Objectives: To assess the impact of rehabilitation in ICU on clinical outcomes.
Data sources: Secondary data analysis of randomized controlled trials published between 1998 and October 2019 was performed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.
Study selection: We have selected trials investigating neuromuscular electrical stimulation or cycling exercises or protocolized physical rehabilitation as compared to standard of care in critically ill adults.
Data extraction: Mortality, length of stay in ICU and at hospital, days on mechanical ventilator, and adverse events.
Data synthesis: We found 43 randomized controlled trials (nine on cycling, 14 on neuromuscular electrical stimulation alone and 20 on protocolized physical rehabilitation) into which 3,548 patients were randomized and none of whom experienced an intervention-related serious adverse event. The exercise interventions had no influence on mortality (odds ratio 0.94 [0.79-1.12], n = 38 randomized controlled trials) but reduced duration of mechanical ventilation (mean difference, -1.7 d [-2.5 to -0.8 d], n = 32, length of stay in ICU (-1.2 d [-2.5 to 0.0 d], n = 32) but not at hospital (-1.6 [-4.3 to 1.2 d], n = 23). The effects on the length of mechanical ventilation and ICU stay were only significant for the protocolized physical rehabilitation subgroup and enhanced in patients with longer ICU stay and lower Acute Physiology and Chronic Health Evaluation II scores. There was no benefit of early start of the intervention. It is likely that the dose of rehabilitation delivered was much lower than dictated by the protocol in many randomized controlled trials and negative results may reflect the failure to implement the intervention.
Conclusions: Rehabilitation interventions in critically ill patients do not influence mortality and are safe. Protocolized physical rehabilitation significantly shortens time spent on mechanical ventilation and in ICU, but this does not consistently translate into long-term functional benefit. Stable patients with lower Acute Physiology and Chronic Health Evaluation II at admission (<20) and prone to protracted ICU stay may benefit most from rehabilitation interventions.