Arthritis & Rheumatism (Arthritis Care & Research)

Vol. 51, No. 5, October 15, 2004, pp 722–729

DOI 10.1002/art.20685
© 2004, American College of Rheumatology
ORIGINAL ARTICLE

Comparison of Manual Therapy and Exercise

Therapy in Osteoarthritis of the Hip: A
Randomized Clinical Trial
HUGO L. HOEKSMA,1 JOOST DEKKER,2 H. KAREL RONDAY,1 ANNET HEERING,1
NICO VAN DER LUBBE,1 CEES VEL,3 FERDINAND C. BREEDVELD,4 AND CORNELIA H. M. VAN DEN ENDE5

Objective. To determine the effectiveness of a manual therapy program compared with an exercise therapy program in
patients with osteoarthritis (OA) of the hip.
Methods. A single-blind, randomized clinical trial of 109 hip OA patients was carried out in the outpatient clinic for
physical therapy of a large hospital. The manual therapy program focused on specific manipulations and mobilization of
the hip joint. The exercise therapy program focused on active exercises to improve muscle function and joint motion. The
treatment period was 5 weeks (9 sessions). The primary outcome was general perceived improvement after treatment.
Secondary outcomes included pain, hip function, walking speed, range of motion, and quality of life.
Results. Of 109 patients included in the study, 56 were allocated to manual therapy and 53 to exercise therapy. No major
differences were found on baseline characteristics between groups. Success rates (primary outcome) after 5 weeks were
81% in the manual therapy group and 50% in the exercise group (odds ratio 1.92, 95% confidence interval 1.30, 2.60).
Furthermore, patients in the manual therapy group had significantly better outcomes on pain, stiffness, hip function, and
range of motion. Effects of manual therapy on the improvement of pain, hip function, and range of motion endured after
29 weeks.
Conclusion. The effect of the manual therapy program on hip function is superior to the exercise therapy program in
patients with OA of the hip.

KEY WORDS. Osteoarthritis; Hip; Rehabilitation; Musculoskeletal manipulations; Exercise therapy.

INTRODUCTION to pain, loss of mobility and muscle function, restriction in

activities of daily living, and decreased quality of life.
Osteoarthritis (OA) is a common cause of disability. OA is In clinical practice, there is a variety of conservative
characterized by progressive loss and degeneration of ar- treatment methods available for patients with OA of the
ticular cartilage, sclerosis of the subschondral bone, and hip, including manual therapy and exercise therapy (4).
formation of osteophytes (1–3). These changes often lead Manual therapy includes manipulation and stretching
techniques. Manual therapy is particularly aimed at the
improvement of elasticity of the joint capsule and the
1
Hugo L. Hoeksma, PT, H. Karel Ronday, MD, PhD, Annet
surrounding muscles (5). Manual therapy is provided by
Heering, PT, Nico van der Lubbe, PT: Leyenburg Hospital, physical therapists (or medical doctors) with a special
The Hague, The Netherlands; 2Joost Dekker, PhD: Institute training in manual therapy. Exercise therapy includes both
for Research in Extramural Medicine (EMGO Institute), active and passive exercises (6 –9). Exercise therapy aims
Vrije Universiteit Medical Center, Amsterdam, The Nether-
lands; 3Cees Vel, PT, The Hague, The Netherlands; 4Ferdi-
at improvement of muscle function, increase of joint range
nand C. Breedveld, MD, PhD: Leiden University Medical of motion, decrease of pain, and increase of walking ability
Center, Leiden, The Netherlands; 5Cornelia H. M. van den (6 –9). Exercise therapy is provided by physical therapists.
Ende, PT, PhD: Netherlands Institute for Health Services Exercise therapy is reported to be effective in patients
Research, Utrecht, The Netherlands.
Address correspondence to Hugo L. Hoeksma, Depart- with OA of the hip (8). Both manual therapy and exercise
ment of Rehabilitation and Health Services. St. Antonius therapy are frequently applied in OA of the hip. It is not
Hospital, PO Box 2500, 3430 EM Nieuwegein, The Nether- known which of these approaches is superior. Therefore,
lands. E-mail: [email protected]. the objective of the present randomized clinical trial was
Submitted for publication February 23, 2003; accepted in
revised form January 17, 2004. to compare the effects of manual therapy and exercise
therapy in OA of the hip.

722
Manual Therapy Versus Exercise Therapy in Hip OA 723

PATIENTS AND METHODS was allowed if it was left unchanged during the study
period. Other treatment by health professionals, such as
Study participants. During the period September 1999 occupational therapists, was to be avoided.
to December 2001, patients with OA of the hip were re-
ferred by orthopedic surgeons or rheumatologists to the Blinding. A single assessor (HLH), who was blinded to
outpatient clinic of the physical therapy department with the allocation of treatment programs, carried out all mea-
complaints due to OA of the hip. Hip OA was defined surements. Patients were instructed by a secretary not to
according to the clinical criteria of the American College of give information about the allocated treatment to the as-
Rheumatology (10). These criteria are hip pain and ⬍15° of sessor. Furthermore, the assessor was not allowed to visit
internal rotation and ⬍115° of flexion in the hip joint or the physical therapy department during treatment hours to
hip pain and ⬎15° of internal rotation and pain on hip further assure blinding. Also, all measurements were per-
internal rotation and morning stiffness of the hip of ⬎60 formed at a location separate from the physical therapy
minutes. Assessment of the criteria was performed by the department, on a different floor of the hospital. Finally, the
referring physician. Exclusion criteria were 1) symptoms assessor was asked to guess the assigned treatment directly
in both hips, 2) fear of manipulative therapy, 3) age ⬍60 after posttreatment measurements (5 weeks).
years or ⬎85 years (to get a maximum spread of 25 years),
4) severe complaints of the lower back, 5) severe cardio- Interventions. Three manual therapists and 3 physical
pulmonary disease, and 6) insufficient knowledge of the therapists performed all treatments. The manual therapists
Dutch language to complete instructions and forms. The were licensed manual therapists. The physical therapists
study was approved by the medical ethics committee of did not receive training in manual therapy or in manipu-
the hospital and all participants provided written in- lation techniques. All participating physical therapists
formed consent. The study was designed as a single-blind, were instructed in training sessions. These training ses-
randomized clinical trial and reported following the CON- sions were repeated every 3 months. All patients were
SORT statement for reporting the results of clinical trials treated twice weekly for a period of 5 weeks with a total of
(11). 9 treatments. The first treatment session was used to tailor
the treatment protocol to the individual patient. Content of
Randomization. After inclusion in the trial, radio- treatment, deviations from the protocol, and compliance
graphs of the hip were taken and scored by a radiologist were registered. In addition, adverse effects were regis-
following a standardized procedure according to a modi- tered.
fied Kellgren/Lawrence scale (12,13). To optimize prog- Each manual therapy session started with stretching
nostic similarity, prestratification was conducted for ra- techniques of identified shortened muscles surrounding
diographic severity. Two strata were constructed: stratum the hip joint (Appendix A). Second, traction of the hip
1, Kellgren/Lawrence score ⱕ1 (no OA or mild OA); and joint was performed, followed by traction manipulation in
stratum 2, Kellgren/Lawrence score ⱖ2 (moderate or se- each limited position (a high velocity thrust technique)
vere OA). (14). All manipulations were repeated during each session
A staff member not involved in the trial prepared the until the manual therapist concluded optimal results of
numbered, nontransparent, sealed envelopes. Permutated the session. Both manual therapists working at the hospi-
blocks were used to optimize equal distribution of patients tal and experts in the field of manual therapy were in-
between the 2 intervention groups. A random sequence of volved in developing the treatment protocol.
permutated blocks of 6 envelopes was generated by using In the exercise treatment group, an exercise therapy
random number tables. program was planned by the physical therapist and ad-
justed to individual symptoms (Appendix B). Exercise
Treatment and measurements. Treatment started therapy included exercises for muscle functions, muscle
within 1 week of baseline. All patients received treatment length, joint mobility, pain relief, and walking ability.
at the outpatient clinic for physical therapy of the hospital. Finally, instructions for home exercises were given. The
At baseline, all participants completed a questionnaire exercise therapy program was an adaptation of the exercise
containing questions on demographic variables, previous therapy program of Van Baar et al (8,9).
complaints, duration of symptoms, cointerventions, and
previous treatment with exercise therapy or manual ther- Outcome assessment. The primary outcome measure,
apy. The use of medication and other treatment was re- general improvement experienced by the patient, was as-
corded at each assessment. sessed using a 6-point Likert scale ranging from “much
Measurements were performed at baseline (week 0), af- worse” to “complete recovery.” This method is often ap-
ter the treatment period (week 5), and after 3 months (week plied as a primary outcome measure in our area of research
17) and 6 months (week 29). General improvement expe- and has shown to be a valid and reliable assessment (15).
rienced by the patient was assessed only at the 5-week Health-related quality of life was assessed using the
followup visit. This was done because we believe that this Short Form 36 (SF-36) (16). We selected the SF-36 sub-
measure, due to the long period between the followup scales for bodily pain, physical functioning, and role phys-
measurements (12 weeks), would not be memorized by ical functioning.
patients in a correct way. The use of nonsteroidal antiin- Hip function was evaluated using the Harris Hip Score
flammatory drugs and pain medication prior to the trial and by a walking test (17–20). The Harris Hip Score is a
724 Hoeksma et al

disease-specific index containing 8 items (20). Items rep- Table 1. Baseline characteristics (n ⴝ 109)*
resent pain, walking function, activities of daily living,
and range of motion of the hip joint. Walking ability was Manual Exercise
assessed using a walking test. This test was set out in a therapy therapy
corridor of the hospital with a marked distance of 10
Total no. patients 56 53
meters. Total walking distance was 80 meters with 7 turn-
Females/males, no. 38/18 38/15
ing points. Patients were instructed to walk fast, but not to Age, mean ⫾ SD years 72 ⫾ 7 71 ⫾ 6
run. Time to complete the test was recorded (walking Duration of complaints
speed). 1 month to 1 year 22 15
At baseline, all patients were asked to formulate his or 1 year to 2 years 12 13
her main complaint. The patient then rated the intensity of 2 years to 5 years 9 15
the main complaint on a visual analog scale (VAS). Also, 5 years to 10 years 10 8
starting stiffness and pain during the walking test were Longer than 10 years 3 2
recorded on a VAS (17,18). Range of hip joint motion from Medication
flexion to extension and from internal to external rotation No medication 41 38
Analgetics or NSAIDs 15 15
was recorded with a long-legged goniometer according to a
Previous treatment
standardized procedure (21). Massage 2 3
Exercise therapy 3 2
Statistical analysis. The target sample size was 120 pa- Manual therapy 0 1
tients. This number yields a power of 90% to detect a Radiographic deterioration
difference between groups of 30% on the primary outcome 0 (no OA) 5 4
measure general improvement at a 2-sided ␣ of 5%, given 1 (mild OA) 7 6
a maximum dropout rate of 20%. Enrollment stopped at 2 (moderate OA) 19 23
109 patients because the actual dropout rate was 6%. The 3 (severe OA) 25 20
Main complaint, no.
intervention was considered successful if a patient re-
Pain 34 33
ported to be improved, much improved, or free of com- Morning stiffness 5 4
plaints. Starting stiffness 4 3
Analyses were performed according to the intention-to- Walking disability 13 14
treat principle (22). In addition, a per-protocol analysis Harris hip score, mean ⫾ SD 54 ⫾ 15 53 ⫾ 14
was performed excluding patients who had received total
hip arthroplasty during the followup period between 5 and * Data presented are means, unless otherwise noted. Radiographic
deterioration, according to modified Kellgren and Lawrence scale.
29 weeks. At first, data of the 2 groups were screened for NSAIDs ⫽ nonsteroidal antiinflammatory drugs; OA ⫽
normal distribution using normal plots. Comparison be- osteoarthritis.
tween groups was made for age, sex, severity of OA, dura-
tion of complaints, and function (Harris Hip Score). To test
between groups on the dichotomized primary outcome assessed on eligibility (Figure 1). Of the 109 participants,
measure general improvement, odds ratios (ORs) were cal- 56 were assigned to the manual therapy program and 53
culated. For all continuous variables, analyses were per- were assigned to the exercise therapy program. After 5
formed with analysis of covariance using the baseline weeks, 6 patients were lost to followup. After 17 weeks,
scores as the covariate (with 95% confidence interval another 9 patients were lost to followup. At 29 weeks, at
[95% CI]). A 2-sided ␣ of 5% was applied in testing overall total of 21 patients were lost to followup (12 in the manual
significance. In addition, effect sizes were calculated by therapy group and 9 in the exercise therapy group).
taking the mean differences of the change scores of the
intervention groups and dividing it by the standard devi- Compliance, cointerventions, and adverse effects. Pa-
ation of the actual scores of the total population. An effect tients who prematurely discontinued the treatment pro-
size of 0.2 is considered small, an effect size of 0.5 is grams were denoted as noncompliant. Seven patients were
considered moderate, and an effect size of 0.8 is consid- considered noncompliant: 3 in the exercise program and 4
ered large (23). in the manual therapy program. In the exercise program, 1
patient withdrew before the first treatment because of trav-
eling distance to the hospital, 2 patients withdrew due to
RESULTS increase of complaints, and 1 patient discontinued be-
cause of cardiorespiratory disease. In the manual therapy
Comparability. In total, 109 patients were included in group, 3 patients discontinued because of increase of com-
the study. The 2 groups were generally similar for baseline plaints. No patients reported to have changed use of med-
characteristics (Table 1); no relevant differences were ication and no patients reported to have had treatment by
found. Mean age was relatively high (72 years). Most pa- other health professionals. No other adverse effects were
tients (80%) had a Kellgren/Lawrence score of 2 or 3, reported.
indicating moderate to severe OA.
Total hip arthroplasty. In total, 18 patients received
Participant flow and followup. One hundred thirteen total hip arthroplasty during the followup period (9 pa-
patients referred to the physical therapy department were tients in each treatment group). Fourteen patients had the
Manual Therapy Versus Exercise Therapy in Hip OA 725

sor correctly guessed the assigned treatment in 53% of the

cases.

Treatment. Exercise therapy was mainly directed to-

ward improvement of muscle function (i.e., strength, en-
durance, and coordination; 90%), improvement of range of
joint motion (82%), and the reduction of pain (74%). The
applied exercises were (assisted) active exercises (95%)
and passive exercises (72%). In 61% of the patients, home
exercise instructions were given. In most cases (90%),
manual therapy was applied according to protocol. How-
ever, in 10% (n ⫽ 5) of the cases, manipulations were not
used in all treatment sessions.

Outcome. In Table 2 it is shown that after treatment (5

weeks) the success rate (primary outcome) of manual ther-
apy was 81% versus 50% for exercise therapy (OR 1.92,
95% CI 1.30, 2.60). Tables 3 and 4 present the results of the
intention-to-treat analysis (including patients who had re-
ceived total hip arthroplasty). Table 3 presents results for
the secondary outcome measures of quality of life and hip
function. Beneficial effects of manual therapy on the Har-
ris Hip Score (Figure 2) and walking speed were found. No
differences in effects were found on the subscales of the
SF-36, except for a beneficial effect of exercise therapy on
the subscale role physical functioning. The effect sizes for
the Harris Hip Score were large and for walking speed,
medium. Table 4 presents the results of the analysis of
outcome measures of impairments. Beneficial effects were
found for manual therapy on pain, stiffness, and range of
motion (Figure 3), effect sizes for pain and stiffness were
medium and effect sizes for range of motion were large.
Figure 1. Patient flow and followup. MT ⫽ manual therapy; ET ⫽ The higher improvement in the manual therapy group
exercise therapy. compared with the exercise therapy group endured for
most measures after 17 and 29 weeks. However in general,
surgery between 5 and 17 weeks followup and 4 patients effects declined as compared with results after 5 weeks. In
had the surgery between 17 and 29 weeks followup. the per-protocol analysis, similar results were found as in
the intention-to-treat analysis (results not reported).
Success of blinding. The outcome assessor reported dis-
closure of the assigned treatment for 2 patients: in 1 case,
a patient dropped the name of the manual therapist and in
DISCUSSION
1 case, the patient was seen with the physical therapist Convincing evidence was found for the effectiveness of the
(exercise therapy group). At 5 weeks followup, the asses- manual therapy program as compared with the exercise

Table 2. Differences between the 2 treatment groups after 5 weeks*

Kellgren and
Main comp Lawrence
Manual Exercise exercise Main comp Odds
therapy therapy therapy manual therapy 0 1 2 3 ratio 95% CI

Worse 3 6 10 ⫾ 15 23 ⫾ 18 0 2 3 4
Little worse 2 3 1⫾8 22 ⫾ 19 1 1 2 1
Stable 5 16 4 ⫾ 12 4 ⫾ 12 2 0 6 13
Improved 27 21 ⫺12 ⫾ 14 ⫺20 ⫾ 14 4 7 20 17
Much improved 15 4 ⫺34 ⫾ 24 ⫺35 ⫾ 17 2 2 10 5
Free of complaints 1 0 ⫺18 0 0 1 0
Improved (%) 43 (81) 25 (50) 1.92 1.30, 2.60
Not improved 10 25

* Improvement of the main complaint (main comp.; visual analog scale, analysis of covariance) on the basis of primary outcome; negative numbers
indicate improvement. Primary outcome on the basis of Kellgren and Lawrence. Odds ratio for improvement of manual therapy versus exercise therapy.
95% CI ⫽ 95% confidence interval.
726 Hoeksma et al

Table 3. Quality of life and hip function*

Manual therapy Exercise therapy

No. Score No. Score Mean differences† 95% CI Effect size

SF-36 bodily pain

Baseline 56 41.1 ⫾ 18 53 37.9 ⫾ 18
Week 5 53 44.0 ⫾ 17 50 42.4 ⫾ 17 ⫺2.1 ⫺4.4, 8.6 0.1
Week 17 49 47.4 ⫾ 25 45 46.1 ⫾ 20 ⫺3.2 ⫺13.1, 6.8 0.1
Week 29 44 51.4 ⫾ 22 44 49.9 ⫾ 24 ⫺1.5 ⫺11.1, 7.7 0.1
SF-36 physical function
Baseline 56 42.1 ⫾ 23 53 41.4 ⫾ 21
Week 5 53 43.6 ⫾ 18 50 41.5 ⫾ 22 1.4 ⫺4.7, 7.4 0.1
Week 17 49 45.3 ⫾ 23 45 46.6 ⫾ 21 ⫺2.1 ⫺11.7, 7.7 0.1
Week 29 44 50.4 ⫾ 22 44 45.3 ⫾ 18 3.1 ⫺4.1, 10.5 0.2
SF-36 role physical function
Baseline 56 27.0 ⫾ 38 53 24.7 ⫾ 36
Week 5 53 23.2 ⫾ 30 50 32.2 ⫾ 24 ⫺11.3 ⫺21.5, ⫺1.1‡ 0.4
Week 17 49 25.4 ⫾ 43 45 29.8 ⫾ 33 ⫺6.4 ⫺23.5, 10.2 0.2
Week 29 44 36.7 ⫾ 44 44 32.4 ⫾ 35 2.2 ⫺16.8, 21.1 0.1
Harris hip score
Baseline 56 54.0 ⫾ 15 53 53.1 ⫾ 14
Week 5 53 69.3 ⫾ 15 50 57.2 ⫾ 11 11.2 6.1, 16.3‡ 0.9
Week 17 49 68.4 ⫾ 17 45 56.0 ⫾ 15 11.1 4.0, 18.6‡ 0.7
Week 29 44 70.2 ⫾ 20 44 59.7 ⫾ 18 9.7 1.5, 17.9‡ 0.5
Walking speed (seconds)
Baseline 56 96.3 ⫾ 37 53 96.1 ⫾ 25
Week 5 53 88.3 ⫾ 23 50 96.5 ⫾ 27 ⫺8.2 ⫺16.7, ⫺0.5‡ 0.3
Week 17 49 86.8 ⫾ 27 45 99.4 ⫾ 21 ⫺12.7 ⫺24.0, ⫺2.0‡ 0.5
Week 29 44 90.5 ⫾ 26 44 102.8 ⫾ 18 ⫺12.1 ⫺20.5, 3.8 0.6

* Data presented as mean scores ⫾ SDs per group. Mean differences adjusted for baseline values as analyzed by analysis of covariance. On Short Form
36 (SF-36) subscales and Harris hip score, positive signs indicate improvement. On walking speed, negative signs indicate improvement. 95% CI ⫽
95% confidence interval.
† Adjusted for baseline values.
‡ P ⬍ 0.05.

therapy program in patients with OA of the hip. The ef- and stiffness and an increase of range of motion (3). This
fects on general improvement, hip function, and pain were could explain the superior effectiveness of manual therapy
significantly better for patients who were treated with on pain, stiffness, and range of motion as compared with
manual therapy. Most of the beneficial effects of manual exercise therapy. However, further research has to be done
therapy endured until 3 months and 6 months after finish- to specifically determine the mechanism behind the effects
ing treatment. of manual therapy on walking ability.
To our knowledge, this is the first study on the contrast As for the design of the study, some comments can be
between exercise therapy and manual therapy in OA. The made. First, the present study was a single-blind study.
only other available study on manual therapy in patients For obvious reasons, it was not possible to blind either
with OA of the lower extremities is a trial on the effective- patients or therapists for the allocated treatment. There-
ness of a combination of manual therapy and exercise fore, extra attention was given to the blinding of the out-
therapy in patients with OA of the knee (5). In that study, come assessor. Second, we cannot exclude the possibility
beneficial effects were found on pain and function. of a placebo effect due to the nature of the interventions.
We observed a beneficial effect (after 5 weeks) of exer- Furthermore, another limitation of the study may be the
cise therapy on the SF-36 subscale role physical function- relatively large number of patients who received total hip
ing. However, in contrast to other findings, this beneficial arthroplasty during the followup period. However, no dif-
effect was the only effect in favor of exercise therapy in ferences were found between the conclusions based on the
⬎10 tested secondary outcome measures on functional intention-to-treat analysis and the per-protocol analysis.
ability. Therefore, in our opinion this finding lacks clinical We chose to standardize the number of treatment ses-
relevance. sions to guarantee equal exposure to attention of the phys-
High intraarticular pressure, due to restriction of the ical therapist in both groups. In both groups, all patients
joint capsule in OA, is associated with pain intensity (1). received 9 treatments sessions. Manual therapy is usually
These pathologic changes lead to reduced range of motion applied in a limited number of treatment sessions, because
of the hip joint (1–3). Manual therapy (manipulation and of the expected immediate and relatively strong effects of
stretching) is particularly aimed at the improvement of manual therapy (25). One could argue that the number of
elasticity of the joint capsule and the surrounding muscles sessions in the exercise therapy group was too small to
(24,25). It is believed that this results in a reduction of pain achieve optimal results. In most trials on the effectiveness
Manual Therapy Versus Exercise Therapy in Hip OA 727

Table 4. Pain, stiffness, and range of joint motion*

Manual therapy Exercise therapy

No. Score No. Score Mean differences† 95% CI Effect size

Pain at rest, VAS, mm

Baseline 56 22.5 ⫾ 23 53 23.0 ⫾ 26
Week 5 53 17.1 ⫾ 22 50 26.7 ⫾ 18 ⫺9.1 ⫺16.4, ⫺1.6‡ 0.5
Week 17 49 19.1 ⫾ 29 45 26.9 ⫾ 28 ⫺7.2 ⫺13.8, ⫺0.5‡ 0.3
Week 29 45 14.0 ⫾ 27 44 21.6 ⫾ 30 ⫺7.0 ⫺20.3, 5.9 0.3
Pain walking, VAS, mm
Baseline 56 34.0 ⫾ 22 53 28.8 ⫾ 22
Week 5 53 22.8 ⫾ 21 50 27.1 ⫾ 21 ⫺9.6 ⫺17.3, ⫺1.8‡ 0.5
Week 17 49 16.4 ⫾ 26 45 23.7 ⫾ 21 ⫺12.1 ⫺22.9, ⫺2.5‡ 0.5
Week 29 44 17.0 ⫾ 22 44 24.3 ⫾ 28 ⫺12.7 ⫺24.0, ⫺1.9‡ 0.5
Main complaint, VAS, mm
Baseline 56 55.2 ⫾ 22 53 56.1 ⫾ 21
Week 5 53 37.7 ⫾ 22 50 50.2 ⫾ 22 ⫺11.7 ⫺20.4, ⫺2.7‡ 0.5
Week 17 49 38.5 ⫾ 22 45 53.0 ⫾ 26 ⫺13.0 ⫺22.5, ⫺2.8‡ 0.5
Week 29 44 35.6 ⫾ 22 44 49.1 ⫾ 30 ⫺12.8 ⫺26.5, 1.8 0.5
Starting stiffness, VAS, mm
Baseline 56 51.2 ⫾ 28 53 46.8 ⫾ 28
Week 5 53 33.3 ⫾ 25 50 41.3 ⫾ 29 ⫺12.1 ⫺23.5, ⫺2.8‡ 0.5
Week 17 49 32.9 ⫾ 33 45 43.0 ⫾ 32 ⫺14.0 ⫺28.1, ⫺0.6‡ 0.4
Week 29 44 44.3 ⫾ 26 44 44.8 ⫾ 30 ⫺4.8 ⫺17.5, 7.7 0.2
ROM, degrees, flex–ext
Baseline 56 101.3 ⫾ 20 53 100.0 ⫾ 30
Week 5 53 115.8 ⫾ 10 50 98.7 ⫾ 23 16.0 8.1, 22.6‡ 1.0
Week 17 49 116.5 ⫾ 13 45 104.4 ⫾ 11 10.7 5.6, 15.2‡ 0.9
Week 29 44 114.3 ⫾ 14 44 104.5 ⫾ 13 8.1 2.7, 13.1‡ 0.6
ROM, degrees exter–inter
Baseline 56 32.1 ⫾ 18 53 27.8 ⫾ 20
Week 5 53 45.5 ⫾ 11 50 29.0 ⫾ 15 12.1 6.1, 17.3‡ 0.9
Week 17 49 43.1 ⫾ 14 45 32.0 ⫾ 12 6.5 ⫺5.4, 18.9 0.5
Week 29 44 39.4 ⫾ 14 44 30.6 ⫾ 12 2.8 ⫺1.1, 7.1 0.2

* Mean scores ⫾ SDs per group. Mean differences adjusted for baseline values as analyzed by analysis of covariance. On visual analog scales (VASs),
negative signs indicate improvement. On range of motion (ROM), positive signs indicate improvement. 95% CI ⫽ 95% confidence interval; flex ⫽
flexon; ext ⫽ extension; exter ⫽ external rotation; inter ⫽ internal rotation.
† Adjusted for baseline values.
‡ P ⬍ 0.05.

of exercise therapy, the number of treatment sessions ex- results were found to be similar: on general improvement
ceeds 9. However, in literature no consensus is reached on as perceived by the patient, 50% of the patients in our
the preferred number and frequency of sessions of exercise study reported to be improved with exercise therapy ver-
therapy (9). Furthermore, the exercise therapy protocol sus 47% of the patients in the van Baar study. Therefore,
that was applied in our trial was an adaptation of the we believe that the exercise protocol that was applied in
exercise therapy protocol of van Baar et al (8). our study was an effective treatment. We believe that both
The protocol of van Baar was proven to be effective (8,9). treatment approaches examined in this trial are common
When we compared the results of the study of van Baar et in the treatment of patients with OA of the hip. Manual
al with the outcome in the exercise group in our trial, the

Figure 3. Results on range of joint motion from flexion to exten-

Figure 2. Results of the Harris Hip Score. sion.
728 Hoeksma et al

therapy was performed according to a standardized proto- 19. Dougados M. Clinical assessment of osteoarthritis in clinical
col, which was developed with experts in the field. The trials. Curr Opin Rheumatol 1995;7:87– 8.
20. Harris H. Traumatic arthritis of the hip after dislocation and
protocol of van Baar was also designed to reflect current
acetabular fracture: treatment by mold arthroplasty. J Bone
practice in exercise therapy. This corroborates the external Joint Surg Am 1969;4:737–55.
validity of our study. 21. Gerhardt JJ, Ripstein J. Measuring and recording of joint
In conclusion, the experimental group received manual motion: instrumentation and techniques. Toronto: Hogrefe
therapy including manipulations and vigorous stretching and Huber; 1990.
22. Newell DJ. Intention to treat analysis: implications for quan-
and the control group received standard exercise therapy, titative and qualitative research. Int J Epidemiol 1992;21:837–
which may have included stretching but did not include 41.
manipulation. The manual therapy program was found to 23. Kaziz LE, Andersson JJ, Meenan RF. Effect sizes for interpret-
be superior to the exercise therapy program. Furthermore, ing changes in health status. Med Care 1989;27:S178 – 89.
24. Janda V. Muscle functions testing. London: Butterworth-
the effects of the manual therapy program lasted up to 6
Heinemann; 1983.
months after the end of therapy. Thus, in patients with OA 25. Hoving JL, Koes BW, de Vet HC, van der Windt DA, Assen-
of the hip, manual therapy seems to be a suitable treatment delft WJ, Van Mameren H, et al. Manual therapy, physical
option. therapy, or continued care by a general practitioner for pa-
tients with neck pain: a randomized, controlled trial. Ann
Intern Med 2002;21:713–22.

REFERENCES
APPENDIX A: MANUAL THERAPY IN
1. Robertsson O, Wingstrand H, Onnerfalt O. Intracapsular pres- OSTEOARTHRITIS OF THE HIP: A PROTOCOL
sure and pain in coxarthrosis. J Arthroplasty 1995;5:632–5.
2. Arnoldi C. Vascular aspects of osteoarthritis. Acta Orthop Developed by the Department of Physical Therapy, Leyen-
Scand Suppl 1994;S2:61– 82. burg Hospital and the Netherlands Institute for Health
3. Duthrie R, Bentley G. Mercer’s orthopaedic surgery. 9th edi-
tion. Oxford: Oxford University Press;1996. Services Research
4. Hochberg MC, Altman RD, Brandt KD, Clark MC, Dieppe PA.
Griffin MR, et al. Guidelines for the medical management of Frequency of sessions and duration of episode
osteoarthritis. Arthritis Rheum 1995;11:1535– 40.
5. Cyriax JH. Illustrated manual of orthopedic medicine. 2nd The duration of a session is 25 minutes. The frequency is
edition. London: Butterworth-Heinemann Medical; 1996. twice a week at a total of 9 treatments.
6. Minor MA. Exercise in the management of osteoarthritis of the
knee and hip. Arthritis Care Res 1994;4:198 –204.
7. Hofmann DF. Arthritis and exercise. Prim Care 1993;20:895– Treatment protocol
910. Manual therapy (manipulation and stretching) is particu-
8. Van Baar ME, Assendelft WJ, Dekker J, Oostendorp RA, Bi-
jlsma JW. The effectiveness of exercise therapy in patients
larly aimed at the improvement of elasticity of the joint
with osteoarthritis of the hip or knee: a randomized clinical capsule and the surrounding muscles.
trial. J Rheumatol 1998;25:2432–9.
9. Fransen M, McConnell S, Bell M. Therapeutic exercise for Muscle stretching. Muscle stretching is an integrated
people with osteoarthritis of the hip or knee: a systematic
review. J Rheumatol 2002;29:1737– 45.
part of the manual therapy program. Each session starts
10. Altman R, Alarcon G, Appelrouth D, Bloch D, Borenstein D, with stretching of shortened muscles. The following mus-
Brandt K. The American College of Rheumatology criteria for cle (groups) are stretched: m. iliopsoas, m. quadriceps
the classification and reporting of osteoarthritis of the hip. femoris, m. tensor fascia latae, m. sartorius, m.m. adduc-
Arthritis Rheum 1991;34:505–14. tors and m. gracilis (1). Starting posture is a supine posi-
11. Altman DG, Schulz KF, Moher D, Egger M, Davidoff F, El-
bourne D, et al. The revised CONSORT statement for report- tion. The patient has to experience a stretching sensation.
ing randomized clinical trials: explanation and elaboration. Actual stretching is applied for 8 to 10 seconds. Repeat
Ann Intern Med 2001;134:663–94. stretching of each muscle (group) 2 times. Total time:
12. Dougados M, Gueguen A, Nguyen M, Berdah L, Lequesne M, 10 –15 minutes.
Mazieres B, et al. Radiographic features predictive of radio-
graphic progression of hip osteoarthritis. Rev Rheum Engl Ed
1997;64:795– 803. Manipulation. Manipulation is performed according to
13. Ravaud P, Dougados M. Radiographic assessment in osteoar- a traction manipulation technique (2). The therapist’s
thritis. J Rheumatol 1997;24:786 –91. hands are placed just above the ankle joint. All manipula-
14. Fryer JA, Mudge JM, McLaughlin PA. The effect of talo crural
joint manipulation on range of motion of the ankle. J Manip-
tions are performed in slight abduction to avoid slamming
ulative Phys Ther 2002;25:384 –90. of the femoral head into the acetabular surface. The first
15. Guyatt GH, Norman GR, Juniper EF, Griffith LE. A critical traction manipulation is performed in the maximum
look at transition ratings. J Clin Epidemiol 2002;55:900 – 8. loosed packed position of the hip joint (2). With each
16. Brazier JE, Harper R, Jones NM, O’Cathain A, Thomas KJ, following manipulation, the hip joint is placed in a more
Usherwood T, et al. Validating the SF-36 health
questionnaire: new outcome measure for primary care. BMJ limited position (which differs per patient). In total, a
1992;305:160 –5. maximum of 5 manipulations can be applied. The final
17. Jensen MP, Miller M, Fisher LD. Assessment of pain during manipulation is performed in the most limited position of
medical procedures: a comparison of three scales. Clin J Pain the hip joint. In between manipulations, active assisted
1998;14:343–9.
18. Fries J, Spitz P, Kraines RG, Holman HR. Measurement of
motions of the hip joint are performed for relaxation.
patient outcome in arthritis. Arthritis Rheum 1980;23:146 – To evaluate the success of manipulation, after each ma-
52. nipulation “end feel” of the hip joint is tested using a
Manual Therapy Versus Exercise Therapy in Hip OA 729

traction test and by passive hip flexion. This is compared Muscle function. Mainly active exercises have to be ap-
with the contralateral hip. When end feel of the treated hip plied to improve muscle function. Exercises consist of
is similar to the contralateral hip, optimal result is con- muscle strengthening exercises with the use of weight or
cluded. strengthening equipment. Endurance is trained by walking
on a treadmill or cycling on a home trainer. Finally, coor-
Patient education and advice. The promotion of phys- dination is trained through walking exercises with in-
ical activities in general is of importance. Main goal is to creased complexity and through balancing exercises.
couple improvement in joint function with physical activ-
ities, such as walking, cycling, and swimming. Further- Range of motion. If regarded necessary, range of joint
more, instruction about load ability of the hip joint has to motion can be increased through both passive and active
be provided. exercises. Active exercises should have the upper hand.
Active exercises consist of 3-dimensional motions of the
Appendix References hip joint that go beyond the range of joint motion that most
patients use in activities of daily living. These exercises
1. Evjenth O, Hamberg J. Autostretching: the complete manual of can be performed in weight-bearing and non–weight-bear-
specific stretching. Chattanooga (TN): Chattanooga Corp.; 1991. ing positions. In addition, these exercises can be applied
2. Cyriax JH. Illustrated manual of orthopedic medicine. 2nd ed. in different positions, such as during standing, sitting on a
London: Butterworth-Heinemann Medical; 1996.
chair, and while lying down.
Passive exercises contain passive movement of the hip
APPENDIX B: EXERCISE THERAPY IN and stretching exercises according to Evjenth and Ham-
OSTEOARTHRITIS OF THE HIP: A PROTOCOL berg. Postures and starting positions for stretching exer-
cises can be found in the book of Evjenth and Hamberg (2).
Developed by the Department of Physical Therapy, Leyen-
burg Hospital and the Netherlands Institute for Health Pain. If regarded necessary, exercises for pain relief can
Services Research be applied. Pain relief is also achieved through active joint
motion exercises and through stretching exercises. In ad-
Introduction dition, second and third degree traction in the maximum
This is a summary of the exercise protocol. The protocol is loosed packed position of the hip can be applied (2).
an adaptation of the protocol of Van Baar et al (1). In
addition, the book of Evjenth and Hamberg is followed on Walking ability. Walking ability is trained by specific
muscle stretching techniques (2). All participating physi- walking exercises with adjustment of gait pattern, use of
cal therapists are instructed in training sessions. These walking aids, and instruction on climbing of stairs.
training sessions will be repeated every 3 months.
Patient education, advice, and home exercises. The
Frequency of sessions and duration of episode promotion of exercise in general is of great importance;
The duration of a session is 25 minutes. The frequency is such activities as walking, cycling, and swimming are
twice a week at a total of 9 treatments. recommended. Concerning home management and social
activities, these are specifically focused to take an active
approach to pain, instead of taking rest and sitting down.
Treatment protocol
Avoidance of prolonged static load and instruction on load
The exercise program is tailored to the individual patient’s ability of the hip should be emphasized. Instructions for
needs by the therapist. The first session is used to compile home exercises, derived from the specific exercises as
exercise therapy treatment goals by questioning, physical performed during the treatment sessions, are provided.
examination, and observation of walking ability. It is of
great importance to identify specific impairments and dis-
abilities that are of high priority to the patient. Appendix References
There are 4 main treatment goals on which exercise
therapy focuses: 1) increase of muscle function, including 1. Van Baar ME, Assendelft WJ, Dekker J, Oostendorp RA, Bijlsma
endurance, strength, and coordination; 2) improvement of JW. The effectiveness of exercise therapy in patients with os-
teoarthritis of the hip or knee: a randomized clinical trial.
range of motion; 3) decrease of pain; and 4) improvement J Rheumatol 1998;25:2432–9.
of walking ability. Furthermore, education and advice 2. Evjenth O, Hamberg J. Autostretching: the complete manual of
need to be provided to the patient. specific stretching. Chattanooga (TN): Chattanooga Corp.; 1991.