The Republic of Korea is growing into an aging society. In Korea, the proportion of the elderly population (65 and older) increased from 11.5% in 2012 to 16.5% in 2021. Additionally, the aging index (the number of elderly people [≥65 years] per 100 individuals younger than 14 years in a specific population) also rapid increased from 76.1% in 2012 to 138.8% in 2021 [1]. Considering the current trend of population change, it is high probable that Korea becomes a super aging society by 2029 [2]. Older age is associated with a higher prevalence of chronic diseases and a higher incidence of hospital admission and intensive care unit (ICU) treatment [3].

ICU and in-hospital mortalities in elderly patients are higher than those in non-elderly patients [4-7]. In a study of very elderly (≥80 years) patients admitted to the medical ICU in Korea, weaning failure and the withdrawal, or withholding of intensive care in the ICU were significantly related to death in patients ≥ 80 years old [8]. In addition, male sex and moderate malnutrition were factors associated with mortality in a study involving patients aged ≥65 years admitted to the ICU without limitations on other therapeutic efforts [9]. However, the clinical characteristics and prognostic factors of very elderly patients admitted to the ICU and changes in the characteristics of the patient population over time in the aging society are not well known.

In this study, we investigated the clinical characteristics of very elderly patients (over 85 years of age) in our ICU over the last decade and compared the first half (2007–2012) and the second half (2013–2017) of the study period. We aimed to determine if age ≥85 years should be regarded as an exclusion criterion in the ICU triage.

On comparing the first half (2007–2012) and the second half (2013–2017) of the study period, the proportion of the very elderly group increased, and the number of patients with underlying disease are increased, but the in-hospital mortality decreased. There were more males in the non-survived group, and there were more medical reasons for admission to the ICU. In addition, mechanical ventilation, vasopressors, and RRT were more required in the non-survived group than in the survived group. High creatinine level, ventilator weaning failure and use of vasopressors were associated with in-hospital mortality.

In Korea and other countries around the world, the proportion of the elderly population is increasing rapidly, and the number of elderly people in need of intensive care is also increasing [1,11-14]. In Korea, a rapidly aging society, there are few studies conducted in very elderly patients admitted to the ICU [14-16]. According to data from 38 ICU training hospitals, 3.7% of patients admitted to the ICU are aged ≥85 years [15]. Additionally, Lee et. al. [8] reported that 21.7% of patients admitted to the ICU were aged ≥80 years. Furthermore, Kim et al. [16] showed that 0.5%–1.9% of patients admitted to the ICU were over 90 years old. In our study, 0.9%–2.0% of the patients admitted to the ICU were ≥85 years old from 2007 to 2017. It is predicted that the number of very elderly patients in the ICU will increase, which is related to the increase in life expectancy in each country [11]. Each study showed various percentages of very elderly patients admitted to the ICU. This may have been influenced by the fact that the age of the study subjects, defined as very elderly patients, varied as 80, 85, and 90 years old, and the ratio of very elderly patients according to the life expectancy of each country varied by period.

The number of very elderly people admitted to the ICU is steadily increasing. As the number of very elderly patients admitted to the ICU increases, there is a change in the patients’ underlying disease and prognosis. On comparing the first half (2007–2012) and the second half (2013–2017) of the study period, the number of ICU admissions for medical reasons increased. During the second half of the study period, the number of underlying diseases such as hypertension, chronic heart disease, solid tumor, and cerebrovascular accident increased. However, in-hospital mortality and hospital LOS decreased in the second half of the study period. Kim et al. [16] reported slight increases in hypertension and cardiogenic shock, among the underlying diseases, as reasons for ICU admission in 2010–2014 compared with that in 2005–2009. Additionally, the mortality rate decreased from 2005 to 2014. In Kim et al.'s study [16] of patients admitted to the medical ICU from 2003 to 2013, the number of patients over 80 years of age increased from 7.3% in 2002 to 23.1% in 2013. However, there was no difference in the annual survival probability according to year (P=0.895) [17]. In Korea, there is a greater interest in intensivists and nursing staff working in the ICU than before, and the presence of intensivists [18] and the number of nursing staff [19] are known to be associated with lower mortality. In addition, there is a tendency for mortality to decrease in acute respiratory distress syndrome [20] and sepsis [21], which are related to mortality in the elderly. These points may have influenced the pattern of decreasing mortality in this study.

In this study, in the very elderly patients admitted to the ICU, hematologic malignancy, diabetes mellitus (DM) were more common in the non-survived group. In addition, in the non-survived group, white blood cell, total bilirubin, creatinine, and C-reactive protein levels were higher, and albumin level was lower than those in the survived group. Additionally, invasive mechanical ventilation and RRT were performed more often in the non-survived group. Roch et al. [22] showed that in-hospital non-survivors, the Simplified Acute Physiology Score (SAPS) II and the SOFA score were higher, and mechanically ventilation and RRT were often performed. Lee et al. [8] showed that the SOFA and Acute Physiology And Chronic Health Evaluation (APACHE) II scores were higher, and renal disease and rates of severe sepsis and septic shock were more common in the non-survivor group. Higher SAPS II [7], SOFA scores [23], and APACHE scores [24] mean that the non-survivor patients were severely ill; thus, treatments such as invasive mechanical ventilation and RRT are more likely to be required.

In the non-survived group, there were more respiratory, infection-related, and hematologic reasons for medical ICU admission. Kim et al. [16] reported that the most common reason for ICU care was acute respiratory failure. Many functional, demographic, and immunologic changes associated with aging are responsible for the increasing incidence and severity of infectious diseases in the elderly [25]. Martin-Loeches et al. [26] showed very elderly patients with sepsis had more pneumonia than elderly patients.

Older age is well-known factor associated with in-hospital mortality in patients admitted to the ICU. However, there are few studies on the characteristics of very elderly patients in the ICU (≥85 years). Lee et al. [8] showed that weaning failure and the withdrawal or withholding of intensive care in the ICU were significantly related to mortality in patients age ≥80 years. Martin-Loeches et al. [26] showed that patients aged ≥80 with sepsis had higher hospital mortality compared with patients aged between 65 and 79 years. In the group of very elderly patients, the predictors of hospital mortality were age, APACHE II score, and prompt adherence to the resuscitation bundle. Giannasi et al. [9] showed that after adjusting for APACHE II score and age, the following independent variables were associated with mortality: male sex, moderate malnutrition, severe malnutrition, and activities of daily living scores <6. Mukhopadhyay et al. [27] showed that in the elderly group, neurological disease, malignancy, mortality prediction model II score, mechanical ventilation, non-invasive ventilation, vasopressor use, and prior RRT were associated with in-hospital mortality. Considering long-term mortality as a prognosis, Atramont et al. [28] showed that age and reason for hospitalization were strongly associated with long-term mortality (9-, 13-, and 20-fold increases in the risk of death 3 years after ICU discharge in patients aged 80–84, 85–89, and ≥90 years, respectively). In this study, like other studies, high creatinine level, ventilator weaning failure and use of vasopressors were associated with in-hospital mortality. The patient's underlying disease was not associated with in-hospital mortality in this study. There was no significant difference in comorbidities except for hematologic malignancy, DM between the survived group and the non-survived group, so it seems to have a influenced the absence of statistically significant risk factor. In some studies, comorbidities did not show as a significant factor even in the elderly. In a systematic review of older patients (≥75 years) by Vallet et al. [29], ICU mortality and comorbidities were associated with one out of six studies, and in-hospital mortality and comorbidities were associated with 10 out of 12 studies. In a review by de Rooij et al. [30], comorbidities are common in very elderly patients, but the effect on outcome is not clearly shown. Therefore, in very elderly patients, ICU management is recommended considering the severity of the patient, the reason for ICU admission, and the functional status in addition to the comorbidities. Therefore, in very elderly patients, it is necessary to help in making a treatment decision considering various factors such as frailty, severity, and laboratory data in addition to comorbidities.

There are some limitations to our study. First, it was a single-center retrospective study and does not represent the entire Korean ICU population. Second, there is a possibility of selection bias as we did not collect data on the ICU’s denial of admission and/or patient/doctor preferences. Therefore, we were unable to estimate the gradual benefits of ICU support for those very old patients rejected admission to the ICU. Third, there is no information about the patients’ functional statuses prior to hospitalization. For elderly patients, the evaluation of the pre-hospital functional status is important, but data could not be collected because the patients’ previous functional evaluations were not clear in the electronic medical records. Fourth, we did an analysis that included patients with documented DNR. In actual situations, there are cases where DNR was received in verbal form or in patients with imminent death who received DNR to avoid cardiopulmonary compression. We were not able to collect additional data of DNR. Therefore, we were unable to conduct a separate analysis on the effect of DNR on patients.

In conclusion, the proportion of very elderly people (≥85 years) in the ICU has increased over the last decade (from 2007 to 2017). In the very elderly admitted in the ICU, the number of patients with underlying diseases slightly increased, but the prognosis was improved. In addition to age, laboratory data and disease severity contribute to patient outcomes. Therefore, we need to consider various factors in addition to age to determine ICU management for very elderly patients.