Predicting 30-day mortality in intensive care unit patients with ischaemic stroke or intracerebral haemorrhage.

BACKGROUND: Stroke patients admitted to an intensive care unit (ICU) follow a particular survival pattern with a high short-term mortality, but if they survive the first 30 days, a relatively favourable subsequent survival is observed. OBJECTIVES: The development and validation of two prognostic models predicting 30-day mortality for ICU patients with ischaemic stroke and for ICU patients with intracerebral haemorrhage (ICH), analysed separately, based on parameters readily available within 24 h after ICU admission, and with comparison with the existing Acute Physiology and Chronic Health Evaluation IV (APACHE-IV) model. DESIGN: Observational cohort study. SETTING: All 85 ICUs participating in the Dutch National Intensive Care Evaluation database. PATIENTS: All adult patients with ischaemic stroke or ICH admitted to these ICUs between 2010 and 2019. MAIN OUTCOME MEASURES: Models were developed using logistic regressions and compared with the existing APACHE-IV model. Predictive performance was assessed using ROC curves, calibration plots and Brier scores. RESULTS: We enrolled 14 303 patients with stroke admitted to ICU: 8422 with ischaemic stroke and 5881 with ICH. Thirty-day mortality was 27% in patients with ischaemic stroke and 41% in patients with ICH. Important factors predicting 30-day mortality in both ischaemic stroke and ICH were age, lowest Glasgow Coma Scale (GCS) score in the first 24 h, acute physiological disturbance (measured using the Acute Physiology Score) and the application of mechanical ventilation. Both prognostic models showed high discrimination with an AUC 0.85 [95% confidence interval (CI), 0.84 to 0.87] for patients with ischaemic stroke and 0.85 (0.83 to 0.86) in ICH. Calibration plots and Brier scores indicated an overall good fit and good predictive performance. The APACHE-IV model predicting 30-day mortality showed similar performance with an AUC of 0.86 (95% CI, 0.85 to 0.87) in ischaemic stroke and 0.87 (0.86 to 0.89) in ICH. CONCLUSION: We developed and validated two prognostic models for patients with ischaemic stroke and ICH separately with a high discrimination and good calibration to predict 30-day mortality within 24 h after ICU admission. TRIAL REGISTRATION: Trial registration: Dutch Trial Registry ( https://www.trialregister.nl/ ); identifier: NTR7438.

Restrictive versus conventional ward fluid therapy in non-cardiac surgery patients and the effect on postoperative complications: a meta-analysis.

BACKGROUND: Diligent fluid management is an instrumental part of Enhanced Recovery After Surgery. However, the effect of a ward regimen to limit intravenous fluid administration on outcome remains unclear. We performed a meta-analysis investigating the effect of a restrictive versus a conventional fluid regimen on complications in patients after non-cardiac surgery in the postoperative period on the clinical ward. STUDY DESIGN: We performed a systematic search in MEDLINE, Embase, Cochrane Library, and CINAHL databases, from the start of indexing until June 2022, with constraints for English language and adult human study participants. Data were combined using classic methods of meta-analyses and were expressed as weighted pooled risk ratio (RR) or odds ratio (OR) with 95% confidence interval (CI). Quality assessment and risk of bias analyses was performed according to PRISMA guidelines. RESULTS: Seven records, three randomized controlled trials, and four non-randomized studies were included with a total of 883 patients. A restrictive fluid regimen was associated with a reduction in overall complication rate in the RCTs (RR 0.46, 95% CI 0.23 to 0.95; P < .03; I2 = 35%). This reduction in overall complication rate was not consistent in the non-randomized studies (RR 0.74, 95% CI 0.53 to 1.03; P 0.07; I2 = 45%). No significant association was found for mortality using a restrictive fluid regimen (RCTs OR 0.51, 95% CI 0.05 to 4.90; P = 0.56; I2 = 0%, non-randomized studies OR 0.30, 95% CI 0.06 to 1.46; P = 0.14; I2 = 0%). A restrictive fluid regimen is significantly associated with a reduction in postoperative length of stay in the non-randomized studies (MD - 1.81 days, 95% CI - 3.27 to - 0.35; P = 0.01; I2 = 0%) but not in the RCTs (MD 0.60 days, 95% CI - 0.75 to 1.95; P = 0.38). Risk of bias was moderate to high. Methodological quality was very low to moderate. CONCLUSION: This meta-analysis suggests restrictive fluid therapy on the ward may be associated with an effect on postoperative complication rate. However, the quality of evidence was moderate to low, the sample size was small, and the data came from both RCTs and non-randomized studies.

Effect of the machine learning-derived Hypotension Prediction Index (HPI) combined with diagnostic guidance versus standard care on depth and duration of intraoperative and postoperative hypotension in elective cardiac surgery patients: HYPE-2 - study protocol of a randomised clinical trial.

INTRODUCTION: Hypotension is common during cardiac surgery and often persists postoperatively in the intensive care unit (ICU). Still, treatment is mainly reactive, causing a delay in its management. The Hypotension Prediction Index (HPI) can predict hypotension with high accuracy. Using the HPI combined with a guidance protocol resulted in a significant reduction in the severity of hypotension in four non-cardiac surgery trials. This randomised trial aims to evaluate the effectiveness of the HPI in combination with a diagnostic guidance protocol on reducing the occurrence and severity of hypotension during coronary artery bypass grafting (CABG) surgery and subsequent ICU admission. METHODS AND ANALYSIS: This is a single-centre, randomised clinical trial in adult patients undergoing elective on-pump CABG surgery with a target mean arterial pressure of 65 mm Hg. One hundred and thirty patients will be randomly allocated in a 1:1 ratio to either the intervention or control group. In both groups, a HemoSphere patient monitor with embedded HPI software will be connected to the arterial line. In the intervention group, HPI values of 75 or above will initiate the diagnostic guidance protocol, both intraoperatively and postoperatively in the ICU during mechanical ventilation. In the control group, the HemoSphere patient monitor will be covered and silenced. The primary outcome is the time-weighted average of hypotension during the combined study phases. ETHICS AND DISSEMINATION: The medical research ethics committee and the institutional review board of the Amsterdam UMC, location AMC, the Netherlands, approved the trial protocol (NL76236.018.21). No publication restrictions apply, and the study results will be disseminated through a peer-reviewed journal. TRIAL REGISTRATION NUMBER: The Netherlands Trial Register (NL9449), ClinicalTrials.gov (NCT05821647).

Feasibility to estimate mean systemic filling pressure with inspiratory holds at the bedside.

Background: A decade ago, it became possible to derive mean systemic filling pressure (MSFP) at the bedside using the inspiratory hold maneuver. MSFP has the potential to help guide hemodynamic care, but the estimation is not yet implemented in common clinical practice. In this study, we assessed the ability of MSFP, vascular compliance (Csys), and stressed volume (Vs) to track fluid boluses. Second, we assessed the feasibility of implementation of MSFP in the intensive care unit (ICU). Exploratory, a potential difference in MSFP response between colloids and crystalloids was assessed. Methods: This was a prospective cohort study in adult patients admitted to the ICU after cardiac surgery. The MSFP was determined using 3-4 inspiratory holds with incremental pressures (maximum 35 cm H2O) to construct a venous return curve. Two fluid boluses were administered: 100 and 500 ml, enabling to calculate Vs and Csys. Patients were randomized to crystalloid or colloid fluid administration. Trained ICU consultants acted as study supervisors, and protocol deviations were recorded. Results: A total of 20 patients completed the trial. MSFP was able to track the 500 ml bolus (p < 0.001). In 16 patients (80%), Vs and Csys could be determined. Vs had a median of 2029 ml (IQR 1605-3164), and Csys had a median of 73 ml mmHg-1 (IQR 56-133). A difference in response between crystalloids and colloids was present for the 100 ml fluid bolus (p = 0.019) and in a post hoc analysis, also for the 500 ml bolus (p = 0.010). Conclusion: MSFP can be measured at the bedside and provides insights into the hemodynamic status of a patient that are currently missing. The clinical feasibility of Vs and Csys was judged ambiguously based on the lack of required hemodynamic stability. Future studies should address the clinical obstacles found in this study, and less-invasive alternatives to determine MSFP should be further explored. Clinical Trial Registration: ClinicalTrials.gov Identifier NCT03139929.

Incidence, Severity and Clinical Factors Associated with Hypotension in Patients Admitted to an Intensive Care Unit: A Prospective Observational Study.

Background: The majority of patients admitted to the intensive care unit (ICU) experience severe hypotension which is associated with increased morbidity and mortality. At present, prospective studies examining the incidence and severity of hypotension using continuous waveforms are missing. Methods: This study is a prospective observational cohort study in a mixed surgical and non-surgical ICU population. All patients over 18 years were included and continuous arterial pressure waveforms data were collected. Mean arterial pressure (MAP) below 65 mmHg for at least 10 s was defined as hypotension and a MAP below 45 mmHg as severe hypotension. The primary outcome was the incidence of hypotension. Secondary outcomes were the severity of hypotension expressed in time-weighted average (TWA), factors associated with hypotension, the number and duration of hypotensive events. Results: 499 patients were included. The incidence of hypotension (MAP < 65 mmHg) was 75% (376 out of 499) and 9% (46 out of 499) experienced severe hypotension. Median TWA was 0.3 mmHg [0−1.0]. Associated clinical factors were age, male sex, BMI and cardiogenic shock. There were 5 (1−12) events per patients with a median of 52 min (5−170). Conclusions: In a mixed surgical and non-surgical ICU population the incidence of hypotension is remarkably high.

Reducing postoperative fasting times by implementing a food service in the Post Anaesthesia Care Unit (PACU).

BACKGROUND & AIMS: Early oral postoperative nutrition is a proven beneficial element of postoperative care, resulting in improvement in quality of life and a shorter hospital stay. Guidelines state that postoperative oral nutrition can safely be started within the first hours after surgery. However, oral nutrition is mainly investigated starting from postoperative day one (POD1). This pragmatic study assessed whether a food service in the Post Anaesthesia Care Unit (PACU) may reduce postoperative fasting times on postoperative day zero (POD0) early after surgery on this high care unit. The primary outcome was postoperative fasting time in the PACU. Secondary outcomes were postoperative patient experiences and nutritional tolerance. METHODS: This pragmatic prospective before-after study, included adult patients with a planned overnight stay in the PACU, and without anticipated postoperative dietary restrictions. A food service was developed, allowing re-uptake of oral nutrition early after surgery, already at POD0. Postoperative fasting time in the PACU was defined as the duration of time between the moment of arrival at the PACU, until first postoperative intake or until 9 a.m. on POD1, when most patients were discharged to the surgical ward. Secondary outcomes such as postoperative patient experiences and nutritional tolerance were scored on POD0 at 8 p.m. and on POD1 at 8 a.m. RESULTS: A total of 235 patients were included, 119 in the pre-implementation (control) and 116 in the post-implementation (intervention) group. Mean postoperative fasting time was reduced by a mean of 6.0 h (95% CI 4.8-7.4, p-value ≤ .001), from 17.5 h in controls to 11.5 h in the intervention group. In the intervention group, 45% of patients had nutritional intake on POD0 (8 p.m.), increasing to 80% of patients on POD1 (8 a.m.), compared to no intake at both time-points in the control group. In the intervention group, average patient satisfaction increased over time. Flatus frequency was higher in the intervention group on POD1 (45% vs. 22%, p-value ≤ .001), while postoperative nausea and/or vomiting, and usage of anti-emetic medication were not different between groups. CONCLUSIONS: In a heterogeneous group of adult post-surgical patients with a planned overnight stay in the PACU, implementation of a food service resulted in a significant and clinically relevant reduction of postoperative fasting time, and increased patient satisfaction without compromising nutritional tolerance. Registered at www.isrctn.org with study ID ISRCTN17976930.

The Effect of Intermittent versus Continuous Non-Invasive Blood Pressure Monitoring on the Detection of Intraoperative Hypotension, a Sub-Study.

Intraoperative hypotension is associated with postoperative complications. However, in the majority of surgical patients, blood pressure (BP) is measured intermittently with a non-invasive cuff around the upper arm (NIBP-arm). We hypothesized that NIBP-arm, compared with a non-invasive continuous alternative, would result in missed events and in delayed recognition of hypotensive events. This was a sub-study of a previously published cohort study in adult patients undergoing surgery. The detection of hypotension (mean arterial pressure below 65 mmHg) was compared using two non-invasive methods; intermittent oscillometric NIBP-arm versus continuous NIBP measured with a finger cuff (cNIBP-finger) (Nexfin, Edwards Lifesciences). cNIBP-finger was used as the reference standard. Out of 350 patients, 268 patients (77%) had one or more hypotensive events during surgery. Out of the 286 patients, 72 (27%) had one or more missed hypotensive events. The majority of hypotensive events (92%) were detected with NIBP-arm, but were recognized at a median of 1.2 (0.6-2.2) minutes later. Intermittent BP monitoring resulted in missed hypotensive events and the hypotensive events that were detected were recognized with a delay. This study highlights the advantage of continuous monitoring. Future studies are needed to understand the effect on patient outcomes.

Prophylactic furosemide to prevent transfusion-associated circulatory overload: a randomized controlled study in rats.

Transfusion-associated circulatory overload (TACO) is the leading cause of transfusion related morbidity and mortality. The only treatment is empirical use of furosemide. Our aim was to investigate if furosemide can prevent TACO. A randomized controlled trial was performed using a previously validated two-hit rat model for TACO. Volume incompliance was induced (first hit) in anemic, anesthetized Lewis rats. Rats were randomized to placebo, low-dose (5 mg kg-1) or high-dose (15 mg kg-1) furosemide-administered prior to transfusion (second-hit) and divided over two doses. Primary outcome was change in left-ventricular end-diastolic pressure (∆LVEDP) pre- compared to post-transfusion. Secondary outcomes included changes in preload, afterload, contractility and systemic vascular resistance, as well as pulmonary outcomes. Furosemide treated animals had a significantly lower ∆LVEDP compared to placebo (p = 0.041), a dose-response effect was observed. ∆LVEDP in placebo was median + 8.7 mmHg (IQR 5.9-11), + 3.9 (2.8-5.6) in the low-dose and 1.9 (- 0.6 to 5.6) in the high-dose group. The effect of furosemide became apparent after 15 min. While urine output was significantly higher in furosemide treated animals (p = 0.03), there were no significant changes in preload, afterload, contractility or systemic vascular resistance. Furosemide rapidly and dose-dependently decreases the rise in hydrostatic pulmonary pressure following transfusion, essential for preventing TACO.

Autologous red blood cell transfusion does not result in a more profound increase in pulmonary capillary wedge pressure compared to saline in critically ill patients: A randomized crossover trial.

BACKGROUND AND OBJECTIVES: Transfusion-associated circulatory overload (TACO) is a major cause of severe transfusion-related morbidity. Transfusion of red blood cells (RBCs) has been shown to induce hydrostatic pressure overload. It is unclear which product-specific factors contribute. We set out to determine the effect of autologous RBC transfusion versus saline on pulmonary capillary wedge pressure (PCWP) change. MATERIALS AND METHODS: In a randomized crossover trial, patients who had undergone coronary bypass surgery were allocated to treatment post-operatively in the intensive care unit with either an initial 300 ml autologous RBC transfusion (salvaged during surgery) or 300 ml saline infusion first, followed by the other. Primary outcome was the difference in PCWP change. Secondary outcome measures were the difference in extra-vascular lung water index (EVLWI) and pulmonary vascular permeability index (PVPI). RESULTS: Change in PCWP was not higher after autologous RBC transfusion compared to saline (ΔPCWP 0.3 ± 0.4 vs. 0.1 ± 0.4 mmHg). ΔEVLWI and ΔPVPI were significantly decreased after autologous RBC transfusion compared to saline (ΔEVLWI -1.6 ± 0.6 vs. 0.2 ± 0.4, p = 0.02; ΔPVPI -0.3 ± 0.1 vs. 0.0 ± 0.1, p = 0.01). Haemodynamic variables and colloid osmotic pressure were not different for autologous RBC transfusion versus saline. CONCLUSION: Transfusion of autologous RBCs did not result in a more profound increase in PCWP compared to saline. RBC transfusion resulted in a decrease of EVLWI and PVPI compared to saline. Our data suggest that transfusing autologous RBCs may lead to less pulmonary oedema compared to saline. Future studies with allogeneic RBCs are needed to investigate other factors that may mediate the increase of PCWP, resulting in TACO.

Differential effects of speed and volume on transfusion-associated circulatory overload: A randomized study in rats.

BACKGROUND AND OBJECTIVES: Transfusion-associated circulatory overload (TACO) is the primary cause of transfusion-related mortality. Speed and volume of transfusion are major risk factors. The aim of this study was to investigate the interaction of red blood cell (RBC) transfusion speed and volume on the development of TACO. MATERIALS AND METHODS: A validated model for TACO in anaemic Lewis rats with an acute myocardial infarction was used. The effect on pulmonary hydrostatic pressure of one, two or four units of packed RBCs transfused in either 30 or 60 min was evaluated (3.3-26.6 ml·kg-1 ·hr-1 ). Pulmonary capillary pressure was measured as left ventricular end-diastolic pressure (LVEDP). Cardiac stress biomarkers atrial natriuretic-peptide (ANP) and N-terminal pro-brain natriuretic peptide (NT-proBNP) were measured 1-h post-transfusion. RESULTS: Thirty animals were included (n = 5 per group). Transfusion of RBCs increased LVEDP in a volume-dependent manner (ΔLVEDP [mmHg]: -0.95, +0.50, +6.26, p < 0.001). Fast transfusion increased overall ΔLVEDP by +3.5 mmHg and up to +11.8 mmHg in the four units' group (p = 0.016). Doubling transfusion speed increased ΔLVEDP more than doubling volume in the larger volume groups. No difference in ANP or NT-proBNP were seen in high transfusion volume or groups. CONCLUSION: Transfusion volume dose-dependently increased LVEDP, with speed of transfusion rapidly elevating LVEDP at higher transfusion volumes. ANP and NT-proBNP were not impacted by transfusion volume or speed in this model. TACO is seen as purely volume overload, however, this study emphasizes that limiting transfusion speed, as a modifiable risk factor, might aid in preventing TACO.

Perioperative Care Standards in Cardiac Surgery Patients Aiming at Enhancing Recovery: A Nationwide Survey in the Netherlands and Belgium.

OBJECTIVE: The aim of this survey was to describe existing perioperative care standards and best practices in the Netherlands and Belgium. DESIGN: An online survey was followed up by an in-depth personal interview. The main outcomes were the existing standards of perioperative care for patients undergoing cardiac surgery. SETTING: The online survey and subsequent interviews were targeted to one representative in the intensive care unit (ICU), cardiac surgery, and anesthesiology department from each cardiac surgical center in the Netherlands and Belgium. PARTICIPANTS: A representative intensive care physician, cardiac surgeon, and cardiac anesthesiologist. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The response rate was 60% (71% for the Netherlands, and 44% in Belgium). Agreement across centers was found for discontinuation of proton-pump inhibitors (80%) and avoiding intra- and postoperative (92%) nonsteroidal antiinflammatory drugs. Additionally, 98% of respondents stated that physiotherapy should be started immediately in the ICU. Major divergence was found for elements such as the discontinuation of angiotensin-converting enzyme inhibitors (55%) or the postoperative use of chest support vests (44%). CONCLUSIONS: The authors demonstrated a wide range of different local protocols. Strategies differed among disciplines, hospitals, and countries. This emphasized the need for the implementation of a more universal protocol to further reduce variance and improve recovery practices. This nationwide survey was the first of its kind simultaneously studying best practices for cardiac surgery through the entire care pathway at the advent of Enhanced Recovery After Surgery (ERAS) Cardiac implementation. A multinational randomized controlled trial to test the implementation of an evidence-based ERAS Cardiac protocol is the next step to pave the way for further outcome improvements in this high-risk population.

Performance of a machine-learning algorithm to predict hypotension in mechanically ventilated patients with COVID-19 admitted to the intensive care unit: a cohort study.

The Hypotension Prediction Index (HPI) is a commercially available machine-learning algorithm that provides warnings for impending hypotension, based on real-time arterial waveform analysis. The HPI was developed with arterial waveform data of surgical and intensive care unit (ICU) patients, but has never been externally validated in the latter group. In this study, we evaluated diagnostic ability of the HPI with invasively collected arterial blood pressure data in 41 patients with COVID-19 admitted to the ICU for mechanical ventilation. Predictive ability was evaluated at HPI thresholds from 0 to 100, at incremental intervals of 5. After exceeding the studied threshold, the next 20 min were screened for positive (mean arterial pressure (MAP) < 65 mmHg for at least 1 min) or negative (absence of MAP < 65 mmHg for at least 1 min) events. Subsequently, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and time to event were determined for every threshold. Almost all patients (93%) experienced at least one hypotensive event. Median number of events was 21 [7-54] and time spent in hypotension was 114 min [20-303]. The optimal threshold was 90, with a sensitivity of 0.91 (95% confidence interval 0.81-0.98), specificity of 0.87 (0.81-0.92), PPV of 0.69 (0.61-0.77), NPV of 0.99 (0.97-1.00), and median time to event of 3.93 min (3.72-4.15). Discrimination ability of the HPI was excellent, with an area under the curve of 0.95 (0.93-0.97). This validation study shows that the HPI correctly predicts hypotension in mechanically ventilated COVID-19 patients in the ICU, and provides a basis for future studies to assess whether hypotension can be reduced in ICU patients using this algorithm.

Effect of Hypotension Prediction Index-guided intraoperative haemodynamic care on depth and duration of postoperative hypotension: a sub-study of the Hypotension Prediction trial.

BACKGROUND: Intraoperative and postoperative hypotension are associated with morbidity and mortality. The Hypotension Prediction (HYPE) trial showed that the Hypotension Prediction Index (HPI) reduced the depth and duration of intraoperative hypotension (IOH), without excess use of intravenous fluid, vasopressor, and/or inotropic therapies. We hypothesised that intraoperative HPI-guided haemodynamic care would reduce the severity of postoperative hypotension in the PACU. METHODS: This was a sub-study of the HYPE study, in which 60 adults undergoing elective noncardiac surgery were allocated randomly to intraoperative HPI-guided or standard haemodynamic care. Blood pressure was measured using a radial intra-arterial catheter, which was connected to a FloTracIQ sensor. Hypotension was defined as MAP <65 mm Hg, and a hypotensive event was defined as MAP <65 mm Hg for at least 1 min. The primary outcome was the time-weighted average (TWA) of postoperative hypotension. Secondary outcomes were absolute incidence, area under threshold for hypotension, and percentage of time spent with MAP <65 mm Hg. RESULTS: Overall, 54/60 (90%) subjects (age 64 (8) yr; 44% female) completed the protocol, owing to failure of the FloTracIQ device in 6/60 (10%) patients. Intraoperative HPI-guided care was used in 28 subjects; 26 subjects were randomised to the control group. Postoperative hypotension occurred in 37/54 (68%) subjects. HPI-guided care did not reduce the median duration (TWA) of postoperative hypotension (adjusted median difference, vs standard of care: 0.118; 95% confidence interval [CI], 0-0.332; P=0.112). HPI-guidance reduced the percentage of time with MAP <65 mm Hg by 4.9% (adjusted median difference: -4.9; 95% CI, -11.7 to -0.01; P=0.046). CONCLUSIONS: Intraoperative HPI-guided haemodynamic care did not reduce the TWA of postoperative hypotension.

Goal-directed fluid therapy vs. low central venous pressure during major open liver resections (GALILEO): a surgeon- and patient-blinded randomized controlled trial.

BACKGROUND: Low central venous pressure (low-CVP) is the clinical standard for fluid therapy during major liver surgery. Although goal-directed fluid therapy (GDFT) has been associated with reduced morbidity and mortality in major abdominal surgery, concerns remain on blood loss when applying GDFT in liver surgery. This randomized trial compared outcomes of low-CVP and GDFT during major liver resections. METHODS: In this surgeon- and patient-blinded RCT, patients undergoing major open liver resections (≥3 segments) were randomized between low-CVP (n = 20) or GDFT (n = 20). Primary outcome was intraoperative blood loss. Secondary outcomes included the quality of the surgical field (VAS scale 0 (worst)-100 (best)) and major morbidity (≥grade 3 Clavien-Dindo). RESULTS: During surgery, CVP was 3 ± 2 mmHg in the low-CVP group vs. 7 ± 3 mmHg in the GDFT group (P < 0.001). Blood loss (1425 vs. 1275 mL; P = 0.640) and the rate of major morbidity (40% vs. 50%, P = 0.751), did not differ between low-CVP and GDFT, respectively. The quality of the surgical field was comparable between groups (low-CVP 83% vs. GDFT 80%, P = 0.955). CONCLUSION: In major open liver resections, GDFT was not associated with differences in intraoperative blood loss, major morbidity or quality of the surgical field, compared to low-CVP. Larger RCTs are needed to confirm this finding. Registration number: NTR5821 (www.trialregister.nl).

Predictors of 30-Day Mortality Among Dutch Patients Undergoing Colorectal Cancer Surgery, 2011-2016.

Importance: Quality improvement programs for colorectal cancer surgery have been introduced with benchmarking based on quality indicators, such as mortality. Detailed (pre)operative characteristics may offer relevant information for proper case-mix correction. Objective: To investigate the added value of machine learning to predict quality indicators for colorectal cancer surgery and identify previously unrecognized predictors of 30-day mortality based on a large, nationwide colorectal cancer registry that collected extensive data on comorbidities. Design, Setting, and Participants: All patients who underwent resection for primary colorectal cancer registered in the Dutch ColoRectal Audit between January 1, 2011, and December 31, 2016, were included. Multiple machine learning models (multivariable logistic regression, elastic net regression, support vector machine, random forest, and gradient boosting) were made to predict quality indicators. Model performance was compared with conventionally used scores. Risk factors were identified by logistic regression analyses and Shapley additive explanations (ie, SHAP values). Statistical analysis was performed between March 1 and September 30, 2020. Main Outcomes and Measures: The primary outcome of this cohort study was 30-day mortality. Prediction models were trained on a training set by performing 5-fold cross-validation, and outcomes were measured by the area under the receiver operating characteristic curve on the test set. Machine learning was further used to identify risk factors, measured by odds ratios and SHAP values. Results: This cohort study included 62 501 records, most patients were male (35 116 [56.2%]), were aged 61 to 80 years (41 560 [66.5%]), and had an American Society of Anesthesiology score of II (35 679 [57.1%]). A 30-day mortality rate of 2.7% (n = 1693) was found. The area under the curve of the best machine learning model for 30-day mortality (0.82; 95% CI, 0.79-0.85) was significantly higher than the American Society of Anesthesiology score (0.74; 95% CI, 0.71-0.77; P < .001), Charlson Comorbidity Index (0.66; 95% CI, 0.63-0.70; P < .001), and preoperative score to predict postoperative mortality (0.73; 95% CI, 0.70-0.77; P < .001). Hypertension, myocardial infarction, chronic obstructive pulmonary disease, and asthma were comorbidities with a high risk for increased mortality. Machine learning identified specific risk factors for a complicated course, intensive care unit admission, prolonged hospital stay, and readmission. Laparoscopic surgery was associated with a decreased risk for all adverse outcomes. Conclusions and Relevance: This study found that machine learning methods outperformed conventional scores to predict 30-day mortality after colorectal cancer surgery, identified specific patient groups at risk for adverse outcomes, and provided directions to optimize benchmarking in clinical audits.

Clinical performance of a machine-learning algorithm to predict intra-operative hypotension with noninvasive arterial pressure waveforms: A cohort study.

BACKGROUND: Intra-operative hypotension is associated with adverse postoperative outcomes. A machine-learning-derived algorithm developed to predict hypotension based on arterial blood pressure (ABP) waveforms significantly reduced intra-operative hypotension. The algorithm calculates the likelihood of hypotension occurring within minutes, expressed as the Hypotension Prediction Index (HPI) which ranges from 0 to 100. Currently, HPI is only available for patients monitored with invasive ABP, which is restricted to high-risk procedures and patients. In this study, the performance of HPI, employing noninvasive continuous ABP measurements, is assessed. OBJECTIVES: The first aim was to compare the performance of the HPI algorithm, using noninvasive versus invasive ABP measurements, at a mathematically optimal HPI alarm threshold (Youden index). The second aim was to assess the performance of the algorithm using a HPI alarm threshold of 85 that is currently used in clinical trials. Hypotension was defined as a mean arterial pressure (MAP) below 65 mmHg for at least 1 min. The predictive performance of the algorithm at different HPI alarm thresholds (75 and 95) was studied. DESIGN: Observational cohort study. SETTING: Tertiary academic medical centre. PATIENTS: Five hundred and seven adult patients undergoing general surgery. RESULTS: The performance of the algorithm with invasive and noninvasive ABP input was similar. A HPI alarm threshold of 85 showed a median [IQR] time from alarm to hypotension of 2.7 [1.0 to 7.0] min with a sensitivity of 92.7 (95% confidence interval [CI], 91.2 to 94.3), specificity of 87.6 (95% CI, 86.2 to 89.0), positive predictive value of 79.9 (95% CI, 77.7 to 82.1) and negative predictive value of 95.8 (95% CI, 94.9 to 96.7). A HPI alarm threshold of 75 provided a lower positive predictive value but a prolonged time from prediction to actual hypotension. CONCLUSION: This study demonstrated that the algorithm can be employed using continuous noninvasive ABP waveforms. This opens up the potential to predict and prevent hypotension in a larger patient population. TRIAL REGISTRATION: Clinical trials registration number NCT03533205.

One of the first validations of an artificial intelligence algorithm for clinical use: The impact on intraoperative hypotension prediction and clinical decision-making.

This review describes the steps and conclusions from the development and validation of an artificial intelligence algorithm (the Hypotension Prediction Index), one of the first machine learning predictive algorithms used in the operating room environment. The algorithm has been demonstrated to reduce intraoperative hypotension in two randomized controlled trials via real-time prediction of upcoming hypotensive events prompting anesthesiologists to act earlier, more often, and differently in managing impending hypotension. However, the algorithm entails no dynamic learning process that evolves from use in clinical patient care, meaning the algorithm is fixed, and furthermore provides no insight into the decisional process that leads to an early warning for intraoperative hypotension, which makes the algorithm a "black box." Many other artificial intelligence machine learning algorithms have these same disadvantages. Clinical validation of such algorithms is relatively new and requires more standardization, as guidelines are lacking or only now start to be drafted. Before adaptation in clinical practice, impact of artificial intelligence algorithms on clinical behavior, outcomes and economic advantages should be studied too.

A Higher Fluid Balance in the Days After Septic Shock Reversal Is Associated With Increased Mortality: An Observational Cohort Study.

OBJECTIVES: Previous studies demonstrated that extensive fluid loading and consequently positive fluid balances during sepsis resuscitation are associated with adverse outcome. Yet, the association between fluid balance and mortality after reversal of shock, that is, during deresuscitation, is largely unappreciated. Our objective was to investigate the effects of fluid balance on mortality in the days after septic shock reversal. DESIGN: Retrospective observational cohort study. SETTING: ICUs of two university-affiliated hospitals in The Netherlands. PATIENTS: Adult patients admitted with septic shock followed by shock reversal. Reversal of septic shock was defined based on Sepsis-3 criteria as the first day that serum lactate was less than or equal to 2 mmol/L without vasopressor requirement. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Reversal of septic shock occurred in 636 patients, of whom 20% died in the ICU. Mixed-effects logistic regression modeling, adjusted for possible confounders, showed that fluid balance in the days after reversal of septic shock (until discharge or death) was an independent predictor of ICU mortality: odds ratio 3.18 (1.90-5.32) per 10 mL/kg increase in daily fluid balance. Similar results were found for 30-day, 90-day, hospital, and 1-year mortality: odds ratios 2.09 (1.64-2.67); 1.79 (1.38-2.32); 1.70 (1.40-2.07); and 1.53 (1.17-2.01), respectively. Positive cumulative fluid balances vs. neutral or negative fluid balances on the final day in the ICU were associated with increased ICU, hospital, 30-day, and 90-day mortality: odds ratios 3.46 (2.29-5.23); 3.39 (2.35-4.9); 5.33 (3.51-8.08); and 3.57 (2.49-5.12), respectively. Using restricted cubic splines, we found a dose-response relationship between cumulative fluid balance after shock reversal and ICU mortality. CONCLUSIONS: A higher fluid balance in the days after septic shock reversal was associated with increased mortality. This stresses the importance of implementing restrictive and deresuscitative fluid management strategies after initial hemodynamic resuscitation. Prospective interventional studies are needed to confirm our results.

Long-Term Mortality Among ICU Patients With Stroke Compared With Other Critically Ill Patients.

OBJECTIVES: Assessment of all-cause mortality of intracerebral hemorrhage and ischemic stroke patients admitted to the ICU and comparison to the mortality of other critically ill ICU patients classified into six other diagnostic subgroups and the general Dutch population. DESIGN: Observational cohort study. SETTING: All ICUs participating in the Dutch National Intensive Care Evaluation database. PATIENTS: All adult patients admitted to these ICUs between 2010 and 2015; patients were followed until February 2017. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Of all 370,386 included ICU patients, 7,046 (1.9%) were stroke patients, 4,072 with ischemic stroke, and 2,974 with intracerebral hemorrhage. Short-term mortality in ICU-admitted stroke patients was high with 30 days mortality of 31% in ischemic stroke and 42% in intracerebral hemorrhage. In the longer term, the survival curve gradient among ischemic stroke and intracerebral hemorrhage patients stabilized. The gradual alteration of mortality risk after ICU admission was assessed using left-truncation with increasing minimum survival period. ICU-admitted stroke patients who survive the first 30 days after suffering from a stroke had a favorable subsequent survival compared with other diseases necessitating ICU admission such as patients admitted due to sepsis or severe community-acquired pneumonia. After having survived the first 3 months after ICU admission, multivariable Cox regression analyses showed that case-mix adjusted hazard ratios during the follow-up period of up to 3 years were lower in ischemic stroke compared with sepsis (adjusted hazard ratio, 1.21; 95% CI, 1.06-1.36) and severe community-acquired pneumonia (adjusted hazard ratio, 1.57; 95% CI, 1.39-1.77) and in intracerebral hemorrhage patients compared with these groups (adjusted hazard ratio, 1.14; 95% CI, 0.98-1.33 and adjusted hazard ratio, 1.49; 95% CI, 1.28-1.73). CONCLUSIONS: Stroke patients who need intensive care treatment have a high short-term mortality risk, but this alters favorably with increasing duration of survival time after ICU admission in patients with both ischemic stroke and intracerebral hemorrhage, especially compared with other populations of critically ill patients such as sepsis or severe community-acquired pneumonia patients.