Delirium as a predictor of long-term cognitive impairment in survivors of critical illness

Abstract

Objective: To test the hypothesis that duration of delirium in the intensive care unit is an independent predictor of long-term cognitive impairment after critical illness requiring mechanical ventilation.

Design: Prospective cohort study.

Setting: Medical intensive care unit in a large community hospital in the United States.

Patients: Mechanically ventilated medical intensive care unit patients who were assessed daily for delirium while in the intensive care unit and who underwent comprehensive cognitive assessments 3 and 12 mos after discharge.

Measurements and main results: Of 126 eligible patients, 99 survived>or=3 months after critical illness; long-term cognitive outcomes were obtained for 77 (78%) patients. Median age was 61 yrs, 51% were admitted with sepsis/acute respiratory distress syndrome, and median duration of delirium was 2 days. At 3-mo and 12-mo follow-up, 79% and 71% of survivors had cognitive impairment, respectively (with 62% and 36% being severely impaired). After adjusting for age, education, preexisting cognitive function, severity of illness, severe sepsis, and exposure to sedative medications in the intensive care unit, increasing duration of delirium was an independent predictor of worse cognitive performance-determined by averaging age-adjusted and education-adjusted T-scores from nine tests measuring seven domains of cognition-at 3-mo (p=.02) and 12-mo follow-up (p=.03). Duration of mechanical ventilation, alternatively, was not associated with long-term cognitive impairment (p=.20 and .58).

Conclusions: In this study of mechanically ventilated medical intensive care unit patients, duration of delirium (which is potentially modifiable) was independently associated with long-term cognitive impairment, a common public health problem among intensive care unit survivors.

Figure 1. Enrollment and follow-up

*One patient who was not reached for testing at 3-month follow-up was fully tested at 12-month follow-up.

Figure 2. Relationship between duration of delirium and average cognitive performance measured at 3- and 12-month follow-up

At 3-month (Panel A) and 12-month follow-up (Panel B), duration of delirium independently predicted average performance on a battery of nine neuropsychological tests after adjusting for age, education, preexisting cognitive function, severity of illness, severe sepsis, ABC Trial treatment group, and total benzodiazepine, opiate, and propofol doses administered in the ICU (p = 0.02 and 0.03, respectively). A mean T-score (shown on the Y-axis) of 50 indicates average performance on nine neuropsychological tests, based on age- and education-adjusted normative data. These results show that, other factors being equal, a patient with 5 days of delirium will score on average nearly one-half of a standard deviation lower (i.e., 5 points lower) across domains of cognitive function at 3-month follow-up (and 7 points lower at 12-month follow-up) than a patient who was delirious for 1 day. The smooth graphs were created using restricted cubic splines. A rug plot indicates the distribution of delirium duration in the cohort; specifically, each patient is represented by a vertical bar sitting on the X-axis, showing the duration of that patient’s delirium (jittering is used to display all patients, though only integers were used to record delirium days).

Figure 2. Relationship between duration of delirium and average cognitive performance measured at 3- and 12-month follow-up

At 3-month (Panel A) and 12-month follow-up (Panel B), duration of delirium independently predicted average performance on a battery of nine neuropsychological tests after adjusting for age, education, preexisting cognitive function, severity of illness, severe sepsis, ABC Trial treatment group, and total benzodiazepine, opiate, and propofol doses administered in the ICU (p = 0.02 and 0.03, respectively). A mean T-score (shown on the Y-axis) of 50 indicates average performance on nine neuropsychological tests, based on age- and education-adjusted normative data. These results show that, other factors being equal, a patient with 5 days of delirium will score on average nearly one-half of a standard deviation lower (i.e., 5 points lower) across domains of cognitive function at 3-month follow-up (and 7 points lower at 12-month follow-up) than a patient who was delirious for 1 day. The smooth graphs were created using restricted cubic splines. A rug plot indicates the distribution of delirium duration in the cohort; specifically, each patient is represented by a vertical bar sitting on the X-axis, showing the duration of that patient’s delirium (jittering is used to display all patients, though only integers were used to record delirium days).