Table 2 Characteristics of the included studies

Binder 2005 [38]

n = 91 from the USA Age, mean ± SD: 83.0 ± 4.0 54 % women

Community dwelling

Screening instruments and procedures [12]: Modified Physical Performance Test score between 18 and 32 (maximum 36); reporting the difficulty and/or assistance with up to two instrumental and/or one basic ADL; and a peak aerobic power between 10 and 18 mL kg^-1 min^-1

Multi-component PRT

9 months

3/week

60-90 min/session

Initial goal:

 1-2 sets

 6-8 repetitions

 65% 1-RM

Final goal:

 3 sets

 8-12 repetitions

 40-60% 1-RM

1-RM strength in six different exercises (knee extension, knee flexion, seated bench press, seated row, leg press, biceps curl), performed bilaterally in a weightlifting machine

Upper and lower extremities

Strength: skeletal muscle strength (maximal voluntary muscle strength for knee extension and flexion)

Body composition: Total body DEXA

Significant increase in knee extension with the intervention (d = 0.62)

Knee flexion strength showed no effect with the intervention

Intervention induced greater increases in total (d = 0.20) and regional (d = 0.19) FFM but no changes in fat mass

Cadore 2014 [13]

n = 24 from Europe Age, mean ± SD: 91.9 ± 4.1 70 % women

Institutionalized

Fried’s criteria [2]

Multi-component PRT

12 weeks

2/week

40 min/session

8-10 repetitions

40-60% 1-RM

Two exercises for the leg extensor muscles (bilateral leg extension

and bilateral knee extension muscles) and one exercise for upper limbs (seated bench press), performed on a resistance variable exercise machine

Upper and lower extremities

Falls: Incidence

Mobility: 5 meter walking tests at their habitual speed; TUG; chair rising ability (the most rises in 30 sec)

Balance: FICSIT-4

Functional disability: ADLs using BI

Strength: Isometric upper and lower limb muscle strength

Body composition: fat muscle infiltration

Exercise training significantly reduced the incidence of falls (d = 2.71)

Walking ability did not change with the intervention

Exercise training significantly improved the time spent on the TUG (d = 0.42)

Significant change in the chair-rising ability test in the intervention group (d = 0.89)

Exercise training improved balance (d = 0.72)

Exercise training improved functional ability (d = 1.17)

Significant increase in knee extension with the intervention (d = 1.74)

Upper body muscle strength did not significantly change with the intervention

Intervention induced a decrease in fat muscle infiltration (Quadriceps femoris, d = 0.20; and knee flexor, d = 0.10)

Faber 2006 [32]

n = 238 from Europe Age, mean ± SD: 84.9 ± 6.0 79 % women

Institutionalized

Frailty indicators adapted from Fried’s criteria [2]

Multi-component functional walking and in balance exercises

20 weeks

1/week for 4 weeks

2/week for 16 weeks

90 min/session

Exercise without machines focused on balance, mobility,

and transfer training. They included standing up from a chair, reaching and stepping forward and sideward, heel and toe stands, walking and turning, stepping on and over an obstacle, staircase walking, tandem foot standing, and single-limb standing

Lower extremities

Tailored to the needs of each participant

Falls: incidence

Mobility: POMA; walking tests: 6 meters as fast as possible; TUG; chair rising ability (the time needed to stand up and sit down 5 consecutive times as fast as possible)

Balance: POMA; FICSIT-4

Functional disability:

ADL and instrumental ADL using GARS

Exercise training significantly reduced the incidence of falls in the pre-frail elderly sample. In the frail subgroup, the risk of becoming a faller increased with the intervention

Positive effect of the intervention on mobility in the pre-frail subgroup. In the frail subgroup, mobility worsened after the intervention

Small, but significant, positive intervention effect in POMA score in the exercise group, compared with the control group

Exercise training showed no effect on functional ability

Fairhall 2014 [33]

n = 241 from Australia/Oceania Age, mean ± SD: 83.3 ± 5.9 68 % women

Community dwelling

Fried’s criteria [2]

Multi-component exercise intervention:

Home program of balance and lower limb training based on the WEBB program

12 months

3-5/week

20-30 min/session

Exercises without machines

Lower extremities

Tailored to the needs of each participant

Falls: incidence; risk of falls (Physiological Profile Assessment [PPA]; short physical performance battery [SPPB]); 4-m walking tests

Strength: Lower body strength (knee-extension strength as a component of the PPA)

Exercise training did not significantly reduce the incidence of falls

Exercise training found a better postural sway (d = 0.09)

Significant increase in leg muscle extension with the intervention (d = 0.03)

Significant improvements in mobility (SPPB score, d = 0.40; and gait speed, d = 0.20)

Giné-Garriga 2010 [34]

n = 51 from Europe Age, mean ± SD: 84.0 ± 2.9 61 % women

Community dwelling

Two tests of physical abilities [51, 52] and according to two questions from the Center for Epidemiological Studies depression scale [2]

Multi-component functional based circuit training

12 weeks

2/week

45 min/session

1-2 sets

6-8 repetitions

1 day of balance-based

activities and 1 day of lower-body strength-based exercises, combined with function-focused activities. Exercises without machines

Lower extremity exercises included activities such as rising from a chair, stair climbing, knee bends, floor transfer, lunges, leg squats, leg extension, leg flexion, calf

raises, and abdominal curls using ankle weights

Mobility: walking tests: 8 meters at their habitual speed and as fast as comfortably possible; MTUG (modified TUG test)

Functional disability: ADL using BI

Strength: lower body strength (knee-extension and flexion strength)

Walking ability improved with the intervention (Balance measures: semitandem d = 4.65, tandem d = 6.62, and single leg d = 7.78; Gait speed measures: normal d = 3.50 and fast d = 3.50)

Exercise training significantly improved the time spent on the MTUG (assessment questionnaire d = 8.24, and total time d = 4.61)

Exercise training improved functional ability (BI score d = 1.08)

Significant increase in leg muscle extension with the intervention d = 3.50)

Giné-Garriga 2013 [35]

n = 51 from Europe Age, mean ± SD: 84.0 ± 2.9 61 % women

Community dwelling

Two tests of physical abilities [51] and according to two questions from the Center for Epidemiological Studies depression scale [2]

Multi-component functionally based circuit training

12 weeks

2/week

45 min/session

1-2 sets

6-8 repetitions

1 day of balance-based

activities and 1 day of lower-body strength-based exercises, combined with function-focused activities. Exercises without machines

Lower extremity

exercises included activities, such as rising from a chair, stair climbing, knee bends, floor transfer, lunges, leg squat, leg extension, leg flexion, calf

raise, and abdominal curl using ankle weights

Falls: fear of falling (Activities-specific Balance Confidence [ABC] scale)

Exercise training improved the fear of falling (d = 1.10)

Kim 2015 [37]

N = 131 from Asia Age, mean ± SD: 80.7 ± 2.8 100 % women

Community dwelling

Fried’s criteria [2]

Physical comprehensive training

12 weeks

2/week

60 min/session

30 minutes of strengthening exercises plus 20 minutes of balance and gait training

Strength exercises performed in progressive sequence from the seated to standing positions,

and progressive resistance was applied through Thera-bands, with increasing repetition

with each exercise

Lower extremities

consisted of leg extensions, hip flexions, and more. Upper body exercises included double-arm pull downs, bicep curls, and others

Mobility: walking speed; TUG

Strength: Grip strength and isometric knee extension strength

Body composition: Total body DEXA

Frailty status

Walking speed did not change with the intervention

Exercise training improved the time spent on the TUG (d = 0.64)

No increase in knee extension with the intervention

Upper body muscle strength did not significantly change with the intervention

No effect on body composition of the intervention

Exercise training and exercise training plus nutrition supplementation significantly improved frailty status

Latham 2003 [36]

n = 243 from Australia/Oceania Age, mean ± SD: 79.1 ± 6.9 53 % women

Teaching hospitals

Winograd’s frailty scale [53]

Home-based resistance training

20 weeks

3/week

Initial goal:

3 sets

8 repetitions

30-40% 1RM for 2 weeks

Final goal:

3 sets

8 repetitions

60-80% 1RM

Accomplished goal:

3 sets

8 repetitions

51% 1RM ±13%

Adjustable ankle cuff weights

Lower extremities

Falls: incidence; fear of falling

Mobility: 4 meter walking tests; TUG

Balance: BBS

Functional disability: ADL (BI) and participation in higher non-ADL levels of activity (Adelaide Activity Profiles)

Strength: maximal isometric knee extensor strength

Exercise training did not significantly reduce the incidence of falls

Walking ability did not change with the intervention

Exercise training did not change TUG measurements

Balance was not affected by the intervention

No intervention effect on ADL with exercise training

No effect on leg muscle extension with the intervention

Lustosa 2011 [31]

n = 48 from Brazil Age, mean ± SD: 72.0 ± 4.0 100 % women

Community dwelling

Fried’s criteria [2]

Body weight resistance training

10 weeks

3/week

60 min/session

3 sets

8 repetitions

70% 1RM

Ankle weights with loads ranging from 0.5 to 3 kg

Lower extremities

Mobility: 10 meter walking tests at their habitual speed; TUG

Strength: Muscle strength of knee extensor

Walking ability improved with the intervention (d = 0.69)

Exercise training significantly improved the time spent on the TUG (d = 0.17)

Significant increase in leg muscle extension with the intervention (d = 0.05)