Introduction

Very elderly critically ill patients (ie, those older than 75 or 80 years) are an increasing population in intensive care units (ICU) [1, 2]. However, there is much controversy on how these patients should be managed. Although selected patients seem to benefit from ICU admission, elderly patients tend to have more ICU rejections [3] and a policy of systematic ICU admission of these patients has no positive effect on both short and long-term outcomes [4].

Recent research has focused on the epidemiology of elderly patients to identify the clinical path they follow after the ICU admission. As expected, those studies have showed increased mortality rates when compared to younger patients. However, mortality rates in the range of 20–40% may be considered acceptable under such patients’ conditions [5–9]. Nevertheless, the vast majority of studies in this population were performed in high-income countries [5–9], which raises concerns on the generalization of results to low-and, more particularly, middle-income countries where half of the world population live and demographic changes have been accelerated [10].

In parallel, a considerable improvement in outcomes of critically ill cancer patients was observed in recent years [11]. Despite these findings, cancer is still considered a condition predisposing denial of ICU admission in elderly critically ill patients [4]. However, as anticancer therapies become safer and more effective, a larger number of elderly individuals are being treated [12, 13]. Moreover, the increased indications of surgical and pharmacologic interventions have expanded the need for intensive care for monitoring after procedures or treatment-related complications. Although there is no evidence to manage elderly critically ill patients with cancer differently from patients without cancer, evidence supporting this premise is very scarce [14]. To bridge this gap, we performed a multicenter study in Brazil to describe clinical characteristics, outcomes and to identify factors associated with hospital mortality of a large cohort of patients aged 80 years or older with cancer admitted to ICUs.

Methods

Results

There were 4604 eligible patients in the two databases during the study periods (Fig 1). Out of them, 856 (18.6%) patients were 90 years or older; 22 (0.5%) were 100 years or older. A total of 980 (21.3%) of patients died in ICU and 1807 (39.2%) died in hospital. Patients who did not survive to hospital stay had more metastatic and hematologic tumors, were more commonly admitted for medical reasons, had worse performance status and a higher burden of comorbidities. On the other hand, age was not associated with higher hospital mortality. Deceased patients used more invasive mechanical ventilation, vasopressor and renal replacement therapies during ICU LOS. Patients who deceased also have longer ICU and hospital length-of-stays (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Study flowchart.

https://doi.org/10.1371/journal.pone.0238124.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Patients’ characteristics.

https://doi.org/10.1371/journal.pone.0238124.t001

Most patients had solid tumors (n = 4186, 90.9%). Prostate, colorectal, breast and lung were the most common site of solid tumors. Lung cancer was more common among patients who deceased, while prostate cancer was more common among patients who survived to hospital discharge. Among patients with hematological tumors (n = 418, 9.1%), lymphomas, leukemias and multiple myeloma were the most common type of tumors and occurred at similar patterns among patients who did and did not survive to hospital discharge (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Type of tumors.

https://doi.org/10.1371/journal.pone.0238124.t002

There was no multicollinearity among the predefined variables to be included in the logistic regression model (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Variation inflation index of the selected variables to be included in the logistic regression model.

https://doi.org/10.1371/journal.pone.0238124.t003

Metastatic cancer and hematologic cancer were independently associated with hospital mortality in comparison with locoregional solid tumors. Emergency surgical admissions were associated with lower hospital mortality than medical admissions. Performance status impairment, need for mechanical ventilation, vasopressors and renal replacement therapy were all also associated with increased hospital mortality (Fig 2). The model had a good discrimination (S1 Fig) power and was well calibrated (S2 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Mixed effect logistic regression analysis for risk factors associated with hospital mortality.

https://doi.org/10.1371/journal.pone.0238124.g002

Discussion

Our study showed that main clinical short-term outcomes for elderly critically ill cancer patients requiring ICU admission are reasonable. The observed mortality rates were comparable to those reported in the literature for younger patients with cancer [11, 18] or undifferentiated elderly critically ill patients [9, 19]. In the present study, type of cancer, type of ICU admission, performance status and acute organ dysfunctions were all associated with increased hospital mortality.

Patients with cancer have been only a minority in large epidemiological studies of very old critically ill patients [5, 8]. Our study suggests elderly critically ill patients with cancer had a similar hospital mortality to those without cancer reported in previous studies [5, 7]. Use of vasopressors and renal replacement therapy in the patients of our study were similar to those of the studies of undifferentiated elderly critically ill patients [5–7], reinforcing similar severity among them.

Among baseline characteristics, performance status and type of cancer were associated with increased mortality. Performance status has widely been known as a prognostic factor in critically ill patients [17] and also, specifically, in elderly [20] and patients with cancer admitted to ICU [21]. As expected, the proportion of patients with at least moderate impairment of performance status in our study (60%) was much higher than that reported in studies of general critically ill patients [17]. However, the rates of performance status impairment found in our study are even higher than that found in some cohorts of elderly critically ill patients [22], suggesting that cancer burden may impact negatively on performance status of very old patients.

One specific finding of our study was that metastatic solid cancer and hematologic cancer were also associated with increased hospital mortality. Previous studies in critically ill patients with cancer have not consistently shown that type of cancer had such an impact [21]. It seems cancer status has a role only in mortality beyond 120 days [23, 24]. Our study suggests that, in a different manner from their younger counterparts, type of cancer has an impact on short-term outcomes of elderly critically ill patients. Therefore, performance status and type of cancer are two known characteristics at ICU admission which may be taken into account on decisions for care and on prognostic information of elderly critically ill patients with cancer.

Understanding the epidemiology and outcomes of elderly critically ill patients in middle-income countries is of paramount importance because these countries have been facing with a more accelerated demographic change than that faced by high-income countries in the end of 20^th century [10]. Additionally, since rational resource utilization has been another challenge faced by middle-income countries, collaboration among intensivists, oncologists [25] and geriatricians [26] to the care of these patients will be fundamental.

Our study has some limitations. First, we could not assess frailty in our study patients. Frailty has been shown to be an important marker of mortality in critically ill, elderly and cancer patients [27]. Second, we also did not assess withholding or withdrawal of life-sustaining therapies. These decisions are obviously associated with increased mortality rates and may have had an impact on study patients’ outcomes [28]. Third, we only have data on patients who were admitted to ICU. We cannot point out whether patients’ which ICU admission was refused were more or less severely ill than those admitted neither they have or not worse outcomes. Finally, data on long-term survival and patient reported outcomes other than short-term mortality were unavailable.

Our study has also some strengths. It is a large multicenter cohort in middle-income country. Therefore, it adds information to those of high-income studies on elderly critically ill patients. Additionally, it sheds light on a population poorly described in the previous studies: elderly critically ill patients with cancer. Although these patients tend to have their ICU admission refused, it seems they also benefit from ICU admission in terms of short-term mortality. Whether this benefit is sustainable to long-term outcomes is current unknown.

In conclusion, elderly critically ill patients with cancer have comparable short-term outcomes after ICU admission to those of elderly patients without cancer reported in literature. Traditional markers of severity, such as medical admissions, use of life-sustaining therapies and performance status are also associated with increased mortality in elderly patients with cancer. In a different manner from their younger counterparts, metastatic solid cancer and hematologic cancer are associated with increased mortality in elderly critically ill patients and should be taken into consideration when planning care of these patients.

Supporting information