INTRODUCTION

This chapter reviews the concept of torsional malalignment of the lower limb and its importance on patellofemoral (PF) pathology. This topic was first well reviewed by Stan James, M.D., in which he introduced the term “miserable malalignment.”¹⁵ His classic description is highly recommended reading.

It is essential in a chapter on osteotomy for treating PF symptoms to place this operation in proper context relative to other surgical procedures. The treatment of PF disease remains without consensus. There is little agreement on definitions. Objective evidence is scant. Most studies lack precise terminology and complete physical examinations, use incomparable imaging studies, and refer to incomplete biomechanical studies. The terms malalignment, instability, subluxation, dislocation, chondromalacia, chondropathy, and anterior knee pain are inappropriately used interchangeably. Measurements lack objectivity and validation.

Confusion exists in the literature regarding the term malalignment. Insall and coworkers¹³ in 1976 defined patellar malalignment as being either an increased Q-angle or a high-riding patella. No other options were considered. Then, Insall and coworkers observed that “an increased Q-angle is usually associated with increased femoral anteversion and external tibial torsion. In the presence of these abnormalities when the hip and ankle joints are normally aligned, the patella faces inward and motion of the knee occurs about an axis which is rotated medially compared with the axes of the hip and ankle joints.” These authors recommended a medial tibial tubercle transfer or “patellar realignment” to treat this torsional malalignment, and it is no small wonder that the investigation reported that 61% of the patients remained symptomatic. Clearly, the underlying pathoanatomy had not been addressed. Tibial tubercle transfer has been shown to cause external rotation of the tibia on the femur rather than pull the patella medially into the trochlea.¹⁷ In addition, this procedure has been shown to increase medial compartment loading²⁰ and medial facet loading. This author does not use the term “patellar malalignment” or “patellar realignment,” and reserves the term malalignment to refer only to a limb that deviates from normal in any one plane or any combination of the three anatomic planes: coronal (frontal), sagittal, or transverse (horizontal). Limb alignment is a direct result of the shape of the bones in the limb, but malalignment may be the result of acquired cartilage or ligament loss in which the deformity exists at the joint and not the bones.

Normal alignment is presumed to be normal because it is the optimal limb geometry for the transfer of the body mass to the ground. Abnormal geometry results in unusual force transmissions through the PF joint during gait. Rarely have published PF clinical studies adequately reviewed limb alignment. The precise variation from the normal skeleton is generally not specified, ranges of motion of all joints in all planes of the lower extremity have not been adequately recorded, and ligamentous stability is not measured.

Over 50 independent observations have been reported in the literature to be associated with anterior knee pain, chondromalacia, patellar instability, or patellar malalignment (Table 40-1). Whereas these factors may be very important, objective measurement tools are lacking. No study has ever tried to assess all these variables in a single cohort of patients, and even if this were undertaken, the lack of objective measurement techniques would make the study invalid. If these variables are indeed important, then any study would need to incorporate accurate measurements of each. The relative importance of each will need to be developed. Without this, clinicians are left with a realization that data do not exist to guide in the knowledge of pathomechanics, diagnosis, and treatment and that we depend only on unrelated clinical observations and some very basic laboratory studies.

TABLE 40-1 Cited Contributions to Patellofemoral Dysfunction

ACL, anterior cruciate ligament; FCL, fibular collateral ligament; MCL, medial collateral ligament; PCL, posterior cruciate ligament; VLO, vastus lateralis obliquus; VMO, vastus medialis obliquus.

Table 40-1 shows the variables related to PF syndrome, anterior knee pain, PF instability, chondromalacia, and PF arthritis. The contributions of these variables need to be defined, measured, and validated before any meaningful outcome studies can be completed. It is critical to simplify the confusing clinical picture created by so many variables. Toward this end, this author considers that there are only four independent but related variables to assess. Each of the conditions just discussed can be fitted into one of these four: skeletal geometry, ligaments, articular cartilage (patella and trochlea), and musculotendinous units.

BASIC ASSUMPTIONS

The geometry of the skeleton determines the direction of the load applied on the PF joint. The PF ligaments acting against the patella in the trochlea are responsible for stability. Articular cartilage reduces friction to conserve energy and to transmit force generated by a combination of body mass, momentum, gravity, and muscles. The musculotendinous units move the skeleton, stop it from moving, or hold it against gravity but are not normally responsible for preventing joint dislocation or subluxation.

Critical Points BASIC ASSUMPTIONS

• The geometry of the skeleton determines the direction of the load applied on the patellofemoral joint.

• The patellofemoral ligaments are responsible for stability.

• Articular cartilage reduces friction to conserve energy and to transmit force generated by a combination of body mass, momentum, gravity, and muscles.

• The musculotendinous units move the skeleton, stop it from moving, or hold it against gravity, but are not normally responsible for preventing joint dislocation or subluxation.

• Assessment

Limb alignment: physical examination, coronal standing radiographs, sagittal radiographs, and rotational CT studies.

Ligaments: physical examination, stress radiographs.

Articular cartilage: double-contrast CT arthrography with fine cuts.

Musculotendinous units: usual clinical tests of strength, orientation, flexibility.

• Patellar instability results from a loss of the patellar stabilizers and often an excess of displacement forces.

• Patellar stabilizers: trochlear groove, medial and lateral patellofemoral ligaments, medial and lateral meniscopatellar ligaments.

• For lateral patellar dislocation to occur, the medial patellofemoral ligament must be injured.

• Arthrosis or chondromalacia exists when articular cartilage fails as the force per unit area exceeds biologic tolerance.

CT, computed tomography.

It has been useful to independently assess these four structures and then try to understand the cause-and-effect relationship between each. This integrated picture is then used to develop an understanding of the pathomechanics, diagnosis, and logical treatment.

This author uses physical examination, coronal standing radiographs, sagittal radiographs, and rotational computed tomography (CT) studies to assess limb alignment. Physical examination and stress radiography⁴³ are used to assess ligament insufficiency. Double-contrast CT arthrography with fine cuts determines the thickness of articular cartilage and the location and size of various focal defects. The usual clinical tests of strength, orientation, and flexibility (e.g., Ober test, prone rectus test, straight leg raising, Achilles contracture) are used to assess musculotendinous units.

Patellar instability results from a loss of the patellar stabilizers and often an excess of displacement forces. The patellar stabilizers are the trochlear groove and the ligaments (both the medial and the lateral patellofemoral ligaments) and the medial and lateral meniscopatellar ligaments. Brattström,² as early as 1964, demonstrated that in recurrent dislocation of the patella, a flattening or dysplasia of the trochlear groove is nearly always present. Dejour and colleagues⁶ demonstrated that this dysplasia is often in the proximal portion of the trochlea not seen on an axial patellofemoral radiograph. This can be measured on a true lateral knee radiograph by recognizing the floor of the groove and the anterior projections of the medial and lateral condyles. For lateral patellar dislocation to occur, the medial patellofemoral ligament (MPFL) must be injured. A number of authors including Christoforakis and associates⁴ and Desio and colleagues⁸ have demonstrated in the laboratory that the lateral retinaculum provides up to 19% of the resistance against lateral displacement; thus, the patella is usually more unstable after lateral retinacular release. Medial dislocation is a distinct possibility and this author reported 70 cases confirmed with stress radiography in 1990 (Annual Meeting of the American Academy of Orthopaedic Surgeons). The diagnosis of instability requires a force and displacement examination and stress radiography may provide documentation (Figs. 40-1 and 40-2), because unstressed radiographs are normal but stress application demonstrates obvious excessive displacement.

FIGURE 40-1 A, An 18-year-old patient complained of anterior knee pain and giving-way, without history of dislocation. Radiographic parameters were normal. B, Stress application was necessary to demonstrate dislocation.

FIGURE 40-2 A, Radiograph of a 38-year-old woman with anterior knee pain after a dashboard injury. Symptoms worsened after a lateral retinacular release. Radiographic parameters were normal. B, Stress radiographs demonstrated medial dislocation.

A frequent contributing cause of patellar dislocation is the existence of an abnormally high displacement force. This most often occurs when the direction of the lateral quadriceps vector is increased. This commonly occurs when the femur twists away from under the patella, most often in the direction of medial rotation. This is usually due to some torsional abnormality of the lower limb (Fig. 40-3). In the face of such excess displacement forces, a PF ligament reconstruction may fail.

FIGURE 40-3 If the knee joint twists inward because the femur twists inward, the lateral pull on the quadriceps is increased, the lateral displacement pull on the patella is increased, the strain on the medial patellofemoral ligament (MPFL) ligament is increased, the compression on the lateral patellar facet is increased, and the compression on the medial patellar facet is decreased. The treatment must be to decrease the inward twist on the knee joint, not to move the tubercle medially. A similar increase of inward pointing of the knee joint occurs in the presence of excess external tibial torsion when the foot is pointed forward.

Arthrosis or chondromalacia exists when articular cartilage fails as the force per unit area exceeds biologic tolerance. For acute fractures of articular cartilage, this may be in the range of 20 MPa. In the case of chronic overload due to a too small articular surface area or too much body weight, muscle loading or long lever arms with chronic pressure exceeding 5 MPa may lead to arthrosis. Articular fractures exist approximately 90% of the time after acute patellar dislocation.³¹ Although magnetic resonance imaging (MRI) is often used to demonstrate articular cartilage lesions, Figure 40-4 gives an example of pathology clearly seen with double-contrast CT arthrography not visualized on MRI. Rotational malalignment is often a major contributing cause of pain, instability, and cartilage injury, but the evidence of how much benefit can be derived from rotational osteotomy in the presence of instability and arthrosis is unknown.

FIGURE 40-4 A, Magnetic resonance imaging (MRI) read as focal osteonecrosis. B, Contrast computed tomography (CT) arthrogram clearly shows a split in the articular cartilage extending down to the cystic lesion with damaged articular cartilage medial to the cartilage split. Most likely, there was an acute subluxation episode that compressed the articular cartilage until it failed with a split relieving ligament tension, with further lateral subluxation damaging the medial articular cartilage into which the contrast dye now moves.

HOW DOES LONG BONE TORSION AFFECT THE PF JOINT?

The knee joint axis moves straight forward during gait, with a small (<10°) turning inward and outward.²⁷ The foot also tends to move in a fairly constant direction (foot progression angle).^28,37 The geometry of the lower extremity skeleton largely determines the direction of load applied at the PF joint, with the amount of load depending on the body mass, the length of the lever arms, the surface area of the PF joint, and the velocity of the moving system. If the knee joint axis is twisted out of its normal plane of flexion and extension while the quadriceps is contracting, a side-directed force is created that acts on the patella, attempting to displace it. This causes an increased strain on the PF ligaments and retinaculum and an imbalanced direction of force on the PF articular surface.

Critical Points HOW DOES LONG BONE TORSION AFFECT THE PATELLOFEMORAL JOINT?

• Knee joint axis moves straight forward during gait, with a small turning inward and outward.

• Geometry of the lower extremity skeleton determines the direction of load applied at the PF joint. Amount of load depends on

Body mass.

Length of the lever arms.

Surface area of the PF joint.

Velocity of the moving system.

• Side-directed force is created on the patella if the knee joint twists out of its normal plane of flexion-extension during quadriceps contraction. Instability or arthrosis may result.

• Deviations in normal limb alignment may result in the knee joint flexion-extension axis advancing sideways while the body moves forward:

Femoral anteversion or retroversion.

Excess internal or external tibial torsion.

Genu valgum or varus.

Hyperpronation.

Achilles contracture.

• Foot progression angle: angle between the long axis of the foot and the direction of body progression, averages 10° to 15°.

• Increase in external foot rotation may account for the apparent increased genu valgus in females, the increased incidence in PF symptoms in females, and even for the increased incidence of ACL tears in females.

ACL, anterior cruciate ligament; PF, patellofemoral.

If these side-directed vectors exceed biologic tolerance, either instability or arthrosis may result. For example, an inward pointing of the knee increases the lateral direction of pull of the quadriceps. Therefore, the pull on the MPFL (and also medial retinaculum and medial meniscopatellar ligament) is increased and the direction of pressure on the patella is altered, creating an increase on the lateral facet and a decrease on the medial facet (Fig. 40-5). The radiographic appearance of subchondral density under the patellar facets is a useful clue to the mechanical environment (Fig. 40-6).

FIGURE 40-5 A, If the knee joint moves forward, the compression on the patellofemoral joint and ligaments tensioned are balanced. B, If the knee joint twists inward from beneath the patella, the MPFL is placed under increased tension, the compression beneath the lateral facet increases, and the compression beneath the medial facet decreases.

FIGURE 40-6 As the femur rotates internally beneath the patella, the lateral patellar articular cartilage is compressed. The increased crescentric density in the lateral patellar subchondral bone is indicative of chronic localized pressure.

Maximum energy conservation during gait is due to precise anatomy and joint kinematics.³⁴ Abnormalities in normal bony anatomy or joint motions distort the normal force vectors needed to transfer the body mass to the ground. Deviations in normal limb alignment may result in the knee joint flexion-extension axis advancing sideways while the body moves forward. These deviations include femoral anteversion or retroversion, excess internal or external tibial torsion, genu valgum or varus, hyperpronation, and Achilles contracture.

The foot progression angle (FPA) is generally defined as the angle between the long axis of the foot and the direction of body progression and averages 10° to 15°.^28,37 It has been shown that the FPA remains similar despite differences in the torsion of the tibia or femur.³⁷ It is likely that this is because proper ankle dorsiflexion cannot occur during gait if the ankle joint axis is not aligned with the direction of forward movement or because this presents the most stable position of the foot on the ground. Hip rotation must vary if the torsion of the long bones changes and the FPA stays constant. Alterations in both femoral and tibial torsion change the effective lever arm of the hip stabilizers¹ and may account for the frequency of soft tissue complaints around the hip and pelvis, as well as the increased pelvic tilt and lumbar lordosis seen in these patients. Examples of the change in the position of the hip and knee with a constant foot progression angle and changes in femoral and tibial torsion are seen in Figures 40-7 to 40-11.

FIGURE 40-7 Normal male with femoral anteversion of 13° and external tibial torsion of 21°. Note that with a foot progression angle of 13°, the knee joint faces slightly outward.

FIGURE 40-8 Normal female with femoral anteversion of 13° and external tibial torsion of 27°. Note that the knee joint is pointing slightly more inward and the greater trochanter is slightly more anterior than in the normal male.

FIGURE 40-9 Female with a 30° increase in tibial torsion. To keep the foot progression angle normal, the knee joint axis points inward nearly 30°, causing increased strain on the knee. The hip appears markedly internally rotated with the greater trochanter pointing somewhat anteriorly.

FIGURE 40-10 Female with a 30° increase in femoral anteversion. The knee joint points the same direction, slightly inward, as in the normal female, but the greater trochanter points posteriorly, giving a poor mechanical advantage. At some point, the hip cannot externally rotate enough to keep the knee joint pointed forward. With fatigue of the hip abductors, the knee points more inward to compensate for hip collapse, placing greater stress on the knee.

FIGURE 40-11 Female with a 30° increase in both femoral anteversion and external tibial torsion. Note that the trochanter is pointed more anterior than normal, and with the foot progression angle normal, the knee joint axis points markedly inward.

Figures 40-7 to 40-11 are drawn with an FPA of 13° (Seber average) for a normal male torsional alignment (Fig. 40-7), a normal female torsional alignment (Fig. 40-8), a female with 30° excess external tibial torsion (Fig. 40-9), a female with 30° excess femoral anteversion (Fig. 40-10), and a female with both 30° excess tibial external torsion and 30° excess femoral anteversion (Fig. 40-11). One may study the variations in knee joint progression with these common abnormal patterns. Yoshioka and associates^47–49 found identical femoral anteversion in both males and females, equal genu valgus in both males and females, but an increase in external tibial torsion and in foot external rotation in females over males. This increase in external foot rotation may account for the apparent increased genu valgus in females, the increased incidence in PF symptoms in females, and even for the increased incidence of ACL tears in females.

BIOMECHANICAL EVIDENCE RELATING LIMB TORSION WITH PF INSTABILITY AND PAIN

A change in PF contact areas and pressures was noted after alteration of femoral or tibial torsion or rotation by Lee and coworkers^22–24 Hefzy and associates,¹² and van Kampen and Huiskes.⁴⁵ Kiljowski and colleagues¹⁸ (unpublished data) placed lower limbs (femoral head to foot) in a frame with a clamp, attaching the femoral head to a vertical post that allowed the knee flexion to be altered by sliding the femoral head up and down the post. The foot was held in a position of normal FPA and an osteotomy was performed in the shaft of the femur, allowing rotation. PF contact pressures and MPFL strain were measured while the osteotomy changed the femoral torsion 15° and 30°, both inward and outward. At 30° knee flexion, lateral facet pressure increased an average of 30%, with a 30° increase in femoral anteversion, while medial facet pressures were decreased. Strain in the MPFL increased 57% after a 30° increase of internal femoral torsion. Fujikawa and coworkers¹¹ in a biomechanical study that measured PF contact pressures concluded that if an angular deformity and a torsional deformity coexist, the rotatory component causes the greater PF changes. These authors also noted that when a varus deformity was created, the greater the varus deformity, the less the PF congruence.

Critical Points BIOMECHANICAL AND CLINICAL EVIDENCE RELATING LIMB TORSION WITH PATELLOFEMORAL INSTABILITY AND PAIN

• Change in PF contact areas and pressures occur after alteration of femoral or tibial torsion or rotation.

• If an angular deformity and a torsional deformity coexist, it is the rotatory component that causes the greater PF changes.

• Shallow Q-angle is present in patients with recurrent patellar dislocation.

• Clinical study correlations

PF arthrosis and increased femoral anteversion.

Patellar dislocation and an increase in medial femoral torsion.

Internal femoral torsion with patellar instability and patellar chondropathy.

External tibial torsion and PF instability.

Decrease in vastus medialis and gluteus medius electromyogram amplitude with increased internal femoral torsion.

PF, patellofemoral.

CLINICAL EVIDENCE RELATING LIMB TORSION TO PF INSTABILITY AND PAIN

Brattström² in 1964 measured trochlear depth in a group of patients with recurrent dislocation of the patella and confirmed that a shallow trochlea was always present. This author also defined the Q-angle and pointed out the change in Q-angle as a result of changes in limb rotation. He reviewed the literature on the rotational femoral osteotomy as described by Graser (1904), Fürmeier (1953), Kiesselbach (1956), and Vinditti and Forcella (1958) and concluded that the “femoral osteotomy is a big operation that demands prolonged after treatment. I have not encountered any instance of recurrent dislocation of the patella that justified my suggesting an operation of this nature to the patient.” Takai and associates⁴¹ measured femoral and tibial torsion in patients with unicompartmental arthrosis in either the medial, the lateral, or the PF compartment and found the highest correlation to be that of PF arthrosis with increased femoral anteversion (23° anteversion in the osteoarthritis group vs. 9° anteversion in the control group). Janssen¹⁶ found a high correlation between patellar dislocation and an increase in medial femoral torsion and speculated that this medial femoral torsion was also responsible for the development of trochlear dysplasia. Lerat and colleagues^25,26 noted a significantly increased association of internal femoral torsion with both patellar instability (P < .0001) and patellar chondropathy (P < .001). Stroud and coworkers⁴⁰ followed 92 patients who at age 5 showed 30° greater medial hip rotation (measured in extension) than lateral rotation. At age 24, PF pain was noted by 30% in the increased medial rotation group compared with only 8% in the control group (P < .001). Winson and associates⁴⁶ found that 70% of adolescents undergoing arthroscopy for anterior knee pain had an increased internal torsion of the femur, compared with only 33% of those undergoing arthroscopy for a meniscal or cruciate injury. Turner⁴⁴ noted increased external tibial torsion in patients with PF instability (25° vs. 19°). Of perhaps great significance to those who believe that muscle strengthening is the key to treating PF symptoms, Nyland and colleagues³² found a significant decrease in vastus medialis and gluteus medius electromyogram amplitude in athletes with clinically increased internal femoral torsion. Arnold and coworkers¹ noted that an increase in femoral anteversion of 30° to 40° and decreased abduction moment arm strength of 40 to 50% was enough to impair normal walking, and therefore, those individuals required turning the knee inward to keep the hip from collapsing. This inward pointing of the knee creates high shearing force on the PF joint.

FEMORAL AND TIBIAL TORSION MEASUREMENTS

Femoral anteversion has been measured by a variety of techniques by many authors and recently by Yoshioka and associates^47,48 (Fig. 40-12). To measure limb alignment in the horizontal (transverse) plane, a CT rotational study is obtained. This is a CT scan that overlaps cuts from the femoral head, base of the femoral neck or lesser trochanter, the knee joint (either tangent to the posterior condyles or between the medial and the lateral epicondyles), the proximal tibia at the joint and the tibial tuberosity, and the ankle joint (Fig. 40-13). These cuts allow measurement of femoral torsion, tibial torsion, knee torsion, and the tibial tubercle–trochlear groove (TT/TG) distance. CT also allows for the observation of trochlear depth or dysplasia, the position of the patella in the trochlea (unstressed), patellar tilt, and shift and the density of the subchondral trabeculae.

FIGURE 40-12 A, Femoral anteversion or antetorsion (angle ϕ here) may be measured along the center of the femoral neck from the center of the femoral head to the center of the femoral shaft at the base of the neck, and distally either along the transepicondylar axis (Tr) 7.4° or along the tangent of the posterior femoral condyles 13.1°. These differ by about 6°, with a range of 11° retroversion to 22° anteversion. B, This measurement by Yoshioka and associates⁴⁸ is taken directly off the bone. Murphy and coworkers³⁰ validated a CT measurement along these same lines.

(A, From Yoshioka, Y.; Cooke, T. D. V.: Femoral anteversion: assessment based on function axes. J Orthop Res 5:86–91, 1987, p. 88. Fig 3A; B, from Yoshioka, Y.; Siu, D.; Cooke, T. D. V.: The anatomy and functional axes of the femur. J Bone Joint Surg Am 69:873–880, 1987, p. 875, Fig. 2B.)

FIGURE 40-13 A, For measuring femoral torsion (anteversion), CT slices taken through the femoral head, the inferior greater trochanter, and the knee joint are superimposed. One line is drawn from the center of the femoral head through the greater trochanter, and a second is tangent to the posterior femoral condyles at the knee joint. B, Hip and knee overlapped, allowing references lines seen in C to be drawn. C, Reference line 1 is taken from the center of femoral head to the center of femur at the level of the lower greater trochanter. Reference line 2 is a tangent to the posterior surface of the femoral condyles. The epicondyles are also readily seen and can be used as a second reference. There is generally a 6° difference between these two choices of reference. D, Tibial torsion can be measured between the axis of the tibial plateau and the ankle joint axis. E, The angle between the epicondylar axis of the femur and the ankle joint axis can be used as a measure of tibial torsion. F, The distance between the tibial tubercle and the center of the trochlear groove (TT/TG) is also provided in the rotational study. Dejour and colleagues⁶ stated that a distance > 20 mm is abnormal and correlates with objective patellar instability.

Critical Points FEMORAL AND TIBIAL TORSION MEASUREMENTS

• CT rotational study measures limb alignment in horizontal (transverse) plane

Femoral torsion

Tibial torsion.

Knee torsion.

TT/TG distance.

• Observation of

Trochlear depth or dysplasia.

Position of the patella in the trochlea (unstressed).

Patellar tilt.

Shift and the density of the subchondral trabeculae.

• Femoral torsion: average anteversion 13.1° when distal measurement is the tangent across the femoral condyles, and 7.4° when measured across the epicondyles. Author has arbitrarily chosen 13° as the goal for correction.

• Tibial torsion: measurement method not standardized; wide variability in normative data.

CT, computed tomography; TT/TG, tibial tubercle–trochlear groove.

To measure torsion of the femur, the recommendations of Murphy and coworkers³⁰ and Yoshioka and Cooke⁴⁷ are followed; these superimpose the CT cuts through the center of the femoral head and the base of the femoral neck where the circular nature of the femoral shaft becomes apparent (see Fig. 40-13A–C). A line drawn through the center of each femoral segment becomes the proximal reference. A CT cut through the distal femur where the anteroposterior (AP) dimension of the medial and lateral condyles is greatest allows a line to be drawn tangent to the posterior condylar surface. A second line can be drawn through the medial and lateral epicondyles. Yoshioka and Cooke⁴⁷ measured femoral torsion (anteversion) directly from the bone in a series of 32 femora and noted an average anteversion of 13.1° in both males and females when the distal measurement was the tangent across the femoral condyles, and 7.4° when measured across the epicondyles (standard deviation [SD], 8°; range, –11°–+22°). It is useful to measure both to avoid an error should dysplasia of the lateral femoral condyle with a short AP distance be present.

Yoshioka and Cooke⁴⁷ reviewed the literature measuring femoral anteversion by a variety of different techniques and noted that 12 of the 18 reviewed papers reported normal values between 8° and 16°. This author has arbitrarily chosen 13° as the goal for correction. Kuo and associates¹⁹ observed various measurement techniques and concluded that CT was accurate, but that the plane radiographic techniques were not. It is often stated that a patient with increased femoral anteversion has a compensatory external tibial torsion. Pasciak and colleagues³³ found this not to be the case and found no relation between femoral anteversion and tibial torsion. In a review of over 300 CT rotational studies, the author of this chapter found that with increased femoral anteversion, tibial torsion may be normal, increased external, or increased internal.

The measurement of tibial torsion by CT has not been standardized. Jakob and coworkers¹⁴ described a common sense method superimposing cuts across the proximal tibia and distal ankle joint. The author of this chapter has attempted to use the anatomic measurements of the tibia described by Yoshioka and colleagues⁴⁹ as a basis for measuring the proximal tibial reference line (Fig. 40-14). Blinded repeat measurements made by an orthopaedic radiologist produced a measurement variation of less than 1°, indicating that reproducible measurements are possible. Yoshioka and colleagues’ anatomic study⁴⁹ reported a lateral torsion of 21° in males and 27° in females, a significant difference. This was the only difference in geometry noted between males and females in both the tibia and the femur. In addition, these authors noticed a gender difference of 16° between the transverse axis across the tibial plateau and the center of the foot. If this observation is confirmed, it is of major importance because it may explain the widely quoted difference between men and women in the frequency of PF disorders. It may also explain the reported increased incidence of anterior cruciate ligament (ACL) tears in women. Le Damany²¹ reported 23.7° to be normal tibial torsion. Seber and associates³⁷ calculated external tibial torsion to be 30° (range, 16°–50°) in a group of 50 “normal” (asymptomatic) men, but their measurement technique was unique in selecting the location for drawing the proximal tibial reference. Eckhoff and coworkers⁹ noted variations between 15° and 30°. Turner⁴⁴ found 19° (SD, 4.8°) in a control group and 24.5° (SD, 6.3°) in a group with unstable patella. The CT measurements of Sayli and coworkers³⁵ averaged 30° to 35°. Tamari and associates⁴² recently demonstrated a lack of reliability between currently used measurement methods, which is a major stumbling block in answering the question of how much of an influence altered tibial torsion plays in PF instability and when osteotomy may be useful.

FIGURE 40-14 The tibial torsion may be measured as the angle μ, which is between the widest distance on the tibial plateau (EF) and the line connecting the malleoli (M_m-M_L). Males = 21°, Females = 27°. Angle σ is the “foot rotation,” the outward deviation of the foot in reference to the tibial plateau transverse axis (EF). Males = –5°, Females = +11°.

(From Yoshioka, Y.; Siu, D. W.; Scudamore, R. A.; Cooke, T. D. V.: Tibial anatomy and functional axes. J Orthop Res 7:132–137, 1989.)

In addition to the just-discussed method taken after Yoshioka and colleagues’ anatomic measurements,⁴⁹ the author of this chapter also uses the femoral epicondylar axis as the proximal reference for tibial torsion, owing to the interest in the relationship of the knee joint axis to the ankle joint axis and the problem incurred when selecting a line across an oval tibial plateau. The error of twisting across the knee joint while obtaining the CT must be avoided to use this reference.

INDICATIONS

Any variation from optimal skeletal alignment may increase the vector forces acting on the PF joint causing either PF ligament failure with subsequent subluxation or cartilage failure as in chondromalacia or arthrosis or both ligament and cartilage failure. Anterior knee pain may result from these abnormal forces or their consequences (Fig. 40-15).

FIGURE 40-15 Miserable malalignment. The patient often presents with medial parapatellar pain with no localizing findings or images indicating pathology. Later, instability or arthrosis may develop.

An inward-pointing knee due to any cause will increase the lateral displacement force on the patella (Figs. 40-16 to 40-18). If this force is great or if the trochlear support is reduced from dysplasia or fracture, the medial ligaments may fail, resulting in lateral patellar instability. The converse is also true. Pain in the medial retinaculum is a common symptom caused by this increased stress. If the trochlear support is normal, the ligament may not fail but the articular load may increase, causing arthrosis. This increased force may be the result of abnormal femoral torsion, abnormal tibial torsion, genu valgum, hyperpronation, Achilles’ contracture, or other conditions. The choice for surgery depends on the relative size of the deformity and the relative surgical risk and morbidity.

FIGURE 40-16 A 17–year-old student with rapid onset of pain and abnormal gait. It is likely that progressive overload due to torsional malalignment resulted in progressive increased foot pronation with loss of foot and ankle stability and secondary weakness of the lateral hip musculature. A, On standing anteroposterior (AP), the squinting patella on the right is well seen. Both femurs have 50° greater than normal anteversion. B, What the patient sees looking down her legs. C, If the knee faces forward, the foot turns out. D, The dynamic picture is much worse, because the anteversion puts the greater trochanter pointing posteriorly so there is no hip abduction power and the pelvis collapses. In an attempt to increase hip power and put the foot forward, the knee joint must point inward, even more when there is an increase in hyperpronation.

FIGURE 40-17 CT rotation study of the same patient yields anteversion of 67°. The hip cannot be rotated externally enough for the knee joint to face forward.

FIGURE 40-18 Six weeks postoperative, the patient shows 50° external rotational intertrochanteric osteotomy on the right hip. The right foot is less pronated than the left foot and the left knee joint is pointed slightly more inward than the right. Gait has already improved and knee pain diminished immediately after surgery.

The MPFL is a rather small structure. This is because a large ligament is not needed to restrain the patella from leaving the center of the trochlea if the limb is straight, the knee joint axis points forward, the quadriceps vector is in the frontal plane, and the trochlea is normal. If the trochlea is dysplastic (shallow), the ligament may not be strong enough to withstand the normal lateral pull of the quadriceps. If there is an abnormal limb torsion (so the knee joint does not move forward in the direction the body is moving), the side-displacing force is increased and the ligament is more likely to be overcome, leading to patellar instability. An MPFL reconstruction may be useful to restore the normal patellar restraint, but in the face of continued increased displacement forces caused by a torsional limb malalignment, such a reconstruction may fail. Conversely, in the face of normal limb torsion, the side-directed force may not be sufficient to displace the patella even with insufficient ligaments. The clinical presentation may be pain, instability, arthrosis, or a combination of these problems.

SURGICAL STRATEGY

The goal of operative treatment is to normalize the biomechanics through restitution of normal anatomy. The morbidity of surgery may dictate otherwise. When multiple anatomic abnormalities exist, it is not known which may be more important. If a patient has recurrent patellar lateral subluxation, articular cartilage damage, trochlear dysplasia, patella alta, an Achilles contracture, a femoral anteversion of 45°, an external tibial torsion of 55°, and a genu valgum of 10°, is it best to perform a varus, external rotational femoral osteotomy; a distalization of the tibial tuberosity; a reconstruction of the MPFL; a trochlear osteotomy; an internal rotation osteotomy of the tibia; or an Achilles lengthening? These abnormal findings are often quite subtle, but combinations are surprisingly common. The surgical approach to correct all abnormalities may be quite logical from the biomechanical view, but excessive from the surgical morbidity standpoint. A biomechanical solution with osteotomy may mean that the ligament reconstruction is not necessary, because the displacing forces are reduced. A cartilage restoration procedure is not necessary because the compressive forces are reduced and changed in direction. As yet, no biomechanical studies indicate which surgery alters biomechanics the most.

Experience suggests that for most cases with torsion of the femur or tibia exceeding 30° from normal, surgery is indicated. For those with torsion exceeding 20° from normal, surgery is thought to be beneficial. For cases with abnormality less than 20°, the accuracy of surgery or the morbidity may not justify the smaller biomechanical changes. Turner’s data⁴⁴ found 19° average external torsion in a control group and 24.5° in a patellar instability group. In another study, Takai and associates⁴¹ found 9° average femoral anteversion in a control group and 23° average in a PF arthrosis group. Therefore, it is possible that very small limb torsional differences are significant.

SURGICAL TECHNIQUE

To correct an excess of internal femoral torsion that is not associated with a genu varus or valgus, an intertrochanteric osteotomy has been used since the late 1980s. Prior to that, the author used the technique of intramedullary saw and fixation with an intramedullary nail. However, the antecurvatum of the femoral shaft is destroyed when the curved segments are rotated and the nail can split the distal fragment attempting to compensate (Fig. 40-19). In the original series of patients, cutting the femur from the intramedullary position and fixing with an intramedullary nail, bone stability appeared to be less with more postoperative pain and blood loss. There was a higher incidence of delayed union and one death resulted from apparent fat embolism. Consequently, the intertrochanteric region using an AO 95° condylar blade plate (Figs. 40-20 and Figs. 40-21) for fixation was selected. The intertrochanteric location was chosen to minimize damage to the quadriceps, avoid scarring that may occur between the distal quadriceps and the supracondylar femur with a supracondylar location, and theoretically reduce the abrupt angular change the quadriceps would make with a supracondylar location. It should be noted these decisions are without scientific evidence. If a genu varum or genu valgum is associated with a torsional abnormality and the tibiofemoral angle is abnormal, this deformity must be corrected in the supracondylar region.

FIGURE 40-19 Postoperative radiograph of a femoral rotational osteotomy produced with an intramedullary saw. Rotation changes the anterior bow of the femur to a corkscrew shape that does not allow the nail to pass congruently, resulting in splitting of the distal fragment.

FIGURE 40-20 The 95° condylar blade plate is designed for fixation of the proximal femur. The U-shaped blade is inserted into the center of the neck and the inferior half of the head of the femur.

(From Müller, M. E.; Allgöwer, M.; Schneider, R.; Willenegger, H.: Manual of Internal Fixation, 2nd ed. New York: Springer-Verlag, 1979; p. 91.)

FIGURE 40-21 Blade is inserted into the center of the femoral neck and head.

(From Müller, M. E.; Allgöwer, M.; Schneider, R.; Willenegger, H.: Manual of Internal Fixation, 2nd ed. New York: Springer-Verlag, 1979; p. 93.)

The standard AO/ASIF (Arbeitsgemainshaft für Osteosynthesefragen/Association for the Study of Internal Fixation) technique for hip osteotomy is used. This has been well described. The ability to solidly compress the two osteotomy fragments ensures stability and reduces pain. This compression remains a major advantage of the classic approach. The osteotomy is placed at the level of the proximal third of the lesser trochanter to increase the osteotomy surface contact area and to leave sufficient purchase for the 95° condylar blade plate in the proximal fragment. The blade plate is a tool that controls the correction. One needs to ensure only that the seating chisel for the blade plate enters the femoral neck at the proper height and orientation. The goal is for the blade of the blade plate to be centered in the femoral neck at such an angle that the side plate will sit on the lateral cortex of the femur without altering flexion-extension or varus-valgus. The standard AO tools aid in this goal (Fig. 40-22).

FIGURE 40-22 A, When a 95° condylar blade plate is to be used for fixation of a proximal femoral osteotomy, the site of insertion can be located with the aid of the condylar plate guide, which is a mirror image of the condylar blade plate. A K-wire inserted parallel to the top of the condylar plate guide will guide the insertion of the seating chisel. The direction of the femoral neck can be approximated with a K-wire placed along its anterior cortex. The neck must be seen for this to be precise and radiographic control with the image intensifier may be easier than opening the hip capsule. B, The seating guide placed in line with the axis of the femur will ensure that there is no change in flexion or extension of the hip.

(A and B, From Müller, M. E.; Allgöwer, M.; Schneider, R.; Willenegger, H.: Manual of Internal Fixation, 2nd ed. New York: Springer-Verlag, 1979; p. 93.)

A standard lateral approach to the hip is altered so the iliotibial track (fascia lata) is entered more anterior than usual to ensure that the trochanter or the plate over the trochanter will not put pressure on the repaired facial incision, causing a dehiscence.

After exposure of the lateral shaft of the femur by elevating the vastus lateralis anteriorly, the condylar plate guide, a mirror image of the 95° condylar plate, is placed along the lateral shaft of the femur. It is moved proximally and distally until an extension along its proximal flat surface lines up just below the center of the femoral head, as seen with x-ray image intensification (Fig. 40-23). A 2.0-mm K-wire is inserted into a 2.5-mm hole drilled into the femoral neck parallel to the top of the condylar plate guide and centered anteriorly and posteriorly in the neck. This hole is placed proximal enough to leave room below it to insert a seating chisel on top of the condylar plate guide. Because of the torsional deformity, the femoral neck is often angled quite anteriorly relative to the knee joint, but its exact direction cannot be seen. Classically, one may expose the neck of the femur anteriorly or slide a K-wire along the anterior cortex of the neck, but this is usually approximate and not exactly in the center line. To find whether this 2-mm K-wire is parallel to the center line of the femoral neck, a frog-leg lateral image is obtained. This normally results in bending of the K-wire against the iliotibial band, but the portion within the femoral neck will indicate whether the direction is appropriate. It is common to need to change these K-wires when they are bent. Obtaining the frog-leg lateral radiographic projection and the ability to slide the 2.0-mm wire in and out of the 2.5-mm hole are useful.

FIGURE 40-23 The condylar plate guide is placed along the lateral femoral shaft. It is moved proximal or distal until a line extending medially parallel to the top of the guide is located just below the center of the femoral head. A K-wire is inserted above and parallel to this line; this K-wire position is checked on the lateral radiograph to ensure that it is in the center of the femoral neck. Once proper positioning of the K-wire is obtained, the seating chisel for the blade of the blade plate is inserted into the femoral neck and head parallel to the K-wire and to the top of the condylar plate guide.

Once the proper position of the K-wire is obtained, the seating chisel for the angled blade is inserted. A window in the lateral cortex of the greater trochanter for the seating chisel may be started with drills (Fig. 40-24). The thickness of the 95° condylar blade is 4.5 mm, allowing a 4.5-mm drill to be used. A triple drill guide is available to attach to the top of the condylar plate guide to ensure the correct drill direction.

FIGURE 40-24 The window for the seating chisel may be created in the lateral cortex of the greater trochanter with a 4.5-mm drill. A triple drill guide (4.5 mm) may be attached to the top of the condylar plate guide.

(From Schauwecker, F.: Practice of Osteosynthesis, 2nd rev. ed. New York: Thieme-Stratton, 1982; p. 187.)

The condylar plate guide and later the seating chisel guide are held parallel to the femoral shaft to ensure the side plate will align with the lateral femoral shaft and there will be no change in flexion or extension (Fig. 40-25). Once the proper position of a K-wire is obtained, a thin chisel is used to cut into the lateral femoral cortex at the top of the properly positioned condylar plate guide. This cut in the cortex will allow the seating chisel for the blade to be advanced without sliding proximally or distally (Fig. 40-26). The condylar plate guide is used repeatedly to check that the varus-valgus angulation is not being changed and also that the flexion-extension angle will not be changed (Fig. 40-27). The seating chisel is advanced slowly until it approaches the center of the femoral head. It is backed out after advancing every 5 mm or so, so that it does not become incarcerated. The slotted hammer is used to prevent twisting of the seating chisel into flexion or extension as it is advanced. The depth of insertion is read off the marks etched on the seating chisel. A chisel is used to remove a small sliver of bone from the inferior edge of the entry site for the blade. This allows the shoulder of the plate to be advanced into the proximal fragment (Figs. 40-28 and Figs. 40-29).

FIGURE 40-25 The seating chisel guide and condylar plate guide must be placed parallel to the shaft of the femur to avoid causing a flexion or extension change. The tendency is for the plate to be inserted horizontally (shown here), forgetting the anterior curve to the femur, thus causing an increased hip extension.

(From Schauwecker, F.: Practice of Osteosynthesis, 2nd rev. ed. New York: Thieme-Stratton, 1982; p. 163.)

FIGURE 40-26 A triple drill guide is attached to the top of the condylar plate guide and 2- x 4.5-mm drill bits used to start the window for the seating chisel in the lateral cortex of the greater trochanter. A thin chisel is used to cut the cortex to begin the entry point for the seating chisel.

FIGURE 40-27 The seating chisel is driven into the femoral neck, parallel to the K-wire previously placed. The seating chisel guide is used to ensure that the plate will lie centered on the lateral shaft of the femur. This flexion and extension are controlled by the slotted hammer.

(From Schauwecker, F.: Practice of Osteosynthesis, 2nd rev. ed. New York: Thieme-Stratton, 1982; p. 188.)

FIGURE 40-28 The condylar plate guide is placed parallel to and touching the lateral femoral cortex, and the seating chisel is kept flush and parallel with the top of the guide. The slotted hammer allows control of the twist of the seating chisel; thus, it controls flexion and extension of the osteotomy.

FIGURE 40-29 A, The shoulder of the condylar blade plate will prevent the plate from seating all the way into the proximal fragment. B, A chisel is used to remove the lower edge of the window for the blade plate so it does not block full insertion.

(A and B, From Schauwecker, F.: Practice of Osteosynthesis, 2nd rev. ed. New York: Thieme-Stratton, 1982; p. 188.)

When the seating chisel is removed, a 2.0-mm K-wire may be inserted into the chisel track and this wire will slide into the U corner (Figs. 40-30 and Figs. 40-31). This guarantees that the direction for the blade will not be lost while the osteotomy is completed.

FIGURE 40-30 A 2-mm-diameter K-wire is inserted in the corner of the U-shaped chisel track so it can be located later when the osteotomy is complete and the blade is ready for insertion.

FIGURE 40-31 Once the seating chisel has been inserted, it is removed and a 2-mm K-wire is placed into the seating chisel track. A 2.5-mm wire is to be placed in the top half of the lesser trochanter and perpendicular to the femoral shaft.

A 2.5-mm drill hole is placed at the location for the osteotomy, usually at the upper third of the lesser trochanter, and drilled from lateral to medial and perpendicular to the femoral shaft coronal axis. A 2.5-mm K-wire, which will be used to guide the saw, is inserted into this hole.

Two anterior-to-posterior K-wires, which are set at the desired angle of correction, are placed in the femur, one proximal and one distal to the selected site for osteotomy (Figs. 40-32 and Figs. 40-33). A 2.5-mm drill is used to create the holes for these two K-wires. These are the markers to indicate the torsional correction. For excess femoral anteversion, the need is to externally rotate the distal fragment. The proximal pin is placed first, far enough anterior and medial that it will not come into conflict with the angled blade plate that will be inserted into the proximal fragment. The distal pin is started in line with the proximal pin in the anterior femoral cortex, but angled from medial to lateral the desired correction amount. Commercially provided triangles of various angles are available to be placed along the proximal K-wire and used as a guide to drill the 2.5-mm hole for the distal 2.5-mm K-wire reference.

FIGURE 40-32 After the level of osteotomy is selected and marked with a K-wire (bottom of photo), K-wires are needed as rotational markers. The first is placed proximal to the level of osteotomy and anterior to the location for the blade plate. For a 50° correction, one side of a 50° triangle is placed along the vertical K-wire and a second K-wire will be placed distal to the osteotomy level (bottom K-wire) and parallel to the medial edge of the 50° triangle.

FIGURE 40-33 The K-wires (right) to indicate the angle of correction are seen to subtend an angle of 50°. To the left are the K-wires in the osteotomy site and in the seating chisel tract.

Once these reference pins are in place, the osteotomy is completed with a thin saw (Fig. 40-34). The saw blade is kept perpendicular to the axis of the femur, and the image helps ensure that the orientation is proper (Fig. 40-35). Elevators are placed anterior and posterior to the bone to protect the soft tissues, and the image is used to judge the medial depth of the cut. Once the cut is completed, the soft tissues adjacent to the cut are stretched by placing a lamina spreader within the osteotomy (Fig. 40-36). This makes it easier to adjust rotation. Next, the blade plate is inserted into the proximal fragment (Fig. 40-37). The previously placed 2.0-mm K-wire in the seating chisel track makes the direction easier to follow. If the blade is properly in line with the track, it can often be pushed in by hand. After it is correctly placed in the femoral neck and head, it is necessary to impact the blade tightly into the proximal fragment (Fig. 40-38). Usually, a five-hole length side plate is sufficient. A longer-length side plate provides greater leverage but requires a longer incision and surgical approach.

FIGURE 40-34 The saw cut is started in the lateral cortex and the saw blade is perpendicular to the shaft in both the AP and the lateral radiographic projections. The K-wires indicate that rotation must be placed prior to completing the osteotomy and should be placed now.

FIGURE 40-35 An oscillating saw is used to cut the femur transversely. The two rotation wires (right) subtend the desired angle of correction, and the lower K-wire is a guide for the saw at the osteotomy level.

FIGURE 40-36 When the osteotomy is complete, it is useful to stretch the soft tissue with a lamina spreader. Four K-wires are shown: a 2-mm-diameter in the track of the seating chisel, 2- x 2.5-mm-diameter rotational wires above and below the osteotomy placed for rotational reference, and the fourth between to serve as a guide for the saw.

FIGURE 40-37 The blade can often be inserted by hand. The inserter handle is set so it is parallel to the blade and the insertion is parallel to the orientation K-wire.

(From Schauwecker, F.: Practice of Osteosynthesis, 2nd rev. ed. New York: Thieme-Stratton, 1982; p. 189.)

FIGURE 40-38 The impactor is used to seat the plate fully into the proximal fragment.

(From Schauwecker, F.: Practice of Osteosynthesis, 2nd rev. ed. New York: Thieme-Stratton, 1982; p. 190.)

The distal fragment is now externally rotated until the two torsional reference K-wires are parallel and the cortices of the osteotomy surface are well aligned circumferentially (Fig. 40-39). If the blade is not driven sufficiently far into the proximal fragment, it may allow the distal fragment to translate more laterally than the proximal fragment such that the surfaces of the osteotomy do not line up. This can be corrected by removing a small amount of bone from beneath the shoulder of the plate and impacting the plate further into the proximal fragment. Hohmann retractors or clamps may be useful to help align the axis of both fragments while maintaining the rotational correction. Once this position is reached, the articular tension device is attached to the femur distal to the plate (Fig. 40-40). This must be in line with the plate or the rotation will be changed as tension is applied to the plate.

FIGURE 40-39 The desired cortical alignment. The oblique lag screw prevents the plate from backing out and contributes extra compression across the osteotomy. The holes for attachment of the articulated tension device are seen distal to the plate.

FIGURE 40-40 The distal fragment has been externally rotated so the rotational wires are parallel, the gap at the osteotomy is seen between the two K-wires. The articulated tension device has been screwed onto the femur distally. Applying tension to the plate will compress the osteotomy surfaces providing stability.

A tension device placed anterior to the plate will cause external rotation of the distal fragment when tension is applied, and a tension device placed posterior to the plate will cause internal rotation of the distal fragment to be introduced. The articular tension device is color-coded, with the red indicating 120 kPa. Using the spanning wrench, an attempt is made to always exceed the 120 kPa indicator (Fig. 40-41). If the screw attaching the tension device to the femoral shaft is bent, the author is confident satisfactory compression of the osteotomy surfaces has been reached. If the blade is not tightly seated in the proximal fragment, it may pull out as tension is placed in the plate. A screw through the proximal plate into the proximal fragment is often required to prevent the blade from backing out with subsequent loss of osteotomy compression (Figs. 40-42 and Figs.40-43). The author has not found that proximal femoral nails or locked plates provide the degree of compression that can be obtained with the angled plates and the articulated tension device.

FIGURE 40-41 The osteotomy is compressed; the two rotational K-wires are parallel.

FIGURE 40-42 The blade has been inserted, and the tensioning of the plate complete and screws inserted. The top cancellous screw is used to keep the plate from backing out of the proximal fragment as it is pulled distally with the tensioning device. The distal two screws hold the plate in tension by preventing the plate from proximal migration. The second screw from the top is both an interfragmentary lag screw and adds additional rotational support. For perfect cortical contact without the medial translation of the proximal fragment, bone under the shoulder of the plate could have been removed so it would not prevent the further advancement of the blade into the proximal fragment. The surgeon felt this would have little or no effect on the limb alignment and advised against taking down the fixation at this point.

FIGURE 40-43 The blade is centered in the femoral neck as seen on the frog-leg lateral.

Rotational osteotomy of the tibia should be performed below the level of the tibial tuberosity. The Q-angle or TT/TG distance is an important biomechanical arrangement that should not be changed. It makes little difference where the osteotomy is performed, because the goal is to realign the knee joint axis with the ankle joint axis in the transverse plane, leaving the trochlea-tubercle relationship normal (Fig. 40-44). The author prefers the proximal tibia below the tubercle for the location of the osteotomy and the angled blade plate for compression, but the geometry of the proximal tibia makes the blade insertion always a problem. With a lateral approach, the side plate will come to lie along the flat surface of the tibia so the angle the blade takes into the proximal fragment will determine rotation, but the precise direction of this is not easily referenced off the flat surface of the tibia. This lack of smooth tibial surface for reference also makes proper insertion in varus-valgus a problem when tension is applied to the plate with the articular tension device. The eccentric loading of the plate may introduce a varus or valgus deformity that must be corrected.

FIGURE 40-44 Excess external tibial torsion is corrected by an internal rotation osteotomy below the level of the tibial tubercle.

The locking plate allows greater ease of rotational control, but still allows a distortion of varus-valgus correction when a compression is applied to the osteotomy. Tensioning of a locked plate that is fixed in the proximal fragment may also result in the distal plate sitting well off the bone, producing an unsatisfactory prominence that can tent the anterior compartment muscle laterally or the sparse subcutaneous tissue medially. Locking plates without high compression of the osteotomy fragments definitely result in delayed union time and more postoperative discomfort than the tight compression possible with an angled blade plate.

Two 2.5-mm K-wires are placed from front to back in the anterior cortex of the tibia, one proximal and one distal to the anticipated level of osteotomy. These may be parallel or subtend the desired angle of correction in the transverse plane. If they are parallel, the angle between the two after manipulation of the osteotomy fragments indicates the correction angle (Fig. 40-45). If they have been placed at an angle to each other for the desired correction, the pins are brought parallel for judging the desired correction. It is important to have these as reference points before the osteotomy is made. Currently, the author uses a navigation system (Praxim, Walpole, MA) that indicates the changed alignment in all three planes (transverse, coronal, and sagittal) (Fig. 40-46). It is impressive that small changes in all three planes are usually present and the manipulation of the fragments to a realignment precise in all three planes is difficult to obtain and difficult to hold while fixation is applied. In the author’s experience, the precision and confidence are greatly improved while the time and effort to obtain the correction is also increased. If a blade plate is used and the blade is inserted at the correct angle, the osteotomy fixation is easier to obtain but more difficult to change should an error have been made. Unfortunately, there are no imageless navigation systems the author has seen for use in the proximal femur.

FIGURE 40-45 A blade of a blade plate has been inserted into the proximal tibial fragment in a direction to prevent changes in varus or valgus, flexion or extension. It is inserted 30° in a posteromedial direction relative to the flat surface on the lateral tibial diaphysis. Two K-wires that were placed parallel before the osteotomy now are changed to an angle of 30° with the distal wire in the distal or shaft fragment having been rotated internally 30°. A clamp is holding the side plate to the lateral tibial cortex while the tension device will be applied.

FIGURE 40-46 Computer screen from a navigation system that allows a direct and accurate reading of the change in position of fragments. This system uses infrared reflectors placed into the femur, the tibia above the osteotomy, and the tibia below the osteotomy. Three planes are represented with the correction in each plane. Left, the frontal correction, the sample shows that 5° of valgus has been introduced with current hip-knee-ankle line measuring 173° and with an increase in leg length of 7 mm. Upper central figure, The horizontal plane correction shows an example of 4° of increased external rotation. Lower central figure, The sagittal plane change, an example of 2° of recurvatum or extension. The green V and T in this program indicate which markers are being visualized by the camera and utilized by the computer. The red P is not being seen by the camera. V: ?, T: ? are reference points on the metaphyseal and diaphyseal reflectors that are used as a test to ensure that there has been no loss of position of these reflectors.

(From Praxim, Walpole, MA.)

COMPLICATIONS

As with all major surgery, complications are possible and include malunion with overcorrection, undercorrection or the addition of a varus or valgus deformity, flexion-extension deformity, nonunion, muscle adhesions, compartment syndrome, vascular or neurologic injury, painful trochanter, and painful implants. Implant removal is generally necessary. Generalized complications are the usual expected and not listed here. This author has not performed a revision osteotomy for overcorrection or undercorrection when the blade plate was used. One revision was required in which an intramedullary nail was used for fixation. There have been no cases in which the external rotation of the femur caused the foot to turn out too far if the femur was not overcorrected (except for the one intramedullary nail case) and if an excess external tibial torsion was not already present. Such complications reported in the literature are rare owing to meticulous and slow surgical technique and use of rigid internal fixation.

The immediate goal of osteotomy is to obtain the planned correction. Failure to obtain the proper correction—that is, over- or undercorrection or secondary deformities of abnormal varus or valgus or flexion or extension—are due to errors in planning, execution, or fixation. The technique of using CT rotational studies for measurement of femoral or tibial torsion is not yet standard. The lines selected for axis measurements may not be accurate. The population normal values need further assessment. The clinical incidence of maltorsion is unknown and symptoms that develop may occur because of activity level. It is common that patients may have abnormal torsion in both the femur and the tibia. Any relationship between these two remains unknown.

The goal of femoral or tibial osteotomy is to put the knee joint axis in line with the forward progression of the body. If an excess femoral anteversion coexists with excess external tibial torsion, then following femoral external rotation osteotomy, the foot may appear to point too far outward and an internal tibial rotational osteotomy may be necessary. This has rarely occurred in this author’s experience. At times, the reverse has been noted in which, during walking after an external femoral rotational osteotomy, the foot may be less pronated and therefore appear to point less outward than before the osteotomy. In some patients, the foot is so pronated that it appears to be pointing outward when in reality the ankle joint is in proper alignment but the subtalar joint is not. This author cannot at this time predict before surgery what the gait will be after either a femoral or a tibial osteotomy or after an osteotomy of both femur and tibia. However, with accurate preoperative measurements, the patient can be informed of the existence of multiple deformities that might eventually require further correction.

Proper execution depends on the use of exacting intraoperative references. The references for rotation are K-wires placed above and below the level selected for osteotomy. When the osteotomy has been completed, it is necessary to manually manipulate the fragments into the newly desired position. The surgeons’ interpretation of the angulation between the reference K-wires is a source of potential error. The rotation of the tibia and femur is equally difficult, but because of a difficulty in maintaining the reduction, malposition of the proximal femoral fragments is less common than malposition of the proximal tibial fragments. This is because the site of the femoral osteotomy is through a relatively circular bone and the exact fit of a condylar blade plate to the proximal femur. At the proximal tibia, there is no implant that matches the bony anatomy both before and after correction. This makes it difficult to use the implant to control both varus-valgus and flexion-extension changes. In addition, the triangular shape of the tibia below the tubercle imposes a reduced contact area between the two cut surfaces, which is inherently more unstable. The reference wires are useful for estimating rotation and, to some extent, for estimating flexion and extension, but are not useful to accurately estimate varus or valgus changes. C-Arm images with a marker placed over the center of the femoral head and the center of the talus with a Bovie cord or an alignment rod placed between the two are more useful, but this method of judging limb alignment is highly sensitive to uncontrolled limb rotation. The knee joint axis must be perpendicular to the x-ray beam while this measurement is being made. The more compression that can be applied to the osteotomy surfaces, the greater stability, but this requires that the pull on the plate be in line with the axis of the limb. With a plate not perfectly contoured to the newly corrected bony anatomy, it is difficult to obtain attachment of the plate to the proximal tibial fragment such that the distal portion of the plate is aligned with the limb axis.

Rotational CT studies postoperatively would be required to measure the accuracy of the postoperative correction; this has been done in a limited number of cases. These limited results suggest that the accuracy of correction is within 5° of the planned correction when a blade plate was used, but up to 10° to 12° when an intramedullary nail was used for fixation.

Fixation without compression has a higher rate of delayed union, nonunion, and loss of correction. In spite of the published suggestion of frequent complications with tibial osteotomy, these have been relatively few when solid fixation through compression has been obtained. In this author’s experience, delayed union or nonunion has always been associated with inadequate fixation and compression. If this occurs without loss of correction, recompression is usually a reliable solution. This author recommends to not “take-down” a biologically active nonunion. In an unpublished review of 72 intertrochanteric osteotomies (author’s series), there was one delayed union that healed uneventfully after reoperation to add compression to the osteotomy by adding tension to the implant.

RESULTS OF TORSIONAL REALIGNMENT

Meister and James²⁹ reviewed 7 patients who had undergone tibial rotational osteotomy for anterior knee pain at an average of 10 years previously, and rated 1 patient excellent, 5 good, and 1 fair. Bruce and Stevens³ reviewed 14 patients with anterior knee pain caused by miserable malalignment treated with combined external femoral rotational osteotomy and internal rotational tibial osteotomy at an average of 5 years and reported that all were satisfied. Cooke and colleagues⁵ reported on 7 patients with anterior knee pain associated with varus and excessive external tibial torsion treated with valgus internal rotation osteotomy in which 5 were rated as excellent. Delgado and coworkers⁷ reported on 9 patients with anterior knee pain and 13 osteotomies of the femur, tibia, or both, with marked improvement seen over 2 years. Server and associates³⁸ reported on 25 patients undergoing internal rotational tibial osteotomy for excessive external tibial torsion and symptoms suggesting patellar subluxation, with 23 of 25 satisfied, including some athletes. Ficat and Hungerford¹⁰ reported 1 case of recurrent patellar subluxation that responded to osteotomy of both the femur and the tibia.

Ruesch in 1995 (unpublished paper) reviewed a series of 35 femoral external rotational osteotomies performed for anterior knee pain and/or instability in 31 patients. Five patients were not located, but 26 (84%) were seen an average of 5.3 years postoperatively. Eighty-eight percent of the group presented with failed PF surgery, with 23 of the patients averaging more than 2 prior surgeries. Only 3 patients had not had prior surgery. Six of the 35 osteotomies were intramedullary fixed with intramedullary nails and 29 were intertrochanteric. Two PF rating scores were used. The average Schwartz and colleagues’ score³⁶ was 24.6 points (of 29 points), with 15 of 35 excellent, 6 good, 5 fair, and 9 poor. The average Shea and Fulkerson score³⁹ was 82.5 points (of 100 points), with 20 excellent, 6 good, 1 fair, and 8 poor. The poor results mostly reflected preexisting arthrosis. However, subjectively, 77% indicated a decrease in the amount of pain, 86% less giving-way, 80% improvement in quality of life, and 42% a reduction in retropatellar crepitation. Of the 3 patients without prior surgery, the Shea and Fulkerson score was 100 points and the Schwartz score was 28.7 points.

Latteier (the author’s series; unpublished data) reviewed 72 intertrochanteric, external rotational osteotomies in 53 patients presenting with PF dysfunction. These included the patients reviewed by Ruesch described earlier. The mean follow-up was 9.7 years (range, 2–17 yr). Follow-up was obtained in 92% of the patients. Preoperative and postoperative Kujala, Lysholm, and Tegner scores were obtained. The mean Tegner activity score increased from 2.2 to 4.0 points, the mean Kyjala score increased from 53 to 86 points, and the mean Lysholm increased from 49 to 89 points. All but 1 patient showed improvement in all postoperative scores. However, in 20%, the Kujala improvement was insufficient to raise the score beyond the fair or poor range, and in 22%, the Lysholm improvement was insufficient to raise the score beyond the fair or poor range. Patients subjectively stated they improved by an average of 91%, with 55% of all respondents reporting improvement at 100%. The patient with the worst result reported improvement of 20%. Three patients would not undergo surgery again. One would not repeat the surgery because the amount of improvement was not enough, and 1 because of a femur fracture after blade plate removal 11 months after osteotomy. In the third patient, the surgery uncovered a previously unknown rare genetic coagulopathy requiring Greenfield filter, permanent anticoagulation, and advice not to have plate removal or further surgery. One patient had a delayed union treated by recompression of the osteotomy.

Each patient presents with different problems and different structural failures; thus, outcome measures fail to reveal the complete patient profile. Pain in the absence of instability or chondral disease is the most common early presentation. The only objective measurement may be that of abnormal limb torsion. Patients after surgery have reported returning to painless cross-country running after not being able to walk to classes before correction, the loss of bunions, the loss of radiating pain down the lateral side of the limb from pelvis to lateral ankle after 40 years without relief, the loss of retropatellar crepitation on stairs, the first ability to ride a bicycle and ski without pain, and the feeling of stability in landing after a jump in basketball. Such observations are not part of any PF scores and the improvements are therefore missed. Exercise that constrains the pelvis and feet such as bicycling, classic cross-country skiing, or elliptical machines is a particular problem for these patients because there is no escape for the inward-pointing knee. Strong hip abduction is a benefit for patients with femoral anteversion, because it may allow the patient to be active with a knee that points forward thus relieving pain. However, because the hip abductor is working at a mechanical disadvantage, it will often fatigue and the knee joint must point medially to gain mechanical leverage at the hip. This inward limb rotation forces an increased pronation at the foot such that medial arch strain, posterior tibial tendon strain, and bunions are more common. This chronic foot pronation may allow an increase in equinus ankle contracture, which then accentuates the increase in internal limb rotation. Excess external tibial torsion with normal femoral torsion presents a different problem. The knee points inward because this is necessary for proper ankle dorsiflexion during gait. The medial knee pain is common because the quadriceps increases the lateral vector on the patella when the foot is pointing forward. The hip abductors, however, are working at near-normal mechanical advantage so pelvic strengthening is less likely to be useful.

CONCLUSIONS