Abstract
Background: Staple epiphysiodesis is an option for the treatment of limb-length discrepancies, but it is not without complications. The purpose of this study was to review the outcomes of staple epiphysiodesis, including changes in the mechanical axis.
Methods: The study included patients who underwent, between 1990 and 2005, staple epiphysiodesis of the femur or tibia, or both, to address limb-length discrepancy. We reviewed preoperative, postoperative, and final long standing anteroposterior radiographs of fifty-four patients to assess limb-length discrepancy, shifts in the mechanical axis, changes in the mechanical axis zone, and changes in the anatomic lateral distal femoral angle and the medial proximal tibial angle. Postoperative radiographs were also reviewed to assess the adequacy of staple placement.
Results: Three staple epiphysiodesis groups were identified: fifteen patients who underwent a distal femoral staple epiphysiodesis, eighteen who underwent a proximal tibial procedure, and twenty-one who underwent combined distal femoral and proximal tibial procedures. Fifty percent (twenty-seven) of the fifty-four patients showed a shift in the mechanical axis of =1 cm as compared with the preoperative measurement. Eighty-nine percent of these large shifts were varus in nature. The proximal tibial and combined epiphysiodeses resulted in significantly larger shifts in the mechanical axis (p = 0.002 and p = 0.006, respectively) and zone changes (p = 0.009 and p = 0.006, respectively) than did the distal femoral procedures. Six patients ultimately underwent a high tibial osteotomy to correct a post-stapling varus deformity. The proximal-lateral aspect of the tibia was by far the most common location for inadequate staple placement.
Conclusions: Mechanical axis deviation is common following staple epiphysiodesis for the treatment of limb-length discrepancy. Proximal tibial and combined distal femoral and proximal tibial staple epiphysiodeses, even if done well technically, lead to clinically relevant shifts in the mechanical axis of the lower extremity more than half of the time. Distal femoral staple epiphysiodesis may still be a safe option for the treatment of limb-length discrepancy, but we advise caution when utilizing proximal tibial staple epiphysiodesis to treat limb-length inequality.
Level of Evidence: Therapeutic Level IV. See Instructions to Authors for a complete description of levels of evidence.
Since Blount and Clarke first reported on the control of bone growth by physeal stapling1, orthopaedic surgeons have modulated both limb-length and angular deformities with use of this technique. Before the era of stapling, the options available to correct a limb-length discrepancy included shoe-lifts, bone-shortening or bone-lengthening, and open epiphysiodesis, as described by Phemister2,3. More recent developments in the treatment of limb-length discrepancy include percutaneous epiphysiodesis and percutaneous epiphysiodesis with use of transphyseal screws4,5. The general indications for epiphysiodesis include a limb-length discrepancy of >2 cm and <5 cm, adequate remaining growth potential of the short limb, and a predicted adult height of >5 ft (1.5 m)6.
In 1971, Frantz reviewed the results of 189 physeal stapling procedures performed for the treatment of both limb-length and angular deformities at his institution and another 577 cases reported in the literature7. He noted a variety of complications, including buried staples; metal reaction; overcorrection; premature closure of the physis; peroneal nerve palsy; knee joint laxity; misplaced, extruded, or fractured staples; knee hyperextension; genu valgum; infection; and false aneurysm. Many of the early complications were due to errors in surgical technique, such as raising a periosteal flap, misplacing staples, and leaving staples in for too long8,9. Other reports have focused on overcorrection and undercorrection of limb lengths with the mention of the occasional angular deformity requiring osteotomy10. To our knowledge, no one has previously analyzed the effect of physeal stapling for the treatment of limb-length discrepancies on the mechanical axis of the stapled limb.
The goal of this study was to evaluate the mechanical axis in a group of patients who had undergone distal femoral, proximal tibial, or combined distal femoral and proximal tibial stapling for the treatment of limb-length discrepancy.
In this retrospective study, which was approved by our institutional review board, we analyzed the cases of patients in whom a limb-length discrepancy had been treated with physeal stapling of a lower extremity with use of the Blount technique between 1990 and 2005 at our institution. We initially screened all cases of physeal stapling for the treatment of limb-length discrepancy during the stated time period. For a patient to be included in the study, the stapling had to have been performed on a normal lower limb or one with overgrowth due to hemihypertrophy. This resulted in the identification of seventy-one patients. Fifty-five patients had adequate preoperative, postoperative, and final follow-up radiographs. Finally, one patient with cerebral palsy was eliminated because of a substantial change in the mechanical axis of the limb following a hamstring release. Ultimately, fifty-four patients (thirty boys and twenty-four girls) with adequate radiographs and a minimum of two years of follow-up were included. Fifteen patients underwent distal femoral stapling (group DF), eighteen patients had proximal tibial stapling (group PT), and twenty-one patients underwent stapling of both the distal femoral and the proximal tibial physis (the pan-genu, or PG group). In seven patients in the PG group, the femoral and tibial stapling procedures were done separately. Twenty-nine left lower extremities and twenty-five right lower extremities underwent stapling. The treated lower extremity was normal in 91% (forty-nine) of the fifty-four patients, and in the other 9% the limb was overgrown as a result of hemihypertrophy or Klippel-Trenaunay syndrome.
Common causes of the limb-length discrepancies in this series included neurologic disorders (eleven patients), post-axial hypoplasia (eight), trauma (five), congenital femoral deficiency (four), hemihypertrophy (four), Klippel-Trenaunay syndrome (two), hip dysplasia (two), and amniotic band syndrome (two). The mean initial limb-length discrepancy was 3.2 cm (range, 0.4 to 7.9 cm). Our indications for epiphysiodesis included a limb-length discrepancy of >2 cm and <5 cm and adequate remaining growth potential of the short limb. The patient with the 0.4-cm difference underwent stapling following trauma to the contralateral distal femoral physis to prevent an increasing limb-length discrepancy, and the patient with the 7.9-cm difference did not want a lengthening procedure. The mean age at the time of the initial stapling was 12.0 years (range, 7.9 to 15.1 years). It is standard practice at our institution to remove staples once a patient reaches skeletal maturity, although patients had staples removed prior to skeletal maturity when they were thought to have obtained clinical limb-length equalization before the physes had closed. The mean duration between the initial stapling procedure and staple removal was 2.8 years (range, 0.7 to 8.3 years).
Surgical Technique
A similar surgical technique was employed by a number of staff physicians between 1990 and 2005. The level of the incision was selected and was marked by placing a Kirschner wire over the physis and confirming the level with fluoroscopy. A 3 to 4-cm-long incision was then made along the Langer lines in an oblique fashion down to the perichondrium overlying the physis with caution taken to avoid the perichondrium. The physis was identified with fluoroscopy. The surgeons used 5/8-in (15.9-mm) cobalt-chromium (Zimaloy) Blount staples (Zimmer, Warsaw, Indiana) for the distal femoral physis and 3/8-in (9.5-mm) Zimaloy Blount staples for the proximal tibial physis. Four staples (two on each side of the knee) were placed so that they straddled the physis. Two staples, rather than the three used in Blount's original technique, were placed across each physis because the current Zimaloy staple is stronger than the original stainless-steel staple used when the procedure was originally described. The prongs of the staple were placed parallel to the physis and at an equal distance from the anterior and posterior margins of the bone. The staple position was verified with use of anteroposterior and lateral fluoroscopic views prior to wound irrigation and closure. No patient had stapling or epiphysiodesis of the proximal part of the fibula. A soft dressing or knee immobilizer was applied at the end of the procedure.
Postoperative limb lengths were evaluated clinically and radiographically and staples were removed at maturity or when the attending physician determined appropriate or maximal correction of limb length had been achieved.
Radiographic Evaluation
Measurements were made on preoperative and postoperative long standing anteroposterior lower-extremity radiographs. The radiographs were evaluated for limb length, the mechanical axis, the zone of the mechanical axis, the distance of the mechanical axis from the center of the tibial spines to define the mechanical axis deviation, the anatomic lateral distal femoral angle, and the medial proximal tibial angle. Placement of the staples was graded as adequate or inadequate depending on their location in relationship to the physis. The position was graded as adequate if the prongs of the staple captured and were parallel to the physis and were completely seated and remained seated until completion of the physeal arrest period. Inadequately placed staples were not parallel to the physis, did not capture the physis, or had dislodged from the starting position during the postoperative period.
Limb lengths were measured from the most superior portion of the femoral head to the center of the tibial plafond. The mechanical axis was defined as a line drawn between the center of the femoral head and the center of the tibial plafond. Zones of the mechanical axis have previously been described, and identification of these zones has been reported to be reproducible with negligible interobserver error11. We made one modification to previously described zones by considering the middle two quadrants as zone 1, with the more medial (varus) zones being designated as -2 and -3 and the more lateral (valgus) zones designated as +2 and +3 (Fig. 1). The zone of the mechanical axis at the level of the knee joint was recorded, as was the distance of the mechanical axis from the center of the tibial spines (the mechanical axis deviation).
We considered all changes in the mechanical axis of =1 cm as being clinically important. One centimeter was chosen as the threshold because it represents a large change in the position of the mechanical axis and it is greater than what would occur as a result of a simple measurement error from one radiograph to the next. A change of 1 cm did not result in a zone change if the preoperative mechanical axis began in the middle of zone 1. The average distance between the middle of zone 1 and the edge of zone 2 was 1.7 cm. We considered a zone change away from zone 1 as being more clinically relevant than just a 1-cm shift in the mechanical axis, as a zone change represents deviation from what is considered to be the normal range for the weight-bearing axis of the lower extremity.
The anatomic lateral distal femoral angle and the medial proximal tibial angle were previously described by Paley12 (Fig. 1). Measurement of the mechanical axis, anatomic lateral distal femoral angle, medial proximal tibial angle, and mechanical axis deviation has excellent intraobserver and interobserver reliability regardless of the experience of the observer13. In our study, two authors performed all measurements, and one author (S.D.S.) was responsible for the grading of staple placement as adequate or inadequate.
At the time of final follow-up, long standing anteroposterior radiographs of the lower extremities, made at what we deemed to be skeletal maturity or just before or after staple removal, were evaluated to determine the final limb lengths, mechanical axis deviation, zone of the mechanical axis, anatomic lateral distal femoral angle, and medial proximal tibial angle (Fig. 2). The mean age at staple removal was 14.1 years for girls and 15.4 years for boys. If it is assumed that girls mature at fourteen years of age and boys mature at sixteen years of age, 75% (eighteen) of the twenty-four girls and 60% (eighteen) of the thirty boys were mature at the time of staple removal and the final radiographic evaluation. Additionally, we reviewed radiographs made just prior to staple removal to check for closure of the stapled physis or physes as well as radiographs of the ipsilateral ankle to document skeletal maturity. This revealed a high correlation between radiographically judged skeletal maturity and chronologic maturity. As judged radiographically, 69% (thirty-seven) of the patients were skeletally mature at the time of staple removal, with 75% (eighteen) of the twenty-four female patients and 63% (nineteen) of the thirty male patients having complete closure of the stapled physis as well as the physis of the ipsilateral ankle. Thirteen percent of the patients (four female and three male) were very near skeletal maturity, with closure of the stapled physis but a partially open ipsilateral ankle physis. Nineteen percent of the patients (eight male and two female) were skeletally immature at the time of staple removal, with both an open stapled physis and an open ipsilateral ankle physis.
Medical records were analyzed for demographic data and complications such as wound infection and limitation of joint motion.
Statistical Analysis
Shifts in the mechanical axis and zone changes were analyzed statistically with use of analysis of variance, with p < 0.05 representing significance. When a significant difference among the groups was demonstrated by analysis of variance, a post hoc (least-significant-difference) test was performed to further compare the groups, with p < 0.05 again representing significance. Additionally, the magnitudes of the mechanical axis shifts and the zone change were evaluated with a univariate analysis of variance with both age and sex as factors and with p < 0.05 representing significance.
Source of Funding
There was no external source of funding for this study.
Mechanical Axis Shifts and Zone Changes
Preoperatively, the mechanical axis was within zone 1 in fifty-two (96%) of the fifty-four patients. At the end of treatment, it was within zone 1 in thirty-six (67%) of the fifty-four patients (Table I). The two patients in whom the preoperative mechanical axis was in zone +2 (valgus) had a varus shift in the axis to zone 1 at the end of treatment. One of these two patients had a shift in the mechanical axis of >1 cm.
Twenty-seven (50%) of the fifty-four patients exhibited a shift in the mechanical axis of =1 cm, and eighteen (33%) of the fifty-four had a clinically relevant zone change. In group DF (distal femoral stapling), three (20%) of the fifteen patients exhibited a shift of =1 cm, but none of these patients had a zone change. In group PT (proximal tibial stapling), ten (56%) of the eighteen patients had a shift of =1 cm, and eight (44%) of the eighteen had a clinically relevant zone change toward varus and away from the midline. In group PG (pan-genu stapling), fourteen (67%) of the twenty-one patients had a shift of =1 cm, and ten (48%) of the twenty-one had a clinically relevant zone change toward varus.
Of the twenty-seven patients who had a shift in the mechanical axis of =1 cm, twenty-four (89%) had a shift into varus. Only three patients (one in group DF and two in group PG) had a shift of =1 cm into valgus. These three valgus shifts in the mechanical axis were not accompanied by a zone change. Only one patient had a zone change that was not associated with a shift in the mechanical axis of =1 cm. This patient was in group PG and had a shift in the mechanical axis of -7 mm, with the axis moving from zone +2 into zone 1.
Groups PT and PG had significantly larger shifts in the mechanical axis than did group DF (Table II and Fig. 3). With the numbers studied, there was no significant difference in the magnitudes of the shifts in the mechanical axis between groups PT and PG. Groups PT and PG also had significantly larger zone changes than did group DF (Table II and Fig. 4). With the numbers studied, there was no significant difference in the magnitudes of the zone changes between groups PT and PG. Although the statistical tests involved comparisons of absolute shifts with neither varus nor valgus propensity taken into account, the great majority (89%) of the large shifts (=1 cm) in our patients were into varus. Of the thirty-nine patients in groups PT and PG, twenty-two (56%) had a =1-cm shift in the mechanical axis toward varus and eighteen (46%) had a clinically relevant zone change. All of the zone changes were to a more varus zone, with the axis changing by one zone in fourteen patients and by two zones in four.
Limb-Length Discrepancy
Ninety-one percent (forty-nine) of the patients had a decrease in the limb-length discrepancy, with a final mean discrepancy of 1.6 cm (range, -0.5 to 5.1 cm), compared with a mean preoperative discrepancy of 3.2 cm, in the entire group. Thirty-seven percent (twenty) of the patients had a final discrepancy of =1 cm, 37% (twenty) had a final discrepancy of between 1 and 2 cm, and 26% (fourteen) had a final discrepancy of >2 cm.
Anatomic Angles
Changes in the anatomic lateral distal femoral angle and the medial proximal tibial angle corresponded to the development of varus or valgus, indicating asymmetric growth in the stapled bones. This is best illustrated by the change in the anatomic angles in the patients who had a =1-cm shift in the mechanical axis toward varus in each of the stapling groups. The ten patients in group PT with a varus shift in the mechanical axis of =1 cm showed the greatest change in the medial proximal tibial angle. In those patients, the mean medial proximal tibial angle decreased from 88.9° preoperatively to 83.7° postoperatively, contributing substantially to genu varum. The mean anatomic lateral distal femoral angle did not contribute to varus in this group of patients, being 82.9° preoperatively and 80.3° postoperatively, and this counteracted the development of proximal tibial varus. In the twelve patients in group PG with a varus shift in the mechanical axis of =1 cm, both the distal part of the femur and the proximal part of the tibia contributed to the varus. The mean anatomic lateral distal femoral angle in these patients changed from 83.5° preoperatively to 86.2° postoperatively, and the mean medial proximal tibial angle changed from 89.3° to 87.4°.
Of the twenty-two patients in groups PT and PG who had a shift of =1 cm in the mechanical axis toward varus, seventeen (77%) had a corresponding change in the medial proximal tibial angle toward varus. These twenty-two patients had a mean preoperative medial proximal tibial angle of 89.1° and a mean postoperative medial proximal tibial angle of 85.7°. The anatomic lateral distal femoral angle exhibited little change, with a mean preoperative value of 83.2° and a mean postoperative value of 83.5°. The change in the anatomic lateral distal femoral angle was inconsistent (+4° and -4°) in the two patients in group DF who had a varus shift in the mechanical axis of =1 cm. The results of this analysis of the anatomic angles suggest that a stapled proximal part of the tibia is more susceptible to varus angulation than is a stapled distal part of the femur.
Staple Placement
In this study, a total of 300 staples had been placed, with sixty inserted in group DF, seventy-two used in group PT, and 168 used in group PG. Overall, 84% of the physes were adequately stapled. The proximal-lateral aspect of the tibial physis was the location for 86% of the staples that were considered to have been inadequately placed. Of these inadequately placed proximal-lateral tibial staples, 67% displaced from their initial starting position, 25% did not capture the physis at the time of insertion, and 8% were not parallel to the physis at the time of insertion.
Age
We investigated the relationship of age and sex to the magnitude of the mechanical axis shift and zone change by first examining our data on scatterplots. Thirteen patients underwent stapling before the age of eleven years, and nine of them were female. These thirteen patients had significantly larger changes in both the mechanical axis (p = 0.046) and the zone (p = 0.025) than did patients who were more than eleven years of age at the time of stapling. Neither sex nor the interaction of age and sex appeared to be significant factors affecting the magnitude of the mechanical axis shift or zone change.
Complications
Six (11%) of the fifty-four patients eventually underwent a high tibial valgus-producing osteotomy to correct excessive varus. Four of these patients were in group PG, accounting for 19% of the twenty-one patients in that group, and two were in group PT, accounting for 11% of the eighteen patients in that group. The mean shift in the mechanical axis in these six patients was —2.5 cm, and the mean zone change was -1.3, with the preoperative axis lying in zone 1 in all six patients. After the stapling treatment, the axis was in zone -2 in four patients and in zone -3 in two. The mean medial proximal tibial angle in these patients changed from 91.1° preoperatively to 82.3° after the stapling. The anatomic lateral distal femoral angle was minimally affected, changing from 84.3° preoperatively to 84.5° postoperatively. Two patients had adequately placed proximal-lateral tibial staples, while the other four patients had inadequately placed proximal-lateral tibial staples.
Other complications included persistent knee pain in one patient, irritation from prominent staples in nine, staples that backed out in four, a retained staple that could not be removed in one, and limb-length overcorrection in one. Of four patients with asymmetric backing out of staples, three had a varus shift of the limb and one had a valgus shift. None of the four patients with backing out of staples had a corrective osteotomy, but two required repeat stapling to treat a resultant angular deformity. There were two cases of a superficial suture abscess, which resolved with a short course of oral antibiotics. There were no cases of deep infection, and no patient had a permanent decrease in knee motion. Two patients with recurvatum as a complication of stapling were identified clinically, but an in-depth evaluation of sagittal plane deformity was not possible because of a lack of follow-up lateral radiographs.
The fifty-four patients in this study had a total of sixty-five stapling operations. Seven of the eleven second stapling procedures were not considered to be complications, as they were part of a two-stage operation in which the femoral and tibial stapling procedures were performed separately within the stapling period. Three of the additional procedures were done to rearrange staples to address angular deformity during the treatment period, and two of these deformities were thought to be due to the backing out of staples. The final additional surgery was for the removal of a completely dislodged staple. All but one patient eventually had the staples removed, resulting in a total of 118 surgical procedures.
Many authors have reported the results of staple epiphysiodesis for the treatment of limb-length discrepancy. Few have mentioned angular changes and, to our knowledge, none have used preoperative and postoperative long standing anteroposterior lower-extremity radiographs to assess changes in the mechanical axis among different groups of patients. Our primary objective in this study was to evaluate the frequency and severity of changes in the mechanical axis following stapling epiphysiodesis in order to better understand the angular deformities that may arise about the knee during the treatment of limb-length discrepancy.
Sengupta and Gupta, in a review of the results of physeal stapling (predominantly of the distal part of the femur), noted a residual limb-length discrepancy of <1 cm in 71% of their patients14. Complications included nine infections. These were controlled with antibiotics, although two patients required staple removal. A change in limb alignment, with clinical angular deformity, occurred in eight (1.6%) of 503 cases. These were treated with removal of staples from the concave side in three patients and an osteotomy in two. The authors did not state if the angular deformities were varus or valgus in nature, and long standing anteroposterior radiographs were not made.
Raab et al. reported that stapling had better results when used in an overgrown limb such as in Klippel-Trenaunay syndrome (a 71% rate of good-to-excellent results) but less encouraging results when employed to treat a limb-length discrepancy due to a short lower extremity15. Of twenty-four patients treated with stapling for limb-length discrepancy, four (17%) had a mild deviation in the mechanical axis (4° to 9°), but none underwent additional surgery to treat the residual angular deformity.
Brockway et al. reported that genu varum developed in four of forty-two patients who had undergone epiphysiodesis, but genu valgum did not develop in any patient16. Three of these four patients had undergone both distal femoral and proximal tibial stapling, and the other patient had had tibial stapling only. Neither genu varum nor genu valgum developed in any of the thirty-two patients who had undergone isolated femoral stapling. These clinical results are similar to our radiographic findings. In the study by Brockway et al., three of the four patients with genu varum went on to have a corrective osteotomy. The authors noted that "the greater flare of the lateral tuberosity of the tibia as compared with that of the medial tuberosity" contributed to the development of genu varum as the lateral staples were not as well embedded in the tibia. Our results, which showed that the proximal-lateral aspect of the tibia was the most likely location for inadequate placement of staples, concur with this conclusion.
Blount expanded on the findings of Brockway et al. in his 1971 paper regarding physeal stapling and noted that stapling of the tibia was technically more challenging than femoral stapling because of the tibia's conical shape17. Blount stated that "tubulation" of the proximal end of the tibia tends to uncover the distal prong of the staple over time, emphasizing that, with faulty insertion, the staple could bend and extrude as the bone molded away from it, leading to angulation. He stressed the importance of having the cross member of the staple perpendicular to the growth plate and the prongs parallel to the physis. Blount commented that, in his series of 426 stapling operations in 185 patients, there were only two deformities that required a corrective osteotomy; lesser deformities were corrected by rearranging staples.
It may be that changes in the mechanical axis associated with physeal stapling for the treatment of limb-length discrepancy were less often recognized in earlier studies because orthoroentgenograms or scanograms were the imaging studies most commonly used to assess limb-length discrepancy. Since scanograms focus only on the hips, knees, and ankles, an accurate assessment of the mechanical axis cannot be made, so clinically relevant shifts in the axis could be overlooked. A teleoroentgenogram, or full-length standing anteroposterior lower-extremity radiograph, is a superior study, enabling evaluation of both limb-length discrepancy and angular deformities18,19.
In our study, substantial deviations (of =1 cm) in the mechanical axis toward varus were common, occurring in 56% of the patients who had undergone physeal stapling of the proximal part of the tibia in isolation or in combination with distal femoral stapling. We attempted to identify the mechanism of varus shifts with staple epiphysiodesis, and it appears that inadequate staple placement plays a role. However, inadequate staple placement does not account for all of the varus shifts in the mechanical axis or zone changes observed in this study. As Blount suggested, the so-called tubulation of the proximal-lateral aspect of the tibia and loss of fixation of the distal prong of the staple can lead to unequal growth modulation and angular deformity. However, since substantial backing out was noted in only four of our fifty-four patients, this cannot be suggested as the predominant cause of varus in our patients.
Anatomically, the tibia is more triangular proximally and narrows quickly below the physis. As a child reaches maturity, the slope of the medial and lateral edges of the physis increases. This "capping" of the growth plate, similar to that seen in the distal part of the radius, makes it more difficult to place staples parallel to the physis. The growth plate forces in the tibia might also be different from those in the distal part of the femur.
The head of the fibula makes placement of tibial staples difficult. The patients in our study did not undergo stapling or epiphysiodesis of the proximal part of the fibula. It is possible that the continued growth of the fibula posterior and lateral to the knee may encourage the development of varus and recurvatum about the knee. We could not adequately analyze fibular growth in this study because of a lack of controlled radiographic data.
It appears that the development of a change in the mechanical axis is centered about the proximal part of the tibia. The triangular/tubular shape and the asymmetric geometry of the proximal part of the tibia, the staple backing away from the lateral-distal aspect of the tibial physis with time, the abnormal strain applied and the resultant forces that occur when the femur is stapled in conjunction with the tibia, and the growth of the proximal part of the fibula may all play a role in the development of post-stapling varus in the tibia.
Limb-length discrepancies have a number of etiologies, and the clinical problem is typically a short limb rather than an overgrown limb. Thus, when the treatment of choice is epiphysiodesis, either by stapling or other methods, it is usually the unaffected, normal lower extremity that undergoes surgery. It is imperative to counsel the patient and the parents regarding the potential risks and complications associated with stapling. We think that it is important to discuss the potential for a mechanical axis deviation to develop and its possible ramifications.
The large number of patients in this study who had a change in the mechanical axis of =1 cm emphasizes the importance of monitoring these patients at regular intervals (every three to four months). Follow-up visits should routinely include a careful analysis of both limb-length discrepancy and angular alignment. We recommend a clinical assessment with a block test and then obtaining a long standing anteroposterior lower-extremity radiograph with the patient placing the foot of the shorter limb on a lift that matches the clinical discrepancy in order to evaluate limb length and angular alignment. From our data, it appears that addressing a limb-length discrepancy with staple epiphysiodesis is safer the closer the child is to maturity. Younger children (less than eleven years of age) require even closer monitoring as they have a greater propensity for larger changes in the mechanical axis.
Other techniques used to address limb-length discrepancy have been reported to have fewer complications than the stapling procedure carried out in this study. Classic open epiphysiodesis has been reported to have overall complication rates ranging from 2.5% to 8.1% in some series20,21. In one large study of open epiphysiodesis, four (2.3%) of 173 patients had a resultant angular deformity requiring osteotomy21. Open epiphysiodesis is also associated with a low reoperation rate, ranging from 0% to 11%, and the final limb-length discrepancy is <1.6 cm in the majority of cases21-23.
Percutaneous epiphysiodesis is a minimally invasive option for treating limb-length discrepancy4,24. Overall, complication rates range from 2.8% to 17%, but some of the complications are very minor25,26. Angular deformities are rarely reported, with only one case identified among 172 reviewed, and reoperations are very infrequent, with a rate of 1.7% (three) of 172 cases reviewed20,22,25-27. Like open epiphysiodesis, percutaneous epiphysiodesis has the advantage of not requiring an additional procedure for hardware removal. Additionally, the final mean limb-length discrepancy was 1.5 cm in one study, and 91% of patients followed to maturity had a final limb-length discrepancy of =1 cm in another study25,26.
Percutaneous epiphysiodesis with use of transphyseal screws has gained popularity in the last decade and has had promising results. In the published literature, overall complication rates range from 3% to 27%5,28,29. Of seventy-two patients who underwent percutaneous epiphysiodesis with use of transphyseal screws for treatment of limb-length discrepancy5,28,29, three were reported to have a varus deformity; recurvatum developed in a fourth patient. The reoperation rates, which ranged from 4.9% to 57%, were higher than those following open and percutaneous epiphysiodesis, as many patients underwent removal of hardware either at the physician's discretion or to prevent overcorrection. The average final limb-length discrepancy ranged from 0.51 to 1.36 cm.
In our study, six (11%) of the patients who underwent stapling had a subsequent osteotomy to treat angular deformity, four patients required an additional stapling procedure to address complications during the treatment period, and all but one patient eventually had an additional procedure for staple removal. Although the final mean limb-length discrepancy was 1.6 cm, 26% of our patients still had a discrepancy in excess of 2 cm. Stapling of the proximal part of the tibia, especially when combined with stapling of the distal part of the femur, with use of the methods described in this study is a questionable treatment for limb-length inequality. Our results suggest that, if one still decides to choose stapling as the treatment method, only the distal part of the femur should be stapled, as it appears to be less susceptible to shifts in the mechanical axis.
There are notable limitations to this study. First, it is retrospective in nature. Second, it is predominantly a radiographic study, and functional outcomes were not evaluated. Thus, we do not know the amount of shift in the mechanical axis that causes a clinical problem. However, shifts in the mechanical axis that are large enough to require a corrective osteotomy are definitely clinically relevant. Additionally, we did not have standardized lateral radiographs, and thus we cannot comment on the frequency of sagittal plane deformities, such as recurvatum. The major strength of this study was the use of long standing anteroposterior radiographs to follow limb-length discrepancies in a large group of patients who had undergone staple epiphysiodesis, as this allowed an evaluation of the mechanical axis and angular deformity about the knee.
We identified a serious complication—namely, that stapling of the proximal part of the tibia is associated with a substantial change in the mechanical axis toward varus in approximately half of the patients. Limbs treated with distal femoral stapling are less susceptible to changes in the mechanical axis, but the patient still must be monitored regularly. The magnitude of the shift in the mechanical axis that is well tolerated is not known. However, zone changes are not acceptable in our practice, and six of our patients who had such a change subsequently underwent an osteotomy to correct a coronal plane deformity. Physeal stapling for the treatment of limb-length discrepancy should not be considered a benign procedure, and follow-up with full-length standing anteroposterior radiographs should be carried out on a regular basis (every three to four months) to identify substantial deviations in the mechanical axis before a corrective osteotomy is the only option left. Our data suggest that distal femoral staple epiphysiodesis is a safe way to address limb-length discrepancy. However, one should give careful consideration to other treatment options such as percutaneous epiphysiodesis with or without screws, as these procedures are minimally invasive, are associated with fewer complications, and can lead to good correction of limb-length discrepancy. Tibial stapling, even if done well technically, can lead to tibial varus. We advise caution when utilizing proximal tibial staple epiphysiodesis for limb-length discrepancy, either as a solitary procedure or when combined with distal femoral staple epiphysiodesis. 
Blount WP, Clarke GR. Control of bone growth by epiphyseal stapling. A preliminary report. J Bone Joint Surg Am.1949;31:464-78.31464
1949
Phemister DB. Operative arrestment of the longitudinal growth of bones in the treatment of deformities. J Bone Joint Surg.1933;15:1-15.151
1933
White JW. Femoral shortening for equalization of leg length. J Bone Joint Surg Am.1935;17:597-604.17597
1935
Ogilvie JW. Epiphysiodesis: evaluation of a new technique. J Pediatr Orthop.1986;6:147-9.6147
1986
[PubMed][CrossRef]
Métaizeau JP, Wong-Chung J, Bertrand H, Pasquier P. Percutaneous epiphysiodesis using transphyseal screws (PETS). J Pediatr Orthop.1998;18:363-9.18363
1998
[CrossRef]
Guille JT, Yamazaki A, Bowen JR. Physeal surgery: indications and operative treatment. Am J Orthop.1997;26:323-32.26323
1997
Frantz CH. Epiphyseal stapling: a comprehensive review. Clin Orthop Relat Res.1971;77:149-57.77149
1971
May VR Jr, Clements EL. Epiphyseal stapling: with special reference to complications. South Med J.1965;58:1203-7.581203
1965
[CrossRef]
Trias A, Mueller K, Ray RD. Epiphysial stapling. Surg Gynecol Obstet.1961;113:315-23.113315
1961
Poirier H. Epiphysial stapling and leg equalisation. J Bone Joint Surg Br.1968;50:61-9.5061
1968
Stevens PM, Maguire M, Dales MD, Robins AJ. Physeal stapling for idiopathic genu valgum. J Pediatr Orthop.1999;19:645-9.19645
1999
[CrossRef]
Paley D. Principles of deformity correction. New York: Springer; 2002.
2002
Gordon JE, Chen RC, Dobbs MB, Luhmann SJ, Rich MM, Schoenecker PL. Interobserver and intraobserver reliability in the evaluation of mechanical axis deviation. J Pediatr Orthop.2009;29:281-4.29281
2009
[CrossRef]
Sengupta A, Gupta P. Epiphyseal stapling for leg equalization in developing countries. Int Orthop.1993;17:37-42.1737
1993
Raab P, Wild A, Seller K, Krauspe R. Correction of length discrepancies and angular deformities of the leg by Blount's epiphyseal stapling. Eur J Pediatr.2001;160:668-74.160668
2001
Brockway A, Craig WA, Cockreli BR Jr. End-result study of sixty-two stapling operations. J Bone Joint Surg Am.1954;36:1063-70; passim.361063
1954
Blount WP. A mature look at epiphyseal stapling. Clin Orthop Relat Res.1971;77:158-63.77158
1971
Machen MS, Stevens PM. Should full-length standing anteroposterior radiographs replace the scanogram for measurement of limb length discrepancy? J Pediatr Orthop B.2005;14:30-7.1430
2005
[CrossRef]
Sabharwal S, Zhao C, McKeon JJ, McClemens E, Edgar M, Behrens F. Computed radiographic measurement of limb-length discrepancy. Full-length standing anteroposterior radiograph compared with scanogram. J Bone Joint Surg Am.2006;88:2243-51.882243
2006
[CrossRef]
Surdam JW, Morris CD, DeWeese JD, Drvaric DM. Leg length inequality and epiphysiodesis: review of 96 cases. J Pediatr Orthop.2003;23:381-4.23381
2003
[CrossRef]
Green WT, Anderson M. Epiphyseal arrest for the correction of discrepancies in length of the lower extremities. J Bone Joint Surg Am.1957;39:853-72; discussion, 872; passim.39853
1957
Liotta FJ, Ambrose TA 2nd, Eilert RE. Fluoroscopic technique versus Phemister technique for epiphysiodesis. J Pediatr Orthop.1992;12:248-51.12248
1992
[CrossRef]
Stephens DC, Herrick W, MacEwen GD. Epiphysiodesis for limb length inequality: results and indications. Clin Orthop Relat Res.1978;136:41-8.13641
1978
Bowen JR, Johnson WJ. Percutaneous epiphysiodesis. Clin Orthop Relat Res.1984;190:170-3.190170
1984
Timperlake RW, Bowen JR, Guille JT, Choi IH. Prospective evaluation of fifty-three consecutive percutaneous epiphysiodeses of the distal femur and proximal tibia and fibula. J Pediatr Orthop.1991;11:350-7.11350
1991
Gabriel KR, Crawford AH, Roy DR, True MS, Sauntry S. Percutaneous epiphyseodesis. J Pediatr Orthop.1994;14:358-62.14358
1994
[CrossRef]
Horton GA, Olney BW. Epiphysiodesis of the lower extremity: results of the percutaneous technique. J Pediatr Orthop.1996;16:180-2.16180
1996
[CrossRef]
Khoury JG, Tavares JO, McConnell S, Zeiders G, Sanders JO. Results of screw epiphysiodesis for the treatment of limb length discrepancy and angular deformity. J Pediatr Orthop.2007;27:623-8.27623
2007
[CrossRef]
Nouth F, Kuo LA. Percutaneous epiphysiodesis using transphyseal screws (PETS): prospective case study and review. J Pediatr Orthop.2004;24:721-5.24721
2004
[CrossRef]