence the improved cosmesis achieved through Syme's-type ankle amputations is of particular importance, especially to females, but we have been impressed by how well some of our patients have performed without amputation. Final evaluation of the merits of conversion amputations will require further time to compare all types of PFFD patients with varying degrees of leg-length discrepancies who have been treated by different methods and followed well into adult life. Ultimately we hope to be able to recommend treatment based on the individual physical, mental, and emotional makeup of the patient.
We have had little experience with the combined ankle disarticulation and knee-arthrodesis procedures by which the patient is converted to a functional above-knee amputee with the ankle of the abnormal limb positioned either at the level of the normal knee or between the knee and ankle. We have functionally "arthrodesed" the patient's knee in extension within the prosthesis, resulting in a long-limb segment which has been attended with the usual problems of a fused knee and is tolerated best by short persons. One young PFFD Type III-D woman who is 5 ft 9 in. in height had a marked abductor lurch. Because of pain in the hip, knee, and back, presumably the result of the long extremity lever transmitting stress to the hip and back, a knee arthrodesis and a below-knee amputation were performed to provide her with a flexible prosthetic knee at the level of the normal knee.
Now that we are able to predict ultimate leg-length discrepancy (4), the surgical treatment could be performed at the appropriate age so that a femoral or tibial epiphysiodesis, or both, could be done at the time of knee arthrodesis and a Syme's-type end-bearing amputation rather than a below-knee amputation could be performed. A moderately severe proximal femoral focal deficiency in which it was predicted that the ankle would otherwise be opposite the middle of the contralateral normal leg would then mature with the prosthetic knee at the same level as the contralateral knee. This type of planning would result in a good end-bearing amputation stump of the ankle disarticulation type as a definitive procedure. Recently we utilized this approach for a PFFD Type III-D with a 46 percent inhibition of femoral growth on whom the greater trochanter and capital femoral epiphysis were excised at eight months and the fibular head unsuccessfully transplanted into the femur. An ankle disarticulation and knee arthrodesis with distal femoral and proximal tibial epiphysiodesis were performed two weeks apart at the age of ten years and eleven months. The extremity was immobilized in a rigid plaster dressing and a pylon attached. The technique of knee arthrodesis and epiphysiodesis using a triangular template as a guide is illustrated (Figure 17).
The timing of the procedure was based on a growth analysis, using the Green and Anderson growth graph (Figure 18). By projection on the graph, the estimated length of the normal femur at maturity was 45.5 cm. Therefore, the arthrodesed femur and tibia of the abnormal limb should have a functional length about five centimeters less to allow space for the prosthetic knee-joint mechanism.
The anticipated length of the unaltered tibia, assuming that the growth pattern would continue along a line about equidistant between the mean and one standard deviation below it, was 37 cm. Transcribing the same spatial relationship of the patient's growth graph line for the tibia on the Green and Anderson graph for remaining growth of the proximal tibia, the length could be diminished 3.75 cm by epiphysiodesis at skeletal age 10 years plus 11 months. The net tibial length would then be approximately 32.2 cm.
The functional length of the abnormal femur was 15.2 cm at the time of operation (18.2 cm length minus 3 cm of proximal migration). The arthrodesis reduced the length an additional 4.3 cm so that the estimated combined length of the extremity will be 43.1 cm (15.2 - 4.3 + 32.2). Our growth studies indicate that less than anticipated growth may occur following ankle disarticulation, although more data must be accumulated to substantiate this indication. Therefore, it is possible that the limb segment may be shorter than predicted and allow enough space for the knee joint mechanism below the long end-bearing Syme's-type stump at maturity and still have level knee centers. The technique of arttirodesis and the final result will merit further study.
When the thigh segment is longer and the leg-length discrepancy is less, preservation of knee function and swingphase control may be desirable despite the long prosthetic lower-limb segment which results. In such cases this approach might be preferable to knee fusion, and the choice would depend not only on the length of the thigh segment but also on the patient's height. Again a short person is functionally better able to accommodate a prosthetically lengthened lower-limb segment in many pursuits ranging from riding a bicycle to sitting in the theater. Undoubtedly, each PFFD patient must be analyzed individually, and the results of knee arthrodesis must be carefully followed.
As yet, there are no reports of the effect of increased stress on the abnormal hip resulting from knee arthrodesis. If fusion is performed, it is recommended that it be done with the knee in extension. Favorable reports indicate (17) that the hip contractures diminish in time. This result is most desirable for improvement of gait.
Knee arthrodesis of the externally rotated and abducted limb would increase the distance from the center of gravity as well as the amount of lateral shift and abductor lurch necessary to balance the unstable hip unless the contractures were corrected.
We have been gratified by the results obtained from fitting two bilaterally asymmetrical PFFD patients with double extension prostheses. The first patient had bilateral paraxial fibular hemimelias with two- and three-rayed feet in addition to PFFD Types III-B and III-D. Ankle disarticulations were performed with immediate postsurgical prosthetics fitting, and the patient's height was gradually increased from 3 ft 10 in. to 5 ft 2 in. to equal her span. At present she walks without support and has demonstrated remarkable improvement in social and intellectual achievement at school. Similar social improvement was noted in another patient who had a knee disarticulation for Type II PFFD and a contralateral Type I-A PFFD.
Classification of proximal femoral focal deficiencies based on the natural history of specific types and subtypes is essential for prognostication of ultimate morphology. The deficient limb segment follows the rule of proportionality of growth in relation to the normal limb segment. Serial scanograms and determination of skeletal age enable the physician to predict ultimate leg-length discrepancy, plan treatment at an early age, and perform surgery at an appropnate age based on individual intellectual, personal, physical, social, and psychological needs.
The methods of treatment of the hip instability, contractures, and leg-length discrepancy associated with PFFD are described, and recommendations, where possible, have been made for specific types and subtypes. However, many forms of treatment which are now being applied to these challenging problems in this clinic and others around the world require continued careful analysis and follow-up. It is our hope that operative procedures for the PFFD patient will be based on a most careful morphological analysis and review of the experiences of many surgeons.
The author would like to thank Mr. George Lloyd-Roberts, F.R.C.S., and Sir Herbert Seddon, and their respective hospital staffs for allowing him to review the records of their patients.
1. Aitken, G. T. Proximal femoral focal deficiency. Personal communication.
2. Amstutz, H. C., and P. D. Wilson, Jr. Dysgenesis of the proximal femur (coxa vara) and its surgical management. J. Bone Joint Sure. 44A(l): 1-24, 1962.
3. Amstutz, H. C. Developmental coxa vara-a distinct entity. To be published.
4. Amstutz, H. C. Prognostication of length for congenital anomalies of the lower limbs. To be published.
5. Amstutz, H. C. The natural history and classification of congenital anomalies of the femur. To be published.
6. Badger, V. M., and C. N. Lambert. Differential diagnosis of an apparent proximal focal femoral deficiency. ICIB V(l): 3-9, October 1965.
7. Bevan-Thomas, W. H., and E. A. Miller. A review of proximal femoral focal deficiencies. J. Bone Joint Surg. 49A(7):13761388, 1967.
8. Blauth, W. Der Kongenitale Femurdefekt. Stuttgart: Ferdinand Enke Veriag, 1967.
9. Burtch, Robert L. Nomenclature for congenital skeletal limb deficiencies: A revision of the Frantz and O'Rahilly classification. Artif. Limbs 10(1):24-35, 1966.
10. Dunn, A. W., and G. E. Aponte. Congenital bowing of the tibia and femur-Case report with autopsy rmdings. J. Bone Joint Surg. 44A(4):737-740,1962.
11. Fock. G., and M. Sulaman. Congenital short femur. Acta Orthoped. Scand. 37:294-300, 1965.
12., Frantz, C. H., and R. O'Rahilly. Congenital skeletal limb deficiencies. J. Bone Joint Surg. 43A(8):1202-1224, 1961.
13. Hall, C. B., M. B. Brooks, and J. F. Dennis. Congenital skeletal deficiencies of the extremities. J.A.M.A. 181:590-599, 1962.
14. Lloyd-Roberts, G. C., and K. H. Stone. Congenital hypoplasia of the upper femur. J. Bone Joint Surg. 45B(3):557-560, 1963.
15. Shands, A. R., Jr., and G. D. MacEwen. Congenital anomalies of the femur. Acta Orthoped. Scand. 32:307-314, 1962.
16. Van Nes, C. P. Rotation-pluty for congenital defects of the femur. J. Bone Joint Surg. 32B: 12-16, 1950.
17. Westin, W. Proximal femoral focal deficiency. Personal communication.
This is Section 3 of 5 from PFFD: A Congenital Anomaly, National Academy of Sciences, 1969.