Comparison of the Effects of Exercise in Water and

on Land on the Rehabilitation of Patients With

Intra-articular Anterior Cruciate Ligament
Reconstructions
Brian J Tovin, Steven L Wolf, Bruce H Greenfield, Jeri
Crouse and Blane A Woodfin
PHYS THER. 1994; 74:710-719.

The online version of this article, along with updated information and services, can be
found online at: http://ptjournal.apta.org/content/74/8/710
Collections

This article, along with others on similar topics, appears

in the following collection(s):
Injuries and Conditions: Knee
Injuries and Conditions: Lower Extremity
Therapeutic Exercise

e-Letters

To submit an e-Letter on this article, click here or click on

"Submit a response" in the right-hand menu under
"Responses" in the online version of this article.

E-mail alerts

Sign up here to receive free e-mail alerts

Correction

A correction has been published for this article. The

correction has been appended to this PDF. The
correction is also available online at:
http://ptjournal.apta.org/content/74/12/1165.full.pdf

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

Research Report

Comparison of the Effects of Exercise in Water and

on Land on the Rehabilitation of Patients With
Intra-articular Anterior Cruciate Ligament
Reconstructions

Brlan J Tovln
Steven L Wolf
Bruce H Greenfield
Jerl Crouse
Blane A Woodfin

Background and Purpose. Exercises in water have been shown to be effective

for improving strength and passive range of motion (PROM). Traditional rehabilitation following intra-articular anterior cruciate ligament (ACL) reconstruction
has taken place o n land. This study was designed to compare the effects of exercises in water o n strengh and girth of the thgb musculature, knee PROM, joint
laxity, e m i o n , and functional outcome with the effects of similar mercises o n
land in subjectsfollowing intra-articular reconstruction of the ACL. Subjects.
Twenty subjects were randomly asstgned to either a group that exercised o n land
or a group that exercised i n water. Metbods. Thigh girth, joint effm'on, and knee
PROM measurements were recorded at 2-week intervalsfor the first 8 weeks postoperatively. Isokinetic and isometric peak torque measurementsfor the thigh musculature, knee joint laxity assessments, and Lysholm scores were obtained at the
end of 8 weeks. Results. Higher outcome scores were recorded in the water
group than in the landgroup, as measured by Lysholm scales. No dtferences were
noted between groups for knee PROM, thigh girth, or quadriceps femoris muscle
p e r f o m w e . In the water group, lessjoint e f m o n was noted aJer the 8 weeks.
In the land group, greater peak torquefor isokinetic knee flexion was recorded.
Concluston and Discussion. Although exercise in water may not be as effective as exercise o n land for regaining maximum muscle perfomnce, rehabilitation in water may minimize the amount of joint effiion and lead to greater
self-reportsof functional improvement in subjects with intra-articular ACL reconstructions. [Tovin BJ, Wolf SL, Greenfield BH, et al. Comparison of the effects of
exercise in water and o n land o n the rehabilitation of patients with intraarticular anterior cruciate ligament reconstmctions. Phys Ther.
199g 74:710-719.1

Key Words: Knee; Ligaments; Lower extremity, knee; Muscle pe$omance, lower
extremity; Rehabilitation.

Rehabilitation following anterior cruciate ligament (ACL) reconstruction

has evolved over the past few decades
and is considered important in guaranteeing a beneficial outcome following surgety.1 Advances in surgical
approaches, such as graft placement
and graft fixation, and the use of arthroscopically assisted procedures

22/710

have influenced rehabilitation, as have

knowledge of stress-strain patterns in
the ACL during various exercises.2
Twelve-month protocols requiring
immobilization and non-weight bearing3 have given way to accelerated
protocols permitting immediate
weight bearing, no immobilization,
and return to activity within 6

months.2 Primary goals continue to be

the recovery of joint range of motion
(ROM), quadriceps femoris muscle
force-generating capability, and ambulatory skills.2 Attaining these goals,
however, may be delayed by postoperative joint effusion and the persistence of pain. Early phases of rehabilitation must minimize the deleterious

Physical Therapy/Volume 74, Number B/August 1994

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

effects of surgery through ROM and

muscle strengthening exercises while
ensuring that each activity is performed without overstressing the ACL
grafts.435
Electromyographic biofeedback5j6 and
neuromuscular electrical stimulation7.8
are two modalities used in the early
phases of rehabilitation following ACL
reconstruction to reduce muscle atrophy and to facilitate strengthening.
The effectiveness of these modalities
in imprc~vingquadriceps femoris
muscle force in subjects with ACL
reconstruction has been measured by
isokinetic dynomometry.6~7These
studies, however, applied feedback o r
neuroml~scularelectrical stimulation
during isometric quadriceps femoris
exercises, and this approach may not
simulate functional activities.
We believe knee extension exercises
should be designed to simulate functional activities. "Closed-chain" knee
extension has been advocated as a
safe exercise for patients after ACL
reconstruction.9 These exercises involve applying resistance through the
terminal joint of a limb segment,
which restrains the joint's free movement (eg, rising from a chair),
whereas "openchain" exercises involve applying resistance to an extremity in a way that the distal joint is
free to move (eg, kicking into the
air).IO Although both of these forms of
exercise can address the physical

impairments of patients following ACL

reconstruction, research suggests that
closed-chain exercises are safer than
open-chain exercises because there is
less stress on the graft.11-13 Despite
this fact, some subjects experience
increased pain and knee effusion
following closed-chain exercises.14
Therefore, performing closed-chain
exercises in an environment in which
the forces around the knee joint are
reduced may aid in reducing knee
pain and joint effusion.
Fkercises in water could expedite
rehabilitation because of the decreased stress on the joints, improved
circulation, and facilitated movement
that occur in water.15.16 Researchers
have analyzed limb movement in
water17-'9 and have compared different aquatic exercise devices,2&22but
few studies have quantified gains in
muscular force that occur following
an aquatic exercise program. Bartow
and Diamond23 have concluded that
exercises performed using water as
resistance can increase the torquegenerating capabilities of the thigh
musculature in healthy subjects.
Gehlsen et a124 have made similar
conclusions in patients with multiple
sclerosis, but no control group was
used for comparison.
NapoletanZ5found that in subjects
with ACL reconstructions, underwater
treadmill ambulation in conjunction
with traditional rehabilitation was

BJ Tovin, IT, ATC, is StafT Physical Therapist, Physiotherapy Associates, 2770 Lenox Rd NE, Ste 102,
Atlanta, GA 30324 (USA), and Director of Rehabilitation, Georgia Tech Athletic Association, Atlanta,
GA 30332. Mr Tovin was a student at Emory University, Atlanta, GA, at the time this study was completed in partial fulfillment of the requirements for his Master of Medical Science degree. Address
all correspondence to Mr Tovin.
SL Wolf, PhD, FT,FAlTA, is Professor and Director of Research, Department of Rehauilitation Medicine, Professor, Division of Geriatrics, Department of Internal Medicine, and Associate Professor,
Department of Anatomy and Cell Biology, Emory University School of Medicine, 1441 Clifton Rd
NE, Atlanta, GA 30322.
BH Greenfield, FT,OCS, is Clinical Coordinator of Education and Clinic Director, Physiotherapy Associates, Jonestmro, GA 30236, and Clinical Instructor, Division of Physical Therapy, Emory University.
J Crouse, IT, is Clinical Coordinator of Physical Therapy, HealthSouth, Atlanta, GA 30342.
BA Woodfin, MD, is Orthopaedic Surgeon, Resurgeons Orthopaedics, and Team Physician, Georgia
Tech Athletic Association.
This study was approved by the Human Investigation Committee of Emory University and Piedmont Hospital.

This article was submitted April 13, 1993, and was accepted Januaty 6, 194.

more effective in retarding thigh atrophy than traditional rehabilitation

alone. Thigh atrophy, however, is only
one measure of recovery. Whether
rehabilitation in water will be different from traditional rehabilitation in
reducing knee joint laxity, enhancing
muscle force, and improving functional outcomes in subjects with intraarticular ACL reconstructions is
uncertain.
The purpose of this study was to
determine whether exercises in a
pool will lead to less joint effusion,
less thigh atrophy, increased ROM
and thigh musculature strength, and
less difficulty with activities of daily
living in patients after intra-articular
ACL reconstruction compared with
exercises on land. An effort was made
to match specific exercises in both
groups so that each program was
identical and only the rehabilitation
environment was manipulated.

Method
Subjects
Twenty subjects (14 male, 6 female)
ranging in age from 16 to 44 years
@=29.0, SD=7.8) participated in this
study. All subjects had undergone
arthroscopically assisted intra-articular
ACL reconstruction using a bonepatellar tendon-bone autograft, performed by the same orthopedic surgeon. Subjects who had prior ACL
surgery to either knee or who had a
meniscus repair at the time of surgery
were excluded from the study.

Procedure
During the preoperative visit, subjects
were familiarized with the study and
postoperative rehabilitation protocols
were explained. Each subject signed
an informed consent statement, written to conform with the guidelines of
Emory University and Piedmont Hospital (Atlanta, Ga), and a questionnaire
was administered. Subjects were assigned to either a traditional rehabilitation (TR) group or a pool rehabilitation (PR) group using the following
method of group assignment. The first
2 subjects were randomly assigned to

Physical Therapy /Volume 74, Number 8/August 1994

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

Table 1 . Rehabilitation Pmgrams

--

Week 1 and Home Program Exercises (Both Groups)

1. Wall slides: 25 repetitions
2. Active-assistive range of motion: 25 repetitions
3. Passive knee extension: 10 minutes

4. Hamstring muscle and calf stretching: 10 minutes each

5. Quadriceps femoris muscle sets
6. Straight leg raisesa: 3 sets x 10 repetitions for hip flexion, abduction, adduction, and extension
7. Active knee flexiona: 3 sets x 10 repetitions

8. Toe raises: 3 sets x 10 repetitions

9. Partial wall squats (usually added to the home program after first week): 3 sets x 10 repetitions

Week 2-8 Exercise Programs

Traditional Rehabilitation Group

Pool Rehabilitation Group

1. Stationary cycling: 10 minutes

1. Stationary cycling: 10 minutesb

2. Gait training without brace, alternating forward and backward

ambulation: 10 min

2. Gait training without brace, alternating forward and backward

ambulation: 10 min

3. Side step-ups, front step-ups, step-downs: beginning with 3 sets

of 10 repetitions, progressing to 3 sets of 15 repetitions

3. Side step-ups, front step-ups, step-downs: beginning with 3 sets of

10 repetitions, progressing to 3 sets of 15 repetitionsC

4. Hip flexion, extension, abduction, adduction in standing using a

4. Hip flexion, extension, abduction, adduction in standing using the
wall pulley with 4.54-kg (10-lb) plates: beginning with 3 sets of 10
Hydrotone resistance boot: beginning with 3 sets of 10 repetitions
repetitions, progressing to 3 sets of 15 repetitions
and progressing to 3 sets of 15 repetitions
5. Knee flexion in sitting: 3 sets of 10 repetitions; boot: beginning
5. Knee flexicn in standing using the Hydrotone resistance boot:
with 3 sets of 10 repetitions, progressing to 3 sets of 15 repetitions
beginning with 3 sets of 10 repetitions and progressing to 3 sets of
15 repetitions

"Cuff weights were added to straight leg raises and knee flexion in increments of 0.91 kg (2 lb).
*stationary cycling in the pool rehabilitation group used a peddling device (see Fig. 1) rather than a stationary bicycle.
'Step-ups in the water were done with 20.32-cm (8-in) and 40.64-cm (16-in) steps.

one of the two groups using a coin

toss. The next 2 subjects recruited
were placed in opposite groups of the
first 2 subjects. This procedure was
continued for every 4 subjects until
20 subjects were recruited. As a result,
6 men and 4 women were placed in
the PR group and 8 men and 2
women were placed in the TR group.
This method of group assignment was
used to evenly distribute subjects
between the two groups over time,
while also incorporating random
assignment to groups.

Week 1 Exercises for Both

Oroups
During the first postoperative session,
patients in both groups were instructed in an identical program (Tab.
I), which they performed at home
twice per day. The first week of postoperative rehabilitation consisted of
three o r four treatment sessions in
24/712

which one of the authors reviewed

the home program to ensure that the
exercises were done safely and independently. To facilitate passive knee
extension, each subject was positioned prone and the involved leg
(from the superior third of the tibia
to the foot) was placed off the side of
a treatment table or bed, letting gravity pull the knee into extension. Resistance for the straight leg raises and
leg curls was added using variableresistance cuff weights. Subjects initiated each exercise, performing three
sets of 10 repetitions without weight
and progressing until they could perform three sets of 15 repetitions without difficulty. Subjects then added 0.9
kg (2 lb) to the cuff weight and repeated the progression starting with
three sets of 10 repetitions. This procedure was continued, and resistance
was added in 0.9-kg increments (most
patients progressed their weight every
2-3 days). Subjects were instructed

how to keep a log of their home

exercise p r o g m , which was checked
by one of the authors to help assess
compliance.

Weight Bearing
Gait training was also initiated on the
first postoperative session with axillary crutches and a hinged knee
brace. The braces were locked in full
extension for the first 4 to 7 days, and
subjects were instructed to bear as
much weight as they could tolerate.
Subjects were progressed from two
crutches to one crutch between the
4th and 7th postoperative days and
were usually off the crutch by the
10th postoperative day. The hinged
knee brace was unlocked at the beginning of the 2nd week, permitting
90 degrees of knee flexion. The ROM
of the braces were increased to 120
degrees by the beginning of the 3rd
postoperative week, and subjects were

Physical Therapy /Volume 74, Number 8/August 1994

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

achieved a maximum height of 30.48

cm (12 in).
Subjects in the PR group initiated
closed-chain exercises on a 20.32-cm
(8-in) step. Subjects began with three
sets of 10 repetitions and progressed
until they could do three sets of 15
repetitions without difficulty. This
progression usually occurred within 1
week of rehabilitation in the water.
Between the second and third weeks,
subjects were advanced to a 40.64-cm
(16-in) step in chest-deep water and
the progression format was repeated.
Between the fourth and eighth weeks,
subjects used the 40.64-cm step in
waist-deep water to reduce the force
of buoyancy on body weight, thereby
increasing resistance. Exercises in
waist-deep water progressed in the
same manner. If subjects were able to
perform three sets of 15 repetitions
on the 40.64-cm step in waist-deep
water without difficulty, they were
positioned on a 40.64-cm step in
thigh-deep water for maximal resistance and the sequencing format was
repeated.

Figure I. Pedalling device used by subjects in the water group.

out of the brace by the 6th postoperative week.

with the same exercises, but used a

pedalling device underwater (Fig. 1)
instead of a stationary bicycle.

Rehabliltation Programs
During the second through the eighth
postope~ztiveweeks, the TR group
performed a land rehabilitation program and the PR group performed a
similar program in the water (Tab. 1).
Both programs were performed three
times per week in the same sequence.
Subjects in the TR group warmed up
with 10 minutes of stationary cycling,
followed by 10 minutes of gait training (alternating forward and backward
walking) and 5 minutes of passive
stretching. The PR group warmed up

Subjects in the TR group initiated

closed-chain exercises on a 5.08-cm
(2-in) step. Three sets of 10 repetitions were performed, progressing to
three sets of 15 repetitions. When
subjects could perform three sets of
15 repetitions comfortably at a given
height, the height was increased by
5.08 cm and they started with three
sets of 10 repetitions again. Subjects
usually advanced every two or three
sessions and continued the same
exercise progression while the height
of the step was increased in increments of 5.08 cm. Subjects usually

'Hydrotone International Inc, 3535 NW 58th St, Ste 1000, Oklahoma City, OK 73112.

The next group of exercises consisted

of standing hip flexion, extension,
abduction, adduction, and knee flexion strengthening. The TR group
performed these exercises using pulleys that contained a stack of 4.5-kg
(10-lb) plates. Subjects initiated each
exercise with a weight they could lift
comfortably for three sets of 10 repetitions and progressed until they
could perform three sets of 15 repetitions without difficulty. Another 4.5-kg
plate was then added, and the exercise was repeated with three sets of
10 repetitions.
Hip strengthening and knee flexion
exercises were done using a Hydrotone exercise boot* (Fig. 2). Exercises
consisted of three sets of 10 repetitions for hip flexion-extension,
abduction-adduction, and knee flexion. Because this study did not intend
to quantify the amount of resistance
in the water or to increase the surface
area of the Hydrotone boot, subjects
were instructed to move their involved legs through the water as fast
as they could. As symptoms decreased

Physical Therapy /Volume 74, Number 8/August 1994

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

ware (version 5.1)* were used to

calculate and record peak torque (in
foot-pounds), and gravity-corrected
measurements were obtained. According to the manufacturer, the dynamometer's accuracy is self-calibrated
through the computer software package. One tester, who was blind to
group assignment, performed all the
testing.
During the testing session, subjects
were positioned with their hips in 80
to 90 degrees of flexion. The hips and
tested limb were stabilized with Velcro@ straps across the pelvis and
over the thigh. Subjects were instructed to grasp the handrails during
the test. The axis of rotation of the
dynamometer was aligned with that of
the knee, and the lever arm pad was
placed 7.62 cm (3 in) below the tibial
tubercle. Subjects were allowed a
short period of familiarization at each
speed.

Flgure 2. Hydrotone resistance boot used by subjects in the water group.

and muscle performance improved,
subjects increased the speed and
created more resistance.

Data Collection
Arthrometric measurements.Joint
laxity was measured preoperatively
and at 8 weeks following surgery.
Measurements were made by one of
two physical therapists (BJT and JC)
using a KT-1000 knee arthrometer.'
This device has the highest diagnostic
accuracy of five different arthrometric
de~ices.~6
Anterior drawer testing was
performed with the knee flexed 30
degrees. Anterior displacement of the

tibia on the femur was measured (in

millimeters) during 6.8-kg (15-lb) and
9.1-kg (20-lb) Lachman tests. Greater
forces were not used in fear of overstressing the graft during this critical
period of graft healing. The testers
maintained 100% agreement, within
0.5 mm, both with a prior reliability
study and throughout this study.

Muscle performance
measurements. Isometric and isokinetic peak knee torques were measured at the end of the eighth week
of rehabilitation and compared between groups. An electromechanical
dynamometer$ and LIDO@AC+ soft-

Isometric testing consisted of three

maximal 5-second repetitions with the
knee flexed 85 degrees to measure
knee extension torque and three
maximal 5-second repetitions with the
knee flexed 60 degrees to measure
knee flexion torque. Subjects were
given a 30-second rest period between repetitions. The highest torque
value was recorded.
Isokinetic testing consisted of three
separate contractions at 90/s with a
30-second rest period between repetitions. Isokinetic extension was tested
from 80 to 40 degrees of knee flexion, and isokinetic flexion was tested
from 0 to 70 degrees of knee flexion.
Isohnetic extension was done separately from isokinetic flexion to prevent possible shearing during changes
in direction. The maximum peak
torque for the three repetitions was
recorded for each of the four tests. AU
subjects were tested in the same
order.

Passlve range of motion

measurements. Passive range of
+Medrnetric,San Diego. CA.
$Loredan Biomedical Inc, 2121-B 2nd St, Ste 107, Davis, CA 95616.
"elcro USA Inc, 406 Brown Ave, Manchester, NH 03108.

26/714

motion (PROM) measurements for

knee flexion and extension were
taken by one of the two physical therapists using a standard plastic goniPhysical Therapy/Volume 74, Number 8/August 1994

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

Table 2. Results of Analysis of Variance of Dzferences in Joint Laxity Measurements

During a 6.8-kg (15-lb) Lachmun Test
Source

df

Between subjects
Groups (A)
Error
Within subjects
Weeks (B)

1
17
1

50.84

50.84

AxB

Error

17

MS

21 .OO

21 .OO

3.43

,082

104.08

6.12

SS

0.003
103.05

ometer (17.78 cm [7 in] long with a

360" scale and 1" increments). The
testers maintained 100% agreement,
within 5 degrees, both with a prior
reliability study and throughout this
study.
Passive range of motion was measured at the beginning of each treatment session at 2, 4, 6, and 8 weeks
postoperatively. Subjects were allowed a 3-minute warm-up, which
consistell of self-stretching within
their available ROM. Both measurements were taken with subjects positioned supine. Knee extension measurements were taken with a towel
roll under the heel of the involved
extremity. Knee flexion measurements
were taken with the hip maintained at
90 degrees of flexion, while the heel
was moved toward the buttocks. Endrange was determined by applying
overpressure until firm resistance was
met. The maximum value of three
measurements was recorded.
Girth measurements. Girth measurements were taken by one of the
two physical therapists during the
preoperative visit and at 2, 4, 6, and 8
weeks following surgery. Measurements were taken at the mid-patella
level and 15.24 cm (6 in) above the
mid-patella using a standard tape
measure (increments of 0.3175 cm
[?h
in]) with subjects positioned supine with their thigh musculature
relaxed. 'These measurement locations
were used to document changes in
knee joint e h s i o n and thigh muscular atrophy. The testers maintained

0.003

8.39

.01

0.00

,984

6.06

100% agreement, within 0.636 cm (Y4

in), both with a prior reliability study
and throughout this study.

The ROM measurements for weeks 2,

4, 6, and 8 were analyzed using a
two-way ANOVA (groups X weeks) for
repeated measures. A Tukey's pairwise comparison post hoc test for
significance was used for withingroup comparisons, and a Bonferroni
pair-wise comparison was used for
between-group comparisons.
Girth measurements were calculated
from measurements of girth at midpatella and 15.24 cm above midpatella. Mean differences were compared at 2, 4, 6, and 8 weeks using
tests identical to those undertaken for
ROM. The alpha level of significance
was set at .05.

Results

Functional questlonnalre.A functional questionnaire was administered

at the end of the eighth postoperative
week. The questionnaire consisted of
a Lysholm scale,27 which quantifies
the functional use of the knee joint
using a scale of 0 to 100. This rating
system is a self-report of the subject's
perceived ability of activities such as
walking, stair climbing, and squatting
and is an accepted method of evaluating functional impairment.27,28Higher
scores indicated better functional use
with fewer symptoms.

Results of the ANOVAs for joint laxity

measurements, presented in Tables 2
and 3, showed no significant difference between groups (F=3.43, 4.04;
df=l,l;P=.08, .06), indicating that
neither program induced more laxity
than the other. A significant effect for
time did exist at both the 6.8-kg
(F=8.39, df=l, P=.01) and 9.1-kg
forces (F=24.0, df= 1,P=.0001), indicating that both groups had sign&cantly less joint laxity at 8 weeks after
surgery compared with before
surgery.

Data Management and Analysis

Side-to-side differences in joint laxity
measurements were calculated and
used to compare the values between
groups prior to surgery and 8 weeks
following surgery. Mean differences
were compared using an analysis of
variance (ANOVA). A Tukey's pair-wise
comparison was used for withingroup comparisons, and a Bonferroni
pair-wise comparison was used for
between-group comparisons.

Comparison of quadriceps femoris

and hamstring muscle isometric and
isokinetic peak torque percentages
(Tab. 4) between groups revealed no
significant differences for isometric
knee flexion, isometric knee extension, and isokinetic knee extension
peak torque percentages. The isokinetic knee flexion peak torque percentage, however, was significantly
higher for the TR group @=96.4,
SD=13.5) than for the PR group
@= 81.7, SD=11.1) (P=.01).

Measurements of isometric and isokinetic peak torque for the quadriceps

femoris and hamstring muscles were
normalized to the values of the uninvolved contralateral musculature and
expressed as a percentage. The mean
peak torque percentage and the mean
Lysholm score were compared between groups using a Student's t test.

Passive range of motion measurements were recorded at weeks 2,4, 6,

and 8. Table 5 shows that there were
no significant differences between
groups at each measurement period
(F=0.38, df=l, P=.546). As expected,
there was a significant effect for time
(F= 116.49,df=3, P=.0001), implying
that knee joint PROM for both groups

Physical 'rherapy/Volume 74, Number

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

Table 3. Results of Analysis of Variance of Differences in Joint Laxity Measurements

During a 9.1-kg (20-lb)Lachrnan Test
Source

df

SS

17

156.20
123.73

MS

between the baseline measurement

and the second postoperative week,
but no difference existed between
groups. Mean Lysholrn scores were
significantly higher in the PR group
@=92.2, SD=4.31) than in the TR
group @=82.4, SD=12.36) (P=.03)

Between subjects
Groups (A)
Error
Within subjects
Weeks (B)

A x B

0.25

Error

17

87.64

improved over the 8 weeks. At 2

weeks following surgery, the first
PROM measurement showed that
both groups had an average of 117
degrees of knee PROM. Both groups
showed progressive increments over
time, averaging 20 degrees between
weeks 2 and 4, 8 degrees between
weeks 4 and 6, and 4 more degrees
between weeks 6 and 8. Mean knee
PROM for both groups at the end of
the 8-week program was 150 degrees.
Post hoc analysis revealed that gains
in PROM were significant for both
groups only during the first 6 weeks.
No significant differences were noted
between groups. There was no significant groupx time interaction, indicating that change in PROM over time
was not dependent on assignment.

Girth measurements taken at midpatella and 15.24 cm above midpatella were compared between
knees to determine mean differences
(Tab. 6). Between-group analysis
showed that the PR group had less
girth than the TR group for each midpatella measurement, but the difference was significant only at 8 weeks.
No significant difference between
groups was noted (F=2.09, df= 1,
P=.l66). A time effect was shown
(F=23.45, df=4, P=.0001), as both
groups had a significant increase in
girth at mid-patella between the baseline measurement and the second
postoperative week. Additionally, both
groups showed a significant decrease
in girth at mid-patella after the second
week, but only until week 4. At 15.54
cm above mid-patella, both groups
had significant decreases in girth

Table 4. Means and Standard Deviations for Group Peak Torque Recovery at the
Eighth Postoperative Week (Percentage of Nonoperative Limb's Peak Torque)

Group

lsometrlc Peak
Torque Percentage
(07s)

lsoklnetlc Peak
Torque Percentage
(901s)

SD

85.1

9.1

96.4

Extension

43.1

11.6

56.1
81.7
50.6

13.5
9.2

Pool rehabilitation (n=10)

Flexion

83.7

10.6

Extension

42.8

12.7

At 8 weeks following surgery, both

groups had less than 3 mm of difference in joint laxity between the involved and uninvolved knees for both
the 6.8- and 9.1-kg Lachman tests.
Neither program induced knee joint
laxity, as a laxity difference of 5 3 mm
is considered normal.29 Although
between-group comparisons revealed
no significant difference, the withingroup means at the end of 8 weeks
indicated that the TR group had
greater than 1.5 mm more laxity for
both tests than the PR group. The
inability to detect a significant difference between groups may have been
due to insufficient sample size. This
result may be due to the increased
stresses on the knee joint during
rehabilitation on land
with
in water.l5 Increased knee joint laxity
in the surgical knee at 8 weeks could
have resulted in increased knee joint
effusion, which may have led to the
lower Lysholm scores.

SD

Traditional rehabilitation (n=9)

Flexion

Lysholm scale measurements showed

that the PR group scored significantly
higher than the TR group at 8 weeks,
indicating that this group had fewer
problems with activities of daily living.
Increased pain, based on the subjects'
self-report, and knee swelling during
activities of daily living were primarily
responsible for lower scores in the
TR group. The results of the laxity
and girth measurements may offer
possible reasons why the PR group
had higher Lysholm scores.

1l.l
18.1

Girth measurements taken at 15.24 kgabove mid-patella showed no signscant difference between groups for
atrophy of the thigh musculature.
Within-group comparison, however,
revealed that both groups followed
the same significant changes from the
presurgical measurement until the
eighth postoperative week. Both

Physical Therapy /Volume 74, Number 8/August 1994

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

Table 5. Results of Analysis of Variance of Dgerences Between Groups in Recovey

of Range of Motion
Source

df

SS

MS

Between subjects
Groups (A)
Error

132.61

132.61

18

6287.63

349.31

0.38

,546

Within subjects
Weeks (B)

13277.84

4425.95

1 16.49

AXB

48.24

16.08

0.42

Error

54

2051.66

37.99

groups experienced the greatest

change between the presurgical measurement and the second postoperative week, with the greatest decrease
in girth occurring at the fourth postoperative week. Thigh musculature
atrophy is commonly observed during
the acutt: postsurgical period due to
muscle inhibition that takes place
from the increased joint effusion and
increased pain. DeAndrade et a13O
have shc~wnthat with increased knee
joint effusion, there is less muscle
output as measured by electromyographic activity.

.0001
,737

Thigh girth began to increase after

the fourth postoperative week, and
the involved extremity was within
1.90 cm (0.75 in) of the contralateral
extremity by the eighth postoperative
week for both groups. Increases in
thigh girth at this time may be attributed to several factors. As postoperative joint effusion and pain decrease
while ROM increases, the thigh musculature can b e exercised through a
greater ROM. As exercises are performed more vigorously, muscle
tissue begins to hypertrophy, resulting

Table 6. Results of Analysis of Variance of Dzffwences Between Groupsfor Girth

Measurenzents at Mid-patella and 40.62cm (6 in) Above Mid-patella
Source

df

SS

MS

Mid-patella
Between subjects
Groups (A)
Error
Within sl~bjects
Weeks (B)
B
A x m
40.64 cm above mid-patella

Betweeri subjects
Groups (A)
Error
Within S

1
17

0.006
15.53

0.006
0.80

L J ~ ~ ~ C ~ S

Weeks (B)
AXB
Error

0.01

,933

in greater circumferential
measurements.
Girth measurements taken at midpatella showed that the girth for the
PR group was consistently less at each
time period, but these differences
were significant only at 8 weeks. As
discussed earlier, the increased ginh
in the TR group may have been
caused by the joint effusion resulting
from greater stress on the joint during land exercises compared with
water exercises. This increased joint
effusion may have led to lower Lysholm scores.
Within-group comparison reveals that
mid-patella ginh measurements
changed similarly for both groups;
that is, measurements at this location
were inversely related to the measurements taken at 15.24 kg above
mid-patella. The greatest increase in
girth was noted between the presurgical measurement and the second
postoperative week, suggesting the
increased joint effusion that typically
occurs following surgery. These results indicate that as joint effusion
decreases, muscle girth increases,
with the transition occurring around 4
to 6 weeks following surgery.
Between-group comparison for peak
torque percentages (PTPs) showed
that the TR group had a significantly
higher PTP for the hamstring muscles
at 90/s, indicating that the traditional
rehabilitation approach was more
effective than the pool rehabilitation
approach for strengthening the hamstring muscles. This result may have
occurred for two reasons. First, resistance in the water was partially determined by the speed of limb movement, which was controlled by each
subject.16 Subject effort can be affected
by pain and motivation. Therefore,
subjects may not have generated
enough resistance to facilitate maximal strengthening. Hamstring muscle
exercises in the TR group were done
using weights, so resistance was not
self-paced. Second, there is a difference in the type of muscle contraction that occurs o n land. Empirical
evidence suggests that an eccentric
muscle contraction is important for

Physical Therapy /Volume 74, Number 8/August 1994

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

restoring muscle perf0rmance.3~This

type of contraction is more likely to
occur on land than in water due to
increased gravitational forces.
In both groups, there was equal effectiveness in restoring quadriceps femoris muscle strength. These results also
showed that greater joint effusion in
the TR group did not significantly
affect peak torque muscle performance. A possible reason for this
finding is that all subjects were tested
in the range of 85 to 40 degrees of
knee flexion, rather than at the endrange where joint effusion has been
shown to s e c t muscle perfonnance.30
The mean PTPs for both groups are
similar to those reported for other
subjects with ACL reconstructions.5
Other studies,7,8 however, have demonstrated higher peak torque values.
Two possible explanations for lower
PTPs in this study are the type of
quadriceps femoris muscle strengthening and methodological factors.

Type of Quadriceps Femoris

Muscle Strengthenlng
The method of quadriceps femoris
muscle strengthening in both groups
focused on closed-chain exercises,
which may not have provided enough
isolated stimulus to the quadriceps
femoris muscle to facilitate maximum
strength gains. Previous studies,7.8
which demonstrated higher strength
gains, applied neuromuscular electrical stimulation during open-chain
knee extension exercises. Both
groups in this study may have benefited from isolated knee extension
exercises through a limited ROM (90"
to 40" of knee flexion to ensure graft
protection), as recent research findings indicate that closed-chain exercises alone may not b e enough to
facilitate maximum muscle performance as measured by isokinetic
dynamometry.32

Methodological Factors
Changing the methodology may have
resulted in higher mean PTPs. Performing three 5-second isometric
quadriceps femoris muscle contrac-

tions resulted in donor site pain (the

anatomical site at which the central
third of the patellar tendon was surgically removed for use as an autograft) in some subjects in both
groups, which may have altered the
remaining tests. Anterior knee pain is
common in the early phases of ACL
rehabilitation if a patellar tendon
autograft is used. Testing isometrically
and at slow speeds increases the joint
reaction forces around the patella, but
usually is a better indicator of
strength. Although strength testing in
this study provided adequate graft
protection, testing at faster speeds first
and slower speeds at the end of the
testing session might have resulted in
better PTP scores. The testing procedure in this study did not take these
factors into account because at the
time the study was proposed, no
published research had incorporated
isokinetic testing at 8 weeks, using
only subjects with patellar tendon
autografts.

step-ups, and the land pulleys appeared to be most beneficial for hamstring muscle and hip strengthening.
Isolated quadriceps femoris muscle
contractions in a safe range using
open-chain exercises may have benefited both groups.

Clinical lmpllcatlons

Although traditional exercises have

been the treatment choice of most
clinicians, the results of this study
suggest that a rehabilitation program
for patients with intra-articular ACL
reconstructions performed in a pool
is more effective in reducing joint
effusion and facilitating recovery of
lower-extremity function as indicated
by Lysholm scores. The results also
suggest that rehabilitation in water is
equally effective as on land for restoring knee ROM and quadriceps femoris muscle strength, but not as effective in restoring hamstring muscle
strength. Clinicians who wish to allow
maximal weight bearing may find the
adjunct of aquatic exercises useful.
Future studies should analyze the
effectiveness of a program that combines traditional and water exercises,
using larger sample sizes and a
longer follow-up period.

Although a primary goal in the rehabilitation of patients with ACL reconstructions is the restoration of quadriceps femoris muscle performance, the
means of achieving this goal must
avoid overstressing the graft and increasing joint ehsion. Addttionally, to
expedite recovery, patients must tolerate the rehabilitation program. Some
patients find postoperative exercises
too uncomfortable because of age,
low presurgical activity level, o r low
pain tolerance, and progression during the early phases of rehabilitation
is limited.
Exercises in water may make the total
rehabilitation program more tolerable. Although a complete aquatic
exercise program may be unnecessary, augmenting a land program with
pool exercises may permit loading
the joint to a greater degree. For
patients who are unable to tolerate
traditional exercises on land, water
can be used to facilitate progression
to more aggressive exercises. In this
study, a water environment was most
beneficial for facilitating closed-chain
exercises, such as gait training and

Patients using a pool for rehabilitation

are likely to tolerate an even more
aggressive rehabilitation program than
that presented in this study. In this
study, however, exercises in both
groups had to be carefully matched to
ensure that the main effect between
rehabilitation programs was due to
the environment. The PR group could
have performed more advanced exercises, but varying the exercises would
have made interpretation of results
unclear because differences between
groups could have then been attributed to the environment, exercises, or
interaction between the two.

Acknowledgments

We thank Lynn Snyder-Mackler, ScD,

PT, for assisting with preparation of
this manuscript; Roberto Infante, PT,
and the staff at Resurgeons Orthopaedics for their assistance with data

Physical Therapy/Volume 74, Number 8/August 1994

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

collection; and Piedmont Hospital for

use of their facilities.
References
1 Paulos LE, Payne FC, Rosenberg TD. Rehabilitation after anterior cruciate ligament surgery.
In: Jackson DW, Drez D Jr, eds. The Anterior
Cmciate LkJcient Knee. St Louis, Mo: CV
Mosby Co; 1987:291-314.
2 Shelbourne KD, Nitz P. Accelerated rehabilitation after anterior cruciate ligament reconstruction. Am ]Sports Med. 1990;18:292-299.
3 Paulos IE, Noyes FR, Grood ES, Butler DL.
Knee rehabilitation after anterior cruciate ligament reconstruction and repair. Am] Sports
Med. 1981 9140-149.
4 Shelbourne KD, Wilckens JH. Current concepts in ar~teriorcruciate ligament rehabilitation. Orthcp Rev. 1990;11:957-964.
5 Draper V, Ballard L. Electrical stimulation
versus electromyographic biofeedback in the
recovery of quadriceps femoris muscle function following anterior cruciate ligament surgery. Phys Thet. 1991;71:455-464.
6 Draper 'V. Electromyographic biofeedback
and rec0vt:r-y of quadriceps femoris muscle
function following anterior cruciate ligament
reconstruction. Phys Thm 1990;70:11-17.
7 Snyder-Mackler L, Ladin 2, Schepsis Aq
Young LC. Electrical stimulation of the thigh
musculature after reconstruction of the anterior cruciate ligament.] Bone Joint Surg [Am].
1991;73:1025-1036.
8 Delitto A, Rose SJ, McKowen JM, et al. Electrical stimulation versus voluntary exercise in
strengthening thigh musculature after anterior
cruciate ligament surgery. Phys Ther. 1988;68:
661-663.
9 Ohkoshi Y, Yasada K. Biomechanical analysis
of shear force exerted to anterior cruciate ligament durirrg half squat exercise. Orrhop Trans.
1989;13:310.
10 Steindler A. Kinesiology of the Human
Body Under Nonnal and Pathological Condi-

tions Springfield, Ill: Charles C Thomas, Publisher; 1955.

1 1 Pope MH, Stankewich CJ, Beynnon BD,
Fleming BC. Effect of knee musculature on
anterior cruciate ligament strain in vivo. Journal of Electromyography and Kinesiology.
1991;1:191-198.
12 Whieldon T, Yack J, Collins C. Anterior tibial translation during weight-bearing and nonweight-bearing rehabilitation exercises in the
anterior cruciate deficient knee. Phys Ther.
1989;69:151.Abstract.
13 Henning CE, Lynch MA, Glick KR.An in
vivo strain gauge study of elongation of the
anterior cruciate ligament. Am] Sports Med.
1985;13:22-26.
14 Reynolds NL, Worrell TW, Perrin DH. Effect
of a lateral step-up exercise protocol on quadriceps isokinetic peak torque values and thigh
ginh. ] Orthop Sports Phys Ther. 1992;15:151155.
15 Golland A. Basic hydrotherapy. Physiotherapy. 1981;67:25%262.
16 Edlich RF,Towler MA, Goitz RJ, et al.
Bioengineering principles of hydrotherapy.
J Bum Care Rehabil. 1987;8:580-584.
17 Hillman MR, Matthews L, Pope J. The resistance to motion through water of hydrotherapy table-tennis bats. Physiotherapy. 1987;73:
570-572.
18 Harrison RA,Allard LL. An attempt to quantify the resistances produced using the bad
ragaz ring method. Physiotherapy. 1982;68:23@231.
19 Harrison Rk A quantitative approach to
strengthening exercises in the hydrotherapy
pool. Physiotherapy. 1980;66:60.
20 Abidin MR, Lobardi SA, Devlin PM, et al. A
new hydrofitness device for strengthening
muscles of the upper extremity.] Bum Care
Rehabil. 1988;9:402-406.
21 Abidin MR, Thacker JG, Becker DG, et al.
Hydrofitness devices for strengthening upper
extremity muscles.] Bum Care Rehabil. 1988;
9:199-202.
22 Goitz RJ, Towler TA, Buschbacher LP, et al.
A new hydrofitness device for leg muscu-

loskeletal conditioning.] Bum Care Rehabil.

1988;9:203-206.
23 Bartow L, Diamond L. Resistance Training
in the Water:An Analysis Comparing the
Hydro-tone System to Water Resistance Without
a Training Tool in Resistance of the Knee Flerors and Extensors. Boston, Mass: Boston University; 1989. Master's thesis.
24 Gehlsen GM, Grigsby SA, Winant DM. Effects of an aquatic fitness program on the
strength and endurance of patients with multiple sclerosis. Phys Ther. 1984;64:653457.
25 Napoletan JC. The Efect of Undenvatet
Treadmill Exercise in the Rehabilitation of Surgical Anterior Cmciate Ligament Repair. Orange, Calif: Chapman College; 1990. Master's
thesis.
26 Anderson AF, Snyder RB,Federspiel CF,
Lipscomb B. Instrumented evaluation of knee
laxity: a comparison of five arthrometers. Am]
Sports Med. 1992;20:135-140.
27 Lysholm J, Gillquist J. Evaluation of knee
ligament surgery results with special emphasis
on use of a scoring scale. Am] Sports Med.
1982;10:150-154.
28 Tegner Y, Lysholm J. Rating systems in the
evaluation of knee ligament injuries. Clin Orthop. 1985;198:43-49.
29 Daniel DM, Malcom LL, Losse G, et al. Instrumented measurement of anterior laxity of
the knee.] Bone Joint Surg [Am] 1985;67.720726.
30 deAndrade JR, Grant C, Dixon AJ. Joint distension and reflex inhibition in the knee.
]Bone Joint Surg [Am]. 1965;47:3542.
31 Albert M. Physiologic and clinical principles of eccentrics. In: Albert M, ed. Eccenmc
Muscle Training in Sports and Orthopaedics.
New York, NY: Churchill Livingstone Inc; 1991:
11-23.
32 Reynolds NL, Worrell TW, Perrin DH. Effect
of a lateral step-up protocol on quadriceps
isokinetic peak torque values and thigh ginh.
J Orthop Sports Phys Ther. 1992;15:151-155.

Physical 'Therapy/Volume 74, Number B/August 1994

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

Comparison of the Effects of Exercise in Water and

on Land on the Rehabilitation of Patients With
Intra-articular Anterior Cruciate Ligament
Reconstructions
Brian J Tovin, Steven L Wolf, Bruce H Greenfield, Jeri
Crouse and Blane A Woodfin
PHYS THER. 1994; 74:710-719.

This article has been cited by 2 HighWire-hosted articles:

Cited by

http://ptjournal.apta.org/content/74/8/710#otherarticles
http://ptjournal.apta.org/subscriptions/

Subscription
Information

Permissions and Reprints http://ptjournal.apta.org/site/misc/terms.xhtml

Information for Authors

http://ptjournal.apta.org/site/misc/ifora.xhtml

Downloaded from http://ptjournal.apta.org/ by guest on May 24, 2014

Correction
In "Comparison of the Effects of
Exercise in Water and o n Land o n
the Rehabilitation of Patients With
Intra-articular Anterior Cruciate
Ligament Reconstructions" by
Tovin et al in the August 1994
issue, an incorrect unit of measurement for girth is shown o n pages
716 and 717 and an incorrect metric conversion is presented in
Table 6. The discussion and the
table should b e corrected to reflect
that thigh girth measurements in
that study w e r e taken at mid-patella
and at 1 5 2 4 c m (6 in) above midpatella. The Journal regrets the
errors.

no chest pain and normal coronary arteries (group 3).

The TENS was set to deliver 300-millisecond constant-current pulses at 150 Hz.
Electrodes were placed 20 cm apart at
the site of the most intense pain for subjects in groups 1 and 2, and over the
pericardium for subjects in group 3. Intensity was adapted individually to just
below the level that produced pain
(10-60 mA). Coronary angiography was
performed for measurement of blood
flow velocity and diameter of the left
anterior descending and circumflex arteries during diastole. After determining
TENS intensity, the stimulator was turned
off and baseline values of resting blood
flow velocity, heart rate (HR), mean and
systolic blood pressures (BP), arterial
diameters, and aortic norepinephrine and
epinephrine concentrations were determined. Transcutaneous electrical nerve
stimulation treatment was then given for
5 minutes, and measurements were immediately repeated. In 10 patients in each
group, the study was repeated after coronary blood flow returned to baseline
values. Results were compared using the
Wilcoxon matched-pairs test.
Resting coronary blood flow velocity
increased and aortic epinephrine decreased significantly in groups 1 and 2,
after TENS treatment. Heart rate, mean or
systolic BP, blood norepinephrine, and
arterial diameters were not altered by
TENS in any group. Results were highly
reproducible. There was a variability in
responses to TENS in groups 1 and 2 that
was independent of threshold stimulation

84/1165

and site of electrode placement, suggesting individual differences in sensitivity to

neurostimulation. The lack of any
changes in group 3 (patients with heart
transplant) may be associated with cardiac denervation or heterogenous reinnervation, immunosuppression therapy,
or higher baseline values of HR.

of the continuous marker and the individual spinal level markers were manually digitized, and the spinal curvature
was calculated through the use of a
GTCO digitizer, two computers, and
custom-designed software. This measurement process involved 2 hours of time
per subject.

The authors suggested that the increase

in coronary blood flow velocity after
TENS in patients with chest pain is due to
dilation at the microcirculatory level. This
may be the result of local production of
vasodilatory substances, reduction in
local sympathetic activity, or both.

The measurement error due to the manual digitization was determined by examining repeated trials of manual tracings
over known curves that were designed to
simulate spinal curvatures. This was
found to be less than 0.02% in relation to
the mean surface curvature (-0.000010
to -0.000041 rad/mm). This degree of
error was found to be similar throughout
the different regions of the curve.

Carol J Weaver, PhD, PT

West Virginia University
Morgantoum, WVa

Validation off a Non-Invasive

Method of Measuring the Surface
Curvature of the Erect Spine
Raine S, Twomey LT (School of Physiotherapy, Curtin University of Technology,
Perth, Western Australia, Australia), Journal of Manual and Manipulative Therapy. 1994;2:11-21.
The authors introduced a noninvasive
method for measurement of spinal curvatures, examined the amount of measurement error for this method, and determined the validity of this method by
comparisons with radiographs. The authors
offered an alternative to measurement
methods using a modified Cobb technique
that reflected changes in the angulation of
the superior and inferior vertebrae in a
particular region without consideration of
the shape of the curve. This method used
photographs, from which the spinal curves
were manually digitized and then calculated by a computer.
Fifteen volunteer subjects from an outpatient scoliosis clinic were used in this
study. The noninvasive measurement
method involved identifying with adhesive dots the C-7, T-6, T-12, L-2, and the
left posterior superior iliac spine. This
was followed by the application of a continuous foam marker (1.27 cm in width
and 2.54 cm in depth) placed along the
subject's spine. The adhesive dots were
then transferred to the foam marker,
delineating the borders of four regions:
upper thoracic, lower thoracic, upper
lumbar, and lower lumbar. Through a
standardized procedure, a profile photograph was taken as the subject stood over
the center of a graphlex platform and
directly behind a plum line. The contours

The validity of this technique was studied

by comparing the measurements taken
from the photographs with measurements taken from the radiographs. Tracings were taken from the curves defined
by the posterior vertebral bodies and
from the spinous processes. These tracings were manually digitized and computer calculated by the same means as
the photographs. These radiographic
curves were then compared with the
photographic, surface curves for each
subject. A Pearson Product-Moment Correlation Coefficient was used for analysis
of the results. The results were varied,
ranging from r=.37 to r=.84. The correlations between the surface measurements
and the radiographs improved if the surface measurement was compared with the
posterior vertebral body measurement in
the thoracic spine (r=.70-.84), and if the
surface measurement was compared with
the spinous process measurement in the
lumbar spine (r=.65-.67). This discrepancy
was explained by the authors in terms of
variations in the radiographic anatomy of
these two spinal regions.
The authors concluded that this noninvasive measurement technique was able to
provide a reasonable indication of the
curvature of the underlying vertebral
column. The authors suggested that this
mathematically derived curvature of the
surface contour of the spine is a more
precise method than the modified Cobb
method. They speculated about modifications in the testing conditions that may
improve the correlations between the
surface measurements and the radiographic measurements.
Karen Maloney Backstrom, PT, OCS
Univ of Colorado
Denver, Colo

Physical Therapy/Volume 74, Number 12/December 1994