AMERICAN JOURNAL OF HUMAN BIOLOGY 12:610–615 (2000)

Assessment of Skeletal Maturity of the Hand-Wrist and Knee:

A Comparison Among Methods
GIORGIO AICARDI,1 MARINA VIGNOLO,1* SILVANO MILANI,2
ARTURO NASELLI,1 PAOLO MAGLIANO,1 AND PATRIZIA GARZIA1
1
DIPE - Università di Genova, Istituto G. Gaslini, Genova, Italy
2
Istituto di Statistica Medica e Biometria, Università di Milano,
Milano, Italy

ABSTRACT This study evaluated the influence of height growth and nu-
tritional status on skeletal maturation of the knee and hand-wrist. Radio-
graphs of 589 subjects (250 girls and 339 boys) from 2 to 15 years were rated
according to Greulich-Pyle, TW-20 bone and TW-RUS, RWT knee, and FELS
hand-wrist methods and a method combining FELS and RWT indicators. The
subjects were referred to the Genoa University Paediatric Department from
1980 to 1987 for short stature, simple obesity, or acute diseases. Bone age was
closer to chronological age using the RWT knee method rather than the
hand-wrist methods, while bone age assessed at the hand-wrist was closely
related to height and BMI. When skeletal maturation was delayed, Greulich-
Pyle, TW-20-bone, TW-RUS, and FELS bone ages tended to be lower than
RWT knee estimates. Conversely, if maturation was advanced the hand-wrist
estimates tended to be higher than RWT knee bone ages. The combined
estimates are close to FELS bone age values. These findings show true in-
traindividual variability of skeletal maturity at the hand-wrist and knee. A
certain “laziness” in knee maturation seems to be confirmed. Am. J. Hum.
Biol. 12:610–615, 2000. © 2000 Wiley-Liss, Inc.

Considerable independence between bone ment of G. Gaslini Children Hospital be-

ages assessed at the hand-wrist and knee
has been reported (Roche et al., 1975), re-
flecting the presence of important intraindi-
vidual differences in maturity levels of dif-
ferent skeletal areas. When examined by
the FELS-RWT system, at least 5% of cases
presented larger differences than could be
attributed to unusually large observer er-
rors (Xi and Roche, 1990). In addition, un-
der conditions involving modifications in
maturation tempo, the RWT method for the
assessment of skeletal age of the knee
proved to be less sensitive than the TW-
RUS hand-wrist method in quantifying ma-
turity advancement or delay (Vignolo et al.,
1989). The aim of the present study was to
evaluate the influence of height growth and
nutritional status on skeletal maturation at
the knee and hand-wrist.
SUBJECTS AND METHODS
The 589 subjects studied were admitted to
the Genoa University Paediatric Depart-
tween 1980 and 1987. They were referred
for: 1) short stature, height below the 10th
percentile using the Tanner and White-
house (1976) growth charts; history, physi-
cal examination, and laboratory tests al-
lowed the diagnosis of genetic short stature
or constitutional growth delay; 2) simple
obesity, weight more than 20% over ex-
pected according to height and sex; subjects
with endocrine or metabolic disorders or ge-
netic syndromes were excluded; 3) acute dis-
eases (gastro-enteritis, cutaneous, upper
and lower respiratory tract or urinary infec-
tions, seizures, psychic or stress headaches,
migraine, minor traumas, fractures); 4)
adult height prediction.
Dorso-palmar X-rays of the left hand-
Contract grant sponsor: “Istituto G. Gaslini,” Genova.
*Correspondence to: Dr. Marina Vignolo, Dipartimento di Pe-
diatria - DIPE, Università di Genova, Istituto G. Gaslini, Largo
G Gaslini 5, 16148 Genova, Italia. E-mail: gaslini@mbox.ulisse.it
Received 3 August 1998; Revision received 23 July 1999; Ac-
cepted 16 September 1999

© 2000 Wiley-Liss, Inc.

PROD #M98076R
 SKELETAL MATURITY AT HAND-WRIST AND KNEE 611

wrist and antero-posterior radiographs of TABLE 1. Height SDS and BMI-SDS distributions for
the knee were taken under standard condi- age and sex (mean and SE)
tions and assessed according to the Sample
Greulich-Pyle (1959), TW-20 bone, and TW- Age size n Height—SDS BMI—SDS
RUS (Tanner et al., 1983), RWT knee years Boys Girls Boys Girls Boys Girls
(Roche et al., 1975), and FELS hand-wrist 2–3 21 25 −0.18 −0.47 −0.03 −0.17
(Roche et al., 1988) methods, and a method (0.30) (0.25) (0.48) (0.29)
4–5 37 24 −0.79 −0.28 −0.22 1.03
using all the FELS hand-wrist and RWT (0.28) (0.31) (0.55) (0.56)
knee maturity indicators (Roche et al., 6–7 44 40 −0.22 −0.14 1.03 2.10
1988). According to the FELS and RWT (0.25) (0.21) (0.45) (0.39)
methods, the graded indicators were se- 8–9 54 54 −0.17 −0.11 0.04 1.91
(0.24) (0.22) (0.49) (0.29)
lected on the basis of chronological age and 10–11 67 52 −0.39 −0.17 1.76 1.29
sex. When the initially assigned skeletal age (0.19) (0.20) (0.31) (0.29)
differed from chronological age by more 12–13 70 47 −1.06 −1.35 1.12 1.08
than ± 1.5 years, further assessments were (0.21) (0.24) (0.31) (0.42)
14–15 46 8 −2.51 −0.60 −0.47 0.84
made using the indicators appropriate for (0.19) (0.54) (0.23) (0.81)
the first assigned skeletal age. The assessor
(MV) was experienced in working with all of
the methods and had a high level of repro- closer to chronological age with the RWT
ducibility (between observers) and repeat- knee method compared to the other meth-
ability (within observer) for all methods ods. The FELS and combined methods yield
(Vignolo et al., 1989, 1990). Radiographs of very similar mean values of relative bone
589 subjects (250 girls and 339 boys) from 2 age in each sex and age class.
to 15 years were rated. Height and weight of The relationships between Greulich-Pyle,
all subjects were measured. Height mea- TW-20 bone, TW-RUS, RWT knee, FELS
surements were standardized according to hand-wrist, and combined FELS-RWT rela-
British reference values (Tanner and White- tive bone ages vs. height-SDS and BMI-SDS
house, 1976). The body mass index (BMI: are illustrated in Figures 2 and 3, respec-
weight/height2) value was calculated for tively.
each subject and standardized using the ref- Relative bone age is closely related to
erence values of Rolland-Cachera et al. height-SDS in all methods based on the
(1982). The BMI appears to be an appropri- hand-wrist, while the dependence of RWT
ate single measure of nutritional status bone age on height-SDS is rather less
(Power et al., 1997). prominent. The contribution of RWT to the
combined estimate does not appear to
RESULTS weaken the dependence of FELS bone age
Sample size, mean and SE of height SDS on height-SDS.
and BMI-SDS for age and sex are reported Hand-wrist relative bone ages appear to
in Table 1. Mean height and mean BMI be also related to BMI-SDS; their increase is
were, respectively, somewhat lower and particularly large in overweight children
higher than the reference values, due to the (Fig. 3). The dependence of knee-RWT bone
high prevalence of subjects who were over- age upon BMI-SDS is less evident. The con-
weight and of short stature. Means ± SE of tribution of RWT to the combined estimate
Greulich-Pyle, TW-20 bone, TW-RUS, RWT, does not appear to weaken the dependence
FELS, and the combined FELS-RWT skel- of FELS bone age on BMI-SDS.
etal age estimates by age group and sex are Furthermore, skeletal maturation delay
reported in Table 2. (relative skeletal age <−1.5 years) or ad-
The relationship between relative bone vancement (relative skeletal age >1.5 years)
age (i.e., difference between bone age and were observed more frequently when skel-
chronological age) and chronological age as etal age was estimated by methods using
assessed by the different methods is shown the hand-wrist than by RWT knee method
in Figure 1. The shape of the relationship is (see Table 3).
roughly similar in boys and girls, but in Figure 4 shows the distribution of
girls the bone age values were more ad- Greulich-Pyle relative bone age conditional
vanced independently of the method used. on RWT relative bone age. When skeletal
The difference between the sexes is particu- maturation is delayed or normal, Greulich-
larly prominent using TW-RUS. Bone age is Pyle bone age tends to be lower than RWT
 TABLE 2. Mean and SE (years) of chronological age, Greulich-Pyle, TW-20 bones, TW-RUS, FELS, and combined FELS-RWT estimates by age groups and sex
Chronological Combined
Age age Greulich-Pyle TW-20 bones TW-RUS Hand-wrist FELS Knee-RWT FELS-RWT
class Boys Girls Boys Girls Boys Girls Boys Girls Boys Girls Boys Girls Boys Girls
2–3 2.99 2.90 2.84 2.61 2.87 2.52 3.03 2.82 2.90 2.61 3.02 2.70 2.91 2.65
(0.11) (0.12) (0.18) (0.12) (0.19) (0.12) (0.19) (0.13) (0.19) (0.16) (0.15) (0.14) (0.17) (0.15)
4–5 4.96 5.01 4.26 4.92 4.24 4.76 4.41 4.87 4.53 5.00 4.50 4.86 4.50 4.98
(0.09) (0.13) (0.19) (0.28) (0.19) (0.28) (0.19) (0.31) (0.20) (0.26) (0.18) (0.17) (0.19) (0.24)
6–7 6.97 6.98 6.59 7.05 6.76 7.01 6.70 7.18 6.87 7.17 6.99 6.85 6.91 7.13
(0.09) (0.09) (0.27) (0.23) (0.31) (0.24) (0.29) (0.25) (0.28) (0.24) (0.24) (0.18) (0.27) (0.23)
8–9 9.17 9.06 9.00 9.29 9.16 9.35 9.06 9.73 9.31 9.65 9.28 9.51 9.29 9.64
(0.07) (0.07) (0.27) (0.21) (0.26) (0.25) (0.27) (0.26) (0.27) (0.25) (0.20) (0.15) (0.26) (0.23)
10–11 11.01 11.05 10.78 11.19 10.95 11.32 10.95 11.85 11.06 11.61 11.34 11.52 11.11 11.61
(0.07) (0.08) (0.21) (0.19) (0.25) (0.18) (0.26) (0.21) (0.24) (0.17) (0.16) (0.15) (0.22) (0.16)
12–13 13.04 12.93 12.25 12.83 12.66 12.79 12.81 13.27 12.63 13.17 13.10 13.22 12.71 13.14
(0.07) (0.08) (0.20) (0.29) (0.23) (0.24) (0.24) (0.26) (0.22) (0.29) (0.13) (0.20) (0.20) (0.26)
14–15 14.75 14.73 12.92 14.51 13.47 13.96 13.51 14.52 13.38 14.71 13.90 14.43 13.50 14.53
(0.09) (0.17) (0.18) (0.46) (0.21) (0.36) (0.22) (0.35) (0.19) (0.72) (0.12) (0.41) (0.17) (0.59)
 SKELETAL MATURITY AT HAND-WRIST AND KNEE 613

Fig. 1. Relationships of relative bone ages (i.e., the difference between bone age and chronological age) to
chronological ages (mean ± SE) in boys (䊐---䊐) and in girls (䊏—䊏).

Fig. 2. Relationships of relative bone ages to height-SDS (mean ± SE) in boys (䊐---䊐) and in girls (䊏—䊏).

bone age. If skeletal maturation is ad-
vanced, Greulich-Pyle bone age tends to be
higher than RWT bone age. Differences of
more than 2 years were observed in about
3% of children. Similar results were ob-
served when the RWT method was com-
pared with the TW-20-bone, TW-RUS, and
FELS methods. As regards the TW-RUS
method, differences of more than 2 years
were found in about 5% of children. When
skeletal maturation is delayed, the com-
bined bone ages (FELS+RWT indicators)
tend to be lower than the RWT bone ages
(Fig. 5). When skeletal maturation is ad-
vanced, the combined bone ages tend to be
higher than RWT bone ages. Differences of
614 G. AICARDI ET AL.

Fig. 3. Relationships of relative bone ages to BMI-SDS (mean ± SE) in boys (䊐---䊐) and in girls (䊏—䊏).

TABLE 3. Subjects with skeletal maturation delay or advancement determined by different methods
TW-20 Hand-wrist Combined
Greulich-Pyle bones TW-RUS FELS Knee-RWT FELS-RWT
Skeletal age % % % % % %
Delay 25,4 21,7 21,7 20,3 7,5 18,3
(relative skeletal age
ⱕ−1.5 years)
Advancement 15,7 15,1 22,0 19,4 10,6 17,6
(relative skeletal age
>1.5 years)

more than 2 years were observed only in
about 1% of children. Interindividual vari-
ability of the combined estimates was lower
than that of the estimates based only on
hand-wrist indicators. The combined esti-
mates were close to FELS bone age values,
the large majority of differences being
within ±6 months. However, when skeletal
maturation is delayed, the combined bone
ages tend to be higher than FELS bone ages.
If skeletal maturation is advanced, the com-
bined bone ages tend to be lower than FELS
bone age.
DISCUSSION
Since there are differences between as-
sessments of various skeletal areas by dif-
ferent methods, biological limitations of the
estimates of maturity of an individual ob-
tained from the evaluation of only one part
of the skeleton should be considered. These
differences are usually related to the inter-
action of factors such as differences between
methods, observer errors, and true biologi-
cal variations in maturity.
In the present study, quite similar differ-
ences were observed between RWT knee es-
timates and hand-wrist assessments ob-
tained by the Greulich-Pyle, TW, and FELS
methods, which are very different in their
source samples, statistical basis, and appli-
cation. This seems to indicate that the ob-
served differences are independent of the
methods used and that there is a true intra-
individual variability in maturity levels of
the hand-wrist and knee.
The bone ages from the RWT method
were closer to the chronological ages than
those from the other methods using the
hand-wrist, while bone age assessed at
 SKELETAL MATURITY AT HAND-WRIST AND KNEE
Fig. 4. Distribution of Greulich-Pyle relative bone
age conditional on knee-RWT relative bone age. Dotted
lines delineate the area from −2 to 2 years of difference
between the two methods.
Fig. 5. Distribution of combined (FELS+RWT) rela-
tive bone age conditional on knee-RWT relative bone
age. Dotted lines delineate the area from −2 to 2 years
of difference between the two methods.
hand-wrist was rather closely related to
height and BMI. Therefore, knee matura-
tion proved to be less affected than hand-
615
wrist bone age by delay or advancement of
height growth as well as by overweight. On
the whole, these results confirm previous re-
sults (Vignolo et al., 1989). At present, fur-
ther data on the comparison between RWT
knee and Greulich-Pyle hand-wrist, TW,
and FELS bone ages are not available, since
both hand-wrist and knee bone ages are not
commonly assessed and these bone ages are
usually used interchangeably.
Under conditions which involve modifica-
tion in the tempo of maturation, the RWT
knee method may be less sensitive than
methods using the hand-wrist in quantify-
ing advanced or delayed maturation. Fur-
ther data are needed to choose the more
suitable assessment area between the hand-
wrist and knee, especially when estimates
of growth potential are required in condi-
tions of advanced or retarded maturation.
LITERATURE CITED
Aicardi G. 1982. L’accrecimento staturale normale e pa-
tologico. Diagnosi differenziale in pediatria. Torino:
Edizioni Medico Scientifiche.
Greulich WW, Pyle SI. 1959. Radiographic atlas of skel-
etal development of the hand and wrist. Stanford:
Stanford University Press.
Power C, Lake JK, Cole TJ. 1997. Measurement and
long-term health risks of child and adolescent fatness.
Int J Obes 21:507–526.
Roche AF, Wainer H, Thissen D. 1975. Skeletal matu-
rity. The knee joint as a biological indicator. New
York: Plenum.
Roche AF, Chumlea WMC, Thissen D. 1988. Assessing
the skeletal maturity of the hand-wrist. FELS
method. Springfield, IL: CC Thomas.
Rolland-Cachera MF, Sempè M, Guilloud-Bataille M,
Patois E, Pèquignot-Guggenbuhl F, Fautrad V. 1982.
Adiposity indices in children. Am J Clin Nutr 36:179–
184.
Tanner JM, Whitehouse RH. 1976. Clinical longitudinal
standard of height, weight, height velocity, weight ve-
locity and the stages of puberty. Arch Dis Child 51:
170–179.
Tanner JM, Whitehouse RH, Cameron N, Marshall W,
Healy MJR, Goldstein H. 1983. Assessment of skel-
etal maturity and prediction of adult height (TW2
method), 2nd ed. London: Academic Press.
Vignolo M, Milani S, Di Battista E, Naselli A, Mostert
M, Aicardi G. 1989. A comparison between TW2-RUS
hand-wrist and RWT knee assessment of skeletal ma-
turity. In: Tanner JM, editor. Auxology 88. London:
Smith-Gordon.
Vignolo M, Milani S, Di Battista E, Naselli A, Mostert
M, Aicardi G. 1990. Modified Greulich-Pyle, Tanner-
Whitehouse, Roche-Wainer-Thissen (knee) methods
for skeletal age assessment in a group of Italian chil-
dren and adolescents. Eur J Pediatr 149:314–318.
Xi H, Roche AF. 1990. Differences between the hand-
wrist and the knee in assigned skeletal ages. Am J
Phys Anthropol 83:95–102.