The Natural History of ACL Injury

The natural history of ACL tears in pediatric patients is not fully understood. However, evaluating the results of nonoperative treatment of ACL tears leads to some understanding of it. Physicians often favor nonoperative treatment for ACL injury because of the risks involved in operating on a skeletally immature patient. Nonoperative treatment may include activity modification, bracing, and rehabilitation. However, studies have shown that these nonoperative approaches have poor efficacy,^‡ predominantly because pediatric patients are noncompliant, especially with modification of sports activity, and patients may be injured during free play. Noncompliance often leads to recurrent and sports-related instability and damage to the menisci.

Evidence suggests that the efficacy of nonoperative treatment is related to the severity of the ACL tear. Kannus and Jarvinen²³ treated 32 patients nonoperatively with grade II (partial) and grade III (complete) ACL tears. At 8-year (on average) follow-up, 25 patients in this series with grade II tears had good to excellent outcomes. Seven patients with complete grade III tears had poor outcomes that included chronic instability and posttraumatic arthritis. These outcomes led the authors to reject nonoperative treatment of grade III ACL tears in pediatric patients. Angel and Hall also reported poor outcomes, including pain and limited activity, with nonoperative treatment of 27 pediatric patients with grade III ACL tears.⁵ Most children younger than 14 years of age (92%) had functional knee disability at follow-up evaluation. Graf and associates found new meniscal tears after 15 months of nonoperative treatment in 87.5% of pediatric patients.¹⁵ In a study of 38 adolescent patients, McCarroll and coworkers reported that 97% of patients experienced episodes of instability, and 71% had symptomatic meniscal tears.³¹ Mizuta and colleagues found degenerative changes in 61% (11/18) of patients within 51 treatment months³⁵ and concluded that these outcomes were unacceptable. A study by Millet and associates found significantly increased meniscal injuries in chronic cases.³⁴ Finally, T. J. Ganley, M.D. (personal communication, July 2009) evaluated 70 pediatric patients with ACL tears to determine independent risk factors for and relative risk of meniscal and chondral injuries. Despite activity modification, bracing, and rehabilitation, patients who delayed surgical reconstruction for longer than 12 weeks had a fourfold increase in irreparable medial meniscal tears, an 11-fold increase in lateral compartment chondral injuries, and a threefold increase in patella-trochlear injuries. A single episode of instability was associated with an 11-fold increase in irreparable medial meniscal tears. This large body of evidence strongly suggests that nonoperative treatment of an ACL injury in children carries a high probability of long-term knee disability.

Although operative treatment has serious risks, these risks can be mitigated by careful evaluation of skeletal and sexual maturity of the patient and by selection of the appropriate surgical technique based on these presurgical evaluations.

Skeletal Maturity

Although chronological age is a good indicator of mean skeletal maturity in large populations, an individual child can vary widely from the mean. The skeletal age of the pediatric patient is a key factor in determining appropriate treatment for an ACL tear. By estimating the skeletal age of the patient, the physician can gauge the potential risks and consequences of iatrogenic injury to the physis. As a rule, the younger the skeletal age (i.e., the more growth remains in the distal femur and proximal tibial physes), the greater the risk of severe treatment-related growth disturbances.

Skeletal age is determined with radiographs. The most common method of determining skeletal age is to compare an anteroposterior radiograph of the patient’s left hand and wrist with the age-specific radiograph in the Greulich and Pyle atlas.¹⁶

Although skeletal age is essential to determining the relative risk of ACL reconstruction, the physiologic age of the patient is also important and should be considered when the treatment planning. Physiologic age can be determined using the Tanner staging of sexual maturation.⁴⁵ Tanner stages can determine whether the child is prepubescent (stages I and II), pubescent (stage III), or postpubescent (stages IV and V) through the presence or absence of secondary sexual characteristics (i.e., pubic and axillary hair and development of breasts and genitalia) (Table 88-1).

Table 88-1 Tanner Stages of Development

Preliminary staging should be assigned before surgery by asking the patient about the onset of menarche or the growth of axillary hair. To spare the child the trauma of genital examination, a thorough examination should be completed after the child is under anesthesia, but before surgery, for precise determination of the Tanner stage.

Growth and Development

The most rapidly growing physes in the body are located on the distal femur and the proximal tibia. The distal femoral physis contributes about 40% of the overall lower extremity length, and the proximal tibial physis contributes about 27%.³ The distal femur grows at the annual rate of 1.3 cm, but slows in the last 2 years of growth to an annual rate of 0.65 cm.⁴⁰ In boys, the mean peak height velocity occurs at age 13.5 years, with a range from 13 to 15 years of age. Peak height velocity in boys usually occurs at Tanner stage IV. However, about 20% of boys do not reach peak height velocity before Tanner stage V. Girls reach peak height velocity earlier than boys. The mean age for girls is age 11.5 years, with a range from 11 to 13 years of age. Onset of menarche typically occurs 1 year after peak height velocity is reached.

The severity of iatrogenic growth disturbance can be predicted by the skeletal maturity of the patient at the time that injury occurred. A 3-cm discrepancy in leg length—nearly three times normal variance—is estimated to occur from complete closure of the proximal tibial physis in an average 12-year-old boy, complete closure of the distal femoral physes in a 13-year-old boy, or complete closure of the femoral and tibial physes in a 14-year-old boy.

Although leg length discrepancy is an undesirable result of surgery, angular deformity is the more serious surgical complication. A valgus/flexion deformity of the distal femur can be caused by an over-the-top femoral groove if the perichondral ring of LaCroix is damaged, and recurvatum of the knee can occur if the anterior tibial physis is damaged. Webster and colleagues estimated that partial tibial physeal arrest in a 14-year-old boy with 2 cm of growth remaining in the distal femur could result in a 14-degree valgus deformity with a lateral femoral epiphysiodesis, or 11-degree recurvatum with a partial tibial physeal arrest.⁴⁶

Each of the studies reviewed in the following section illustrates the potential consequences of iatrogenic injury to the physis during surgical treatment. Patients at greatest risk are prepubescent (Tanner stages I and II), followed by pubescent patients (Tanner stage III). Patients at least risk are those nearing and those who have reached sexual maturity (Tanner stages IV and V).

Basic Research on Physeal Injury

Although there is a dearth of basic research on physeal injury in pediatric patients, several animal studies have evaluated the consequences of drill hole damage to the physis and of insertion of a soft tissue graft through a transphyseal hole. In 1988, Makela and coworkers studied the effects of 2.0- and 3.2-mm transphyseal femoral drill holes in rabbits.²⁹ The cross-sectional area of the physis destroyed was 3% for the 2-mm drill hole and 7% for the 3.2-mm drill hole. Results showed that 7% cross-sectional destruction of the physis resulted in permanent disruption of growth.

Guzzanti and colleagues evaluated the effects of placing a soft tissue graft across the physis in immature rabbits.¹⁷ ACL reconstruction was performed with the semitendinosus tendon using 2-mm transphyseal femoral and tibial holes. Drill hole damage to the femoral physis was seen in 11% of the transverse diameter and 3% of the cross-sectional diameter. The extent of damage to the tibial physis was 12% of the transverse diameter and 4% of the cross-sectional area. A valgus deformity developed in about 9% (2/21) of tibiae, and one incident of tibial growth disruption was noted. Based on these data, the authors recommended extreme caution when transphyseal reconstruction is considered in pediatric patients.

Transphyseal ACL reconstruction in a rabbit model using four tunnel diameters ranging from 1.95 to 3.97 mm was conducted by Houle and associates.²⁰ Larger drill hole size was associated with increased and substantial deformity, and physeal arrest occurred despite the soft tissue graft. This study suggests that no more than 1% of the physis should be disrupted in children during an ACL reconstruction. In a rabbit model, Babb and coworkers evaluated the potential for growth arrest in three groups.⁷ Group 1 was the control group; tunnels were drilled in the femur and tibia and were left open. In group 2, the tunnels were filled with a soft tissue autograft, and in group 3, the autograft was seeded with mesenchymal stem cells. Angular deformity and growth arrest were prevented only in group 3.

In contrast to these three studies, which found that soft tissue provided no protection, the following two studies demonstrated that a soft tissue graft across the physis prevents growth disturbance. In rabbit femurs, drill holes of 1.7, 2.5, and 3.4 mm were evaluated, where one hole was left empty and the contralateral one was filled with an autograft of soft tissue.²² Growth was retarded when 7% to 9% of the distal femoral physis was destroyed, but not when 4% to 5% of the cross-sectional area of the physis was destroyed. Bone cylinders were observed around the soft tissue grafts, but solid bone bridging did not occur. Prevention of bony bridge development followed by growth disturbance was also found in a canine model subsequent to a soft tissue graft placement in transphyseal drill holes.⁴³

Other researchers have evaluated the effects of graft tension. Edwards studied the effect of tensioning a graft across open physes in a canine model at 80 N.¹³ This technique resulted in the development of valgus femoral and varus tibial deformities without radiographic or histologic evidence of physeal bar formation. Chudik and colleagues also tensioned autografts at 80 N using transepiphyseal, transphyseal, and over-the-top femoral positions.¹⁰ They found growth disturbances with each technique. However, the transepiphyseal technique was more anatomic and caused less growth disturbance. These results are predicted by the Hueter-Volkman principle, that is, when compressive force is applied perpendicular to the physes, longitudinal growth is inhibited. This suggests that even physeal-sparing procedures pose a risk for ACL reconstruction in pediatric patients.

Causes of Iatrogenic Growth Disturbance

Decisions about the surgical technique used in ACL reconstruction of a skeletally immature knee should be made after the potential for growth disturbance is weighed. Basic research, although incomplete and not entirely generalizable to humans, provides some evidence to assess the risk factors. Studies by Guzzanti and coworkers¹⁷ and by Houle and associates²⁰ found that the risk for arrested growth is greater in the proximal tibial physis than in the femoral physis.

The risk of growth disturbance is generally associated with the extent of damage to the cross-sectional area of the physis. It is not completely understood, in animal models or in children, which drill hole size and orientation can be used without risk of disturbing growth. In animal models, the threshold for drill size growth disturbance appears to be between 1% and 7% of the cross-sectional area of the physis.^{7,17,20,22,29} Damage to the cross-sectional area of the physis can be diminished by making drill holes perpendicular rather than oblique to the surface of the physis. Although study results are not uniform, soft tissue grafts placed across the physis are probably protective against bone bridging and arrested growth. The physes are sensitive to compression forces,¹³ so excessive ACL graft tension should be avoided.

Rare complications, such as angular deformity and significant leg length discrepancies, have been reported in children who underwent ACL reconstruction.^24,26,27 Kocher and colleagues found 15 cases of growth disturbance in a survey of 140 physicians.²⁴ Lipscomb and Anderson²⁷ had one case of valgus deformity following ACL reconstruction in skeletally immature patients. One patient’s deformity followed placement of a staple across the lateral femoral physis. The authors evaluated another patient in consultation—a 12-year-old boy who had undergone transphyseal ACL reconstruction with an Achilles tendon allograft. At the 6-month follow-up, the graft had failed, resulting in a 3-degree valgus alignment of the nonoperative knee and a 7-degree valgus alignment of the ACL-reconstructed knee without physeal arrest.

Treatment Options

Reports of iatrogenic growth disturbance following intra-articular transphyseal replacement in both basic research and case studies illustrate the risks of using adult ACL reconstruction techniques on pediatric patients. Delaying surgery with nonoperative treatment in immature patients has some appeal and benefits. A more physically and psychologically mature patient is typically more compliant with postoperative rehabilitation. Additionally, delaying surgery until the patient reaches skeletal maturity allows the physician to use traditional surgical procedures. However, as discussed earlier, nonoperative treatments usually have undesirable outcomes.^4,14,26,32

Some operative approaches in pediatric patients, such as primary repair^11,13 and extra-articular replacements,^15,31 have also resulted in poor outcomes. It is possible to minimize the risk of physeal injury using a modified physeal-sparing intra-articular replacement.³⁶ Parker and coworkers reconstructed an ACL by passing hamstring tendons through a groove in the anterior aspect of the tibia and over the top of the lateral femoral condyle.³⁸ In 44 Tanner stage I or II patients, Kocher and associates used a combined intra-articular/extra-articular ACL reconstruction technique.^25,33 This technique places the iliotibial band around the lateral femoral condyle extra-articularly and passes it through the intercondylar notch. It is then sutured to the periosteum of the proximal tibia. In 42 of the 44 patients, a mean International Knee Documentation Committee (IKDC) subjective score of 96.7 was reported. Mean growth from surgery to follow-up was 21 cm. Lachman examinations were normal in 23 patients, nearly normal in 18, and abnormal in one patient. Pivot-shift test results were normal in 31 patients. Functional outcomes were excellent, and growth disturbance was minimal. Two patients had graft failure and subsequent reconstruction. Transphyseal tibial holes and over-the-top femoral positions with autografts^8,28 and allografts⁴ have also been used.

In Tanner stage I patients, Guzzanti and colleagues suggested reconstructing the ACL using single-stranded semitendinosus and gracilis tendon grafts with a transepiphyseal tibial hole and an over-the-top femoral position.¹⁸ No growth disturbances have been reported with over-the-top procedures, but lack of isometry can be an issue with this technique. The femoral over-the-top position has resulted in a mean graft elongation of 10 mm as the knee approaches full extension.³⁷ Avoid rasping with the over-the-top femoral position, as this may damage the perichondral ring of LaCroix.

Controversy continues over ACL replacement procedures that use intra-articular transphyseal graft placement, because of deficiencies in basic science and clinical literature. Clinical studies that demonstrate the safety of transphyseal replacements have included postmenarchal girls and postpubescent boys with physes near closure.^4,6,12,30,31 Intra-articular replacements were performed by Pressman and coworkers in a series of 18 patients, 7 with open physes and 11 with closed or nearly closed physes.³⁹ Other surgeons have performed intra-articular ACL replacements as well, but patients in these cohorts had only 2.3 cm to 4.5 cm of postoperative growth.^4,32 In other case series, average patient age was greater than 14 years at the time of surgery, so the risks of angular deformity and leg length discrepancy were low compared with those in younger children.^6,30,42

Children in Tanner I and II stages are at greatest risk for growth disturbance as a consequence of ACL surgery. Few patients in the early Tanner stages have participated in studies of transphyseal procedures; therefore, the safety of these procedures in preadolescent patients is not documented in the clinical literature. Further, basic research has not proven the safety of drilling across the physis, or of placing a soft tissue graft across the physis.

In an effort to minimize physeal trauma in Tanner stage II and III patients, Guzzanti and associates used a semitendinosus graft passed through 6-mm or smaller transphyseal femoral holes and transepiphyseal tibial holes.¹⁹

Anderson performed transepiphyseal replacement in a series of 12 patients (Tanner stage I, n = 3; Tanner stage II, n = 4; Tanner stage III, n = 5) using a modified adult ACL reconstruction procedure that did not transgress the physes of the tibia or the femur.² At the 4-year follow-up, mean growth from surgery was 16.5 cm. No clinically significant differences were noted in lower leg lengths, as determined by long leg radiographs. The mean IKDC Subjective Knee Form score was 96.5. The ligament laxity testing performed using a KT-1000 arthrometer showed a mean side-to-side difference of 1.5 mm at 134 N. According to the criteria of the Objective 2001 IKDC Knee Form,²¹ the rating was normal for seven patients and nearly normal for the remaining five. At 4 years post surgery, one patient, who rated 100 on the IKDC Subjective Score at follow-up year 2, ruptured his ACL graft during a sporting event.

This technique was subsequently performed on an additional 26 patients (Tanner stage I, n = 6; Tanner stage II, n = 7; Tanner stage III, n = 13). Two patients re-ruptured their ACL grafts. One was in a motorcycle accident 8 weeks post surgery, sustaining a grade III injury to the ACL graft, as well as an injury to the medial collateral ligament. In the other patient, the graft failed, and no history of trauma was reported. The only other complication was seen in a patient who had a break in the EndoButton continuous loop 1 year after surgery. This patient had an excellent recovery after removal of the washer without residual pathologic laxity.