ARTICLE IN PRESS

Manual Therapy 13 (2008) 112–121

www.elsevier.com/locate/math

Original article

Postpartum characteristics of rectus abdominis on

ultrasound imaging
Yvonne Coldrona,, Maria J. Stokesb, Di J. Newhamc, Katy Cookd
a
Department of Basic Medical Sciences, St George’s, University of London, UK
b
School of Health Professions and Rehabilitation Sciences, University of Southampton, UK
c
Division of Applied Biomedical Research, School of Biomedical & Health Sciences, King’s College London, UK
d
Fetal Medicine Unit, St. George’s Hospital NHS Trust, London, UK

Received 7 November 2005; received in revised form 22 August 2006; accepted 20 October 2006

Abstract

This cross-sectional and partial longitudinal study aimed to characterize changes in rectus abdominis (RA) and provide reference
ranges for the first year postpartum. Ultrasound scanning was used at four stages postnatally to measure cross-sectional area (CSA),
thickness, width (indirectly using a shape value) and inter-recti distance (IRD). One hundred and fifteen postnatal women (though
some postnatal subjects appeared in more than one postnatal group thus giving a total of 183 data points) and 69 age-matched
nulliparous female controls were recruited. Postnatal subjects were studied at Day 1 (PN1; n ¼ 63) and at 2 (PN2; n ¼ 55), 6 (PN3;
n ¼ 39) and 12 (PN4; n ¼ 26) months postpartum. Longitudinal data were analysed for CSA, thickness, shape (indirect width
measurement) (df ¼ 67) and IRD (df ¼ 62). The mean CSA of the PN1 group was significantly larger (Po0.001) than in controls
and decreased (Po0.0021) by 12 months. In all postnatal groups, RA was significantly thinner (Po0.0001, PN1–PN3; Po0.0478,
PN4), wider (Po0.0001, PN1–PN3; P ¼ 0.0326, PN4) and the IRD was significantly larger (Po0.0001, PN1–PN4) than in controls.
Over 2 months postpartum, RA became thicker (P ¼ 0.0003) and the width and IRD decreased (Po0.0001 and P ¼ 0.0002,
respectively) but did not return to control values by 12 months. These results have implications for strength of RA postpartum and
anterior abdominal wall stiffness, which together with other muscle characteristics could inform development of effective postnatal
exercise programmes.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Rectus abdominis; Postpartum; Ultrasound scanning; Muscle characteristics

1. Introduction et al., 2005) and poor control of the abdominal muscles

The main musculoskeletal problems encountered in
postnatal women are low back and pelvic girdle pain
(Mantle et al., 1977; Ostgaard et al., 1991, 1996, 1997;
Ostgaard and Andersson 1992; Ostgaard, 1997; Noren
et al., 2002), diastasis recti abdominis (DRA) (Bursch,
1987; Boissonnault and Blaschak, 1988; Potter et al.,
1997a; Lo et al., 1999; van Uchelen et al., 2001; Nahas
Corresponding author. Physiotherapy Department, Mayday
Healthcare NHS Trust, London Road, Thornton Heath, Surrey CR7
7YE, UK. Tel.: +442084013093.
E-mail address: yvonne.coldron@mayday.nhs.uk (Y. Coldron).
(Gilleard and Brown, 1996; Potter et al., 1997b). As well
as consulting physiotherapists for clinical problems,
postnatal women often attend exercise groups to restore
their figure and fitness. The exercises used in physiother-
apy departments or exercise classes are not based on
evidence of changes to the abdominal muscles in
pregnancy, and very little literature about muscular
changes during and after pregnancy is available. One of
the muscles thought to undergo change in pregnancy is
the rectus abdominis (RA) (Gilleard and Brown, 1996;
Lo et al., 1999).
The two bellies of RA extend the whole length of the
anterior abdomen and are connected by the linea alba.

1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.10.001
 ARTICLE IN PRESS
Y. Coldron et al. / Manual Therapy 13 (2008) 112–121
The main function of RA is to flex the trunk on a fixed
pelvis or flex the pelvis on a fixed trunk (Williams et al.,
1989; Kendall et al., 1993). Working isometrically it
contributes to trunk stability to allow limb movement
(Negrao Filho et al., 1997). The distribution of muscle
fibres in a skeletal muscle have been classified on the
basis of their content of different myosin heavy chain
(MHC) isoforms and identification of three main human
muscle fibre types (I, IIA and IIX (previously called
IIB)) have been established (Bottinelli et al., 1999). A
faster Type II (IIB) MHC isoprotein may also be
expressed in skeletal muscle (Graziotti et al., 2001). In a
histological study of the abdominal muscles, Caix et al.
(1984) reported approximately twice as many Type I as
Type II muscle fibres and relatively few Type IIX fibres
compared with Type IIA in the RA muscle. However,
Haggmark and Thorstensson (1979) found the distribu-
tion of the two main fibre types, Types I and II, to be
similar (mean 55–58% Type I, 15–23% Type IIA,
21–28% Type IIX fibres) but a large inter-individual
variation was found.
As the fetus grows, the RA of the mother elongates as
her abdominal wall expands. The linea alba softens and
the two bellies curve round the abdominal wall with
most separation occurring at the umbilicus (Boisson-
nault and Blaschak, 1988; Fast et al., 1990; Gilleard and
Brown, 1996). This gap, the inter-recti distance (IRD),
may vary from 2 to 3 cm wide and 2 to 5 cm long to
20 cm wide and involving the whole length of RA (see
Polden and Mantle, 1990). This increased IRD is often
referred to as a diastasis or divarication of RA (DRA).
Imaging techniques using computerized tomography
(CT) scans have been used to set a pathological DRA at
an IRD of 42.7 cm at the level of the umbilicus (Rath
et al., 1996). Using ultrasound imaging, this criterion
has been used to measure IRD after abdominoplasty for
postpartum DRA and determine those patients who
needed further surgery (van Uchelen et al., 2001).
Table 1
Contribution of data points to repeated measures trend estimation
No. of data points No. of subjects No. of data points
available
1 70 70
2 29 58
3 9 27
4 7 28
Total 115 183
The numbers of data points used to estimate the common slope (i.e. average rate of change per month) of ultrasound measurements of thickness,
shape value (indirect width measurement) and cross-sectional area of the rectus abdominis (RA) muscle are shown. The degrees of freedom (df) are
calculated from those subjects who contributed at least two data points. There were 115 subjects in total contributing 68–1 ¼ 67 df to the estimation
and testing of the average rate of change.
a
One more df used to estimate the common slope giving 67 df overall.
113
Normative data for postpartum RA thickness, width
and cross-sectional area (CSA) are not available.
Knowledge of changes in RA postpartum is important
for the development of rational postnatal exercise
programmes and general postnatal advice.
This study aimed to characterize the size and shape of
RA and the IRD postpartum and also to produce
normative data for the resolution of these variables
during the first postnatal year.
2. Methodology
2.1. Study design
A prospective cross-sectional cohort study design was
utilized and repeated measures were used (where
possible) to provide a partial longitudinal design.
2.2. Subjects
A total of 69 nulliparous female controls (CTL) and
115 postnatal women (both primiparous [n ¼ 72] and
multiparous [n ¼ 43]) participated in the study. Some
postnatal subjects appeared in more than one postnatal
group thus giving a total of 183 data points. Control
subjects were nulliparous women of childbearing age
(mean age 27, range 18–45 years). Postnatal subjects
(mean age 32, range 19–46 years) were studied on the
first day (PN1; n ¼ 63); 8 weeks (PN2; n ¼ 55); 6
months (PN3; n ¼ 39) and 12 months (PN4; n ¼ 26)
after delivery. A small proportion of women (45)
included in the PN1–4 groups formed the longitudinal
part of the study and contributed 67 df to the trend
estimation and testing of the average rate of change
of RA thickness, shape variable (indirect width mea-
surement) CSA over a 12-month postpartum period
(Table 1). Due to problems imaging the IRD/DRA in a
No. of df Data points contributing to
the trend estimation
0 Single points contributing
nothing
29 Pairs contributing 1 df each
18 Triplets contributing 2 df
each
21 Quadruplets contributing 3
df each
68a
 ARTICLE IN PRESS
114 Y. Coldron et al. / Manual Therapy 13 (2008) 112–121
minority of subjects, the IRD was not measured on
every subject at every occasion so slightly fewer data
points were available for analysis (df ¼ 62).
2.3. Equipment
An Aloka SSD ultrasound scanner (Aloka Co. Ltd,
Mitaka-shi, Tokyo, Japan), with a 5 MHz linear probe
(11 cm footprint) was used. The image was frozen and
then downloaded to a computer using a frame grabber
and measured off-line with on-screen callipers using
UltraSound Image Concatenation and Analysis (USI-
CA) software (developed by the Department of Medical
Physics, St. George’s Hospital, London).
2.4. Procedure
With the subject in crook lying and knees flexed over
two pillows, measurements were taken of resting RA
thickness, CSA, shape of RA and the IRD. The bottom
edge of the probe was placed centrally on the skin just
cephalad to the umbilicus. To determine the IRD the
two medial ends of the recti were identified and two
images taken (Fig. 1). The transducer was then moved
laterally until the full image of the left RA was seen on
the screen (Fig. 2); two scans were taken and the
procedure repeated for the right RA. The mean of two
measurements was recorded. Mean values for right and
left RA were combined, the mean calculated and used
for analysis.
Cross-sectional area of rectus abdominis: The circum-
ference of RA was traced around the inside of its border
using the on-screen cursor and the CSA computed.
Fig. 1. Ultrasound images showing the inter-recti distance (IRD) and linea alba (LA) between the medial ends of the right and left rectus abdominis
(RA) muscles in cross section. (A) A control subject and with thick LA and an IRD of approximately 1 cm. (B) A postnatal subject with a thin LA
and an IRD of approximately 2.6 cm. *Medial borders of the left and right rectus abdominis.
Thickness of rectus abdominis: This was measured
vertically at the mid-point of the width of the belly,
between the inside edges of the superior and inferior
fascial borders.
Shape value (indirect width measurement) of RA: True
linear width (W) of postpartum RA was not always
possible to measure because the shape of some muscles
(particularly at Day 1) made it difficult to bisect the
muscle using on screen callipers; however, the shape of
the RA was approximately an ellipse (Fig. 2). The area
of an ellipse is calculated from the formula Area ¼ pAB
where A is the semi-major axis (width ¼ W) and B is the
semi-minor axis (thickness ¼ T) (Fig. 3). Consequently,
the width can be calculated if the CSA and thickness are
known and a shape variable (S) can be derived.
Therefore if A ¼ width/2 ¼ W/2 and B ¼ thickness/2
¼ T/2
pWT
Area ¼ pðW =2ÞðT=2Þ ¼ ¼ CSA
4
and
pW p W
CSA=T 2 ¼ ¼ .
4T 4T
Therefore S ¼ CSA/T2 is proportional to the ratio of
W/T and can be used to describe the cross-sectional
shape of a muscle. For the purposes of this paper, S has
been defined as the ‘shape variable’ and it can be seen
that the larger the shape variable, the wider and thinner
the muscle cross-section, in this case RA, would be. If
the CSA of RA were a true ellipse then S would be p/4
and the width would be CSA/(Tp/4), but since the cross-
section of the RA muscles scanned only approximated
an ellipse, the widths estimated this way are themselves
 ARTICLE IN PRESS
Y. Coldron et al. / Manual Therapy 13 (2008) 112–121 115

Fig. 2. Ultrasound images of the left rectus abdominis (RA) muscle showing the cross-sectional area in (A) a control subject—external oblique (EO)
internal oblique (IO) and transversus abdominis (TA) are seen on the right side of the image forming the aponeurosis around RA. (B) A postnatal
subject at Day 1—width of RA is almost that of the probe and the lateral abdominal muscles cannot be seen on the right side of the image. *Medial
and lateral borders of RA.

individuals, contributed to more than one group.

B
A Significance level for all tests was set at Po0.05.
Fig. 3. Ellipse with semi-major axis of length A and semi-minor axis of
length B. The area of an ellipse is calculated from the formula:
Area ¼ pAB, where A is the semi-major axis (width ¼ W) and B is the
semi-minor axis (thickness ¼ T). Consequently, the width can be
calculated if the CSA and thickness are known and a shape variable (S)
can be derived. Thus the formula S ¼ CSA/T2 is proportional to the
ratio of W/T and can be used to describe the cross-sectional shape of a
muscle. Therefore the larger the shape variable, the wider and thinner
the muscle (in this case rectus abdominis) would be.
Student’s unpaired two-tailed t-tests were used to
compare CSA, thickness, shape value and IRD between
the control group and each of the four postnatal groups.
Repeated measures: Because the pattern of change in
individuals over time may be distorted by plotting group
means, repeated measures analysis was used. This
included those who contributed at least two measure-
ments over time. The analysis consisted of fitting
individual regression line models to each individual’s
data, but constraining them to be parallel. This common
slope of the regression lines, indicating change in muscle
size over time (rate of change per month), was calculated
for RA CSA, thickness, shape (width) (df ¼ 67) and
IRD (df ¼ 62). The significance level for testing the
value of the gradient of the slope against zero was set at
Po0.05.

only approximate. Nonetheless the shape constant [S]

and the estimated widths were useful for establishing
3. Results
whether the relationship between width and thickness
altered during the postpartum period and to investigate
RA was significantly thinner, wider (had a higher
how they compared with controls over the postpartum
shape value) and had a larger IRD at 12 months
follow-up period.
postpartum compared with controls, whilst the CSA was
Inter-recti distance: The probe was placed over the
larger at Day 1 postpartum only, and was similar to that
midline of the abdomen with the lower border just
of controls after 8 weeks.
cephalad to the umbilicus. The medial ends of the two
RA muscle bellies were identified and the gap between
the two was measured. 3.1. Cross-sectional data

2.5. Data analysis 3.1.1. Cross-sectional area of RA

The CSA of Day 1 postnatal group (PN1) only was
Cross-sectional data: Due to the partial longitudinal significantly greater (Po0.0001) than that of controls
design of the study, ANOVAs were not performed on (CTL) (CTL: 5.2870.97 cm2, n ¼ 69; PN1: 6.557
data for the postnatal groups, as some but not all 1.43 cm2, n ¼ 63) (Table 2 and Fig. 4).
 ARTICLE IN PRESS
116 Y. Coldron et al. / Manual Therapy 13 (2008) 112–121

Table 2
Rectus abdominis thickness, cross-sectional area, shape ratio and inter-recti distance during the first postnatal year

Group n Thickness (mm) CSA (cm2) Shape value (arbitrary IRD (mm)
units)

CTL 69 9.77 (1.62) 5.28 (0.97) 5.74 (1.41) 11.17 (3.62)

PN1 63 8.09 (1.40)*** 6.55 (1.43)*** 10.38 (2.55)*** 42.22 (20.28)***
PN2 54 7.97 (1.44)*** 5.02 (1.05) 8.26 (2.19)*** 22.84 (9.44)***
PN3 39 8.44 (1.48)*** 4.93 (0.95) 7.25 (1.99)*** 20.73 (7.31)***
PN4 26 9.03 (1.64)* 5.09 (1.11) 6.44 (1.35)* 23.32 (8.38)***

Mean values (71 SD) for thickness, cross-sectional area (CSA), shape value (width) and inter-recti distance (IRD) of RA for control (CTL) and
postnatal (PN) groups. Differences between CTL and PN values were tested using two tailed unpaired t- tests. At 12 months postpartum RA was
significantly thinner and wider (had a larger shape value) and a larger IRD than controls but CSA was larger at Day 1 postpartum only.
*Significantly different from controls *Po0.05 ***Po0.0001.
Post-natal groups: PN1 ¼ Day 1; PN2 ¼ 8 weeks; PN3 ¼ 6 months; PN4 ¼ 12 months.

a b 8
11

10 7
Thickness (mm)

CSA (cm2)

9 6

8 5

7 4
CTL PN1 PN2 PN3 PN4 CTL PN1 PN2 PN3 PN4

c d
12 50

40
10
Shape value

IRD (mm)

30
8
20

6
10

4 0
CTL PN1 PN2 PN3 PN4 CTL PN1 PN2 PN3 PN4

Fig. 4. Cross-sectional data for rectus abdominis (RA) characteristics in control and postnatal groups at different times over the first postnatal year
(means795% confidence limits for: (a) thickness (mm), (b) cross-sectional area (CSA; cm2), (c) shape value (arbitrary units) (indirect width
measurement), (d) inter-recti distance (IRD; mm). Time postpartum—PN1 ¼ Day 1, PN2 ¼ 8 weeks, PN3 ¼ 6 months, PN4 ¼ 12 months.
*Significantly different from controls; ***Po0.0001; *Po0.05.
 ARTICLE IN PRESS
Y. Coldron et al. / Manual Therapy 13 (2008) 112–121 117

3.1.2. Thickness of RA 3.2.3. Shape value (width) of RA

In all four postnatal groups, RA was significantly
thinner than in controls (PN1, PN2, PN3 groups
Po0.0001; PN4 group P ¼ 0.0478; Table 2 and Fig. 4).
3.1.3. Shape value (width) of RA
RA became wider and thinner during pregnancy. The
mean value for shape (S) of the control group was
significantly smaller than in each of the four postnatal
groups (PN1, PN2, PN3 Po0.0001; PN4 P ¼ 0.0326;
Table 2 and Fig. 4).
3.1.4. Inter-recti distance
The IRD in all four postnatal groups was significantly
greater (Po0.0001) than in controls (Table 2 and
Fig. 4), suggesting that it did not return to normal
values by 12 months postpartum.
3.2. Longitudinal data: changes over 12 months
postpartum
3.2.1. Cross-sectional area of RA
Regression analysis revealed that the CSA became
significantly smaller during the 12 months postpartum
(P ¼ 0.0021) (Table 3) with strong evidence of curvature
(P ¼ 0.0001), i.e. the initial CSA values at Day 1
postpartum were greater than controls, followed by a
sharp decrease at 8 weeks, returning to similar values to
those of controls (Fig. 5).
3.2.2. Thickness of RA
Regression analysis demonstrated that RA became
steadily significantly thicker over the 12-month period
(P ¼ 0.0003; Table 3) but values did not return to those
of controls (Fig. 5).
Repeated measures analysis revealed that the shape
value (width) became significantly smaller (Po0.0001)
during the 12 months postpartum (Table 3) with strong
evidence of curvature (P ¼ 0.0003), i.e. the initial width
of RA at Day 1 postpartum was greater than controls
followed by a steady decrease at 8 weeks postpartum
that flattened off by 6–12 months and did not return to
those of controls at 12 months (Fig. 5).
3.2.4. Inter-recti distance
Linear regression analysis showed that the IRD
became significantly smaller during the 12 months
postpartum (P ¼ 0.0002) (Table 3) with strong evidence
of curvature (P ¼ 0.0003), i.e. the initial IRD values at
Day 1 postpartum were greater than those of controls
with a sharp decrease in values at 8 weeks postpartum.
These values plateaued at 8 weeks and did not return to
those of controls at 12 months (Fig. 5).
4. Discussion
Alteration in RA thickness, shape (width) and IRD
occurred during the study period and none of these
variables had returned to control values by 12 months
postpartum. Only the CSA had returned to normal over
an 8-week period.
4.1. Change in shape of rectus abdominis
Maintenance of a thinner and wider RA muscle
postpartum suggests that stretch-induced changes occur
during pregnancy in the contractile and connective
tissue components of the muscle belly, fascial aponeuro-
sis surrounding RA and the underlying transversalis
fascia.

Table 3
Repeated measures of rectus abdominis thickness, shape ratio and inter-recti distance

Muscle Units Gradient (b) 95% CLs t P

characteristic estimated change
per month
Low High

Thicknessa mm 0.08651 0.04164 0.1314 3.78 0.0003b

CSAa cm2 0.08237 0.1328 0.03191 3.20 0.0021b
Shapea (width) arbitrary units 0.288 0.370 0.206 6.92 0.0001b
IRDc mm 1.458 2.193 0.7233 3.89 0.0002b

Rate of change per month (plus 95% confidence limits—CLs) of RA thickness, cross-sectional area (CSA), shape value (width) and inter-recti
distance (IRD) during the 12 months post-partum. RA became significantly thicker and narrower during the 12-month period whereas the CSA and
IRD became significantly smaller. These changes per month were all significantly different from zero. The t values from unpaired t-tests are shown
with significance (P) values.
n ¼ 45.
a
df ¼ 67.
b
Gradient significantly different from zero.
c
df ¼ 62.
 ARTICLE IN PRESS
118 Y. Coldron et al. / Manual Therapy 13 (2008) 112–121

a 16 b 12

14
10
12
*** ***
Thickness (mm)

RA CSA (cm2)
10

8 6

6
4
4
2
2

0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12
Time since delivery (months) Time since delivery (months)

c d 140
20
Shape Value/Width (Arbitrary units)

18 120
16
100
14 ***
IRD (mm)

12 80 ***

10
60
8
6 40
4
20
2
0 0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Time since delivery (months) Time since delivery (months)

Fig. 5. Repeated measures results for rectus abdominis showing changes in thickness, cross-sectional area (CSA), shape value (width) and inter-recti
distance (IRD). The data set in black on the left of the plots is that of the control group, the remaining four data sets in grey are those of the four
postnatal groups. a ¼ RA thickness. The bold extended line is the fitted line for a representative subject to illustrate the average slope obtained from
least-squares regression analysis fitting separate but parallel straight lines to every subject. Since every subject was allowed their own intercept, their
trend line equations were different even though the slope was same. The general formula is: Muscle thickness ¼ individual starting value+b.time in
months since delivery, where b is the same for every subject (b ¼ beta). No evidence of a significant curvilinear tend was seen. ***Slope significantly
different from zero Po0.0001. b ¼ CSA, c ¼ shape value (width) and d ¼ IRD. The bold curved line on the plots is the fitted line for a representative
subject to illustrate the average-fitted curved trend from least-squares regression analysis fitting separate but parallel curved lines to every subject.
The general formula is: muscle variable (CSA, Shape or IRD) ¼ individual starting time+b1.time+b2.time2 (where time is time in months since
delivery). ***Significant curve Po0.0001.

There could be a selective hypertrophy of Type I Prolonged stretch is known to increase the number of
fibres in response to stretch, as suggested by animal sarcomeres in series at the ends of muscle (Williams and
studies. Prolonged muscle stretch in rats produced Goldspink, 1978) with increased protein synthesis and
activation of slow genes and repression of fast Type addition of myotubes (Dix and Eisenberg, 1990),
IIx genes (Goldspink et al., 1991). In the RA of pregnant resulting in an increase in total muscle volume. It is
rats the diameter of slow Type I fibres increased in the possible that this mechanism occurs in postnatal women
latter half of pregnancy while that of the fast glycolitic and the addition of sarcomeres in series to RA could
fibres (Type IIX) decreased, with no change to the Type reduce the ability of postpartum women to perform
IIA fibres (Martin, 1979). In contrast, both Types I and trunk flexion and hold an inner range contraction. The
II fibres increased in diameter in porcine and rabbit combination of fibre type and architectural changes may
studies, with a concomitant increase in the number of also reduce the ability of RA to produce active tension.
Type I fibres and decrease in Type II fibres (Lalatta et However following damage or exercise, skeletal muscle
al., 1987, 1988). Since these studies were performed on is able to repair itself by replenishing cell nuclei from
different animal species, it is not known how results satellite cells stimulated by the action of a local tissue
translate to humans but it seems that alterations in RA repair mechanism, the mechanogrowth factor (Gold-
muscle fibre types might occur during pregnancy. spink, 2006). The effects of stretch on postnatal
 ARTICLE IN PRESS
Y. Coldron et al. / Manual Therapy 13 (2008) 112–121
abdominal muscle have not yet been investigated
histochemically and would be the subject of further
study.
Anecdotally, many women reported feeling a ‘‘lack of
support’’ in their abdomen. This could be caused by a
change in fascia as well as muscle. Prolonged stretch of
the tibialis anterior in rats produced increased deposi-
tion of collagen Type III, damage to the perimysial and
endomysial network and indicated that intramuscular
connective tissue did not adapt to stretch as well as the
contractile components (Williams et al., 1998). Further
study into the effects of prolonged stretch on different
types of connective tissue in pregnancy and factors
affecting recovery are indicated.
Measurement of the width of RA on the ultrasound
image in control subjects using on-line callipers was
straightforward as the muscle cross-section is approxi-
mately elliptical. Thus, the approximate ratio of width
to thickness was easy to calculate. As the RA appeared
to be wider and thinner in most Day 1 postpartum
subjects and it was not known whether this dispropor-
tion continued further into the postpartum period, it
was decided to explore this width/thickness ratio
further. However, it was not always possible to bisect
the RA and establish width because its shape was not so
regular. An approximate estimate of width was therefore
derived from the formula for establishing the area of an
ellipse (see Fig. 3). This approach to estimation of width
to thickness ratio has not been used in any studies on
postnatal RA muscles and from the present results it is
recommended that the formula be used in future studies
of this subject group when width cannot be measured
accurately.
4.2. Cross-sectional area of RA
As the CSA was larger than in controls in Day 1
postpartum subjects, it could be assumed that there was
hypertrophy of RA during pregnancy, perhaps in
response to bearing an increased load. A healthy muscle
would show low echogenicity on ultrasound imaging,
i.e. it would be darker (Stokes et al., 1997). Increased
echogenicity of the contractile parts of muscle has been
found in patients with age-related deterioration (Maur-
its et al., 2003), denervated muscles (Gunreben and
Bogdahn, 1991) or myopathic muscles (Lamminen,
1991; Udd et al., 1991; Maurits et al., 2003), whereas
with neuropathic muscles there is inhomogeneity due to
pathological disruptions of muscle architecture (Zuberi
et al., 1999; Maurits et al., 2003). In the current study,
observations on RA appearance were not analysed
quantitatively, but visual assessment suggested that
some postnatal RA muscles showed inhomogeneity
and areas with some increased echogenicity. High
echogenicity has been attributed to infiltration of non-
contractile tissue such as fat or collagen (Reimers et al.,
119
1993; Elliott et al., 2005). It is not possible to determine
whether there was any neuropathy present in RA but it
is interesting to note that it has been suggested that
stretch-induced denervation of the abdominal wall may
occur during pregnancy and delivery (Stelzner et al.,
1993). Therefore increased CSA at Day 1 postpartum
may reflect replacement of muscle tissue with fat and
connective tissue rather than hypertrophy of muscle
fibres.
4.3. Implications for muscle strength
Long-term change in muscle thickness and width may
affect the strength of RA but strength of individual
abdominal muscles cannot be measured directly. Altera-
tion in the ratio between thickness and width of the
quadriceps in patients with myositis was associated with
abnormal muscle properties and a decrease in muscle
force (Chi-Fishman et al., 2004). This finding could be
extrapolated to the current study, implying that altera-
tion in thickness and width postnatally could result in a
reduced force generation capacity.
Trunk flexion exercises are known to recruit RA
(Sarti et al., 1996; Negrao Filho et al., 2003; Clark et al.,
2003) and an attempt to perform a trunk curl at 8 weeks
after delivery was reported to be successful or moder-
ately successful in 5 out of 6 women (Gilleard and
Brown, 1996) and successful in 80% at 6 weeks (Spence,
1978). However, these two studies used qualitative
measurement tools and a robust study of postpartum
trunk flexor strength is required.
Using isokinetic dynamometry, Potter et al. (1997b)
found that concentric and eccentric trunk flexion
strength was significantly lower in women at 24 weeks
postpartum than in controls. The reduction in RA
thickness found in the current study indicates a decrease
in strength. If this were the case, study of the effect of
postnatal exercise programmes that do not load the
spine would be informative.
Repeated measures analysis demonstrated that RA
recovery of CSA, shape and thickness continued during
the 12-month postpartum period. Further study of the
time for full recovery of thickness and shape postpartum
is indicated.
4.4. Inter-recti distance
The IRD postpartum was significantly wider than in
controls in all four postnatal groups but most recovery
occurred between Day 1 and 8 weeks, when the IRD
reached a plateau. The mean IRD at 12 months was
22.3 mm, but measurements ranged from 10.2 to
42.1 mm with a third of subjects presenting with a wider
gap than the mean.
Definitions of the width of a postnatal diastasis of RA
(DRA) at the umbilicus using external skin markers
 ARTICLE IN PRESS
120 Y. Coldron et al. / Manual Therapy 13 (2008) 112–121
vary. It has been defined as a width of 41.5 cm (Gilleard
and Brown, 1996), 42 cm (Potter et al., 1997a; Lo et al.,
1999), 42.5 cm (Candido et al., 2005) or 42 finger
widths during a partial sit-up (Bursch, 1987; Sheppard,
1996). DRA has been noted by the 26th week of
gestation (Gilleard and Brown, 1996) and in 66% of
women during the third trimester (Boissonnault &
Blaschak). Previous studies have noted that there is a
partial resolution of a DRA by 4 weeks (Gilleard and
Brown, 1996) and 8 weeks (Boissonnault and Blaschak,
1988) postpartum and these findings agree with those of
the current study. Risk factors have been found to
include multiparity, maternal age (434 years), larger
babies, greater weight gain, caesarean section and
multiple gestation (Lo et al., 1999). This has been
disputed by Candido et al. (2005) who found that
women with a diastasis were more likely to be providing
child care (although the nature of child care was not
specified) and there was an association with Caucasian
ethnicity and a lack of regular exercise during preg-
nancy.
It is suggested that those with DRA may have a
reduced capacity for force generation. In support of this
hypothesis, in a small study of six subjects at 8 weeks
postpartum, subjects with a DRA of 435 mm were
significantly less successful at performing a pelvic tilt or
a trunk curl than those without a DRA (Gilleard and
Brown, 1996).
Together with the persistence of decreased thickness
and increased width of RA plus potential changes to the
fascia surrounding RA at 12 months postpartum; an
increased IRD/DRA may lead to decreased anterior
abdominal wall stiffness and a possible mechanical
disadvantage. These changes in passive stiffness might
lead to muscle imbalance and thus poor co-ordination
with the other three abdominal muscles due to their
connections to RA via the surrounding aponeurosis and
the linea alba. These findings have implications for
postnatal exercises and the identification of factors that
reduce a persistent DRA, increase RA thickness and
reduce RA width would be valuable.
4.5. Limitations of the study
The main limitation of the study was the discrepancy
between cross-sectional and longitudinal subject num-
bers. Given the time frame of this project, a full
longitudinal study could not be undertaken. In addition,
a proportion of the women became pregnant again in
less than 1 year, some moved from the area and others
returned to work, hence many could not attend a follow-
up appointment. As the results from the limited long-
itudinal data showed change in characteristics of RA
during the 12-month postnatal period, a further long-
itudinal study using greater numbers is indicated.
5. Conclusions
Characteristics of RA thickness and width, and the
IRD had not returned to normal values by 12 months
postpartum. A thinner, wider and longer RA has
implications for strength and fascial support. Persistent
increased IRD may cause decreased stiffness of the
anterior abdominal wall and predispose to a mechanical
disadvantage. Studies of the postpartum histochemical
and architectural changes, trunk flexion tension and
production would inform the development of effective
postnatal exercise programmes. Exercises that target the
return of normal IRD, RA width, thickness and length
without loading and compressing the lumbar spine are
required.
Acknowledgements
The authors thank the subjects who took part in the
study, Dr. Anthony Swan for statistical advice, Dr.
Basky Thilaganathan for use of facilities in the Fetal
Medicine Unit at St. George’s Hospital, London and the
Neuro-disability Research Trust for financial support.
Part of this work was undertaken at the Royal Hospital
for Neuro-disability, which received a proportion of its
funding from the NHS Executive; the views expressed in
this publication are those of the authors and not
necessarily those of the NHS Executive.
References
Boissonnault JS, Blaschak MJ. Incidence of diastasis recti abdominis
during the childbearing year. Physical Therapy 1988;68(7):1082–6.
Bottinelli R, Pellegrino MA, Canepari M, Rossi R, Reggiani C.
Specific contributions of various muscle fibre types to human
muscle performance: an in vitro study. Journal of Electromyo-
graphy and Kinesiology 1999;9:87–95.
Bursch SG. Interrater reliability of diastasis recti abdominis measure-
ment. Physical Therapy 1987;67(7):1077–9.
Caix M, Outrequin G, Descottes B, Kalfon M, Pouget X. The muscles
of the abdominal wall: a new functional approach with anatomo-
clinical deductions. Anatomia Clinica 1984;6:101–8.
Candido G, Lo T, Janssen PA. Risk factors for diastasis of the recti
abdominis. Journal of the Association of Chartered Physiothera-
pists in Women’s Health 2005;97(Autumn):49–54.
Chi-Fishman G, Hicks JE, Cintas HM, Sonies BC, Gerber LH.
Ultrasound imaging distinguishes between normal and weak
muscle. Archives of Physical Medicine and Rehabilitation
2004;85(6):980–6.
Clark KM, Holt LE, Sinyard J. Electromyographic comparison of the
upper and lower rectus abdominis during abdominal exercises.
Journal of Strength and Conditioning Research 2003;17(3):475–83.
Dix DJ, Eisenberg BR. Myosin mRNA accumulation and myofi-
brillogenesis at the myotendinous junction of stretched muscle
fibers. Journal of Cell Biology 1990;111(5, Part 1):1885–94.
Elliott JM, Galloway GJ, Jull GA, et al. Magnetic resonance imaging
analysis of the upper cervical spine extensor musculature in an
asymptomatic cohort: an index of fat within muscle. Clinical
Radiology 2005;60:355–63.
 ARTICLE IN PRESS
Y. Coldron et al. / Manual Therapy 13 (2008) 112–121
Fast A, Weiss L, Ducommun EJ, Medina E, Butler JG. Low-back pain
in pregnancy: abdominal muscles, sit-up performance, and back
pain. Spine 1990;15(1):28–30.
Gilleard WL, Brown JM. Structure and function of the abdominal
muscles in primigravid subjects during pregnancy and the
immediate postbirth period. Physical Therapy 1996;76(7):750–62.
Goldspink G. Impairment of IGF-I gene splicing and MGF expression
associated with muscle wasting. The International Journal of
Biochemistry and Cell Biology 2006;38(3):481–9.
Goldspink G, Scutt A, Martindale J, Jaenicke T, Turay L, Gerlach
GF. Stretch and force generation induce rapid hypertrophy and
myosin isoform gene switching in adult skeletal muscle. Biochem-
istry Society Transactions 1991;19(2):368–73.
Graziotti GH, Rios CM, Rivero JL. Evidence for three fast myosin
heavy chain isoforms in type II skeletal muscle fibers in the adult
llama (Lama glama). Journal of Histochemistry and Cytochemistry
2001;49:1033–44.
Gunreben G, Bogdahn U. Real-time sonography of acute and chronic
muscle denervation. Muscle Nerve 1991;14(7):654–64.
Haggmark T, Thorstensson A. Fibre types in human abdo-
minal muscles. Acta Physiologica Scandanavica 1979;107(4):
319–25.
Kendall FP, McCreary E, Provance D. Muscles: testing and function,
4th ed. Baltimore, USA: Williams & Wilkins; 1993. pp. 131–176
[Chapter 6].
Lalatta CG, Barazzoni AM, Lucchi ML, Bortolami R. Histochemical
types and sizes of fibers in the rectus abdominis muscle of guinea
pig: adaptive response to pregnancy. Anatomy Record
1987;217(1):23–9.
Lalatta CG, Barazzoni AM, Lucchi ML, Bortolami R. Fiber sizes and
histochemical characteristics of the rectus abdominis muscle of the
rabbit under conditions of pregnancy and mechanically induced
stress. Anatomy Record 1988;222(2):136–44.
Lamminen A. Low field magnetic resonance imaging and ultrasono-
graphy in muscle diseases. Acta Radiologica Supplement
1991;377:42–5.
Lo T, Candido G, Janssen P. Diastasis of the recti abdominis in
pregnancy: risk factors and treatment. Physiotherapy Canada
1999;44:32–7.
Mantle MJ, Greenwood RM, Currey HL. Backache in pregnancy.
Rheumatology and Rehabilitation 1977;16(2):95–101.
Martin WD. A study of the effect of pregnancy on muscle fibers of the
rectus abdominis muscle of the rat. Anatomical Record
1979;195(3):455–62.
Maurits NM, Bollen AE, Windhausen A, De Jager AE, Van Der
Hoeven JH. Muscle ultrasound analysis: normal values and
differentiation between myopathies and neuropathies. Ultrasound
in Medicine and Biology 2003;29(2):215–25.
Nahas FX, Ferreira LM, Augusto SM, Ghelfond C. Long-term
follow-up of correction of rectus diastasis. Plastic and Reconstruc-
tive Surgery 2005;115(6):1736–41.
Negrao Filho RF, Berzin F, Souza GC. Electromyography study of
the portions of the abdominal rectus muscle. Electromyography
and Clinical Neurophysiology 1997;37(8):491–501.
Negrao Filho RF, Berzin F, Souza GC. Quantitative end qualitative
analysis of the electrical activity of rectus abdominis muscle
portions. Electromyography and Clinical Neurophysiology
2003;43(5):305–14.
Noren L, Ostgaard S, Johansson G, Ostgaard HC. Lumbar back and
posterior pelvic pain during pregnancy: a 3-year follow-up.
European Spine Journal 2002;11(3):267–71.
Ostgaard HC. Lumbar and pelvic pain in pregnancy. In: Vleeming A,
Mooney V, Dorman T, Snijders C, Stoeckart R, editors. Move-
ment, stability and low back pain. Edinburgh: Churchill Living-
stone; 1997. p. 411–20 [Chapter 33].
121
Ostgaard HC, Andersson GB. Postpartum low-back pain. Spine
1992;17(1):53–5.
Ostgaard HC, Andersson GB, Karlsson K. Prevalence of back pain in
pregnancy. Spine 1991;16(5):549–52.
Ostgaard HC, Roos-Hansson E, Zetherstrom G. Regression of back
and posterior pelvic pain after pregnancy. Spine 1996;21(23):
2777–80.
Ostgaard HC, Zetherstrom G, Roos-Hansson E. Back pain in relation
to pregnancy: a 6-year follow-up. Spine 1997;22(24):2945–50.
Polden M, Mantle J. Physiotherapy in obstetrics and gynaecology.
Oxford: Butterworth Heinemann; 1990. p. 222–80 [Chapter 7].
Potter HM, Downey JL, Jones ST. Effect of an Intense training
programme for the deep antero-lateral abdominal muscles on
rectus abdominis diastasis: a single case study. In: Proceedings of
the 10 Biennial Conference of the Manipulative Physiotherapists
Association of Australia (MPAA). Victoria, Australia: MPAA;
1997a. p. 153–5.
Potter HM, Randall HF, Strauss GR. Effect of pregnancy and
motherhood on trunk muscle strength: an examination of
isokinetic trunk strength at 24 weeks postpartum. In: Proceedings
of the 10 Biennial Conference of the Manipulative Physiotherapists
Association of Australia (MPAA). Victoria, Australia: MPAA;
1997b. p. 151–2.
Rath AM, Attali P, Dumas JL, Goldlust D, Zhang J, Chevrel JP. The
abdominal linea alba: an anatomo-radiologic and biomechanical
study. Surgical and Radiological Anatomy 1996;18(4):281–8.
Reimers K, Reimers CD, Wagner S, Paetzke I, Pongratz DE. Skeletal
muscle sonography: a correlative study of echogenicity and
morphology. Journal of Ultrasound Medicine 1993;12(2):73–7.
Sarti MA, Monfort M, Fuster MA, Villaplana LA. Muscle activity in
upper and lower rectus abdominus during abdominal exercises.
Archives of Physical Medicine and Rehabilitation
1996;77(12):1293–7.
Sheppard S. The role of transversus abdominis in postpartum
correction of gross divarication recti. Manual Therapy
1996;1(4):214–6.
Spence M. Postnatal survey. Australian Journal of Physiotherapy
1978;34:151–60.
Stelzner F, Beyenburg S, Hahn N. [Acquired disorders of peritoneal
cavity muscles. Abdominal wall denervation in pregnancy,
denervation incontinence, and continent and incontinent constipa-
tion]. Langenbecks Archiv fuer Chirugie 1993;378(1):49–59.
Stokes MJ, Hides JA, Nassiri DK. Musculoskeletal ultrasound
imaging; diagnostic and trreatment aid in rehabilitation. Physical
Therapy Reviews 1997;2:73–92.
Udd B, Lamminen A, Somer H. Imaging methods reveal unexpected
patchy lesions in late onset distal myopathy. Neuromuscular
Disorders 1991;1(4):279–85.
van Uchelen JH, Kon M, Werker PM. The long-term durability of
plication of the anterior rectus sheath assessed by ultrasonography.
Plastic and Reconstructive Surgery 2001;107(6):1578–84.
Williams PE, Goldspink G. Changes in sarcomere length and
physiological properties in immobilized muscle. Journal of
Anatomy 1978;127(3):459–68.
Williams PL, Warwick R, Dyson M, Bannister LH. Gray’s anatomy,
37th ed. Edinburgh: Churchill Livingstone; 1989. p. 546–69
[Chapter 5].
Williams P, Kyberd P, Simpson H, Kenwright J, Goldspink G. The
morphological basis of increased stiffness of rabbit tibialis anterior
muscles during surgical limb-lengthening. Journal of Anatomy
1998;193(Pt 1):131–8.
Zuberi SM, Matta N, Nawaz S, Stephenson JB, McWilliam RC,
Hollman A. Muscle ultrasound in the assessment of suspected
neuromuscular disease in childhood. Neuromuscular Disorders
1999;9(4):203–7.