Early and Late Mortality Following Discharge From the ICU

A Multicenter Prospective Cohort Study

Regis G. Rosa, PhD; Maicon Falavigna, PhD; Caroline C. Robinson, PhD; Evelin C. Sanchez, RN; Renata Kochhann, PhD; Daniel Schneider, BSc; Daniel Sganzerla, BSc; Camila Dietrich, MSc; Mirceli G. Barbosa, MSc; Denise de Souza, MSc; Gabriela S. Rech, BSc; Rosa da R. dos Santos, RN; Alice P. da Silva, RN; Mariana M. Santos, MSc; Pedro Dal Lago, PhD; Tarek Sharshar, PhD; Fernando A. Bozza, PhD; Cassiano Teixeira, PhD


Crit Care Med. 2020;48(1):64-72. 

In This Article

Abstract and Introduction


Objectives: To identify the frequency, causes, and risk factors of early and late mortality among general adult patients discharged from ICUs.

Design: Multicenter, prospective cohort study.

Setting: ICUs of 10 tertiary hospitals in Brazil.

Patients: One-thousand five-hundred fifty-four adult ICU survivors with an ICU stay greater than 72 hours for medical and emergency surgical admissions or greater than 120 hours for elective surgical admissions.

Interventions: None.

Measurements and Main Results: The main outcomes were early (30 d) and late (31 to 365 d) mortality. Causes of death were extracted from death certificates and medical records. Twelve-month cumulative mortality was 28.2% (439 deaths). The frequency of early mortality was 7.9% (123 deaths), and the frequency of late mortality was 22.3% (316 deaths). Infections were the leading cause of death in both early (47.2%) and late (36.4%) periods. Multivariable analysis identified age greater than or equal to 65 years (hazard ratio, 1.65; p = 0.01), pre-ICU high comorbidity (hazard ratio, 1.59; p = 0.02), pre-ICU physical dependence (hazard ratio, 2.29; p < 0.001), risk of death at ICU admission (hazard ratio per 1% increase, 1.008; p = 0.03), ICU-acquired infections (hazard ratio, 2.25; p < 0.001), and ICU readmission (hazard ratio, 3.76; p < 0.001) as risk factors for early mortality. Age greater than or equal to 65 years (hazard ratio, 1.30; p = 0.03), pre-ICU high comorbidity (hazard ratio, 2.28; p < 0.001), pre-ICU physical dependence (hazard ratio, 2.00; p < 0.001), risk of death at ICU admission (hazard ratio per 1% increase, 1.010; p < 0.001), and ICU readmission (hazard ratios, 4.10, 4.17, and 1.82 for death between 31 and 60 days, 61 and 90 days, and greater than 90 days after ICU discharge, respectively; p < 0.001 for all comparisons) were associated with late mortality.

Conclusions: Infections are the main cause of death after ICU discharge. Older age, pre-ICU comorbidities, pre-ICU physical dependence, severity of illness at ICU admission, and ICU readmission are associated with increased risk of early and late mortality, while ICU-acquired infections are associated with increased risk of early mortality.


The lasting burden of critical illness after ICU discharge is a reason for concern[1,2] with increasing evidence showing reduced long-term survival in ICU survivors as compared with the general population.[1–5] Nevertheless, despite the need for research on risk factors of mortality after critical illness,[6] few prospective studies have evaluated this aspect, and little information is available on the frequency and causes of post-ICU mortality.[3,7] Additionally, most published studies have failed to address issues such as the impact of pre-ICU health on long-term survival, or whether different factors may predict early or late mortality after ICU discharge.[8] Notably, this evidence gap may constitute a barrier to better understand the weight of critical illness on long-term survival after ICU discharge and to implement both therapeutic and palliative interventions that will improve survival and/or quality of life after critical illness. The present study aimed to assess the frequency and causes of early and late mortality after discharge from the ICU and to investigate which sociodemographic, pre-ICU state of health, and critical illness variables were associated with early and late mortality among general adult patients who were discharged from ICUs.