SAPSII Score Research Articles

Left ventricular diastolic dysfunction is prevalent but not associated with mortality in patients with septic shock.

Prognostic impact of left ventricular diastolic dysfunction (LVDD) in septic shock patients has not been determined using current diagnostic guidelines. We assessed the relation between LVDD during the first 3days following intensive care unit (ICU) admission for septic shock and Day-28 mortality. This prospective, multicenter, observational study enrolled 402 patients (age: 63 ± 13year; 59% male; SAPS II: 59 ± 20; SOFA: 9.4 ± 3.6; mechanical ventilation: 74%) with septic shock (Sepsis-3 definition). Patients were echocardiographically assessed within 12h after admission (Day 1), on Day 2, Day 3, at ICU and at hospital discharge (or Day 28 whichever occurred first), using 2016 American-European guidelines. LVDD was present at least once between Day 1 and 3 in 304 patients (76%), and in 56% and 44% of patients at ICU discharge and on Day 28 (or hospital discharge), respectively (43% of patients with follow-up). Seventy-eight of 172 patients (45%) exhibited similar LV diastolic properties throughout the study period while 58 patients (34%) improved their LVDD at follow-up (lower grade: n = 9, regression: n = 49). Day-28 mortality was not statistically different between patients with and without LVDD (80/304 [26%] vs. 25/88 [28%]; OR: 0.900 [0.530-1.527]; p = 0.696). Similar results were obtained when adjusting the multivariate model on SAPSII or SOFA score on admission, together with fluid balance during the first three days of ICU stay (OR: 0.838 [0.471-1.491]: p = 0.547 and OR: 0.887 [0.513-1.534]: p = 0.668, respectively). LVDD was highly prevalent in patients with septic shock but not associated with mortality. It appeared improving in one-third of survivors. PRODIASYS study registered on ClinicalTrials (September 27, 2016, number NCT02918214).

Comparison of methods to normalize urine output in critically ill patients: a multicenter cohort study

BackgroundOliguria diagnosis includes the normalization of urine output (UO) by body weight. However, the rational and the method to apply to normalize UO to body weight are unclear. We aimed to explore the impact of the method applied to normalize UO on oliguria incidence and association with outcomes.MethodsWe included all adult patients admitted to a Swiss (derivation cohort) and a US (MIMIC-IV database, validation cohort) ICU, except those on maintenance hemodialysis, who declined consent or had < 6 consecutive UO measurements. Among a panel of candidate variables (ideal body weight, body mass index, body surface area and adjusted body weight), we identified the best predictor for UO (i.e. the variable that was most closely associated with mean UO during ICU stay). We then compared oliguria incidence and association with 90-day mortality and acute kidney disease (AKD) at hospital discharge, according to whether UO was normalized by actual body weight (ABW) or the identified best UO predictor.ResultsThe derivation and validation cohorts included respectively 15 322 and 28 610 patients. Those in the validation cohort were heavier (mean ABW 81 versus 75 kg) older (65 versus 62 years) and had a lower SAPS-II score (38 versus 43). The best UO predictor was ideal body weight (IBW). Oliguria incidence increased almost linearly across weight categories with ABW normalization but remained constant with IBW normalization. Using IBW for UO normalization rather than ABW improved the association between oliguria and 90-day mortality and AKD. It increased the proportion of patients correctly classified from 37.6 to 48.3% (mortality) and from 37.8 to 47% (AKD). All findings persisted after correction for sex and SAPS-II score and were confirmed in sensitivity analyses.ConclusionUO normalization by IBW lead to a stable incidence of oliguria across categories of weight and improved the association between oliguria and outcomes. IBW should be preferred to normalize UO in critically ill patients.

Explainable machine learning for early prediction of sepsis in traumatic brain injury: A discovery and validation study.

People with traumatic brain injury (TBI) are at high risk for infection and sepsis. The aim of the study was to develop and validate an explainable machine learning(ML) model based on clinical features for early prediction of the risk of sepsis in TBI patients. We enrolled all patients with TBI in the Medical Information Mart for Intensive Care IV database from 2008 to 2019. All patients were randomly divided into a training set (70%) and a test set (30%). The univariate and multivariate regression analyses were used for feature selection. Six ML methods were applied to develop the model. The predictive performance of different models were determined based on the area under the curve (AUC) and calibration curves in the test cohort. In addition, we selected the eICU Collaborative Research Database version 1.2 as the external validation dataset. Finally, we used the Shapley additive interpretation to account for the effects of features attributed to the model. Of the 1555 patients enrolled in the final cohort, 834 (53.6%) patients developed sepsis after TBI. Six variables were associated with concomitant sepsis and were used to develop ML models. Of the 6 models constructed, the Extreme Gradient Boosting (XGB) model achieved the best performance with an AUC of 0.807 and an accuracy of 74.5% in the internal validation cohort, and an AUC of 0.762 for the external validation. Feature importance analysis revealed that use mechanical ventilation, SAPSII score, use intravenous pressors, blood transfusion on admission, history of diabetes, and presence of post-stroke sequelae were the top six most influential features of the XGB model. As shown in the study, the ML model could be used to predict the occurrence of sepsis in patients with TBI in the intensive care unit.

Relationship between coagulopathy score and ICU mortality: Analysis of the MIMIC-IV database

ObjectiveCoagulopathy score has been applied as a new prognostic indicator for sepsis, heart failure and acute respiratory failure. However, its ability to forecast intensive care unit (ICU) mortality in patients with an acute cerebral hemorrhage (ICH) has not been assessed. The purpose of this study was to clarify the relationship between ICU mortality and early coagulation problem score. MethodsData from the Medical Information Mart for Intensive Care (MIMIC-IV) (v2.0) database were used in this retrospective cohort analysis. The association between the coagulation disorder score and ICU mortality was examined using multivariate logistic regression. Furthermore, the impact of additional variables on the results was investigated by a subgroup analysis. Results3174 patients (57.3 % male) were enrolled in total. The ICU mortality reached 18.2 %. After adjusting for potential confounders, the ICU mortality of patients rose with the increase of coagulation disorder score. The ROC curve revealed the predictive accuracy of coagulation dysfunction score to mortality in patients with ICU. The coagulation disorder score had a lower AUC value (0.601, P < 0.001) than the SAPSII(AUCs of 0.745[95 % CI, 0.730–0.761]) and the combined indicators(AUCs of 0.752[95 % CI, 0.737–0.767]), but larger than single indicators platelet, INR and APTT. In the subgroup analysis, most subgroups showed no significant interaction, but only age showed significant interaction in the adjusted model. ConclusionThe coagulopathy score and ICU mortality were found to be strongly positively correlated in this study, and its ability to predict ICU mortality was better than that of a single measure (platelet, INR, or APTT), but worse than that of the SAPSII score, GCS system.

Association between estimated plasma volume status and acute kidney injury in patients who underwent coronary revascularization: A retrospective cohort study from the MIMIC-IV database.

Acute kidney injury (AKI) remains a common complication of coronary revascularization and increases poor outcomes in critically ill surgical patients. Compared to the plasma volume status (PVS), estimated plasma volume status (ePVS) has the advantages of being noninvasive and simple and has been shown to be associated with worse prognosis in patients undergoing coronary revascularization. This study was to evaluate the association of ePVS with the risk of AKI in patients who underwent coronary revascularization. In this retrospective cohort study, data of patients who underwent coronary revascularization were extracted from the Medical Information Mart for Intensive Care (MIMIC)-IV database (2008-2019). The outcome was the occurrence of AKI after ICU admission. The covariates were screened via the LASSO regression method. Univariate and multivariate Logistic regression models were performed to assess the association of ePVS and PVS and the odds of AKI in patients who underwent coronary revascularization, with results shown as odds ratios (ORs) and 95% confidence intervals (CIs). Subgroup analyses of age, surgery, and anticoagulation agents and sequential organ failure assessment (SOFA) score were performed to further explore the association of ePVS with AKI. A total of 3,961 patients who underwent coronary revascularization were included in this study, of whom 2,863 (72.28%) had AKI. The high ePVS was associated with the higher odds of AKI in patients who received coronary revascularization (OR = 1.06, 95%CI: 1.02-1.10), after adjusting for the covariates such as age, race, SAPS-II score, SOFA score, CCI, weight, heart rate, WBC, RDW-CV, PT, BUN, glucose, calcium, PH, PaO2, mechanical ventilation, vasopressors, and diuretic. Similar results were found in patients who underwent the CABG (OR = 1.07, 95%CI: 1.02-1.11), without anticoagulation agents use (OR = 1.07, 95%CI: 1.03-1.12) and with high SOFA score (OR = 1.10, 95%CI: 1.04-1.17). No relationship was found between PVS and the odds of AKI in patients who underwent the coronary revascularization. The ePVS may be a promising parameter to evaluate the risk of AKI in patients undergoing coronary revascularization, which provides a certain reference for the risk stratification management of ICU patients who underwent coronary revascularization.

#236 Immediate or delayed initiation of renal replacement therapy in patients with leptospirosis and acute kidney injury: a target trial emulation

Abstract Background and Aims Leptospirosis is a widespread zoonosis responsible for 60,000 deaths annually worldwide [1]. Acute kidney injury (AKI) stands out as a hallmark of the disease. Leptospirosis-induced AKI (AKI-L) has been entertained as a potential cause of chronic kidney injury of unknown etiology, which is a significant burden in tropical rural settings [2]. Anecdotal evidence suggests that early renal replacement therapy (RRT) may improve mortality associated with AKI-L [3]. Conversely, several randomized controlled trials (RCTs) conducted in intensive care units have refuted the positive impact of early RRT on mortality in patients with other causes of AKI and sepsis [4]. Given these contradicting results, the optimal timing for initiating RRT in patients with leptospirosis and severe AKI remains uncertain. In this emulated RCT, we aimed to evaluate the impact of the timing of RRT on mortality and worsening renal failure (WRF) in patients hospitalized with leptospirosis, stage 3 AKI on admission, and no immediate need for RRT. Method We conducted a retrospective cohort study in two academic centers in Réunion island between 2010 and 2020. During this period, we included adult patients with documented leptospirosis and stage 2 or 3 AKI. Then we selected patients with stage 3 AKI to perform a target trial emulation, adapted from the AKIKI trial's design. Patients were categorized according to the allocated treatment strategy. Patients assigned to the “early” group received RRT within 48 hours following hospital admission while patients in the “delayed” group did not. A propensity score was derived for each imputed dataset using multivariate logistic regression. The propensity score estimates the probability of treatment allocation as a function of covariates assessed upon hospital entry. These covariates included those used in the SAPSII score and patients' baseline renal function. The primary outcome was a composite outcome assessed at one year: death or WRF. Results The cohort included 380 patients. We selected 295 patients with stage 3 AKI to perform the emulated RCT. Eight two (28%) patients were started on RRT within 48 hours of admission and were assigned to the early group. The delayed group included the remaining 213 patients, with 25% undergoing subsequent RRT and 75% not requiring such intervention. The median time to RRT initiation was 2 (1-3) days. Fifty-nine patients (20%) met the primary outcome: 14 patients (4.7%) died during their hospital stay or within 1 year after discharge, and 45 (15.3%) developed WRF. In the main analysis, the adjusted odds ratio (OR) for primary outcome occurrence was of 2.12 (95% confidence interval: 1.03 to 4.37, p = 0.04; reference group: “delayed” strategy). In secondary analyses, there was a significantly higher probability of WRF at one year in patients assigned to the “early strategy” (OR = 2.79 [1.28-6.11], p = 0.010). Mortality at one year according to RRT timing strategy did not differ (p = 0.6). Conclusion Early initiation of RRT was not found to improve the combined outcome of death and development of WRF in patients with leptospirosis and stage 3 AKI. A strategy based on a delayed initiation of RRT should be formally recommended for patients with severe AKI-L as in the general population of critically ill patients.

Features and outcomes of patients admitted to the ICU for chimeric antigen receptor T cell-related toxicity: a French multicentre cohort

BackgroundChimeric antigen receptor T-cell (CAR-T) therapy is increasingly used in patients with refractory haematological malignancies but can induce severe adverse events. We aimed to describe the clinical features and outcomes of patients admitted to the intensive care unit (ICU) after CAR-T therapy.MethodsThis retrospective observational cohort study included consecutive adults admitted to either of two French ICUs in 2018–2022 within 3 months after CAR-T therapy.ResultsAmong 238 patients given CAR-T therapy, 84 (35.3%) required ICU admission and were included in the study, a median of 5 [0–7] days after CAR-T infusion. Median SOFA and SAPSII scores were 3 [2–6] and 39 [30–48], respectively. Criteria for cytokine release syndrome were met in 80/84 (95.2%) patients, including 18/80 (22.5%) with grade 3–4 toxicity. Immune effector cell-associated neurotoxicity syndrome (ICANS) occurred in 46/84 (54.8%) patients, including 29/46 (63%) with grade 3–4 toxicity. Haemophagocytic lymphohistiocytosis was diagnosed in 15/84 (17.9%) patients. Tocilizumab was used in 73/84 (86.9%) patients, with a median of 2 [1–4] doses. Steroids were given to 55/84 (65.5%) patients, including 21/55 (38.2%) given high-dose pulse therapy. Overall, 23/84 (27.4%) patients had bacterial infections, 3/84 (3.6%) had fungal infections (1 invasive pulmonary aspergillosis and 2 Mucorales), and 2 (2.4%) had cytomegalovirus infection. Vasopressors were required in 23/84 (27.4%), invasive mechanical ventilation in 12/84 (14.3%), and dialysis in 4/84 (4.8%) patients. Four patients died in the ICU (including 2 after ICU readmission, i.e., overall mortality was 4.8% of patients). One year after CAR-T therapy, 41/84 (48.9%) patients were alive and in complete remission, 14/84 (16.7%) were alive and in relapse, and 29/84 (34.5%) had died. These outcomes were similar to those of patients never admitted to the ICU.ConclusionICU admission is common after CAR-T therapy and is usually performed to manage specific toxicities. Our experience is encouraging, with low ICU mortality despite a high rate of grade 3–4 toxicities, and half of patients being alive and in complete remission at one year.

Epidemiology and outcomes of early versus late septic acute kidney injury in critically ill patients: A retrospective cohort study

BackgroundIt was recently proposed to distinguish early from late sepsis-associated acute kidney injury (SA-AKI). We aimed to determine the relative frequency of these entities in critically ill patients and to describe their characteristics and outcomes. MethodsWe included in this retrospective cohort study all adult patients admitted for sepsis in a tertiary ICU between 2010 and 2020. We excluded those on chronic dialysis or without consent. We extracted serum creatinine, hourly urinary output, and clinical and socio-demographic data from medical records until day 7 or ICU discharge. AKI presence and characteristics were assessed daily using KDIGO criteria. We compared patients with early (occurring within 2 days of admission) or late (occurring between day 2 and day 7) SA-AKI. We conducted sensitivity analyses using different definitions for early/late SA-AKI. ResultsAmong 1835 patients, 1660 (90%) fulfilled SA-AKI criteria. Of those, 1610 (97%) had early SA-AKI, and 50 (3%) had late SA-AKI. Similar proportions were observed when only considering AKI with elevated sCr (71% vs. 3%), severe AKI (67% vs. 6%), or different time windows for early SA-AKI. Compared with early SA-AKI patients, those with late SA-AKI were younger (median age [IQR] 59 [49–70] vs. 69 [58–76] years, p < 0.001), had lower Charlson comorbidity index (3 [1–5] vs. 5 [3–7], p < 0.001) and lower SAPSII scores (41 [34–50] vs. 53 [43–64], p < 0.001). They had similar (24% vs. 26%, p = 0.75) in-hospital mortality. ConclusionsAKI is almost ubiquitous in septic critically ill patients and present within two days of admission. The timing from ICU admission might not be relevant to distinguish different phenotypes of SA-AKI. Ethics approvalEthics Committee Vaud, Lausanne, Switzerland (n°2017-00008).

Admission Heart Rate and Mortality in Critically Ill Patients with Acute Aortic Dissection.

The association between admission heart rate (HR) and the mortality of critically ill patients with acute aortic dissection (AAD) remains unclear.The data were extracted from the Medical Information Mart for Intensive Care (MIMIC-III) database. Cox regression models and Kaplan-Meier (KM) survival curve were used to explore the association between admission HR and 90-day, 1-year, and 3-year mortality in patients with AAD. Sensitivity analyses were conducted to assess potential bias.A total of 374 eligible AAD patients were included and divided in 4 groups according to admission HR (HR ≤ 70, 71-80, 81-90, and > 90 beats per minute (bpm) ). The patients with AAD in the group with HR > 90 bpm had higher 90-day, 1-year, and 3-year mortality than those in the groups with HR ≤ 70, 71-80, and 81-90 bpm. After adjusting for age, sex, BMI, systolic blood pressure, diastolic blood pressure, SOFA score, SAPSII score, Stanford type, hypertension, coronary artery disease, liver disease, atrial fibrillation, valvular disease, intensive care unit mechanical ventilation, aortic surgery, and thoracic endovascular aortic repair, patients with admission HR > 90 bpm had a higher risk of 90-day, 1-year, and 3-year mortality [adjusted hazard ratio, 95% confidence interval, 5.14 (2.22-11.91) P < 0.001; 4.31 (2.10-8.84) P < 0.001; 3.01 (1.66-5.46) P < 0.001] than those with HR 81-90 bpm. The 90-day, 1-year, and 3-year mortality were similar among the groups with HR ≤ 70, 71-80, and 81-90 bpm.Admission HR > 90 bpm was independently associated with all-cause mortality in critically ill AAD patients, either type A or B aortic dissection.

Development of a nomogram for predicting 90-day mortality in patients with sepsis-associated liver injury

The high mortality rate in sepsis patients is related to sepsis-associated liver injury (SALI). We sought to develop an accurate forecasting nomogram to estimate individual 90-day mortality in SALI patients. Data from 34,329 patients were extracted from the public Medical Information Mart for Intensive Care (MIMIC-IV) database. SALI was defined by total bilirubin (TBIL) > 2 mg/dL and the occurrence of an international normalized ratio (INR) > 1.5 in the presence of sepsis. Logistic regression analysis was performed to establish a prediction model called the nomogram based on the training set (n = 727), which was subsequently subjected to internal validation. Multivariate logistic regression analysis showed that SALI was an independent risk factor for mortality in patients with sepsis. The Kaplan‒Meier curves for 90-day survival were different between the SALI and non-SALI groups after propensity score matching (PSM) (log rank: P < 0.001 versus P = 0.038), regardless of PSM balance. The nomogram demonstrated better discrimination than the sequential organ failure assessment (SOFA) score, logistic organ dysfunction system (LODS) score, simplified acute physiology II (SAPS II) score, and Albumin–Bilirubin (ALBI) score in the training and validation sets, with areas under the receiver operating characteristic curve (AUROC) of 0.778 (95% CI 0.730–0.799, P < 0.001) and 0.804 (95% CI 0.713–0.820, P < 0.001), respectively. The calibration plot showed that the nomogram was sufficiently successful to predict the probability of 90-day mortality in both groups. The DCA of the nomogram demonstrated a higher net benefit regarding clinical usefulness than SOFA, LODS, SAPSII, and ALBI scores in the two groups. The nomogram performs exceptionally well in predicting the 90-day mortality rate in SALI patients, which can be used to assess the prognosis of patients with SALI and may assist in guiding clinical practice to enhance patient outcomes.

Predictive factors and prognostic value of candiduria in critically-ill patients with solid and hematological malignancies.

To investigate the epidemiology of candiduria in critically-ill patients with solid/hematological malignancies and to assess its predictive factors and prognostic value. All adult patients with confirmed solid/hematological malignancy admitted in the intensive care units (ICUs) for more than 48h were retrospectively included. Urine cultures were sampled on admission and then whenever signs of sepsis were identified. Two groups were compared: (candiduria (+)) and (candiduria (-)). One-hundred-seventy-three patients were included. Solid cancer was the underlying oncological disease for 147 patients (85%) while 26 patients (15%) had hematological malignancies. Twenty-nine patients (16.8%) were diagnosed with candiduria, and 31 urinary samples grew Candida spp. Candida spp represented 55.8% of the total urinary isolates. Fourteen isolates (45.2%) of Candida albicans were identified. Among the 17 non-albicans isolates, Candida tropicalis was the most predominant (41.9%). Six patients (3.5%) had candidemia with no significant difference between candiduria(+) and candiduria(-) groups (respectively, 6.9% and 2.8%; p=0.264). In multivariate analysis, previous exposure to quinolones (OR=3.8, CI95% [1.4-8.3]; p=0.008), mechanical ventilation (OR=4.1, CI95% [1.1-14.7]; p=0.034) and renal replacement therapy (OR=3.5, 95%CI [1.2-9.7]; p=0.017) were identified as independent factors predicting candiduria. Candiduria was associated with significantly higher ICU-mortality after adjusting for SAPSII score on admission (OR=2.9 CI95% [1.3-6.8]; p=0.009). Candiduria is common in cancer critically-ill patients. We reported an increased rate of non-albicans species, over albicans species. Patients with candiduria had higher ICU mortality, probably related to higher frailty and clinical severity.

Developing an Interpretable Machine Learning Model to Predict in-Hospital Mortality in Sepsis Patients: A Retrospective Temporal Validation Study.

Risk stratification plays an essential role in the decision making for sepsis management, as existing approaches can hardly satisfy the need to assess this heterogeneous population. We aimed to develop and validate a machine learning model to predict in-hospital mortality in critically ill patients with sepsis. Adult patients fulfilling the definition of Sepsis-3 were included at a large tertiary medical center. Relevant clinical features were extracted within the first 24 h in ICU, re-classified into different genres, and utilized for model development under three strategies: "Basic + Lab", "Basic + Intervention", and "Whole" feature sets. Extreme gradient boosting (XGBoost) was compared with logistic regression (LR) and established severity scores. Temporal validation was conducted using admissions from 2017 to 2019. The final cohort included 24,272 patients, of which 4013 patients formed the test cohort for temporal validation. The trained and fine-tuned XGBoost model with the whole feature set showed the best discriminatory ability in the test cohort with AUROC as 0.85, significantly higher than the XGBoost "Basic + Lab" model (0.83), the LR "Whole" model (0.82), SOFA (0.63), SAPS-II (0.73), and LODS score (0.74). The performance in varying subgroups remained robust, and predictors, such as increased urine output and supplemental oxygen therapy, were crucially correlated with improved survival when interpretability was explored. We developed and validated a novel XGBoost-based model and demonstrated significantly improved performance to LR and other scores in predicting the mortality risks of sepsis patients in the hospital using features in the first 24 h.

Agreement between pulse oximetry and arterial oxygen saturation measurement in critical care patients during COVID-19: a cross-sectional study

Some publications suggest that pulse oximetry measurement (SpO2) might overestimate arterial oxygen saturation (SaO2) measurement in COVID-19 patients. This study aims to evaluate the agreement between SpO2 and SaO2 among COVID-19 and non-COVID-19 patients. We conducted a multicenter, prospective study including consecutive intensive care patients from October 15, 2020, to March 4, 2021, and compared for each measurement the difference between SpO2 and SaO2, also called the systematic bias. The primary endpoint was the agreement between SpO2 and SaO2 measured with the Lin concordance coefficient and illustrated using the Bland and Altman method. Factors associated with systematic bias were then identified using a generalised estimating equation. The study included 105 patients, 66 COVID-19 positive and 39 COVID-19 negative, allowing for 1539 measurements. The median age was 66 [57; 72] years with median SOFA and SAPSII scores of, respectively, 4 [3; 6] and 37 [31; 47]. The median SpO2 and SaO2 among all measurements was respectively 97 [96–99] and 94 [92–96] with a systematic bias of 0.80 [− 0.6; 2.4]. This difference was, respectively, 0.80 [− 0.7; 2.5] and 0.90 [− 0.3; 2.0] among COVID-19 positive and negative patients. Overall agreement measured with the Lin correlation coefficient was 0.65 [0.63; 0.68] with 0.61 [0.57; 0.64] and 0.53 [0.45; 0.60] among the COVID-19 positive and negative groups, respectively. Factors independently associated with the variation of the SpO2-SaO2 difference were the PaO2/FiO2 ratio and need for mechanical ventilation. In our population, agreement between SpO2 and SaO2 is acceptable. During the COVID-19 pandemic, SaO2 remains an efficient monitoring tool to characterise the level of hypoxemia and follow therapeutic interventions. As is already known about general intensive care unit patients, the greater hypoxemia, the weaker the correlation between SpO2 and SaO2.

Stress-induced Hyperglycemia Ratio as an Independent Risk Factor of In-hospital Mortality in Nonresuscitation Intensive Care Units: A Retrospective Study

PurposeTo determine whether the stress-induced hyperglycemia ratio (SHR) is independently associated with in-hospital mortality in critically ill patients in nonresuscitation ICUs. MethodsIn this retrospective cohort study, clinical- and laboratory-related data from patients first admitted to nonresuscitation ICUs were extracted from an open-access database of >50,000 ICU admissions. Patients were assigned to one of two groups according to an SHR threshold of 1.1. The primary end point of this study was the in-hospital mortality rate. The associations between SHR and length of stay in the ICU and hospital, duration of mechanical ventilation use, and vasopressor use were secondary end points. Logistic regression models were established in the analysis of in-hospital mortality risk, and areas under the receiver operating characteristic curve (AUC) were analyzed to investigate the association between the primary end point and SHR used alone or together with the Simplified Acute Physiology Scale (SAPS) II score. The Youden index, specificity, and sensitivity of SHR and SAPS-II were also assessed. FindingsIn this study, 1859 patients were included, 187 of whom (10.06%) died during hospitalization. The group with an SHR of ≥1.1 had a greater in-hospital mortality rate (13.7% vs 7.4%; P < 0.001), longer length of stay both in the ICU and in the hospital, a longer duration of mechanical ventilation use, and a greater rate of vasopressor use. On adjustment for multivariate risk, a 0.1-point increment in SHR was significantly associated with in-hospital mortality (OR = 1.08; 95% CI, 1.00–1.16; P = 0.036). The AUC of the association between risk and the SAPS-II score was significantly greater than that with SHR (0.797 [95% CI, 0.576–0.664] vs 0.620 [95% CI, 0.764–0.830]; P < 0.001). The AUC with SAPS-II + SHR was significantly greater than that with SAPS-II used alone (0.802 [95% CI, 0.770–0.835] vs 0.797 [95% CI, 0.764-0.830]; P = 0.023). The Youden index, specificity, and sensitivity of SAPS-II + SHR were 0.473, 0.703, and 0.770, respectively. ImplicationsStress-induced hyperglycemia, as evaluated using the SHR, was associated with increased in-hospital mortality and worse clinical outcomes in these critically ill patients in nonresuscitation ICUs. SHR was an independent risk factor for in-hospital mortality, and when used together with the SAPS-II, added to the capacity to predict mortality in these patients in nonresuscitation ICUs. Prospective data are needed to validate the capacity of SHR in predicting in-hospital mortality in patients in the nonresuscitation ICU.

Prediction and risk assessment of sepsis-associated encephalopathy in ICU based on interpretable machine learning

Sepsis-associated encephalopathy (SAE) is a major complication of sepsis and is associated with high mortality and poor long-term prognosis. The purpose of this study is to develop interpretable machine learning models to predict the occurrence of SAE after ICU admission and implement the individual prediction and analysis. Patients with sepsis admitted to ICU were included. SAE was diagnosed as glasgow coma score (GCS) less than 15. Statistical analysis at baseline was performed between SAE and non-SAE. Six machine learning classifiers were employed to predict the occurrence of SAE, and the adjustment of model super parameters was performed by using Bayesian optimization method. Finally, the optimal algorithm was selected according to the prediction efficiency. In addition, professional physicians were invited to evaluate our model prediction results for further quantitative assessment of the model interpretability. The preliminary analysis of variance showed significant differences in the incidence of SAE among patients with pathogen infection. There were significant differences in physical indicators like respiratory rate, temperature, SpO2 and mean arterial pressure (P < 0.001). In addition, the laboratory results were also significantly different. The optimal classification model (XGBoost) indicated that the best risk factors (cut-off points) were creatinine (1.1 mg/dl), mean respiratory rate (18), pH (7.38), age (72), chlorine (101 mmol/L), sodium (138.5 k/ul), SAPSII score (23), platelet count (160), and phosphorus (2.4 and 5.0 mg/dL). The ranked features derived from the best model (AUC is 0.8837) were mechanical ventilation, duration of mechanical ventilation, phosphorus, SOFA score, and vasopressin usage. The SAE risk prediction model based on XGBoost created here can make very accurate predictions using simple indicators and support the visual explanation. The interpretable model was effectively evaluated by professional physicians and can help them predict the occurrence of SAE more intuitively.

A Regional Observational Study on COVID-19-Associated Pulmonary Aspergillosis (CAPA) within Intensive Care Unit: Trying to Break the Mold.

The reported incidence of COVID-19-associated pulmonary aspergillosis (CAPA) ranges between 2.4% and 35% in intensive care unit (ICU) patients, and awareness in the medical community is rising. We performed a regional retrospective observational study including patients diagnosed with CAPA defined according to the Modified AspICU Dutch/Belgian Mycosis Study Group and CAPA-EECMM, from five different ICUs, admitted between March, 2020 and September, 2021. Forty-five patients were included. The median age was 64 (IQR 60-72), mostly (73%) males. At ICU admission, the median Charlson comorbidity index was 3 (2-5), and the simplified acute physiology score (SAPS)-II score was 42 (31-56). The main underlying diseases were hypertension (46%), diabetes (36%) and pulmonary diseases (15%). CAPA was diagnosed within a median of 17 days (IQR 10-21.75) after symptoms onset and 9 days (IQR 3-11) after ICU admission. The overall 28-day mortality rate was 58%, and at univariate analysis, it was significantly associated with older age (p = 0.009) and SAPS-II score at admission (p = 0.032). The use of immunomodulatory agents, p = 0.061; broad-spectrum antibiotics, p = 0.091; positive culture for Aspergillus on BAL, p = 0.065; and hypertension, p = 0.083, were near reaching statistical significance. None of them were confirmed in multivariate analysis. In critically ill COVID-19 patients, CAPA acquired clinical relevance in terms of incidence and reported mortality. However, the risk between underdiagnosis-in the absence of specific invasive investigations, and with a consequent possible increase in mortality-and over-diagnosis (case identification with galactomannan on broncho-alveolar fluid alone) might be considered. Realistic incidence rates, based on local, real-life epidemiological data, might be helpful in guiding clinicians.

Methylene blue dosing strategies in critically ill adults with shock-A retrospective cohort study.

Shock increases mortality in the critically ill and the mainstay of therapy is the administration of vasopressor agents to achieve hemodynamic targets. In the past, studies have found that the NO-pathway antagonist methylene blue improves hemodynamics. However, the optimal dosing strategy remains elusive. Therefore, we investigated the hemodynamic and ICU outcome parameters of three different dosing strategies for methylene blue. We performed a retrospective cohort study of patients in shock treated with methylene blue. Shock was defined as norepinephrine dose >0.1 μg/kg/min and serum lactate level >2 mmol/l at the start of methylene blue administration. Different demographic variables, ICU treatment, and outcome parameters were evaluated. To compare the differences in the administration of vasopressors or inotropes, the vasoactive inotropic score (VIS) was calculated at different time points after starting the administration of methylene blue. Response to methylene blue or mortality at 28 days were assessed. 262 patients from July 2014 to October 2019 received methylene blue. 209 patients met the inclusion criteria. Three different dosing strategies were identified: bolus injection followed by continuous infusion (n = 111), bolus injection only (no continuous infusion; n = 59) or continuous infusion only (no bolus prior; n = 39). The groups did not differ in demographics, ICU scoring system, or comorbidities. In all groups, VIS decreased over time, indicating improved hemodynamics. Cardiogenic shock and higher doses of norepinephrine increased the chance of responding to methylene blue, while bolus only decreased the chance of responding to methylene blue treatment. 28-day mortality increased with higher SAPSII scores and higher serum lactate levels, while bolus injection followed by continuous infusion decreased 28-day mortality. No severe side effects were noted. In this cohort, methylene blue as a bolus injection followed by continuous infusion was associated with a reduced 28-day mortality in patients with shock. Prospective studies are needed to systematically evaluate the role of methylene blue in the treatment of shock.

Association between type II diabetes mellitus and 90-day mortality in a large multicenter prospectively collected cohort. A FROG ICU post-hoc study

PurposeFactors associated with adverse outcomes in ICU patients with type II (T2DM) are poorly defined. The main goal of this study is to determine the impact of pre-existing T2DM on 90-day mortality post ICU admission. MaterialPost-hoc analysis from the FROG-ICU cohort. All patients admitted to ICU who were ventilated and/or treated by a vasoactive agent for >24 h were included.Association between T2DM and 90-day mortality was analyzed in unmatched, and populations matched by propensity score (PS) method to balance confounders recorded before ICU admission. Analysis was performed in non-imputed and imputed datasets. Results2002 patients were included, and 16% had a history of T2DM. The latter were at inclusion more severely ill (SAPSII score 51(39–67) vs 48(35–61), p < 0.0001; Charlson score 2(1–3) vs 0(0–2), p < 0.0001). In the unmatched cohort, T2DM patients had a higher 90-day risk of death compared to no-DM patients (HR 1.35(1.1–1.65)). The 90-day risk of death was not significantly different T2DM and no T2DM patients after PS matching (HR: 0.81 (0.56–1.18). Results were similar with the analysis performed on imputed datasets (pooled HR: 0.95 (0.69–1.30)). ConclusionsIn the present study, T2DM was not associated with 90-day mortality post ICU admission.

Patient mobilisation in the intensive care unit and evaluation of a multifaceted intervention including Facebook groups: A quasi-experimental study

AimsTo describe prevalence and time to mobilisation in intensive care unit patients defined as a minimum sitting in an upright position in bed, and evaluate the impact of a multifaceted quality improvement campaign on likelihood of patients being mobilised. Research methodology/designQuality improvement project using a quasi-experimental study design, comparing patient cohorts before (Before) and after (Intervention) a campaign including educational sessions, audit and feedback of intensive care unit quality indicators via closed Facebook-groups and e-mail and local opinion leaders. Secondary analysis of mobilisation data from adult intensive care patient stays extracted from electronical medical charts. Likelihood of being mobilised was analysed with Multivariate Cox-regression model and reported as Sub-hazard Ratio (SHR). SettingFour intensive care units in a university hospital. Main outcome measuresPrevalence and time to first documented mobilisation, defined as at least “sitting in bed” during the intensive care unit stay. ResultsOverall, 929 patients were analysed, of whom 710 (76 %) were mobilised; 73 % (356/ 489) in Before vs 81 % (354/ 440) in Intervention (p = 0.007). Median time to mobilisation was 69.9 (IQR: 30.0, 149.8) hours; 71.7 (33.9, 157.9) in Before and 66.0 (27.1, 140.3) in Intervention (p = 0.104). Higher SAPS II-scores were associated with lower likelihood (SHR 0.98, 95 % CI 0.97–0.99), whereas admissions due to gastroenterological failure (SHR 2.1, 95 % CI 1.4–3.0), neurological failure (SHR 1.5, 95 % CI 1.0–2.2) and other causes (intoxication, postoperative care, haematological-, and kidney failure) (SHR 1.7, 95 % CI 1.13–2.6) were associated with higher likelihood of mobilisation vs respiratory failure. ConclusionA quality improvement campaign including use of Facebook groups is feasible and may improve mobilisation in intensive care unit patients. Most patients were mobilised within 72 hours following intensive care unit admission, and SAPS II scores and causes for intensive care unit admission were both associated with likelihood of being mobilised.

Red blood cell distribution width as prognostic factor in sepsis: A new use for a classical parameter.

To evaluate Red blood cell distribution width (RDW) as a sepsis prognostic biomarker. 203 septic patients admitted to the ICU. Analysis of RDW dynamics, hospital mortality discrimination ability and the added value when incorporated to the SOFA, LODS, SAPS-II and APACHE-II scores using the AUC-ROC. Non-survivors presented higher RDW values during the first week after ICU admission (p = 0.048). Only SOFA and RDW were independently associated with mortality when adjusted by Charlson, immunosuppression, nosocomial infection, NEWS2, SAPS-II, septic shock and haemoglobin (p < 0.05). After adjustment, AUC-ROC was 0.827, 0.822, 0.824, 0.834 and 0.812 for each model including admission, 24, 48 and 72-h and 7-days RDW, respectively. When added to the scores, 24-h RDW and admission RDW improved their discrimination ability (SOFA AUC-ROC = 0.772 vs 0.812 SOFA + admission RDW, p = 0.041; LODS AUC-ROC = 0.687 vs 0.710, p = 0.002; SAPS-II AUC-ROC = 0.734 vs 0.785, p = 0.021; APACHE-II AUC-ROC = 0.672 vs 0.755, p = 0.003). Admission RDW with SOFA presented the better discrimination ability for mortality. RDW is an independent prognostic marker of death in septic patients admitted in the ICU that improves SOFA, LODS, APACHE-II and SAPS-II discrimination ability. This parameter could be incorporated to the prognostic scores as a marker of systemic dysfunction and dysregulated inflammatory response.