Assessment of performance of four mortality prediction systems in a Saudi Arabian intensive care unit

A Novel Composite Indicator of Predicting Mortality Risk for Heart Failure Patients With Diabetes Admitted to Intensive Care Unit Based on Machine Learning

BACKGROUND

Patients with heart failure (HF) with diabetes may face a poorer prognosis and higher mortality than patients with either disease alone, especially for those in intensive care unit. So far, there is no precise mortality risk prediction indicator for this kind of patient.

METHOD

Two high-quality critically ill databases, the Medical Information Mart for Intensive Care IV (MIMIC-IV) database and the Telehealth Intensive Care Unit (eICU) Collaborative Research Database (eICU-CRD) Collaborative Research Database, were used for study participants' screening as well as internal and external validation. Nine machine learning models were compared, and the best one was selected to define indicators associated with hospital mortality for patients with HF with diabetes. Existing attributes most related to hospital mortality were identified using a visualization method developed for machine learning, namely, Shapley Additive Explanations (SHAP) method. A new composite indicator ASL was established using logistics regression for patients with HF with diabetes based on major existing indicators. Then, the new index was compared with existing indicators to confirm its discrimination ability and clinical value using the receiver operating characteristic (ROC) curve, decision curve, and calibration curve.

RESULTS

The random forest model outperformed among nine models with the area under the ROC curve (AUC) = 0.92 after hyper-parameter optimization. By using this model, the top 20 attributes associated with hospital mortality in these patients were identified among all the attributes based on SHAP method. Acute Physiology Score (APS) III, Sepsis-related Organ Failure Assessment (SOFA), and Max lactate were selected as major attributes related to mortality risk, and a new composite indicator was developed by combining these three indicators, which was named as ASL. Both in the initial and external cohort, the new indicator, ASL, had greater risk discrimination ability with AUC higher than 0.80 in both low- and high-risk groups compared with existing attributes. The decision curve and calibration curve indicated that this indicator also had a respectable clinical value compared with APS III and SOFA. In addition, this indicator had a good risk stratification ability when the patients were divided into three risk levels.

CONCLUSION

A new composite indicator for predicting mortality risk in patients with HF with diabetes admitted to intensive care unit was developed on the basis of attributes identified by the random forest model. Compared with existing attributes such as APS III and SOFA, the new indicator had better discrimination ability and clinical value, which had potential value in reducing the mortality risk of these patients.

Development of a Novel Simple Model to Predict Mortality in Patients With Systemic Lupus Erythematosus Admitted to the Intensive Care Unit

Background: Patients with systemic lupus erythematosus (SLE) may sometimes require admission to the intensive care unit (ICU), and the outcome is poor. The aim of this study was to explore the clinical features of patients with SLE in the ICU, identify prognostic factors, and develop and evaluate a prognostic model to predict in-ICU mortality of patients with SLE.

Patients and Methods: This was a single center retrospective study in a tertiary medical institution in China. A total of 480 SLE patients with 505 ICU admissions from 2010 to 2019 were screened, and 391 patients were enrolled. The clinical feature and outcomes of the patients were analyzed. According to the random number table, patients were divided into two mutually exclusively groups named derivation (n = 293) and validation (n = 98). Prognostic factors were identified by a Cox model with Markov Chain Monte Carlo simulation and evaluated by latent analysis. The risk score was developed based on the derivation group and evaluated using the validation group.

Results: Among the 391 patients, 348 (89.0%) patients were females. The median age of patients was 34 years, and the median course of SLE was 6 months. The median APACHE II and SLEDAI were 17 and 10, respectively. The average in-ICU mortality was 53.4% (95% CI, 48.5–58.4%). A total of 186 patients were admitted to the ICU due to infection. Pneumonia (320/391, 81.8%) was the most common clinical manifestation, followed by renal disease (246/391, 62.9%). Nine prognostic factors were identified. The model had C statistic of 0.912 (95% CI, 0.889–0.948) and 0.807 (95% CI 0.703–0.889), with predictive range of 5.2–98.3% and 6.3–94.7% for the derivation and validation groups, respectively. Based on distribution of the risk score, 25.3, 49.5, and 25.2% of patients were stratified into the high, average, and low-risk groups, with corresponding in-ICU mortality of 0.937, 0.593, and 0.118, respectively.

Conclusion: Nine prognostic factors including age, white blood cell count, alanine transaminase, uric acid, intracranial infection, shock, intracranial hemorrhage, respiratory failure, and cyclosporin A/tacrolimus usage were identified. A prognostic model was developed and evaluated to predict in-ICU mortality of patients with SLE. These findings may help clinicians to prognostically stratify patients into different risk groups of in-ICU mortality, and provide patients with intensive and targeted management.

Development and validation of a new score for predicting functional outcome of neurocritically ill patients: The INCNS score

Aims

To develop and validate a novel score for prediction of 3‐month functional outcome in neurocritically ill patients.

Methods

The development of the novel score was based on two widely used scores for general critical illnesses (Acute Physiology and Chronic Health Evaluation II, APACHE II; Simplified Acute Physiology Score II, SAPS II) and consideration of the characteristics of neurocritical illness. Data from consecutive patients admitted to neurological ICU (N‐ICU) between January 2013 and June 2016 were used for the validation. The modified Rankin Scale (mRS) was used to evaluate 3‐month functional outcomes. APACHE II scores, SAPS II scores, and our novel scores at 24 hours and 72 hours in N‐ICU were obtained. We compared the prognostic performance of our score with APACHE II and SAPS II.

Results

We developed a 44‐point scoring system named the INCNS score, and it includes 19 items which were categorized into five parts: inflammation (I), nutrition (N), consciousness (C), neurological function (N), and systemic function (S). We validated the INCNS score with a cohort of 941 N‐ICU patients. The 72‐hours INCNS score achieved an area under the receiver operating characteristic curve (AUC) of 0.828 (95% CI: 0.802‐0.854), and the 24‐hours INCNS score achieved an AUC of 0.788 (95% CI: 0.759‐0.817). The INCNS score exhibited significantly better discriminative and prognostic performance than APACHE II and SAPS II at both 24 hours and 72 hours in N‐ICU.

Conclusion

We developed an INCNS score with superior predictive power for functional outcome of neurocritically ill patients.

Association of Obesity with Increased Mortality in the Critically Ill Patient

The impact of obesity on critical care outcomes has been an issue for debate in the literature. Variable data and conflicting results have been reported. The purpose of our study is to examine the impact of obesity on the outcome of patients admitted to a tertiary closed Intensive Care Unit (ICU) in Saudi Arabia.

Data was obtained from a prospectively collected database from September 2001 to May 2004. Patients younger than 18, those with burns, brain death and readmissions were excluded. The study population was stratified into six groups according to their Body Mass Index (BMI). Primary endpoints were ICU and hospital mortality, duration of mechanical ventilation and ICU length of stay.

A total of 1835 patients were included in the analysis. Baseline characteristics were similar among the six groups including severity of illness scores, reflected by Acute Physiology and Chronic Health Evaluation II (APACHE II) scores. The ICU mortality was not statistically different among the groups. Hospital mortality was lower in patients with BMI 35–39.9 kg/m2 and BMI >40 kg/m2 compared to those with BMI 18.5–24.9 kg/m2. Multivariate analysis showed that a BMI >40 kg/m2 was an independent predictor of lower hospital mortality (odds ratio 0.51, 95% confidence interval 0.28–0.92, P 0.025) after adjustment for other confounding factors.

In conclusion, mortality of obese critically ill patients was not higher than patients with normal weight. In fact, the hospital mortality was lower for patients with BMI >40 kg/m2 compared to the normal BMI group despite similar severity of illness. Obesity might have a protective effect, although further studies are needed to substantiate this finding.

Simplified prognostic model for critically ill patients in resource limited settings in South Asia

BACKGROUND

Current critical care prognostic models are predominantly developed in high-income countries (HICs) and may not be feasible in intensive care units (ICUs) in lower- and middle-income countries (LMICs). Existing prognostic models cannot be applied without validation in LMICs as the different disease profiles, resource availability, and heterogeneity of the population may limit the transferability of such scores. A major shortcoming in using such models in LMICs is the unavailability of required measurements. This study proposes a simplified critical care prognostic model for use at the time of ICU admission.

METHODS

This was a prospective study of 3855 patients admitted to 21 ICUs from Bangladesh, India, Nepal, and Sri Lanka who were aged 16 years and over and followed to ICU discharge. Variables captured included patient age, admission characteristics, clinical assessments, laboratory investigations, and treatment measures. Multivariate logistic regression was used to develop three models for ICU mortality prediction: model 1 with clinical, laboratory, and treatment variables; model 2 with clinical and laboratory variables; and model 3, a purely clinical model. Internal validation based on bootstrapping (1000 samples) was used to calculate discrimination (area under the receiver operating characteristic curve (AUC)) and calibration (Hosmer-Lemeshow C-Statistic; higher values indicate poorer calibration). Comparison was made with the Acute Physiology and Chronic Health Evaluation (APACHE) II and Simplified Acute Physiology Score (SAPS) II models.

RESULTS

Model 1 recorded the respiratory rate, systolic blood pressure, Glasgow Coma Scale (GCS), blood urea, haemoglobin, mechanical ventilation, and vasopressor use on ICU admission. Model 2, named TropICS (Tropical Intensive Care Score), included emergency surgery, respiratory rate, systolic blood pressure, GCS, blood urea, and haemoglobin. Model 3 included respiratory rate, emergency surgery, and GCS. AUC was 0.818 (95% confidence interval (CI) 0.800-0.835) for model 1, 0.767 (0.741-0.792) for TropICS, and 0.725 (0.688-0.762) for model 3. The Hosmer-Lemeshow C-Statistic p values were less than 0.05 for models 1 and 3 and 0.18 for TropICS. In comparison, when APACHE II and SAPS II were applied to the same dataset, AUC was 0.707 (0.688-0.726) and 0.714 (0.695-0.732) and the C-Statistic was 124.84 (p < 0.001) and 1692.14 (p < 0.001), respectively.

CONCLUSION

This paper proposes TropICS as the first multinational critical care prognostic model developed in a non-HIC setting.

Applicability of the APACHE II model to a lower middle income country

PURPOSE

To determine the utility of APACHE II in a low-and middle-income (LMIC) setting and the implications of missing data.

MATERIALS AND METHODS

Patients meeting APACHE II inclusion criteria admitted to 18 ICUs in Sri Lanka over three consecutive months had data necessary for the calculation of APACHE II, probabilities prospectively extracted from case notes. APACHE II physiology score (APS), probabilities, Standardised (ICU) Mortality Ratio (SMR), discrimination (AUROC), and calibration (C-statistic) were calculated, both by imputing missing measurements with normal values and by Multiple Imputation using Chained Equations (MICE).

RESULTS

From a total of 995 patients admitted during the study period, 736 had APACHE II probabilities calculated. Data availability for APS calculation ranged from 70.6% to 88.4% for bedside observations and 18.7% to 63.4% for invasive measurements. SMR (95% CI) was 1.27 (1.17, 1.40) and 0.46 (0.44, 0.49), AUROC (95% CI) was 0.70 (0.65, 0.76) and 0.74 (0.68, 0.80), and C-statistic was 68.8 and 156.6 for normal value imputation and MICE, respectively.

CONCLUSIONS

An incomplete dataset confounds interpretation of prognostic model performance in LMICs, wherein imputation using normal values is not a suitable strategy. Improving data availability, researching imputation methods and developing setting-adapted and simpler prognostic models are warranted.

Continuous Mandatory Onsite Consultant Intensivists in the ICU: Impacts on Patient Outcomes

OBJECTIVE

The aim of this study was to compare the impacts on patient outcomes of continuous versus on-demand access to certified consultant intensivists in the intensive care unit (ICU).

METHODS

Two general adult ICUs within the same health-care organization were compared in terms of patient outcomes. One unit featured continuous mandatory presence of a consultant intensivist (unit A), whereas the other had continuous access to a consultant intensivist during daytime hours but only on-demand access during the night-time hours (unit B). The data collected from these 2 units over the same 12-month period included sex, age, APACHE II score, disease category (medical, surgical, or traumatic), ICU mortality, and length of stay. A subgroup analysis was undertaken to assess the impact of disease severity, age, sex, and disease category on mortality.

RESULTS

When adjusted for disease severity, mortality was significantly lower in unit A with continuous mandatory 24-hour presence of a consultant intensivist compared with unit B with on-demand access to a consultant intensivist after working hours. Old age, female sex, and a higher APACHE II score were associated with poorer outcomes at both sites. The subgroup analysis revealed that the difference in mortality was only significant among medical patients but not among surgical or trauma patients.

CONCLUSIONS

An improved survival rate was observed only among medical patients admitted to the ICU with mandatory continuous access to a consultant intensivist, despite the presence of greater disease severity in the population admitted to this unit.

Comparison of APACHE II and SAPS II Scoring Systems in Prediction of Critically Ill Patients' Outcome

INTRODUCTION

Using physiologic scoring systems for identifying high-risk patients for mortality has been considered recently. This study was designed to evaluate the values of Acute Physiology and Chronic Health Evaluation II (APACHE II) and Simplified Acute Physiologic Score (SAPS II) models in prediction of 1-month mortality of critically ill patients.

METHODS

The present prospective cross sectional study was performed on critically ill patients presented to emergency department during 6 months. Data required for calculation of the scores were gathered and performance of the models in prediction of 1-month mortality were assessed using STATA software 11.0.

RESULTS

82 critically ill patients with the mean age of 53.45 ± 20.37 years were included (65.9% male). Their mortality rate was 48%. Mean SAPS II (p < 0.0001) and APACHE II (p = 0.0007) scores were significantly higher in dead patients. Area under the ROC curve of SAPS II and APACHE II for prediction of mortality were 0.75 (95% CI: 0.64 - 0.86) and 0.72 (95% CI: 0.60 - 0.83), respectively (p = 0.24). The slope and intercept of SAPS II were 1.02 and 0.04, respectively. In addition, these values were 0.92 and 0.09 for APACHE II, respectively.

CONCLUSION

The findings of the present study showed that APACHE II and SAPS II had similar value in predicting 1-month mortality of patients. Discriminatory powers of the mentioned models were acceptable but their calibration had some amount of lack of fit, which reveals that APACHE II and SAPS II are partially perfect.

Better prognostic marker in ICU - APACHE II, SOFA or SAP II!

Objectives:

This study was designed to determine the comparative efficacy of different scoring system in assessing the prognosis of critically ill patients.

Methods:

This was a retrospective study conducted in medical intensive care unit (MICU) and high dependency unit (HDU) Medical Unit III, Civil Hospital, from April 2012 to August 2012. All patients over age 16 years old who have fulfilled the criteria for MICU admission were included. Predictive mortality of APACHE II, SAP II and SOFA were calculated. Calibration and discrimination were used for validity of each scoring model.

Results:

A total of 96 patients with equal gender distribution were enrolled. The average APACHE II score in non-survivors (27.97+8.53) was higher than survivors (15.82+8.79) with statistically significant p value (<0.001). The average SOFA score in non-survivors (9.68+4.88) was higher than survivors (5.63+3.63) with statistically significant p value (<0.001). SAP II average score in non-survivors (53.71+19.05) was higher than survivors (30.18+16.24) with statistically significant p value (<0.001).

Conclusion:

All three tested scoring models (APACHE II, SAP II and SOFA) would be accurate enough for a general description of our ICU patients. APACHE II has showed better calibration and discrimination power than SAP II and SOFA.

Comparison of Glasgow Coma Scale, Full Outline of Unresponsiveness and Acute Physiology and Chronic Health Evaluation in Prediction of Mortality Rate Among Patients With Traumatic Brain Injury Admitted to Intensive Care Unit

Background

Traumatic brain injury (TBI) is a common cause of mortality and disability worldwide. Choosing an appropriate diagnostic tool is critical in early stage for appropriate decision about primary diagnosis, medical care and prognosis.

Objectives

This study aimed to compare the Glasgow coma scale (GCS), full outline of unresponsiveness (FOUR) and acute physiology and chronic health evaluation (APACHE II) with respect to prediction of the mortality rate of patients with TBI admitted to intensive care unit.

Patients and Methods

This diagnostic study was conducted on 80 patients with TBI in educational hospitals. The scores of APACHE II, GCS and FOUR were recorded during the first 24 hours of admission of patients. In this study, early mortality means the patient death before 14 days and delayed mortality means the patient death 15 days after admitting to hospital. The collected data were analyzed using descriptive and inductive statistics.

Results

The results showed that the mean age of the patients was 33.80 ± 12.60. From a total of 80 patients with TBI, 16 (20%) were females and 64 (80%) males. The mortality rate was 15 (18.7%). The results showed no significant difference among three tools. In prediction of early mortality, the areas under the curve (AUCs) were 0.92 (CI = 0.95. 0.81 - 0.97), 0.90 (CI = 0.95. 0.74 - 0.94), and 0.96 (CI = 0.95. 0.87 - 0.9) for FOUR, APACHE II and GCS, respectively. In delayed mortality, the AUCs were 0.89 (CI = 0.95. 0.81-0.94), 0.94 (CI = 0.95. 0.74 - 0.97) and 0.90 (CI = 0.95. 0.87 - 0.95) for FOUR, APACHE II and GCS, respectively.

Conclusions

Considering that GCS is easy to use and the FOUR can diagnose a locking syndrome along same values of subscales. These two subscales are superior to APACHI II in prediction of early mortality. Conversation APACHE II is more punctual in the prediction of delayed mortality.

Profile, outcome and predictors of mortality of abdomino-pelvic trauma patients in a tertiary intensive care unit in Saudi Arabia

INTRODUCTION

The Kingdom of Saudi Arabia (KSA) is one of countries with the world's highest number of deaths per 100,000 populations from road traffic accidents (RTAs). Numerous trauma victims sustain abdomino-pelvic injuries, which are associated with considerable morbidity and mortality. The purpose of this study was to describe profile, outcomes and predictors of mortality of patients with abdomino-pelvic trauma admitted to the intensive care unit (ICU) in a tertiary care trauma centre in Riyadh, KSA.

METHODS

This was a retrospective analysis of prospectively collected ICU database. All consecutive patients older than 14 years with abdomino-pelvic trauma from March 1999 to June 2013 were included. The followings were extracted: demographics, injury severity, mechanism and type of injury, associated injuries, use of vasopressors and mechanical ventilation, and worst laboratory results in the first 24h. The primary outcome was hospital mortality. We compared profile and outcomes between survivors and non-survivors and reported predictors of mortality.

RESULTS

Of the 11,374 trauma patients who were admitted to the hospital during the study period, 2120 (18.6%) patients had abdomino-pelvic injuries, out of which 702 (33.1%) patients were admitted to the ICU. The mean age was 30.7 (SD 14.4) years and the majority was male (89.5%). RTA was the most common cause of abdomino-pelvic trauma (70.4%). Pelvis (46.2%), liver (25.8%), and spleen (23.1%) were the most frequently injured organs; and chest (55.6%), head (41.9%), and lower extremities (27.5%) were the most commonly associated injuries. Mechanical ventilation was required in 89.6% with a mean duration of 9.1 (SD 9.2) days and emergency surgery was performed in 45.0% of the patients with prolonged ICU and hospital length of stay (10.8 [SD 10.8], 56.9 [SD 96.7] days; respectively). Of the 702 patients with abdomino-pelvic trauma, 115 (16.4%) patients did not survive. Associated head trauma and retroperitoneal haematoma, higher level of lactic acid on admission and ISS, and advanced age were potential risk factors for hospital mortality.

CONCLUSIONS

Abdomino-pelvic injuries are common in trauma patients, affecting mainly young male victims, and are associated with significant morbidity and mortality, and resource utilisation.

A Prospective Study of Comparison of APACHE-IV & SAPS-II Scoring Systems and Calculation of Standardised Mortality Rate in Severe Sepsis and Septic Shock Patients

CONTEXT

Severe sepsis and septic shock are major causes of mortality in the Intensive Care Unit (ICU) Illness Scoring Systems can help in the prediction of outcome of these patients.

AIM

To calculate and compare APACHE-IV and SAPS-II Scoring Systems along with calculation of Standardised Mortality Rate (SMR) in patients of severe sepsis and septic shock in the ICU.

STUDY DESIGN

Observational-analytical prospective study.

MATERIALS AND METHODS

The study was conducted on 84 patients with severe sepsis and septic shock admitted to the Medical ICU of a tertiary care teaching hospital.

RESULTS

Mean of Predicted Mortality Rate (PMR) for APACHE-IV was 37.85% and for SAPS-II, it was 72.36% which shows that APACHE-IV had under-predicted overall mortality while SAPS-II had over-predicted overall mortality of patients with severe sepsis and septic shock. Standardised Mortality Rate for APACHE-IV was 1.60 and for SAPS-II, it was 0.83.

CONCLUSION

Predicted Mortality of APACHE-IV and SAPS-II Scoring Systems did not correlate with the observed mortality for patients with severe sepsis and septic shock.

Assessment of performance and utility of mortality prediction models in a single Indian mixed tertiary intensive care unit

OBJECTIVES

To assess the performance and utility of two mortality prediction models viz. Acute Physiology and Chronic Health Evaluation II (APACHE II) and Simplified Acute Physiology Score II (SAPS II) in a single Indian mixed tertiary intensive care unit (ICU). Secondary objectives were bench-marking and setting a base line for research.

MATERIALS AND METHODS

In this observational cohort, data needed for calculation of both scores were prospectively collected for all consecutive admissions to 28-bedded ICU in the year 2011. After excluding readmissions, discharges within 24 h and age <18 years, the records of 1543 patients were analyzed using appropriate statistical methods.

RESULTS

Both models overpredicted mortality in this cohort [standardized mortality ratio (SMR) 0.88 ± 0.05 and 0.95 ± 0.06 using APACHE II and SAPS II respectively]. Patterns of predicted mortality had strong association with true mortality (R (2) = 0.98 for APACHE II and R (2) = 0.99 for SAPS II). Both models performed poorly in formal Hosmer-Lemeshow goodness-of-fit testing (Chi-square = 12.8 (P = 0.03) for APACHE II, Chi-square = 26.6 (P = 0.001) for SAPS II) but showed good discrimination (area under receiver operating characteristic curve 0.86 ± 0.013 SE (P < 0.001) and 0.83 ± 0.013 SE (P < 0.001) for APACHE II and SAPS II, respectively). There were wide variations in SMRs calculated for subgroups based on International Classification of Disease, 10(th) edition (standard deviation ± 0.27 for APACHE II and 0.30 for SAPS II).

INTERPRETATION AND CONCLUSION

Lack of fit of data to the models and wide variation in SMRs in subgroups put a limitation on utility of these models as tools for assessing quality of care and comparing performances of different units without customization. Considering comparable performance and simplicity of use, efforts should be made to adapt SAPS II.

Modification of Acute Physiology and Chronic Health Evaluation II score through recalibration of risk prediction model in critical care patients of a respiratory disease referral center

Background:

Several models have been developed to measure the severity of illness in intensive care unit (ICU) patients, It is suggested that the models should be customized depending on the characteristics of different population of patients. This study is aimed to assess and modify the performance of Acute Physiology and Chronic Health Evaluation II (APACHE-II) model in a respiratory diseases referral center.

Materials and Methods:

A total of 730 patients, admitted to an intensive care unit during one year, were divided into two sets (71% training and 29% test). Our modified APACHE-II model was developed and calibrated on training set. Then, the integrity of the customized model was checked and compared to the original APACHE-II, on the test set. Logistic regression was used to develop ROC analysis, F-measure and kappa coefficient and were employed to calibrate the model.

Results:

Both Original and Our modified APACHE-II scores performed acceptable discriminative power (AUC = 0.908: 95%CI 0.861-0.854; and AUC = 0.856: 95%CI 0.789-0.923, respectively); the difference was not significant (P = 0.132). Our modified APACHE-II showed improved accuracy (87.9% vs. 84.1%) and sensitivity (56.4% vs. 16.3%) compared to the original model. F-measure and Kappa also gave the impression of improvement for our modified APACHE-II system.

Conclusion:

The results demonstrated that a modified APACHE-II system in a local ICU of respiratory disease could have similar discrimination and comparable calibration to the original model.

Comparison of the predictive power of the LODS and APACHE II scoring systems in a neurological intensive care unit

OBJECTIVE

A prospective study to compare the power of the Logistic Organ Dysfunction System (LODS) and the Acute Physiology and Chronic Health Evaluation II (APACHE II) scoring systems to predict survival, in patients admitted to the neurological intensive care unit (NICU).

METHODS

Clinical data from 521 consecutive NICU patients were collected during the first 24 h of admission and were used to compare the predictive power of both scoring systems.

RESULTS

The observed mortality rate was 10.0% compared with predicted mortality rates of 7.2% and 4.8% according to LODS and APACHE II, respectively. Both scoring systems had excellent discrimination but LODS had superior calibration.

CONCLUSION

The LODS scoring system was more stable than the APACHE II scoring system in the NICU setting.

Etomidate and mortality in cirrhotic patients with septic shock

BACKGROUND

Clinical effects and outcomes of a single dose etomidate prior to intubation in the intensive care setting is controversial. The aim of this study is to evaluate the association of a single dose effect of etomidate prior to intubation on the mortality of septic cirrhotic patients and the impact of the subsequent use of low dose hydrocortisone.

METHODS

This is a nested-cohort study within a randomized double blind placebo controlled study evaluating the use of low dose hydrocortisone in cirrhotic septic patients. Cirrhotic septic patients ≥ 18 years were included in the study. Patients who received etomidate prior to intubation were compared to those who did not receive etomidate for all cause 28-day mortality as a primary outcome.

RESULTS

Sixty two intubated patients out of the 75 patients randomized in the initial trial were eligible for this study. Twenty three of the 62 intubated patients received etomidate dose prior to intubation. Etomidate use was not associated with all cause 28-day mortality or hospital mortality but was associated with significantly higher ICU mortality (91% vs. 64% for etomidate and controls groups, respectively; p = 0.02). Etomidate patients who received subsequent doses of hydrocortisone required lower doses of vasopressors and had more vasopressor-free days but no improvement in mortality.

CONCLUSIONS

In this group of septic cirrhotic patients with very high mortality, etomidate increased ICU mortality. Subsequent use of hydrocortisone appears to have no benefit beyond decreasing vasopressor requirements. The lowest mortality was observed in patients who did not receive etomidate but received hydrocortisone.

Low-dose hydrocortisone in patients with cirrhosis and septic shock: a randomized controlled trial

BACKGROUND

Recent studies have reported a high prevalence of relative adrenal insufficiency in patients with liver cirrhosis. However, the effect of corticosteroid replacement on mortality in this high-risk group remains unclear. We examined the effect of low-dose hydrocortisone in patients with cirrhosis who presented with septic shock.

METHODS

We enrolled patients with cirrhosis and septic shock aged 18 years or older in a randomized double-blind placebo-controlled trial. Relative adrenal insufficiency was defined as a serum cortisol increase of less than 250 nmol/L or 9 μg/dL from baseline after stimulation with 250 μg of intravenous corticotropin. Patients were assigned to receive 50 mg of intravenous hydrocortisone or placebo every six hours until hemodynamic stability was achieved, followed by steroid tapering over eight days. The primary outcome was 28-day all-cause mortality.

RESULTS

The trial was stopped for futility at interim analysis after 75 patients were enrolled. Relative adrenal insufficiency was diagnosed in 76% of patients. Compared with the placebo group (n = 36), patients in the hydrocortisone group (n = 39) had a significant reduction in vasopressor doses and higher rates of shock reversal (relative risk [RR] 1.58, 95% confidence interval [CI] 0.98-2.55, p = 0.05). Hydrocortisone use was not associated with a reduction in 28-day mortality (RR 1.17, 95% CI 0.92-1.49, p = 0.19) but was associated with an increase in shock relapse (RR 2.58, 95% CI 1.04-6.45, p = 0.03) and gastrointestinal bleeding (RR 3.00, 95% CI 1.08-8.36, p = 0.02).

INTERPRETATION

Relative adrenal insufficiency was very common in patients with cirrhosis presenting with septic shock. Despite initial favourable effects on hemodynamic parameters, hydrocortisone therapy did not reduce mortality and was associated with an increase in adverse effects. (Current Controlled Trials registry no. ISRCTN99675218.).

Developing severity adjusted quality measures for intensive care units

PURPOSE

Intensive care unit performance evaluation is usually affected by variations in the severity of inpatients' health status. This paper aims, therefore, to standardize two performance measures: intensive care unit survival and length of stay using the Acute Physiology and Chronic Health Evaluation II (APACHE II) severity of illness score.

DESIGN/METHODOLOGY/APPROACH

A records study in three Ain Shams University Hospital intensive care units, from January 1-December 31, 2003 was carried out to examine illness severity effect using APACHE II, length of stay and survival. Retrospective data were used to model length of stay in days and the survival using the APACHE II score as a predictor. This was followed by a prospective study to monitor the standardized measures in two intensive care units for one year.

FINDINGS

APACHE II scores predicted length of stay of those who were discharged and control charts for severity-adjusted length of stay were drawn up. The APACHE II score predicted survival for those with APACHE II score >16. The model is significant with a specificity of 89.9 percent while sensitivity was 25 percent. Control charts for severity-adjusted mortality were drawn up to monitor mortality.

RESEARCH LIMITATIONS/IMPLICATIONS

Only 60 percent of the files examined in the retrospective part of the study had enough data to calculate APACHE II scores.

PRACTICAL IMPLICATIONS

Standardized APACHE II severity of illness score can monitor intensive care unit length of stay and mortality.

ORIGINALITY/VALUE

The paper underlines the need to implement a standardized measurement system to evaluate intensive care patient outcomes.

The performance of customised APACHE II and SAPS II in predicting mortality of mixed critically ill patients in a Thai medical intensive care unit

The aim of this study was to evaluate and compare the performance of customised Acute Physiology and Chronic Health Evaluation HII (APACHE II) and Simplified Acute Physiology Score HII (SAPS II) in predicting hospital mortality of mixed critically ill Thai patients in a medical intensive care unit. A prospective cohort study was conducted over a four-year period. The subjects were randomly divided into calibration and validation groups. Logistic regression analysis was used for customisation. The performance of the scores was evaluated by the discrimination, calibration and overall fit in the overall group and across subgroups in the validation group. Two thousand and forty consecutive intensive care unit admissions during the study period were split into two groups. Both customised models showed excellent discrimination. The area under the receiver operating characteristic curve of the customised APACHE II was greater than the customised SAPS II (0.925 and 0.892, P < 0.001). Hosmer-Lemeshow goodness-of-fit showed good calibration for the customised APACHE II in overall populations and various subgroups but insufficient calibration for the customised SAPS II. The customised SAPS II showed good calibration in only the younger, postoperative and sepsis patients subgroups. The overall performance of the customised APACHE II was better than the customised SAPS II (Brier score 0.089 and 0.109, respectively). Our results indicate that the customised APACHE II shows better performance than the customised SAPS II in predicting hospital mortality and could be used to predict mortality and quality assessment in our unit or other intensive care units with a similar case mix.

The performance and customization of SAPS 3 admission score in a Thai medical intensive care unit

PURPOSE

The aim of this study was to evaluate the performance of Simplified Acute Physiology Score 3 (SAPS 3) admission scores, both the original and a customized version, in mixed medical critically ill patients.

METHODS

A prospective cohort study was conducted over a 2-year period in the medical intensive care unit (MICU) of a tertiary referral university teaching hospital in Thailand. The probability of hospital mortality of the original SAPS 3 was calculated using the general and customized Australasia version (SAPS 3-AUS). The patients were randomly divided into equal calibration and validation groups for customization.

RESULTS

A total of 1,873 patients were enrolled. The hospital mortality rate was 28.6%. The general equation of SAPS 3 had excellent discrimination with an area under the receiver operating characteristic curve of 0.933, but poor calibration with the Hosmer-Lemeshow goodness-of-fit H = 106.7 and C = 101.2 (P < 0.001), and it overestimated mortality with a standardized mortality ratio of 0.86 (95% confidence interval, 0.79-0.93). The calibration of SAPS 3-AUS was also poor. The customized SAPS 3 showed a good calibration of all patients in the validation group (H = 14, P = 0.17 and C = 11.3, P = 0.33) and all subgroups according to main diagnosis, age, gender and co-morbidities.

CONCLUSIONS

The SAPS 3 provided excellent discrimination but poor calibration in our MICU. A first level customization of the SAPS 3 improved the calibration and could be used to predict mortality and quality assessment in our ICU or other ICUs with a similar case mix.

Weekend and weeknight admissions have the same outcome of weekday admissions to an intensive care unit with onsite intensivist coverage

OBJECTIVE

Several reports have indicated increased mortality for weekend and nighttime admissions to the intensive care unit. This increase has been attributed to differences in staffing levels. The impact of onsite 24-hr/7-day intensivist staffing on weekend and weeknight outcomes has not been examined before. The objective of this study was to determine whether weekend and nighttime admissions compromise patient outcome in an intensive care unit staffed by an onsite intensivist 24 hrs a day and 7 days a week.

DESIGN

Cohort study.

SETTING

Tertiary care medical-surgical intensive care unit staffed 24 hrs/7 days by onsite consultant intensivists with predominantly North American Critical Care Board certifications.

PATIENTS

We included all emergency admissions over 4 yrs (March 1999 to February 2003) from a prospectively collected intensive care unit database. Admissions were grouped into weekday, weeknight, and weekend admissions.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Predicted mortality rates were calculated using Mortality Probability Models II0 and II24. The primary outcome was hospital mortality. Standardized mortality ratios were calculated. Secondary end points included intensive care unit mortality, duration of mechanical ventilation, intensive care unit length of stay, and the need for renal replacement therapy, tracheostomy, and pulmonary artery catheter during the intensive care unit course. A total of 2,093 admissions were included in the study, of which 31% were admitted on weekdays, 35% on weeknights, and 34% on weekends. The three groups were similar in baseline characteristics. There was no significant difference in hospital mortality rates among the three time periods (36%, 36%, and 37%, respectively, p=.90). There were also no significant differences in any of the secondary end points.

CONCLUSIONS

In an intensive care unit staffed by onsite certified intensivists 24 hrs/7 days, we found no compromise in the care of patients admitted during weekends and weeknights. These findings suggest that such coverage helps in ensuring consistency of care and therefore represents a potentially improved model for intensive care unit practice.

The comparison of the efficacy of scoring systems in organophosphate poisoning

The purpose of this study was to evaluate the impact of the Glasgow Coma Scale (GCS), Acute Physiology and Chronic Health Evaluation (APACHE) II and Simplified Acute Physiology Score (SAPS) II scoring systems for organophosphate poisoning (OPP) in an intensive care unit (ICU). The following data were collected on all consecutive patients who were admitted to the ICU between June 1999 and December 2004. Demographic data, GCS, APACHE II and SAPS II scoring systems were recorded. Predicted mortality was calculated using original regression formulas. Standardized mortality ratio (SMR) was computed with 95% confidence intervals (CI). The sensitivity and specificity for each scoring system were evaluated by calculating the Area Under the Receiver Operating Characteristic Curves. The actual mortality in OPP was 21.9%. Predicted mortality by all systems was not significantly different from actual mortality [SMR and 95% CI for GCS: 1.00 (0.65 1.35), APACHE II: 0.87 (0.54-1.03), SAPS II: 1.40 (0.98-1.82)]. The area under the ROC curve for APACHE II is largest, but there is no statistically significant difference when compared with SAPS II and GCS (GCS 0.900 +/- 0.059, APACHE II 0.929 +/- 0.045 and SAPS II 0.891 +/- 0.057). In our ICU group of patients, in predicting the mortality rates in OPP, the three scoring systems, which are GCS, APACHE II and SAPS II, had similar impacts; however, GCS system has superiority over the other systems in being easy to perform, and not requiring complex physiologic parameters and laboratory methods.

Prediction of mortality in an Indian intensive care unit. Comparison between APACHE II and artificial neural networks

OBJECTIVE

To compare hospital outcome prediction using an artificial neural network model, built on an Indian data set, with the APACHE II (Acute Physiology and Chronic Health Evaluation II) logistic regression model.

DESIGN

Analysis of a database containing prospectively collected data.

SETTING

Medical-neurological ICU of a university hospital in Mumbai, India.

SUBJECTS

Two thousand sixty-two consecutive admissions between 1996 and 1998.

INTERVENTIONS

None.

MEASUREMENTS AND RESULTS

The 22 variables used to obtain day-1 APACHE II score and risk of death were recorded. Data from 1,962 patients were used to train the neural network using a back-propagation algorithm. Data from the remaining 1,000 patients were used for testing this model and comparing it with APACHE II. There were 337 deaths in these 1,000 patients; APACHE II predicted 246 deaths while the neural network predicted 336 deaths. Calibration, assessed by the Hosmer-Lemeshow statistic, was better with the neural network (H=22.4) than with APACHE II (H=123.5) and so was discrimination (area under receiver operating characteristic curve =0.87 versus 0.77, p=0.002). Analysis of information gain due to each of the 22 variables revealed that the neural network could predict outcome using only 15 variables. A new model using these 15 variables predicted 335 deaths, had calibration (H=27.7) and discrimination (area under receiver operating characteristic curve =0.88) which was comparable to the 22-variable model (p=0.87) and superior to the APACHE II equation (p<0.001).

CONCLUSION

Artificial neural networks, trained on Indian patient data, used fewer variables and yet outperformed the APACHE II system in predicting hospital outcome.