Methods

Study design: randomized controlled trial

Participants

Clinically stable COPD (n = 20), never received arm exercise training previously

Baseline characteristics

Experimental

• Age (yrs): 61 (14)

• FEV₁(% pred): 27 (16)

• FEV₁(L): 0.65 (0.37

Control

• Age (yrs): 63 (13)

• FEV₁(% pred): 37 (18)

• FEV₁(L): 0.96 (0.44)

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: unsupported arm exercise at home for 5 days per week for 8 weeks (but supervised × 1 day per week).

Control

• Exercise Prescription: no arm training (phone calls × once per week to match attention).

Outcomes

Endurance Upper Limb Capacity: unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (8 weeks)"

• Scale: rings moved in 6 min

• Unit of measure: no. of rings

• Direction: higher is better

Identification

Sponsorship source: Theta Mu Chapter, Sigma Theta Tau, Pittsburgh, Pennsylvania.

Country: USA

Setting: home‐based training

Authors name: Gerene S. Bauldoff

Institution: University of Pittsburgh Medical Center

Address: University of Pittsburgh Medical Center, Pittsburgh, PA 15213.

Notes

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

The randomisation process involved assembling 20 envelopes from which subjects blindly selected 1 that determined their group assignment.

Allocation concealment (selection bias)

Low risk

The randomization process involved assembling 20 envelopes from which the subjects blindly selected 1 that determined their group assignment.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no indication that participants or personnel were blind from the knowledge of which intervention the participant was receiving.

Blinding of outcome assessment (detection bias)
All outcomes

High risk

All phone calls, home testing, and training were performed by the same investigator. There was no blinding of the outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Data reported for all outcomes stated (presumed to be complete for all participants but not explicitly stated).

Selective reporting (reporting bias)

Low risk

Data reported for all outcomes stated.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable COPD (n = 43) with age > 45 yrs

Baseline characteristics

Experimental

• Age (yrs): 71.0 (8.5)

• FEV₁(% pred): 55.9 (17.1)

• Gender (male/female): 17/5

Control

• Age (yrs): 71.5 (7.5)

• FEV₁(% pred): 58.2 (16)

• Gender (male/female): 18/3

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: health education + 16 week exercise programme, twice weekly in laboratory for 1.5 hours and home exercise once per week. Maintenance exercise for a further 12 months, 3 times weekly. Exercise programme was arm resistance training in the laboratory with a cable crossover system using 8 lifts: shoulder shrug, modified latissimus dorsi pull down, overhead pull down, front pull down, front raise, upright row, biceps curl, and triceps extension. Training was initiated at 70% of the 1 repetition maximum (1RM) at a training volume of 2 sets of 8 to 10 repetitions. Training intensity was increased as tolerated to 80% of the 1RM over the first 4 weeks of training and adjusted to maintain an intensity of 80% of the current 1RM for the remaining weeks. Training volume was increased to 3 sets of 8 to 10 repetitions for weeks 5 to 16. For the home training, subjects performed 1 set of 10 to 20 repetitions using dumbbells at a weight (range 2 to 15 lbs) to elicit a rating of perceived exertion equal to 12 (between light and somewhat hard) using the Borg Rate of Perceived Exertion scale (range 6 to 20).

Control

• Exercise Prescription: health education + sham training (gentle chair exercise with stretching of major joints). At home subjects performed similar chair exercises using a videotape.

Outcomes

Upper Limb Strength

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (16 weeks)"

• Scale: sum of scores for all 8 types of lifts

• Unit of measure: kg

• Direction: higher is better

• Data value: endpoint

(NB: this could not be used in the pooled analysis as data was not available for the experimental group alone)

Respiratory Muscle Strength

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (16 weeks)"

• Scale: Pimax

• Unit of measure: cmH₂O

• Direction: higher is better

• Data value: endpoint

Physical Activity Level: Subjective

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (16 weeks)"

• Scale: Functional Performance Inventory

• Range: 0 to 3

• Unit of measure: mean of items

• Direction: higher is better

• Data value: endpoint

Identification

Sponsorship source: the sources of support for this research were The National Institute of Nursing Research R01‐NR08037 and the University of Illinois at Chicago General Clinical Research Center M01‐RR‐13987.

Country: USA

Setting: lab‐based training

Authors name: Margaret K. Covey

Institution: University of Illinois at Chicago, USA

Email: [email protected]

Address: Department of Biobehavioral Health Science, University of Illinois at Chicago, M/C 802, 845 S. Damen Avenue, Chicago, IL, 60612, USA.

Notes

There was a third intervention group who performed upper body resistance training with self‐efficacy training. As this group does not reflect the typical delivery of arm training in a pulmonary rehabilitation programme, the group with arm training and health education only was chosen as the experimental group. However, some data in this paper combined these 2 groups so was not able to be used in the pooled analysis (e.g. upper limb strength where differences between the 2 groups were reported). For dyspnoea, respiratory muscle strength and physical activity level the pooled analysis used a sample size of n = 22, reflecting the experimental group numbers as the paper had reported no significant differences between the 2 upper limb training groups for these outcomes.
This is the same study as the Larson 2014 study but different outcomes are reported.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: "Randomization to group was stratified by gender and disease severity (GOLD stages II, III, and IV) with a software program (biased coin algorithm to ensure equivalent groups)."

Allocation concealment (selection bias)

Low risk

Quote: "This was a concealed allocation process."

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Whilst subjects were not informed of the intent of the 3 groups, they were aware what group they were in.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote: "Data collectors were blinded to the group assignment."

Incomplete outcome data (attrition bias)
All outcomes

High risk

There was a 25% drop‐out rate for the study (n = 7 in each group × 3 groups).

Selective reporting (reporting bias)

High risk

Data has been presented across 2 different publications.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable COPD (n = 23)

Baseline characteristics

Experimental

• Age (yrs): 65 (SE ± 2)

• FEV₁(L): 0.74 (SE± 0.06)

• Gender (male/female): 11/1

Control

• Age (yrs): 65 (SE± 2)

• FEV₁(L): 0.98 (SE± 0.14)

• Gender (male/female): 10/1

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: unsupported arm exercises, supervised, 3 times weekly for 7 to 8 weeks (21 to 24 sessions completed)

Control

• Exercise Prescription: no arm training but included supervised, low‐intensity resistive breathing training using a threshold resistor set at 30% maximal inspiratory pressure, 3 times weekly for 7 to 8 weeks.

Outcomes

Endurance Upper Limb Exercise Capacity: unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post Intervention (8 weeks)"

• Scale: move weighted dowel 10 cm

• Unit of measure: time (s)

Respiratory Muscle Strength

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post Intervention (8 weeks)"

• Scale: Pimax

• Unit of measure: cmH₂O

Identification

Country: USA

Setting: outpatient and inpatient

Authors name: Scott K Epstein.

Institution: New England Medical Center and St. Elizabeth's Medical Center, Tufts University.

Address: Pulmonary and Critical Care Division, New England Medical Center, Box 369, 750 Washington St, Boston, MA 02111.

Notes

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method used to generate the allocation sequence was not provided.

Allocation concealment (selection bias)

Unclear risk

The method used to conceal the allocation sequence was not provided.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no indication that participants or personnel were blind from the knowledge of which intervention the participant was receiving.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

The investigators performing the measurements were blinded to the method of training.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Low numbers of drop‐out with reasons provided for incomplete data (i.e. 1 participant withdrew and 2 participants developed respiratory failure).

Selective reporting (reporting bias)

Low risk

Data reported for all outcomes stated.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Severe to very severe, clinically stable COPD (n = 38)

Baseline characteristics

Experimental

• Age (yrs): 66.6 (8.4)

• FEV₁(% pred): 34.2 (10.2)

• FEV₁(L): 0.97 (0.36)

• Gender (male/female): 14 / 8

Control

• Age (yrs): 69.4 (6.6)

• FEV₁(% pred): 39.8 (10.4)

• FEV₁(L): 1.02 (0.32)

• Gender (male/female): 10/6

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: unsupported arm exercises, twice wkly for 6 weeks supervised and daily at home plus lower limb endurance training, twice wkly for 6 weeks.

Control

• Exercise Prescription: no arm training except sham training of a pegboard finger dexterity task twice wkly for 6 weeks supervised plus lower limb endurance training, twice wkly for 6 weeks.

Outcomes

Peak Upper Limb Exercise Capacity: unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (6 wks)"

• Unit of measure: time (s)

• Direction: higher is better

• Data value: endpoint

Dyspnoea from questionnaire

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post intervention (6 wks)"

• Unit of measure: dyspnoea score from Chronic Respiratory Disease Questionnaire

Health‐Related Quality of Life

• Outcome type: continuous outcome

• Measure names:"Baseline", "Post intervention (6 wks)"

• Direction: higher is better

• Notes: total score from Chronic Respiratory Disease Questionnaire

Identification

Sponsorship source: Alfred Research Trust Small Projects Grant.

Country: Australia

Setting: outpatient and home programme

Authors name: Anne E. Holland

Institution: Department of Physiotherapy, Alfred Hospital, Melbourne and University of Melbourne, Melbourne, Victoria, Australia.

Email: [email protected]

Address: Department of Physiotherapy, Alfred Hospital, Commercial Road, Melbourne, Australia 3004

Notes

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method used to generate the allocation sequence was not provided.

Allocation concealment (selection bias)

Unclear risk

The method used to conceal the allocation sequence was not provided.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no indication that participants or personnel were blind from the knowledge of which intervention the participant was receiving.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote: "Measurements were obtained by an independent data collector blinded to group allocation."

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Two participants did not complete the study so overall completion rate was high at 95%.

Selective reporting (reporting bias)

Unclear risk

Data not reported for arm fatigue during the arm exercise test.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable, moderate to very severe COPD (n = 12), aged > 50 yrs

Baseline characteristics

Experimental

• Age (yrs): 67.8 (7.4)

• FEV₁(% pred): 30.9 (18.9)

• Gender (male/female): 4/1

Control

• Age (yrs): 70.4 (8.5)

• FEV₁(% pred): 34.3 (8.1)

• Gender (male/female): 5/2

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: 3 weekly sessions, lasting 40 min each, for 6 consecutive weeks of arm exercises with warm‐up (5 min) followed by 3 series of 8 repetitions of 2 arm strength exercises starting with load of 80% of 1 repetition maximum test (1RM).

Control

• Exercise Prescription: 3 weekly sessions, lasting 40 min each, for 6 consecutive weeks of bronchial hygiene and respiratory functional rehabilitation.

Outcomes

Upper Limb Strength (pulley and bench press)

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (6 wks)"

• Scale: 1RM

• Unit of measure: kg

Endurance Upper Limb Exercise Capacity: Unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (6 wks)"

• Scale: rings moved in 6 min

• Unit of measure: no. of rings

Identification

Country: Brazil

Authors name: Daniela Ike

Notes

Study reported in Portugese and required translation to English.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method used to generate the allocation sequence was not provided.

Allocation concealment (selection bias)

Unclear risk

The method used to conceal the allocation sequence was not provided.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no indication that participants or personnel were blind from the knowledge of which intervention the participant was receiving.

Blinding of outcome assessment (detection bias)
All outcomes

High risk

There was no indication that there was blinding of the outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Complete data available for all outcomes.

Selective reporting (reporting bias)

Low risk

Data reported for all outcomes stated.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable, mild to very severe COPD (n = 36)

Baseline characteristics

Experimental

• Age (yrs): median IQR 67 (11)

• FEV₁(% pred): 37.8 (16.2)

• FEV₁(L): median IQR 0.8 (0.4)

• Gender (male/female): 9/8

Control

• Age (yrs): median IQR 67 (11)

• FEV₁(% pred): 32.5 (14)

• FEV₁(L): median IQR 0.8 (0.5)

• Gender (male/female): 12/7

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: resistance arm training programme 3 times a week for 6 weeks, for a total of 18 sessions supervised. The following muscle groups were targeted using free weights and a multi‐station gym: biceps brachii, triceps brachii, pectoralis major and minor, latissimus dorsi, deltoids, and rhomboids. Training was started using loads equivalent to 10 to 12 repetition maximum. The loads were increased when the patients could manage 12 repetitions for both sets on 2 consecutive training sessions. The group also performed the standard leg endurance and strength training (6 weeks × 5 wkly if inpatient or 12 weeks × 3 wkly if outpatient).

Control

• Exercise Prescription: sham training that consisted of upper limb flexibility and stretching exercises 3 times a week for 6 weeks supervised. The group also performed the standard leg endurance and strength training (6 weeks × 5 wkly if inpatient or 12 weeks × 3 wkly if outpatient).

Outcomes

Health‐Related Quality of Life

• Outcome type: continuous outcome

• Measure names: "Baseline", "post intervention"

• Direction: higher is better

• Notes: total score from Chronic Respiratory Disease Questionnaire

Peak Upper Limb Exercise Capacity: unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (6 wks)"

• Unit of measure: time (s)

• Direction: higher is better

• Data value: endpoint

Upper Limb Strength (shoulder flexion and abduction)

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (6 wks)"

• Scale: hand‐held dynamometer

• Unit of measure: kg

• Direction: higher is better

Endurance Upper Limb Exercise Capacity: unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (6 wks)"

• Scale: rings moved in 6 min

• Unit of measure: no. of rings

Identification

Sponsorship source: this study was supported by the Ontario Thoracic Society, West Park Healthcare Centre Foundation, Canada Research Chair Program, and the Swedish Heart and Lung Foundation.

Country: Canada

Setting: outpatient and inpatient

Authors name: Tania Janaudis‐Ferreira

Institution: West Park Healthcare Centre, Toronto, ON, Canada

Email: [email protected]

Address: Department of Physical Therapy, University of Toronto, 160‐500 University Ave, Toronto, ON, Canada, M5G1V7

Notes

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: "Patients were randomly (in blocks of 4) assigned to an intervention or control group. Randomization was stratified according to the presence or absence of the use of supplemental oxygen at rest."

Allocation concealment (selection bias)

Low risk

Quote: "The sequence was kept in opaque envelopes by an investigator who was not involved in the recruitment process. These envelopes were drawn by the trainer after the subjects had completed their pre assessment session, allowing for concealed allocation."

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Quote: "the patients remained unaware of the group allocation."

However, the trainers were aware of group allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote: "The outcome assessor and the patients remained unaware of the group allocation."

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Data not available in 14% of participants. Reasons were provided, mostly unrelated to the study.

Selective reporting (reporting bias)

Low risk

Data reported for all outcomes stated.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable, severe COPD (n = 26)

Baseline characteristics

Upper Limb Training Only

• Age (yrs): 62.0 (7.0)

• FEV₁(L): 0.97 (0.29)

• Gender (male/female): 6/0

Control ‐ No Training

• Age (yrs): 65.7 (3.5)

• FEV₁(L): 0.97 (0.29)

• Gender (male/female): 6/1

Combined Upper Limb and Lower Limb Training

• Age (yrs): 66.3 (6.8)

• FEV₁(L): 0.83 (0.25)

• Gender (male/female): 4/3

Lower Limb Training Only

• Age (yrs): 71.8 (3.3)

• FEV₁(L): 0.73 (0.24)

• Gender (male/female): 6/0

Interventions

Intervention characteristics

1. Upper Limb Training Only versus No Training

Experimental

• Exercise Prescription: the participants performed a warm‐up, 20 min of upper‐limb circuit training, and ten min of cool off, 3 times a week for an hour for 8 weeks. The exercises were arm cycle ergometry, throwing a ball against a wall, passing a bean bag over the head, pulling on ropes and pulleys, and playing an accuracy game of moving a ring across a horizontal wire. Each exercise was performed for 40 s followed by 20 s of rest (repeated 3 times in 3 min).

Control

• Exercise Prescription: no training group

2. Combined Upper Limb Training and Lower Limb Training versus Lower Limb Training Alone

Combined

Experimental

• Exercise Prescription: the combined group had 10 min of warm‐up, 15 min of upper‐limb circuit training, 15 min of walking training, and ten min of cool off.

Control (Lower Limb only)

• Exercise Prescription: the lower limb group had ten minutes of warm‐up, 20 min of walking training, and ten min of cool off.

Outcomes

Peak Upper Limb Exercise Capacity: supported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (8 wks)"

• Scale: arm ergometer peak work

• Unit of measure: kpm/min

Respiratory Muscle Strength

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (8 wks)"

• Scale: Pimax

• Unit of measure: cmH₂O

Identification

Country: Australia

Setting: Outpatient

Authors name: Fiona R.Lake

Institution: Sir Charles Gairdner Hospital, Perth, Western Australia.

Address: Sir Charles Gairdner Hospital, Verdun Street, Nedlands West Australia 6009.

Notes

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method used to generate the allocation sequence was not provided.

Allocation concealment (selection bias)

Unclear risk

There is no description of how they performed the allocation concealment.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no indication that participants or personnel were blind from the knowledge of which intervention the participant was receiving.

Blinding of outcome assessment (detection bias)
All outcomes

High risk

There was no indication that there was blinding of the outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Data not available in 7% of participants. Reasons were provided, mostly unrelated to the study.

Selective reporting (reporting bias)

Unclear risk

Data not reported for resting blood pressure.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable, moderate to severe COPD (n = 34), age > 45 yrs

Baseline characteristics

Experimental

• Age (yrs): 72 (9)

• FEV₁(% pred): 54 (17)

Control

• Age (yrs): 71 (8)

• FEV₁(% pred): 56 (17)

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: the upper body resistance training included 3 sets of 8 lifts (plus education) using a cable crossover system twice a week in the laboratory and with hand weights, once a week at home over 16 weeks.

Control

• Exercise Prescription: sham chair exercises which were not aerobic but stretching and toning (plus education). Gentle chair exercises were performed twice a week in the laboratory and once a week at home over 16 weeks.

Outcomes

Physical Activity Level ‐ Objective

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (16 wks)"

• Scale: time spent in moderate intensity activity

• Unit of measure: min per day

• Direction: higher is better

Identification

Sponsorship source: The National Institute of Nursing Research R01‐NR08037 and the University of Illinois at Chicago General Clinical Research Center M01‐RR‐13987.

Country: USA

Authors name: Janet L Larson

Institution: University of Michigan, Ann Arbor

Email: [email protected]

Address: University of Michigan, 400 N Ingalls, Ann Arbor, MI 48109, USA

Notes

There was a third intervention group who performed upper body resistance training with self‐efficacy training. As this group does not reflect the typical delivery of arm training in a pulmonary rehabilitation programme, the group with arm training and health education only was chosen as the experimental group.

This was the same study as the Covey 2012 study but the Larson 2014 paper includes an objective measure of physical activity in 34 COPD patients (which was not part of the Covey 2012 paper).

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: "Randomization was stratified by sex and disease severity (Global Initiative on Obstructive Lung Disease stages II, III, and IV) 8 using a customized computer program that blinded investigators to group assignment."

Allocation concealment (selection bias)

Low risk

Customized computer program would have concealed the allocation.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Whilst all participants were assigned to an active intervention and were not told which was the experimental group, all participants knew which intervention they were performing.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

The data collectors were blinded to group assignment.

Incomplete outcome data (attrition bias)
All outcomes

High risk

42% participants were not included at 16 weeks.

Selective reporting (reporting bias)

High risk

Data has been presented across 2 different publications.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable COPD (n = 14)

Baseline characteristics

Experimental

• Age (yrs): 65 (9.8)

• FEV₁ (% pred): 45 (11.5)

• Gender (male/female): 8/0

Control

• Age (yrs): 68 (10.4)

• FEV₁(% pred): 42 (11.6)

• Gender (male/female): 6/0

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: upper limb strength exercises, 30 min sessions, 3 times per week for 6 weeks. The 6 upper limb strength exercises were elbow flexion, elbow extension, shoulder press, shoulder adduction, diagonal upper limb exercises × 2 types. A progressive load was used for 3 sets of 10 repetitions performed for each of the 6 exercises.The first set was stipulated at 50% of the 10RM obtained during the assessment, followed by the second set at 75% of 10RM, and the third set at 100% of 10RM.

Control

• Exercise Prescription: bronchial hygiene therapy along with diaphragmatic pattern breathing orientations and stretching of the cervical, UL and LL muscles, 45 min sessions, 3 times weekly for 6 weeks.

Outcomes

Endurance Upper Limb Exercise Capacity: unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (6 wks)"

• Scale: 5‐min blackboard erasing task but metabolic data only so not included in meta‐analysis

Upper Limb Strength

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (6 wks)"

• Scale: 5 min weight‐lifting task with 5 kg weight but metabolic data only so not included in meta‐analysis

Identification

Country: Brazil

Setting: outpatient (assumed)

Authors name: Kamilla Tays Marrara,

Institution: Federal University of São Carlos – UFSCar

Email: [email protected]

Address: Federal University of São Carlos – UFSCar, Av. Filomeno Rispoli no. 179, Parque Santa Marta, CEP: 13564‐200, São Carlos, São Paulo, Brazil

Notes

Data not used in meta‐analysis as only metabolic data presented.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method used to generate the allocation sequence was not provided.

Allocation concealment (selection bias)

Unclear risk

The method used to conceal the allocation sequence was not provided.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no indication that participants or personnel were blind from the knowledge of which intervention the participant was receiving.

Blinding of outcome assessment (detection bias)
All outcomes

High risk

There was no indication that there was blinding of the outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes

High risk

24% drop out in the study.

Selective reporting (reporting bias)

Unclear risk

Only metabolic outcomes reported so data not available for meta‐analysis.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable COPD (n = 35)

Baseline characteristics

Experimental

• Age (yrs): median (range): 67 (55 to 75)

• FEV₁(% pred): median (range) 34 (15 to 56)

• Gender (male/female): 7/11

Control

• Age (yrs): median (range): 65 (43 to 75)

• FEV₁(% pred): median (range) 30 (16 to 81)

• Gender (male/female): 7/10

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: unsupported upper limb training, 3 times weekly for 10 weeks. Unsupported exercises were performed with a wooden dowel: shoulder flexion; shoulder flexion/extension in conjunction with elbow flexion/extension; isolated elbow flexion/extension; horizontal abduction/adduction; and shoulder circles. The duration of each exercise was increased by 30 s increments to a total time of 3.5 min. The duration of exercise and/or weight of the dowel were increased weekly as tolerated. All participants also did lower limb training from 10 to 30 min (Borg 3, RPE 12‐14).

Control

• Exercise Prescription: supported upper limb training, 3 times weekly for 10 weeks. Supported upper limb exercises performed on an arm cycle ergometer. A load achieving an RPE of 12 to 14 and dyspnoea rating of approximately 3 was chosen. Workload and exercise duration were increased weekly as tolerated up to 15 min. All participants also did lower limb training from 10 to 30 min (Borg 3, RPE 12 to 14).

Outcomes

Endurance Upper Limb Exercise Capacity: unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (10 wks)"

• Scale: endurance time (s) from weighted dowel test to exhaustion

• No mean data of the groups provided in the paper (only individual data in figures) so not used in the meta‐analysis.

Peak Upper Limb Exercise Capacity: supported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (10 wks)"

• Scale: peak work from arm ergometer test

• No mean data of the groups provided in the paper (only individual data in figures) so not used in the meta‐analysis.

Respiratory Muscle Strength

• Outcome type: continuous outcome

• Measure names: ("Baseline", "Post‐intervention (10 wks)")

• Scale: Pimax

• Unit of measure: cmH₂O

• No mean data of the groups provided in the paper (only individual data in figures) so not used in the meta‐analysis.

Identification

Sponsorship source: no description

Country: USA

Setting: outpatient

Authors name: Fernando J. Martinez

Institution: University of Michigan Medical Center

Email: [email protected]

Address: 1500 East Medical Center Drive, Ann Arbor MI 48109‐0026

Notes

Authors contacted for group data but no response.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method used to generate the allocation sequence was not provided.

Allocation concealment (selection bias)

Unclear risk

The method used to conceal the allocation sequence was not provided.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no indication that participants or personnel were blind from the knowledge of which intervention the participant was receiving.

Blinding of outcome assessment (detection bias)
All outcomes

High risk

There was no indication that there was blinding of the outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Data were not available for 12% of participants but reasons were provided for this, which was unrelated to the study.

Selective reporting (reporting bias)

High risk

Mean group data on main outcomes not provided so data not used in meta‐analysis.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

COPD (n = 33)

No baseline characteristics provided.

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: 10‐week programme (do not know how many times per week). Upper limb and thoracic muscle strength exercise (with lower limb training)

Control

• Exercise Prescription: 10‐week programme (do not know how many times per week). Lower limb endurance and strength training

Outcomes

Health‐Related Quality of Life

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (10 wks)"

• Scale: St George's Respiratory Questionnaire

• No mean data of both groups provided in the abstract so not used in the meta‐analysis

Respiratory Muscle Strength

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (10 wks)"

• Scale: Pimax

• Unit of measure: cmH₂O

• No mean data of the groups provided in the paper so not used in the meta‐analysis

Identification

Country: Japan

Setting: there is no description of setting.

Comments: abstract only so few details

Authors name: H. Matsunaga

Institution: Nippon Steel Yawata Memorial Hospital, Kitakyushu, Japan,

Notes

Abstract only, no email to contact authors so data cannot be included in meta‐analysis.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

No information provided.

Allocation concealment (selection bias)

Unclear risk

No information provided.

Blinding of participants and personnel (performance bias)
All outcomes

Unclear risk

No information provided.

Blinding of outcome assessment (detection bias)
All outcomes

Unclear risk

No information provided.

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Insufficent information provided.

Selective reporting (reporting bias)

Unclear risk

Insufficient information provided.

Other bias

Unclear risk

Insufficient information provided.

Methods

Study design: randomized controlled trial

Participants

Clinically stable COPD (n = 38)

Baseline characteristics

Endurance Training Group

• Age (yrs): 66 (8)

• FEV₁(% pred): 57 (17)

• FEV₁(L): 1.58 (0.6)

• Gender (male/female): 9/2

Resistance Training Group

• Age (yrs): 65 (9)

• FEV₁ (% pred): 60 (17)

• FEV₁(L): 1.58 (0.7)

• Gender (male/female): 4/5

Combined Endurance and Resistance Training Group

• Age (yrs): 66 (6)

• FEV₁(% pred): 54 (16)

• FEV₁(L): 1.53 (0.7)

• Gender (male/female): 6/3

Control

• Age (yrs): 65 (7)

• FEV₁(% pred): 57 (12)

• FEV₁(L): 1.43 (0.4)

• Gender (male/female): 4/5

Interventions

Intervention characteristics

All groups had exercise training 3 times per week for 8 weeks and performed standard leg endurance and strength training during this period.

Experimental (Endurance Training)

• Exercise Prescription: the training consisted of arm cranking and unsupported arm exercise. Arm cranking was performed for 15 min (continuous or intermittent) at 60% of the peak work rate achieved on the peak arm crank test. Intensity progression was made according to breathlessness and arm perceived exertion in order to maintain symptom scores at a moderate level of 3 on the 0 to 10 modified Borg category ratio scale. Unsupported arm exercise was performed for 5 min at 1 level below the maximal level achieved on the unsupported arm test, with the level progressed according to breathlessness and arm perceived exertion in order to maintain symptom scores at a moderate level of 3 on the 0 to 10 modified Borg category ratio scale.

Experimental (Resistance Training)

• Exercise Prescription: the resistance group trained the muscles that are involved both as accessory breathing muscles and as muscles that move the arm on 3 fixed weight machines (i.e. a pec deck, lat pull‐down, and bicep lifts) with resistance loading initially consisting of 2 sets (of 10 repetitions) at 60% of 1RM. Training was progressed in subsequent sessions to 3 sets. The training work load was also progressed to 80% of 1RM.

Experimental (Combined Training)

• Exercise Prescription: a combination of the endurance training protocol and resistance training protocol described above.

Control

• Exercise Prescription: no arm training.

Outcomes

Endurance Upper Limb Exercise Capacity: supported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (8 wks)"

• Scale: arm ergometer time

• Unit of measure: seconds

Peak Upper Limb Exercise Capacity: Supported

• Outcome type: Continuous Outcome

• Measure names: "Baseline", "Post‐intervention (8 wks)"

• Scale: arm ergometer power output

• Unit of measure: watts

Health‐Related Quality of Life

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (8 wks)"

• Scale: total score on St George's Respiratory Questionnaire

• Range: 0 to 100

• Direction: lower is better

Respiratory muscle strength

• Outcome type: Continuous Outcome

• Measure names: "Baseline", "Post‐intervention (8 wks)"

• Scale: Pimax

• Unit of measure: cmH₂O

• Baseline data in paper and authors provided post‐intervention data.

Peak Upper Limb Exercise Capacity: Unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (8 wks)"

• Scale: time to exhaustion

• Unit of measure: seconds

• Direction: higher is better

Identification

Sponsorship source: this article was supported by the Australian Respiratory Council.

Country: Australia

Setting: outpatient

Authors name: Zoe J McKeough

Institution: University of Sydney

Email: [email protected]

Address: Discipline of Physiotherapy, Faculty of Health Sciences, The University of Sydney, 75 East St Lidcombe, NSW 2141, Australia

Notes

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: "Randomisation was via computerised random number generation."

Allocation concealment (selection bias)

Low risk

Quote: "...with the sequence concealed using opaque envelopes prepared by an investigator not involved in the study."

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no blinding of participants and personnel.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote: "An assessor, blinded to group allocation, performed the outcome measures at baseline and at the end of the intervention period."

Incomplete outcome data (attrition bias)
All outcomes

High risk

Quote: "A total of 52 subjects were recruited and 38 (73%) completed the study with 11 in the endurance group, 9 in the strength group, 9 in the combined group, and 9 in the control group". Therefore outcome data was not available in 27%.

Selective reporting (reporting bias)

High risk

Data not reported in the paper for respiratory muscle strength but authors have provided this data for the meta‐analysis. Data not reported for upper limb strength although stated in trials registry.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable COPD (n = 28)

Baseline characteristics

Endurance Training Group

• FEV₁(% pred): 46 (20)

• FEV₁(L): 1.08 (0.63)

Resistance Training Group

• FEV₁₁ (% pred): 46 (20)

• FEV₁(L): 1.08 (0.63)

Control

• FEV₁(% pred): 31 (8)

• FEV₁(L): 0.78 (0.23)

Interventions

Intervention characteristics

All groups did lower limb training such as walking over 8 weeks with weekly supervision.

Experimental (Endurance Training)

• Exercise Prescription: upper limb training, 8 weeks, once weekly including both isokinetic arm cycle training for 15 min and arm endurance training of 5 low‐resistance high‐repetition exercises. Each exercise was coordinated with breathing in progressively increasing levels of: number of arm motions per exhalation (1 to 6); the number of sets (1 or 2) of 10 repetitions of each exercise; and added hand weights (1 to 5 lb).

Experimental (Resistance Training)

• Exercise Prescription: Upper limb training, 8 weeks, 4 exercises with weights (1 to 5 lbs based on the maximum weight repeated 6 times). Each exercise was performed in 3 sets of 4 to 10 repetitions. Training was performed every other day for 1 week and then once daily. The weight was increased when completion of 3 sets of at least 6 repetitions of each exercise was achieved.

Control

• Exercise Prescription: no arm training

Outcomes

Endurance Upper Limb Exercise Capacity: unsupported

• Outcome type: continuous outcome

• Unit of measure: time (s)

• Notes: 4 lifts (shld flexion, shld abduction of each arm separately) ‒ the number of lifts in 1 min averaged across the 4 tests.

Peak Upper Limb Exercise Capacity: supported

• Outcome type: continuous outcome

• Scale: peak power output from arm ergometer

• Unit of measure: watts

• Direction: higher is better

Endurance Upper Limb Exercise Capacity: Supported

• Outcome type: continuous outcome

• Scale: arm ergometer

• Unit of measure: time (s)

• Notes: just looked at arm cycle (isotonic)

Activities of Daily Living Function

• Outcome type: continuous outcome

• Scale: time

• Unit of measure: time (s)

• Notes: time to complete 3 tasks. The 3 tasks were the following: (1) dishwashing (patients washed, rinsed, dried, and put on a shelf 2 place settings of dishes, plus a pot); (2) blackboard task (patients covered a dustless blackboard with continuous writing and then erased it; this activity simulates cleaning a window, wall, or mirror); and (3) grocery shelving (patients carried, unpacked, and shelved 2 bags of groceries weighing 8 lb each).

Identification

Sponsorship source: supported in part by the Easter Seal Research Foundation of the National Easter Seal Society and National Institutes of Health grant RR00827 from the Division of Research Resources for the Clinical Research Center.

Country: USA

Setting: outpatient

Authors name: Andrew L Ries

Institution: Department of Medicine, University of California, San Diego

Address: UCSD Medical Center H‐772, 225 Dickinson Street, San Diego, 92103

Notes

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

The method used to generate the allocation sequence was not provided.

Allocation concealment (selection bias)

Unclear risk

The method used to conceal the allocation sequence was not provided.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no indication of blinding of participants and personnel.

Blinding of outcome assessment (detection bias)
All outcomes

High risk

There was no indication that there was blinding of the outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes

High risk

38% of participants did not complete the study. Reasons were given and one may have related to the study.

Selective reporting (reporting bias)

Unclear risk

Data from baseline analysis not available.

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable, severe COPD (n = 28)

Baseline characteristics

Experimental

• Age (yrs): 66.07 (9.24)

• FEV₁(% pred): 37.29 (11.47)

• FEV₁(L): 1.08 (0.38)

• Gender (male/female): 11/3

Control

• Age (yrs): 63.07 (9.41)

• FEV₁(% pred): 34.93 (17.13)

• FEV₁(L): 0.93 (0.38)

• Gender (male/female): 12/2

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: unsupported upper limb endurance exercise, 3 times weekly for 8 weeks. Five different types of unsupported upper limb exercise using an intermittent protocol of 45 s exercise, 15 s rest for no more than 5 min. Also, lower limb cycle exercise at 75% peak work for 45 min.

Control

• Exercise Prescription: no arm training; lower limb cycle exercise at 75% peak work for 45 min, 3 times weekly for 8 weeks.

Outcomes

Endurance Upper Limb Exercise Capacity: unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (8 wks)"

• Scale: arms lifted in 6 min

• Unit of measure: total no. of lifts in 6 min

Health‐Related Quality of Life

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (8 wks)"

• Scale: total score from Chronic Respiratory Disease Questionnaire

• Unit of measure: score (max 140)

• Direction: higher is better

Respiratory Muscle Strength

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (8 wks)"

• Scale: Pimax

• Unit of measure: cmH₂O

Hospitalisation index

• Outcome type: continuous outcome

• Measure names: "Baseline", "12 months post training"

• Scale: 12 months pre and 12 months post

• Unit of measure: no. hospitalisation days × no. patient

Identification

Sponsorship source: none

Country: Argentina

Setting: Outpatient

Authors name: Martin L Sivori

Institution: Servicio de Neumonologia Policlinico Bancário.

Email: [email protected]

Address: Dr Martin L Sivori, Lafayette 124 1872 Avellaneda, Pcia Buenos Aires, Argentina

Notes

Study reported in Spanish and required translation to English.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Randomisation by simple table.

Allocation concealment (selection bias)

Unclear risk

The method used to conceal the allocation sequence was not provided.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no indication that participants or personnel were blind from the knowledge of which intervention the participant was receiving.

Blinding of outcome assessment (detection bias)
All outcomes

High risk

There was no indication that there was blinding of the outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes

High risk

35% did not complete the study. Reasons were given not directly related to study.

Selective reporting (reporting bias)

Unclear risk

Data not provided for the quality of life domains (only a total score was given).

Other bias

Low risk

Study appears to be free of other sources of risk.

Methods

Study design: randomized controlled trial

Participants

Clinically stable COPD (n = 17), age 45 to 75 yrs

Baseline characteristics

Experimental

• Age (yrs): 60 (9)

• FEV₁(% pred): 46 (11)

• Smoking history: 34 (11)

• Gender (male/female): 9/0

Control

• Age (yrs): 57 (8)

• FEV₁(% pred): 49 (11)

• Smoking history: 35 (11)

• Gender (male/female): 8/0

Interventions

Intervention characteristics

Experimental

• Exercise Prescription: unsupported endurance upper limb exercise, 5 times per week for 4 weeks of training. Four types of upper limb exercises were performed for 40 s with 20 s rest for 4 min, repeated 4 times. Also, lower limb training was performed (alternated on different days).

Control

• Exercise Prescription: lower limb training only of walking for 20 min, 5 times per week for 4 weeks of training.

Outcomes

Dyspnoea from questionnaire

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (4 wks)"

• Scale: dyspnoea score from Chronic Respiratory Disease Questionnaire

Peak Upper Limb Exercise Capacity: unsupported

• Outcome type: continuous outcome

• Measure names: "Baseline", "Post‐intervention (4 wks)"

• Unit of measure: time (s) from the incremental unsupported arm exercise test

Identification

Country: India

Setting: outpatient

Authors name: Vaishali Rao Subin

Institution: Department of Physiotherapy, Kasturba Medical College

Email: [email protected]

Address: Ms. Vaishali Rao, Department of Physiotherapy, Kasturba Medical College, Mangalore, India

Notes

A third group was involved and performed upper limb training only but was not included in this review as there was no control group with no training.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Block randomisation was used.

Allocation concealment (selection bias)

Unclear risk

Block randomisation used but unclear whether it is concealed.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

There was no indication that participants or personnel were blind from the knowledge of which intervention the participant was receiving.

Blinding of outcome assessment (detection bias)
All outcomes

High risk

There was no indication that there was blinding of the outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Appears to be missing data for 10% of participants only.

Selective reporting (reporting bias)

Low risk

Data reported for all outcomes stated.

Other bias

High risk

Block randomisation in an unblinded trial.