Effectiveness of Fulkerson Osteotomy with Femoral Nerve Stimulation for Patients with Severe Femoral Trochlear Dysplasia

Abstract

Background

Patients with femoral trochlear dysplasia are at risk for chronic recurrent patellofemoral dislocations, with extreme cases often requiring a surgical procedure. Anteromedialization of the tibial tubercle with intraoperative femoral nerve stimulation and concurrent medial patella-femoral ligament (MPFL) reconstruction is a previously reported method of maximizing patello-femoral congruency. We hypothesize the Fulkerson osteotomy with intraoperative femoral nerve stimulation and concurrent MPFL reconstruction in patients with severe trochlear dysplasia provides equivalent postoperative clinical outcomes to the same procedure in patients with low level trochlear dysplasia.

Methods

48 knees underwent Fulkerson osteotomy with intraoperative femoral nerve stimulation and concurrent MPFL reconstruction for recurrent lateral patellar dislocations. MRI, surgeon intraoperative assessment, and X-ray were used to assess degrees of trochlear dysplasia; inter-observer and intra-observer error were measured. The knees positive for severe dysplasia on MRI, intraoperative assessment, and X-ray were considered as a comparison cohort to the rest of the study population. We considered postoperative dislocation events and patellar tracking kinematics as outcome measures. Independent student t tests and Fisher exact tests were used to evaluate differences between groups. Significance was set at P<0.05.

Results

11 knees were positive for severe dysplasia (SD) by combined MRI, surgeon intraoperative assessment, and X-ray with the remaining 37 knees categorized as low dysplasia (LD). No patients in either group exhibited apprehension or required re-operation. Mean sulcus angle in the SD group was 175.8 +−2.45 degrees (95% CI 171.0–180.6); the LD group mean sulcus angle was 154.3 +− 0.98 degrees (95% CI 152.4–156.2) (P<.001). Postoperatively there was no significant difference in dislocation events between the SD group (0/11) and the LD group (2/37) (P>0.999). Patellar maltracking decreased in both groups and there were no significant differences in estimates of patellofemoral congruency between the SD (2/11) and LD (8/37) (P>0.999) groups.

Conclusion

The Fulkerson osteotomy with femoral nerve stimulation aimed at maximizing patellofemoral congruency may be an equally effective procedure for patients with either severe or mild trochlear dysplasia.

Level of Evidence

Level III, Retrospective comparative study

Introduction

Previous authors have described various etiologies of patellar maltracking and instability including an imbalance between soft tissue, muscular action and bone morphology. The medial patellofemoral ligament (MPFL) contributes approximately 50–60 % of the total restraining force against lateral patellar displacement^1-8. The trochlea can be described as a concave trough at the distal end of the femur with an average sulcus angle of 138±6° and the shallowness of this groove is described as trochlear dysplasia⁹. A sulcus angle of greater than 145 degrees is suggestive of trochlear dysplasia and has been found to be an important risk factor for recurrent patellar dislocation^10-13. Similarly, a lateral trochlear inclination (LTI) of less than 11 degrees has been described as 93% sensitive and 87% specific for trochlear dysplasia¹⁴. Previous authors identified that 96% of patients with a prior history of a patellar dislocation had evidence of trochlear dysplasia radiographically¹⁰.

Various imaging modalities and techniques have been described in an effort to characterize the bony architecture and pathology associated with patellofemoral instability. A computed tomography (CT) scan or magnetic resonance imaging (MRI) can be used to determine the lateral offset of the tibial tubercle relative to the trochlear groove (TT-TG distance) ^15,16. Patella alta has been described as a risk factor for recurrent patella dislocation. The Caton-Deschamps method is a recognized method of assessing Patella alta and is defined as the ratio between the patellar articular facet length and the distance between the facet and the anterior corner of the superior tibial epiphysis^12,17. A Caton-Deschamps ratio of > 1.2 is suggestive of Patella alta and may indicate a disruption of the engagement between the proximal trochlea and the patella¹⁸. Dejour proposed a classification system of trochlear dysplasia based on 2-dimensional radiographs and 3-dimensional computed tomography (CT) scans with type A described by a fairly shallow trochlea, Type B described as a flat or convex trochlea with a supratrochlear spur creating a prominence and type C as demonstrating an asymmetry of trochlear facets with a hypoplastic medial condyle. Finally, type D is described as an asymmetry of trochlear facets plus a vertical join and cliff pattern^9,19-21.

Physical exam maneuvers including palpation, range of motion and provocative maneuvers can be used to evaluate the patient with suspected patella instability. The most widely used is the apprehension test where the examiner attempts to elicit an apprehensive response from the patient by pushing the patella laterally. Other kinematic tests of patellofemoral congruency are also employed. A positive J-sign describes the movement of the patella over the lateral femoral condyle when the knee is extended during active quadriceps contraction. A J-sign measuring greater than 5 mm may have clinical value in predicting the amount of medialization required during a Fulkerson osteotomy²². The J-sign is very common in the population of patients with lateral patellar dislocations and can be used as a gross measure of patellar tracking²³. An “S-sign” has also been previously described as the initial medial movement of the patella prior to the lateral movement over the lateral femoral condyle during active extension²³.

There are a variety of conservative and surgical methods that have been used to treat patients with lateral patellar dislocations. Conservative treatment of patellofemoral instability includes immobilization, muscle reeducation (quadriceps isometrics, straight leg raises, and single-plane motion exercises) bracing, and taping^24-30. Patellar instability manifested by recurrent dislocations that have failed conservative treatment can be treated by various surgical interventions including soft tissue balancing, tibial tubercle transfer, trochleoplasty, as well as rotational osteotomy of the femur^15,31-33. Most notably, lateral retinaculum release, MPFL reconstruction, trochleoplasty and various bony realignment procedures have all been discussed in the literature with varying degrees of success^31,34-42. In the face of poor patellofemoral congruency, isolated MPFL reconstruction has been shown to be inadequate and result in a high rate of patellar redislocation⁴³. Previous work has cited indications for a sulcus deepening trochleoplasty as reserved for severe cases that includes a prominent supratrochlear spur, high-grade trochlear dysplasia, and patellar instability or abnormal tracking^19,21,44. Fulkerson et al previously described a technique that we commonly use at our institution involving surgical anteromedialization of the tibial tubercle to align the extensor mechanism of the knee thus allowing the patella to move appropriately within the confines of the trochlear groove³¹. Additionally, the senior author of this study has previously used intraoperative femoral nerve stimulation to estimate the amount of correction required to achieve congruency of the patella tracking from 0° to 30° of flexion when performing a Fulkerson osteotomy^45,46. This technique is based on achieving maximal patellofemoral congruency during active quadriceps extension when the femoral nerve is stimulated intraoperatively. This technique also takes into consideration the TT-TG distance to help guide placement of the tibial tubercle, however, the final placement of the tibial tubercle attempts to achieve complete congruency in the dynamic condition of active quadriceps contraction^23,45,46,47. Finally, the technique used by the senior author utilizes MPFL reconstruction concomitantly with anteromedialization of the tibial tubercle to correct the loss of the medial soft tissue checkrein restraints sustained in recurrent lateral patellar dislocations¹³.

The aims of this study are to evaluate the preoperative imaging and postoperative outcomes of patients who underwent Fulkerson osteotomies with intraoperative femoral nerve stimulation and MPFL reconstruction for trochlear dysplasia. We hypothesize there will be no difference in postoperative patellar maltracking and re-dislocation events between patients with severe trochlea dysplasia and patients with low level dysplasia after undergoing Fulkerson osteotomy with intraoperative femoral nerve stimulation and MPFL reconstruction.

Methods

This study is a retrospective comparative study of 48 knees (42 patients: 18 males, 24 females; range 16 to 54 years old) who underwent Fulkerson osteotomies for recurrent lateral patellar dislocations. Exclusion criteria included patients without adequate MRI scans or those with insufficient information in the pre- and postoperative clinical notes. Lateral trochlear inclination (LTI; Figure 1) and Sulcus angle (Figure 2) were determined on die most proximal image of an axial MRI in which the entire width of the femoral trochlea was observed. Our methods for measuring sulcus angle and lateral trochlear inclination index have been previously described⁴⁸ with good reliability previously demonstrated⁴⁹. Patellar height was measured using the Caton-Deschamps method on lateral MRI and TT-TG distance using axial MRI. Inter-observer error was measured between three observers and intra-observer error was measured by each observer repeating the measurements in a different order after a one month period. We used axial MRI slices and a two group classification for trochlear dysplasia which differentiated type A (Figure 3) low-level dysplasia from types B (Figure 4), C, and D (Figure 5) which were all considered to represent severe trochlear dysplasia. This two grade analysis has been previously described using the Dejour trochlear dysplasia classification¹⁹. Additionally the senior author examined and described the intraoperative appearance of each trochlear groove in a graphic illustration of the trochlear contour from the most superior contact point for the patella from through the first 30 degrees of flexion of the knee. Trochlear grooves that were described intraoperatively as flat or having a bump were categorized as severely dysplastic; all others as low-level dysplasia. For each trochlea that was both classified as severely dysplastic (types B, C, or D) on MRI and described intraoperatively as either flat or having a bump, a lateral X-ray was assessed for the presence of an obvious supratrochlear bump or crossing sign. To be classified into the severe dysplasia group (SD) a knee must have been classified as type B, C, or D on axial MRI in at least 66% of observations, described intraoperatively as flat or having a bump, and shown to have a supratrochlear bump or crossing sign on a true lateral X-ray. The remainder of knees that did not meet all three of these criteria were considered as the low-level dysplasia group (LD). Pre-operative and post-operative clinical notes were assessed for the presence of patellar maltracking and re-dislocation. Patellar maltracking was characterized as a clinically gross J-sign, a clinically gross S-sign, or a J-sign of >5mm if measured. Simple kappa coefficients were calculated to describe intra observer and interobserver agreement when determining two group axial MRI classification of trochlear dysplasia¹⁹. Independent student t tests and Fisher exact tests were used to evaluate differences between groups. Significance was set at P<0.05.

Figure 1. Axial MRI - Lateral trochlear inclination.

Figure 2. Axial MRI - Sulcus angle.

Figure 3. Type A - Shallow trochlea.

Figure 4. Type B - Flat trochlea.

Figure 5. Type D - Spur trochlea.

Table 1.

Pre and postoperative clinical findings in patients with severe trochlear dysplasia (SD) and low-level or absent trochlear dysplasia (LD).

Pre- and Post-operative Maltracking	SD Group (n=11)	LD group (n=37)
Total Pre-operative Maltracking	11	34
Post-operative J-sign > 5 mm	1	0
Post-operative Gross J-sign	0	1
Post-operative Gross S-sign	1	7
Total Post-operative Maltracking	2	8

Table 2.

Imaging findings in patients with severe trochlear dysplasia (SD) and low-level or absent trochlear dysplasia (ID).

MRI Measurements	SD group	LD group	p-value
Sulcus Angle (degrees)	175.8 +/− 2.45	154.3 +/− 0.98	1.15×10−12
Lateral Trochlear Inclination (degrees)	−4.8 +/− 1.87	11.0 +/− 0.80	1.12×10−5
Caton-Deschamps Index	1.25 +/− 0.23	1.26 +/− 0.182	0.42
TT-TG Distance (mm)	21.25 +/− 3.06	19.44 +/− 3.77	0.059

Results

The SD group consisted of 11 knees which were positive for severe dysplasia by combined MRI, intraoperative assessment, and lateral view plain X-rays. The remaining 37 knees were placed into the LD group. Mean follow up time was 8.96+/−1.41 months for SD group and for 13.02+/−1.97 months for the LD group. Minimum followup was 1 month and maximum followup was 45 months. Mean sulcus angle was 175.8 +/−2.45 (95% CI 171.0–180.6) degrees and lateral trochlear inclination was −4.8 +/− 1.87 (95% CI −8.4 − −1.1) degrees in the SD group. The LD group mean sulcus angle was 154.3 +/−0.98 (95% CI 152.4–156.2) and with a lateral trochlear inclination of 11.0 +/−0.80 (95% CI 9.4–12.6) degrees. Both the mean sulcus angle and mean lateral trochlear inclination differed significantly between the SD and LD groups (P<.001). Mean Caton-Deschamps index was 1.25+/− 0.23 and 1.26+/−0.182 in the SD and LD groups respectively (P<0.42). 7 out of the 11 knees in the SD group surpassed the cutoff for patella Alta (Caton-Deschamps ratio >1.2) while 21 out of the 37 knees in the LD group had an index of >1.2. TT-TG distances were 21.25+/−3.06 mm and 19.44+/−3.77 mm in the SD and LD groups respectively (P=0.059). Simple kappa coefficients were calculated to describe intraobserver (Table 3) and interobserver agreement when determining two group axial MRI classification of trochlear dysplasia (Table 4). Intraobserver agreement analysis were as follows: Rater 1 (Kappa coefficient [KC], 0.7500 (95% confidence interval [CI], 0.5635–0.9365); rater 2 (KC, 0.4615; 95% CI, 0.2615–0.6616); rater 3 (KC, 0.4621; 95% CI, 0.2175–0.7067). Interobserver agreement for Time 1 are as follows: Rater 1 versus rater 2 (KC, 0.4167; 95% CI, 0.1677–0.6657); rater 1 versus rater 3 (KC, 0.5833; 95% CI, 0.3568–0.8099); rater 2 versus rater 3 (KC, 0.3913; 95% CI, 0.1269–0.6557) (Table 3). Interobserver agreement for Time 2 were as follows: Rater 1 versus rater 2 (KC, 0.4857; 95% CI, 0.2432–0.7283); rater 1 versus rater 3 (KC, 0.7909; 95% CI, 0.6177–0.9642); rater 2 versus rater 3 (KC, 0.4507; 95% CI 0.2077–0.6937). There were no dislocation events in the SD group (0/11) after surgical intervention while two patients from the LD group had a dislocation of the patella postoperatively (2/37) (P>0.999). Patellar maltracking decreased from 11 to 2 in the SD group and from 34 to 8 in the LD group; there was no difference between groups (P>0.999).

Table 3.

Intraobserver error in measurements assessing the two group axial MRI classification of trochlear dysplasia

Intraobserver Simple Kappa Coefficients	Rater 1	Rater 2	Rater 3
Kappa Value	0.7500 (95% CI 0.5635–0.9365).	0.4615 (95% CI 0.2615–0.6616)	0.4621 (95% CI 0.2175–0.7067)

Table 4.

Interobserver error in measurements assessing the two group axial MRI classification of trochlear dysplasia

Interobserver Simple Kappa Coefficients	Rater 1 - Rater 2	Rater 1 - Rater 3	Rater 2 - Rater 3
Kappa Value Run 1	0.4167 (95% CI 0.1677–0.6657)	0.5833 (95% CI 0.3568–0.8099)	0.3913 (95% CI 0.1269–0.6557)
Kappa Value Run 2	0.4857 (95% CI 0.2432–0.7283)	0.7909 (95% CI 0.6177–0.9642)	0.4507 (95% CI 0.2077–0.6937)

Discussion

Femoral nerve stimulation is a previously described successful method for intraoperative assessment of patellar tracking when performing anteromedialization of the tibial tubercle; however its effectiveness in those with severe femoral dysplasia has not been directly assessed^22,23,45,46. Our work primarily focused on the ability of our unique method of maximizing patellofemoral congruency to stabilize even the most dysplastic femoral trochlea. However, in doing so we also included measurements for many of the known radiographic risk factors for lateral patellar dislocations including sulcus angle, lateral trochlear inclination, TT-TG distance, and Caton-deschamps index in addition to qualitative indices of pre- and post-operative patellofemoral function. We also describe intra and interobserver differences when using the 2-grade analysis for trochlear dysplasia. In the setting of anteromedialization of the tibial tubercle with intraoperative femoral nerve stimulation and MPFL reconstruction, we report there were no significant differences in postoperative dislocation events or patellar tracking between patients with severe trochlear dysplasia and low-level dysplasia.

Trochlear dysplasia has been found to be an important risk factor for recurrent patellar dislocation with a sulcus angle of greater than 145 degrees suggestive of trochlear dysplasia^50,51,52. Authors have also reported a Caton-Deschamps ratio greater than 1.2 and LTI less than 11 degrees are suggestive of patella maltracking and dysplasia^53,54. Previous authors identified that 96% of patients with a prior history of a patellar dislocation had evidence of trochlear dysplasia radiographically⁵⁰. Previous work has also shown intraob server and interobserver reliability using the two group axial classification of trochlear dysplasia to be 70–90% and 62–86% respectively indicating good to excellent agreement¹⁹. We report sulcus angle and LTI differed significantly in the SD versus LD groups of our cohort. The mean sulcus angle for the SD group was 175.8 +/−2.45 degrees while the LD group mean sulcus angle was 154.3 +/− 0.98 degrees. LTI was −4.8 +/− 1.87 and 11.3 +/− 0.8 and degrees in the SD and LD groups respectively (p <0.001). The LTI LD group mean was close to the numeric cutoff of 11 degrees for trochlear dysplasia⁵³. The LTI SD group mean was calculated to be negative reflecting the extreme amount of subluxation and lateral tilt that often moved the patella completely outside of the femoral trochlea. TT-TG distances were not significantly different between the two groups (P=0.059), however, the distances were greater in the severely dysplastic group. The mean Caton-Deschamps indices were 1.25+/−0.23 and 1.26+/−0.82 for the SD and LD groups respectively with no significant difference between groups. Using a standard cutoff of >1.2 for Caton-Deschamps index both groups had a majority of knees reaching the radiographic cutoff for patella alta⁵⁴. We speculate that variation in LTI measurements are due to the observed difficulty in measuring angles in extremely dysplastic trochleas as well as differences between observers in choosing the most proximal image of the entire trochlea. The majority of values for intraobserver (Table 3) and interobserver (Table 4) kappa coefficients were calculated between 0.4 and 0.6 indicating moderate agreement which is slightly lower than previous reports¹⁹. While direct visualization of the trochlear surface by surgeons will provide the most information, radiographs and MRIs can provide important planning information regarding the architecture of the femoral trochlea before any operation is done. Axial MRI measurements reveal the architecture of the trochlea, evidence of articular damage, and allow the surgeon to estimate the amount of anteromedialization needed. Still, the placement of the tibial tubercle is finalized after direct observation of active quadriceps contraction using intraoperative femoral nerve stimulation. The utility of sulcus angle, LTI, and the Dejour grading method should be further assessed. While we have described significant differences in both sulcus angle and LTI between those with severely dysplastic trochleas and those with low-grade dysplasia, we have not qualified these data into clinically useful measures.

Femoral nerve stimulation allows for the reproduction of active forces and increases a surgeon's ability to simulate patellofemoral tracking intraoperatively^45,46. The tibial tubercle placement is often modified after quadriceps stimulation as the lateral forces on the patella become apparent and increased medialization is required. Our described technique allows for the surgeon to avoid damaging the trochlear cartilage or underlying tissue as may be necessary in sulcus deepening trochleoplasty⁵⁵. Our results demonstrate that no knee in the SD group (n=11) and only 2 knees in the LD group (n=37) experienced a patellar dislocation following surgery (P>0.999). In addition, total patellar maltracking decreased from 11 to 2 in the SD group and from 34 to 8 in the LD group following surgery (P>0.999) The absence of dislocations in the SD group suggests that the Fulkerson Osteotomy with intraoperative femoral nerve stimulation and concurrent MPFL reconstruction may be as effective in stabilizing the patellofemoral joint in cases of severely dysplastic trochlear architecture as compared to cases of mild dysplasia. We also note the resolution of gross patellar maltracking in 9 of the 11 knees in the SD group which was not significantly different from the LD group; we report there was an improved kinematic patellofemoral relationship in both the SD and LD groups after our surgical technique. Our results suggest that the Fulkerson Osteotomy with intraoperative femoral nerve stimulation and MPFL reconstruction is as effective in stabilizing the patellofemoral joint in those with severe trochlear dysplasia as it is in those with low or absent trochlear dysplasia.

Limitations of the present work include the retrospective nature of this study. Patients were not contacted to obtain any new information and follow up times were variable. As previously discussed, axial MRI analysis was done by choosing the most proximal part of the trochlea in which the entire trochlear surface could be visualized. Difficulty was encountered between observers in choosing similar axial MRI cuts to make measurements. This may account for some of the variability in grades of dysplasia as well as angle measurements between observers. In addition, highly dysplastic trochleas often had no true sulcus to measure. The convex and laterally displaced trochlear surface often had sulcus angles of greater than 180 degrees and lateral trochlear inclinations of less than 0 degrees when referencing to a line through the posterior condyles. Additionally, information was derived from clinical notes written by different orthopaedic clinicians at different levels of training who presumably had varying exam techniques, skill level, clinical knowledge and thresholds for determining certain clinical characteristics. In particular, some clinicians reported patellar tracking as a gross J- or S-sign while others measured the J-sign objectively. The clinical cutoffs for a gross J- or S-signs for these particular clinicians are therefore not possible to ascertain. Lastly, our followup is limited, with some patients completing only 1 month. It is possible that some patients might go on to have late dislocations, and thus further follow-up is warranted in the future.

Conclusion

We report sulcus angle, lateral trochlear inclination, Caton-Deschamps Index, and TT-TG distance measurements in a cohort of 48 knees that underwent Fulkerson osteotomy with femoral nerve stimulation and MPFL reconstruction for patellar instability. The SD group consisted of 11 patients with severe dysplasia that have been reported to be candidates for the extreme measure of reshaping the trochlea^10,44,55. We report moderate intraobserver and interobserver reliability for the two group axial MRI classification of trochlear dysplasia. We find significant differences in both sulcus angle and lateral trochlear inclination between the SD and LD groups. We find no difference in dislocation events or patellar tracking between severely and mildly dysplastic trochleas after Fulkerson osteotomy with femoral nerve stimulation and MPFL reconstruction. We report that Fulkerson osteotomy with femoral nerve stimulation and MPFL reconstruction may be an equally effective procedure for those with both severe and mild trochlear dysplasia. The restricted chart review follow-up of this cohort will require functional outcome and activity based assessments in the setting of a long term follow-up to confirm the success of this approach.