The cancer control status and APACHE II score are prognostic factors for critically ill patients with cancer and sepsis

Outcome of Cancer Patients with an Unplanned Intensive Care Unit Admission: Predictors of Mortality and Long-term Survival

Background

Understanding the characteristics and outcomes of cancer patients with unplanned ICU admission is imperative for therapeutic decisions and prognostication purposes.

Objective

To describe the clinical characteristics of patients with hematological and non-hematological malignancies (NHM) who require unplanned ICU admission and to determine the predictors of mortality and long-term survival.

Methods

This retrospective study included all patients with cancer who had an unplanned ICU admission between 2011 and 2016 at a tertiary hospital in Saudi Arabia. The following variables were collected: age, gender, ICU length of stay (LOS), APACHE II score, type of malignancy, febrile neutropenia, source and time of admission, and need for mechanical ventilation (MV), renal replacement therapy (RRT), and treatment with vasopressors (VP). Predictors of mortality and survival rates at 28 days and 3, 6, and 12 months were calculated.

Results

The study included 410 cancer patients with 466 unplanned ICU admissions. Of these, 52% had NHM. The average LOS in the ICU was 9.6 days and the mean APACHE score was 21.9. MV was needed in 73% of the patients, RRT in 15%, and VP in 24%, while febrile neutropenia was present in 24%. There were statistically significant differences between survivors and non-survivors in the APACHE II score (17.7 ± 8.0 vs. 25.6 ± 9.2), MV use (52% vs. 92%), need for RRT (6% vs. 23%), VP use (42% vs. 85%), and presence of febrile neutropenia (18% vs. 30%). The predictors of mortality were need for MV (OR = 4.97), VP (OR = 3.43), RRT (OR = 3.31), and APACHE II score (OR = 1.10). Survival rates at 28 days, 3, 6, and 12 months were 52%, 28%, 22%, and 15%, respectively.

Conclusion

The survival rate of cancer patients with an unplanned admission to the ICU remains low. Predictors of mortality include need for MV, RRT, and VP and presence of febrile neutropenia. About 85% of cancer patients died within 1 year after ICU admission.

A simplified machine learning model utilizing platelet-related genes for predicting poor prognosis in sepsis

Background

Thrombocytopenia is a known prognostic factor in sepsis, yet the relationship between platelet-related genes and sepsis outcomes remains elusive. We developed a machine learning (ML) model based on platelet-related genes to predict poor prognosis in sepsis. The model underwent rigorous evaluation on six diverse platforms, ensuring reliable and versatile findings.

Methods

A retrospective analysis of platelet data from 365 sepsis patients confirmed the predictive role of platelet count in prognosis. We employed COX analysis, Least Absolute Shrinkage and Selection Operator (LASSO) and Support Vector Machine (SVM) techniques to identify platelet-related genes from the GSE65682 dataset. Subsequently, these genes were trained and validated on six distinct platforms comprising 719 patients, and compared against the Acute Physiology and Chronic Health Evaluation II (APACHE II) and Sequential Organ-Failure Assessment (SOFA) score.

Results

A PLT count <100×109/L independently increased the risk of death in sepsis patients (OR = 2.523; 95% CI: 1.084-5.872). The ML model, based on five platelet-related genes, demonstrated impressive area under the curve (AUC) values ranging from 0.5 to 0.795 across various validation platforms. On the GPL6947 platform, our ML model outperformed the APACHE II score with an AUC of 0.795 compared to 0.761. Additionally, by incorporating age, the model's performance was further improved to an AUC of 0.812. On the GPL4133 platform, the initial AUC of the machine learning model based on five platelet-related genes was 0.5. However, after including age, the AUC increased to 0.583. In comparison, the AUC of the APACHE II score was 0.604, and the AUC of the SOFA score was 0.542.

Conclusion

Our findings highlight the broad applicability of this ML model, based on platelet-related genes, in facilitating early treatment decisions for sepsis patients with poor outcomes. Our study paves the way for advancements in personalized medicine and improved patient care.

Identification of a novel gene signature for the prognosis of sepsis

Sepsis is a life-threatening organ dysfunction caused by the host's dysfunctional response to infection, and its pathogenesis is still unclear. In view of the complex pathological process of sepsis, finding suitable biomarkers is helpful for the research and treatment of sepsis. This study determined the potential prognostic markers of sepsis by analyzing the molecular characteristics of patients with sepsis. During this study, bioinformatics analysis was conducted on the RNA sequencing data and DNA methylation sites from the public database to determine the prognostic genes related to sepsis, and a 9-gene prognostic signature for sepsis was constructed. According to the risk score, all sepsis samples were divided into two groups. Then, the prediction effect of the 9-gene signature was verified in two cohorts, and the association between these genes and sepsis was further revealed through immune infiltration analysis, gene set enrichment analysis and the relationship between clinical phenotype and survival rate. Our study provided a reliable prognostic signature for sepsis. The signature could predict the survival of patients with sepsis and serve as a predictor.

Prediction of survival time after terminal extubation: the balance between critical care unit utilization and hospice medicine in the COVID-19 pandemic era

BACKGROUND

We established 1-h and 1-day survival models after terminal extubation to optimize ventilator use and achieve a balance between critical care for COVID-19 and hospice medicine.

METHODS

Data were obtained from patients with end-of-life status at terminal extubation from 2015 to 2020. The associations between APACHE II scores and parameters with survival time were analyzed. Parameters with a p-value ≤ 0.2 in univariate analysis were included in multivariate models. Cox proportional hazards regression analysis was used for the multivariate analysis of survival time at 1 h and 1 day.

RESULTS

Of the 140 enrolled patients, 76 (54.3%) died within 1 h and 35 (25%) survived beyond 24 h. No spontaneous breathing trial (SBT) within the past 24 h, minute ventilation (MV) ≥ 12 L/min, and APACHE II score ≥ 25 were associated with shorter survival in the 1 h regression model. Lower MV, SpO2 ≥ 96% and SBT were related to longer survival in the 1-day model. Hospice medications did not influence survival time.

CONCLUSION

An APACHE II score of ≥ 25 at 1 h and SpO2 ≥ 96% at 1 day were strong predictors of disposition of patients to intensivists. These factors can help to objectively tailor pathways for post-extubation transition and rapidly allocate intensive care unit resources without sacrificing the quality of palliative care in the era of COVID-19. Trial registration They study was retrospectively registered. IRB No.: 202101929B0.

Predictive Value of Heart-Type Fatty Acid-Binding Protein for Mortality Risk in Critically Ill Patients

Objective

Our study assessed the predictive value of heart-type fatty acid-binding protein (H-FABP) for critically ill patients.

Methods

150 critically ill patients admitted to the emergency department of Beijing Chaoyang Hospital, Capital Medical University, were included in our study from August 2021 to April 2022. Serum H-FABP, procalcitonin (PCT), lactate (LAC), and other markers were determined within 1 h after admission. The Sequential Organ Failure Assessment (SOFA) score and the Acute Physiology and Chronic Health Evaluation II (APACHE II) were calculated. The independent predictors of 28-day mortality in critically ill patients were analyzed by logistic regression, and the receiver operating characteristic curve (ROC) was used to analyze the predictive value for 28-day mortality in critically ill patients.

Results

Age, APACHE II, SOFA, GCS, LAC, H-FABP, IL-6, Scr, and D-dimer were significantly different in the nonsurvivor vs. survivor groups (P < 0.05), with H-FABP correlating with cTNI, Scr, PCT, and SOFA scores (P < 0.05). Logistic regression analysis showed that H-FABP, APACHE II, LAC, and age were independent predictors for 28-day mortality in critically ill patients (P < 0.05). The AUC of ROC curve in H-FABP was 0.709 (sensitivity 72.9%, specificity 66.1%, and cut-off 4.35), which was slightly lower than AUC of ROC curve in LAC (AUC 0.750, sensitivity 58.3%, specificity 76.1%, and cut-off 1.95) and APACHE II (AUC 0.731, sensitivity 77.1%, specificity 58.7%, and cut-off 12.5). However, statistically, there was no difference in the diagnostic value of H-FABP compared with the other two indicators (Z ₁ = 0.669, P = 0.504; Z ₂ = 0.383, P = 0.702). But H-FABP (72.9%) has higher sensitivity than LAC (58.3%). The combined evaluation of H-FABP+APACHE II score (AUC 0.801, sensitivity 71.7%, and specificity 78.2%; Z = 2.612, P = 0.009) had better diagnostic value than H-FABP alone and had high sensitivity (71.7%) and specificity (78.2%).

Conclusion

H-FABP, LAC, APACHE II, and age can be used as independent risk factors affecting the prognosis of critically ill patients. Compared with using the above indicators alone, the H-FABP+APACHE II has a high diagnostic value, and the early and rapid evaluation is particularly important for the adjustment of treatment plans and prognosis.

A Prognostic Model for In-Hospital Mortality in Critically Ill Patients with Pneumonia

Purpose

To determine the utility of a novel serum biomarker for the outcome prediction of critically ill patients with pneumonia.

Patients and Methods

A retrospective analysis of critically ill patients was performed at an emergency department. The expression and prediction value of parameters were assessed. Binary logistic regression analysis was utilized to determine the indicators associated with in-hospital mortality of pneumonia patients. The Last Absolute Shrinkage and Selection Operator was used to further determine the independent predictors, which were validated by multiple logistic regression. The receiver operator characteristic curve was performed to assess their prediction values. A prognostic nomogram model was finally established for the outcome prediction for critically ill patients with pneumonia.

Results

Retinol-binding protein (RBP) was significantly reduced in non-survived and pneumonia patients. CURB-65 score, levels of RBP, and blood urea nitrogen (BUN) were associated with in-hospital mortality of critically ill patients with pneumonia. Their combination was determined to be an ideal prognostic predictor (area under the curve of 0.762) and further developed into a nomogram prediction model (c-index 0.764).

Conclusion

RBP is a novel in-hospital mortality predictor, which well supplements the CURB-65 score for critical pneumonia patients.

All-cause mortality in cancer patients treated for sepsis in intensive care units: a systematic review and meta-analysis

Purpose

Sepsis is a common complication in patients with cancer, but studies evaluating the outcomes of critically ill cancer patients with sepsis on a global scale are limited. We aimed to summarize the existing evidence on mortality rates in this patient population.

Methods

Prospective and retrospective observational studies evaluating critically ill adult cancer patients with sepsis, severe sepsis, and/or septic shock were included. Studies published from January 2010 to September 2021 that reported at least one mortality outcome were retrieved from MEDLINE (Ovid), Embase (Ovid), and Cochrane databases. Study selection, bias assessment, and data collection were performed independently by two reviewers, and any discrepancies were resolved by a third reviewer. The risk of bias was assessed using the Newcastle–Ottawa scale. We calculated pooled intensive care unit (ICU), hospital, and 28/30-day mortality rates. The heterogeneity of the data was tested using the chi-square test, with a P value < 0.10 indicating significant heterogeneity.

Results

A total of 5464 citations were reviewed, of which 10 studies met the inclusion criteria; these studies included 6605 patients. All studies had a Newcastle–Ottawa scale score of 7 or higher. The mean patient age ranged from 51.4 to 64.9 years. The pooled ICU, hospital, and 28/30 day mortality rates were 48% (95% CI, 43– 53%; I² = 80.6%), 62% (95% CI, 58–67%; I² = 0%), and 50% (95% CI, 38– 62%; I² = 98%), respectively. Substantial between-study heterogeneity was observed.

Conclusion

Critically ill cancer patients with sepsis had poor survival, with a hospital mortality rate of about two-thirds. The substantial observed heterogeneity among studies could be attributed to variability in the criteria used to define sepsis as well as variability in treatment, the severity of illness, and care across settings. Our results are a call to action to identify strategies that improve outcomes for cancer patients with sepsis.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00520-022-07392-w.

Prognostic Value of Albumin-to-Fibrinogen Ratio for 28-Day Mortality among Patients with Sepsis from Various Infection Sites

Purpose

This study investigated the prognostic value of the albumin-to-fibrinogen ratio (AFR) in patients with sepsis as a consequence of infection at various sites.

Methods

A total of 300 patients with sepsis caused by various infection sites, who met the diagnostic criteria for sepsis hospitalized in the intensive care unit, were enrolled in this study. The observational endpoint was 28-day mortality. Cox proportional hazard regression analysis was performed to determine the potential prognostic factors for 28-day mortality in these septic patients. Receiver operating characteristic (ROC) curve analysis was used to evaluate and compare the prognostic factors for 28-day mortality.

Results

Of 300 participants, 147 died, corresponding to a 28-day mortality of 49% (147/300). Baseline Acute Physiology and Chronic Health Evaluation (APACHE II) score (hazard ratio (HR) 1.18 (95% confidence interval (CI) 1.07–1.30); P < 0.001), baseline lactic acid level (HR 1.27 (95% CI 1.08–1.50); P = 0.005), the presence of septic shock (HR 21.44 (95% CI 2.51–182.76); P = 0.005), and baseline AFR (HR 0.70 (95% CI 0.62–0.80); P < 0.001) were independent prognostic factors for 28-day mortality in patients with sepsis according to multivariate Cox analysis. Baseline AFR was an effective predictor of 28-day mortality, with an area under the ROC curve (AUC) of 0.700, and a specificity and sensitivity of 90.8% and 42.1%, respectively. A low baseline AFR level was associated with increased 28-day sepsis-related mortality. The quadruple index, which included the APACHE II score, lactic acid, septic shock, and AFR, showed a more accurate predictive value for septic patients than the APACHE II score, lactic acid, septic shock, and AFR alone, with an AUC of 0.922, and specificity and sensitivity of 86.9% and 83.6%, respectively. Moreover, the triple index, which included the APACHE II score, lactic acid, and septic shock, showed a significantly lower prognostic value for 28-day mortality compared with the ROC curve of the quadruple index and triple index, with an AUC of 0.877 and specificity and sensitivity of 77.8% and 82.3%, respectively.

Conclusions

The results of this study demonstrate that AFR is an independent protective factor for predicting 28-day mortality in patients with sepsis due to various infection sites. AFR combined with the APACHE II score, lactic acid, and septic shock showed a higher prognostic value for sepsis prognosis.

E-CatBoost: An efficient machine learning framework for predicting ICU mortality using the eICU Collaborative Research Database

Improving the Intensive Care Unit (ICU) management network and building cost-effective and well-managed healthcare systems are high priorities for healthcare units. Creating accurate and explainable mortality prediction models helps identify the most critical risk factors in the patients' survival/death status and early detect the most in-need patients. This study proposes a highly accurate and efficient machine learning model for predicting ICU mortality status upon discharge using the information available during the first 24 hours of admission. The most important features in mortality prediction are identified, and the effects of changing each feature on the prediction are studied. We used supervised machine learning models and illness severity scoring systems to benchmark the mortality prediction. We also implemented a combination of SHAP, LIME, partial dependence, and individual conditional expectation plots to explain the predictions made by the best-performing model (CatBoost). We proposed E-CatBoost, an optimized and efficient patient mortality prediction model, which can accurately predict the patients' discharge status using only ten input features. We used eICU-CRD v2.0 to train and validate the models; the dataset contains information on over 200,000 ICU admissions. The patients were divided into twelve disease groups, and models were fitted and tuned for each group. The models' predictive performance was evaluated using the area under a receiver operating curve (AUROC). The AUROC scores were 0.86 [std:0.02] to 0.92 [std:0.02] for CatBoost and 0.83 [std:0.02] to 0.91 [std:0.03] for E-CatBoost models across the defined disease groups; if measured over the entire patient population, their AUROC scores were 7 to 18 and 2 to 12 percent higher than the baseline models, respectively. Based on SHAP explanations, we found age, heart rate, respiratory rate, blood urine nitrogen, and creatinine level as the most critical cross-disease features in mortality predictions.

Value of CRP, PCT, and NLR in Prediction of Severity and Prognosis of Patients With Bloodstream Infections and Sepsis

Objective

To investigate the value of C-reactive protein (CRP), procalcitonin (PCT), and neutrophil to lymphocyte ratio (NLR) in assessing the severity of disease in patients with bloodstream infection and sepsis, and to analyze the relationship between the levels of three inflammatory factors and the prognosis of patients.

Methods

The clinical data of 146 patients with bloodstream infection and sepsis admitted to our intensive care unit (ICU) from October 2016 to May 2020 were retrospectively analyzed. The differences in the levels of inflammatory indicators such as CRP, PCT, and NLR within 24 h in patients with bloodstream infection sepsis with different conditions (critical group, non-critical group) and the correlation between these factors and the condition (acute physiology and chronic health evaluation II, APACHE II score) were analyzed. In addition, the prognosis of all patients within 28 days was counted, and the patients were divided into death and survival groups according to their mortality, and the risk factors affecting their death were analyzed by logistic regression, and the receiver operating characteristic (ROC) curve was used to analyze the value of the relevant indicators in assessing the prognosis of patients.

Results

The levels of NLR, CRP, PCT, total bilirubin (TBIL), glutamic oxaloacetic transaminase (AST), and serum creatinine (Scr) were significantly higher in the critically ill group than in the non-critically ill group, where correlation analysis revealed a positive correlation between CRP, PCT, and NLR and APACHE II scores (P < 0.05). Univariate logistic regression analysis revealed that CRP, PCT, NLR, and APACHE II scores were associated with patient prognosis (P < 0.05). Multi-factor logistic regression analysis found that PCT, NLR, and APACHE II scores were independent risk factors for patient mortality within 28 days (P < 0.05). ROC curve analysis found that PCT and NLR both had an AUC area > 0.7 in predicting patient death within 28 days (P < 0.05).

Conclusion

Inflammatory factors such as NLR, CRP, and PCT have important clinical applications in the assessment of the extent of disease and prognosis of patients with bloodstream infection and sepsis.

Values of serum PCT, suPAR combined with severity scores for evaluating prognosis of septic shock patients

Background: To explore the values of serum procalcitonin (PCT), soluble urokinase-type plasminogen activator receptor (suPAR) combined with APACHE II and SOFA scores for evaluating the prognosis of septic shock patients.

Materials and Methods: A total of 118 eligible patients admitted from August 2017 to January 2021 were divided into survival and death groups. Serum PCT and suPAR levels were detected. APACHE II and SOFA scores were evaluated. A combination predictor pre1 was constructed. The predictive efficacy of the indicator alone or in combination was compared using receiver operating characteristic curve. Risk factors leading to death were analyzed, and a predictive model was established.

Results: Serum PCT and suPAR levels as well as APACHE II and SOFA scores of death group significantly exceeded those of the survival group (P<0.05). PCT, suPAR, SOFA and APACHE II scores were valuable for predicting death. The area under curve (AUC) constructed by predictor pre1 for predicting death was largest. PCT, suPAR, APACHE II, and SOFA scores were independent risk factors for death. The model had AUC of 0.828, with the sensitivity of 86.54%, specificity of 89.03%, and accuracy of 82.47%. The death risk predicted by the model had a high concurrence with the actual one.

Conclusion: PCT, suPAR, APACHE II, and SOFA scores are closely related to the prognosis of septic shock patients. The combined predictor pre1 is more effective than a single index for predicting prognosis. The combined prediction model of septic shock based on PCT, suPAR, APACHE II, and SOFA scores has higher predictive efficiency.

Reducing Futile Acute Care Services for Terminally Ill Patients With Cancer: The Dignity Project

PURPOSE:

Patients with terminal diseases frequently undergo interventions that are futile and may be detrimental to their quality of life. We conducted a quality improvement project aimed to reduce the utilization of futile acute care services (ACSs) for patients with cancer treated with a palliative intent.

METHODS:

A multidisciplinary team reviewed the records of terminally ill patients with cancer who died between November 2017 and May 2018, during their admission at our institution. The review aimed to assess the magnitude of improper utilization of ACSs and admission to the intensive care unit (ICU). Lack of timely documentation of the goals of care (GOCs) was the main reason for this problem. We defined timely documentation as the availability of electronic documentation of patients' GOC before the need for ACSs. Interventions were implemented to improve the process; postintervention data were captured and compared with the baseline data.

RESULTS:

After the delivery of staff education and the implementation of mandatory documentation of the GOCs in the healthcare electronic record system, the timely documentation of the GOCs for patients with a palliative intent increased significantly from 59% at baseline to 83% in the postintervention phase. The impact of this intervention led to a decrease in admissions to the ICU from 26% to 12% and an estimated annual cost saving of $777,600 in US dollars.

CONCLUSION:

Our interventions resulted in improved documentation of the GOCs and decrease in the utilization of ACSs including ICU admissions and the associated cost.

Risk factors for postoperative pneumonia and prognosis in lung cancer patients after surgery: A retrospective study

Postoperative pneumonia (POP) is one of the most frequent complications following lung surgery. The aim of this study was to identify the risk factors for developing POP and the prognostic factors in lung cancer patients after lung resection.We performed a retrospective review of 726 patients who underwent surgery for stages I-III lung cancer at a single institution between August 2017 and July 2018 by conducting logistic regression analysis of the risk factors for POP. The Cox risk model was used to analyze the factors influencing the survival of patients with lung cancer.We identified 112 patients with POP. Important risk factors for POP included smoking (odds ratio [OR], 2.672; 95% confidence interval [CI], 1.586-4.503; P < .001), diffusing capacity for carbon monoxide (DLCO) (40-59 vs ≥80%, 4.328; 95% CI, 1.976-9.481; P < .001, <40 vs ≥80%, 4.725; 95% CI, 1.352-16.514; P = .015), and the acute physiology and chronic health evaluation (APACHE) II score (OR, 2.304; 95% CI, 1.382-3.842; P = .001). In the Cox risk model, we observed that age (hazard ratios (HR), 1.633; 95% CI, 1.062-2.513; P = .026), smoking (HR, 1.670; 95% CI, 1.027-2.716; P = .039), POP (HR, 1.637; 95% CI, 1.030-2.600; P = .037), etc were predictor variables for patient survival among the factors examined in this study.The risk factors for POP and the predictive factors affecting overall survival (OS) should be taken into account for effective management of patients with lung cancer undergoing surgery.

Changes in plasma HDL and its subcomponents HDL2b and HDL3 regulate inflammatory response by modulating SOCS1 signaling to affect severity degree and prognosis of sepsis

OBJECTIVES

To explore if SOCS1 is regulated by plasma HDL and its subcomponents HDL2b and HDL3 to affect inflammatory reaction then to influence the severity degree and prognosis of sepsis.

METHODS

One hundred sepsis patients in ICU and 85 normal control persons from October 2018 to October 2019 in our hospital were enrolled. Adult male C57BL/6 mice were used to establish sepsis model by CLP method. HDL, CRP, and WBC count of human were measured using an auto-analyzer. Plasma HDL, IL-1β, and TNF-α proteins levels of mice were measured with ELISA. Microfluidic chip was used for plasma HDL2b and HDL3 detections. SOCS1 in liver and spleen of mice were measured by qRT-PCR. The relationship between plasma HDL//HDL2b and inflammatory indices/SOCS1 in liver/spleen was analyzed with spearman correlation coefficient method. The sepsis patients/mice were divided into non-survival and survival groups. The sepsis patients were divided into severe and mild sepsis patients based on the SOFA score or divided into high and low score groups according to the APACHE II score. The sepsis mice were divided into high and low score group based on the modified sepsis severity score criterion.

RESULTS

Plasma HDL and HDL2b levels were significantly declined (P < 0.01), while HDL3 was normal in both sepsis patients and mice (P > 0.05). Plasma HDL and HDL2b were negatively associated with the serum CRP concentration and positively correlated with the prognosis and severity in sepsis patients (P < 0.05). Moreover, the downregulated plasma HDL but not HDL2b was negatively related to increased SOCS1 mRNA levels in liver and spleen of mice, which were positively connected with TNF-α and IL-1β protein levels (P < 0.05).

CONCLUSIONS

Plasma HDL is downregulated in sepsis, which may facilitate inflammatory reaction then activate the SOCS1 signaling to regulate the severity and affect prognosis of sepsis. The decline of plasma HDL2b content could aggravate the severity and poor prognosis of sepsis through facilitating inflammatory reaction. The plasma HDL3 is not involved in sepsis. The more and further explorations may be needed.