PII:

SOO21-9290(96)00136-X

J. Bwmrhunim.
Vol. 30, No. 3. pp 233 250. 1997
,cm 1997 Elsev~er Scmm Ltd. All rights reserved
Pnnted I Great Rrltam
0021 9290!9; %I?.00 + .oO

THE RELATIONSHIP
OF STATIC FOOT STRUCTURE
DYNAMIC
FOOT FUNCTION

TO

P. R. Cavanagh,*.E E. Morag,* A. J. M. Boulton,$ M. J. Young,$ K. T. Deffner* and S. E. Pammert

*Center
University,

for Locomotion
Studies, 10 IM Building.
and TDepartment
of Statistics, The Pennsylvania
State
University
Park, PA 16802, U.S.A.; and :The University
Department
of Medicine,
Manchester
Royal
Infirmary,
Manchester.
U.K.

Abstract-Many
theories have been advanced concerning
the relationship
between structure
and function in the
human foot, yet few of these theories have been subjected to quantitative
examination.
In this study, foot structure
was characterized
by 27 measurements
taken from standardized
lateral and dorsi-plantar
weight-bearing
plain
radiographs
of 50 healthy adult subjects. Regional plantar pressure distribution
data collected from the same feet
were chosen as the functional
measures. A stepwise regression analysis was performed
to (1) explore what portion
of the variance in peak plantar pressure during walking can be explained by the radiographic
measurements,
and
(2) identify structural
characteristics
of the foot which are significant
predictors
of peak plantar pressure under the
heel and the first metatarsal
head (MTHl).
Most of the radiographic
measurements
were highly reliable. However,
only 3 1 and 38% of the variance in peak
plantar pressure at the heel and MTHl,
respectively,
could be explained using multiple regression analyses with the
radiographic
measurements
as independent
variables.
Among the structural
predictors
that were identified,
soft
tissue thickness (e.g. calcaneus or sesamoid heights), and arch-related
measurements
were the strongest predictors
of plantar pressure under both the heel and the first metatarsal
head. We conclude that, in normal subjects. only
about 35% of the variance in dynamic plantar pressure can be explained
by the measurements
of foot structure
derived from radiographs.
This implies that the dynamics of gait are likely to exert the major influence on plantar
pressure during walking.
(c-i 1997 Elsevier Science Ltd. All rights reserved.
Kepwrds:

Foot;

Radiographs;

Plantar

pressure;

Walking;

1990; Otremski et al., 1987). The measurement of plantar

pressure distribution has been widely used to characterize the functional aspects of foot-floor and foot-shoe
interaction (Alexander et al., 1990). Since elevated peak
plantar pressure has been shown to be associated with
tissue damage to insensate feet (Boulton et al., 1983) and
pain in rheumatoid arthritis (Lord et al., 1986) there has
been considerable speculation concerning structural factors that may predispose a foot to elevated peak pressure.
We hypothesize that elevated pressure may result from
both functional and structural factors. The purpose of the
present study is to determine to what degree foot structure as determined from radiographic measurements can
predict peak plantar pressure in the symptom-free foot
during walking.

INTRODUCTION

Historically, many opinions have been expressed regarding the relationship between structure and function in the
foot. For example, factors such as soft tissue thickness
(Gooding et al., 1986) relative metatarsal length (Fox,
1950; Morton, 1935), the configuration of the medial
longitudinal arch of the foot (Rodgers, 1995) bony prominences (Duckworth et al., 1985), and the presence of claw
and hammer toes (Habershaw and Donovan, 1984;
Myerson and Shereff, 1989) have all been implicated as
leading to elevated plantar pressure. It is also a common
clinical dictum that the higher the arch (pes cavus) the
higher the predisposition for overuse injuries (James
et al., 1978; Simkin et al., 1989). Few quantitative studies
have examined these hypotheses.
Radiography offers a convenient method for assessing
the structure of the foot during weight bearing. Saltzman
et al. (1994) have recently shown that high reliability
can be obtained from selected measurements. In a comprehensive quantitative
description
of thirty-one
measurements from a lateral view and twenty-eight
measurements from an antero-posterior (AP) view, Steel
et ul. (1980) established both the means and standard
deviations for symptom free feet (see also Renton and
Stripp, 1982). Valley and Reese (1991) also defined normal ranges of various radiographic measurements of the
foot in order to help preoperative evaluation and surgical
planning. Such measurements have been used to gain
insight into a number of foot pathologies (Nestor et al.,
Received
7 Author

in finul form 5 August 1996.

to whom correspondence
should

Structure

METHODS

This study was performed with the approval of the

Manchester Central Hospitals and Community
Care
Trust Ethical Committee, Manchester, U.K. All subjects
gave informed consent before participation.
Standardized lateral and dorsi-plantar weight bearing plain radiographs of the right foot and ankle of 50 symptom-free
subjects (mean age 63.3 + 13.1 yr, height 169.4 f 8.1 cm,
weight 70.0 f 10.5 kg, 15 females) were taken by a single
radiographer. The radiographic
methods have been
previously reported in detail (Cavanagh et al., 1994).
Briefly, for the lateral view, a focus-film distance of
100 cm was used with a horizontal beam at 63 kV,
25 mA s. The corresponding parameters for the dorsiplantar view were 80 cm with a 20 anterior tilt to the
vertical and an exposure of 63 kV, 20 mA s. A plasticine

be addressed.
243

144

P. K. Cavanagh

wedge filter was used for the dorsi-plantar view to improve the uniformity of the image. The films were scanned with a high resolution video camera and captured
with a frame grabber card and Color Snap+ software
(Computer Friends Inc., Portland, OR) on to a Macintosh Quadra 950 (Apple, Cupertino, CA). Fifteen lateral
and 12 dorsi-plantar angular and linear measurements
(Table 1 and Figs 1 and 2) were then made from the
digitized image using the NIH Image (written by Wayne
Rasband at the U.S. National Institutes of Health and
available from the Internet by anonymous FTP from
zippy.nimh.nih.gov).
Plantar pressure distribution from
the initial contact in each of three first step walking trials
was collected at 30 Hz using an optical pedobarograph
(Baltimore Therapeutic Equipment Co., Hanover, MD.
U.S.A.). The system has been reported to be reliable by
Hughes et al. (1987). In subsequent analysis, the peak
pressures at 10 discrete sites (heel, mid-foot, five metatarsal regions, hallux, second toe, and lateral toes) were
determined for each step, and averaged. Only the results
from the heel and first metatarsal head are considered
here. The natural log transformation was applied to the
resulting peak plantar pressure values to stabilize the
error variance and to normalize the distribution of the
response variables in order to satisfy the assumptions of
regression analysis. Pressure data for two subjects were
lost due to technical problems and thus, the regressions
were conducted using 48 subjects. In order to assess the
reliability of the radiographic measurements, ten subjects
dismounted from the X ray stand, were then repositioned, and repeat radiographs were taken. For each
of the resulting 20 lateral and 20 dorsi-plantar radiographs, three repetitions were taken for every radiographic measurement (all 60 measurements were randomly ordered). Intraclass correlation coefficients (ICCs)
were calculated to assess the reliability of measurement
repetition (i.e. measurements taken from the same film at
different time points) and the radiographic reliability (i.e.
measurements taken from two different films of the same
subject). The following model was used for each of the 27
radiographic measurements:
Yi,j,k

= /d + ai +

xj

(ax)i,j

Ci.j,k,

(1)

where ~1is the overall mean, ai is a random subject effect,

xj is a random radiograph effect, (ax)i,j is a random
subject by radiograph effect, and
is a random error
term or repetition effect. Restricted maximum likelihood
estimates were calculated for each of the variance components using the SAS procedure PROC VARCOMP
(SAS Institute, 1989). In order to estimate measurement
repetition reliability the following ICC was calculated:
&i,j,k

This term measures the correlation between measurements taken on the same film on the same day (repetition
effect). The ICC representing radiographic reliability, the
correlation between measurements taken on the same
person from different films, is given by
r=

2
Qa

er ctl

t
1. length

6. Cnlc ht

11 1.1sfranc

2. Ses ht

7. Inf Calc

12 M5

8. sup cnic

13 Ml

4. 5 base ht

9. Chop

14. PI

5. Nawc ht

3. MM5

ht

NavCun

* AU angles are measured

Fig. 1. Schematic

behveen

15. Talus
the ax,s that bmcb

the bone and the horizon

diagram of radiographic
measurements
view (see Table 1 for further details).

from

lateral

Fourteen measurements from the lateral film were

chosen as potential predictors of heel pressure. Similarly,
12 measurements from the lateral film and eight from the
dorsi-plantar film which had anatomical relevance were
chosen as potential predictors of MTHl pressure. We
hypothesized that, in certain cases, the relationships between ln(pressure) and foot structure variables would be
nonlinear and, thus, offered squared terms to the regression equation. A stepwise regression approach of the
form

+ p2n-

1 Xin

B2nxiZn

Ei3

(4)

where
Yi = ln(pressure) for the ith subject (i = 1, . . . ,48),
/Ij = coefficient (j = 1, . . ,28 for the heel and
j = 1, . . . ,40 for MTH l),
Xij = value of the jth predictor variable for the ith
subject,
n = number of potential predictors offered to the
equation. (n = 14 for the heel and 11= 20 for
MTH l),
.si = error term for the ith subject,
was applied to predict the ln(pressure) under the heel and
MTHl from the radiographic measurements using the
SAS procedure PROC REG (SAS Institute, 1989). The
criterion for entry into and staying in the model for the
regressors was set at p = 0.05. Two measurements with
low reliability were not offered to the model (talar inclination [talus] and intermetatarsal
1-2 angle Cl-2
angle]).
RESULTS

Reliability of the radiographic measurements was, in

general, high (Table 1). Measurement repetition ICCs
for only two of the 27 radiographic measurements (talar

Foot length
Sesamoid heigth
Fifth MT head height
Fifth MT base height
Navicular height
Calcaneal height
Inferior calcaneal inclination
Superior calcaneal inclination
Choparts joint angle
Navicular lst-cuneiform angle
Lisfrancs angle
MT 5 inclination
MT 1 inclination
Proximal 1st phalanx inclination
Talar inclination
Intermetatarsal l-5 angle
Intermetatarsal I-2 angle
MTl-phalangeal
1 angle
Interphalangeal 1 angle
Medial sesamoid x deviation
Medial sesamoid y deviation
Lateral sesamoid x deviation
Lateral sesamoid y deviation
Mortons Index
Metatarsal 1 thickness
Metatarsal 2 thickness
Metatarsal 3 thickness

Term

Note: *KC1 = measurement repetition ICC.

*ICC2 = radiographic ICC.

Length
Ses ht
MTH 5 ht
5 base ht
Navic ht
Calc ht
Inf Calc
Sup Calc
Chop
Nav-Cun
Lisfranc
M5
Ml
PI
Talus
I-5 angle
1-2 angle
M-P angle
I-P angle
MSx
MSy
LSX
LSY
Mort
Tl
T2
T3

Code
the laterul

view

Total length of foot

Distance between floor and lower surface of the 1st MTH sesamoids
Distance between floor and lower surface of the fifth MTH
Distance between floor and lower surface of the fifth MT base
Distance between floor and lower surface of the navicular tuberosity
Distance between floor and lower surface of the calcaneus
Angle between the horizon and the inferior surface of the calcaneus
Angle between the horizon and the superior surface of the calcaneus
Angle between the horizon and the proximal navicular surface
Angle between the horizon and the proximal cuneiform 1 surface
Angle between the horizon and the proximal MT 1 surface
Angle between the horizon and the midline of the fifth MT diaphysis
Angle between the horizon and the midline of the first MT shaft
Angle between the horizon and the proximal first phalanx shaft
Angle between the horizon and the diaphysis of the talus
Angle between the first and fifth MT diaphyses
Angle between the first and second MT diaphyses
Angle between the first MT diaphysis and the proximal phalangeal 1 shaft
Angle between the proximal and distal phalangeal 1 diaphyses
Deviation of the med sesamoid from the origin in the med-lat direction
Deviation of the med sesamoid from the origin in the ant-post direction
Deviation of the lat sesamoid from the origin in the med-lat direction
Deviation of the lat sesamoid from the origin in the ant-post direction
Distal protrusion of MTH2 relative to MTHl (see Fig. 2)
Minimum thickness of the first metatarsal diaphysis
Minimum thickness of the second metatarsal diaphysis
Minimum thickness of the third metatarsal diaphysis

Measitrernentsfronl

Description

Table I. List of radiographic measurements and their statistics

mm

mm
mm
mm
mm
mm
mm

deg
de
de
deg
de

mm
mm
mm
mm
mm
mm
de
deg
deg
deg
deg
de
de
deg

Units

249.1
1.6
5.1
12.4
40.2
9.8
22.5
21.3
62.0
62.3
63.2
11.0
21.3
7.3
28.8
22.9
7.8
14.4
10.4
2.4
11.8
7.6
14.4
1.3
13.6
6.9
6.1

Mean

13.6
2.8
1.7
3.4
8.2
2.4
6.1
6.1
4.9
4.8
4.3
3.0
4.6
1.7
4.5
4.3
2.7
8.8
4.4
3.9
2.9
4.3
2.1
3.2
1.8
0.8
1.0

SD.

0.98
0.96
0.92
0.97
0.99
0.98
0.98
0.95
0.90
0.85
0.75
0.87
0.83
0.88
0.54
0.94
0.46
0.95
0.84
0.95
0.89
0.98
0.83
0.99
0.98
0.93
0.86

*ice1

0.98
0.87
0.84
0.87
0.98
0.96
0.97
0.95
0.90
0.70
0.71
0.64
0.75
0.87
0.51
0.92
0.44
0.88
0.70
0.95
0.86
0.98
0.80
0.98
0.98
0.90
0.84

*ICC2

P. R. Cavanagh

1. 1-5 angle

2. 1-2 angle

4. I-P angle

5. MSx

7. LSX
10. Tl

8. LSy
11. T2

3. M-P angle
6. MSy
9. Mort
12. T3

(a)
Fig. 2. (a) Schematic

er cl/.

(b)

diagram of radiographic
measurements
from AP view (see Table
(b) An expanded
view on the metatarsal
head region.

inclination and the angle between the first and second

metatarsal diaphyses) were less than 0.75. Furthermore,
the majority of the coefficient values were greater than
0.90. The radiographic
ICCs were not as high as the
measurement repetition ICCs, with only 21 of the 27
radiographic ICCs being greater than or equal to 0.75.
However, the majority of the values for the radiographic
ICCs were greater than 0.85, and the differences between
measurement repetition and radiographic ICCs were minor (less than 0.1 in 24 of the 27 radiographic measurements).
In the regression models, only three and four variables
(from the potential 14 and 20 for heel and MTHl, respectively) entered the model as significant predictors (Table
2). Both models contained a plantar soft tissue thickness
related measurement
(calcaneal height or sesamoid
height in the heel and MTHl equations, respectively) and
a medial longitudinal arch related measurement (e.g. first
metatarsal inclination in both models and calcaneal inclination for the MTHl model).

The regression
ln(pressure)

1 for further

details)

equation for heel pressure is

= 6.680 - O.O69(calcht) - O.O44(m5)
+ 0.001(m1)2

(5)

with a coefficient of multiple correlation

of 0.56 (an
explained variance of 31%). A lower calcaneal height
(calcht) is associated, as expected, with higher plantar
pressure. This feature, which represents less plantar soft
tissue thickness during compression, is the most dominant structural measurement associated with high plantar pressure in the heel (Table 3). In addition, greater
inclination of the first metatarsal (ml, this is entered as
a squared term) and lower inclination of the fifth metatarsal (m5) contribute to higher plantar pressures under
the heel. An appreciation for the effect of variation in the
predictor variables on plantar pressure can be gained
from Table 3, where the effect of an increase and decrease
of one standard deviation (and of one unit-degree
or
mm) in each of the predictors on the estimated peak

247

The relationship of static foot structure to dynamic foot function

pressure in the heel is summarized. These results demonstrate, for example, that an individual with a calcaneus
height which is 1 mm below the average for this sample
will have a predicted peak plantar pressure under the
heel which is 28.1 kPa larger than the mean value. All
other predictors were kept at their mean values while the
variable under consideration was manipulated.
The regression equation that describes first metatarsal
head pressure is
ln(pressure) = 5.799 - O.O63(mort) - O.l37(sesht)
+ O.O94(ml) - 0.001(infcalc)2

(6)

with a coefficient of multiple correlation of 0.61 (an

explained variance of 38%). The interpretation
of the
equation is that the factors which are associated with
higher peak pressure under the first metatarsal head
during walking were: (1) lower Mortons index (mort) (i.e.
small or no difference in the A-P alignment of the first
and the second metatarsal heads), (2) a smaller measured
thickness of piantar soft tissue under the sesamoids
(sesht) during standing, (3) a greater first metatarsal inclination (with reference to the standing surface) and (4)
a lower inferior calcaneal inclination (inf talc, squared
term). A decrease in one standard deviation of sesamoid
height (2.8 mm) is predicted to increase plantar pressure
by 225.9 kPa (Table 4). Again, all other predictors were

kept at their mean values while the variable under consideration was manipulated.
DISCUSSION

The ICC values presented in Table 1 indicate high

reliability
of the vast majority of the radiographic
measurements used in this study (for both measurement
repetition and radiographic reliability). High measurement repetition ICCs suggest a high reliability of the
method used to measure angles and distances from plain
radiographs, and consistency of the observer. Even more
impressive were the high radiographic reliabilities of
measurements taken from two different weight bearing
films of the same subject. The latter finding suggests that
when the radiographic
procedure is standardized,
measurements of most of the angles and distances of
interest here are objective and reliable. These findings
confirm the results of Saltzman et al. (1994).
Notabie exceptions to the findings of high reliability
were the talar inclination (TI) and the angle (in the A-P
radiograph) between the first and second metatarsals
diaphyses (M12), both of which had ICC values much
lower than that of the other measurements. Examination
of the variances used to compute the ICCs indicated
that TI reliability was low because of a larger coefficient
of variation ((cr,/Z) 100) compared with other angles,

Table 2. Five number summaries (minimum, Ql, median, Q3, maximum) of mean peak pressure, predictor
list, model R, and proportion of explained variance (R) under the heel and the first metatarsal head
(n = 48)

Minimum
25th percentile (Ql)
Median
75th percentile (Q3)
Maximum
Predictors

Heel

First metatarsal head

183.1 kPa
318.8 kPa
387.5 kPa
458.6 kPa
2053.6 kPa
Calcaneal height
(First metatarsal inclination)*
Fifth metatarsal inclination

98.1 kPa
255.1 kPa
390.8 kPa
879.0 kPa
3668.9 kPa
Mortons index
Sesamoid height
First metatarsal inclination
fnf calcaneal inclination
0.61
0.38
0.69

Model R
Model R
Intraclass correlation
coefficient (ICC)

0.56
0.3 1
0.15

Note: The intraclass correlation coefficient value is the ratio: (between subject variance/total variance).
The total variance is the sum of the between-subject variance and within-subject variance.
Table 3. Contribution

Variable

Description

calcht
m5

Calcaneus height
Metatarsal
5 inclination
Metatarsal
1 inclination

(ml)

of various predictors to peak plantar pressure under the heel during walking
AP
per
+ unit

-??.D.

Pi

AP
per
+ 1S.D.
Ha)

&Pa)

Wa)

BP
per
- unit
@Pa)

0.13
0.07

0.07
0.044

- 60.8
- 48.6

- 26.2*
- 16.9f

71.9
55.4

28.1*
17.7t

0.11

0.001

96.6

- 63.3

- l6.0~

Effect of
increase on
pressure

Mean

S.D.

part.
I4

part.
R*

9.8
11.0

2.4
3.0

0.36
0.26

21.3

4.6

0.33

175t

AP

Note: Peak plantar pressure decreases by 26.2 kPa as the thickness of the soft tissue increases by 1 mm. Note that the effect of both
increase and decrease of ISD is shown because of logarithmic transformation.
*kPamm-.
tkPa*deg-.

24x

P. R. Cavanagh et ul.
Table 4. Contribution

of various predictors to peak plantar pressure under the first metatarsal head

Part.
SD.

Part.
IRI

RZ

(@a)

WW

AP
pe,
- ISD.
(kPa)

1.3

3.2

0.26

0.07

0.063

- 86.5

~ 29.3*

105.6

31.2*

7.6

0.137
0.094

225.9

70.4*

21.3

0.10
0.13

- 61.4*

0.31
0.37

- 153.9

Metatarsal
1 inclination

2.8
4.6

47.3t - 169.5

~ 4XOt

Inferior calcaneal
inclination

22.5

6.1

0.29

0.08

0.001

- 129.1

Variable

Description

Effect Mean

mort

Mortons index

sesht

Sesamoid height

ml
(infcalc)

AP
per
+ 1S.D.

AP
per
+ unit

262.1

- 21.6t

129.1

AP
per
~ unit
(kPa)

21.61-

Note: Peak plantar pressure decreases by 61.4 kPa as the thickness of the soft tissue between the sesamoid and the ground
increases by 1 mm. Note that the effect of both increase and decrease of 1SD is shown because of logarithmic transformation.
*kPamm-.
tkPa*degg.

whereas Ml2 was low because of the relatively small

variation in this angle between subjects. Therefore, we do
not recommend the use of talar inclination as an objective measure from radiographs. The variability in this
measure is due, apparently, to the lack of distinct bony
landmarks for use in measurement of the talar angle. We
can recommend, however, the use of the angle between
the first and second metatarsal diaphyses because consistent results are found when it is repeatedly measured.
Although it has been argued that the validity of length
measurements taken from radiographs is poor (Perry
et al., 1992), this is probably due to variable magnification. We did not correct for magnification
errors
(Camasta et al., 1991) because there was little variation in
positioning between subjects as a result of the careful
standardization
of technique that was employed. Our
results suggest that standardized weight bearing radiography can be used as an objective tool for assessment of
foot structure.
The results of the regression analysis indicate that
a small number of radiographically
obtained structural
measurements can explain approximately
35% of the
variance in peak plantar pressure under the heel and
MTHl during walking. This confirms that static structural variables are significant predictors of dynamic foot
function in walking. It was also noted that prediction of
plantar pressure under MTHl
met with somewhat
greater success than the prediction of heel pressure
(R* = 38% vs 31%). This trend, though small, should be
investigated further since we have observed that these
differences are even greater in pathological feet.
Two dominant factors in the prediction of pressure for
both regions were compressed soft tissue thickness (as
expressed by calcaneal or sesamoid height) and medial
longitudinal arch height (expressed by ml inclination).
While each additional millimeter of compressed soft tissue thickness at the heel is predicted to reduce heel
pressure by 26.2 kPa, a thickness change of the same
amount at MTHl
results in a reduction of 61.4 kPa
pressure at that location. The same equation predicts
that a normal individual with a tissue thickness that is
one standard deviation below the mean will experience
MTHl
plantar pressures in excess of 650 kPa during
walking.

The effect of ml inclination on heel pressure is substantial, with an increase of 17.5 kPa accompanying each
one degree of increase in ml inclination. However, the
effect of ml and calcaneal inclination on MTH 1 pressure
must be considered together because both measures contribute to the same property of the foot (arch height), yet
they both appear in the regression equation and have
opposite signs. At low ml angles (low medial longitudinal arch) MTHl pressure is small and (Fig. 3) the effect
of calcaneal inclination is negligible. However, as ml
inclination increases (indicating a high arch), calcaneal
inclination would also increase and, thus, there will be
a diminishing effect of ml inclination on MTHl pressure as a pressure proportional to the squared calcaneal
angle term is subtracted from the predicted result. Thus,
the squared calcaneal term modulates the effect of ml
angle so that, at higher ml angles (which will usually be
accompanied by high calcaneal angles-the correlation
between the two was 0.51), the predicted pressure will not
be as high as expected due to the increase in ml inclination alone.
Since several of the measurements were correlated, it is
not surprising that some of the more commonly used
radiographic measurements were not represented in the
regression equations. For example, the height of the
medial longitudinal
arch is often characterized by
navicular height, which was one of the radiographic
measurements made in the present study. Navicular
height correlated 0.8 with ml inclination (Table 5) and,
with a slightly different dataset, navicular height may
have been included in the regression equation. Future
studies may benefit from exploring factor analysis to
identify related groups of measurements.
In this study, the widely held belief that a cavus foot
type is associated with high plantar pressure has been
confirmed: a greater ml inclination, which occurs in
a cavus foot type, is predicted by our equations to lead to
elevated pressure under the rearfoot and the forefoot.
The respective increases per degree are predicted to be
17.5 and 47.3 kPa under the heel and MTHl, respectively
(Tables 3 and 4).
This study also provides insight into the role of shortness of the first metatarsal relative to the second (which
we have characterized by Mortons index) in determining

The relationship of static foot structure to dynamic foot function

Effect

of

Calcaneus

249

Angle

1800
1600
1400
1200
1000
6oo

Pressure

(kPa)

600
400
200
0

Fig. 3. Contribution of first metatarsal and inferior calcaneat inclinations to the pressure under the first
metatarsal head. Note that at low metatarsal 1 angles, MTHl pressure is also low and the effect of calcaneal
inclination is negligible. However, as ml inclination increases, the effect of calcaneal inclination on MTHI
pressure is more marked. Other predictors in equation (6) were kept constant at their mean values. See text
For further details,
Table 5. Pearson correlation coefficients between arch related
radiographic measures of the foot
Radiographic measures
Navicular height and metatarsal 1 inclination
Navicuiar height and inferior calcaneal
inclination
Metatarsal 1 inclination and inferior
calcaneal inclination

r value
0.80
0.56
0.51

pressure under MTHl. Historically, a Morton foot structure is defined as a foot in which (a) the head of the
second metatarsal is more distally placed than the head
of the first, (b) the diaphysis of the second metatarsal is
thickened, and (c) there is hypermobility
at the first
cuneiform-first
metatarsal joint (Morton, 1935). Stokes
et a2. (1979) have suggested that the load on the metatarsal beads is related to the relative lengths of the metatarsals, and Rodgers and Cavanagh (1989) found
a greater pressure differential between MTHl
and
MTH2 in subjects with a Morton foot structure. The
present results confirm that when MTH2 protrudes more
distally than MTHl,
a lower pressure results under
MTHl than when the heads are aligned or when MTHl
is the more distally placed. Functionally, it is likely that
the angle made by the foot and the direction of walking is
also a factor in this relationship.
Surprisingly, our results do not support another component of Mortons hypothesis: that high pressure under
the second metatarsal head leads to hypertrophy of the
diaphysis of the metatarsal. The correlation between
MTH2 pressure and the second metatarsal thickness (T2)
was - 0.33, suggesting an inverse relationship between
pressure and metatarsal thickness. It should be recalled
that the pressure measurements were made during barefoot walking and that the type of footwear habitually
used by the subjects will have a significant and different
effect on the daily stress histories of MTH2 in different
subjects. Also, we have no information on the activity
profiles for the different subjects.

In summary, this study has shown that a small number

of structural variables can explain approximately 35% of
the variance in plantar pressure under the heel and the
first metatarsal head during walking. Since factors such
as walking speed and the kinematics of gait were not
measured in the present study, it is likely that the addition of these and other dynamic variables would result in
more comprehensive models to predict plantar pressure.
The present approach also appears promising in understanding the factors which contribute to elevated pressure in individuals with mechanically induced foot ulceration, and we are in the process of exploring the prediction of plantar pressure from structural factors in a group
of patients with diabetic sensory neuropathy.
Acknowledgements-This work was supported in part by the National
Institutes of Health under grant R01 DK 42912 from NIDDKD and by
the Peter Kershaw trust. The authors thank Karen Vickers for her
assistance in taking the radiographs.
REFERENCES
Alexander, I. J.. Chao, E. Y. S. and Johnson, K. A. (1990) The assessment of dynamic foot-to-ground contact forces and plantar pressure
distribution: a review of the evolution of current techniques and
clinical applications. Foor Ankle 11, 152% 167.
Boulton, A. J. M., Hardisty, C. A., Betts, R. P., Franks, C. I., Worth. R.
C., Ward, J. D. and Duckworth, T. (1983) Dynamic foot pressure and
other studies as diagnostic and management aids in diabetic neuropathy. Diabetes Care 6, 26-33.
Camasta, C. A., Pontious. J. and Boyd, R. B. (1991) Quantifying magnification in pedal radiographs. f. Am. Podiat. Med. Assoc. 81, 545-548.
Cavanagh, P. R.. Young, M. J., Adams, J. E.. Vickers, K. L. and
Boulton, J. M. (1994) Radiographic abnormalities in the feet of
patients with diabetic neuropathy. Diabetes Care 17, 201.-209.
Duckworth, T., Boulton, A. J. M.? Betts, R. P., Franks, C. I. and Ward,
J. D. (3985) Plantar pressure measurements and the prevention of
ulceration in the diabetic foot. J. Bone fr Sury. 67, 79-85.
Fox, M. G. (1950) The effect of shortness of the first metatarsal bone on
foot function. Res. Q., 277-286.
Gooding, G. A. W., Stess, R. M. and Graf, P. M. (1986) Sonography of
the sole of the foot: evidence for loss of foot pad thickness in diabetes
and its relationship to ulceration of the foot. Invest. Rudiol. 21,45-48.
Habershaw, G. and Donovan. J. C. (1984) Biomechanical considerations of the diabetic foot. In Management ofDiabetic Font Problems

P. R. Cavanagh

-.Ii0

(Edited by Ko~ac. G. P., Hoar, C. S., Rowbotham.

J. L., Wheelock.
I-. C.. Gibbons.
G. W. and Campbell, D.). pp. 32 44. W.B. Saunders.
Philadelphia.
Hughes. J., Kriss, S. and Klenerman,
L. (1987) A clinicians view of foot
pressure: a comparison
of three different methods of measurement.
F-001 .Ardi/r I. 277 284.
James. S. L.. Bates, B. T. and Ostering,
L. R. (1978) Injuries to runners.
3rn. J. Sports Med. 6, 40 50.
Lord. M.. Reynolds.
D. P. and Hughes. J. R. (1986) Foot pressure
measurement:
a review of clinical findings.
J. hiomed. Enyny 8,
283

~294.

Morton.
D. J. (1935) Tile Hzrmun Foot. Columbia
University,
New
York.
Myerson.
M. S. and Shereff, M. J. (1989) The pathological
anatomy
of
claw and hammer toes. J. Bone Jr Sury. (Am) 71-A, 45-49.
Nestor.
B. J.. Kitaoka,
H. B., Ilstrup,
D. M., Berquist.
T. H. and
Bergmann
A. D. (1990) Radiologic
anatomy
of the painful bunionette. Foot Ankle 1, 6- Il.
Otremski.
I., Salama,
R., Khermosh,
0. and Wientroub.
S.
I 1987) Residual
adduction
of the forefoot.
A review
of the
Turco procedure
for congenital
club foot. J. Bone Jt Sury. (Br) 69,
831-834.

Perry. M. D.. Mont. M. A., Einhorn,

T. A. and Waller, J. D. (1992) The
validity
of measurements
made on standard
foot orthoroentgenograms. Foot Ankle 13, 502-507.

et (I/

Renton, P. and Strlpp, W. J. (1982) The radiology

and I-adiograph)
01
the foot. In Tllr Fm)t crnd Its Disr~,v/rrs (Edited by Klenerman.
S.). pp.
305-399.
Blackwell
Scientific. Oxford.
Rodgers,
M. M. (1995) Dynamic
foot blomcchanics.
./OSP7
21,
306.

316.

Rodgers. M. M. and Cavanagh.

P. R. (1YXY) Pressure distribution
m
Mortons
fool structure.
Med. Sci. Sports E.\~~isc 21, 23 3X.
Saltzman, C. L.. Brandser, E. A.. Berbaum,
K. S.. DeGnorc.
L.. Holmes.
J. R., Katcherian,
D. A., Teasdall, R. D. and Alexander.
I. J. (1994)
Reliability
of standard foot radiographic
measurements.
Fool .4&/c,
15, 661.. 665.
SAS Institute (1989) SASISTAT
C&rs Glride (4th edn.) Version 6. Vol.
2, SAS Institute
Inc., Cary, NC.
Simkin. A., Leichter,
I., Giladi, M.. Stein, M. and Milgrom,
C. (1989)
Combined
effect of foot arch structure
and an orthotic
device on
stress fractures. Foot Ankle 10, 25-29.
Steel III, M. W., Johnson, K. A., DeWitz.
M. A. and Listrup.
D. M.
(1980) Radiographic
measurements
of the normal adult foot. Foot
Ankle 1, 151 158.
Stokes, I. A. F., Hutton,
W. C., Stott, J. R. R. and Lowe, L. W. (1979)
Forces under the hallux valgus foot before and after surgery. Clin.
Orthop. 142, 64-72.
Valley, B. A. and Reese. H. W. (1991) Guidelines
for reconstructing
the
metatarsal
parabola
with the shortening
osteotomy.
J. Am. Podint.
Med. Assoc. 81, 406 413.