Table 1 Detailed summary of eligible studies in the review

Setting

Cardiac rehabilitation facility

Geriatric nursing home

University indoor exercise facility

Two senior welfare centres

Design

RCT, parallel

RCT, parallel, four-armed

RCT, parallel

RCT, parallel, three-armed

Aims of the study

To evaluate the value of resistance training on measures of physical performance in older women with coronary heart disease

To evaluate the short-term effects of three different low-intensity exercise interventions on physical performance, muscle CSA and ADL.

To evaluate whether a multidimensional exercise training program can significantly reduce frailty in community-dwelling older men and women

To investigate if there are differences in the effects of an exercise intervention due to the applied intensity of supervision

Sample size (analyzed), n

IG: 19, CG: 14

IG: 22, CG: 23

IG: 66, CG: 49

IGc: 32, CG: 17

Female gender, n

Overall: 100%

IG: 68.1%, CG: 65.2%

IG: 52%, CG: 53%

IGc: 30/32 (92%), CG: 15/17 (88%)

Mean age (SD), years, range

IG: 73.2 (6.0), CG: 72.2 (5.7)

IG: 85.5 (4.7), 83.6 (5.6), 75–96

IG: 83 (4), CG: 83 (4)

IG1: 80,0 (6,7), IG2: 80,8 (5,3), CG: 77,2 (6,5) (completers only)

Participant health status (functional limitation criteria)

Patients had CHD diagnosed for > 6 months, MOS SF > 36, physical function domain score < 85

Residents in geriatric nursing home

Defined frailty criteria including: Objective test, reported ADL and IADL dependency

Difficulty in rising from a chair and unilateral knee extensor strength below 25 kgf.

Residential status

Community-dwelling

Geriatric nursing home

Community-dwelling

Apartments for elderly connected to welfare centres

Description, intensity, duration and total number of sessions

8 RT exercises focusing on leg, arm, and shoulder. Progressive program updated monthly

Low intensity RT program targeted major knee extensor muscles. 40% 1RM, 16 weeks, 48 session

Progressive whole-body RT program in weightlifting machines. 65–100% 1RM, 12 weeks, 36 sessions

9 thigh muscles exercises. Resistance provided by body weight and elastic bands. 4–8 RM (elastic band exercises) 10 weeks, 30 sessions

Control condition

Control patients met 3 times per week performing stretching, calisthenics, deep-breathing progressive-relaxation exercises, and light yoga

No intervention. Refrain from participation in exercise programs

Sham intervention: 9 flexibility exercises

No intervention. Maintain habitually active

Self-reported measure of ADL-disability/function

MOS SF-36, physical function domain

Barthel Index

Functional Status Questionnaire

The Groningen Activity Restriction Scale (ADL/IADL)

Drop-out from intervention, n

IG: 5 (21%), CG: 4 (22%)

IG: 4 (18%), CG: 4 (17%)

IG: 20 (30%), CG: 8 (16%)

IGc: 18 (36%), CG: 5 (23%)

Compliance, % (criteria)

Patients were required to attend at least 54/72 sessions (75%). 2 patients failed, and were recorded as dropouts

78% (mean attendance at sessions)

100% (attendance at sessions. Less than 100% attendance led to exclusion)

IG1: 79%, range: 57–100%, IG2: 72% range 20–93% (mean attendance at sessions)

Direction of the effect on self-reported disability/function

No effect

Positive effect

Positive effect

No effect

Adverse events

No adverse events

No adverse events

One: rotator cuff injury, and one: RT exacerbating shoulder problem

Not reported

Notes

RT is the second of three 3-months intervention phases. We consider 3- and 6-month time points as baseline and endpoint test respectively

Two eligible RT-intervention groups. Degree of supervision varied between groups

Data notes

Published and unpublished data

Data from the two intervention groups were collapsed in all analysis

Included in primary meta-analysis

Yes

Yes

Yes

Yes

First author, year, country

Buchner et al., 1997, USA [35]

Cadore et al., 2014, Spain [36]

Chandler et al., 1998, USA [16]

Chin A Paw, et al., 2006, the Netherlands [37]

Setting

Enrolees in a health maintenance organization

–

The home of the elderly

Long-term care facilities

Design

RCT, parallel, four-armed

RCT, parallel

RCT, parallel

RCT, parallel, four-armed

Aims of the study

To investigate the effect of strength and endurance training on gait, balance, physical health status, fall risk, and health service’s use in older adults

To investigate the effects of multicomponent exc. Intervention on muscle power output, muscle mass, tissue attenuation, fall risk and functional outcomes

To determine whether strength gain is associated with improvement in physical performance and disability

To evaluate the effectiveness of three different training protocols on functional performance and self-rated disabilities of older adults living in long-term care facilities.

Sample size (analyzed), n

IG: 22, CG: 29

IG: 11, CG: 13

IG: 44, CG: 43

IG: 40, CG: 31

Female gender, n

IG: 52%, CG:50%

17/24 (70%) (completers only)

Overall: 50%

IG:29/40 (73%), CG26/31 (84%)

Mean age (SD), years, range

IG: 74, CG: 75 No SD

IG: 93,4 (3,2), CG: 90,1 (1,1)

IG: 77,5 (7,1), CG 77,7 (7,8)

IG: 80,9 (5,7), CG: 81,2 (4,4)

Participant health status (functional limitation criteria)

Unable to do an 8-step tandem gait without errors, below the reference 50th percentile in KE strength

Frieds frailty criteria, institutionalized

Inability to descent stairs step by step without holding the railing

Living in long-term care facilities. The population is referred to by the authors as frail

Residential status

Community-dwelling

Institutionalized

Community-dwelling

Nursing home/residential care

Description, intensity, duration and total number of sessions

RT of the upper and lower body using Cybex Eagle weight machines. Including training at the ankle joint using adjustable weights

3 RT-exercises. 2 for knee extensors + chest press in machines (20 min). Gait and balance exercises (10 min). 8–10 RM, 12 weeks, 24 sessions

Home-based low-moderate intensity RT-programme using elastic band. Exercises target lower extremity muscles with slow velocities of movement. 10 RM, 10 weeks, 30 sessions

Long term care facility-based. 5 RT-exercises using machines, free weights and ankle/wrist weights. 60–80% 1RM, 24 weeks, 48 sessions

Control condition

Instructed to maintain usual activity

No intervention. Routine care and activities

No intervention. Controls were offered RT after the end of the trial

Attention control. Educational program led by professional creative therapist. 45–60 min twice weekly.

Self-reported measure of ADL-disability/function

Sickness Impact Profile, body care and movement subscale

Barthel Index

MOS SF-36, physical function domain

Disability in 17 ADLs

Drop-out from intervention, n

IG: 5 (20%), CG: 1 (3%)

IG: 5 (31%), CG: 3 (19%)

Overall: 13 (13%)

IG: 21 (37%), CG: 23 (45%)

Compliance, % (criteria)

IG: 95% (mean attendance at sessions)

90% (attendance at sessions. Attendance was defined as ≥90% of prescribed exercises completed)

Not reported

76% (mean attendance at sessions)

Direction of the effect on self-reported disability/function

No effect

Positive effect

No effect

No effect

Adverse events

No adverse events

Not reported

Not reported

No adverse events reported. N = 8 dropped out because the program was too intensive

Notes

One intervention group was eligible for inclusion in the analysis

One intervention group was eligible for inclusion in the analysis

Data notes

Ceiling effects of the Sickness Impact Profile, body care and movement-subscale was reported

Published and unpublished data

Post data not available

Included in primary meta-analysis

Yes

Yes

No

Yes

First author, year, country

Clemson et al, 2012, Australia [38]

Danilovich et al., 2016, USA [39]

Fahlman et al., 2007, USA [40]

Hewitt et al., 2018, Australia [41]

Setting

Residents in metropolitan Sydney, Australia

Home-based, Illinois

University facilities, Urban area

Long -term residential aged care facilities

Design

RCT, parallel, three-armed

RCT, parallel

RCT, parallel, three-armed

RCT, Cluster

Aims of the study

To determine if a lifestyle integrated approach to balance and strength training is effective in reduces the rate of falls in high risk people

To test the effect of an RT-program on the physical performance and self-rated health of older adults receiving home and community-based services

To determine whether RT or a combination of RT and aerobic training resulted in the most improvement in measures of functional ability in functionally limited elders

To test the effect of published best practice exercise in long-term aged care, and determine if combined balance and progressive RT is effective in reducing the rate of falls

Sample size (analyzed), n

IG: 79, CG: 80

IG: 24, CG: 18

IG: 39, CG: 33

IG: 93, CG: 82

Female gender, n

IG: 57/105 (54,3%), CG: 58/105 (55,2%)

Overall: 83%

Not reported

IG: 71 (62.8%), CG: 73 (68.2%)

Mean age (SD), years, range

IG: 84,03 (4,38), CG: 83,47 (3,81)

CG: 74,1, CG: 75,6

IG: 74,6 (SE;1,0), CG: 76,5 (SE 1,4)

IG: 86, 65–100, CG: 86, 65–99

Participant health status (functional limitation criteria)

Two or more falls or one injurious fall in the past 12 months

Homebound, receiving long-term ADL-assistance and home management

Score < 24 on the SF-36 PFD (reference score = 30)

High- or low-care requirements (daily assistance by nurse / some assistance but not complex care-needs)

Residential status

Community-dwelling

Community-dwelling, homebound

Community-dwelling

Long-term residential care

Description, intensity, duration and total number of sessions

Structured home-based programme. 7 exercises for balance + 6 exercises for lower limb strength 3 times a week, 1 year

Health care assistant and DVD-delivered, RT program with elastic bands

RT program consisting of 13 exercises using resistive bands. Low-moderate intensity, 16 weeks, 48 sessions

Moderate intensity progressive RT program consisting of 5 exercises combined with high-progressive level balance program. 25 weeks, 50 sessions

Control condition

Sham intervention: 12 gentle flexibility exercises

No intervention. Usual care

No intervention. They were instructed to maintain their current level of activity

No intervention. Usual care

Self-reported measure of ADL-disability/function

The National Health and Nutrition Examination Surveys independence measure for Activities of Daily Living (NHANES ADL)

Patient-Reported Outcomes Measurement Information System (PROMIS), physical summary score/ADL

MOS SF-36, physical function domain

MOS SF-36, physical function domain

Drop-out from intervention, n

IG: 22 (21%), CG: 16 (15%)

IG: 3 (13%)

Not reported

IG: 16 (14%), CG: 15 (14%)

Compliance, % (criteria)

IC: 35% (SD: 29), CG: 47% (SD: 34) (adherence to programmes)

Not reported

Not reported

54% (SD: 14.3) attended at least 30 sessions (60% adherence). Median attendance: 35 sessions

Direction of the effect on self-reported disability/function

No effect

No effect

No effect

No effect

Adverse events

One Surgery for inguinal hernia due to groin strain

No adverse events

Not reported

No major events. N = 3 reported short-term musculoskeletal pain, n = 1 non-injurious fall

Notes

RT-program based on Jette 1996

One intervention group was eligible for inclusion in the analysis

48.9% of participants had a diagnosis of mild to moderate cognitive impairment

Data notes

Pre and post results are presented for different subsamples

Extraordinary small sizes of variability distorted the meta-analysis of SMDs

Pre and post results are presented for different subsamples

Included in primary meta-analysis

No

Yes

No

No

First author, year, country

Latham et al., 2003, New Zealand [42]

McMurdo and Johnstone, 1995, USA [43]

Mihalko and McAuley 1996, USA [44]

Sahin et al., 2018, Turkey [45]

Setting

Five urban hospitals in New Zealand/Australia

The home of elderly receding in sheltered housing

Nursing home or senior citizen facility

Not reported

Design

RCT, parallel, four-armed

RCT, parallel, three armed

RCT, parallel

RCT, parallel, three armed

Aims of the study

To determine the effectiveness of vitamin D and home-based quadriceps resistance exercise on reducing falls and improving physical health of frail older people after hospital discharge

To develop a low technology approach to home exercise provision for elderly people with restricted mobility

To examine the effects of upper body high-intensity strength training on muscular strength levels, ADLs, and subjective well-being in elderly males and females.

To evaluate changes in the functioning of frail older adults after undergoing RT 3 days a week for 8 weeks

Sample size (analyzed), n

IG: 112, CG: 110

IG: 21, CG: 28

IG: 29, CG: 29

IGc: 32, CG: 16

Female gender, n

IG: 55%, CG: 51%

IG: 19/21 (90%), CG: 25/28 (89%)

Overall: 83%

Not reported

Mean age (SD), years, range

IG: 80 (range: 79–81), CG: 78 (range: 77–80)

IG: 81,4 (3,4), CG:81,9 (4,7)

Overall: 82.67 (7.72)

IG1: 84.18 (6.85), IG2: 84.50 (4.81), CG: 85.37 (4.70)

Participant health status (functional limitation criteria)

Frail according to criteria (Winograd). Admitted to geriatric rehabilitation unit.

Limited mobility, dependence in ADL

19 used a wheelchair, 13 used walking assistance

Frailty according to Fried criteria

Residential status

Not specified

Sheltered housing

Nursing home

Nursing home

Description, intensity, duration and total number of sessions

Home-based quadriceps resistance program using adjustable ankle cuff weights. 3 sets of 8 reps of knee extensions in a seated position.

Low technology, low cost home exercise program using elastic bands. Emphasis on safety and respect for pain. 6 months with training on daily basis. No data on intensity

Upper body RT program with one exercise for the following muscle-groups: pectorals, latissimus dorsi, deltoids, biceps, and triceps. Performed with dumbbells

11 RT exercises for upper and lower body. 1 set of 6–10 reps at a slow speed (6–8 s/rep). IG1: 70% 1RM IG2: 40% 1RM. 8 weeks, 24 sessions

Control condition

Received frequency-matched telephone calls and home visits from physical therapist who inquired about patient’s recovery, gave general advice.

Frequency and duration matched health education program. Informal discussions on exercise, diet, sleep, meditation, stress foot care and safety

Upper body, no-stress exercise program: Breathing techniques; movement of the neck, shoulder, arms, hands, and torso; and mild stretching activities

Instructed to continue usual daily routines

Self-reported measure of ADL-disability/function

MOS SF-36, physical function domain

Barthel Index

Barthel Index, tailored

Barthel Index

Drop-out from intervention, n

IG: 8 (7%), CG: 13 (10%)

Overall: 20%

Not reported

IGc: 0, CG: 0

Compliance, % (criteria)

82% (mean attendance at sessions)

Not reported

Not reported

Not reported

Direction of the effect on self-reported disability/function

No effect

No effect

Positive effect

Positive effect

Adverse events

The exercise group had an increased risk of musculoskeletal injury and higher scores of fatigue.

No adverse events

Not reported

Not reported

Notes

One intervention group was eligible for inclusion in the analysis

Two eligible RT-intervention groups. Work load intensity varied between groups.

Data notes

Missing baseline data

ANCOVA test applied to account for baseline imbalances

Data from the two intervention groups were collapsed in all analysis but the sub-analysis for training intensity

Included in primary meta-analysis

No

Yes

Yes

Yes

First author, year, country

Seyennes et al., 2004, France [17]

Timonen et al., 2006, Finland [46]

Venturelli et al., 2010, Italy [47]

Westhoff et al., 2000, the Netherlands [18]

Setting

Public nursing homes

Primary care health centre

Geriatric institute

Home-based/community centre-based

Design

RCT, parallel, three-armed

RCT, parallel

RCT, parallel

RCT, parallel

Aims of the study

To measure dose-response effect of a free weight-based RT program on KE muscle function, functional limitation and self-reported disability.

To determine the effects of a group-based exercise program on ADL and IADL activities relevant to daily life after discharge from hospital

To evaluate the feasibility of upper-body circuit-RT program, and to verify if arm training improves physical outcomes, ADL-function and cognitive outcomes.

To investigate if a 10-week low-intensity strength training program can improve strength of the knee extensors and functional ability in frail elderly.

Sample size (analyzed), n

IGc: 14, CG: 8

IG: 26, CG: 30

IG: 12, CG: 11

IG: 10, CG: 11

Female gender, n

Not reported

IG: 100%, CG: 100%

IG: 100%, CG: 100%

Not reported

Mean age (SD), years, range

IG1: 83.3 (2.8), IG2: 80.7 (2.3), CG: 80.3 (2.0)

IG: 83.5 (4.1) CG: 82.6 (3.7)

IG: 83,3 (6,7), CG: 84,1 (5,8)

IG: 75.9 (6.8), CG: 77.5 (8.1)

Participant health status (functional limitation criteria)

Institutionalized. Characterised by authors as frail. Objective measure not reported

Hospitalized due to an acute illness and mobility-impaired

Dependent in one or more ADL (BI), serious mobility limitation, MMSE > 15 < 25

Difficulty in rising from a chair

Residential status

Public nursing home

Community-dwelling

Geriatric institute

Residents of assistant living facilities

Description, intensity, duration and total number of sessions

Classical progressive RT of the KE muscles using ankle cuffs. IG1: 80% 1RM, IG2: 40% 1RM, 10 weeks, 30 sessions

Group based progressive RT with weight training equipment plus functional exercises. 8–10 RM, 10 weeks, 20 sessions

Group based upper body RT program using dumbbells, looped, elastic bands, sticks and sponge balls. Progression by raising number of repetitions and or load

Individually tailored RT program for the KE using bodyweight and elastic band to provide resistance. 9 exercises. 4 RM (elastic band exercises), 10 weeks, 30 sessions

Control condition

Placebo: similar program with empty ankle cuffs

Instructions for a home exercise training program, including functional exercises. No further encouragement to exercise.

Kept their habits unaltered throughout the study. Were provided physiotherapy as usual

No intervention. Asked to continue with their normal activities

Self-reported measure of ADL-disability/function

Health Assessment Questionnaire

Tailored ADL/IADL function scale

Barthel Index

The Groningen Activity Restriction Scale (ADL/IADL), lower extremity-specific domain

Drop-out from intervention, n

Overall: 5 (19%)

IG: 8 (23%), CG: 4 (12%)

IG: 3 (20%), CG: 4 (27%)

IG: 4 (29%), CG: 1 (8%)

Compliance, % (criteria)

99% (criteria not stated)

90%, range 55–100% (mean attendance at sessions)

75% (SD: 16%) (mean attendance to sessions)

87% (mean attendance to sessions)

Direction of the effect on self-reported disability/function

No effect

No effect

Positive effect

Positive effect

Adverse events

No adverse events

Not reported

No adverse events

No adverse events

Notes

Two eligible RT-intervention groups. Work load intensity varied between groups. 5 drop outs in total. Number of dropouts on group-level is not reported.

ADL/IADL measured by proxy (health care personnel)

Very frail subjects - many are wheelchair users

Data notes

Published and unpublished data. Data from two intervention groups were collapsed in all analysis but the sub-analysis for training intensity

Data not suitable for meta-analysis

Included in primary meta-analysis

Yes

No

Yes

Yes