Southwick-Fulkerson Osteotomy with Intraoperative Femoral Nerve Guidance

Abstract

Background

Patellofemoral instability is a complex problem with most previous treatment plans addressing static alignment and static stabilizers. Although the quadriceps muscles are known to affect the tracking of the patella, they are rarely taken into account during a surgical procedure.

Purpose

The purpose of this study is to determine the two year minimum results of 37 knees which received a Southwick-Fulkerson Osteotomy and MPFL repair or reconstruction both under the guidance of femoral nerve stimulation

Methods

Patients underwent a Southwick-Fulk- erson Osteotomy and either medial patellofemoral ligament (MPFL) repair or reconstruction using femoral nerve stimulation as a means of dynamic intraoperative evaluation of patello-femoral con- gruity in terminal extension. Two year minimum outcomes of 26 patients, 31 knees (84% return rate) were evaluated using KOOS and IKDC scores, and physical exam features of apprehension and assessment of dynamic tracking in the last 30 degrees of knee extension. Variables were evaluated with t-tests and ANOVA.

Results

29/31 knees reported they were happy with the procedure and reported they would do it again. One knee (3%) reportedly “redislocated”, but did not return for verification by exam. 30/31 had non-pathologic tracking. One knee displayed a small but residual J sign. 4/16 knees with MPFL repair only and 0/15 with MPFL repair and reconstruction exhibited a positive apprehension sign. Increased age and apprehension were correlated with lower outcome scores.

Conclusions

Intraoperative femoral nerve stimulation is an effective way of evaluating patellar tracking intraoperatively that leads to 97% stable patellae with near congruent patello-femoral tracking. MPFL reconstruction is superior to MPFL repair in eliminating the persistence of the apprehension sign.

Introduction

The Southwick-Fulkerson Osteotomy is a novel surgical procedure designed to fully correct patellofemoral incongruity in a manner that is sufficiently stable to withstand quadriceps contraction under anesthesia. Chronic recurrent patellofemoral instability is a pathologic condition often involving maltracking of the patella in the groove of the femur (trochlea) that leads to subluxation or frank dislocation episodes. Patellofemoral instability can stem from a multitude of problems which usually include a variety of anatomical defects of the lower extremity. These defects may involve dynamic and passive forces¹.

The mix of issues causing this problem often makes patellofemoral instability difficult to treat. Over 100 different surgical procedures have been devised and used to treat this single issue of patellar instability². The primary problem with obtaining success with any one treatment method stems from the multi-factorial nature of each patient’s pathology. Not only are the contributing factors of each patient’s instability different, but current clinical evaluation techniques focus on static restraints of the patellofemoral joint and do not take into account the role of the dynamic muscle-tendon anatomy. Based on our review of the English speaking literature, other than VMO advancements, the role of dynamic movement associated with muscle contraction has never been a major focus¹.

The procedure introduced in this case series combines the advantage of anterior-medialization of the tibial tubercle (Fulkerson) with the added tubercle stability realized by the (Southwick) proximal dovetail shelf architecture³. To correct for high incidence of apprehension associated with isolated distal transfer techniques ⁴, in this study the tubercle transfer was done in combination with either a medial patellofemoral ligament (MPFL) repair alone or reconstruction to supplement the repair.

Currently, transfer distance is calculated preoperatively through the tibial tubercle – trochlear groove (TT-TG)⁵ measurement of the bony alignment on CT or MRI or intraoperatively with passive attempts to force the patella to escape the trochlear groove through various combinations of varus and valgus and internal and external rotation. Interestingly, proponents of this technique are hesitant to fully correct the tracking of the patella in the trochlear groove for fear of overcorrection and medial dislocation. However, we have hypothesized that for many patients, residual post-operative maltracking stems from dynamic forces created by the muscle which cannot be evaluated through these methods. To avoid this problem, we teamed with the anesthesia department in developing a method of intraoperative femoral nerve stimulation⁶ that allows for dynamic tracking to be assessed throughout the case. Using this approach, it is the intraoperative goal to achieve complete congruity of the patella in the confines of the trochlear borders in the initial phases of flexion of the knee. We also hypothesize that reconstruction of the MPFL will eliminate the sensation of apprehension often found in patients where congruity is established but only repair of the compromised static stabilizer is attempted.

The purpose of this study is to retrospectively evaluate the two year minimum results of 37 knees which received a Southwick-Fulkerson Osteotomy and MPFL repair or reconstruction under the guidance of femoral nerve stimulation in order to evaluate the effectiveness of this technique in determining tubercle transfer distance.

Methods

I. Patient Population

This study received institutional IRB approval. Thirty- seven knees (17 right and 20 left) in twenty-six patients underwent a Southwick-Fulkerson Osteotomy between November 2002 and April 2007. This included all patients who underwent a Southwick-Fulkerson Osteotomy and MPFL repair or reconstruction under the guidance of femoral nerve stimulation and were at least two years postoperative. The decision to repair or reconstruct the MPFL was made on a chronological basis, creating a prospective randomized study. In total, the first 16 patients received an isolated MPFL repair and the next 21 patients received an MPFL repair augmented with an allograft reconstruction. All patients received the same post-operative instructions and rehabilitation.

Inclusion criteria included a history of recurrent dislocations with evidence of maltracking through patellar subluxation in the merchant view⁷ and/or a grossly positive J-sign. TT-TG measurements were consistent with lateral malalignment and ranged from 12-24 mm on CT imaging. This study only included patients undergoing the combination of Southwick-Fulkerson Osteotomy and MPFL repair or reconstruction under the guidance of femoral nerve stimulation. Exclusion criterion included: no patient requiring rotational femoral osteotomy, independent Fulkerson-Southwick osteotomy, independent MPFL reconstruction or repair, and no patient with a previous diagnosis of a generalized soft tissue hyperlaxity syndrome.

II. Catheter Placement

As previously described by Lavery et al⁶. under local anesthesia, a femoral nerve catheter is placed directly on the nerve through a needle and attached to a small stimulator. Ultrasound guidance is used at the discretion of the anesthesia team. The catheter is adjusted until the extensor muscles twitch with stimulation between 0.3-0.5 mA. The femoral nerve catheter is not dosed prior to surgery in order to be able to stimulate the nerve during the procedure. With the frequency set to 50 Hertz, the intensity is slowly increased until a tetanic stimulation was obtained. The intensity is increased slowly to prevent post-operative neuropathy, which can be inflicted with high intensities.

III. Surgical Correction

An incision is made extending from the proximal portion of the medial femoral condyle to the distal end of the tibial tubercle. The patellar tendon is identified and clearly defined. At this point, lateral structures up to the mid-portion of the patella and medial structures up to the vastus medialis are released and the femoral nerve is stimulated. This exposure is necessary to directly assess the tracking of the patella in relationship to the trochlear borders while not under the influence of soft tissue structures. Tracking is dynamically assessed with the quadriceps muscle extending the knee from 30 degrees of flexion to full extension. A pin is placed through the frontal plane of the tibial tubercle at an angle varying between 0 and 30 degrees in order for the tubercle to move anteriorly as it is medialized. The pin is used as a guide for the oscillating saw. The osteotomy separates the bone into approximate anterior one third and posterior two thirds, see figure 1. This cut begins at the proximal portion of the patellar tendon insertion and extends to the distal portion of the tubercle. The cut narrows as it extends distally, but is never completed in order to have a distal hinge bridging the tibial crest with the osteotomy. A small sagittal saw is then used to create a dovetail at the proximal portion of the initial cut. This overhanging ledge provides an additional point of stabilization. A bone tamp is used to transfer the tubercle anteriomedially to the TT-TG measurement previously measured on CT. Three temporary pins are placed to hold the tubercle in position and manual pressure is applied while the femoral nerve is being stimulated.

Figure 1. Fulkerson Osteotomy with Southwick dovetail architecture and MPFL repair. used with permission from Techniques in Knee surgery.

IV. Nerve Stimulation

Prior to the osteotomy, the nerve is stimulated in order to evaluate the corrected tracking of the patella. Stimulation is begun with the knee flexed 90 degrees and the leg is guided through the contraction to full extension. This replicates the force of the quadriceps as if the patient were extending their knee while awake. Throughout range of motion, the surgeon directly visualizes the patella in relation to the lateral and medial trochlear edges in an effort to determine the ideal tubercle position. Once the osteotomy has been made, the action of the quadriceps is only tested through the terminal 30 degrees of extension to closely assess the achievement of congruity in this critical range.

V. Secure Tendon Fixation

Once appropriate placement has been determined, two or three bicortical screws are placed through the osteotomy fragment and into intact bone. Fluoroscopy is used to evaluate the adequacy of screw purchase on the tibia.

VI. Medial Patellofemoral Ligament Reconstruction

Prior to beginning reconstruction of the MPFL in a two-tailed manner as described by Fithian⁸ , the original medial and lateral ligaments and capsule are plicated to allow for tissue balancing adjustments to be made that take into account the new tracking of the extensor mechanism. This completes the repair of the lateral structures. For MPFL reconstruction, the isometric anchor point on the distal medial femoral condyle is first identified. It is located distal to the adductor tubercle at the border of the vastus medialis and the anterior margin of the MCL. A pin is passed through the femur for placement of the ligament. If pin placement is acceptable, a suture is looped around the pin with the ends extending onto the medial patellar edge at a) its midportion and b) the junction between the middle and superior thirds. The knee is passively moved to test for the adequacy of tissue balancing and also evaluated dynamically through nerve stimulation. If these are satisfactory, a 6-7 mm tunnel for the hamstring allograft is drilled and it is fixed in place with an endobutton on the lateral side of the femur. Suture anchors are then drilled into the patella. The stimulator is again used to validate the balanced points for attachment of the allograft tendon and to confirm the patella is not over-constrained. When these points have been determined, the ends of the allograft are sutured to the top of the patella using Krakow stitches.

If a repair is elected, the medial capsule including the original MPFL is imbricated.

VI. Post-operative Protocol

Patients are able to partial weight bear immediately following the procedure. Crutch use is recommended for two weeks and brace use for six weeks. Range of motion from 0 to 90 degrees is allowed immediately post-operatively. A safe, rapid return to full weight bearing is allowed for two structural reasons. First, the osteotomy is not as deep in this procedure as originally described by Fulkerson. Second, superior fixation achieved by the five point fixation proven intraoperatively to be able to withstand the forces of the quadriceps activation assures security of the construct in the early post surgical period. This allows the patient to be ambulatory in the early post-operative period compared to the six week delay reported with the Fulkerson technique of fixation. As a precaution, patients are cautioned to remain within the envelope of comfort and evaluated radiographically at 2 and 6 weeks post-operative to assure they are not developing a fatigue fracture and to ensure proper healing of the tibia. Physical therapy, consisting of range of motion exercises and gradual muscle strengthening, is recommended for six weeks with at home continuation.

VII. Assessment of Results

These patients were asked to return to clinic for follow-up and to complete functional evaluation forms, including the International Knee Documentation Committee functional evaluation form (IKDC)⁹ and Knee Injury and Osteoarthritis Outcome Score (KOOS)¹⁰.

The more aggressive fixation used in this procedure has been observed to lead to a dramatically more comfortable early post-operative course including earlier return to weight bearing and everyday activities such as work and driving. We therefore included questions to reflect these outcome measures in our evaluation.

Patients also received a physical exam to determine clinical outcomes by both the primary surgeon (JPA) and a member of the research team (CJM). If disagreements occurred, the patient was re-evaluated by both to determine an agreed upon result. Patellar tracking was observed in a sitting and supine position, and while walking. The “J-sign”, which was invariably present before the surgery was evaluated as a sign of corrected tracking. The “J-sign” is the movement of the patella past the lateral femoral condyle when the knee is in full extension. Patients were also evaluated for an “S-sign”, which we defined as the movement of the patella medially prior to moving laterally past the lateral femoral condyle. With the patient relaxed and with their knee flexed slightly, the examiner looks for apprehension when attempting to displace the kneecap laterally. This was done with maximum effort, placing 15-20 pounds of force directly lateral at both 20 degrees of flexion and full extension. Range of motion and the presence of crepitus were also recorded.

Statistical Analysis

All information was analyzed through independent variable t-tests. Multi-variable analysis was done with the use of Chi squares. A p-value of less than .005 was determined to be significant.

Results

Six knees were lost to follow-up, leaving thirty one included in the study (84% return). Sixeen received MPFL repair and fifteen were supplemented with an

MPFL reconstruction. The patient group consisted of seven males and nineteen females. Average time of follow-up was 3.72 years, ranging from 2 years to 5.95 years. The average age at the time of surgery was 24.06 years old (range 14-47). Twenty-four knees returned to clinic. The remaining seven completed the survey through mail correspondence with clinical exam findings taken from our most recent clinic note and confirmed through phone interviews.

I. Patient Satisfaction:

Twenty-nine knees (93.5%) reported they considered the surgery successful and would repeat it.

II. Transfer distance:

In 8 of the 31 cases, the pre-operatively determined distance was altered due to intraoperative tracking. This added distance varied from up to 5mm in this study group. None were decreased.

III. Exam findings:

One patient, or 3.22%, reported a subsequent dislocation. She did not return for evaluation, but was seen by an outside orthopedic surgeon who reported her patella to be adequately stabilized. None reported subluxation. Thirty patients had correction of their J-sign to the point of being grossly described as having “relatively normal” tracking, 96.78% of the 31 followed-up. “Perfect tracking” (total absence of J-sign) was present in 14 of the 31 patients, 45.16%. A slight J-sign, lateralization of only a few millimeters with hyperextension, was noted in eight patients, 25.8%, and an “S” sign (where the patella was observed to move medially from 30 to 10 degrees of flexion and then laterally to neutral in complete extension) was noted in another eight patients, 25.8%. This left one patient, 3.22%, demonstrating an obvious J sign of more than five millimeters. The existence of the obvious J sign was present only after suffering a post-operative trauma.

Four patients, 12.9%, had apprehension at follow-up exam. All four had undergone an MPFL repair, not a reconstruction.

IV. Functional Outcome Scores:

Postoperative scores were obtained for all 31 knees. Overall, the mean KOOS scores were: symptom 52.4 (sd ±13.9), pain 79.9 (sd±20.28), ADL 85.34 (sd±18.24), sport 61.61 (sd±31.36) and QOL 64.11(sd±24.68). For the entire group, the mean postoperative IKDC score was 68.4 (sd±22.78). Unfortunately, preoperative KOOS and IKDC scores were available in only 13 of our 31 knees.

For the 13 knees with pre-operative scores, preoperative averages were symptom 55.65, pain 55.69, ADL 62.53, sport 27.66 and QOL 36.64. Post-operative averages improved to symptom 52.64, pain 84.85, ADL 60.35, sport 64.6 and QOL 68.4.

V. MPFL repair vs. MPFL reconstruction:

An MPFL ligament reconstruction was found to be statistically superior to an MPFL repair in every section of KOOS except quality of life and IKDC outcomes (table 1). MPFL reconstruction also correlated with decreased patient apprehension. Apprehension was associated with a decreased KOOS symptom score (p=.0284).

Table 1.

	MPFL Repair	MPFL Reconstruction	p-value
KOOS Symptom	47.99	57.14	.066
KOOS Pain	73.09	87.22	.05
KOOS ADL	78.21	92.94	.0215
KOOS Sport	49.37	74.66	.0221
KOOS QOL	57.42	71.25	.11
IKDC	59.15	78.27	.0168

VI. Recovery time milestones:

Recovery periods were measured in days. Patients continued to use their crutches, at least part time, for an average of 35.72 (sd±24.98) days and also continued brace use for an average of 59.24 (sd±39.15) days. Patients returned to driving and work an average 52 (sd±67.65) and 73.44 (sd±72.15) days after surgery, respectively. Patients reached their self-declared full recovery at an average 190.03 (sd±122.33) days.

All recovery time milestones were biased to two patients who reported far outlying results as a consequence of neuropathy as a result of the femoral nerve block. Both patients reached a full recovery within a year.

Discussion

In the modern literature, aside from outcome scores, the results of reporting any one technique for patellofemoral maltracking have been limited to a) re-dislocation and b) need for re-operation¹. There have been no studies in which visually determined, subjective assessment of patellar tracking or apprehension have been reported. When these classic criteria are applied to our study, our results are equal to or superior to most current literature on the treatment of patellofemoral instability at 96.68%, due to one patient reporting a dislocation. However, given our focus on the dynamics of the extensor mechanism, a broader criterion for success was necessary which includes the direct assessment of patellar tracking and the existence of apprehension. Our study has a success rate of 80.6%, when success is defined as near normal tracking, no apprehension, no subluxations or dislocations and did not require a repeat operation. When only the patients receiving the MPFL reconstruction are evaluated, this success rate raises to 93.75%.

Patients were generally pleased with the outcome of their surgery with 93.5% reporting that they both considered the surgery successful and would undergo the procedure again. While the same number of patients were unhappy with one of these outcomes, it is important to note that there was one patient who considered their surgery successful, but would not do it again as well as one patient who did not consider their surgery successful, but would choose to have the procedure again. While this is not a common outcome measure, Morshuis et al² found only 80% of patients subjectively satisfied at first evaluation, with this number dropping to 65% at an average follow-up of 30.4 months.

Only one patient in our study, 3.22%, experienced an episode of dislocation post-operatively. This matches outcomes of MPFL repair¹² and improves upon the instability reportedly found in patients who received the Elmslie-Tillat procedure¹³. Post-operative KOOS and IKDC scores were improved to a significant level, however, comparison to other studies is difficult due to the difference in outcome measures. Additionally, this procedure was only performed on patients who experienced repeated patellar dislocations and did not include patients who experienced only patellofemoral pain or were one time dislocators, as many MPFL studies include. The closest patient cohort is described in Ebinger et al¹⁴. In their small pilot study of seven cases, the same criterion of success was used and one dislocation was present. Although this matches the number of dislocations present in this study, it represents a higher percentage of their cohort at 14.28%.

An important finding in the study was the improved outcome of MPFL reconstruction over repair. While most studies agree that attention should be paid to the MPFL¹⁵, the difference between reconstruction and repair has not been effectively evaluated. MPFL reconstruction has been proven alone to be an effective method of patellar stabilization^12,16,17, however, due to the tendency of the MPFL to tear near its insertion into the femur, repair of the ligament near the patella has been questioned as an effective treatment⁸. Although there is discussion of the downfalls of MPFL reconstruction through over tensioning of the ligament causing increased stress rates on the medial femoral condyle and medial facet of the patella^1,15, our results show the superiority in stability and outcome scores. We believe we eliminate this problem through the additional use of the femoral nerve stimulator as an evaluation of ligament placement and graft tensioning on the already transferred extensor mechanism by observing that there is no added tension on the reconstruction imposed by the contracting quadriceps. In addition to post-operative outcome scores, apprehension was decreased with MPFL reconstruction, a sign of more severe malalign- ment¹⁸. A previous study of dovetail procedures reported by the senior author found a 20% incidence of retained apprehension signs despite no history of redislocation⁴. The elimination of apprehension found in this study is an important outcome of the procedure as it allows patients to experience more everyday activities without the fear of subluxation or dislocation. A decrease in apprehension also improved KOOS symptom scores. Although these statistics could be interpreted as an MPFL reconstruction being sufficient for realignment without the additional osteotomy, the authors of this paper continue to believe static repairs will not last if the dynamics behind them remain unchanged¹⁴.

While early weight-bearing can have negative effects¹⁹, patients included in this study were able to return to successive milestones at equal or improved time than patients undergoing other realignment procedures without complication. While these patients averaged about five weeks (range less than one to 13 weeks) on crutches, most studies describe a protocol of six to eight weeks²⁰. Return to everyday activities, such as driving, were not evaluated in other studies despite their impact on the patients’ perception of the surgery. This study reports an average 190 days, or about six months to a self-reported full recovery. At six months post-operatively in other studies, patients would only be starting to return to running with return to sports occurring at nine to twelve months¹³. Other recovery milestones could not be found in existing literature, but are provided here as a basis of comparison for future studies. While it is agreed that over-ambitious post-operative protocol can lead to further complications, it is believed that this procedure’s improved placement of the tubercle, in addition to dovetail and double screw fixation led to an improved recovery time. Only a review of MPFL reconstruction without distal realignment matched the post-operative recovery times of this study¹⁵.

In conclusion, we feel the intra-operative evaluation of the dynamic knee extensor mechanism during tibial tubercle realignment and MPFL repair/reconstruction provides important information to supplement existing techniques in this difficult patient population. This study has several limitations. These include, but are not limited to; lack of a control group which had the same procedure without the femoral nerve catheter and lack of pre-operative objective scoring measures for all patients included in this study. Further study into the soft tissue anatomy of this cohort is needed to improve upon the understanding of patellar maltracking and instability.