Table 1.

Summary of included studies.

Authors/Yr/Country	Methodology study design sample size	Sample characteristics	Dependent Variable(s)/Measure	Independent Variable(s)	Major findings/Outcomes	Limitations	Sackett’s/PEDro score
Martin et al, 2005 United States	Retrospective Medical record review N = 49	Average age of sample was 58.5 ± 7 years Patients were ventilated for ≥ 14 consecutive days, and at least two failed weaning attempts prior to start of study Patients were intubated for 18.1 ± 7.7 days (range 9–29 days) before Tracheostomy Average admission APACHE II score was 20	Pulmonary mechanics: - RSBI - NIFF Peripheral muscle strength: - 5-Point muscle strength score Function status: - FIM Weaning status: - Tolerance of 48 h of unassisted breathing	Whole body rehabilitation program: - Conducted by physical therapists - Progressed from one session per day to two sessions per day once able to consistently tolerate > 45 min in one session	Time to wean from mechanical ventilation, measured from the initial day of admission, was 16.0 ± 9 days 15 (30%) of patients weaned in ≤7 days Patients who weaned ≤7 days had higher upper limb motor strength scores than patients who took ≥7 days to wean, difference was not statistically significant All patients had significant improvement in FIM scores (admit score 1, discharge score 3) All patients were bed-bound at the time of admission; ambulation went from 0 to 52 ± 18 feet for the group (p = .005). A total of 40 patients (81%) were able to ambulate at the time of discharge Improvement in respiratory muscle strength evidenced by significant increase in maximal NIF (admission 24, discharge 35) A significant correlation was found between upper limb motor strength at the time of admission and time to wean (R = 0.72, R2 = 0.54, p < .001) On admission, the RSBI was <105 in 40% of patients. On the day of successful weaning, only 63% of patients had an RSBI <105 Three variables were found to be statistically significant in terms of predicting weaning time: upper motor strength (−7.1, SE 1.19, t −5.9, p = .001), exposure to neuromuscular blocking (4.6, SE 2.14, t 2.2, p = .03), and systemic steroids (4.8, SE 1.76, t 2.7, p = .0092)	Retrospective analysis of data collected prospectively without a control group Sample is from a single facility Reliability and validity of 5-point muscle strength score not discussed Strength was not objectively quantified Did not perform routine electromyograms to rule out the presence of critical illness polyneuropathy	4
Chiang et al, 2006 Taiwan	Single-blinded, non-randomized prospective, repeated measures design. N = 32 (n = 17 treatment group	Average age of control group was 79 (72.5 −82.8), average age of treatment group was 75 (63.0–80.3) All patients received mechanical ventilation via a tracheostomy tube All patient had received mechanical ventilation for > 14 days prior to start of study Strength of the 3 tested muscle groups were the same at baseline in both groups.	Pulmonary mechanics: - PImax - PEmax - TLC Limb muscle strength: - Upper and lower extremity muscle strength via hand-held dynamometer Weaning status: - Time in hours off ventilator Functional status: - BI - FIM - 2-min walk test (only if off ventilator)	Physical training program - Bedside upper extremity strengthening exercises including ROM exercises and use of weights - Bedside lower extremity strengthening exercises including ROM exercises and straight leg raises - Initially exercises were performed in supine position, progressed to seated position as tolerated - Progressed to side-to-side turning in bed, transfers from bed to chair, standing, and eventually, ambulation	Limb strength increased in the treatment group (p = .001) at the 3rd and 6th weeks compared with baseline. After 3 and 6 weeks of physical training, strength of all 3 muscle groups was greater in the treatment group than in the control group (p = .05) Limb strength in control group deteriorated at both 3rd and 6th weeks of the study period compared with baseline (p < .05) PI max was greater in the treatment over control group at 6 weeks (p < .05); PEmax was greater in the treatment group over the control group at both 3 and 6 weeks (p < .05). After 6 weeks of physical training, BI and FIM scores were greater in the treatment group than in the control group (BI scores control group 0, treatment group 35 and total FIM score control group 26, treatment group 49; p < .05) All subjects at the time of enrollment into the study were unable to walk. After 6 weeks of physical training, 53% of the subjects in the treatment group regained their ambulation ability After 6 weeks of intervention, the average distance walked during the 2-min walk test was 42.9 ± 12.7 m (n = 9) in treatment group All subjects in control group remained bedridden and none could ambulate at the end of the 6-week study period Ventilator-free hours improved in the treatment group (p < .01) to an average of 8.9 h and increased to 4.8 h in the control group, however this change was not statistically significant (p = .10). There were no between group differences. At the end of the 6-weeks study period, 8 patients (47%) in the treatment group and 3 patients (20%) in the control group could tolerate unassisted breathing for at least 12 h	Wide variety of patients with different diagnoses and etiologies Lack of randomization, patients were assigned in an alternating fashion to the treatment group of the control group, possible source of systematic bias Sample is from a single facility Small sample size	2a/4
Yang et al, 2010	Prospective controlled clinical trial N = 126 (n = 55 treatment group and n = 71 control group)	Average age of sample was 68.9 ± 16.1 years Patients were ventilated for 20.4 ± 6.6 days before entering study Only 71(56%) of the sample had a tracheostomy at the start of the study Average admission APACHE II score was 16.4 ±6.5 Baseline difference in APACHE II severity of illness scores (χ2 = 2.23, p = .028), indicating that the severity of illness scores were lower in the therapy group than the non-therapy group (15.0 vs 17.5). Baseline differences also noted in BI score (t = 3.00, p = .004), indicating that mobility was higher in the therapy group than the non-therapy group (0.6 vs 0.2)	Pulmonary mechanics: - RSBI Functional status: - BI	Physical therapy - One session per day - Included abdominal breathing exercises, respiratory muscle weight training, passive and active joint exercises, upper and lower limb exercises, progressive mobility training. Patients averaged 14.8 ± 7.9 days of therapy	RSBI before therapy was 75.7 ± 37.9, and after therapy was 80.0 ± 48.5, this was not significant (t = 0.540, p = .57), indicating that therapy did not have a significant impact on RSBI No significant difference in weaning rates between therapy group (32 ± 58.2) and control group (29 ± 40.9) at completion of study (t = 3.73, p = .054) There was a significant increase in BI in the therapy group increased from 0.8 ± 1.4, to 1.9 ± 2.5 at completion of therapy program (t = 0.004, p < .005). Therapy was positively associated with improved mobility Passive range of motion was most frequently provided therapy intervention (n = 37, 23.7%), only 18 patients (11.5%) progressed to progressive mobility	No discussion of how patients were assigned to intervention or control group Sample is from a single facility Heterogenous sample population Most of exercises in the therapy group involved only passive range of motion of upper extremities, this could possibly explain lack of significant result	2b
Chen, S. et al, 2011 Taiwan	Randomized control trial, prospective, repeated measures N= 34 (n = 18 treatment group and n = 16 control group)	Average age of control group was 79 (72.5 −82.8), average age of treatment group was 75 (63.0–80.3) All patient had received mechanical ventilation for > 14 days prior to start of study, requiring mechanical ventilation for at least 6 h/day and had been attempting a spontaneous breathing trial each day All patients were tracheostomized prior to start of study	Functional status: - FIM - BI 1-year survival rate Weaning status not defined	Physical training program - Followed by independent exercise continuation - UE and LE strengthening - Active chair transfer and maintenance of sitting position for at least 20 min/day - Standing-ambulating with assist devices as needed	In the rehab group FIM scores improved: Total FIM scores baseline 34 (30.3 −38.3); 6 weeks 49 (45–66.3); after independent therapy 78 (62 −126) Motor domain baseline 4.5 (13 −18.3); after independent therapy 47 (34–91) Cognitive domain baseline 19.5 (16.5–20.3); after independent therapy 33 (28–35) FIM scores remained unchanged for the control group The eating, comprehension, expression, and social interaction subscales reached the 7-point complete independence level at 6 months in the rehabilitation group, but not in the control group The 1-year survival rate for the rehabilitation group was 70%, which was significantly higher than the 25% for the control group (p = .0151) Five patients in the therapy group were discharged from the hospital during the one-year study, while only one was discharged in the control group No significant differences between groups for the percentage of survivors freed from mechanical ventilation after training or at the three follow-up times	Sample size was relatively small Sample is from a single facility At the 3rd follow-up, there were only four patients in the Ctrl group and 11 patients in the Rehab group. The uneven sample size may have had an impact on the statistical analysis of the final data collection time point Entry criteria required that patients had undergone mechanical ventilation for at least 14 days. However, no restriction was placed on the duration of mechanical ventilation, which varied considerably in the study Lacked compliance records from caregivers for independent exercise training in the treatment group	2a/5
Clini et al, 2011 Italy	Prospective cohort study N = 77	Average age of sample was 75 ± 7 years Male predominated sample (n = 46 60%) Average admission APACHE II score was 11.5 ± 4.4 Mean short term ICU hospitalization prior to admission to rehabilitation center was 24 days with SD ± 3 days Mean onset of acute respiratory failure was 24 days (range 18–29 days) Mean tracheostomy day was day 16 (range 12–21 days) No patients at the time of admission could tolerate a 2-h spontaneous breathing trial Mean length of hospitalization in the rehabilitation center was 51 days (range 12–115 days)	Functional status: - BADL - FIM - 6-Point Kendall muscle testing scale Pulmonary mechanics: - PImax - PEmax Weaning status: Successfully weaned when could tolerate spontaneous respiration for at least 7 consecutive days	Physical training program - Started 48 h after admission - Began with active movement of the limbs - Progressed to active muscular intervention consisting of trunk control, maintenance of body posture, and both upper and lower limb activities to facilitate transfer from bed to chair and standing up - As soon as possible patient begin to use a wheeled walker - Supported or unsupported limb training was incremented daily	The mean ± SD Δ BADL was 2.5 ± 2.0 points for the entire cohort Statistically significant difference in survival and weaning rate between the groups: The group with the lowest Δ BADL had the worst clinical outcome 64 of the 77 patients improved in at least one BADL activity. Patients with less improvement or no change were less likely to be weaned and to live All the patients in the ΔBADL >2 category survived, and over 80% of them (51% of total) successfully weaned There were 67 respiratory-ICU survivors, and at respiratory-ICU discharge 23 patients went home, 15 were transferred to a medical ward, and 29 were transferred to a rehabilitation unit	A single-center observational study Small sample size No control group	4
Chen, Y-H. et al, 2012 Taiwan	Randomized, single-blinded, prospective, repeated measures design. N = 27 (n = 12 treatment group and n = 15 control group)	Average age of treatment group was 64.9 ± 21.3 and average age of control group was 66.5 ± 18.7 Significant difference in pulmonary mechanics between groups prior to intervention larger Vt in control group (230 ± 95.6 vs 143.6 ± 79.4)	Pulmonary mechanics: - Vt - MV - PImax - RSBI - f, breaths/min Functional status: - BI - FIM Hospitalization outcomes: - LOS - Ventilator weaning rate: (Weaning from ventilation considered successful if subject free from ventilation continuously for > 5 days) - # of Total ventilator days Mortality rate:	Exercise training program - Cardio/pulm endurance training with cycle ergometer - Muscle strength training with resisted arm activities with weights and weighted sand bag on abdomen for respiratory muscle training - Stretching exercises of cervical, upper limb, and upper chest	After exercise intervention, treatment group had significant improvement in Vt(192.5 ± 75.0), while the control group was essentially unchanged (257.6 ± 125.1) Mid-study RSBI was decreased, when compared to pre-study measurements in the treatment group (110.6 ± 31.5 vs 162.2 ± 70.1), while essentially unchanged in the control group (123.1 ± 56.2vs 136.2 ± 48.8) In the treatment group, both FIM (28.1 ± 14.2 vs 44.6 ± 10.0, P = .005) and BI(4.3± 9.5 vs 19.3 ± 18.6, P = .004) scores improved after the exercise intervention. These were statistically significant for FIM, but not BI. Subjects in the training group had a higher weaning rate (75% vs 53.3%) and a lower mortality rate (0% vs 20%) during their hospitalization Patient in the training group had fewer days of mechanical ventilation than those in the control group, although not statistically significant. (32.7 ± 23.4d vs 54.6 ± 46.2d. p = .15) The mean Los for subject in the training group (36 d)was shorter than that for the control group(57 d)	Small sample size Sample is from a single facility Baseline differences in some pulmonary mechanics measures between groups Did not account for the severity of illness, length of previous ICU stay, or etiologies for requiring PMV	lb/6
Patmanetal, 2012 Australia	Retrospective Medical record review N = 190	Mean age was 52 ± 18 years Male predominated sample (n = 126 66%) Average APACHE II score was 20 ± 8 Median quartile ICU LOS 14 (10) days Median quartile acute care LOS 42 (38) days	Functional Milestones: - Sit - Stand - Ambulate		On discharge from acute care, 89 (47%; 95% CI, 40%−54%) were ambulating independently, of whom 54 (61%) did not require a gait aid 183 patients (96%) sat out of bed during their acute care stay. Of these. 124 (65%) achieved this milestone during the ICU admission Median time between admission to the ICU and when the patient first sat out of bed was 13 (8) days 163 patients (86%) stood during acute care stay. Of these. 42 (22%) achieved this during the ICU admission Median time between admission to the ICU and when the patient first stood was 19 (19) days 155 patients (82%) walked during their acute care stay. Of these. 15 (8%) achieved this during the ICU admission The median time between admission to the ICU and when the patient first walked was 23 (21) days Compared with those who stood within 30 days of admission to ICU. a delay in standing of between 30 and 60 days increased the odds of not being able to ambulate independently at the time of discharge 5-fold (95% CI. 2–11) A delay in standing of greater than 60 days increased the odds of not being able to ambulate independently at the time of acute care discharge 28-fold (95% CI, 6–122)	Limited by retrospective data collection Sample is from a single facility No intervention Dependent variables measured at the nominal level resulting in potential loss of statistical power	4
Hill etal, 2013 Australia	Retrospective Descriptive N = 181	Mean age was 52 ± 19 years 22 patients (12%) did not survive the 12-month period after discharge from the acute care facility	Mortality: - Death in the 12 month period after discharge from hospital Morbidity: - Hospital admission rate after discharge from hospital Functional Milestones: - Sit - Stand - Ambulate		Those who died were older (50 [32] vs 69 [16] years, p = .001). had higher APACHE II scores (19 [10] vs 24 [9]. p = .008), and a longer hospital length of stay (40 [38] vs 64 [61] days, p = .005) In the 12-month period prior to the study. 66 (36%) patients had at least 1 admission and 11 (6%) had been admitted to an ICU and received MV. 2 (18%) had been mechanically ventilated for 7 days or longer. Median number of admissions was 2, median length of stay was 11 days In the 12-month period after discharge from the hospital, 148 (82%) patients had at least 1 admission and 17 (9%) had been admitted to an ICU and received MV, 7 (41%) had been mechanically ventilated for 7 days or longer. Median number of admissions was 2 and the median length of stay was 40 days With patients grouped according to functional status (independent ambulation vs dependent ambulation) there was no difference in the proportion of people who survived vs deceased over the 12-month period (χ2 = 0.023, P =.880) No difference when patients first sat out of bed (13 [8] days vs 12 [6] days, P = .282) or stood (19 [19] days vs 18 [23] days, P = .663) between those who survived and those who had deceased A longer time lapse between admission to the ICU and when the patient first stood was associated with greater health care use in the 12 months after the index admission When compared with those who could ambulate independently at the time of discharge, those who could not ambulate had 81% more admissions in the subsequent 12-month period	Retrospective data collection methods Sample is from a single facility May lack statistical power for some of the analyses	4

APACHE II + Acute Physiology and Chronic Health Evaluation; P_Imax = Maximum inspiratory pressure; P_Emax = Maximum expiratory pressure; Vt = Tidal Volume; RSBI = Rapid Shallow Breathing Index; MV = Minute Volume; f, breaths/min = respiratory rate; TLC = Total Lung Capacity; NIFF = negative inspiratory force; FIM = Function Independence Measurement; BI = Barthel Index; BADL = Basic Activities of Daily Living; LOS = length of stay.