Chapter 63 Surgery of the Patellofemoral Joint

Proximal Realignment

W. Norman Scott, Giles R. Scuderi, Gabriel Levi

Disorders of the patellofemoral joint are numerous and are of great importance because they seriously limit patients’ function. Treatment of these conditions is highly dependent on an accurate diagnosis and determination of the correct cause.⁵⁵ Within the spectrum of disorders of the patellofemoral joint, one of the more common is patellar instability, which will be the focus of this chapter. Instability can present with a variety of symptoms that mimic other pathology of the knee, such as pain, mechanical “clicking,” feelings of instability or “giving out,” weakness, and limited range of motion. We begin with a review of the presentation and pathophysiology of patellofemoral disorders, including patellofemoral instability. We then review physical examination of the patellofemoral joint, imaging modalities, and the surgical technique of proximal realignment.⁸⁰ The techniques of patellofemoral arthroplasty and distal patellar realignment are beyond the scope of this chapter and are discussed elsewhere in the text. A review of the biomechanics of the patellofemoral joint is recommended in conjunction with this chapter.

Pathophysiology of Patellar Instability

The stability of the patellofemoral joint is dependent on many factors. Specifically, it depends on the congruence of the trochlea and patella, static ligamentous stabilizers, and dynamic stabilizers. Bony and cartilaginous constraint is determined by the shape of the trochlea and the patella, which can be extremely variable. For example, a hypoplastic lateral condyle or a shallow trochlea can predispose to acute patellar dislocation after trauma, or can lead to chronic subluxation and dislocation. In addition to bony and cartilaginous constraints are the static ligamentous constraints. We now understand that the medial patellofemoral ligament (MPFL) is the greatest static stabilizer for lateral translation of the patella.^24,^33,^34,⁹⁵ Finally, and perhaps most important, is the quadriceps mechanism. The coordinated contraction of the quadriceps mechanism centralizes the patella in the trochlea throughout a range of motion (Fig. 63-1). It is well known now that the vastus medialis obliquus (VMO) is the primary dynamic stabilizer of the patellofemoral joint (Fig. 63-2). Dysfunction of the VMO, whether by trauma or by atrophy, may lead to patellar instability. It is interesting to note that as we learn more about the anatomy of medial restraints, we are learning to appreciate the important relationship between the MPFL and the VMO.⁹¹ The sum of these factors defines the ultimate stability of the joint. If any single factor or multiple factors are deficient or dysplastic, the remaining structures must compensate. The ability of these factors to collectively stabilize the joint determines the stability or instability of the joint. In the remainder of this chapter, we will describe how the history, physical findings, and radiographic studies are used to determine the relative contribution of each of these factors to the instability of the patellofemoral joint.

Figure 63-1 The patella is anchored and stabilized to the knee by four structures in a cruciform fashion: the patellar tendon inferiorly, the quadriceps tendon superiorly, and the retinacula medially and laterally.

Figure 63-2 The vastus medialis obliquus becomes tendinous just a few millimeters proximal to the patella. Because of the oblique direction of its fibers, it is best suited to resist lateral displacement of the patella. The patellotibial ligament is visible as a distinct structure medial to the patellar tendon.

Presentation of Patellofemoral Disorders

Patellofemoral disorders, and especially instability, may present with a variety of symptoms. The history is important in determining the cause of the instability. Oftentimes, but not always, a patient with instability will present with a history of trauma. Commonly, a valgus and external rotational injury is the culprit. First-time dislocators may describe a traumatic event in which the patella dislocates and then self-reduces, or is reduced by the patient. Other first-time dislocators feel “their knee pop out of place.” Recurrent dislocators often feel catching, locking, or a feeling of instability. However, many patients with patellar instability do not describe instability on presentation, and for this reason, it should be considered in all patients with patellofemoral discomfort and mechanical knee pain.

Physical Examination

Thorough examination of the knee with a focus on the patellofemoral joint is essential to diagnosing patellofemoral instability or any associated patellofemoral or intra-articular pathology that often accompanies patellofemoral instability. During the physical examination, the patient should be examined sequentially while standing, walking, sitting, supine, and prone.

The examination begins with the patient standing with the feet together. Genu varum or valgum can be observed readily, as can rotatory malalignment such as in-facing or “squinting” of the patellae in patients with an increased quadriceps (Q) angle and hip anteversion (Fig. 63-3). Quadriceps tone and development can be appreciated in the standing position or during a half-squat. Hypoplasia of the VMO should be noted. Normally, the vastus medialis inserts on the upper third or half of the medial border of the patella. In knees with patellofemoral dysplasia, the muscle belly may end a few centimeters short of the superior patellar margin. The presence of quadriceps atrophy implies decreased dynamic muscle control on the patella.

Figure 63-3 Squinting of the patellae caused by rotational malalignment of the limb. This phenomenon is accompanied by an increase in the Q angle.

The position of the foot also deserves attention. Eversion at the subtalar joint is accompanied by compensatory internal tibial torsion, which increases the Q angle and consequently stress on the patellofemoral joint. The subtalar joint is a single-axis joint that acts like a hinge connecting the talus to the calcaneus. The axis of the subtalar joint deviates an average of 23 degrees medially and anteriorly to the long axis of the foot and 41 degrees inferiorly and posteriorly in relation to the horizontal plane. Therefore, internal rotation of the leg causes eversion of the heel and depression of the medial side of the foot. External rotation of the leg produces the opposite effect.²² Subtalar joint eversion may be primary or secondary, as in knees with varus alignment or tibia vara, wherein compensatory subtalar joint eversion is required to produce a plantigrade foot. This phenomenon is probably more important in long-distance runners.⁶² Eversion of the heel (heel valgus) is readily appreciated by looking at the patient in the standing position from the back side. Abduction of the forefoot is evaluated in the standing position by palpation of the talar head on the anterior aspect of the ankle. The neutral position is defined when the head of the talus can be equally palpated on the medial and lateral sides. During weight bearing, a normal foot is in mild pronation, and additional pronation should still be possible.¹⁸

Gait is observed, and if possible, the patient is asked to squat and hold the halfway position briefly because pain in this position is usually patellar in origin (half-squat test). Whenever possible, stair climbing and descending should be observed because this activity also provokes patellar symptoms.

With the patient seated on the examining table, the position of the patella is first checked. Normally, it sinks between the femoral condyles with the knee at 90 degrees of flexion. If patella alta is present, its anterior surface points to the ceiling with the knee in the same 90-degree position. Active extension is observed, and the presence of patellar crepitus and painful catching, as well as abnormal patellar tracking, is recorded.

Ficat and Hungerford ³⁶ stress the importance of observing the entrance and exit of the patella into and out of the sulcus between 10 and 30 degrees of flexion. They describe four common abnormalities in patellar tracking. Normal patellar tracking is present when the patella glides smoothly into the sulcus, and only minimal lateral displacement may be appreciated in the final extension when the patella exits the trochlear groove. We define more marked lateral displacement as lateralization, whereas greater degrees of pathologic tracking are defined as subluxation or dislocation. This finding is also called the J sign because the path resembles an upside-down J. (See video on the website.) Furthermore, evaluation of the tilt of the patella should be attempted. In normal knees, the medial border of the patella should be at the same level as the lateral border, with a minor lateral tilt in full extension. It should be noted that most abnormalities in patellar tracking involve lateral displacement and lateral tilt of the patella in extension, which reduces in flexion. Therefore, we find it useful to roughly estimate patellar subluxation and tilt during the physical examination (Fig. 63-4) and to verify this later with radiographic axial views or computed tomography (CT). Other abnormalities in patellar tracking, including medial dislocation or subluxation of the patella in flexion (after over-release of the lateral structures and excessive medial displacement of the tibial tuberosity) or lateral dislocation in flexion (as in habitual or permanent dislocation), may be encountered more rarely.

Figure 63-4 Clinical evaluation of patellar tracking. The patient is sitting on a firm examining table with the knee flexed at 90 degrees. The examiner places his hand on the knee so that the medial and lateral borders of the patella are palpated with the index finger and thumb. The patient is asked to actively extend the knee. An effort is made to detect the presence of lateral subluxation (displacement) or lateral tilt (lateral border of the patella lower than the medial border) during tracking.

Patellar crepitation is appreciated during active extension and is recorded as absent, mild, moderate, or severe. It should be evaluated in the sitting position and is enhanced by the application of manual resistance on the lower part of leg. Because crepitation beyond 90 degrees of flexion cannot be evaluated in the sitting position, it is better assessed during a full squat.

Hughston and Walsh ⁵¹ described a lateral position of the patella in the flexed knee for which they coined the term frog-eye patella. This seems to be associated with patella alta, which can be suspected clinically when the fat pad is unusually prominent. In fact, the fat pad may, on inspection, be mistaken for the patella because it occupies the femoral sulcus with the knee in extension while the patella is situated in the supracondylar pouch.

With the patient supine on the examining table, tenderness around the patella is evaluated systematically. Tenderness and swelling in the prepatellar area may be indicative of prepatellar bursitis. Evaluation of tenderness around the patella includes assessment of the medial and lateral retinacula, quadriceps, and patellar tendons, including their insertions. Joint line tenderness should also be evaluated.

The Q Angle (Quadriceps Angle)

The Q angle is measured by drawing an imaginary line connecting the center of the patella and the anterior superior iliac spine to produce a surface marking that approximates the line of pull of the quadriceps tendon (Fig. 63-5). A second line drawn from the center of the patella to the center of the tibial tubercle indicates the direction of the patellar tendon. The intersection of these two imaginary lines forms the Q angle. Because this measurement is affected by rotation of the hip, an effort is made to note the position of the medial border of the patient’s foot during walking and to reproduce this position during measurement.

Figure 63-5 The quadriceps angle is measured by drawing two lines from the center of the patella. The first line is drawn up to the anterior superior iliac spine and represents the line of pull of the quadriceps muscle. The second line is drawn down to the tibial tubercle and indicates the line of the patellar tendon.

Active pronation or supination of the foot should be avoided because, as mentioned earlier, these movements are associated with internal and external rotation of the leg, respectively,⁹¹ and consequently increase and decrease the Q angle.

Aglietti and associates measured the Q angle in 150 normal subjects and found it to be 15 degrees (range, 6 to 27 degrees; standard deviation [SD], 3 degrees) (Table 63-1).³ It was lower in men (14 degrees; SD, 3 degrees) than in women (17 degrees; SD, 3 degrees), and the difference was significant (P = .001). Only 11 subjects, all women, had Q angles greater than 20 degrees (7%). Therefore, it seems that a Q angle greater than 20 degrees may reasonably be considered abnormal. The Q angle was also measured in a group of pathologic knees, including 53 patients with patellar pain and 37 with recurrent subluxation or dislocation. In the knees with patellar pain, the Q angle was significantly increased to 20 degrees; this was true for both men and women. In contrast, the Q angle was not significantly different from normal in knees with recurrent subluxation or dislocation (average, 15 degrees), and the same applied to both men and women. In patients with patellar subluxation or dislocation, the Q angle is usually underestimated for at least two reasons: first, the patella is displaced laterally in extension; second, the quadriceps tendon frequently lies more lateral than predicted when the superior iliac spine is used as the surface marking (Fig. 63-6).

Table 63-1 Measurement of Quadriceps Angle and Radiographic Measurements of Patellar Height and Patellofemoral Congruence

Figure 63-6 The quadriceps tendon often lies more laterally than the surface marking of the Q angle predicts.

To overcome the problem of lateral patellar displacement and consequent underestimation of the Q angle, Fithian and colleagues ³⁷ measured the Q angle with the knee in 30 degrees of flexion. They simultaneously applied a posteriorly directed force so that the patella symmetrically contacted the trochlea. With this method, the Q angle was 12 degrees in control subjects (11.2 degrees in men and 13.4 degrees in women). A significantly higher value was found in a group of knees with patellar dislocation (average, 19.2 degrees). The contralateral knee of patients suffering from patellar dislocation also showed an increased value of the Q angle (average, 18.4 degrees).

It is debated whether the Q angle is better measured in the supine or the standing position. Woodland and Francis ¹¹⁸ measured the Q angle in the supine and standing positions in a large number of normal men⁷⁸ and women.⁷⁶ Average values in the supine position were 12.7 degrees for men and 15.8 degrees for women. Changing to the standing position increased the Q angle 0.9 degree in men and 1.2 degrees in women. The difference was statistically significant but probably is less significant clinically. It is relevant to note that the values determined by Woodland and Francis¹¹⁸ are close to those detected by Aglietti and coworkers.³

It has been suggested that the Q angle should be measured at 30 degrees of knee flexion and with maximum external tibial rotation ¹⁰⁷ because this position would give a more reliable measurement of the maximal valgus vector imposed on the patella. However, the difficulty involved in achieving standardized knee flexion and hip rotation makes the reproducibility of this measurement less reliable.

We recommend measurement of the tubercle-sulcus angle (TSA), as described by Kolowich and coworkers.⁶⁵ The TSA is measured with the knee at 90 degrees of flexion (Fig. 63-7). This allows the patella to engage in the sulcus and highlights any rotational abnormalities. It is defined as the angle formed by a line perpendicular to the transepicondylar axis and a line from the tibial tubercle to the center of the patella. The normal value for the TSA is 0 degrees, and values greater than 10 degrees are considered pathologic. However, consistent identification of the transepicondylar line is not easy, especially in obese patients.

Figure 63-7 Tubercle-Sulcus Angle (TSA) is measured with the knee at 90 degrees of flexion. The patella is engaged in the sulcus. The angle is formed by a line perpendicular to the transepicondylar axis and a line form the tibal tubercle to the center of the patella. Normal TSA is 0 degrees and abnormal is greater than 10 degrees.

(From Scuderi GR: The patella, New York, 1995, Springer-Verlag.)

In the supine position and with the knee in extension, the patient is asked to contract the quadriceps, and upward movement of the patella is noted (lateral pull test). In a normal knee, the patella is pulled predominantly upward with an associated minor lateral displacement. The lateral pull test result is considered abnormal if lateral displacement is excessive.⁶⁵

Patellar mobility should be evaluated with the knee in full extension and at 30 degrees of flexion. With the knee in extension, the patella is out of the trochlear groove and may be easily displaced medially and laterally. Gross hypermobility, as seen in patients with patella alta and a dysplastic extensor apparatus, is easily detected in this position. With the knee flexed 20 to 30 degrees, the patella is normally drawn into the trochlear groove and stabilized. Excessive displacement in the lateral direction indicates laxity of the medial retinaculum, or vice versa. On the other hand, reduced medial mobility indicates the presence of a tight lateral retinaculum. Kolowich and colleagues ⁶⁵ suggested that patellar mobility is best evaluated by dividing the patella into longitudinal quadrants (Fig. 63-8). With the knee at 20 to 30 degrees of flexion, mobility in the medial or lateral direction should not exceed two quadrants. A medial glide of one quadrant or less suggests a tight lateral retinaculum, which may also be investigated by trying to lift the lateral border of the patella with the knee in extension (passive patellar tilt) (Fig. 63-9). If the transverse axis of the patella cannot be elevated beyond the horizontal plane, a tight lateral retinaculum is demonstrated.⁶⁵

Figure 63-8 The mobility of the patella is best evaluated with the knee flexed to 20 to 30 degrees to engage the patella into the femoral sulcus. Displacement of the patella in the medial or lateral direction is best recorded in quadrants. Displacement in the medial direction of one quadrant or less indicates a tight lateral retinaculum. Displacement in the lateral direction over two quadrants indicates weakened medial stabilizers.

(From Kolowich PA, Paulos LE, Rosenberg TD, Farnsworth S: Lateral release of the patella: indications and contraindications. Am J Sports Med 18:359, 1990.)

Figure 63-9 A and B, Passive patellar tilt test. In a normal knee in full extension, it is possible to lift the transverse axis of the patella beyond the horizontal. An inability to perform this maneuver indicates a tight lateral retinaculum.

(From Kolowich PA, Paulos LE, Rosenberg TD, Farnsworth S: Lateral release of the patella: indications and contraindications. Am J Sports Med 18:359, 1990.)

Medial and lateral displacement of the patella has been measured with a displacement transducer ³⁷ or axial radiographs.¹¹² Fithian and associates³⁷ used a displacement sensor to record motion in the coronal plane with the knees bent at 30 degrees. Forces of 2.5 and 5 lb were applied with a hand-held force applicator with a load cell. Under a 5-lb force, medial patellar displacement averaged 9.2 ± 3.5 mm, and lateral displacement was 7.7 ± 2.6 mm in a group of 188 normal knees. In a group of 22 patients with symptomatic lateral patellar dislocation, medial displacement at 5 lb was 8.3 ± 4.5 mm, and lateral displacement was 11.5 ± 4.7 mm. Using the lateral minus medial displacement index, control knees had an average value of −2.1 ± 2.8 mm. Symptomatic knees had a lateral minus medial displacement of 3.2 ± 3.4 mm. The difference was statistically significant. In summary, total mediolateral displacement of the patella in normal knees under a 5-lb force was close to 17 mm in normal knees and almost 20 mm in knees with patellar instability. The lateral minus medial displacement was useful for differentiating between normal and unstable patellae. Medial displacement was larger than lateral displacement in 81% of control subjects. In unstable patellae, lateral displacement was larger than medial displacement. Asymptomatic knees in patients with unilateral patellar dislocation could not be used as controls because they showed abnormal patellar mobility, similar to symptomatic knees. The authors offered the concept of balance between medial and lateral structures. A normal knee should have greater medial than lateral displacement. If lateral displacement exceeds medial displacement, the restraining structures are unbalanced.

Teitge and colleagues ¹¹² evaluated medial and lateral patellar mobility with an axial radiograph at 30 to 40 degrees of flexion and a spring-loaded scale to apply a 16-lb force (7.3 kg). In a group of 20 asymptomatic knees, medial displacement averaged 11.1 ± 3.6 mm, and lateral displacement was 11.6 ± 8.1 mm. These values are higher than those reported in the study by Fithian and coworkers,³⁷ but the displacing force was also greater (16 lb vs. 5 lb). In a group of knees with lateral instability, medial displacement was comparable (11.6 mm on average), whereas lateral displacement was highly increased to 21.9 mm on average.

Torsional abnormalities of the femur and tibia have been described as a possible factor leading to patellar pain or instability. The pattern would involve increased femoral neck anteversion so that the trochlear groove faces inward, the Q angle is increased, and the patellae are squinting. Compensatory external tibial torsion is required to produce a foot aligned in the sagittal plane.

Turner and Smillie ¹¹⁴ measured tibial torsion in 836 patients with a tropometer. They found an average lateral tibial torsion of 19 degrees in control knees, which was increased to 24.5 degrees in knees with patellofemoral instability, and to 24 degrees in those with chondromalacia. Because increased tibial rotation was also noted in Osgood-Schlatter disease, this finding does not seem to be specific for patellofemoral joint disorders.

The amount of femoral neck anteversion has often been indirectly estimated by measuring the proportion of internal to external rotation of the hips in extension (Fig. 63-10), which would be its major determinant.^9,^109,¹¹⁰ Carson and colleagues¹⁸ suggested that if internal rotation of the hip in extension exceeds external rotation by more than 30 degrees, femoral neck anteversion is increased. Insall and colleagues⁵⁸ suggested that increased femoral neck anteversion may be present in knees with patellofemoral malalignment. Hvid and Andersen⁵² measured the Q angle and internal rotation of the hip in 29 patients with patellofemoral complaints. They found that both the Q angle and internal hip rotation were higher in women than in men. A significant correlation between the Q angle and hip rotation was noted, thus suggesting that the Q angle is in fact increased because of excessive femoral neck anteversion. Other authors,³⁵ however, have failed to identify any significant differences in Q angle, genu valgum, and anteversion of the femoral neck between normal adolescents and adolescents or adults with anterior knee pain. They concluded that because those affected by knee pain are also those most interested in sports activities, the probable cause is chronic overloading rather than faulty mechanics.

Figure 63-10 Excessive internal rotation of the hips in a patient with recurrent patellar dislocation. Increased femoral neck anteversion is usually present in these cases.

Dejour and associates ³⁰ reported CT measurements of femoral anteversion and tibial torsion in normal controls and in knees with instability of the patella. They found that femoral neck anteversion was increased in knees with patellofemoral instability (15.6 degrees) as compared with controls (10.8 degrees). Tibial torsion was a less important factor. It was 33 degrees in the control group and 35 degrees in the knees with patellar instability.

In light of the data reported in the literature and our own clinical experience, we think that torsional abnormalities of the lower limb, including femoral neck anteversion and, less significantly, external tibial torsion, may contribute to and play a role in patellofemoral disorders by increasing lateral pull on the patella. However, because these deformities are often less marked and remote from the knee, their importance is minor from therapeutic and surgical points of view.

The ultimate goal, in addition to an accurate diagnosis, is to identify any malalignment, medial-lateral restraint imbalance, and patellofemoral articular disease. The history and physical examination, in combination with the correct imaging studies, are essential in developing an appropriate treatment plan for the patient with patellofemoral instability.

Summary of Physical Examination for Patellofemoral Instability

Alignment of the limb is examined in standing and supine positions. Special attention is paid to the Q angle, varus/valgus alignment, and rotation of the limb. Increased Q angle, excessive femoral anteversion, genu valgum, pronation of the foot, and external tibial rotation all lead to excessive lateral force on the patella.

The TSA is measured with the knee at 90 degrees of flexion. The patella is engaged in the sulcus. The angle is formed by a line perpendicular to the transepicondylar axis and a line from the tibial tubercle to the center of the patella. The normal TSA is 0 degrees, and an abnormal TSA measures greater than 10 degrees (see Fig. 63-7).¹⁰⁰

Atrophy of the quadriceps muscle, especially the VMO, is noted.

Tightness and laxity of the medial and lateral soft tissue restraints are assessed. With the knee flexed to 20 to 30 degrees, the patella is displaced medially and laterally. Medial displacement of less than one quarter of the patella, or 5 mm, indicates a tight lateral retinaculum. Lateral displacement of more than three quadrants indicates medial soft tissue incompetence (see Figs. 63-8 and 63-9). Inability to tilt the patella past horizontal is indicative of a tight lateral retinaculum.⁶⁵

A patella apprehension test is performed with the relaxed patient lying supine and the knee flexed to about 20 degrees. The examiner attempts to gently displace the patella laterally. Patients with instability will feel uncomfortable and apprehensive, and will inadvertently contract the quadriceps muscle to attempt to stabilize the patella.

Patellar tracking is assessed throughout a range of motion. A J sign,³⁶ in which the patella is tracking in an inverted J pattern, represents the patella that is tracking centrally and then shifts laterally as the knee is extended. (See video on the website.) At 90 degrees, the patella should be centered between the femoral condyles in a normal patellofemoral joint.

Crepitus in the patellofemoral joint should be noted throughout range of motion. Pain is noted with patellofemoral compression. Tenderness to palpation should be noted with direct palpation of the medial/lateral facets, medial/lateral retinacula, and medial and lateral femoral condyles.

Ligamentous examination of the entire knee, as well as provocative tests for meniscal injury, should be performed.

Radiographic Studies in the Diagnosis and Treatment of Patellofemoral Instability

Standard radiographic evaluation, including anteroposterior, lateral, and axial views, should be obtained in each patient with patellofemoral disorders to assess the height and congruence of the patella and to exclude other bone disorders. CT has been used widely to study the patellofemoral joint, as it provides substantially more detail about the bony anatomy. MRI has also been used to investigate the patellofemoral relationship in extension and early flexion, as well as to detect cartilage lesions and ligamentous injuries.

Radiography

Anteroposterior View

The anteroposterior view does not allow visualization of the patellofemoral joint; however, it does provide valuable information regarding overall alignment of the limb and the presence of degenerative changes in the tibiofemoral joint. Weight-bearing anteroposterior views of the knees, as well as standing full-length mechanical axis views, are recommended. In addition, because the outline of the patella is visible, abnormalities such as patella magna or parva, bipartite patellae, and fractures can be seen. Marked lateral subluxation of the patella can also be detected in the anteroposterior view.

Lateral View

The lateral view is taken with the knee in at least 30 degrees of flexion to place the patellar tendon under tension and to demonstrate the functional relationship between the patella and the femur. Excessive rotation should be avoided because it may obscure some of the bony landmarks, such as the tibial tubercle, and may make interpretation difficult. The patella is not visualized on the lateral view in rare cases of congenital absence and when it is completely displaced laterally, as in habitual dislocation. In children younger than 5 years, the ossific nucleus has not yet appeared, and thus the patella is invisible.

Patellar position is related to the length of the patellar tendon. Patella alta, in particular, is associated with patellar instability, dislocation, and abnormalities of the trochlear groove. Patella alta, in particular, has been identified as a strong contributor to patellofemoral instability. Several methods of measurement have been described, including those reported by Blumensaat,¹⁵ Insall and Salvati,⁵⁹ Blackburne and Peel,¹⁴ Caton et al,²⁰ Rünow,⁹³ and Grelsamer and Meadows.⁴²

Blumensaat’s Line

Blumensaat ¹⁵ states that on a lateral radiograph with the knee flexed 30 degrees, the lower pole of the patella should be on a line projected anteriorly from the intercondylar notch (Blumensaat’s line). It is difficult to obtain routine radiographs with the knee flexed exactly the required number of degrees; this limits the usefulness of the method. Blumensaat’s method is also inaccurate. Of 44 radiographs of the knee that were flexed exactly 30 degrees, in no case did the lower pole of the patella lie on Blumensaat’s line; rather, the patella was positioned above this line (Fig. 63-11).

Figure 63-11 Insall-Salvati measurements to determine the height of the patella. HI, Height of insertion, or perpendicular distance between the joint line and the insertion of the patellar tendon; P, greatest diagonal length of the patella; T, length of the tendon measured on its deep posterior surface; WCBL, width of the femoral condyles at Blumensaat’s line.

(From Insall JN, Salvati E: Patella position in the normal knee joint. Radiology 101:101, 1971.)

Insall-Salvati Method

Insall and Salvati ⁵⁹ sought a method that would fit the following requirements: (1) simple and practical, as well as accurate; (2) applicable to the range of knee positions used during routine radiography, which in the lateral view is usually 20 to 70 degrees of flexion; and (3) independent of the size of the joint and the degree of magnification of the radiograph. Because the ligamentum patellae is not elastic, its length determines the position of the patella, provided that the point of insertion into the tibial tubercle is constant.

Insall and Salvati ⁵⁹ describe an expression for normal patellar height in terms of the length of the patellar tendon. Measurements were made on 114 knees in which the diagnosis of a torn meniscus had been clearly established by clinical history and examination, by positive arthrographic results, and by the finding of a meniscal tear at arthrotomy. Any case in which the slightest doubt existed was excluded, and it was assumed that the joints examined were structurally normal before a traumatic episode produced a torn meniscus. All patients were adults, and none showed radiologic evidence of osteoarthritis. The following measurements were taken (see Fig. 63-11):

1 T (length of tendon): The length of the patellar tendon was measured on its deep or posterior surface from its origin on the lower pole of the patella to its insertion into the tibial tubercle. The point of insertion is usually represented on the radiograph by a clearly defined notch.

2 P (length of patella): The greatest diagonal length of the patella was measured.

3 WCBL (width of the femoral condyles at Blumensaat’s line): Both condyles were measured at the level of Blumensaat’s line, and an average value was obtained. This measurement determined whether great variation in patellar size was present. Patellar size was considered to be acceptably constant.

4 HI (height of insertion): The perpendicular distance from the level of the tibial plateau surface to the point of insertion of the patellar tendon was measured to determine whether a constant relationship could be seen between the level of the tibial tubercle and the tibial plateau. Great variation in tendon insertion would have invalidated the measurements, but the insertion appeared to be acceptably constant. This measurement, therefore, may be disregarded in clinical evaluation.

The length of the patellar tendon (T) was found to be approximately equal to patellar length (P), and this was expressed as a ratio (because of variations in the size of individual knee joints and their projection on radiographs). The average value of the ratio T/P was 1.02, with a mean SD of 0.13. It was concluded that in a normal knee, the length of the patellar tendon should not differ from that of the patella by more than 20%.

Measurements were repeated on bilateral knee radiographs in 50 asymptomatic volunteers by Jacobsen and Bertheussen,⁶¹ who confirmed a similar degree of accuracy. Agletti and associates also measured patellar height according to the Insall-Salvati method in a group of 150 normal knees³ (see Table 63-1). The T/P ratio was found to be 1.04 on average (range, 0.8 to 1.38; SD, 0.11). The patella was significantly higher in women (1.06) than in men (1.01). In the same group of knees, the distance from the plateau level to the tibial tuberosity and the diagonal length of the patella varied only with sex, with larger values in men than in women. In 53 knees with patellar pain, the patella was slightly but significantly higher than in normal knees, with an average T/P ratio of 1.08 (range, 0.88 to 1.29; SD, 0.09). In a group of 37 knees with recurrent subluxation, the average T/P ratio was clearly increased to an average of 1.23 (range, 0.78 to 1.60; SD, 0.18). In conclusion, according to the Insall-Salvati method, an index over 1.2 indicates patella alta, whereas an index below 0.8 indicates a low patella.

Blackburne-Peel Ratio

Blackburne and Peel ¹⁴ criticized the T/P ratio on the basis of two observations:

1 The radiographic marking on the tibial tubercle may be indistinct or even unrecognizable when the tibial tuberosity has been affected by Osgood-Schlatter disease.

2 The nonarticular portion of the lower pole of the patella varies considerably in size; it is instead the position of the articular surface that is of greatest clinical significance.

To overcome these difficulties, the authors suggested a ratio between the perpendicular distance from the lowest articular margin of the patella to the tibial plateau (A) and the length of the articular surface of the patella (B) as measured on a lateral view of the knee in at least 30 degrees of flexion (Fig. 63-12). The A/B ratio in 171 normal knees was 0.80 (SD, 0.14). No difference between the sexes was noted.

Figure 63-12 Blackburne and Peel method of measuring patellar height. Height is expressed as the ratio between A (perpendicular distance between the lowest part of the articular surface and the joint line level) and B (length of the articular surface of the patella).

(From Blackburne JS, Peel TE: A new method of measuring patella height. J Bone Joint Surg Br 59:241, 1977.)

Aglietti and associates measured the A/B ratio in a group of 150 normal knees ³ and found a slightly higher value than the original authors did: an A/B ratio of 0.95 on average (range, 0.65 to 1.38; SD, 0.13), with an insignificant difference between men and women (see Table 63-1). In a group of patients with anterior knee pain, the A/B ratio was 0.91, an insignificant difference from the ratio in control knees. On the other hand, the A/B ratio was significantly increased in knees with recurrent subluxation (average, 1.08; range, 0.76 to 1.89; SD, 0.19).

Modified Insall-Salvati Ratio

Grelsamer and Meadows ⁴² observed that the Insall-Salvati index does not account for the shape of the patella. They found that patients with patella alta and a long distal nose may have a falsely normal Insall-Salvati index. The presence of patella alta in these patients can be easily verified by indices that use the patellar articular surface and the upper part of the tibia as landmarks (the Blackburne and Peel or the Caton ratio). The variable relationship between the length of the patella and the length of the articular surface as expressed by the morphology index⁴³ has been presented in the section on anatomy (Fig. 63-13).

Figure 63-13 Different shapes of the patella in the sagittal plane according to Grelsamer and colleagues.⁴³ Patellar shape is described by the morphology ratio, that is, the ratio of patellar length to length of the articular surface. A, A normal type I patella has a morphology ratio between 1.2 and 1.5. B, Type II patella, with a morphology ratio greater than 1.5, that is, with a long inferior pole (Cyrano appearance). C, Type III patella with a morphology ratio below 1.2, that is, with a short inferior pole.

To overcome the problem of variable morphology of the patella, the Insall-Salvati ratio was modified. It was suggested that the ratio of the distance between the inferior articular facet of the patella and the tibial tuberosity and the length of the articular surface be used (Fig. 63-14). In other words, this method uses the same distal reference point as the Insall-Salvati method (the tibial tuberosity) and the same proximal reference point as the Caton method. In a group of 100 control knees, the modified Insall-Salvati ratio was 1.5 on average (range, 1.2 to 2.1). Ninety-seven percent of control knees had a ratio less than 2.0. Therefore, for practical purposes, a ratio of 2 or more can be used as an index of patella alta.

Figure 63-14 The modified Insall-Salvati method as described by Grelsamer and Meadows.⁴² The distance between the lowest point of the articular surface of the patella and the tibial tuberosity (distance a) is divided by the length of the articular facet (distance b). A value of 2 or more indicates patella alta.

Lyon School

The Lyon School ²⁰ criticized the previously existing methods of measuring the height of the patella. Members of this group found it difficult to define the insertion of the patellar tendon into the tibial tuberosity in knees with previous transposition of the tuberosity. They further observed that the use of a tangent to the tibial plateaus in the Blackburne and Peel method¹⁴ may be a source of significant error. Perfect superimposition of the tibial plateaus is necessary to draw the line. The posterior slope of the tibial plateaus is not constant and may vary 15 degrees or more in subjects who have undergone anterior tibial epiphysiodesis. To overcome these difficulties, they tried to develop an easy method that could be used on lateral radiographs in flexion between 10 and 80 degrees and that was not influenced by radiographic magnification, by previous transposition of the tibial tuberosity, or by fractures of the tip of the patella. In this method, a ratio is calculated between the distance AT from the inferior point of the articular surface of the patella to the anterosuperior edge of the tibia and the length AP of the articular surface of the patella (Fig. 63-15). The AT/AP ratio was calculated in 141 normal subjects and was found to be 0.960⁷ in 80 men and 0.990⁶ in 61 women. Based on these findings, the authors considered the patella to be infera with a ratio of 0.6 or less and alta if 1.3 or more.

Figure 63-15 Caton’s method to measure the height of the patella. Height is expressed as the ratio between the distance (AT) from the lowest point of the articular facet to the most prominent part of the tibial plateau and the length (AP) of the articular facet of the patella. See text for discussion of the ratio AT/AP.

(From Caton G, Deschamps G, Chambat P, et al: Les routules basses: á propos de 128 observations. Rev Chir Orthop 68:317, 1982.)

Norman Index

Norman and coworkers ⁸⁹ observed that the Insall-Salvati method does not describe the relationship between the patella and the femoral sulcus, and that proximal or distal transposition of the tibial tuberosity may be performed without affecting the T/P ratio. To overcome these difficulties, Norman and associates described a method wherein a lateral radiograph is obtained with the knee in full extension (hyperextension) and in quadriceps contraction to straighten the patellar tendon. The film–focus distance should be kept constant (1 m) and the cassette placed in contact with the lateral aspect of the knee. Various parameters were measured in this radiograph, including the length of the tendon, patella, and articular facet and the vertical position of the patella—that is, the distance from the lowest point of the articular facet to the joint line. These measurements were related to the height of the patient, and it was found that the vertical position of the patella was constant without sex-related differences (Fig. 63-16). The Norman index, defined as the ratio between the vertical position of the patella (in millimeters) and body length (in centimeters), is 0.21 in a normal knee (SD, 0.02). In patients with recurrent dislocation without associated generalized laxity, the index is 0.23 on average; in patients with associated generalized laxity, it is 0.25.

Figure 63-16 Norman’s method to measure the height of the patella. A lateral radiograph is obtained with perfect superimposition of the femoral condyles, the knee hyperextended, and the quadriceps contracted to straighten the patellar tendon. A line is drawn tangential to the distal third of the anterior femoral cortex. Two perpendicular lines are then drawn that pass through the femoral–tibial contact point and through the inferior point of the articular facet. The distance between the two perpendicular lines, defined as the vertical position (VP) of the patella, is related to the height of the patient in centimeters. The Norman index is expressed as follows: Vertical position of the patella (mm)/Body height (cm). In a normal knee, its average value is 0.21.

(From Norman O, Egund N, Ekelund L, Rünow A: The vertical position of the patella. Acta Orthop Scand 54:908, 1983.)

MRI Method

Biedert ¹³ proposed a different method by which to evaluate patellar height based on MRI scans in extension. Measurements are taken in the central or just lateral sagittal slice, where the articular cartilage is thicker. Patellar articular cartilage is projected on trochlear cartilage, and the height of its projection is calculated. The ratio between trochlear articular cartilage height and patellar articular cartilage height is deemed the patellar index. Normal values are 12.5% to 50%. An index less than 12.5% signifies patella alta, whereas an index higher than 50% indicates patella baja. The main advantage of this method is that it takes into consideration only the articular cartilage very precisely and is not affected by the patellar tendon.

Conclusions and Author’s Recommendations

Various methods of measuring the height of the patella have been described. In our opinion, the Insall-Salvati T/P ratio remains a reliable and reproducible method. It does not require perfect alignment of the knee in the lateral view and does not need correction for magnification. Values greater than 1.2 are diagnostic of patella alta. The method may not be applicable in knees with previous Osgood-Schlatter disease, and the method is not suitable for evaluating patellar position after distal transfer of the tibial tuberosity. In this setting, one may use the Blackburne and Peel or the Caton method; these methods measure the height of the patella in relation to the tibial plateau and joint line. The main drawback of the Blackburne and Peel method is that it requires a lateral view with superimposition of the femoral condyles to accurately identify the joint line. An image amplifier is required to consistently obtain this degree of accuracy. The same problem is encountered with the Norman technique, which has the additional difficulty of requiring an effective quadriceps contraction. Furthermore, the film–focus distance (1 m) must be kept constant, and the patient’s height must be known. In the presence of abnormal patellar morphology with a long or short distal pole, the modified Insall-Salvati method should be used.

Evaluating for Patellofemoral Dysplasia

Attention has been drawn to evaluation of the anatomy of the trochlea and subluxation of the patella as seen on the lateral view. Maldague and Malghem ⁷⁴ first described the radiographic appearance of the patella and trochlea femoralis on lateral views of normal knees and knees with patellar instability. It is necessary to obtain lateral views with satisfactory superimposition of the posterior and distal femoral condyles, which requires the use of an image amplifier. In the lateral view of a normal knee, the posterior aspect of the patella is represented by two lines: the most posterior one is the patellar ridge, the other is the lateral facet (Figs. 63-17 and 63-18). In knees with mild lateral tilt of the patella, the two lines superimpose. When the patella is more markedly tilted, the lateral facet overhangs the patellar ridge line posteriorly, and the anteroposterior diameter of the patella is greatly increased. The normal trochlea is composed of three lines: the two anterior lines are projections of the top of the medial and lateral facets of the trochlea; the posterior line, in continuation with the intercondylar roof line, represents the deepest point of the sulcus (Fig. 63-19). The distance between the two anterior lines and the posterior line represents the depth of the sulcus. Maldague and Malghem observed that its depth is normally greater than 1 cm as it is measured 1 cm distal to the upper part of the trochlea. In knees with patellar instability, the depth of the trochlea is reduced throughout the length of the sulcus (totally deficient sulcus) or only in its upper part (focally deficient sulcus). The authors emphasized the importance of the lateral view in patients with clinically suspected patellar instability and axial views showing negative results. Because axial views are often obtained in more than 30 degrees of flexion, a lateral view taken at 15 degrees of flexion allows exploration of the patellofemoral congruence at a degree of flexion that cannot be visualized with conventional axial views.

Figure 63-17 Patellar shape on lateral view with the knee flexed 10 to 15 degrees and good superimposition of the femoral condyles. A, In a normal knee, the posterior profile of the patella is represented by two lines: the most posterior one is the patellar ridge (PR); the anterior one is the lateral facet (LF). B, If the patella is slightly tilted laterally, the two lines superimpose. C, If the patella is severely tilted, the lateral facet overhangs the ridge line posteriorly, and the anteroposterior diameter of the patella is increased.

(A, Modified from Maldague B, Malghem J: Apport du cliché de profil du genou dans le dépistage des instabilites rotuliennes: Rapport préliminaire. Rev Chir Orthop 71[Suppl 2]:5, 1985.)

Figure 63-18 Lateral views. A, The patella in a normal knee. B, A knee with a mildly tilted patella. C, A knee with a markedly tilted patella.

Figure 63-19 Morphology of the trochlea in a lateral radiograph with superimposition of the femoral condyles. The two anterior lines are the projection of the medial (M) and lateral (L) facets of the trochlea. The posterior line is in continuation with the intercondylar roof line and corresponds to the deepest part of the sulcus (S).

(Modified from Maldague B, Malghem J: Apport du cliché de profil du genou dans le dépistage des instabilites rotuliennes: rapport préliminaire. Rev Chir Orthop 71[Suppl 2]:5, 1985.)

These concepts have been carried a step farther by the Lyon School.³⁰ Members of this group examined the lateral views of 143 knees with recurrent or acute dislocation of the patella and compared these with the radiographs of 190 control knees. They studied two quantitative measurements—trochlear bump and trochlear depth—and one qualitative sign—the crossing sign.

On a lateral view with superimposition of the femoral condyles, a line is drawn tangent to the last 10 cm of the anterior cortex of the femur. The line of the femoral sulcus may end in front of (positive value), over, or behind (negative value) the line of the anterior cortex. The distance between the anterior cortex line and the sulcus (saille or bump) is measured in millimeters (Fig. 63-20). The bump of the sulcus line in relation to the anterior femoral cortex was found to be highly useful in differentiating between knees with instability (average, +3.2 mm) and normal knees (average, −0.8 mm). A pathologic threshold value for measurement of trochlear bump was identified: 3 mm. Sixty-six percent of knees with patellar instability had anterior trochlear translation of 3 mm or more as compared with only 6.5% of control knees.

Figure 63-20 Quantification of the trochlear bump. A line (ac-a) is drawn tangential to the last 10 cm of the anterior femoral cortex. The line of the sulcus at its most anterior point X may pass in front of (positive value) or behind (negative value) the tangent line. The bump is measured as the distance between the femoral cortex line and the sulcus line in millimeters.

The depth of the trochlea was measured as the distance between the floor of the trochlea and the most anterior condylar contour line. First, a line tangent to the posterior cortex of the femur was drawn. A second line was drawn perpendicular to the posterior femoral cortex line and tangent to the posterior aspect of the femoral condyles. A third line was finally drawn that subtended an angle of 15 degrees to the second line and passed through the intersection of the first and second lines. Trochlear depth was measured along this third line (Fig. 63-21). Trochlear depth was 7.8 mm in the control group and 2.3 mm in knees with patellar instability. A trochlear depth of 4 mm or less was considered pathologic. This value was found in 85% of knees with patellar instability and in only 3% of controls.

Figure 63-21 Quantification of trochlear depth. It is the distance A-B measured in millimeters along a line subtended 15 degrees from the perpendicular to the posterior femoral cortex line and crossing a line tangential to the posterior femoral condyles and perpendicular to the posterior cortex line.

Dysplasia of the trochlea can be divided into three types according to the point at which the sulcus line crosses the lines of the condyles (the croisement, or crossing, sign) The crossing sign is a simple qualitative criterion defined as the crossing between the floor line and the lateral condylar line. At that level, the trochlea is considered flat (Figs. 63-22 and 63-23)³¹:

Type I dysplasia: This form is the mildest. The lines of the condyles are symmetrical, and they are crossed at the same point in the proximal part of the trochlea by the floor line. Only the very proximal part of the trochlea is flat.

Type II dysplasia: The lines of the condyles are not superimposed; the line of the sulcus crosses the medial condyle line first, and the lateral one crosses at a higher level. Separate crossing of the medial and lateral condyle lines is characteristic of this type.

Type III dysplasia: This form is the most severe. The condyle lines are superimposed, but they are crossed low on the trochlea by the sulcus line. Most of the trochlea is therefore flat.

Figure 63-22 The croisement (crossing sign). In normal knees, the sulcus line may be posterior to the condyle lines throughout its length (type A) or may join the medial condyle line only in the upper part of the trochlea (type B). Therefore, there is no crossing between the sulcus line and the condyle line in normal knees. The presence of the crossing sign indicates dysplasia of the femoral trochlea. See text for further discussion.

(From Dejour H, Walch G, Neyret P, Adeleine P: La dysplasie de la trochlee femorale. Rev Chir Orthop 76:45, 1990.)

Figure 63-23 Trochlear dysplasia, types I, II, and III. See text for explanation.

(From Dejour H, Walch G, Neyret P, Adeleine P: La dysplasie de la trochlee femorale. Rev Chir Orthop 76:45, 1990.)

Two types of normal trochlea were identified:

Type A (50%): The sulcus line is posterior to the condyle lines throughout its length (see Fig. 63-22).

Type B (50%): The sulcus line joins the line of the medial condyle, but only in the highest part of the trochlea.

Figure 63-34 The position of the patella as visualized by computed tomography may present the following abnormalities: A, Subluxation with a positive congruence angle persisting beyond 10 degrees of flexion. B, Tilt with a lateral patellofemoral angle less than 8 degrees in the first 30 degrees of flexion. C, Subluxation and tilt when both abnormalities are present.

Dysplasia of the femoral sulcus, as evidenced by the crossing sign, was present in the majority of knees with patellar instability (96%) (Fig. 63-24); the same was true in only 3% of control knees. In light of this study, the authors concluded that in trochlear dysplasia, the trochlea is flat in a zone of variable length and has a shallow groove more distally. According to the authors, trochlear dysplasia is better evaluated on a lateral radiograph than on CT. Such dysplasia is best recognized with the crossing sign, a qualitative factor that was present in 96% of knees with patellar instability. Two other quantitative measurements, trochlear bump (positive when ≥3 mm) and trochlear depth (positive when ≤4 mm), were both positive in 85% of cases.