Postdischarge Survival After Sepsis: A Cohort Study.

BACKGROUND: After hospital discharge, patients who had sepsis have increased mortality. We sought to estimate factors associated with postdischarge mortality and how they vary with time after discharge. METHODS: This was a retrospective study of hospital survivors of sepsis using time-varying Cox proportional hazard models, which produce a baseline hazard ratio (HR) and a second number (δHR) that reflects the amount by which the baseline HR changes with time. RESULTS: Of the 32,244 patients who survived sepsis at hospital discharge, 13,565 patients (42%) died (mean ± standard deviation: 1.41 ± 1.87 years) after discharge from the index hospitalization, while 18,679 patients were still alive at follow-up (4.98 ± 2.86 years). The mortality rate decreased with time after discharge: approximately 8.7% of patients died during the first month after discharge, 1.1% of patients died during the 12th month after discharge, and 0.3%% died during the 60th month; after Kaplan-Meier analysis, survival was 91% (95% confidence interval [CI], 91%-92%) at 1 month, 76% (95% CI, 76%-77%) at 1 year, 57% (95% CI, 56%-58%) at 5 years, and 48% (95% CI, 47%-48%) at 10 years after discharge. Organ dysfunction at discharge was associated with worse survival. In particular, elevated urea nitrogen at discharge (HR, 1.10 per 10 mg/dL, 95% CI, 1.08-1.12, P < .001) was associated with increased mortality, but the HR decreased with time from discharge (δHR, 0.98 per 10 mg/dL per year, 95% CI, 0.98-0.99, P < .001). Higher hemoglobin levels were associated with lower mortality (HR, 0.92 per g/dL, 95% CI, 0.91-0.93, P < .001), but this association increased with increasing time after discharge (δHR, 1.02 per g/dL per year, 95% CI, 1.01-1.02, P < .001). Older age was associated with an increased risk of mortality (HR, 1.29 per decade of age, 95% CI, 1.27-1.31, P < .001) that grew with increasing time after discharge (δHR, 1.01 per year of follow-up per decade of age, 95% CI, 1.00-1.02, P < .001). Compared to private insurances Medicaid as primary insurance was associated with an increased risk of mortality (HR, 1.17, 95% CI, 1.10-1.25, P < .001) that did not change with time after discharge. In contrast, Medicare status was initially associated with a similar risk of mortality as private insurance at discharge (HR, 1), but was associated with greater risk as time after discharge increased (δHR, 1.04 per year of follow-up, 95% CI, 1.03-1.05, P < .001). CONCLUSIONS: Acute physiologic derangements and organ dysfunction were associated with postdischarge mortality with the associations decreasing over time.

Differences between patients in whom physicians agree versus disagree about the preoperative diagnosis of heart failure.

STUDY OBJECTIVE: To quantify preoperative heart failure (HF) diagnostic agreement and identify characteristics of patients in whom physicians agreed versus disagreed about the diagnosis. DESIGN: Observational cohort study. SETTING: Patients undergoing major non-cardiac surgery at an academic center between 2015 and 2019. PATIENTS: 40,659 patients undergoing major non-cardiac surgery, among which a stratified subsample of 1018 patients with and without documented HF was reviewed. INTERVENTIONS: Via a panel of physicians frequently managing patients with HF (cardiologists, cardiac anesthesiologists, intensivists), detailed chart reviews were performed (two per patient; median review time 32 min per reviewer per patient) to render adjudicated HF diagnoses. MEASUREMENTS: Adjudicated diagnostic agreement measures (percent agreement, Krippendorf's alpha) and univariate comparisons (standardized differences) between patients in whom physicians agreed versus disagreed about the preoperative HF diagnosis. MAIN RESULTS: Among patients with documented HF, physicians agreed about the diagnosis in 80.0% of cases (consensus positive), disagreed in 13.8% (disagreement), and refuted the diagnosis in 6.3% (consensus negative). Conversely, among patients without documented HF, physicians agreed about the diagnosis in 88.0% (consensus negative), disagreed in 8.4% (disagreement), and refuted the diagnosis in 3.6% (consensus positive). The estimated agreement for the 40,659 cases was 91.1% (95% CI 88.3%-93.9%); Krippendorff's alpha was 0.77 (0.75-0.80). Compared to patients in whom physicians agreed about a HF diagnosis, patients in whom physicians disagreed exhibited fewer guideline-defined HF diagnostic criteria. CONCLUSIONS: Physicians usually agree about HF diagnoses adjudicated via chart review, although disagreement is not uncommon and may be partly explained by heterogeneous clinical presentations. Our findings inform preoperative screening processes by identifying patients whose characteristics contribute to physician disagreement via chart review. Clinical Trial Number / Registry URL: Not applicable.

Acute Cricoarytenoid Joint Fixation Following Routine Intubation: A Case Report.

Prolonged intubation is a common cause of injury to the posterior larynx often resulting in cricoarytenoid joint (CAJ) fixation and posterior glottic stenosis (PGS). We present a case of respiratory failure due to acute bilateral CAJ fixation and PGS following only 2 days of intubation for routine cardiac surgery. A tracheostomy was placed due to critical airway obstruction. Clinicians should remain vigilant for laryngeal injury presenting as CAJ fixation and PGS. Prompt surgical consultation is advised as early intervention is associated with reduced morbidity.

Identification of intraoperative management strategies that have a differential effect on patients with reduced left ventricular ejection fraction: a retrospective cohort study.

BACKGROUND: There are few data to guide the intraoperative management of patients with reduced left ventricular ejection fraction (LVEF). This study aimed to describe how patients with reduced LVEF are managed differently and to identify and treatments had a different risk profile in this population. METHODS: We performed a retrospective cohort study of adult patients who underwent general anesthesia for non-cardiac surgery. The effect of anesthesia medications and fluid balance was compared between those with and without a reduced preoperative LVEF. The primary outcome was a composite of acute kidney injury, myocardial injury, pulmonary complications, and 30-day mortality. Multivariable logistic regression was used to adjust for confounders. Treatments that affected patients with reduced LVEF differently were defined as those associated with the primary outcome that also had a significant interaction with LVEF. RESULTS: A total of 9420 patients were included. Patients with reduced LVEF tended to have a less positive fluid balance. Etomidate, calcium, and phenylephrine were use more frequently, while propofol and remifentanil were used less frequently. Remifentanil affected patients with reduced LVEF differently than those without (interaction term OR 2.71, 95% CI 1.30-5.68, p = 0.008). While the use of remifentanil was associated with fewer complications in patients with normal systolic function (OR 0.54, 95% CI 0.42-0.68, p < 0.001), it was associated with an increase in complications in patients with reduced LVEF (OR = 3.13, 95% CI 3.06-5.98, p = 0.026). CONCLUSIONS: Patients with a reduced preoperative LVEF are treated differently than those with a normal LVEF when undergoing non-cardiac surgery. An association was found between the use of remifentanil and an increase in postoperative adverse events that was unique to this population. Future research is needed to determine if this relationship is secondary to the medication itself or reflects a difference in how remifentanil is used in patients with reduced LVEF.

Genetic variants associated with sepsis.

BACKGROUND: The variable presentations and different phenotypes of sepsis suggest that risk of sepsis comes from many genes each having a small effect. The cumulative effect can be used to create individual risk profile. The purpose of this study was to create a polygenic risk score and determine the genetic variants associated with sepsis. METHODS: We sequenced ~14 million single nucleotide polymorphisms with a minimac imputation quality R2>0.3 and minor allele frequency >10-6 in patients with Sepsis-2 or Sepsis-3. Genome-wide association was performed using Firth bias-corrected logistic regression. Semi-parsimonious logistic regression was used to create polygenic risk scores and reduced regression to determine the genetic variants independently associated with sepsis. FINDINGS: 2261 patients had sepsis and 13,068 control patients did not. The polygenic risk scores had good discrimination: c-statistic = 0.752 ± 0.005 for Sepsis-2 and 0.752 ± 0.007 for Sepsis-3. We found 772 genetic variants associated with Sepsis-2 and 442 with Sepsis-3, p<0.01. After multivariate adjustment, 100 variants on 85 genes were associated with Sepsis-2 and 69 variants in 54 genes with Sepsis-3. Twenty-five variants were present in both the Sepsis-2 and Sepsis-3 groups out of 32 genes that were present in both groups. The other 7 genes had different variants present. Most variants had small effect sizes. CONCLUSIONS: Sepsis-2 and Sepsis-3 have both separate and shared genetic variants. Most genetic variants have small effects sizes, but cumulatively, the polygenic risk scores have good discrimination.

Endothelial Cell-Activating Antibodies in COVID-19.

OBJECTIVE: While endothelial dysfunction has been implicated in the widespread thromboinflammatory complications of COVID-19, the upstream mediators of endotheliopathy remain, for the most part, unknown. This study was undertaken to identify circulating factors contributing to endothelial cell activation and dysfunction in COVID-19. METHODS: Human endothelial cells were cultured in the presence of serum or plasma from 244 patients hospitalized with COVID-19 and plasma from 100 patients with non-COVID-19-related sepsis. Cell adhesion molecules (E-selectin, vascular cell adhesion molecule 1, and intercellular adhesion molecule 1 [ICAM-1]) were quantified using in-cell enzyme-linked immunosorbent assay. RESULTS: Serum and plasma from COVID-19 patients increased surface expression of cell adhesion molecules. Furthermore, levels of soluble ICAM-1 and E-selectin were elevated in patient serum and correlated with disease severity. The presence of circulating antiphospholipid antibodies was a strong marker of the ability of COVID-19 serum to activate endothelium. Depletion of total IgG from antiphospholipid antibody-positive serum markedly reduced the up-regulation of cell adhesion molecules. Conversely, supplementation of control serum with patient IgG was sufficient to trigger endothelial activation. CONCLUSION: These data are the first to indicate that some COVID-19 patients have potentially diverse antibodies that drive endotheliopathy, providing important context regarding thromboinflammatory effects of autoantibodies in severe COVID-19.

Endothelial cell-activating antibodies in COVID-19.

OBJECTIVE: While endothelial dysfunction has been implicated in the widespread thrombo-inflammatory complications of coronavirus disease-19 ( COVID-19 ), the upstream mediators of endotheliopathy remain for the most part cryptic. Our aim was to identify circulating factors contributing to endothelial cell activation and dysfunction in COVID-19. METHODS: Human endothelial cells were cultured in the presence of serum or plasma from 244 patients hospitalized with COVID-19 and plasma from 100 patients with non-COVID sepsis. Cell adhesion molecules (E-selectin, VCAM-1, and ICAM-1) were quantified by in-cell ELISA. RESULTS: Serum and plasma from patients with COVID-19 increased surface expression of cell adhesion molecules. Furthermore, levels of soluble ICAM-1 and E-selectin were elevated in patient serum and tracked with disease severity. The presence of circulating antiphospholipid antibodies was a strong marker of the ability of COVID-19 serum to activate endothelium. Depletion of total IgG from antiphospholipid antibody-positive serum markedly restrained upregulation of cell adhesion molecules. Conversely, supplementation of control serum with patient IgG was sufficient to trigger endothelial activation. CONCLUSION: These data are the first to suggest that some patients with COVID-19 have potentially diverse antibodies that drive endotheliopathy, adding important context regarding thrombo-inflammatory effects of autoantibodies in severe COVID-19.

Use of the Change in Weaning Parameters as a Predictor of Successful Re-Extubation.

OBJECTIVE: Weaning parameters are well studied in patients undergoing first time extubation. Fewer data exists to guide re-extubation of patients who failed their first extubation attempt. It is reasonable to postulate that improved weaning parameters between the first and second extubation attempt would lead to improved rates of re-extubation success. To investigate, we studied a cohort of patients who failed their first extubation attempt and underwent a second attempt at extubation. We hypothesized that improvement in weaning parameters between the first and the second extubation attempt is associated with successful reextubation. INTERVENTIONS: Rapid shallow breathing index (RSBI), maximum inspiratory pressure (MIP), vital capacity (VC), and the blood partial pressure of CO2 (PaCO2) were measured and recorded in the medical record prior to extubation along with demographic information. We examined the relationship between the change in extubation and re-extubation weaning parameters and re-extubation success. MEASUREMENTS AND MAIN RESULTS: A total of 1283 adult patients were included. All weaning parameters obtained prior to re-extubation differed between those who were successful and those who required a second reintubation. Those with reextubation success had slightly lower PaCO2 values (39.5 ± 7.4 mmHg vs. 41.6 ± 9.1 mmHg, p = 0.0045) and about 13% higher vital capacity volumes (1021 ± 410 mL vs. 907 ± 396 mL, p = 0.0093). Lower values for RSBI (53 ± 32 breaths/min/L vs. 69 ± 42 breaths/min/L, p < 0.001) and MIP (-41 ± 12 cmH2O vs. -38 ± 13 cm H2O), p = 0.0225) were seen in those with re-extubation success. Multivariable logistical regression demonstrates lack of independent associated between the change in parameters between the 2 attempts and re-extubation success. CONCLUSIONS: The relationship between the changes in extubation parameters through successive attempts is driven primarily by the value obtained immediately prior to re-extubation. These findings do not support waiting for an improvement in extubation parameters to extubate patients who failed a first attempt at extubation if extubation parameters are compatible with success.

Preoperative clinical diagnostic accuracy of heart failure among patients undergoing major noncardiac surgery: a single-centre prospective observational analysis.

Background: Reliable diagnosis of heart failure during preoperative evaluation is important for perioperative management and long-term care. We aimed to quantify preoperative heart failure diagnostic accuracy and explore characteristics of patients with heart failure misdiagnoses. Methods: We performed an observational cohort study of adults undergoing major noncardiac surgery at an academic hospital between 2015 and 2019. A preoperative clinical diagnosis of heart failure was defined using keywords from the history and clinical examination or administrative documentation. Across stratified subsamples of cases with and without clinically diagnosed heart failure, health records were intensively reviewed by an expert panel to develop an adjudicated heart failure reference standard using diagnostic criteria congruent with consensus guidelines. We calculated agreement among experts, and analysed performance of clinically diagnosed heart failure compared with the adjudicated reference standard. Results: Across 40 555 major noncardiac procedures, a stratified subsample of 511 patients was reviewed by the expert panel. The prevalence of heart failure was 9.1% based on clinically diagnosed compared with 13.3% (95% confidence interval [CI], 10.3-16.2%) estimated by the expert panel. Overall agreement and inter-rater reliability (kappa) among heart failure experts were 95% and 0.79, respectively. Based upon expert adjudication, heart failure was clinically diagnosed with an accuracy of 92.8% (90.6-95.1%), sensitivity 57.4% (53.1-61.7%), specificity 98.3% (97.1-99.4%), positive predictive value 83.5% (80.3-86.8%), and negative predictive value 93.8% (91.7-95.9%). Conclusions: Limitations exist to the preoperative clinical diagnosis of heart failure, with nearly half of cases undiagnosed preoperatively. Considering the risks of undiagnosed heart failure, efforts to improve preoperative heart failure diagnoses are warranted.

The effect of timing of initiation of renal replacement therapy on mortality: A retrospective case-control study.

PURPOSE: To determine if earlier initiation of renal replacement therapy (RRT) is associated with improved survival in patients with severe acute kidney injury. METHODS: We performed a retrospective case-control study of propensity-matched groups with multivariable logistic regression using Akaike Information Criteria to adjust for non-matched variables in a surgical ICU in a tertiary care hospital. RESULTS: We matched 169 of 205 (82%) patients with new initiation of RRT (EARLY group) to 169 similar patients who did not initiate RRT on that day (DEFERRED group). Eighteen (11%) of DEFERRED eventually received RRT before discharge. By univariate analysis, ICU mortality was higher in EARLY (n = 60 (36%) vs. n = 23 (14%), p < 0.001) as was hospital mortality (n = 73 (43%) vs. n = 44 (26%), p = 0.001). Of the 18 RRT patients in DEFERRED, 12 (67%) died in ICU and 13 (72%) in hospital. After propensity matching and logistic regression, we found that EARLY initiation of RRT was associated with a more than doubling of ICU mortality (aOR = 2.310, 95% confidence interval = 1.254-4.257, p = 0.007). However, after similar adjustment, there was no difference in hospital mortality (aOR = 1.283, 95% CI = 0.753-2.186, p = 0.360). CONCLUSIONS: While ICU mortality was increased in the EARLY group, there was no difference in hospital mortality between EARLY and DEFERRED groups.

Reduced Echocardiographic Inotropy Index after Cardiopulmonary Bypass Is Associated With Complications After Cardiac Surgery: An Institutional Outcomes Study.

OBJECTIVE: Despite advances in echocardiography and hemodynamic monitoring, limited progress has been made to effectively quantify left ventricular function during cardiac surgery. Traditional measures, including left ventricular ejection fraction (LVEF) and cardiac index, remain dependent on loading conditions; more complex measures remain impractical in a dynamic surgical setting. However, the Smith-Madigan Inotropy Index (SMII) and potential-to-kinetic energy ratio (PKR) offer promise as measures calculable during cardiac surgery and potentially predictive of outcomes. Using echocardiographic and hemodynamic monitoring data, the authors aimed to calculate SMII and PKR values after cardiopulmonary bypass and understand associations with postoperative outcomes, adjusting for previously identified risk factors. DESIGN: Observational cohort study. SETTING: Tertiary care academic hospital. PATIENTS: The study comprised 189 elective adult cardiac surgical procedures from 2015-2016. INTERVENTION: None. MEASUREMENTS AND MAIN RESULTS: The primary outcome was postoperative mortality or organ system complication (stroke, prolonged ventilation, reintubation, cardiac arrest, acute kidney injury, new-onset atrial fibrillation). After adjustment, SMII <0.83 W/m2 independently predicted the primary outcome (adjusted odds ratio 2.19, 95% confidence interval 1.08-4.42); whereas PKR, LVEF, and cardiac index demonstrated no associations. When SMII and PKR were incorporated into a EuroSCORE II risk model, predictive performance improved (net reclassification index improvement 0.457; p = 0.001); whereas a model incorporating LVEF and cardiac index demonstrated no improvement (0.130; p = 0.318). CONCLUSION: The present study demonstrated that SMII, but not PKR, as a measure of cardiac function was associated with major complications. The study's data may guide investigations of more suitable perioperative goal-directed therapies to reduce complications after cardiac surgery.

Association of Both High and Low Left Ventricular Ejection Fraction With Increased Risk After Coronary Artery Bypass Grafting.

BACKGROUND: While reduced left ventricular ejection fraction (LVEF) is a known risk factor for complications after coronary artery bypass grafting (CABG), the relevance of higher LVEF values has not been established. Currently, most risk stratification tools consider LVEF values above a certain point as normal. However, since this does not account for insufficient ventricular filling or increased adrenergic tone, higher values may have clinical significance. To improve our understanding of this situation, we investigated the relationship of preoperative LVEF values with short- and long-term outcomes after CABG using a strategy that allowed for the identification of nonlinear relationships. We hypothesised that both higher and lower values are independently associated with increased postoperative complications and death in this population. METHODS: We performed a single-centre retrospective cohort study of patients undergoing isolated CABG surgery. All patients had a preoperative measurement of their LVEF. Surgery involving mitral valve repair was excluded in order to eliminate the impact of mitral regurgitation. The primary outcome was long-term mortality; secondary outcomes included atrial fibrillation, operative mortality, and a composite outcome including any postoperative adverse event. Fractional polynomial equations were used to model the relationship between LVEF and outcomes so we could account for nonlinear relationships if present. Adjustments for confounders were made using multivariable logistic regression and Cox models. RESULTS: A total of 7,932 subjects were included in the study. After adjusting for patient and surgical characteristics, LVEF remained associated with the primary outcome as well as the composite outcome of any postoperative adverse event. Both these relationships were best described by a J-shaped curve given that higher LVEF values were associated with increased risk, albeit not as high has lower values. Regarding long-term mortality, individuals with a preoperative LVEF of 60% demonstrated the longest survival. A statistically significant relationship was not found between LVEF and operative mortality or atrial fibrillation after adjustment for confounders. CONCLUSIONS: Higher preoperative LVEF values may be associated with increased risk for patients undergoing CABG surgery. Future studies are needed to better characterise this phenotype.

Association of four-factor prothrombin complex concentrate with subsequent plasma transfusion: A retrospective cohort study.

OBJECTIVE: To assess whether patients prescribed four-factor prothrombin complex concentrate (4FPC) received less plasma during the following 24-hour period than those treated for the same indications who received only plasma. INTRODUCTION: It is unclear whether 4FPC is associated with a reduction in subsequent plasma transfusion. This is important for minimising transfusion-associated risks and for inventory management. MATERIALS AND METHODS: We retrospectively studied patients treated for bleeding or coagulopathy. Individuals receiving 4FPC were matched by indication to patients treated with only plasma. Blood products received during 24-hour follow up were compared between 4FPC and plasma-only patients. RESULTS: There was no difference in the number of patients receiving additional plasma (19 (21%) 4FPC patients vs 31 (34%) plasma-only patients, P = .07) nor in the median number of additional plasma units received (0 units for both groups, interquartile range [0, 0] for 4FPC patients vs [0, 1] for plasma-only patients, P = .09). Subgroup analysis comparing patients who received 4FPC for on-label vs off-label indications found no difference in the number of patients receiving plasma nor in the median number of plasma units received. CONCLUSION: 4FPC was prescribed to a diverse set of patients, and administration was not associated with reduced plasma transfusion at our institution.

Feasibility of Anesthesiologist-Performed Preoperative Echocardiography for the Prediction of Postinduction Hypotension: A Prospective Observational Study.

PURPOSE: To determine if left ventricular or inferior vena cava (IVC) measurements are easier to obtain on point-of-care ultrasound by anesthesiologists in preoperative patients, and to assess the relationship between preoperative cardiac dimensions and hypotension with the induction of general anesthesia. METHODS: This prospective observational study was conducted at a large academic medical center. Sixty-three patients undergoing noncardiac surgeries under general anesthesia were enrolled. Ultrasound examinations were performed by anesthesiologists in the preoperative area. To ensure that hypotension represented both a relative and absolute decrease in blood pressure, both a mean arterial pressure (MAP) < 65 mmHg and a MAP decrease of >30% from preoperative value defined this outcome. RESULTS: Left ventricular measurements were more likely to be acquired than IVC measurements (97% vs. 79%). Subjects without adequate images to assess IVC collapsibility tended to have a higher body mass index (33.6 ± 5.5 vs. 28.5 ± 4.5, p=0.001). While high left ventricular end-diastolic diameter values were associated with a decreased odds of MAP < 65 mmHg (OR: 0.24, 95% CI: 0.07-0.83, p=0.023) or a MAP decrease of >30% from baseline alone (OR: 0.25, 95% CI: 0.07-0.83, p=0.023), the primary endpoint of both relative and absolute hypotension was not associated with preoperative left ventricular dimensions. CONCLUSIONS: Preoperative cardiac ultrasound may be a more reliable way for anesthesiologists to assess patients' volume status compared to ultrasound of the IVC, particularly for patients with a higher body mass index.

A Comparison of Sepsis-2 (Systemic Inflammatory Response Syndrome Based) to Sepsis-3 (Sequential Organ Failure Assessment Based) Definitions-A Multicenter Retrospective Study.

OBJECTIVES: Recently, the definition of sepsis has changed from a physiologic derangement (Sepsis-1 and -2) to organ dysfunction (Sepsis-3) based. We sought to determine the concordance between the different sepsis phenotypes and how that affected mortality. DESIGN: Retrospective, multicenter study. SETTING: Three academic medical centers. PATIENTS: 29,459 patients who had suspected infection, defined as obtaining blood cultures and receiving antibiotics: 18,183 (62%) had either Sepsis-2 or Sepsis-3. MEASUREMENTS AND MAIN RESULTS: Kappa was used to show agreement between phenotypes. Conditional logistic regression was used to create models of associations between factors and phenotypes and between factors and mortality. About 12,981 patients had Sepsis-2; 12,043 had Sepsis-3; and 6,841 patients had both Sepsis-2 and Sepsis-3. Fifty-three percent of Sepsis-2 patients also had Sepsis-3, whereas 57% of Sepsis-3 patients also had Sepsis-2. Agreement between the two phenotypes was poor: kappa = 0.213 ± 0.006. Mortality was 6% in patients with only Sepsis-2, 10% with only Sepsis-3, and 18% in patients who had both phenotypes. Combining the variables in Sepsis-2 and Sepsis-3 improved the discrimination (C-statistic = 0.742 ± 0.005, p < 0.001) of mortality. CONCLUSIONS: We found that Sepsis-2 and Sepsis-3-based sepsis diagnoses represent separate phenotypes with poor agreement. Patients who have both phenotypes are at increased risk of mortality compared with having either phenotype alone. Inclusion of both systemic inflammatory response syndrome and Sequential Organ Failure Assessment criteria in the same model improves the discrimination of mortality.

Early Detection of Heart Failure With Reduced Ejection Fraction Using Perioperative Data Among Noncardiac Surgical Patients: A Machine-Learning Approach.

BACKGROUND: Heart failure with reduced ejection fraction (HFrEF) is a condition imposing significant health care burden. Given its syndromic nature and often insidious onset, the diagnosis may not be made until clinical manifestations prompt further evaluation. Detecting HFrEF in precursor stages could allow for early initiation of treatments to modify disease progression. Granular data collected during the perioperative period may represent an underutilized method for improving the diagnosis of HFrEF. We hypothesized that patients ultimately diagnosed with HFrEF following surgery can be identified via machine-learning approaches using pre- and intraoperative data. METHODS: Perioperative data were reviewed from adult patients undergoing general anesthesia for major surgical procedures at an academic quaternary care center between 2010 and 2016. Patients with known HFrEF, heart failure with preserved ejection fraction, preoperative critical illness, or undergoing cardiac, cardiology, or electrophysiologic procedures were excluded. Patients were classified as healthy controls or undiagnosed HFrEF. Undiagnosed HFrEF was defined as lacking a HFrEF diagnosis preoperatively but establishing a diagnosis within 730 days postoperatively. Undiagnosed HFrEF patients were adjudicated by expert clinician review, excluding cases for which HFrEF was secondary to a perioperative triggering event, or any event not associated with HFrEF natural disease progression. Machine-learning models, including L1 regularized logistic regression, random forest, and extreme gradient boosting were developed to detect undiagnosed HFrEF, using perioperative data including 628 preoperative and 1195 intraoperative features. Training/validation and test datasets were used with parameter tuning. Test set model performance was evaluated using area under the receiver operating characteristic curve (AUROC), positive predictive value, and other standard metrics. RESULTS: Among 67,697 cases analyzed, 279 (0.41%) patients had undiagnosed HFrEF. The AUROC for the logistic regression model was 0.869 (95% confidence interval, 0.829-0.911), 0.872 (0.836-0.909) for the random forest model, and 0.873 (0.833-0.913) for the extreme gradient boosting model. The corresponding positive predictive values were 1.69% (1.06%-2.32%), 1.42% (0.85%-1.98%), and 1.78% (1.15%-2.40%), respectively. CONCLUSIONS: Machine-learning models leveraging perioperative data can detect undiagnosed HFrEF with good performance. However, the low prevalence of the disease results in a low positive predictive value, and for clinically meaningful sensitivity thresholds to be actionable, confirmatory testing with high specificity (eg, echocardiography or cardiac biomarkers) would be required following model detection. Future studies are necessary to externally validate algorithm performance at additional centers and explore the feasibility of embedding algorithms into the perioperative electronic health record for clinician use in real time.

Impact of the pre-illness lipid profile on sepsis mortality.

PURPOSE: To determine if baseline lipid levels contribute to the relationship between lipid levels during sepsis and outcomes. MATERIALS AND METHODS: We conducted a retrospective cohort study at a tertiary-care academic medical center. Multivariable logistic regression models were used to adjust for confounders. Both Systemic Inflammatory Response Syndrome (SIRS) and Sequential Organ Failure Assessment (SOFA) score-based definitions of sepsis were analyzed. MEASUREMENTS AND MAIN RESULTS: After adjusting for patient characteristics and severity of illness, baseline values for both low density lipoprotein (LDL) cholesterol and triglycerides were associated with mortality (LDL cholesterol odds ratio [OR] 0.44, 95% confidence interval [CI] 0.23-0.84, p = .013; triglyceride OR 0.54, 95% CI 0.37-0.78, p = .001) using a SIRS based definition of sepsis. An interaction existed between these two variables, which resulted in increased mortality with higher baseline low density lipoprotein (LDL) cholesterol values for individuals with triglycerides below 208 mg/dL and the opposite direction of association above this level (interaction OR 1.48, 95% CI 1.02-2.16, p = .039). When using a SOFA score-based definition, only triglycerides remained associated with the mortality (OR 0.55, 95% CI 0.35-0.86, p = .008). CONCLUSIONS: Baseline lipid values, particularly triglyceride concentrations, are associated with hospital mortality in septic patients.

Risk Factors, Etiologies, and Screening Tools for Sepsis in Pregnant Women: A Multicenter Case-Control Study.

BACKGROUND: Given the significant morbidity and mortality of maternal sepsis, early identification is key to improve outcomes. This study aims to evaluate the performance characteristics of the systemic inflammatory response syndrome (SIRS), quick Sequential [Sepsis-related] Organ Failure Assessment (qSOFA), and maternal early warning (MEW) criteria for identifying cases of impending sepsis in parturients. The secondary objective of this study is to identify etiologies and risk factors for maternal sepsis and to assess timing of antibiotics in patients diagnosed with sepsis. METHODS: Validated maternal sepsis cases during the delivery hospitalization from 1995 to 2012 were retrospectively identified at 7 academic medical centers in the United States and Israel. Control patients were matched by date of delivery in a 1:4 ratio. The sensitivity and specificity of SIRS, qSOFA, and MEW criteria for identifying sepsis were calculated. Data including potential risk factors, vital signs, laboratory values, and clinical management were collected for cases and controls. RESULTS: Eighty-two sepsis cases during the delivery hospitalization were identified and matched to 328 controls. The most common causes of sepsis were the following: chorioamnionitis 20 (24.4%), endometritis 19 (23.2%), and pneumonia 9 (11.0%). Escherichia coli 12 (14.6%), other Gram-negative rods 8 (9.8%), and group A Streptococcus 6 (7.3%) were the most commonly found pathogens. The sensitivities and specificities for meeting criteria for screening tools were as follows: (1) SIRS (0.93, 0.63); (2) qSOFA (0.50, 0.95); and (3) MEW criteria for identifying sepsis (0.82, 0.87). Of 82 women with sepsis, 10 (12.2%) died. The mortality rate for those who received antibiotics within 1 hour of diagnosis was 8.3%. The mortality rate was 20% for the patients who received antibiotics after >1 hour. CONCLUSIONS: Chorioamnionitis and endometritis were the most common causes of sepsis, together accounting for about half of cases. Notable differences were observed in the sensitivity and specificity of sepsis screening tools with the highest to lowest sensitivity being SIRS, MEW, and qSOFA criteria, and the highest to lowest specificity being qSOFA, MEW, and SIRS. Mortality was doubled in the cohort of patients who received antibiotics after >1 hour. Clinicians need to be vigilant to identify cases of peripartum sepsis early in its course and prioritize timely antibiotic therapy.