• induction agent
• nomogram
• post-intubation hypotension
• prognosis
• severe pneumonia

• work was supported by a grant from the Peking Union Medical Foundation - Rui E
• Special Fund for Emergency Medicine Research

• multi-drug resistant bacteria
• prediction model
• risk factors

• supported by The Applied Medical Research Project of Hefei Municipal Health Commission in 2022

• Intensive care unit
• Machine learning
• Mortality
• Pneumonia
• Severity scores

• Humans
• Pneumonia
• Hospital Mortality
• Intensive Care Units
• Machine Learning

• elastic net penalized Cox proportional hazard algorithm
• explainability
• herpes simplex virus
• pneumonia
• risk stratification

• CMRPG3M0851 Linkou Chang Gung Memorial Hospital
• CMRPG3L1011 Linkou Chang Gung Memorial Hospital
• 111-2320-B-182A-002-MY2 Ministry of Science and Technology

• Drug-resistant organism
• Early effective therapy
• Novel diagnostics
• Window

This study was performed to develop and validate machine learning models for early detection of ventilator-associated pneumonia (VAP) 24 h before diagnosis, so that VAP patients can receive early intervention and reduce the occurrence of complications.

This study was based on the MIMIC-III dataset, which was a retrospective cohort. The random forest algorithm was applied to construct a base classifier, and the area under the receiver operating characteristic curve (AUC), sensitivity and specificity of the prediction model were evaluated. Furthermore, We also compare the performance of Clinical Pulmonary Infection Score (CPIS)-based model (threshold value ≥ 3) using the same training and test data sets.

In total, 38,515 ventilation sessions occurred in 61,532 ICU admissions. VAP occurred in 212 of these sessions. We incorporated 42 VAP risk factors at admission and routinely measured the vital characteristics and laboratory results. Five-fold cross-validation was performed to evaluate the model performance, and the model achieved an AUC of 84% in the validation, 74% sensitivity and 71% specificity 24 h after intubation. The AUC of our VAP machine learning model is nearly 25% higher than the CPIS model, and the sensitivity and specificity were also improved by almost 14% and 15%, respectively.

We developed and internally validated an automated model for VAP prediction using the MIMIC-III cohort. The VAP prediction model achieved high performance based on its AUC, sensitivity and specificity, and its performance was superior to that of the CPIS model. External validation and prospective interventional or outcome studies using this prediction model are envisioned as future work.

The online version contains supplementary material available at 10.1186/s12890-022-02031-w.

• MIMIC database
• Predictive modeling
• Risk factors
• Ventilator-associated pneumonia

Background

Stenotrophomonas maltophilia is a rapidly emerging nosocomial pathogen with intrinsic or acquired resistance mechanisms to several antibiotic classes. It can cause life-threatening opportunistic pneumonia, particularly among hospitalized patients. Incidence of infections by S. maltophilia has been reported as 0.07-0.4% of hospital discharges, but its mortality is 20 -60%. This is the first study from Qatar indexing the clinical and epidemiological characteristics and antibiotic susceptibility of S. maltophilia.

Materials and methods

This retrospective descriptive epidemiological study was conducted in 6 tertiary care hospitals under Hamad Medical Corporation in Doha, Qatar, analyzing inpatient respiratory isolates of S. maltophilia during 2016-17. Out-patients, children below 14 years, and non-respiratory samples except blood cultures in patients with pneumonia were excluded. Clinical records were reviewed to identify possible risk factors. Infection and colonization were identified using the Centers for Disease Control and Prevention (CDC) algorithm for clinically defined pneumonia and statistically analyzed using the chi-square test and Pearson's correlation.

Results

S. maltophilia was isolated from 2.07% (317/15312) of all respiratory samples received in the microbiology lab during our study period. Three hundred seventeen patients studied had a mean age of 60 ± 20 years, and 68% were men. Most of the isolates were from sputum (179), followed by tracheal aspirate (82) and bronchoscopy (42). Fourteen blood culture samples from patients diagnosed with pneumonia were also included. 67% were hospitalized for more than two weeks, 39.1% were on mechanical ventilators, and 88% had received a broad-spectrum antibiotic before the event. 29.1% were deemed to have an infection and 70.9% colonization. Incidence of infection in those with Charlson’s Co-morbidity Index (CCI) ≥ 3 was 36.5% compared to 24.2% in those with CCI < 3 (Relative Risk (RR)=1.52; 95% CI: 1.04,2.18; p=0.01). Patients with recent chemotherapy, immunosuppressant, or steroid use had a significantly higher infection risk than those without (69.2% v/s 23.3% RR=2.96; 95% CI:2.2,3.9; p<0.005). The most common symptoms in patients with infection were fever (96%) and expectoration (61.9%). The most common radiological finding was lobar consolidation (71.6%). Mean CRP and procalcitonin were 106.5±15.5 mg/l and 12.3 ± 14 ng/ml. Overall mortality was 16.3%. Patients on mechanical ventilator with IBMP-10 score ≥ 2 had 22.8% mortality compared to 5.7% in those with score < 2 (RR=3.9;95%CI:0.9,16.6; p<0.015). As per The US Clinical and Laboratory Standards Institute (CSLI) breakpoint values, Trimethoprim-Sulfamethoxazole (TMP-SMX) showed the highest sensitivity (97.8%), followed by levofloxacin (71.6%). 0.3% of samples were pan-drug resistant.

Conclusions

S. maltophilia is a frequent nosocomial colonizer, but it can cause nosocomial pneumonia in almost one-third of cases, specifically in immunocompromised and patients with CCI ≥ 3 with a high risk of mortality due to ventilator-associated pneumonia (VAP) in those with IBMP-10 ≥ 2. Prolonged hospital stay is a risk factor for colonization by S. maltophilia, while recent chemotherapy, immunosuppressant, or steroid use are risk factors for hospital-acquired pneumonia due to S. maltophilia. TMP-SMX and levofloxacin are the only reliable agents for monotherapy of respiratory infections due to S. maltophilia in Qatar.

Ventilator-associated pneumonia (VAP) is the most common nosocomial infection. Red cell distribution width (RDW) and neutrophil-lymphocyte ratio (NLR) are prognostic factors to mortality in different diseases. The aim of this study is to evaluate prognostic efficiency RDW, NLR, and the Sequential Organ Failure Assessment (SOFA) score for mortality prediction in respiratory patients with VAP.

One hundred thirty-six patients mechanically ventilated and developed VAP were included. Clinical characteristics and SOFA score on the day of admission and at diagnosis of VAP, RDW, and NLR were assessed and correlated to mortality. The average age of patients was 58.80 ± 10.53. These variables had a good diagnostic performance for mortality prediction AUC 0.811 for SOFA at diagnosis of VAP, 0.777 for RDW, 0.728 for NLR, and 0.840 for combined of NLR and RDW. The combination of the three parameters demonstrated excellent diagnostic performance (AUC 0.889). A positive correlation was found between SOFA at diagnosis of VAP and RDW (r = 0.446, P < 0.000) and with NLR (r = 0.220, P < 0.010).

NLR and RDW are non-specific inflammatory markers that could be calculated quickly and easily via routine hemogram examination. These markers have comparable prognostic accuracy to severity scores. Consequently, RDW and NLR are simple, yet promising markers for ICU physicians in monitoring the clinical course, assessment of organ dysfunction, and predicting mortality in mechanically ventilated patients. Therefore, this study recommends the use of blood biomarkers with the one of the simplest ICU score (SOFA score) in the rapid diagnosis of critical patients as a daily works in ICU.

• Advance Directives
• Ageism
• Decision Making
• Humans
• Pandemics
• Resource Allocation

• Hospital-acquired pneumonia
• hospital infections
• hospital mortality
• ventilator-associated pneumonia

• Anti-Bacterial Agents/therapeutic use
• Antimicrobial Stewardship/methods
• Biomarkers/analysis
• Bronchoalveolar Lavage/methods
• Bronchoalveolar Lavage Fluid/microbiology
• Female
• Humans
• Male
• Middle Aged
• Pneumonia, Ventilator-Associated/drug therapy
• Pneumonia, Ventilator-Associated/microbiology
• Process Assessment, Health Care
• State Medicine
• United Kingdom

• IS-BRC-1215-20001 Department of Health
• MR/L004933/1 Medical Research Council
• MR/L004933/2 Medical Research Council
• HICF-510-078 Department of Health
• MR/N02995X/1 Medical Research Council
• Wellcome Trust
• 094949/Z/10/X Wellcome Trust
• 205214/Z/16/Z Wellcome Trust

• Critical care
• Healthcare–associated infections
• Risk factors
• Surveillance
• Time-to-event models

To evaluate the predictive capability of respiratory mechanics for the development of ventilator-associated pneumonia (VAP) and mortality in the intensive care unit (ICU) of a hospital in southern Brazil.

A cohort study was conducted between, involving a sample of 120 individuals. Static measurements of compliance and resistance of the respiratory system in pressure-controlled ventilation (PCV) and volume-controlled ventilation (VCV) modes in the 1^st and 5^th days of hospitalization were performed to monitor respiratory mechanics. The severity of the patients’ illness was quantified by the Acute Physiology and Chronic Health Evaluation II (APACHE II). The diagnosis of VAP was made based on clinical, radiological and laboratory parameters.

The significant associations found for the development of VAP were APACHE II scores above the average (P = 0.016), duration of MV (P = 0.001) and ICU length of stay above the average (P = 0.003), male gender (P = 0.004), and worsening of respiratory resistance in PCV mode (P = 0.010). Age above the average (P < 0.001), low level of oxygenation on day 1 (P = 0.003) and day 5 (P = 0.004) and low lung compliance during VCV on day 1 (P = 0.032) were associated with death as the outcome.

The worsening of airway resistance in PCV mode indicated the possibility of early diagnosis of VAP. Low lung compliance during VCV and low oxygenation index were death-related prognostic indicators.

• Respiratory mechanics
• Respiratory tract infection
• Ventilator-associated pneumonia

• Area under the curve
• Intensive care unit
• Mortality
• Receiver operator characteristic curve
• Severity score
• Ventilator-associated pneumonia

VAP is defined as pneumonia in patients who use ventilators. The acute physiology and chronic health evaluation (APACHE II) scoring system was originally developed for predicting mortality in patients who were admitted to the intensive care unit. Due to the complexity, a simpler score called IBMP-10 was developed. We designed the study to confirm and further investigate these two methods.

This cross-sectional and analysis-descriptive study was done at the moment of VAP diagnosis on 60 patients in intensive care units. APACHE II and the IBMP-10 scores were calculated. ROC curves were generated to compare the new prediction rule with the APACHE II score. Results were reported as adjusted odds ratios with 95% confidence intervals (CIs). Analyses were performed using SPSS, version 20 and P values of 0.05 were considered to be statistically significant.

APACHE II Score means (P < 0.001) and IBMP-10 score (P < 0.001) means had significant increase in Non-survivor patient than in patients who survived. APACHE II can be used as a good prediction measure for mortality rate. In IBMP-10 method, specificity and PPV were greater than APACHE II, but in mc-nemar test, there was no significant difference between the two methods (P = 0.55). Both prediction rules had high NPV. In our study, survivors’ prediction value in APACHE II was 46.7%, and in IBMP-10, it was 46.7%.

IBMP-10, compared to APACHE II, has greater sensitivity, specificity, and AUC to predict mortality. So the consequence of the use of IBMP-10 was better than APACHE II.

• Acute Physiology and Chronic Health Evaluation II
• blood pressure
• immunodeficiency
• mortality
• multilobar infiltrates
• platelet count and 10-day hospitalization-10
• scoring system
• ventilator-associated pneumonia

• APACHE II
• CPIS
• Mortality
• Ventilator-associated pneumonia

Purpose. The long pentraxin 3 (PTX3) is a key component of the humoral arm of the innate immune system. PTX3 is produced locally in response to proinflammatory stimuli. We reviewed the usefulness of systemic levels of PTX3 in critically ill patients with systemic inflammatory response syndrome (SIRS), sepsis, and bacteremia, focusing on its diagnostic and prognostic value. Methods. A PubMed search on PTX3 was conducted. The list of papers was narrowed to original studies of critically ill patients. Eleven papers on original studies of critically ill patients that report on PTX3 in SIRS, sepsis, or bacteremia were identified. Results. Systematic levels of PTX3 have little diagnostic value in critically ill patients with SIRS, sepsis, or bacteremia. Systemic levels of PTX3, however, have superior prognostic power over other commonly used biological markers in these patients. Systemic levels of PTX3 correlate positively with markers of organ dysfunction and severity-of-disease classification system scores. Finally, systemic levels of PTX3 remain elevated in the acute phase and decreased on recovery. Notably, the age of the patients and underlying disease affect systemic levels of PTX3. Conclusions. The diagnostic value of PTX3 is low in patients with sepsis. Systemic levels of PTX3 have prognostic value and may add to prognostication of patients with SIRS or sepsis, complementing severity-of-disease classification systems and other biological markers.

Systemic levels of soluble urokinase-type plasminogen activator receptor (suPAR) positively correlate with the activation level of the immune system. We reviewed the usefulness of systemic levels of suPAR in the care of critically ill patients with sepsis, SIRS, and bacteremia, focusing on its diagnostic and prognostic value.

A PubMed search on suPAR was conducted, including manual cross-referencing. The list of papers was narrowed to original studies of critically ill patients. Ten papers on original studies of critically ill patients were identified that report on suPAR in sepsis, SIRS, or bacteremia.

Systematic levels of suPAR have little diagnostic value in critically ill patients with sepsis, SIRS, or bacteremia. Systemic levels of suPAR, however, have superior prognostic power over other commonly used biological markers in these patients. Mortality prediction by other biological markers or severity-of-disease classification system scores improves when combining them with suPAR. Systemic levels of suPAR correlate positively with markers of organ dysfunction and severity-of-disease classification system scores. Finally, systemic levels of suPAR remain elevated for prolonged periods after admission and only tend to decline after several weeks. Notably, the type of assay used to measure suPAR as well as the age of the patients and underlying disease affect systemic levels of suPAR.

The diagnostic value of suPAR is low in patients with sepsis. Systemic levels of suPAR have prognostic value, and may add to prognostication of patients with sepsis or SIRS complementing severity-of-disease classification systems and other biological markers.