Vol 7, No 3, May/June 1999 143 Assessment of a skeletally imma- ture patient with a limb-length dis- crepancy and formulation of a treat- ment plan require an understanding of the etiology of the disorder and the natural history of the condition, as well as the ability to predict the ultimate discrepancy at maturity. The purpose of this article is to pro- vide a systematic approach to the patient with limb-length inequality and to discuss the potential pitfalls of assessment and the options for treatment. Mechanisms of Compensation Limb-length inequality is common in the general population. 1 A vari- ety of mechanisms are used to compensate for the resultant gait asymmetry. 2-4 Adults tend to walk in a plantigrade fashion, ÒvaultingÓ over the long leg. Children may use either this mechanism or toe walking on the short side, which levels the pelvis and decreases the effective trunk sway during gait. Despite the prevalent belief that limb-length discrepancy may be deleterious to the spine or the hip, there is little evidence to support this assumption. While increased trunk shift, vaulting, and toe-walking all increase the energy expenditure involved in walking, these mecha- nisms appear to have little effect on otherwise-healthy individuals. The data relating to the possibility that limb-length discrepancy causes low back pain in adults are contradic- tory. 5,6 Back pain is usually not a complaint in children with limb- length discrepancy. The effect of limb-length inequality on spinal alignment and the hip can be noted only when the individual is bearing weight equally on both legs. The effort to produce an erect trunk results in functional scoliosis. 7 According to the literature, howev- er, the convexity of the curve is vari- able. 8 The center-edge angle of the hip of the long leg will be decreased due to the compensatory pelvic obliquity. The long-term effects of these functional changes are undoc- umented and largely speculative. In the course of normal daily activity, most people spend little time stand- ing on both legs with their weight evenly distributed. The significance of limb-length differences remains controversial. In general, individuals with con- genital or acquired inequalities that have developed over the course of many years accommodate more readily than those with acute ac- quired differences due, for exam- ple, to trauma. 2,3 The literature suggests that individuals with limb-length disparities less than 2.0 to 2.5 cm usually require no active intervention or, at most, a shoe lift. 1 Dr. Stanitski is Professor, Department of Orthopaedic Surgery, Medical University of South Carolina, Charleston. Reprint requests: Dr. Stanitski, Department of Orthopaedic Surgery, Medical University of South Carolina, Suite 708, 96 Jonathan Lucas Street, Charleston, SC 29425. Copyright 1999 by the American Academy of Orthopaedic Surgeons. Abstract Assessment and treatment of limb-length inequality, particularly in the grow- ing child, is a challenging task. Evaluation of the discrepancy requires an understanding of the significance of the disparity, as well as the natural history of the disorder, before formulation of a treatment plan. In the immature patient, consistent longitudinal data are essential to avoid pitfalls in the projection of ultimate length difference. Therapeutic options range from no treatment or use of a simple shoe lift to a surgical shortening or lengthening procedure. The cur- rent indication for lengthening is a disparity exceeding 5 to 6 cm. Epiphys- iodesis or femoral shortening is useful for smaller discrepancies or for residual differences following a contralateral lengthening. Lengthening is done with a circular or cantilever external fixator, which may be combined with an intramedullary rod. J Am Acad Orthop Surg 1999;7:143-153 Limb-Length Inequality: Assessment and Treatment Options Deborah F. Stanitski, MD Clinical Assessment The causes of limb-length inequality are summarized in Table 1. While not exhaustive, this list includes the most commonly seen entities. The patientÕs history will most often elucidate the etiology of the limb-length discrepancy, whether congenital or acquired. The family history may be helpful in identifying inherited disorders, such as neurofi- bromatosis and multiple hereditary exostoses. The birth history and time of onset may be important. Dis- crepancies noted at birth are most commonly due to the congenital hypoplasia syndromes. Hemihyper- trophy, Klippel-Trenaunay-Weber syndrome, Proteus syndrome, and neurofibromatosis are frequently noted in the perinatal period. The occurrence of generalized sepsis, a septic joint, or osteomyelitis can be a contributing factor. Other common causes of acquired deformities are trauma, inflammatory disorders, and neurologic injury. Skin examination may reveal vascular or pigmentation abnor- malities or scarring. Abnormalities overlying the spine, such as a dim- ple, sinus, or hairy patch, should prompt investigation of the under- lying spine and spinal cord. Ex- amination of the limbs should re- veal differences in size and muscle strength. In hemihypertrophy and hemangiomatous conditions, ab- normalities may be confined to the lower extremity or may involve the entire side of the body. With the patient supine, the lower extremities should be fully extended with the pelvis level to best assess the relative amount of shortening. Tape measurement is generally useless due to the im- precision of finding reproducible landmarks, particularly at the anterior superior iliac spine. 5 If no difference in the limbs can be appreciated clinically by noting the relative relationship of the medial malleoli, the difference is usually small and may be insignif- icant. The Galeazzi test should be performed by flexing the hips 90 degrees and noting relative knee height (Fig. 1). This will elucidate whether limb-segment involve- ment is femoral or tibial. The patient with limb-length inequality should then be examined while standing with blocks placed under the short leg to level the pelvis. This gives the examiner a reasonably accurate clinical mea- surement of limb-length inequality, including the potential contribution of the foot height. Palpation of the iliac wings and observation of the two posterior dimples overlying the sacrum can also be helpful. With the pelvis level, the spine is exam- ined for evidence of frontal- or sagittal-plane deformity. Coexistent spinal deformity can be identified by examining the spine with the patient seated, which eliminates any potential contribution of limb- length difference. The contribution of foot height to limb-length dis- crepancy is assessed clinically by measuring the distance from the floor to the medial malleolus with the patient standing. This is espe- cially helpful in virtually all con- genital conditions distal to the knee in which foot height is reduced on the affected side. Examples of this are the fibular hypoplasia syn- dromes and congenital posterome- dial bowing of the tibia. Motor and sensory examinations should be performed to rule out any neuromuscular abnormalities. Joint range of motion and stability should be assessed clinically and abnormalities, such as contractures, noted. A flexion contracture of the knee or hip will produce a func- tional limb-length inequality. Hip adduction or abduction contrac- tures will also produce a functional limb-length inequality. The patientÕs gait should be examined while walking bare- foot and also with any shoes, lifts, or orthoses. Observation of rapid walking or running is use- ful to magnify mild gait asym- metries. Limb-Length Inequality Journal of the American Academy of Orthopaedic Surgeons 144 Table 1 Causes of Limb-Length Discrepancy Congenital causes Limb hypoplasia syndromes Proximal Proximal femoral focal deficiency Congenital short femur Hypoplastic femur Distal Fibular hemimelia Tibial hemimelia Congenital posteromedial bowing Hemihypertrophy or atrophy Idiopathic Klippel-Trenaunay-Weber syndrome Proteus syndrome Skeletal dysplasias Ollier disease Fibrous dysplasia Multiple hereditary exostoses Neurofibromatosis Chondrodysplasia punctata Acquired causes Trauma Acute bone loss Physeal fracture Burns Irradiation Iatrogenic Infection Osteomyelitis Septic arthritis Purpura fulminans Inflammation Juvenile rheumatoid arthritis Hemophilia Pigmented villonodular synovitis Neurologic Closed head injury Polio Spinal cord injury or tumor Peripheral nerve injury Myelomeningocele Cerebral palsy Radiologic Assessment A variety of radiologic techniques are available for the assessment of limb-length discrepancy. Past stan- dards have been scanography, orthoradiography, and teleradiogra- phy. The teleradiograph is a single exposure on a long 14×36- or 14×54-in film, taken from a 6-ft distance with the patient standing with a ruler placed (ideally) in the center of the cassette. It has the advantage of demonstrating axial deformity but is subject to magnification error. In the authorÕs experience, this aver- ages 6% and can be easily calculated by using a magnification marker of known size on the film. Another advantage of this technique is the demonstration of frontal-plane deformity as well as limb-length discrepancy on one film. Ortho- radiography avoids magnification by using separate exposures of the hip, knee, and ankle. 9 Scanography follows the same technique as orthoradiography, but the film size is reduced by moving the cassette beneath the patient between expo- sures. The difficulty with the latter two techniques is that patient move- ment between exposures produces measurement error. All three tech- niques are inaccurate if there is a fixed hip- or knee-flexion contrac- ture. In the past decade, computed tomographic (CT) scanogram tech- niques have been reported by a number of centers. The images ob- tained entail considerably less radi- ation exposure than conventional radiographs, 10,11 but they have not been shown to be more accurate, except in patients with a significant knee-flexion contracture. 10 De- pending on the institutional avail- ability of CT, the study may need to be scheduled for a second visit. The CT study is more expensive than a standard radiographic exam- ination (e.g., approximately $620 in our institution for technical and interpretation fees, compared with $120 for a teleradiographic study). Ultrasound has been utilized as a tool for assessment of limb lengths. Although it has the benefit of being performed without the use of ioniz- ing radiation, it is less accurate than standard radiologic techniques and may be useful only as a screening tool. 12 A variety of pitfalls are present in the projection of limb-length dis- crepancy in a child. Many of these are directly related to the vagaries of the various radiologic techniques. Regardless of the method chosen, the same type of examination (e.g., scanography) should be performed at each visit, preferably in the same radiographic suite to provide stan- dardization of technique. Skeletal age determination based on comparison with the Greulich and Pyle atlas has traditionally been used along with lower-extremity radiographs to predict ultimate limb-length discrepancy in children. This technique has two inherent flaws. The first is that the bone age obtained is accurate only within approximately 12 months, and bone ages are notoriously inaccurate before the age of 6 years. Ulti- mately, however, if evaluations are done sequentially over a number of years, the intrinsic inaccuracy is reduced. For example, if distal femoral and proximal tibial epi- physiodeses were performed in an adolescent and the bone age deter- mination was in error by 12 months either way, the maximum resulting disparity would likely be no more than 16 mm (10 mm/yr for the dis- tal femur, 6 mm/yr for the proxi- mal tibial physis). From a practical point of view, this is probably not a serious concern. The second flaw is related to bone age determination on the basis of measurement of the left hand and wrist. In conditions in which the left is the abnormal side (e.g., hemihypertrophy and hemi- atrophy), there may be a consequen- tial difference between the bone Deborah F. Stanitski, MD Vol 7, No 3, May/June 1999 145 Fig. 1 The Galeazzi sign signifies shortening of the thigh segment, which may be sec- ondary to hip dislocation or femoral shortening. The Galeazzi test is performed with the patient supine, hips flexed 90 degrees, and knees flexed. The relative relationship of the knee heights can then be assessed. (Adapted with permission from Tachdjian MO: Pediatric Orthopaedics, 2nd ed. Philadelphia: WB Saunders, 1990, vol 1, p 326.) ages as determined on the left and right sides. Radiographs of both hands should be obtained and compared in this situation. Prediction of Discrepancy In the skeletally mature individual, there is no need to analyze sequen- tial data, as the situation is static. The growing child, however, pre- sents a challenge in predicting the need for treatment and selecting from the variety of treatment op- tions. The importance of obtaining reproducible data cannot be over- emphasized. Currently, there are four different methods incorporat- ing three techniques for the predic- tion of limb-length discrepancy: the arithmetic method, the Eastwood- Cole method, the Green-Anderson growth-remaining method, and the Moseley straight-line graph meth- od. 13-16 The potential accuracy of any of these methods is enhanced by hav- ing longitudinal data. Obtaining data at 6-month to yearly intervals over a number of years is much more helpful than using numerous data points over a relatively short time frame. The same technique should be used for each radiographic assessment to avoid the vagaries of magnification. The fact that there are a number of recognized patterns of limb-length discrepancy, as de- scribed by Shapiro, 17 further empha- sizes the importance of minimizing error. The arithmetic method, or rule- of-thumb method, was first de- scribed by White and evaluated by Westh and Menelaus. 13 It is based on four assumptions about growth: (1) boys stop growing at age 16; (2) girls stop growing at age 14; (3) the distal femoral physis grows 10 mm yearly; and (4) the proximal tibia grows 6 mm yearly. This method is useful only during the later years, not in young children. A potential disadvantage lies in using chronologic rather than skele- tal age, which may present prob- lems in assessing individuals who mature very early or very late. Eastwood and Cole 16 published a scheme using a graphic arith- metic method. These data were confirmed with CT scanning and skeletal age measurements in mid- dle and late childhood. Reference slopes indicate the most appropri- ate time for epiphysiodesis. Using this technique, the authors pre- dictably achieved limb-length equality within 1 cm. The growth-remaining method is based on growth tables pub- lished by Green and Anderson. 15 Graphs relate the limb lengths of boys and girls to chronologic age and can be used to determine a childÕs growth percentile. Other graphs demonstrate the remaining proximal tibial and distal femoral physeal growth and can be used to predict the effect of epiphysiodesis. Because only the most recent skele- tal age determination is used, any innaccuracy in its assessment will cause the resultant estimation of limb-length discrepancy to be prone to imprecision. This method has the greatest longevity of use but is cumbersome due to the ne- cessity of referring to two sets of graphs. The straight-line graph method described by Moseley 14 is a distilla- tion of the Green-Anderson data graphically displayed in a straight line over time on a single graph. It is based on two principles: (1) a nomogram can be used to deter- mine the growth percentile from limb length and skeletal age, and (2) the growth of both limbs can be represented graphically by two straight lines. The difference in slope between the long and short limbs indirectly represents the growth inhibition (or stimulation) of the abnormal extremity. An advantage of this method is that a single-page graph represents the entire limb-growth history. In addi- tion, the vagaries of interpreting skeletal age studies and their intrin- sic inaccuracy become less impor- tant over a number of estimations. In a recent study, Little et al 18 compared the accuracy of the Anderson-Green, Westh-Menelaus, and Moseley methods of predicting limb-length discrepancy. No im- portant differences were revealed. Disparities of up to 2.5 cm in the foot height itself can be seen in patients with congenital limb shortening. To date, no radio- graphic measurement method that provides a reproducible standing foot height has been described. Any clinical measurement discrep- ancy should be added to the ulti- mate projected limb discrepancy. Accuracy is enhanced by having a single observer remeasure values on all radiographs, regardless of the source. Using measurements taken from different reference points creates unnecessary errors. With accurate longitudinal data, the goal of producing reasonable limb symmetry with accuracy with- in 1 cm should be readily achiev- able. If the data are inadequate, inaccurate, or confusing, an epi- physiodesis should be avoided, and another method of limb equaliza- tion should be selected at skeletal maturity. Treatment Options The broad spectrum of therapeutic options available for the patient with a limb-length discrepancy includes no treatment at all; simple shoe modification; shortening proce- dures, such as percutaneous epi- physiodesis (Fig. 2) and intra- medullary shortening (Fig. 3); lengthening procedures, and combi- nations thereof. It is essential to establish the goals of treatment before embarking on any of these Limb-Length Inequality Journal of the American Academy of Orthopaedic Surgeons 146 options. In general, these goals are equal limb lengths, normal axial alignment with a level pelvis, and enhanced function. These goals may be modified, depending on var- ious clinical variables. The patient with a stiff knee or hip or weakness of the involved extremity should be left slightly short on that side to allow the foot to clear the floor in swing phase without the need for circumduction or excessive Òhip hike.Ó In patients with a fixed pelvic obliquity, functional and actual limb lengths may differ significantly. If the pelvic obliquity cannot be elimi- nated, functional limb-length equali- ty should be the goal. Data obtained by Gross 1 and others suggest that projected dis- crepancies of less than 2 cm require no treatment. In a recent article, Kaufman et al 2 demonstrated by gait analysis that subjects with a limb-length disparity of less than 2.0 cm had no greater gait asymme- try than the general population. Song et al 3 reported increased work done by the long side and greater vertical displacement of the center of body mass in patients with dis- crepancies greater than 5.5% com- pared with the opposite limb. In general, patients whose ulti- mate inequality will be in the range of 2 to 6 cm should undergo a short- ening procedure, either by epiphys- iodesis or femoral shortening. There are several potential excep- tions. One is the patient in whom the short extremity has a major angular deformity. In such a case, simultaneous deformity correction and lengthening should be consid- ered. Another possible exception is the patient with pathologically short stature in whom further height reduction would compromise func- tion. Yet another potential excep- tion is the patient with shortening below the knee who presents either at maturity or too late for an epi- physiodesis and in whom contralat- eral femoral shortening would pro- Deborah F. Stanitski, MD Vol 7, No 3, May/June 1999 147 A B C D E Fig. 2 A, Percutaneous drilling of the dis- tal femur is performed from both the medi- al and the lateral sides. B, Curettage is then performed to remove all growth carti- lage. C, An anterior approach to the proxi- mal fibular physis provides direct visual- ization and avoids potential peroneal nerve injury. The incision can then be utilized to drill and curette the lateral proximal tibial physis. D, As in the distal femur, both medial and lateral approaches to the proxi- mal tibial physis are recommended to ensure symmetrical growth arrest. E, Introduction of contrast material confirms adequate physeal excision. duce additional knee-height asym- metry. A review of the literature indicates that there is no functional or cosmetic disability as a result of knee-height disparities of less than 4 cm. If the difference is greater than this, lengthening of the involved tibia may be preferable. The patient with a discrepancy exceeding 5 to 6 cm is best treated by limb lengthening or a combina- tion of limb lengthening and con- tralateral shortening. Limb abla- tion and/or prosthetic fitting should be reserved for patients whose problems are unmanageable by current surgical techniques. Shoe Modification A shoe lift remains an excellent treatment for small discrepancies. Unfortunately, even with the new lightweight orthotic materials, all shoe lifts render the sole stiff. Tapering at the toe is necessary to approximate normal gait. This is the least morbid and least expen- sive method of limb-length equal- ization and is preferable for patients with discrepancies of less than 2.0 to 2.5 cm. Nearly half of the dispar- ity can be accommodated inside the shoe, which may be sufficient to provide adequate patient comfort. Although modern orthotic technol- ogy has decreased shoe-lift weight, most patients with larger discrepan- cies shun the lift because of cosme- sis and prefer a surgical option despite the potential morbidity. Shortening Procedures Epiphysiodesis and acute femoral shortening are both length-reducing procedures. In the growing child with adequate longitudinal data, normal axial alignment, and a pro- jected discrepancy of between 2 and 5 cm, epiphysiodesis remains the procedure of choice. Various techniques have been described, including epiphyseal stapling and the Blount and Phemister tech- niques. Epiphyseal stapling should be used cautiously. In order to pro- duce physeal arrest, three medial and three lateral staples are placed in the distal femur and the proximal tibia. The most common complica- tion reported is staple extrusion. 19 The method currently preferred is the percutaneous technique ini- tially reported by Canale et al. 20,21 Small medial and lateral physeal incisions allow percutaneous drilling, followed by physeal curet- tage under image intensifier con- trol (Fig. 2). Postoperative immobi- lization is not required. Excellent and reproducible results have been achieved with this technique. 21-23 The choice of limb segment (i.e., distal femur or proximal tibia or both) should be selected primarily on the basis of the location of the contralateral shortening. If the shortening is idiopathic, both limb segments will be involved. Under these circumstances, knee height Limb-Length Inequality Journal of the American Academy of Orthopaedic Surgeons 148 A B C D E F Fig. 3 In intramedullary shortening, the intramedullary canal is first reamed over a guide wire. A cam saw of appropriate size is then introduced into the femoral diaphysis (A) and deployed gradually while being rotated to produce an osteotomy (B). The saw is then moved the appropriate distance to achieve the amount of shortening desired proximally, and the procedure is repeated (C). The saw is removed, and a J-shaped osteotome is inserted to split the intercalary segment (D). This must be done twice, ideally at 180 degrees with respect to each longitudinal osteotomy (E). The guide wire is reintroduced, the femur is shortened, and the intramedullary nail is inserted (F). symmetry will be maintained if epiphysiodesis is performed on both the distal femur and the proxi- mal tibia. Acute tibial shortening has major potential complications, including nonunion and compartment syn- drome, 24,25 which preclude its com- mon use for limb-length equaliza- tion. Femoral shortening is useful for patients who present after matu- rity and for those with insufficient data or inadequate growth remain- ing for an epiphysiodesis. The two basic described techniques are closed intramedullary shortening, as de- scribed by Winquist 26 and Kempf et al, 27 and open subtrochanteric short- ening performed with use of either a blade plate or large-fragment plate fixation. An intramedullary saw is used for the first technique, with dia- physeal osteotomies, splitting of the intercalary segment, and insertion of a locked intramedullary rod (Fig. 3). This method is technically de- manding, requiring familiarity with the instrumentation. Its success depends on several anatomic as- sumptions that may not be true. The cam-deployed saw works in a circu- lar fashion, but the femur is not always cylindrical and of uniform thickness throughout its circumfer- ence. A small incision may be re- quired to complete the osteotomy. There are concerns as well about the use of this technique in adolescents because of reports of osteonecrosis of the hip after femoral nailing. 28,29 The open subtrochanteric tech- nique is generally easier than the diaphyseal one. Fixation can be achieved by using either a blade plate or a contoured conventional plate (Fig. 4). Nordsletten et al 30,31 have demonstrated a possible max- imum of 10% length reduction in middiaphyseal shortening as opposed to subtrochanteric short- ening. In their experience, thigh muscle strength never returned to normal in patients with diaphyseal shortening greater than 10%. This suggests that the open proximal technique of shortening may be a more physiologically sound proce- dure than closed intramedullary diaphyseal shortening. Limb Lengthening Lengthening has significantly evolved over the past decade in North America due to the introduc- tion of the Ilizarov technique. 32,33 The biologic principles of gradual incremental distraction have con- tributed greatly to the ability to form excellent bone in the distrac- tion gap while avoiding the prob- lems of the need for bone graft and plate fixation, which plague the Wagner and other techniques. Despite the improvements in gradual-distraction lengthening techniques, the complications of limb lengthening exceed those of epiphysiodesis or acute shortening. These include joint contracture, joint subluxation or dislocation, muscle weakness, vascular injury, nerve palsy, bone regenerate defor- mation, and pin-site infection. 34,35 Limb lengthening is indicated for length discrepancies exceeding 5 to 6 cm and those associated with significant angular and/or rotation- al deformity of the short extremity. Limb lengthening can be easily combined with epiphysiodesis as part of the overall strategy for man- agement of limb-length inequality. For example, if a patient with con- genital limb hypoplasia has a pro- jected discrepancy of 18 to 20 cm and is a reasonable candidate for limb elongation, two lengthenings plus a contralateral epiphysiodesis may be a more reasonable strategy than three lengthening procedures. Deborah F. Stanitski, MD Vol 7, No 3, May/June 1999 149 Fig. 4 A, Preoperative scanogram of a skeletally mature 28-year-old woman with 3.3 cm of left femoral shortening due to a previous fracture. B, Open subtrochanteric shortening of the right femur was performed. Fixation was achieved with use of a 90-degree adoles- cent blade plate. A B The currently utilized technique involves a percutaneous osteotomy, with care to avoid periosteal strip- ping, followed by gradual incremen- tal distraction. 32,33 This is accom- plished with the use of external skeletal fixation. Lengthening with temporary external fixation over an intramedullary nail may be used in selected circumstances (Fig. 5). 36 The external fixator may be either a multiplanar (circular) or a uniplanar (cantilever) type. Bone fixation may be achieved with transosseous ten- sioned wires, half pins, or a combi- nation of both, depending on the fix- ator type. Circular, Ilizarov-type fixators allow application to almost any limb segment or size and can be adjusted to correct angular, rota- tional, and translational deformi- ties as well as to achieve lengthen- ing (Fig. 6). The devices can be extended to adjacent limb seg- ments when necessary to protect potentially unstable joints during lengthening and to avoid tendon contracture. However, Ilizarov fix- ators are neither user- nor patient- friendly. There is a steep learning curve before one can consistently avoid iatrogenic errors and major complications related to their use. 34 Uniplanar devices are easier to apply and are usually well tolerated by the patient. Due to their configu- ration, they have some limitations in application in small patients and in patients with multifocal or multi- planar deformities (Fig. 7). Align- ment adjustment in the pediatric patient usually requires general anesthesia. Lengthening of the fe- mur with a uniplanar device causes elongation along the anatomic bone axis, producing medialization of the knee. 33,36 Because the extent to which this occurs is dependent on the extent of lengthening, this factor should be considered before choos- ing a cantilever device. Lengthening over an intra- medullary nail probably has its greatest application in the mature patient. The advantage of this tech- nique is limiting the time of exter- nal fixation. 36 Once the desired length has been achieved, the nail is locked distally, and the external fix- ator is removed. The most signifi- cant potential risk is intramedullary sepsis due to communication of Limb-Length Inequality Journal of the American Academy of Orthopaedic Surgeons 150 Fig. 5 A, Radiographs of a 21-year-old man who was injured in a lawn-mower accident at age 2. Multiple surgical procedures, including a left knee arthrodesis, resulted in an 8-cm limb-length inequity. B, Lengthening of the femur over a proximally locked femoral nail was initiated through a subtrochanteric osteotomy. A cantilever fixator was used. C, Radiographic appearance at the conclusion of gradual distraction to achieve lengthening by 6 cm. The nail was locked distally, and the external fixator was removed. D, Radiographic appear- ance after consolidation of the distraction gap. (Courtesy of John E. Herzenberg, MD, Baltimore.) A B C D external fixator pins with the intra- medullary device. Juxta-articular deformity (such as in the distal femoral metaphysis) cannot be easi- ly corrected with this technique because the nail ascends within the femur during lengthening. How- ever, diaphyseal deformity can be easily corrected acutely prior to nail insertion. A completely implantable inter- nal lengthening device would be ideal. The Albizzia nail works by a ratchet mechanism. 37 The nail is implanted and locked proximally and distally. Rotation of the pa- tientÕs lower extremity creates dis- traction with an audible click. This device is currently under develop- ment and is considered experimen- tal in North America. A hydraulic mechanism would, in theory, be advantageous to eliminate the rota- tion necessary with this system. Controversies in Limb Lengthening The ability to lengthen a limb is now no longer limited by the abili- ty to produce bone that will heal reliably. Soft tissues and joint sta- bility currently limit the ability to lengthen a limb and produce a functionally as well as cosmetically acceptable result. The prior histori- cal constraints of 15 to 18 cm of maximum lengthening may no longer be valid. Patients with severe fibular and tibial hemimelia are probably still best treated by limb ablation in in- fancy. For the patient with fibular hemimelia and a foot with fewer than three rays, there are currently no effective means of producing a reasonably functional weight-bear- ing foot. Patients with acceptable foot function and a moderately mobile ankle can be treated with soft-tissue releases, resection of the fibrous fibular anlage (when pres- ent), Achilles tendon lengthening, and subsequent use of an articulated ankle-foot orthosis until the length discrepancy becomes unmanageable (6 to 8 cm), at which time the first tibial lengthening is performed. The child can be reevaluated to plot the developing discrepancy, and a sec- ond lengthening and contralateral epiphysiodesis can be done if neces- sary. The important feature of these patients is not the presence or ab- sence of the fibula, but rather the morphology and potential function of the foot and ankle. A patient with tibial hemimelia and an absent or dysfunctional knee extensor mechanism is best treated by an early knee disarticu- lation. Limb reconstruction can be a viable option if the proximal tibia is present (as determined by clini- cal examination and ultrasound or magnetic resonance imaging), the knee actively extends and is rea- sonably stable, and the foot can be made functional by early compre- hensive soft-tissue release. Ulti- mately, symptomatic ankle insta- bility in either tibial or fibular hemimelia can be managed with an ankle arthrodesis without sacrific- ing the foot. Severe forms of proximal fe- moral focal deficiency in which there is little femur present (type D in the Aitken classification system) or in which the hip cannot be ren- dered stable are still not amenable to lengthening. However, if hip stability can be achieved, femoral lengthening can be done. If the foot is at the level of the contralat- Deborah F. Stanitski, MD Vol 7, No 3, May/June 1999 151 A B Fig. 6 A, Preoperative radiograph of a skeletally mature woman with Ollier disease and a 14-cm limb-length inequality. Circular external fixation was used to gradually correct the proxi- mal and distal tibial deformities and to lengthen the tibia by 8 cm. B, Teleradiograph at the completion of the tibial lengthening. (Subsequent femoral lenghtening is illustrated in Figure 7.) eral knee and the ankle has a func- tional range of active motion, a Van Nes rotationplasty may provide an alternative to foot ablation. Each limb-lengthening proce- dure should probably be confined to no more than 15% to 20% of the limb-segment length. The rate of distraction should be adjusted according to the appearance of the regenerate bone formation as well as the range of motion of adjacent joints. The Ò0.25 mm four times a dayÓ guideline need not be fol- lowed rigidly. The potential com- plications of joint stiffness due to cartilage injury and/or musculo- tendinous contracture can be avoided by careful assessment dur- ing the distraction phase. 38,39 Limb function should not be sacrificed in the attempt to gain excessive length. Residual discrepancy can be treated by additional lengthening at a later date, shortening of the contralateral extremity, or both. Summary The management of the growing patient with limb-length inequality requires careful assessment, se- quential limb-length evaluations, and formation of a strategy based on the individual patientÕs needs. Treatment may involve a single procedure or a series of proce- dures, depending on the etiology and magnitude of the discrepancy and associated deformities. Limb-Length Inequality Journal of the American Academy of Orthopaedic Surgeons 152 A E B C D Fig. 7 A, Photograph of the patient in Figure 6 after tibial lengthening and deformity correction but before initiation of femoral lengthening and correction of the distal femoral valgus deformity. B, Anteroposterior and lateral radiographs of the femur after acute correction of the distal femoral deformity and initiation of gradual distraction. C and D, Antero- posterior and lateral radiographs at the completion of treatment. E, Final appear- ance of the patient at the completion of limb-length equalization. The original shoe lift is shown on the right.