OBJECTIVES: To describe the effect of dementia on hospital outcomes of individuals requiring invasive mechanical ventilation (IMV).

DESIGN: Retrospective cohort study.

SETTING: 2011 Nationwide Inpatient Sample.

PARTICIPANTS: Hospitalized individuals with and without dementia undergoing IMV.

MEASUREMENTS: The adjusted predicted probability of undergoing IMV was examined in individuals with and without dementia. Then the dataset was limited to individuals who received IMV, and a multivariable logistic regression model was created in which dementia was the primary predictor and mortality was the outcome.

RESULTS: Of the 13,816,586 hospitalizations of older adults in the United States in 2011, 2,204,506 (16%) with a dementia diagnosis code were identified. Individuals with dementia had statistically significantly lower predicted probability of undergoing IMV (5.7%, 95% confidence interval (CI) = 5.6-5.8% than those without (6.5%, 95% CI = 6.4-6.6%). When the dataset was limited to individuals undergoing IMV, those with dementia were older (mean age 80 vs 76, P < .001) and had a higher combined Gagne comorbidity score (4.4 vs 4.1, P < .001) than those without. In a multivariable model, dementia was associated with greater likelihood of survival to hospital discharge (odds ratio (OR) = 0.79, P < .001). Individuals with dementia also had shorter mean length of stay (12.5 +/- 0.2 vs 13.1 +/- 0.2, P = .01) and lower cost per hospitalization for survivors ($37,213 vs $44,557, P < .001).

CONCLUSION: Older critically ill adults with dementia undergoing IMV had better in-hospital outcomes than those without dementia. Because a lower adjusted percentage of individuals with dementia underwent IMV, it is likely that patient selection drove outcome differences. These findings suggest that individuals, families, and clinicians are carefully considering prognosis, quality of life, and appropriate use of intensive care unit resources when deciding whether to use IMV in individuals with dementia.