Malcolm M. Glasgow, Lee Shepstone. From Norfolk and Norwich Health Care NHS Trust, Norwich, England. Anterior knee pain due to dysplasia of the extensor.
The sulcus angle and malalignment of the extensor mechanism of the knee Andrew P. Davies, Matthew L. Costa, Simon T. Donnell, Malcolm M. Glasgow, Lee Shepstone From Norfolk and Norwich Health Care NHS Trust, Norwich, England
nterior knee pain due to dysplasia of the extensor mechanism is common. We have studied 137 knees (103 patients) in order to identify a rapid and reproducible radiological feature which would indicate the need for further analysis. Overall, 67 knees (49%) had at least one radiological abnormality; 70 (51%) were considered normal. There were five cases of Dejour type-3 dysplasia of the femoral trochlea, nine of type-2 and 12 of type-1. There were 49 cases of patella alta and five of patella infera. Four knees had an abnormal lateral patellofemoral angle (patellar tilt), and in 15 knees there was more than one abnormality. Classification of trochlear dysplasia was difficult and showed poor reproducibility. This was also true for the measurement of the lateral patellofemoral angle. Patellar height was more easily measured but took time. The sulcus angle is an easily and rapidly measurable feature which was reproducible and was closely related to other features of dysplasia of the extensor mechanism. The finding of a normal sulcus angle suggested that seeking other radiological evidence of malalignment of the extensor mechanism was unlikely to reveal additional useful information. The severity of other features of dysplasia of the extensor mechanism correlated with increasing sulcus angle.
A
J Bone Joint Surg [Br] 2000;82-B:1162-6. Received 10 December 1999; Accepted after revision 6 June 2000
A. P. Davies, MRCS, Orthopaedic Research Fellow Flat 32, Cheverton House, Yeovil District Hospital, Yeovil, Somerset BA21 4AT, UK. M. L. Costa, MRCS, Orthopaedic Trauma Fellow S. T. Donnell, FRCS, Consultant Orthopaedic Surgeon M. M. Glasgow, FRCS, Consultant Orthopaedic Surgeon Orthopaedic and Trauma Department, Norfolk and Norwich Health Care NHS Trust, Brunswick Road, Norwich NR1 3SR, UK. L. Shepstone, PhD, Lecturer School of Health and Policy Practice, University of East Anglia, Norwich NR4 7TJ, UK. Correspondence should be sent to Dr A. P. Davies. ©2000 British Editorial Society of Bone and Joint Surgery 0301-620X/00/810833 $2.00 1162
Anterior knee pain is common and usually arises from the 1 patellofemoral joint. Malalignment of the extensor mechanism causes both objective and subjective instability of the knee, and is a frequent cause of referral. The challenge is to identify the small subgroup of patients with abnormalities of the patellofemoral joint which are amenable to surgical correction. Factors contributing to malalignment of the extensor mechanism include dysplasia of the femoral trochlea, dysplasia of the quadriceps represented by abnormal tilting of the patella, the height of the patella, and an excessive lateral distance between the tibial tubercle and 2 the trochlear groove. These represent the full range of dysplasias of the extensor mechanism. Plain radiography is likely to be the primary investigative tool available and dysplasia of the trochlea can be characterised by various radiological features, including the 3 intersection sign, the trochlear depth, and the boss height. 4 They are, however, difficult to measure reproducibly. Measurement of patella alta and patella infera can be made 5,6 by a number of techniques, and patellar tilt by the lateral 7 patellofemoral angle, but these methods are also difficult to reproduce accurately. The sulcus angle reflects that part of the trochlea seen on a skyline radiograph and thus is directly linked to trochlear dysplasia. It can be measured rapidly and reproducibly. There are, however, no published data on the relationship between the sulcus angle and other elements of malalignment of the extensor mechanism. We therefore investigated whether there is such a relationship, and whether measurement of the sulcus angle could act as a reliable and clinically useful indicator of malalignment.
Patients and Methods Between October 1998 and March 1999, 103 new patients aged between 12 and 30 years presented to a single orthopaedic consultant (MMG) with an established knee practice. Of these, 34 had bilateral symptoms, giving a total of 137 knees in the study. All had symptoms of pain or instability (Table I). The patients had three routine radiological examinations: weight-bearing, posteroanterior, lateral and skyline (tangential patellar views) performed in 25° of flexion according to 8 the method of Laurin, Dussault and Levesque. In order to THE JOURNAL OF BONE AND JOINT SURGERY
THE SULCUS ANGLE AND MALALIGNMENT OF THE EXTENSOR MECHANISM OF THE KNEE
Table I. Details of the 137 knees Patellar tilt Normal Abnormal
133 (97%) 4 (3%)
Patellar height diagnosis Infera Normal Alta
5 (4%) 83 (58%) 49 (38%)
Dejour classification Normal Type 1 Type 2 Type 3
111 (81%) 12 (9%) 9 (7%) 5 (4%)
Mean (SD) sulcus angle in degrees
142.4 (8.0)
Mean (SD) patellar height ratio
1.09 (0.19)
Fig. 1 Skyline radiograph of a knee showing the technique of measurement of the sulcus angle (normal range 138°, SD 6°).
Fig. 2 Skyline radiograph of a knee showing the measurement of the lateral patellofemoral angle. It is normal if the angle opens laterally.
guarantee consistency, all examinations were carried out by the same small group of experienced radiographers, using the same equipment. A goniometer was used to measure the angles of flexion. The films were then studied by two observers (APD, MLC) and measurements made from each set included the sulcus angle (Fig. 1), the lateral patello6 femoral angle (Fig. 2) and the patellar height. They were made blinded on two occasions separated by two weeks by VOL. 82-B, NO. 8, NOVEMBER 2000
1163
one observer (AD) and then on a further occasion by a 4 second observer (MLC). A pilot study of Remy et al has shown that the classification of trochlear dysplasia is difficult and has poor reproducibility. We therefore adopted a consensus opinion obtained in conjunction with a senior author (STD), which was reached in the absence of any knowledge of other measurements from the knee con3 cerned. The classification of Dejour et al was used to categorise trochlear dysplasia. Statistical analyses. We calculated the intraclass correlation coefficient (ICC) to assess the reliability of measurements scored on a continuous scale (sulcus angle and patellar height ratio) and the kappa statistic to assess agreement with respect to the diagnosis of patella alta and infera. A generalised Fisher’s exact test (an exact test for contingency tables larger than 2 2) was used to determine an association between the categorical variables of abnormal patellar tilt, patella alta and the classification of 3 Dejour et al . Student’s t-test was used to test for the mean difference in the sulcus angle and the patellar height ratio between normal and abnormal patellar tilt groups. A oneway analysis of variance (one-way ANOVA) was used to determine the mean difference in sulcus angle between the diagnosis of patella alta and the classification groups. Pearson’s correlation coefficient was used to assess the strength of the relationship between the patellar height ratio and the sulcus angle. Statistical significance was set at a level of 5%.
Results There was trochlear dysplasia in 26 knees and type-3 (severe) dysplasia in five (Fig. 3). Of the latter, patella alta was present in four and an abnormal lateral patellofemoral angle in one. Nine knees had type-2 (intermediate) trochlear dysplasia. Of these, one had both an abnormal lateral patellofemoral angle and patella alta and four patella alta. Twelve knees had type-1 (mild) trochlear dysplasia (Fig. 4). Four of these had patella alta and one patella infera. Of the 49 knees with patella alta, 13 also had trochlear dysplasia. There were four cases of an abnormal patellar tilt angle of which three had another abnormality. The intraobserver and interobserver ICCs for the measurement of the sulcus angle were 0.94 and 0.92, respectively, indicating excellent reliability. For the patellar height ratios the values were 0.72 and 0.59, respectively, indicating good reliability. This was also reflected in the kappa statistics for the diagnosis of patella alta which were 0.83 (intraobserver) and 0.60 (interobserver), respectively. The relationship between the sulcus angle and the three categorical variables is shown in Table II. There was a significant difference in the mean sulcus angle between patients with abnormal patellar tilt and those without (p = 0.003). The former group had larger sulcus angles. A significant difference was found between the groups with different patellar height ratios (one-way ANOVA,
1164
A. P. DAVIES, M. L. COSTA, S. T. DONNELL, M. M. GLASGOW, L. SHEPSTONE
Fig. 3b
Lateral (a) and skyline (b) radiographs of a knee showing Dejour type-3 dysplasia of the femoral trochlea. There is a prominent anterior boss.
Fig. 3a
Fig. 4b
Lateral (a) and skyline (b) radiographs of a knee showing Dejour type-1 dysplasia of the femoral trochlea.
Fig. 4a
p = 0.002). Dunnett’s post-hoc comparisons showed a significant difference between the normal and patella alta group but not between the normal and patella infera group. The value of the Pearson correlation coefficient (r) between the patellar height ratio and the sulcus angle was 0.294 which was statistically significant (p < 0.001). Figure 5 shows the relationship between the patellar height and the
sulcus angle. As expected, there was a significant relationship between trochlear dysplasia and the sulcus angle, and there was an overall significant difference between the 3 groups classified according to Dejour et al (one-way ANOVA, p < 0.001). Dunnett’s post-hoc comparisons revealed a significant difference between the normal group and type-3 dysplasia and between the normal group and type-2 dysplaTHE JOURNAL OF BONE AND JOINT SURGERY
THE SULCUS ANGLE AND MALALIGNMENT OF THE EXTENSOR MECHANISM OF THE KNEE
sia but none between the normal group and type-1 dysplasia. There was a significant linear trend (p < 0.001) across the groups from normal to type-3 dysplasia (Fig. 6). There was a nearly statistically significant relationship between trochlear dysplasia and the diagnosis of patella alta (generalised Fisher’s exact test, p = 0.051). Those with Dejour classification type 2 or type 3 were more likely to be diagnosed as having patella alta, while those with patella infera were either normal or Dejour type 1. There was a significant difference between the Dejour classification groups when considering patellar height ratios (one-way ANOVA, p = 0.004); a higher ratio related to a higher Dejour classification. Given the inter-relationships between the categorical variables, a three-way analysis of variance was used to assess the relationship between each variable and the sulcus angle in the presence of the other two variables. The relationship between trochlear dysplasia and sulcus angle remained highly significant (p < 0.001). There was also a significant relationship between the diagnosis of patella alta and sulcus angle, although this was marginal (p = 0.045). There was no evidence of a relationship between patellar tilt angle and sulcus angle when allowing for the other explanatory variables.
1165
Table II. Relationship of the mean (SD) sulcus angle (degrees) to the three categorical variables Sulcus angle
p value
Patellar tilt Normal Abnormal
142.0 (7.61) 153.8 (11.62)
0.003
Patellar height ratio Infera Normal Alta
135.0 (1.76) 141.5 (5.66) 146.1 (12.14)
0.002
Dejour classification Normal Type 1 Type 2 Type 3
140.4 143.1 149.0 171.2
< 0.001
(5.20) (4.36) (5.66) (5.32)
Discussion Fig. 5
For the patient with pain in the anterior knee attributable to the patellofemoral joint, plain radiography is most likely to be the first investigation, although it is increasingly being supplemented by additional imaging techniques such as CT and MRI. These are costly and time-consuming and will not be available to every patient. Many of the plain radiological features of dysplasia of the extensor mechanism described previously, particularly the classification of tro4 chlear dysplasia, show poor reproducibility and are of limited value. Our study suggests that the sulcus angle may represent a simple radiological feature which is easy and quick to measure from plain radiographs, and which will reliably predict the severity or otherwise of dysplasia of the extensor mechanism. Factors contributing to malalignment of the extensor mechanism are quadriceps dysplasia manifested by an abnormal patellar tilt angle, trochlear dysplasia, the height of the patella and the sulcus angle. The lateral distance between the tibial tubercle and the trochlea is also important but cannot be measured from plain radiographs. From our analysis of the various features in our series, we have drawn some conclusions about the varying severity of each factor and its likely relationship to other abnormalities. Overall, 67 knees (49%) had an abnormality of the trochlea, the patellar height, the patellar tilt or a combination of these. The other 70 (51%) were considered normal. While the patients in our series must represent a selected group, this high rate (49%) of radiological abnormalities emphasises the importance of plain radiography as the primary VOL. 82-B, NO. 8, NOVEMBER 2000
Graph showing the relationship between the patellar height and the sulcus angle (degrees).
Fig. 6 Graph showing the relationship between the Dejour grade of dysplasia and the sulcus angle (degrees).
investigation of choice. 3 Dejour et al described type-3 trochlear dysplasia as ‘severe’ and this is reflected in the fact that every knee with type-3 dysplasia also had another abnormality. Type-2 dysplasia was described as ‘intermediate’ and of the nine cases of type-2 dysplasia, five had another abnormality of the same knee. Five of the 11 knees with a type-1 ‘mild’ dysplasia of the trochlea also had another abnormality. We agree that the classification of trochlear dysplasia from plain radiographs is difficult and gives poor reproducibil4 ity, and for this reason a consensus diagnosis was adopted. Lateral patellofemoral angles were time-consuming to
1166
A. P. DAVIES, M. L. COSTA, S. T. DONNELL, M. M. GLASGOW, L. SHEPSTONE
measure. Measured angles varied widely, although both observers were in agreement on the four cases which fell outside the normal range. We also confirmed the finding that measurement of patellar height is reproducible, but not without difficulty. Our interobserver kappa stastistic of 0.59 for the patellar height ratio and of 0.60 for the diagnosis of patella alta reflects good reproducibility, but falls a long way short of the values achieved for measurement of the sulcus angle. The latter was both the easiest and the quickest feature to measure and is by definition directly related to dysplasia of the trochlea. Excellent intra- and interobserver correlation coefficients of 0.94 and 0.92, respectively, were achieved. The normal sulcus angle has previously been defined as 7 138° (SD 6°). This is in close agreement with the ‘normal’ group from this study who were symptomatic but had no abnormal radiological features and had a mean sulcus angle of 140.4° (SD 5.2°). Our study demonstrated a linear progression of increasing sulcus angles as other features of dysplasia of the extensor mechanism increased in severity. Thus for patients with patella infera the mean sulcus angle was 135°. For the normal group it was 141.5° and for patients with patella alta 146.1° (Fig. 3). When considering trochlear dysplasia, the mean sulcus angle for normal knees was 140.4°, for type-1 dysplasia 143.1°, for type-2 149° and for type-3 171° (Fig. 4). Considering the lateral patellofemoral angle, the mean angle for the normal group was 142° and for the abnormal group 153.8°. The relative ease of measurement of the sulcus angle, combined with its reproducibility and close relationship to the other features of dysplasia of the extensor mechanism, make it an ideal radiological feature with which to ‘screen’ knees in a clinical setting. The finding of a normal sulcus angle can reassure the clinician that seeking the other radiological markers of dysplasia of the extensor mechanism is unlikely to reveal additional useful information. In this context, it is unlikely that an extensor malalignment syndrome is the cause of the patient’s symptoms and other diagnoses should be sought.
Reliable measurements can only be made from radiographs which are produced by skilled radiographers using standardised techniques. Skyline radiographs in particular require close attention to detail in order to achieve reliable results. The need for small groups of specialised radiographers experienced in radiology of the knee cannot be overemphasised. If the sulcus angle is abnormal it is likely that other features of the dysplasia will also be present. It is this subset of patients who may have a surgically correctable abnormality of the patellofemoral joint. In this situation, the measurement of the other parameters of malalignment such as the angle of patellar tilt, patellar height, type of trochlear dysplasia and the distance between the tibial tuberosity and the trochlear groove will be most likely to yield further information helpful to the surgeon. If the sulcus angle is measured first, the other measurements can be confined to the group of patients most likely to benefit. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
References 1. Fairbank JC, Pynsent PB, Poorvliet JA, Phillips H. Mechanical factors in the incidence of knee pain in adolescents and young adults. J Bone Joint Surg [Br] 1984;66-B:685-93. 2. Aichroth PM, Al-Duri Z. Dislocation and subluxation of the patella: an overview. In: Aichroth PM, Cannon WD Jr, Patel DV, eds. Knee surgery: current practice. Martin Dunitz, 1992:354-79. 3. Dejour H, Walch G, Neyret Ph, Adeleine P. Dysplasia of the femoral trochlea. Rev Chir Orthop Reparatrice Appar Mot 1990;76:45-54. 4. Remy F, Chantelot C, Fontaine C, et al. Inter- and intraobserver reproducibility in radiographic diagnosis and classification of femoral trochlear dysplasia. Surg Radiol Anat 1998;20:285-9. 5. Blackburne JS, Peel TE. A new method of measuring patellar height. J Bone Joint Surg [Br] 1977;59-B:241. 6. Dejour H, Neyret P, Walch G. Factors in patellar instability. In: Aichroth PM, Cannon WD Jr, Patel DV, eds. Knee surgery: current practice. Martin Dunitz, 1992:408-12. 7. Merchant AC. Patellofemoral disorders: biomechanics, diagnosis and nonoperative treatment. In: McGinty JB, ed. Operative arthroscopy. New York: Raven Press, 1991:261-75. 8. Laurin CA, Dussault R, Levesque HP. The tangential x-ray investigation of the patellofemoral joint. Clin Orthop 1979;144:16-26.
THE JOURNAL OF BONE AND JOINT SURGERY
