Does Unsupported Upper Limb

Exercise Training Improve

Symptoms and Quality of Life for
Patients With Chronic Obstructive
Pulmonary Disease?
Anne E. Holland, BAppSc; Catherine J. Hill, BAppSc; Elizabeth Nehez, BAppSc;
George Ntoumenopoulos, PhD

■ PURPOSE: Many patients with chronic obstructive pulmonary disease K E Y W O R D S

(COPD) report dyspnea and fatigue when performing upper limb
activities. Unsupported upper limb training has been shown to improve chronic obstructive pulmonary
upper limb endurance, but its effects on symptoms and quality of life disease
have not been examined. The aim of this study was to compare the
effects of upper limb and lower limb training with lower limb training exercise
alone on exercise capacity, symptoms, and quality of life with COPD.
upper limb
■ METHODS: For this study, 38 patients with moderate to severe COPD
were randomly allocated to unsupported upper limb endurance training quality of life
or to a control group that completed a sham training task. All the
patients underwent lower limb endurance training. The 6-minute walk
test, the Incremental Unsupported Upper Limb Exercise Test, and the
Chronic Respiratory Disease Questionnaire (CRQ) were completed
before training and then 6 weeks afterward. Both patients and
assessors were blinded to group allocation.
From the Department of
■ RESULTS: All the patients reported symptoms associated with upper limb
Physiotherapy, Alfred Hospital (Ms
activities on the initial CRQ. Both groups showed significant
Holland, Ms Nehez, Mr
improvements in all domains of the CRQ and in the 6-minute walk test Ntoumenopoulos); the Department of
after training. Only the upper limb training group showed improvement Physiotherapy, Austin Health (Ms Hill);
in upper limb endurance time (57
75 vs 2
58 seconds; P  .02). and the School of Physiotherapy,
There were no significant differences between the groups for 6-minute University of Melbourne (Ms Holland),
walk test or any domain of the CRQ. Melbourne, Victoria, Australia.
Address correspondence to:
■ CONCLUSIONS: Unsupported upper limb training for patients COPD
Anne E. Holland, BAppSc, Department
improves upper limb exercise capacity, but has no additional effect on of Physiotherapy, Alfred Hospital,
symptoms or quality of life, as compared with leg training alone. This Commercial Road, Melbourne,
type of upper limb training may not adequately address the complex Australia 3004 (e-mail: a.holland@
interaction between respiratory mechanics and upper limb function. alfred.org.au).

The upper extremities play an important role in many upper limb muscles also may be required to act as
activities of daily living such as bathing, dressing, hang- accessory muscles of respiration. During unsupported
ing out the wash, and gardening. Patients with chronic arm exercise, the participation of these muscles in ven-
obstructive pulmonary disease (COPD) frequently tilation decreases, and there is a shift of respiratory
experience marked dyspnea and fatigue when per- work to the diaphragm. This is associated with thora-
forming these simple tasks.1 Upper limb activities com- coabdominal dyssynchrony, severe dyspnea, and termi-
monly require unsupported arm exercise, which poses nation of exercise at low workloads, especially in
a unique challenge for patients with COPD whose patients with more severe bronchial obstruction.2-4

422 / Journal of Cardiopulmonary Rehabilitation 2004;24:422-427

 Upper limb exercise training for patients with COPD formed in a sitting position with a 500-g stick until each
has been shown to increase upper limb work capacity, exercise could be performed for 3 minutes. The weight
improve endurance, and reduce oxygen consumption at then was increased by 0.5-kg increments to maintain a
a given workload.5-8 Unsupported arm training results in load that achieved a perceived exertion rating of 12 to
greater improvements in endurance and oxygen con- 14 on the Borg scale13 and a dyspnea rating of 3 on the
sumption than supported arm training with a cycle modified Borg scale.14 The exercises were supervised by
ergometer.6 The effects of upper limb training on symp- a physiotherapist twice a week. The patients were
toms, functional status, and quality of life, however, instructed to perform the same exercises daily at home,
remain unclear. Only two authors have addressed these and to record their exercise in a home exercise diary.
questions, and the results of their studies are conflict- The control group performed the Purdue pegboard
ing,5,7 possibly because of the low patient numbers. test of finger dexterity.15 This test consists of four timed
The effectiveness of lower limb (LL) exercise training tasks in which as many pins as possible are placed on
for patients with COPD has been well documented, a pegboard within a defined time. The test is performed
with consistent and clinically significant improvements in a sitting position with the arms supported on the
in exercise capacity, symptoms, and quality of life.9,10 table, and thus would be expected to have no effect on
Training effects are specific to the limb trained,11 and it unsupported upper limb endurance. However, because
seems that upper limb training alone is less effective in normal individuals performance of the tasks
than LL training in improving overall function.7 The improves with practice,16 the pegboard test was
benefits of combined upper limb and LL training, how- regarded as sufficiently motivating intrinsically to facili-
ever, are less well defined. It is not clear whether the tate repeated performance over the 6-week period. The
addition of upper limb training improves symptoms and control task was supervised by a physiotherapist twice
quality of life more than leg training alone. a week. The two groups performed the training tasks in
The aim of this study was to compare the effects of physically separate locations at the same time and were
combined upper limb and LL training with LL training instructed not to discuss the nature of the training with
alone on upper limb and LL exercise capacity, symp- one another until the end of the program.
toms, and quality of life in patients with COPD. Upper limb exercise tolerance was assessed before
training and 6 weeks afterward with the Incremental
Unsupported Upper Limb Exercise Test.17,18 The
METHODS patients lifted a 200-g stick in time with a metronome at
a rate of 30 lifts per minute. The height of the lift was
Patients with severe to very severe COPD (GOLD stage increased by 15 cm every minute according to levels on
3 or 4;12 forced expiratory volume in 1 second [FEV1]  a wall chart. Once the patient reached his or her maxi-
50% predicted) referred to a pulmonary rehabilitation mal vertical height with the lift, the weight of the bar
program at a tertiary hospital were recruited. These was increased to 0.5 kg, 1 kg, 1.5 kg, and 2 kg in each
patients were receiving optimal medical therapy and successive minute.
were clinically stable at the time of their entry into the The patients were instructed to continue the test as
program. None of the patients had significant bron- long as possible. The test was terminated by the patient
chodilator reversibility at lung function testing. The because of dyspnea or arm fatigue, or by the therapist
study was approved by the institutional ethics commit- if the patient was unable to continue performing the
tee, and each patient gave written informed consent test correctly. Oxygen saturation and pulse rate were
before participation. monitored continuously during the test with a pulse
All the patients underwent a standard pulmonary oximeter and ear probe (Oxypleth, Novametrics;
rehabilitation program involving 6 weeks of LL Wallingford CT), and blood pressure was measured
endurance training. Lower limb training involved before and after the test. The duration of the test, the
30 minutes of treadmill walking or stationary cycling as Borg scale scores for breathing and arm fatigue before
well as stair training. Intensity of exercise was pro- and after the test, and the reason for termination of the
gressed to maintain a perceived exertion rating of 12 to test were recorded.
14 on the Borg scale13 and a dyspnea rating of 3 on the Lower limb functional exercise capacity was meas-
modified Borg scale.14 The patients attended two super- ured at the same time points with the 6-minute walk
vised sessions each week, and a home exercise pro- test using standardized encouragement.19 A minimum
gram was given. of two tests were performed, with a break of at least
In addition, the patients were randomly allocated to 30 minutes between tests. A third trial was performed if
perform either unsupported upper limb endurance train- the distances recorded differed by more than 10%. The
ing or a control training task. The unsupported upper highest recorded distance was reported.
limb endurance training followed a previously described Health-related quality of life was assessed with the
protocol.6 It consisted of five upper limb exercises per- Chronic Respiratory Disease Questionnaire (CRQ)20

Unsupported Upper Limb Exercise Training / 423

before training and 6 weeks afterward. Scores for the
domains of dyspnea, fatigue, mastery, and emotional
function were recorded. In addition, the five most
important dyspnea-inducing activities, as specified by
each patient, were examined, and the number of these
tasks requiring the use of the upper limbs was
recorded.
The three tests were administered in random order.
Measurements were obtained by an independent data
collector blinded to group allocation. Differences
between groups were compared with independent-
samples t tests or Mann-Whitney U tests as appropriate.
Alpha was set at 0.05.

RESULTS

The baseline characteristics for the two groups are shown

in Table 1. Altogether, 40 patients were recruited and ran-
domized to groups. Two patients withdrew because of
respiratory exacerbations before completing the training
program: one each from the control and upper limb train-
ing groups. Follow-up data were obtained from the sub-
ject who withdrew from the upper limb training group.
One patient in the training group did not attend for final
assessment and could not be contacted. One patient in
the control group whose primary rehabilitation goal was
to improve his golf swing crossed over and began per-
forming upper limb resistive training independently. This
was established at the time of the final assessment, and
his results were included in the control group. No other
failure of blinding was identified by questioning at the
final assessment.
An intention-to-treat analysis was performed with the
available data from 38 patients: 16 in the control group Figure 1. Upper limb and lower limb exercise tests. Results are
mean and standard error of the mean. UL, upper limb.
and 22 in the upper limb training group. There were no
differences between the groups at baseline for any of
the demographic or outcome variables (Table 1).
The results for UL and LL exercise capacity are
shown in Figure 1. Both groups showed a significant
improvement in 6-minute walk test distance after training,
with a mean improvement of 60.9 m in the control
group (95% confidence interval [CI], 35.7-86.2 m) and
Table 1 • SUBJECT DEMOGRAPHICS 68.8 m in the upper limb training group (95% CI, 44.6-
Control Upper Limb P 93.1 m). There was no significant difference between
Number 16 22 the two groups in the degree of improvement in the 6-
Men/women 10 / 6 14 / 8 .90 minute walk test after training. The upper limb training
Age, y 69.4 (6.6) 66.6 (8.4) .45 group showed significant improvement in upper limb
FEV1, L 1.02 (0.32) 0.97 (0.36) .67 endurance after training. However, there was no
FEV1 % 39.8 (10.4) 34.2 (10.2) .11 change in upper limb endurance in the control group.
predicted The mean difference in upper limb endurance between
FER 42.0 (11.3) 38.6 (15.7) .18
the groups after training was 55.3 seconds (95% CI, 8.3-
BMI, kg/m2 24.3 (5.0) 22.9 (4.6) .46
102.4 seconds; P  .02). There was a trend toward a
Data are mean (SD); P values are comparison of groups at baseline--no reduced Borg score for dyspnea in the upper limb train-
significant differences.
FEV1, forced expiratory volume in 1 second; FER, forced expiratory ratio ing group at the end of the test (Mean
SD 4.0
2.0
(FEV1/FVC); BMI, body mass index. in the upper limb training group vs 2.9
1.6 in the

424 / Journal of Cardiopulmonary Rehabilitation 2004;24:422-427

control group; P  .06) despite an increase in the dura- DISCUSSION
tion of upper limb exercise. There were no differences
between the groups in arm fatigue scores. Whereas 34 The findings of this study show that unsupported upper
patients terminated the Incremental Unsupported limb endurance training for patients with severe to very
Upper Limb Exercise Test because of arm fatigue, 4 severe COPD results in improved upper limb exercise
patients terminated the test because of dyspnea. capacity, a finding that accords with previous studies.5,6
The CRQ dyspnea domain showed that all the However, no additional improvements in health-related
patients experienced dyspnea with at least one upper quality of life or symptoms over leg training alone could
limb activity, and 80% of the patients reported dyspnea be identified. Although there was a trend toward
with two or more upper limb tasks. Frequently reported reduced dyspnea at upper limb testing, this did not
activities included carrying (for 42% of patients), show- translate into an identifiable reduction in dyspnea dur-
ering (54%), vacuuming (32%) and gardening (26%). ing activities of daily living.
Both groups showed statistically significant improve- Upper limb endurance training is a recommended
ments for all domains of the CRQ after rehabilitation. component of pulmonary rehabilitation programs.9,10,21
However, there were no differences between the groups These recommendations are based on impaired upper
for any of the four domains (Fig. 2). The dyspnea limb performance of patients with COPD,3 reduction in
domain also was analyzed only for the activities that proximal muscle strength,22 and the favorable physio-
involved the upper limbs. The improvement in dyspnea logic outcomes associated with upper limb training.6,8,23
for upper limb activities after rehabilitation was similar However only two previous studies, with small patient
to that seen for the total dyspnea score, and there was numbers, have evaluated the effect of upper extremity
no significant difference between the groups when only training on symptoms or functional outcomes. The addi-
upper limb activities were included (mean difference tion of upper limb training to LL training has been shown
between the upper limb training group and the control to improve self-efficacy significantly on the Bandura
group, 0.19 points; 95% CI, 0.31-0.68; P  .42). scale,7 which measures quality of life, self-confidence,

Figure 2. Chronic respiratory

disease questionnaire. Results are
mean and standard error of the
mean. UL, upper limb.

Unsupported Upper Limb Exercise Training / 425

and self-esteem. However, a lack of self-efficacy Given the close association between the muscles of
improvement in the group that performed LL training the shoulder girdle and the ribcage, it is likely that res-
alone is not consistent with other findings.24 Other inves- piratory mechanics play an important role in determin-
tigators were not able to show any reduction in dyspnea ing the response to upper limb exercise in individuals
on a functional test simulating activities of daily living with COPD. Celli et al3 noted a strong association
(dishwashing, cleaning, and grocery shelving) despite between dyspnea during upper limb exercise and the
improved upper limb endurance after training.5 onset of thoracoabdominal dyssynchrony, which
To date, there is no evidence that upper limb train- occurred during arm, but not leg, exercise. In five of the
ing is able to improve these important patient outcomes seven patients who terminated their upper limb
above what can be provided by LL training alone. endurance test because of dyspnea, a dyssynchronous
Although the improvements in upper limb strength seen breathing pattern was observed within 2 minutes of
after training suggest that it remains a component of exercise commencement. In the remaining five patients,
pulmonary rehabilitation programs, more work is exercise was terminated at higher workloads because of
required to establish whether there are indeed func- arm fatigue, and was not associated with dyssynchro-
tional and symptom-related benefits. nous breathing. Patients with dyssynchrony had more
The current study used a test of upper limb severe airflow obstruction, as measured by FEV1/forced
endurance that involved stereotypical, repetitive move- vital capacity, and their duration of arm exercise was
ments closely related to the training task used. This significantly shorter.
strategy, which has been used for other studies in Because only four patients in the current study ter-
which increased upper limb endurance was found,5,11,23 minated the test because of dyspnea, it was not possi-
may be an example of task-specific training that allows ble to determine any differences between this small
patients to perform the test more easily. Epstein et al23 group and the rest of the patients in terms of demo-
postulated that improved upper limb endurance after graphics or exercise performance. Our patients, how-
such training is attributable to increased coordination of ever, had less severe lung disease overall than the
accessory muscle action rather than an aerobic training group described by Celli et al3 (mean, FEV1, 0.99 L, as
effect, suggesting that carryover to more complex upper compared with 0.68 L). It is possible that the functional
limb actions may not occur. implications of upper limb training may differ for
Most of the dyspnea-inducing tasks in the CRQ patients with severe COPD who exhibit dyssynchro-
reported by the patients in the current study were com- nous breathing during upper limb exercise, as com-
plex upper limb activities incorporating a variety of pos- pared with those in whom respiratory mechanics are
tures and muscle groups and frequently performed in a comparatively less affected.
standing position. The static muscle work required to sta- The instrument used to detect changes in symptoms
bilize the trunk during upper limb movements may be and health-related quality of life, the CRQ, may have
greater in the standing position than in the sitting position been less responsive to changes in functional activities
with the trunk supported by a chair. Upper limb training performed outside the rehabilitation program.
performed in the sitting position therefore may not ade- However, this tool has been recommended for evalua-
quately address the complex coordination involved in tion of the effects that dyspnea has on functional sta-
functional tasks performed by patients with COPD. tus and daily activities.9 In addition, the dyspnea
Further evidence that the specific nature of the upper domain of this questionnaire requires that patients
limb task influences the perception of dyspnea is pro- specify their five most important dyspnea-inducing
vided by a recent study examining metabolic and venti- activities, a feature allowing confirmation that all the
latory responses to functional upper limb tasks per- current patients did indeed experience significant and
formed by patients with COPD.25 A task involving distressing symptoms during one or more upper limb
coordination, such as screwing in a light bulb, provoked tasks. Nonetheless, it may be that the CRQ is not sen-
greater dyspnea than the repetitive, stereotypic task of sitive to changes in upper extremity exercise capacity,
lifting pots, despite lower peak exercise oxygen con- and that as a result, no improvements in symptoms or
sumption. The authors postulate that such coordination quality of life were evident. It is possible that instru-
tasks are more dyspnea-inducing because of breath hold- ments with a more specific focus on functional status,
ing or a reduction in blood flow to muscle during static such as the Pulmonary Functional Status Scale, may
contractions. The upper limb test did not incorporate a provide additional information. However, as yet, there
coordination component or static muscle contraction, are limited data regarding its responsiveness to exer-
which may explain the low incidence of dyspnea. Only cise training.26 Additional information also could be
four patients in the current study terminated the provided by a upper limb endurance test that more
Incremental Unsupported Upper Limb Exercise Test closely reproduces activities of daily living. However,
because of breathlessness, with the remainder terminat- to date, a reliable and valid test of this nature has not
ing the test because of arm fatigue. been developed.

426 / Journal of Cardiopulmonary Rehabilitation 2004;24:422-427

 It is possible that including a larger number of patients with severe chronic airflow obstruction. Chest. 1993;103:
patients in this study may have shown significant dif- 1397-1402.
7. Lake FR, Henderson K, Briffa T, et al. Upper-limb and lower-limb
ferences in symptoms or quality of life. It was noted, exercise training in patients with chronic airflow obstruction.
however, that the data showed no trends with very high Chest. 1990;97:1077-1082.
P values (P  .5) for all of the CRQ domains. In addi- 8. Couser JI Jr, Martinez FJ, Celli BR. Pulmonary rehabilitation that
tion, the current study represents the largest trial of includes arm exercise reduces metabolic and ventilatory require-
upper limb training for patients with COPD. Previous ments for simple arm elevation. Chest. 1993;103:37-41.
9. American Thoracic Society. Pulmonary rehabilitation–1999. Am J
controlled trials upon which training recommendations Respir Crit Care Med. 1999;159:1666-1682.
have been made9,10 have included 7 to 32 patients.6,11 It 10. British Thoracic Society Standards of Care Subcommittee on
may be useful to repeat this study with larger patient Pulmonary Rehabilitation. Pulmonary Rehabilitation. Thorax.
numbers to validate the findings. However, the need to 2001;56:827-834.
do so shows that any effect size related to upper limb 11. Belman MJ, Kendregan BA. Exercise training fails to increase
skeletal muscle enzymes in patients with chronic obstructive pul-
training is likely to be small, contrasting with the rela- monary disease. Am Rev Respir Dis. 1981;123:256-261.
tively few patients required to find an effect of LL train- 12. Fabbri LM, Hurd SS. Global strategy for the diagnosis, management,
ing on symptoms and quality of life experienced by and prevention of COPD: 2003 update. Eur Respir J. 2003;22:1-2.
patients with COPD.9 13. Borg GA. Perceived exertion. Exerc Sport Sci Rev. 1974;2:131-153.
In conclusion, this randomized double-blind controlled 14. Borg G. Psychophysical bases of perceived exertion. Med Sci
Sports Exerc. 1982;14:377-381.
trial shows that although unsupported upper limb train- 15. Tiffin J, Asher EJ. The Purdue Pegboard: norms and studies of
ing enhances upper limb exercise capacity for patients reliability and validity. J Appl Psychol. 1948;32:234-247.
with moderate to severe COPD, no additional improve- 16. Reddon JR, Gill DM, Gauk SE, et al. Purdue Pegboard: test-retest
ment in symptoms or health-related quality of life beyond estimates. Percept Mot Skills. 1988;66:503-506.
that obtained with LL training alone is evident. Further 17. Takahashi T, Jenkins SC, Strauss GR, et al. A new unsupported
upper limb exercise test for patients with chronic obstructive
research is required to determine the clinical significance pulmonary disease. J Cardiopulm Rehabil. 2003;23:430-437.
of upper limb training for patients with COPD. 18. McKeough ZJ, Allison JA, Bye PTP. Arm exercise capacity and
dyspnoea ratings in patients with chronic obstructive pulmonary
Acknowledgments disease. J Cardiopulm Rehabil. 2003;23:281-225.
19. ATS statement: guidelines for the six-minute walk test. Am J
This study was supported by an Alfred Research Respir Crit Care Med. 2002;166:111-117.
Trust Small Projects Grant. The authors thank the 20. Guyatt GH, Berman LB, Townsend M, et al. A measure of quality
Occupational Therapy Department at the Alfred of life for clinical trials in chronic lung disease. Thorax. 1987;42:
Hospital for assistance with data collection. 773-778.
21. Skeletal muscle dysfunction in chronic obstructive pulmonary
disease: a statement of the American Thoracic Society and
References European Respiratory Society. Am J Respir Crit Care Med. 1999;
159:S1-S40.
1. Tangri S, Woolf CR. The breathing pattern in chronic obstructive 22. Gosselink R, Troosters T, Decramer M. Distribution of muscle
lung disease during the performance of some common daily weakness in patients with stable chronic obstructive pulmonary
activities. Chest. 1973;63:126-127. disease. J Cardiopulm Rehabil. 2000;20:353-360.
2. Martinez FJ, Couser JI, Celli BR. Respiratory response to arm ele- 23. Epstein SK, Celli BR, Martinez FJ, et al. Arm training reduces the
vation in patients with chronic airflow obstruction. Am Rev Respir VO2 and VE cost of unsupported arm exercise and elevation in
Dis. 1991;143:476-480. chronic obstructive pulmonary disease. J Cardiopulm Rehabil.
3. Celli BR, Rassulo J, Make BJ. Dyssynchronous breathing during 1997;17:171-177.
arm but not leg exercise in patients with chronic airflow obstruc- 24. Ries AL, Kaplan RM, Limberg TM, et al. Effects of pulmonary
tion. N Engl J Med. 1986;314:1485-1490. rehabilitation on physiologic and psychosocial outcomes in
4. Criner GJ, Celli BR. Effect of unsupported arm exercise on venti- patients with chronic obstructive pulmonary disease. Ann Intern
latory muscle recruitment in patients with severe chronic airflow Med. 1995;122:823-832.
obstruction. Am Rev Respir Dis. 1988;138:856-861. 25. Velloso M, Stella SG, Cendon S, et al. Metabolic and ventilatory
5. Ries AL, Ellis B, Hawkins RW. Upper extremity exercise train- parameters of four activities of daily living accomplished with
ing in chronic obstructive pulmonary disease. Chest. 1988;93: arms in COPD patients. Chest. 2003;123:1047-1053.
688-692. 26. Haggerty MC, Stockdale-Woolley R, ZuWallack R. Functional sta-
6. Martinez FJ, Vogel PD, Dupont DN, et al. Supported arm exer- tus in pulmonary rehabilitation participants. J Cardiopulm
cise vs unsupported arm exercise in the rehabilitation of Rehabil. 1999;19:35-42.

Unsupported Upper Limb Exercise Training / 427