Intensive care unit outcomes and prognostic factors of colorectal cancer

Abstract

BACKGROUND

Colorectal cancer (CRC) is one of the most common cancers and CRC patients are among the most common intensive care unit (ICU) admitted cancer patients. However, their prognosis and evaluation methods are rarely studied.

AIM

To determine the short-term mortality outcome and identify the potential prognostic factors of CRC cancer patients admitted to the ICU.

METHODS

A multicenter cross-sectional study was performed from May 10, 2021 to July 10, 2021 at the ICU departments of 37 cancer specialized hospitals in China, and included patients aged ≥ 14 years with ICU duration ≥ 24 hours. Clinical records of patients with a primary CRC diagnosis were reviewed. Patients were separated into groups according to 90-day survival. Characteristics between groups were compared. Univariate and multivariate regression tests were used to analyze the correlated factors of ICU outcomes. Predictive values of disease severity scores were assessed using receiver operating characteristic curve analysis.

RESULTS

In total, 189 CRC patients were included in the study. The 90-day mortality was 12.2%. Patients who died showed differences compared to patients who survived mostly in terms of disease severity and ICU complications. It appears that patients admitted to the ICU from a clinical ward due to emergencies may have a higher risk of mortality while surgical management was associated with better survival. In multivariate analysis, only chemotherapy, elective surgery and conventional oxygen therapy were identified as independently correlated with 90-day mortality. Sequential organ failure assessment and acute physiology and chronic health evaluation II scores had moderate accuracy in predicting short-term mortality.

CONCLUSION

ICU admitted CRC patients appear to have low short-term mortality which requires further confirmation in prospective studies. The prognostic tools for these patients need further optimization.

Key Words: Intensive care unit; Colorectal cancer; Prognosis; Cancer-specialized hospitals; Clinical records

Core Tip: This study advances the literature by providing robust data on the prognosis of intensive care unit-admitted colorectal cancer patients, clarifying the role of clinical management on survival outcomes, and highlighting the need for improved prognostic tools. These findings are actionable for clinicians managing colorectal cancer patients and foundational for future research aimed at optimizing care and reducing mortality in this population.

INTRODUCTION

Advances in cancer treatment have significantly improved survival rates among cancer patients in recent years. During disease progression and treatment, intensive care unit (ICU) management may be warranted due to complications or treatment-associated side effects[1,2]. Alongside improved survival outcomes, the number of cancer patients requiring ICU care has risen substantially[3,4]. Balancing potential medical outcomes, individual rights and desires, and economic burdens necessitates identifying patients most likely to benefit from ICU admission and establishing timely prognostic evaluations, both of which hold significant clinical importance[5,6]. Although various studies on the characteristics and outcomes of ICU-admitted cancer patients have been conducted[7-10], available data remain fragmented, and consensus on admission criteria is still lacking.

Colorectal cancer (CRC) is the third most common cancer worldwide and the second leading cause of cancer-related deaths globally[11,12]. In China, CRC was the second most common cancer and the fourth most common cause of cancer death in 2020[13]. CRC patients may require ICU admission for postoperative complications, cancer-related emergencies (e.g., obstruction and perforation), treatment-related toxicities (e.g., sepsis and chemotherapy-induced organ dysfunction), systemic complications, or palliative care crises[14,15]. CRC is among the most common malignancies in ICU settings and represents the most frequent surgical admission to the ICU for cancer patients[16,17]. Compared to non-cancer ICU patients, CRC patients exhibit higher complication rates, longer hospital stays, poorer survival outcomes, and greater healthcare costs[18-20].

Understanding the characteristics and outcomes of CRC patients requiring ICU care is critical for optimizing management and improving prognoses. First, risk stratification is needed to identify high-risk populations and prioritize clinical monitoring[21,22]. Second, given the frequent constraints on ICU resources, effective risk assessment tools are essential to guide clinical decision-making and avoid unnecessary admissions[23,24]. Furthermore, accurate risk prediction of complications may facilitate tailored care, reducing mortality and morbidity[22,25].

Numerous studies have addressed risk assessment in CRC patients, but most focus on postoperative outcomes, as ICU transfer following surgery remains a common strategy despite clinical debate over its utility[26-28]. However, postoperative care is not the sole reason for ICU admission among CRC patients, and few studies have comprehensively analyzed the general characteristics and outcomes of this population[29]. This study utilizes data from a multicenter investigation conducted across 37 ICU departments in specialized cancer hospitals to analyze the characteristics and outcomes of CRC patients admitted to the ICU.

MATERIALS AND METHODS

Study design and population

Patients admitted into ICU departments of 37 cancer specialized hospitals in China from May 10, 2021 to July 10, 2021 were screened for a cross-sectional study. The clinical records of all admitted patients in the ICUs of participating centers were screened and considered eligible if the following information was complete: Reasons for admission to the ICU and underlying medical history, severity evaluation at the time of ICU admission, incidence of sepsis, acute respiratory distress syndrome, acute kidney injury (AKI), treatment of intensive care medicine-related diagnosis and treatment regimen, in-hospital outcomes, and 90-day survival at follow-up. Patients aged < 14 years and those with an ICU stay < 24 hours were excluded. Patients were separated into two groups according to the 90-day mortality outcome. Differences between the groups in terms of characteristics at baseline and during ICU management were analyzed to explore the predictors of short-term mortality. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. This study was approved by the Ethic Committee of Tianjin Medical University Cancer Institute and Hospital (No. bc2021065). The requirement for written informed consent was waived due to the retrospective study design. The study process is depicted in Figure 1.

Figure 1 Flowchart of research scheme. ICU: Intensive care unit; BMI: Body mass index; SOFA: Sequential organ failure assessment; APACHE: Acute physiology and chronic health evaluation; ROC: Receiver operating characteristic.

Data collection

This study retrospectively gathered data from patient clinical records. The extracted information encompassed demographics, clinical history, and illness severity scores (SOFA and APACHE II) at ICU admission. Furthermore, the data collection captured ICU monitoring information, critical condition reports, as well as treatment records. The occurrence of delirium, duration in the ICU, and mortality in the ICU, hospital, as well as a 90-day follow-up period were collected as outcomes.

Medical charts of individuals admitted to ICUs across participating centers were reviewed for eligibility, requiring comprehensive documentation of key parameters: Admission rationale and pre-existing comorbidities, illness severity scores at ICU entry, records on ICU complications such as occurrence of sepsis, acute respiratory distress syndrome, or AKI, therapeutic interventions aligned with critical care diagnoses, hospitalization outcomes, and 90-day post-discharge survival outcomes. Exclusion criteria comprised individuals under 14 years of age or those with an ICU duration shorter than 24 hours. The study was coordinated with data quality control. An electronic data collection Excel form was sent to participating centers. All data were extracted and filled in collection forms according to pre-set standards and then checked and confirmed by quality control officers. The final data of this study were then shared by all the participating units. This information has been added in the revised methods.

Statistical analysis

To assess the normality of continuous variables, the Shapiro-Wilk test was employed. As the data exhibited a non-normal distribution, continuous variables were summarized as median with interquartile range. Group comparisons of these variables were conducted using the Mann-Whitney U test. The categorical variables are presented as frequency counts and percentages (%). The Wilcoxon rank sum test was utilized for ordinal categorical data, and either the χ² test or Fisher’s exact test was applied for nominal categorical comparisons. Factors associated with 90-day mortality were evaluated using both univariate and multivariate Cox regression analyses. Variable selection for the multivariable model was performed via backward stepwise regression. The predictive performance of the SOFA and APACHE II scores was compared by generating time-dependent receiver operating characteristic (ROC) curves using the R package time ROC, and the area under the curve was computed for each. All analyses were performed using R version 4.3.1 (R Foundation for Statistical Computing), employing two-sided tests with a significance threshold of α = 0.05.

RESULTS

Baseline profile and 90-day mortality of CRC patients admitted to the ICU

The study consisted of 189 patients admitted to the ICU with a primary diagnosis of CRC. As detailed in Table 1, the 90-day survival rate for this group was 87.8% (n = 166). No significant differences were found in baseline demographics (age, gender, and BMI) when compared between the surviving and deceased patients. Deceased patients had a higher percentage of targeted therapy and chemotherapy history. In the deceased group, more patients were transferred unplanned to the ICU from a clinical ward, and surviving patients were more likely to undergo elective or emergency surgery. Deceased patients had higher severity scores and were more likely to experience complications including sepsis, respiratory failure, AKI, and shock on admission to the ICU. In addition, deceased patients tended to receive more anti-infection treatment and a lower percentage received conventional oxygen therapy instead of mechanical ventilation. These results indicated that the overall survival of CRC patients admitted to the ICU was good with a low mortality rate. Patients who died showed differences to those who survived mostly in terms of disease severity and ICU complications. It appears that patients admitted to the ICU from clinical wards due to emergencies may have a higher risk of death while surgical management was associated with better survival.

Table 1 Comparison of characteristics between intensive care unit colorectal cancer patients with different 90-day survival outcomes, n (%).

Variables	All (n = 189)	Survived at 90 days (n = 166)	90-day mortality (n = 23)	P value
Demographics
Age (years), median (IQR)	69.0 (60.0, 76.0)	69.0 (61.2, 76.0)	66.0 (57.5, 79.5)	0.663
Gender				0.650
Female	58 (30.7)	50 (30.1)	8 (34.8)
Male	131 (69.3)	116 (69.9)	15 (65.2)
BMI (kg/m2), median (IQR)	22.5 (19.8, 25.1)	22.9 (19.8, 25.2)	22.0 (20.9, 22.7)	0.460
Treatment history
Target therapy	16 (8.5)	9 (5.4)	7 (30.4)	0.001a
Immunotherapy	5 (2.6)	5 (3.0)	0 (0.0)	1.000
Chemotherapy	33 (17.5)	24 (14.5)	9 (39.1)	0.007a
Radiotherapy	8 (4.2)	6 (3.6)	2 (8.7)	0.252
Transferring source				< 0.001a
Operation room	119 (63.0)	114 (68.7)	5 (21.7)
Emergency department	7 (3.7)	6 (3.6)	1 (4.3)
Clinical ward	61 (32.3)	46 (27.7)	15 (65.2)
Other hospitals	2 (1.1)	0 (0.0)	2 (8.7)
Planned transfer	95 (50.3)	77 (53.6)	18 (21.7)	0.004a
Elective or emergency surgery				< 0.001a
No surgery	33 (17.5)	19 (11.4)	14 (60.9)
Elective	131 (69.3)	124 (74.7)	7 (30.4)
Emergency	25 (13.2)	23 (13.9)	2 (8.7)
Severity scores
SOFA, median (IQR)	3.0 (2.0, 5.0)	3.0 (1.0, 5.0)	5.0 (3.0, 9.5)	0.002a
APACHE II, median (IQR)	11.0 (8.0, 15.0)	10.0 (8.0, 14.0)	18.0 (14.5, 21.5)	< 0.001a
ICU diagnosis
Sepsis	104 (55.0)	85 (51.2)	19 (82.6)	0.005a
ARDS	17 (9.0)	14 (8.4)	3 (13.0)	0.441
Respiratory failure	48 (25.4)	38 (22.9)	10 (43.5)	0.034a
AKI				< 0.001a
None	171 (90.5)	155 (93.4)	16 (69.6)
Grade I	9 (4.8)	5 (3.0)	4 (17.4)
Grade II	3 (1.6)	0 (0.0)	3 (13.0)
Grade III	6 (3.2)	6 (3.6)	0 (0.0)
Shock	55 (29.1)	44 (26.5)	11 (47.8)	0.035a
Anti-infection treatment
Carbapenems	60 (31.7)	47 (28.3)	13 (56.5)	0.006a
β-lactam	59 (31.2)	50 (30.1)	9 (39.1)	0.382
Glycopeptides	28 (14.8)	22 (13.3)	6 (26.1)	0.119
Tigecycline	10 (5.3)	8 (4.8)	2 (8.7)	0.349
Echinocandins	2 (1.1)	1 (0.6)	1 (4.3)	0.229
Triazoles	8 (4.2)	5 (3.0)	3 (13.0)	0.059
Other treatment
Mechanical ventilation	76 (40.2)	65 (39.2)	11 (47.8)	0.427
Conventional oxygen therapy	174 (92.1)	158 (95.2)	16 (69.6)	0.001a
Sedation treatment	48 (25.4)	39 (23.5)	9 (39.1)	0.106

^aP < 0.05.

IQR: Interquartile range; BMI: Body mass index; SOFA: Sequential organ failure assessment; APACHE: Acute physiology and chronic health evaluation; ICU: Intensive care unit; ARDS: Acute respiratory distress syndrome; AKI: Acute kidney injury.

Multivariate analysis of predictors associated with mortality of ICU admitted CRC patients

Cox regression analysis, at both the univariate and multivariate setting, were performed to further explore potential outcome predictors. In univariate analysis, factors correlated with higher mortality risk included treatment history, transfer from a clinical ward, unplanned transfer, higher severity scores, ICU complication diagnosis (sepsis, respiratory failure, AKI, and shock) and certain anti-infection treatments (Table 2). Surgical management and conventional oxygen therapy in the ICU were associated with a lower risk of death. In multivariate analysis, only chemotherapy, elective surgery and conventional oxygen therapy were statistically identified as independent correlators of 90-day mortality.

Table 2 Multivariate regression analysis of predictors for 90-day survival of intensive care unit admitted colorectal cancer patients.

Variables, demographics	Univariate analysis1		Multivariate analysis
	HR (95%CI)	P value	HR (95%CI)	P value
Age (years)	0.99 (0.96, 1.03)	0.643
Gender
Female	Reference
Male	0.82 (0.35, 1.94)	0.658
BMI (kg/m2)	0.97 (0.86, 1.08)	0.549
Treatment history
Target therapy	5.43 (2.23, 13.22)	< 0.001
Immunotherapy	-	-
Chemotherapy	3.38 (1.46, 7.80)	0.004	2.66 (1.04, 6.80)	0.041a
Radiotherapy	2.18 (0.51, 9.28)	0.294
Transferring source
Operation room	Reference
Emergency department	3.85 (0.45, 32.96)	0.218
Clinical ward	6.65 (2.41, 18.30)	< 0.001
Other hospitals	-	-
Unplanned transfer	3.83 (1.42, 10.31)	0.008
Elective or emergency surgery
No surgery	Reference
Elective	0.10 (0.04, 0.25)	< 0.001	0.20 (0.07, 0.58)	0.003a
Emergency	0.15 (0.04, 0.68)	0.013	0.26 (0.06, 1.24)	0.091
Severity scores
SOFA	1.19 (1.10, 1.30)	< 0.001
APACHE II	1.12 (1.07, 1.17)	< 0.001	1.06 (1.00, 1.12)	0.071
ICU diagnosis
Sepsis	4.16 (1.42, 12.24)	0.010
ARDS	1.59 (0.47, 5.37)	0.451
Respiratory failure	2.46 (1.08, 5.62)	0.032
AKI	5.16 (2.12, 12.56)	< 0.001
Shock	2.41 (1.06, 5.46)	0.035
Anti-infection treatment
Carbapenems	3.10 (1.36, 7.07)	0.007
β-lactam	1.42 (0.61, 3.28)	0.411
Glycopeptides	2.19 (0.86, 5.56)	0.098
Tigecycline	1.84 (0.43, 7.85)	0.410
Echinocandins	4.19 (0.56, 31.13)	0.161	0.15 (0.01, 1.43)	0.099
Triazoles	3.95 (1.17, 13.31)	0.027
Other treatment
Mechanical ventilation	1.41 (0.62, 3.20)	0.410
Conventional oxygen therapy	0.15 (0.06, 0.38)	< 0.001	0.21 (0.07, 0.62)	0.005a
Sedation treatment	2.03 (0.88, 4.70)	0.097

^aP < 0.05.

¹In the univariate analysis section, blank sections for the variables immunotherapy and transferring source represent confidence intervals of no significance.

HR: Hazard ratio; CI: Confidence interval; BMI: Body mass index; SOFA: Sequential organ failure assessment; APACHE: Acute physiology and chronic health evaluation; ICU: Intensive care unit; ARDS: Acute respiratory distress syndrome; AKI: Acute kidney injury.

Predictive value of disease severity scores for survival of ICU admitted CRC patients

The multivariate analysis results indicated that the current data showed a lack of dependable predictors for short-term mortality of ICU admitted CRC patients. General functional status and disease severity score are commonly applied in clinical practice for evaluation of ICU patients. We therefore tested the predictive value of SOFA and APACHE II scores for short-term mortality of ICU admitted CRC patients. As shown in Figure 2 and Table 3, both scores showed moderate accuracy in predicting short-term mortality, with the highest area under the curve of time ROC at 0.797 when death within 80 days was predicted by the APACHE II score. These findings demonstrated that these scores may not be sufficient for the prognosis of ICU admitted CRC patients. Further exploration of predictors needs to be carried out in the future.

Figure 2 Receiver operating characteristic curves of 90-day survival prediction models for different time periods. A: The model at 30 days; B: The model at 60 days; C: The model at 80 days. SOFA: Sequential organ failure assessment; APACHE: Acute physiology and chronic health evaluation; ROC: Receiver operating characteristic.

Table 3 Area under the curve and 95% confidence intervals for prediction of mortality within 90 days of intensive care unit admitted colorectal cancer patients.

Time periods	AUC	95%CI
30 days
SOFA	0.708	0.550-0.865
APACHE II	0.779	0.6493-0.908
60 days
SOFA	0.697	0.559-0.835
APACHE II	0.785	0.673-0.897
80 days
SOFA	0.719	0.591-0.848
APACHE II	0.797	0.693-0.901

SOFA: Sequential organ failure assessment; APACHE: Acute physiology and chronic health evaluation; AUC: Area under the curve; CI: Confidence interval.

DISCUSSION

An analysis of data from 37 ICUs within specialized cancer hospitals identified 189 admissions for CRC over a two-month period. Following ICU care, the recorded 90-day mortality for these patients was 12.2%. Patients with 90-day mortality generally had more severe conditions before admission when compared to surviving patients, as well as more ICU complications and intensive treatment. Chemotherapy history, elective surgery and conventional oxygen therapy were the only significant associations identified in multivariate analysis. SOFA and APACHE II scores only had moderate predictive values for short-term mortality. These findings showed that ICU admitted CRC patients generally had a good survival prognosis; however, there is currently still a lack of dependable predictors for short-term mortality outcome.

CRC patients generally have good survival prognosis. A multi-national study with a large sample size identified a 5-year overall survival rate of 83.4% in screening diagnosed CRC patients[30]. The average in-hospital mortality was reported to be only 4.9%[31]. Specific data on ICU admitted CRC patients are rarely reported, a single center study suggested that a quarter of ICU admitted CRC patients died during hospitalization[29], which was comparable to 20%-30% overall in-hospital short-term mortality of ICU patients with solid tumors[32-34]. The 90-day mortality after ICU care of CRC patients was found to be 12.2% in the present study. These results again appear to suggest that CRC patients were among the lower risk populations even when intense clinical management is required. However, due to the cross-sectional nature of the current study, it may be premature to draw conclusions. The general prognosis of ICU admitted CRC patients, especially the stratified risks of patients admitted for different reasons, require further confirmation in future longitudinal studies.

Post-surgery care is the main reason for many CRC patients admitted to the ICU. Previous studies have found that the risk factors for planned and unplanned post-surgery admission to the ICU included older age, male gender, lower BMI, type 2 diabetes mellitus, coronary heart disease, and advanced tumor stages[35]. In this study, the majority of ICU admitted CRC patients were transferred from the operation room. Compared to those transferred from clinical wards, patients who received post-surgery care in the ICU, especially those with a planned transfer, had significantly better survival outcomes. A higher percentage of clinical ward transfer was found in deceased patients, indicating that an emergency during disease progression or treatment was more likely to be associated with mortality risk. A previous study showed that the main emergencies related to CRC included intestinal obstruction, hemorrhage, and perforation, with advanced age, ethnicity, comorbidities, and more advanced stage as associated factors[36]. The reviewed data did not include detailed descriptions of cancer-related causes of ICU admission, therefore the risk factors for more severe ICU conditions and higher mortality risks require further exploration in future studies.

Previous studies assessing risk factors for post-operative adverse outcomes in CRC patients have suggested multiple potential predictors. Malnutrition has been suggested to be an important factor. For example, the modified frailty index was suggested to predict adverse outcomes in colon cancer patients undergoing surgical intervention[26]. Subjective global assessment was among predictors of postoperative recovery and survival after CRC surgery[37,38]. Controlled nutritional status score was significantly associated with postoperative patient outcomes and complications, including polyacrylate polyalcohol copolymer[28]. The geriatric nutritional risk index combined with calf circumference is a good predictor of prognosis in patients undergoing surgery for gastric cancer or CRC[39]. Other suggested predictors included age, pre-existing cardiovascular disease or anemia, serum lactate and lactate dehydrogenase levels, and inflammatory markers[15,40-42]. In this study, no specific factors were identified to predict short-term mortality. These results may indicate that ICU admitted CRC patients, who were not only transferred for post-operative complication management, may have more complicated conditions and need to be assessed more carefully for risk stratification.

The prognosis of CRC in the ICU may be influenced by multiple factors. Generally, CRC prognosis is significantly affected by comorbidities and frailty[43]. Age is associated with increased in-hospital mortality[44]. Clinical symptoms and pathological factors significantly impact the prognosis of CRC. Factors associated with worse prognosis included tumor spread beyond the bowel wall and regional lymph node involvement[45], as well as poorly differentiated adenocarcinomas, lymphangitic type[46], while vegetant gross tumors, papillary microscopic forms, well to moderately differentiated tumors may be favorable predictors[47]. In addition, CRC treatment relies on pathological assessment of resected specimens. Molecular subtyping of CRC can enhance understanding of tumor biology and aid in personalizing treatment strategies[48]. Due to the limited sample size, detailed analysis of associated factors could not generate useful information facilitating personalized risk stratification and prognosis. More in-depth analysis is warranted in future studies.

There are some limitations in this study, for example, sample size and the completeness of available data. Moreover, due to limited data, we were unable to identify specific indicators for adverse prognosis of ICU admitted CRC patients. The correlations found in both univariate and multivariate analysis with mortality were mostly indicators of disease severity, such as treatment history and elective surgery which may indicate the manageability of the disease. More records of ICU complications and related treatment also may only reflect the seriousness of conditions on ICU admission. No specific predictor was identified from the current data. The prognosis evaluation may still depend on current available disease severity evaluation tools such as SOFA and APACHE II, although their predictive values were also moderate. However, given the scarcity of research focusing on this specific patient group, our study addresses a gap in the literature regarding the provision of further prognostic information on ICU-admitted CRC patients.

CONCLUSION

A low short-term mortality rate of ICU admitted CRC cancer patients was observed, which needs to be further confirmed by longitudinal studies with a larger sample size. Mortality risk was higher in patients transferred from clinical wards. The prognostic tools for these patients need to be further optimized.