Association of medication-induced deep sedation and emotional distress during mechanical ventilation with loss of independent living: an observational cohort study

Abstract

Background: Deep sedation can be used during invasive mechanical ventilation without proven indication to treat the signs and symptoms of emotional distress or insomnia. However, medication-induced deep sedation is associated with delayed recovery and increased mortality. We tested the hypothesis that medication-induced deep sedation, but not emotional distress, is associated with loss of independent living.

Methods: In this retrospective cohort study, we included adult patients (age ≥18 years) who lived independently before hospital admission and were mechanically ventilated for at least 24 h in any of 20 ICUs at an academic health system in the Bronx, New York (NY, USA), and adjacent counties. The primary exposure was the proportion of time spent in medication-induced deep sedation (defined as a Richmond Agitation Sedation Score of -3 to -5) within the first week of ICU admission. The secondary exposure was the proportion of time with indicators of emotional distress within the first week of ICU admission. The exposures were categorised as none, a low proportion, or a high proportion using the median (42·9% for medication-induced deep sedation and 10·7% for emotional distress) as the cutoff between low and high. The primary outcome was loss of independent living (defined as in-hospital death or postoperative discharge to a long-term skilled nursing facility). Modified Poisson regression with an a priori-defined confounder control model were used to assess the association between exposures and outcomes. Mediation analysis was done to evaluate whether patient mobilisation level during mechanical ventilation contributed to the association between deep sedation and loss of independent living.

Findings: Among 10 204 patients receiving invasive mechanical ventilation between Jan 30, 2016 and July 11, 2023, 6369 (62·4%) had a loss of independent living. The proportion of patients who received deep sedation was a mean 2·84-fold (SD 1·51) higher than the proportion of patients who had an order for deep sedation. 7289 (71·4%) patients had at least one episode of medication-induced deep sedation within the first week of mechanical ventilation in the ICU. A high proportion of medication-induced deep sedation was associated with an increased risk of loss of independent living (adjusted risk ratio [RR_adj] 1·18 [95% CI 1·13-1·23], p<0·0001) compared with no medication-induced deep sedation. There were 30 022 documented episodes of emotional distress. Having a high proportion of emotional distress was associated with a decreased risk of loss of independent living (RR_adj 0·88 [0·84-0·92], p<0·0001) compared with no emotional distress. Clinicians provided deeper sedation in response to emotional distress and during the night (p<0·0001 for both comparisons). Symptom control with antipsychotics or non-opioid analgesics was associated with a decreased risk of loss of independent living compared with no treatment (RR_adj 0·95 [0·91-0·99], p=0·013). Absence of patient mobility during mechanical ventilation was associated with loss of independent living (RR_adj 1·32 [1·28-1·36], p<0·0001) and mediated 32·5% (p<0·0001) of the effects of deep sedation on loss of independent living.

Interpretation: Among mechanically ventilated adults in an ICU, the proportion of time spent in medication-induced deep sedation was independently associated with loss of independent living. Lighter sedation, allowing for recognition of emotional distress and for patient mobility, was associated with a lower risk of loss of independent living.

Funding: Department of Anesthesiology, Montefiore Medical Center, Albert Einstein College of Medicine (New York, NY, USA).