The global incidence of cancer is increasing [1]. Previous studies indicated that patients with cancer account for approximately 10–15% of all intensive care unit (ICU) admissions, and there is an increasing trend in the number of metastatic cancer cases for ICU admissions [2–5]. The mortality rate of patients with advanced cancer admitted to the ICU was high [6,7]. Therefore, patients with advanced cancer were often considered a disqualifier for ICU admission in the past [8]. However, recent studies have shown a decrease in the mortality rate in patients with cancer admitted to the ICU, and this has been attributed to advances in overall critical care and sepsis management [9–11]. Despite these improvements, the decision to treat cancer patients with limited life expectancy in the ICU, where beds and resources are limited, remains a complex and challenging issue [6,12]. In Korea, there is a paucity of data on the outcomes of critically ill patients with solid malignancies. This study aimed to better understand the risk factors for mortality and the short- and long-term outcomes in patients with solid malignancies who had unplanned ICU admissions.

A total of 955 patients with solid malignancies were included. They were followed through July 19, 2023. The baseline characteristics of the patients are described in Table 1. The overall ICU mortality was 23.5%, which significantly differed based on the baseline cancer types (Fig. 1). For all types of solid malignancies, lung cancer was the most frequent type (34.2%) of cancer in our study cohort. Most patients had metastatic cancer (79.1%), and respiratory failure was the most common cause of ICU admission (46.2%), followed by shock (29.3%). The proportion of patients who received chemotherapy within 6 months was 83.4%. The SOFA score at ICU admission was 11. At the time of ICU admission, in 61.6% of the patients, the PaO₂/FiO₂ ratio was lower than 300, and vasopressors were used in 72.0% of patients. When comparing baseline characteristics between the ICU survivor and non-survivor groups, respiratory failure (60.3% vs. 41.9%) and cardiac arrest (10.7% vs. 3.7%) were more common in the non-survivor groups, while shock (33.4% vs. 16.1%) and post-operative/procedure (6.7% vs. 1.3%) were more common in the survivor groups (p < 0.001). The median time from cancer diagnosis to ICU admission was shorter (8.6 mo vs. 12.2 mo, p = 0.015), and the ECOG performance status was poorer (ECOG performance status 2–4; 33.5% vs. 25.3%, p = 0.020) in the non-survivor group. In the ICU non-survivor group, the PaO₂/FiO₂ ratio was lower (240 ± 193 vs. 289 ± 171, p = 0.001), higher doses of vasopressor were administered (0.21 ± 0.23 μg/kg/min vs. 0.14 ± 0.18 μg/kg/min, p <0.001), and the median SOFA score was higher (13 [interquartile range, 11–16] vs. 10 [interquartile range, 7–13], p < 0.001) than the ICU survivor group. The rates of neutropenia were not significantly different between the two groups (12.1% vs. 9.0%, p = 0.227). In lung and breast cancer patients, respiratory failure was the most common reason for ICU admission (67.9% and 64.8%, respectively), whereas in patients with pancreas biliary cancer, shock was the most common reason for ICU admission (Supplementary Table 1, p < 0.001). The missing variables and their number are shown in Supplementary Table 2.

During the ICU stay, 67.7% of the study patients received mechanical ventilation, and the median duration of mechanical ventilation was 6.0 days (3–13 days) (Table 2). RRT was applied in 23.2% of patients, and tracheostomy was performed in 17.4% of patients. The median length of ICU and post-ICU hospital stays was 6.0 days (3–12 days) and 9.0 days (0–19 days), respectively. The hospital mortality rate was 42.1%, and the 1-year mortality rate was 70.3%. The ICU non-survivor group was more likely to receive mechanical ventilation (95.5% vs. 59.2%, p < 0.001) and RRT (42.9% vs. 17.2%, p < 0.001), compared to the ICU survivor group.

In multivariable logistic regression analysis (Table 3), lung cancer (odds ratio [OR] 1.58, p = 0.030), SOFA score at ICU admission (OR 1.11, p < 0.001), mechanical ventilation during the ICU stay (OR 6.74, p < 0.001), and RRT during the ICU stay (OR 2.49, p < 0.001) were independent risk factors for ICU mortality. Pancreas-biliary cancer was associated with lower ICU mortality (OR 0.54, p = 0.040). In Supplementary Table 3, compared to those with respiratory failure, patients with shock (OR 0.35, p < 0.001) and post-operative/procedure (OR 0.10, p = 0.001) had a lower risk of ICU mortality.

When 1-year survival after ICU admission was compared in patients requiring mechanical ventilation or RRT, the survival rates were significantly lower in patients who received mechanical ventilation compared to those who did not receive mechanical ventilation (26.3% vs. 37.4%, p < 0.0001), and in patients who received RRT compared to those who did not (17.5% vs. 33.4%, p < 0.0001) (Fig. 2A, B). When 1-year survival rates among patients with lung, breast, urogenital, and pancreas biliary cancers were compared, the survival rates were significantly different at 24.1%, 28.8%, 31.5%, and 34.7%, respectively (p = 0.0049) (Fig. 2C).

Of the 731 patients who survived in the ICU, 178 (24.4%) died in the hospital. Of these, 141 (79.2%) were not readmitted to the ICU and died in the hospital because the patient or family decided to withhold life-sustaining treatment after experiencing deterioration on the general ward (Supplementary Table 4). Among patients who survived in the ICU, the one-year mortality rate after ICU discharge was 61.2%. Patients with metastatic cancer had the highest one-year mortality (63.8%) compared to those with other stages of cancer (p = 0.029, Supplementary Fig. 1A). Among patients with metastatic cancer who survived in the ICU (n = 568), 246 (45.9%) patients received chemotherapy after ICU discharge. Those who received post-ICU chemotherapy had significantly lower one-year mortality than those who did not receive chemotherapy (49.1% vs. 75.1%, p < 0.0001), and their median survival was 383 days and 42 days, respectively (Supplementary Fig. 1B). Among ICU survivors, those who received RRT during ICU stay had higher one-year mortality than those who did not receive RRT (69.1% vs. 60.0%, p = 0.012) (Supplementary Fig. 2B).

In this single-center, retrospective cohort study of critically ill patients with solid malignancy, the ICU mortality was not high, but the 1-year mortality of the cohort was above 70%. The types of underlying cancer and treatments received during the ICU stay were significantly associated with ICU mortality. This is the first study to report the long-term outcomes of critically ill patients with solid malignancies in Korea.

Our study showed ICU and hospital mortality rates of 23.5% and 42.1%, respectively, which are similar to the statistics previously reported from a Korean dataset and a single-cohort study from France, which included patients with solid malignancies. The Korean multi-center data from 2018 showed an ICU mortality rate of 24.6% and a hospital mortality rate of 38.6% [4]. Similarly, a study from France reported an ICU mortality rate of 22.6% [5]. Conversely, our ICU mortality rate was significantly higher than a study on national data from the Netherlands, which reported an ICU mortality rate of 13.6% for solid malignancies [8]. In the Dutch study, the rates of organ failure such as acute renal failure, the need for mechanical ventilation, and the use of vasoactive drugs were reported to be 10.2%, 36.5%, and 36.7%, respectively. In comparison, our study showed that the rates of RRT initiation, mechanical ventilation use, and use of vasopressors were 23.2%, 67.7%, and 72.0%, respectively. In addition, the median SOFA score of the patients at ICU admission was 11, which was higher than the reported study [4,5]. The difference in patient severity observed in our study and the Dutch study findings could indicate a higher baseline severity in our patient cohort. In addition, our study showed that patients who required mechanical ventilation or RRT during their ICU stay had significantly higher ORs for mortality of 6.74 and 2.49, respectively. Our study showed a significant association between a higher SOFA score on ICU admission and an increased risk of ICU death. This highlighted the fact that ICU mortality rates were higher in patients with multiple organ failure on admission and in those who required organ support therapy such as mechanical ventilation or RRT.

In our cohort, we observed that the type of cancer had a significant impact on ICU mortality. Patients with lung cancer had a higher OR of 1.58 for ICU mortality, whereas patients with pancreas biliary cancer had a lower OR of 0.54. Qian et al. [13], reported that the main reason for ICU admission in lung cancer patients was pneumonia or respiratory failure. Similarly, most lung cancer patients in our study were admitted to the ICU for respiratory failure. The discovery of various molecular oncogenes and advances in targeted and immunotherapies have led to significantly increased life expectancy in patients with lung cancer because these treatments often allow patients (even those in advanced stages) to remain stable for a longer period of time [14]. However, this increase in life expectancy has also resulted in more cases of respiratory failure due to drug-induced pneumonitis or immune-related adverse events following targeted or immunotherapy [5]. Although ICU outcomes in patients who develop adverse events as a result of targeted therapy or immunotherapy may be better than those of conventional cytotoxic chemotherapy [15,16], patients manifesting respiratory failure had worse outcomes than patients with other causes of ICU admission in our cohort. In contrast, patients with pancreas biliary cancer were more likely to be admitted to the ICU for septic shock due to biliary infection or hemorrhagic shock due to bleeding [17]. Our study showed that approximately 2/3 of the patients with pancreas biliary cancer were admitted to the ICU due to shock, and patients with shock had a lower OR for ICU mortality than those with respiratory failure. Despite the poor oncological prognosis for pancreas-biliary cancer, interventions such as biliary drainage and embolization can often be useful for the rapid reversal of sepsis, potentially leading to a higher ICU survival rate. Therefore, our findings suggest that the ICU mortality in patients with solid malignancies is closely related to their cancer type and the cause of ICU admission.

Based on the high long-term mortality, the findings could be helpful to both intensivists and oncologists, providing important data to assist with the decision-making process for ICU admission of patients with solid malignancies. A previous Korean multi-center study that studied ICU patients with malignancies did not investigate the long-term outcomes [4]. In our cohort, the majority of the patients who required ICU admission did not survive beyond a year. Only a few studies have reported long-term outcomes between cancer and non-cancer patients after ICU admission. The Dutch National Cohort study reported that patients with cancer had a significantly worse one-year mortality rate than patients without cancer [8]. The authors suggested that the underlying cancer types or their treatments were important factors in recovery from critical illness. Puxty et al. [18] reported that the 6-month and 4-year mortality rates of surgical ICU patients with cancer were higher than those of the group without cancer. They suggested that long-term survival was related to the underlying cancer, although short-term outcomes were more related to critical illness. Our study presented one-year mortality rates from ICU admission and from ICU discharge. Even after surviving from the ICU, cancer patients in our cohort still had high one-year mortality. We found that cancer stages, patients’ reception of further chemotherapy after ICU discharge, and patient and family’s willingness to receive life-sustaining treatment were associated with the long-term mortality of cancer patients after critical illness. In patients with solid malignancies, the treatment provided in the ICU can disrupt the ongoing cancer treatments due to a decline in patients’ functional status, active infections, or the development of new organ dysfunction, thereby delaying further cancer treatment until the patients’ general condition and the infection or organ dysfunction improve [19]. Such delays can lead to poor outcomes in patients with advanced cancer. Given the significant difference in one-year mortality based on whether chemotherapy was restarted after ICU discharge in our patients with metastatic cancer, evaluating the possibility of restarting chemotherapy after critical illness could be helpful in making decisions on life-sustaining treatment. Our results suggested that ICU admission itself should be considered as critical point for prognosis in patients with solid malignancy. Further, when considering life-sustaining treatments for these patients, decision-making should be based on an understanding of the expected short- and long-term prognosis based on that ICU admission and an informed discussion between the patients, their family, and the treating physician.

This study has certain limitation. Given that this was a single-center study, the generalizability of the study results is limited. However, our study included a large number of patients with data collected from one of the largest referral hospitals in Korea over 7 years after the introduction of targeted therapy and immunotherapy. We believe that these results can be indicative of contemporary clinical courses for critically ill patients with solid malignancy.

In conclusion, this study showed that critically ill patients with solid malignancy had poor 1-year survival despite relatively low ICU mortality. The types of underlying cancer and organ support therapies received during the ICU stay were significantly associated with ICU mortality. These findings highlight the need for careful consideration of ICU admission in patients with solid malignancy, and decision-making should be based on an understanding of the short- and long-term prognosis of ICU admission after an informed discussion between the patients, their families, and the treating physician.