The value of shoe size for prediction of the timing of the pubertal growth spurt
© Busscher et al; licensee BioMed Central Ltd. 2011
Received: 12 July 2010
Accepted: 20 January 2011
Published: 20 January 2011
Skip to main content
© Busscher et al; licensee BioMed Central Ltd. 2011
Received: 12 July 2010
Accepted: 20 January 2011
Published: 20 January 2011
Knowing the timing of the pubertal growth spurt of the spine, represented by sitting height, is essential for the prognosis and therapy of adolescent idiopathic scoliosis. There are several indicators that reflect growth or remaining growth of the patient. For example, distal body parts have their growth spurt earlier in adolescence, and therefore the growth of the foot can be an early indicator for the growth spurt of sitting height. Shoe size is a good alternative for foot length, since patients can remember when they bought new shoes and what size these shoes were. Therefore the clinician already has access to some longitudinal data at the first visit of the patient to the outpatient clinic.
The aim of this study was to describe the increase in shoe size during adolescence and to determine whether the timing of the peak increase could be an early indicator for the timing of the peak growth velocity of sitting height.
Data concerning shoe sizes of girls and boys were acquired from two large shoe shops from 1991 to 2008. The longitudinal series of 242 girls and 104 boys were analysed for the age of the "peak increase" in shoe size, as well as the age of cessation of foot growth based on shoe size.
The average peak increase in shoe size occurred at 10.4 years (SD 1.1) in girls and 11.5 years (SD 1.5) in boys. This was on average 1.3 years earlier than the average peak growth velocity of sitting height in girls, and 2.5 years earlier in boys. The increase in shoe size diminishes when the average peak growth velocity of sitting height takes place at respectively 12.0 (SD 0.8) years in girls, and 13.7 (SD 1.0) years in boys.
Present data suggest that the course of the shoe size of children visiting the outpatient clinic can be a useful first tool for predicting the timing of the pubertal growth spurt of sitting height, as a representative for spinal length.
This claim needs verification by direct comparison of individual shoe size and sitting height data and than a step forward can be made in clinical decision making regarding adolescent idiopathic scoliosis.
In adolescent idiopathic scoliosis in particular it is highly important to know when the peak growth velocity of spinal length takes place. Since a close relationship exists between the spinal growth of the patient and the angle progression of the idiopathic scoliosis, it is essential to know the timing of the pubertal growth spurt in order to determine the optimum treatment strategy for the individual child [1–4]. However, exact spinal length is hard to obtain and therefore sitting height can be a representative alternative.
It is still highly difficult to predict when the individual child will have his or her pubertal growth spurt. 95% of the girls experience their peak growth velocity between ages ten and fourteen, and 95% of the boys between age twelve and sixteen [5–7]. This range is too wide to be able to make an accurate prediction in the individual patient.
It is known that different parts of the body each have their own typical growth pattern. Cameron et al., Dimeglio, and Welon and Bielicki all confirm the distal-to-proximal growth gradient as described earlier by Tanner, meaning that more distal body parts have their pubertal growth spurt earlier in adolescence. Furthermore, it is known that the sequence of growth spurts of different body parts is similar in individual children, regardless of being an "early" or "late" maturer. Therefore, the timing of the growth spurt of foot length could be an early indicator for the timing of the growth spurt of sitting height .
For an accurate determination of the timing of the peak growth velocity of foot length, longitudinal growth data are needed. However, during a first visit to the outpatient clinic patients usually don't have information concerning their foot lengths over the last years. Therefore the clinician can not make an accurate assessment at that moment, as would be preferable. To overcome this lack of information concerning the longitudinal growth of the foot, it is useful to have a representative alternative for foot length. This could be the shoe size. Pilot work for present study in the orthopaedic outpatient clinic confirmed that patients and their parents remember to a large extent when they bought new shoes and what size these shoes were. This could therefore be a very easy, first clinical tool that can help in determining the timing of the growth spurt of the foot, and therefore in predicting the timing of the growth spurt of sitting height.
Information concerning longitudinal increases in shoe size is lacking and therefore the aim of this study was to describe the increase in shoe size during adolescence. It was hypothesized that an earlier "peak increase" in shoe size would occur in comparison to the general timing of the peak growth velocity of sitting height, and that therefore the increase in shoe size could be an early indictor for the timing of the pubertal growth spurt.
Translation of shoe sizes for different regions.
38.5 - 39
39 - 39.5
41 - 41.5
43.5 - 44
44 - 44.5
The shoe sizes used for data-analysis were the standard measured sizes, and not the size of the shoes which were sold to the clients.
The age at the time of measurement was determined by subtracting the date of birth from the date of the measurement.
The data were anonymized by an external organisation before transfer to the researchers.
A customized program in Matlab® (Mathworks, Natick MA, USA) was used to plot the data of each individual client, and the results were checked visually for outliers. The period of peak increase was calculated by the program, as well as the age of the peak increase and the peak velocity of shoe size increase.
Secondary outcome measures were related to the cessation of foot growth. The age at the end of foot growth was determined in the individual graphs as the age from where the shoe size did not increase any further. This plateau period should at least continue for 1 year for the client to be included in the data-analysis of the plateau phase (plateau 2 in Figure 1). Furthermore, the start-age of the plateau phase should be higher than the age of the peak growth velocity. In this way temporary and short plateau phases, or phases which occurred before the peak growth, were excluded (plateau 1 in Figure 1).
Clients who had more than 4 measurements above the age of 8 years, as well as a follow up time of more than two years were selected. Furthermore the last measurement of shoe size should be above 12 years in girls and 13 years in boys. Only data above 8 years were used for analysis of the peak increase in shoe size since it was not expected that the growth spurt would occur earlier in healthy children, and to exclude boys and girls with precocious puberty. The selection of the clients was performed to make sure the pubertal growth spurt in shoe size was not missed. For example, if a client had data until age 10 or 11 years, the Matlab program could still calculate a shortest period in which the shoe size increased two sizes. However, this would result in a "false growth spurt" which would not represent the actual pubertal growth spurt.
A database was collected with longitudinal data of standardized shoe sizes of 636 girls and 513 boys. The individual series of ages of the clients ranged from 10 months to 17.1 years in girls, and 10 months to 17.2 years in boys.
General demographics of the used database.
Number of clients
Follow up time after age 8 in years
Number of measurements after age 8
Results of database-study consisting longitudinal series of shoe sizes of 242 girls and 104 boys
Age at peak growth (years)
Growth speed at time of peak growth (shoe sizes/year)
Age at start plateau-phase
Time between age at peak growth velocity and start of plateau-phase
Shoe size at plateau-phase
The average age for girls (n = 138) to reach a plateau phase in their shoe size was 12.0 years (SD 0.8). At that time the shoe size was on average 38.9 (SD 1.5). Boys (n = 36) reached a plateau at the age of 13.7 years (SD 1.0), with a shoe size of 41.1 (SD 1.8) (Table 3).
The average longitudinal curves for peak increase of the shoe size of girls and boys are shown in Figure 2.
Many researchers have tried to predict the timing of the pubertal growth spurt, or peak growth velocity of total body height in the individual patient. Less researchers have investigated the prediction of the growth spurt in spinal length or sitting height, but this knowledge is of great value for optimizing treatment strategies in patients with adolescent idiopathic scoliosis.
This study showed that the timing of the peak increase in shoe size was 10.4 years in girls, and 11.5 years in boys. Furthermore, the shoe size did not increase any further for at least 1 year after the age of 12.0 (SD 0.8) in girls, and 13.7 years (SD 1.0) in boys. Stavlas et al. also showed a different growth potential of the feet in boys and girls in the sense that the physiological process of foot development occurred earlier in girls in comparison to boys.
To determine the relationship between the timing of the peak increase in shoe size and the timing of the pubertal growth spurt in sitting height, a comparison was made to reliable growth data of a comparable Dutch population. Secular trends in the pubertal growth spurt should be taken into account when comparing present data with known data for sitting height. Therefore growth data from Gerver and de Bruin were taken which were collected from 1995 to 1999, in comparison to the shoe sizes in present study which were collected from 1991 to 2008. Gerver and De Bruin found a peak growth velocity of sitting height to occur at 11.7 years in girls (SD 0.8), and at 14.0 years in boys (SD 0.9). Thus, the average peak increase in shoe size generally occurred 1.3 years and 2.5 years before the average peak growth of sitting height, in girls and boys respectively. A second useful finding in the present study is that on average, the increase in shoe size diminishes when the peak growth velocity of sitting height occurs. The plateau phase of shoe size began at 12 years in girls, and a little under 14 years in boys, nearly similar to when the growth velocity of sitting height in both groups is largest (Figure 2).
These results suggest that the longitudinal course in shoe size and the timing of the peak increase in shoe size can be helpful as a first indication for the timing of the pubertal growth spurt of sitting height. A major advantage of using shoe size as an alternative for actual foot length is that patients and their parents can recall when they bought new shoes, and what the size was. Therefore the clinician already has access to some longitudinal data at the first visit of the patient, which are missing for actual foot length. Furthermore, this is a very easy and non time consuming way of getting a first impression on the stadium of growth of the patient. It is important to ask for the course in shoe size. Children and their parents remember to a large extent that they first had to buy a larger size after 1 year and than suddenly they had to buy a larger size after 0.5 year. When the increase in shoe size is approximately 2.5 sizes per year (in both girls and boys), the physician knows that on average the peak growth velocity of sitting height will occur 1.3 or 2.5 years later in girls and boys respectively. A pilot study in the orthopaedic outpatient clinic revealed that from 20 patients and their parents, 16 knew the course of their shoe size for at least 1.5 years back. However, repeated measurements should be performed in a coming longitudinal study to reveal a possible recall bias.
Comparison of the present results with previous studies
Present study 2009
Anderson et al 1956
Cameron et al 1982
Liu et al 1998
Welon et al 1979
Anderson et al and Liu et al found a cessation of the foot growth around 13.5 years in girls and 16 years in boys (Table 4). This age was higher compared to the present study for the reason that Anderson and Liu used a limit of less than 2 mm growth of the feet. One shoe size represents approximately 7 mm, so the feet can still grow more than 2 mm before a larger shoe size is needed. There might be additional factors to explain the difference of 2.3 years in boys. This will be discussed further in the next section.
A disadvantage of the initial available database was that less data were available for higher ages. Many clients only had measurements until age 10 or 11. As a result of the strict selection criteria, the initial group of clients was therefore decreased by respectively 62% and 80% for girls and boys. However, no indications were found in the non-selected group for differences in increases of shoe size. Furthermore, as will be outlined below, the results were unaffected by changing the selection criteria, thereby suggesting that the results were not biased by selection.
The minimum age for the last measurement was higher for boys and therefore fewer boys were included than girls. The minimum age was chosen to prevent the inclusion of "false" growth spurts, as described earlier in the methods section. It was shown during the analysis of the data that the ages of the peak increase did not change when only the clients were selected who had a last measurement above a higher age. For example, when only the male clients were selected who had the last measurement of shoe size above 14 years of age (instead of 13) the peak growth velocity still occurred at 11.5 years (n = 55). When the group was selected who had their last measurement above 15 years (n = 27), the age at peak growth velocity was 11.4 years. A similar test was done for the female clients and 12 years appeared to be an adequate cut-off point (n = 131 for the group with the last measurement above age 13 and age at peak growth velocity 10.4 years; n = 50 for the group with the last measurement above age 14 and age at peak growth velocity 10.6 years). Thus, for both boys and girls, the primary outcome measure, i.e. the age of peak foot growth velocity based on shoe size, seems reliable.
This study showed a plateau phase in shoe size of at least 1 year in boys starting on average at 13.7 years. The shoe size at that time was on average 41.4, which was smaller than expected.
There are some limitations to the use of the shoe size. Different brands do not always use similar size-length ratios. E.g. size 39 in one brand is not necessarily the same length as size 39 in a different brand. Pilot work for the present study showed that clients recognize this difference and know their "general" shoe size. However, further research should be performed to evaluate the reliability of the use of shoe size in a clinical setting and to evaluate a possible recall bias of patients and their parents.
For further validation of the use of shoe size for prediction of the peak growth velocity of sitting height it is essential to combine both measurements in a single study. Present work of our study group will concern these measurements, but the results will be available only after collection of longitudinal data for several years.
The present data suggest that the course of the shoe size of children visiting the outpatient clinic can be useful as a first indicator for the timing of the pubertal growth spurt of sitting height. This claim needs verification by direct comparison of individual shoe size and sitting height data. If such work supports our claim, a step forward can be made in clinical decision making regarding adolescent idiopathic scoliosis.
The authors would like to thank the shoe shops for providing the data used in the present study.
This study was supported by the Dutch Technology Foundation STW, applied science division of NWO and the Technology Program of the Ministry of Economic Affairs. This sponsor had no involvement in the study design, the collection and analysis of the data, the writing of the paper, or the decision to submit this manuscript. No author has received any form of payment for producing of the manuscript. There are no other financial or other relationships which might have led to a conflict of interest.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.