Long-term recovery of sensorimotor functions and prediction of participation in survivors of critical illness: a prospective cohort study.

Background: Survivors of critical illness frequently face physical, cognitive and psychological impairments after intensive care. Sensorimotor impairments potentially have a negative impact on participation. However, comprehensive understanding of sensorimotor recovery and participation in survivors of critical illness is limited. Therefore, the aims of this study were to quantify long-term sensorimotor recovery in survivors of critical illness, to examine participation in daily life 1.5 years after illness onset, and to assess the predictive capacity of sensorimotor assessments for future participation.

Methods: Survivors of critical illness who were mechanically ventilated ≥ 5 days on the ICU and who were admitted with weakness to neurorehabilitation were included in this single-center prospective cohort study. Time effects on sensation, muscle strength, balance, walking and dexterity were described at admission to and at discharge from rehabilitation, and 1.5 years after critical illness onset. Participation was assessed with the Reintegration to Normal Living Index. A multiple linear regression with sensorimotor outcomes at rehabilitation admission was conducted to find predictive associations with participation. The model was compared to an extended regression model containing demographic variables and factors known to be associated with participation or quality of life.

Results: All sensorimotor outcomes among participants (n = 250, median age 63 (54-73) years) improved over time. However, in most patients some deficits remained after rehabilitation and on long-term follow-up. Good participation (≥ 75%) was achieved by 60.2% of survivors 1.5 years after critical illness onset. Concerning participation, the Mini Balance Evaluation Systems Test (Mini-BESTest) together with the Box-and-Block-Test, the Five-Times-Sit-to-Stand-Test, and the Medical Research Council score at rehabilitation admission formed a predictive model (R² = 0.157, p < 0.001). The extended regression analysis resulted in a model (R² = 0.357, p < 0.001) with the variables depression, duration of mechanical ventilation, cognitive function, Mini-BESTest, comorbidities, sex and cerebral lesion.

Conclusions: We observed significant improvements in sensorimotor function, albeit with lingering deficits in sensation, strength, balance, dexterity and participation. Sensorimotor functions at rehabilitation start have limited explanatory power in predicting participation 1.5 years after disease onset. Trial registration German Clinical Trial Register, DRKS00021753. Date of registration: September 03, 2020.