There are several methods to predict growth and limb length discrepancy in children [

4,

19–

21], and the accuracy of these methods is debatable. The most commonly used method to predict limb length discrepancy and determine maturation is the Moseley straight-line graph method [

1,

3,

5,

7,

10,

12,

14,

19,

22], which is a derivation of the data published by Anderson et al. [

16,

21,

22], and which is easy to visualise because of the graphic format [

14]. The accuracy of predicting limb length discrepancy by the Moseley straight-line graph is supported by only a small number of outcome studies [

1,

7–

10,

22]. Unfortunately, these papers have patients with a variety of different aetiologies, some being congenital deformities and others being acquired conditions [

3,

5,

10,

14,

19,

22]. The accuracy of this method is less reliable under the age of ten years [

23] and it varies because skeletal age determination is difficult and because many cases do not have a linear pattern of growth [

12–

15,

17,

24,

25].

According to the results of Timperlake et al., 17 patients from a group of 35 achieved an acceptable limb length discrepancy at maturity (limb length discrepancy less than 1.5 cm) after epiphyseodesis times using the Moseley straight-line graph [

9]. A good result was achieved in 90% (limb length discrepancy less than 0.7 cm) of the patients treated by epiphyseodesis in the series of Poratz et al. [

7]. Inan et al. found that the Moseley straight-line method accurately predicted the timing for percutaneous epiphysiodesis in all of their 97 patients but one who had hemihypertrophy [

10]. Lampe at al. found that accuracy of the Moseley straight-line graph can be limited by a variable pattern of skeletal maturation in some patients [

15,

17,

24]. They emphasised the importance of a sufficient number of preoperative measurements to develop an accurate graph. For this reason, they suggest the child be referred to a “limb length clinic” at an early age. In the series of Blair et al., there were 45 failed epiphyseodeses, and in ten the cause of failure was an inadequate operative epiphyseodesis. In the remaining 35 cases, the causes of failure were secondary to errors in estimating the proper skeletal age for operation due to inaccurate or incorrect use of Green and Anderson growth prediction tables [

1]. Stephens et al. found that the Green-Anderson charts and the straight-line graph of Moseley both provide an accurate mean of predicting future growth and determination of skeletal age of epiphyseodesis [

8]. Aguilar et al. found that the Moseley method is as accurate as the multiplier method in predicting bone maturity lengths but less accurate in predicting limb length discrepancies at maturity after epiphyseodesis [

26,

27]. Dewaele and Fabry were not able to significantly improve their results in timing the skeletal age for epiphyseodesis by using the Moseley straight-line graph, and their principal source of error was in estimating the bone age [

3]. Kelly et al. emphasised that determination of puberty is essential to reduce errors in timing of epiphyseodesis [

28]. Little at al. found that the Gruelich and Pyle skeletal age data could not be shown to increase their accuracy in predicting outcome over serial chronological data, and thus its value in predicting limb length inequality is thought to be limited, regardless of the method used, and unpredictable results occur in a proportion of patients [

14]. Evaluation of these papers allows us to conclude that skeletal age determination and variation in normal growth patterns affect the ability to predict growth of the longer (normal) limb. In the abnormal (shorter) limb additional factors such as disease process and certain treatments may affect the predictability of the growth. For example, Sharma et al. reported severe and unpredictable growth retardation of the tibia after limb lengthening [

29].

Unilateral fibular hemimelia is thought to be one of the most predictable diseases for determining limb length discrepancy. Shapiro showed that hemiatrophy (anisomelia) presents a type I growth pattern, which is linear in the Moseley straight-line method [

15]. Our data shows patients with unilateral fibular hemimelia have a different pattern of skeletal maturation from normal children. If children with fibular hemimelia are plotted on the Moseley straight-line graph as if they were normal children, the prediction of limb length discrepancy at maturity will be inaccurate. However, if they are plotted using our newly formed reference nomogram line to create the best fit parallel, the prediction will be much more accurate. In this study we had a large series of patients with unilateral fibular hemimelia who had sufficient skeletal age data to evaluate the pattern of maturation. In our group of patients between the ages of four and six the mean skeletal age nomogram line shows immaturity when compared to the normal skeletal age nomogram (Fig. ). Between the ages of six to maturity, the skeletal age nomogram declines to below the mean age line of the Moseley straight-line graph (Fig. ). This demonstrates an acceleration of the skeletal maturation process. The mean skeletal age nomogram shows the same declining character in both girls and boys after six years of age (Fig. ). This declining skeletal age line crosses the mean age line of the Moseley straight-line graph at 10.5 years in girls and 12 years in boys, and continues to decline until maturity. This makes us wonder if unilateral fibular hemimelia involves a general skeletal maturation process rather than being limited to just one limb.

The strengths of this paper are that this series consists of a relatively large group of children with unilateral fibular hemimelia, the children were followed up for many years, and multiple skeletal ages and orthoroentgenograms were obtained yearly. The weakness of the paper is that we were not able to determine the ultimate natural history of limb growth in these children with unilateral fibular hemimelia. Ethically, we could not follow them to maturity without treatment of such a severe deformity.

In conclusion, younger children with unilateral fibular hemimelia have an abnormally immature skeletal age, and as they grow their skeletal age reaches maturity more rapidly than normal. The skeletal age nomogram of patients with unilateral fibular hemimelia consistently declines when compared to the normal skeletal age graph. If the Moseley straight-line graph or the Green-Anderson charts are used for children with unilateral fibular hemimelia, the prediction of limb length at maturity will be inaccurate because they have a different pattern of skeletal maturity from normal children. The skeletal age nomogram from our data allows a more accurate prediction of skeletal maturation in children with unilateral fibular hemimelia.