BACKGROUND: Patients who have received mechanical ventilation can have prolonged cognitive impairment for which there is no known treatment. We aimed to establish whether early mobilisation could reduce the rates of cognitive impairment and other aspects of disability 1 year after critical illness. METHODS: In this single-centre, parallel, randomised controlled trial, patients admitted to the adult medical-surgical intensive-care unit (ICU), at the University of Chicago (IL, USA), were recruited. Inclusion criteria were adult patients (aged ≥18 years) who were functionally independent and mechanically ventilated at baseline and within the first 96 h of mechanical ventilation, and expected to continue for at least 24 h. Patients were randomly assigned (1:1) via computer-generated permuted balanced block randomisation to early physical and occupational therapy (early mobilisation) or usual care. An investigator designated each assignment in consecutively numbered, sealed, opaque envelopes; they had no further involvement in the trial. Only the assessors were masked to group assignment. The primary outcome was cognitive impairment 1 year after hospital discharge, measured with a Montreal Cognitive Assessment. Patients were assessed for cognitive impairment, neuromuscular weakness, institution-free days, functional independence, and quality of life at hospital discharge and 1 year. Analysis was by intention to treat. This trial was registered with ClinicalTrials.gov, number NCT01777035, and is now completed. FINDINGS: Between Aug 11, 2011, and Oct 24, 2019, 1222 patients were screened, 200 were enrolled (usual care n=100, intervention n=100), and one patient withdrew from the study in each group; thus 99 patients in each group were included in the intention-to-treat analysis (113 [57%] men and 85 [43%] women). 65 (88%) of 74 in the usual care group and 62 (89%) of 70 in the intervention group underwent testing for cognitive impairment at 1 year. The rate of cognitive impairment at 1 year with early mobilisation was 24% (24 of 99 patients) compared with 43% (43 of 99) with usual care (absolute difference -19·2%, 95% CI -32·1 to -6·3%; p=0·0043). Cognitive impairment was lower at hospital discharge in the intervention group (53 [54%] 99 patients vs 68 [69%] 99 patients; -15·2%, -28·6 to -1·7; p=0·029). At 1 year, the intervention group had fewer ICU-acquired weaknesses (none [0%] of 99 patients vs 14 [14%] of 99 patients; -14·1%; -21·0 to -7·3; p=0·0001) and higher physical component scores on quality-of-life testing than did the usual care group (median 52·4 [IQR 45·3-56·8] vs median 41·1 [31·8-49·4]; p<0·0001). There was no difference in the rates of functional independence (64 [65%] of 99 patients vs 61 [62%] of 99 patients; 3%, -10·4 to 16·5%; p=0·66) or mental component scores (median 55·9 [50·2-58·9] vs median 55·2 [49·5-59·7]; p=0·98) between the intervention and usual care groups at 1 year. Seven adverse events (haemodynamic changes [n=3], arterial catheter removal [n=1], rectal tube dislodgement [n=1], and respiratory distress [n=2]) were reported in six (6%) of 99 patients in the intervention group and in none of the patients in the usual care group (p=0·029). INTERPRETATION: Early mobilisation might be the first known intervention to improve long-term cognitive impairment in ICU survivors after mechanical ventilation. These findings clearly emphasise the importance of avoiding delays in initiating mobilisation. However, the increased adverse events in the intervention group warrants further investigation to replicate these findings. FUNDING: None.