Rationale

Bowing of one or both legs in children aged 2–4 years is a common concern that prompts parents to seek orthopedic care. In 1975, Salenius and Vannka studied the natural history of the tibiofemoral angle in 979 children admitted to hospital for various reasons.⁷ They noted that the tibiofemoral angle in newborn children was pronounced varus that slowly corrected until 18–24 months, when the angle changed to a valgus position between ages two and three years. In a few patients the varus alignment of the knees does not correct, but increases or persists beyond the third year of life and sometimes requires surgical intervention.⁸ Below the age of two years, it can be difficult to differentiate between patients with physiologic varus alignment, which is expected to correct spontaneously, from those who will develop infantile Blount disease, a radiographic diagnosis based on specific changes to the medial proximal tibia characterized by Langenskiold stages, which are expected to develop a progressive deformity and require treatment. Prior to the onset of these radiographic changes, certain physical and other radiographic criteria may be helpful in predicting those patients who can be safely observed versus those who are likely to develop true Blount disease who might warrant early intervention.

Clinical comment

Having noted improved patient outcomes (less chance of recurrent deformity) with surgical correction prior to age four years, some authors have emphasized early surgical realignment of children with infantile Blount disease.⁸–¹⁰ This has typically been a valgus producing osteotomy to unload the medial physis to give it the best chance to recover. However, an osteotomy is an invasive surgical procedure with substantial risks such as neurovascular injury and compartment syndrome. The efficacy of bracing for Blount disease has been questioned by some authors and is discussed in greater detail in the next section.¹¹,¹² It is still not well established as to which radiographic and physical characteristics in these young children are associated with a progressive deformity and those which will spontaneously correct. Understanding these characteristics can help determine which patients warrant early intervention versus continued observation and may help avoid unnecessary surgery in patients whose deformity could correct spontaneously.

Available literature and quality of the evidence

• Level III: 5 retrospective cohort studies and 1 prognostic study.
  
• Level IV: 3 case series.

Findings

The radiographic analysis of a patient with bowlegs should begin with a standing, full‐length anteroposterior radiograph of both lower extremities with the patellae facing forward to help assess whether the deformity is primarily in the proximal tibia or if there is contribution from the distal femur as well.⁸,¹³,¹⁴ Historically, the severity of a bowleg deformity was determined by comparing the anatomic tibiofemoral angle of affected patients with the normal angle for their age,¹ and examining for the classic radiographic changes of infantile Blount disease as described by Langenskiold.¹⁵ In 1982, Levine and Drennan measured an angle between the transverse plane of the proximal tibial metaphysis and a line perpendicular to the long axis of the tibial shaft, named the metaphyseal diaphyseal angle (MDA), to determine its usefulness in differentiating physiologic bowing from infantile Blount disease.² They compared the MDA patients with infantile Blount disease to those with physiologic bowing and found that only 3 of 58 extremities with an MDA <11° developed progressive deformity compared to 29 of 30 with MDA >11° (p <0.001).² Feldman and Schoenecker further analyzed the MDA and found that an MDA ≥11° resulted in a false‐positive rate of 33% and an MDA of <11° had a false‐negative rate of 9%.³ They determined that an MDA ≤9° had a false‐negative of <5% and an MDA ≥16° had a false‐positive rate of <5%. They recommended observation for those with MDA ≤9° and early intervention for those with MDA ≥16°. Patients with an MDA between 10° and 15° are in a gray area which requires close observation for progression.

Another study done in 1982 by O’Neill and MacEwen measured both the FMDA and tibial metaphyseal diaphyseal angle (TMDA),⁴ and determined that patients whose TMDA was greater than their FMDA were at increased risk for disease progression. McCarthy et al. compared the FMDA to TMDA ratio, or femoral/tibial ratio (FTR), to the tibial MDA in determining which patients would develop progressive deformity.⁵ The authors found that a TMDA >13° or an FTR <1° was prognostic for developing infantile Blount disease; however, the false‐negative and false‐positive rates were lower for the FTR. They also found that the FTR was affected less by rotation of the x‐ray than the MDA. The authors concluded that the FTR was more accurate than the MDA in detecting which deformities would progress. Similarly, in 2002, Bowen et al. examined 98 patients with bowlegs and calculated the total limb varus (LV) by measuring the mechanical axes of the femur and tibia as well as the femoral varus (FV) and tibial varus (TV) by measuring a horizontal line through the joint line with the femoral and tibial mechanical axes, respectively.⁶ They then calculated the percent deformity in the tibia (%DT) by dividing TV by LV. They found that a %DT >50% had a 100% sensitivity and 96% specificity as a predictor of future progression, compared to 64% sensitivity and 93% specificity for MDA ≥16°. They concluded that the %DT was more accurate than the MDA in predicting future progression but emphasized that the only true way to know was to follow a patient with serial radiographs until a trend toward progression or resolution became clear.

In 2000, Mukai et al. used MRI to try to differentiate physiologic bowing from infantile Blount disease.¹⁶ They found that all patients with bowlegs had a high‐intensity area in the medial epiphyseal cartilage on T2 weighted imaging compared to normal tibiae, but that certain patients also exhibited an abnormality in metaphysis, and hypothesized that these patients were at higher risk to progress. At final follow‐up, 5 of 11 patients with metaphyseal signal changes went on to develop the characteristic findings of infantile Blount such as medial metaphyseal beaking and fragmentation, while none of the 14 tibiae that lacked metaphyseal changes went on to progress. While magnetic resonance imaging (MRI) can be helpful, the increased cost of the study and need for sedation in a young child make it difficult to obtain such advanced imaging routinely.⁸

Scott et al. investigated the role of body mass index (BMI) in addition to radiographic findings to help determine which patients would progress.¹⁷ They stated that the role of BMI in clinical decision‐making is especially important in those patients whose MDA falls in the gray zone of 10–15°. The authors found that a TMDA ≥10° and a BMI ≥22 together had a 95% sensitivity and 100% specificity in predicting disease progression and recommended early treatment for these patients. The BMI of 22 corresponds to >99th percentile for children aged 2–4 years.

However, several other studies investigating the natural history of infantile Blount disease have questioned the predictive utility of the MDA and other radiographic parameters. Hagglund et al. followed 13 unoperated children with infantile Blount disease and found that their MDA averaged 15° at presentation but decreased to an average of 7° prior to skeletal maturity.¹⁸ They noted that sometimes a bowing deformity with an MDA >20° could spontaneously correct. They cautioned against making treatment decisions based on a single measurement of the MDA and recommended following such patients clinically until a trend toward progression or resolution was clearer. Shinohara et al. investigated 46 patients with an MDA ≥11° and Langenskiold stage I–III changes on initial radiographs and found that 100% of those with Langenskiold stage I and 75% of those with Langenskiold stage II–III changes resolved spontaneously.¹¹ They noted that all six patients that required surgical intervention showed persistent increase in the FTA and MDA at ages three and four, whereas those who went on to spontaneous resolution showed improvements by age four. The authors concluded that the only way to determine which deformities would progress was by performing serial examinations at six‐month intervals until a clear trend was established. Laville et al. examined 26 patients with bowlegs and Langenskiold stage I radiographic changes and determined there was no difference in the FTA or MDA between patients that developed a progressive deformity and those who spontaneously corrected at presentation, but at subsequent visits a clear trend was detectable.¹² With increasing age, those tibiae which would spontaneously correct showed improvement in the FTA and MDA, whereas those with progressive deformity worsened, further emphasizing the need for clinical surveillance prior to making treatment decisions.

Resolution of clinical scenario

Our patient from the clinical scenario has several risk factors for disease progression. His MDA is 14°, which lies in the gray area of 10–15°; however, his FTR is <1 and %DT is >50%, indicating a high likelihood his deformity will progress. Furthermore, his BMI of 24 puts him at even higher risk for progression. Despite this, he is still less than three years old with Langenskiold stage II radiographic changes, so it is plausible that his condition could improve with time. As such, the patient can be followed closely with serial radiographs to determine if his deformity will normalize over time or if he will progress and require treatment. If his deformity continues to progress over the next six months then surgery should not be delayed since realignment prior to age four is associated with better long‐term outcomes.⁸–¹⁰