Nosocomial acquisition of influenza is associated with significant morbidity and mortality: Results of a prospective observational study

Constructing a screening model to identify patients at high risk of hospital-acquired influenza on admission to hospital

Objective

To develop a machine learning (ML)-based admission screening model for hospital-acquired (HA) influenza using routinely available data to support early clinical intervention.

Methods

The study focused on hospitalized patients from January 2021 to May 2024. The case group consisted of patients with HA influenza, while the control group comprised non-HA influenza patients admitted to the same ward in the HA influenza unit within 2 weeks. The 953 subjects were divided into the training set and the validation set in a 7:3 ratio. Feature screening was performed using least absolute shrinkage and selection operator (LASSO) and the Boruta algorithm. Subsequently eight ML algorithms were applied to analyze and identify the optimal model using a 5-fold cross-validation methodology. And the area under the curve (AUC), area under the precision-recall curve (AP), F1 score, calibration curve and decision curve analysis (DCA) were applied to comprehensively assess the predictive effectiveness of the selected models. Feature factors were selected and feature importance's were assessed using SHapley's additive interpretation (SHAP). Furthermore, an interactive web-based platform was additionally developed to visualize and demonstrate the predictive model.

Results

Age, pneumonia on admission, Chronic renal failure, Malignant tumor, hypoproteinemia, glucocorticoid use, admission to ICU, lymphopenia, BMI were identified as key variables. For the eight ML algorithms, ROC values ranging from 0.548 to 0.812 were observed in the validation set. A comprehensive analysis showed that the XGBoost model predicted the highest accuracy (AUC: 0.812) with an F1 score of 0.590 and the highest A p value (0.655). Evaluating the optimal model, the AUC values were 0.995, 0.826, and 0.781 for the training, validation and test sets. The XGBoost model showed strong robust. SHapley's additive interpretation (SHAP) was utilized to analyze the contribution of explanatory variables to the model and their correlation with HA influenza. In addition, we developed a practical online prediction tool to calculate the risk of HA influenza occurrence.

Conclusion

Based on the routine data, the XGBoost model demonstrated excellent calibration among all ML algorithms and accurately predicted the risk of HA influenza, thereby serving as an effective tool for early screening of HA influenza.

Systematic Molecular Influenza A/B Screening Upon Hospital Admission in Belgium, January-April 2022: Positivity Ratios and Viral Loads According to Symptomatology, Age, and Vaccination Status

Three hospitals implemented molecular point-of-care tests (POCTs) to screen patients for SARS-CoV-2 infection upon admission during the 2021/2022 influenza season, which in Belgium lasted from January to April 2022. The samples were simultaneously tested for influenza A/B. Influenza positivity at admission was examined in relation to patient characteristics and symptomatology. Influenza POCTs were performed on all patients requiring urgent hospitalization, regardless of the admission reason. A total of 9327 patients were included in the study, of which 411 (4.4%) tested positive for influenza A/B. Asymptomatic infection and mild illness accounted for respectively 11.2% (95% CI: 8.5%-14.6%), and 43.3% (95% CI: 38.6%-48.1%) of the cases. A total of 66% (95% CI: 60%-72%) of all patients in these symptom categories (asymptomatic and mild illness) showed a high viral load (cycle threshold [Ct] < 24). Only in 30 (7.3%, 95% CI: 5.2%-10.2%) of all cases and in two (4.4%, 95% CI: 1.2%-14.5%) of the asymptomatic cases, the symptomatology worsened during hospital stay. Coinfections with both influenza and SARS-CoV-2 occurred in 35 patients (8.5% of all influenza positive patients). There was no difference in symptomatology between patients with co-infections and those with an influenza mono-infection. Patients could not be reliably categorized into carriers with low versus high viral loads based on symptomatology, age, and vaccination status. More than half of the influenza-positive individuals were either asymptomatic or had mild symptoms upon admission, while often carrying high viral loads. Our results show that without screening of patients at hospital admission, a considerable number of patients with a high viral load may be incorrectly classified as being not infectious.

Testing and Masking Policies and Hospital-Onset Respiratory Viral Infections

This cohort study examines the ratio between hospital- and community-onset respiratory viral infections at different levels of testing and masking from 2020 to 2024.

In Support of Universal Admission Testing for SARS-CoV-2 During Significant Community Transmission

Many hospitals have stopped or are considering stopping universal admission testing for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). We discuss reasons why admission testing should still be part of a layered system to prevent hospital-acquired SARS-CoV-2 infections during times of significant community transmission. These include the morbidity of SARS-CoV-2 in vulnerable patients, the predominant contribution of presymptomatic and asymptomatic people to transmission, the high rate of transmission between patients in shared rooms, and data suggesting surveillance testing is associated with fewer nosocomial infections. Preferences of diverse patient populations, particularly the hardest-hit communities, should be surveyed and used to inform prevention measures. Hospitals' ethical responsibility to protect patients from serious infections should predominate over concerns about costs, labor, and inconvenience. We call for more rigorous data on the incidence and morbidity of nosocomial SARS-CoV-2 infections and more research to help determine when to start, stop, and restart universal admission testing and other prevention measures.

Influenza A subtype H3N2 is associated with an increased risk of hospital dissemination - an observational study over six influenza seasons

BACKGROUND

Previous studies on hospital-acquired influenza (HAI) have not systematically evaluated the possible impact of different influenza subtypes. HAI has historically been associated with high mortality, but clinical consequences may be less severe in a modern hospital setting.

AIMS

To identify and quantify HAI for each season, investigate possible associations with varying influenza subtypes, and to determine HAI-associated mortality.

METHODS

All influenza-PCR-positive adult patients (>18 years old) hospitalized in Skåne County during 2013-2019, were prospectively included in the study. Positive influenza samples were subtyped. Medical records of patients with suspected HAI were examined to confirm a nosocomial origin and to determine 30-day mortality.

RESULTS

Of 4110 hospitalized patients with a positive influenza PCR, 430 (10.5%) were HAI. Influenza A(H3N2) infections were more often HAI (15.1%) than influenza A(H1N1)pdm09, and influenza B (6.3% and 6.8% respectively, P<0.001). The majority of HAI caused by H3N2 were clustered (73.3 %) and were the cause of all 20 hospital outbreaks consisting of ≥4 affected patients. In contrast, the majority of HAI caused by influenza A(H1N1)pdm09 and influenza B were solitary cases (60% and 63.2%, respectively, P<0.001). Mortality associated with HAI was 9.3% and similar between subtypes.

CONCLUSIONS

HAI caused by influenza A(H3N2) was associated with an increased risk of hospital dissemination. Our study is relevant for future seasonal influenza infection control preparedness and shows that subtyping of influenza may help to define relevant infection control measures. Mortality in HAI remains substantial in a modern hospital setting.

Comparison of 28-Day Mortality Between Hospital- and Community-Acquired Influenza Patients

Objective

This study aimed to compare 28-day mortality between patients have hospital-acquired influenza (HAI) and those have community-acquired influenza (CAI) during the 2017-2019 influenza seasons in a tertiary care center in İstanbul, Türkiye.

Materials and Methods

This retrospective cohort included all hospitalized patients who had confirmed influenza infection and were over 17 years old. HAI was defined as a case of influenza that tested negative in a PCR test or had no signs of influenza on admission but with a positive test result at any point after 72 hours of admission. CAI was defined as a case of influenza diagnosed within 72 hours of admission or before admission. The primary outcome was 28-day mortality after diagnosis. Biological sex, admission to the intensive care unit (ICU), presence of chronic obstructive pulmonary disease, cardiovascular and immunosuppressive comorbidities, influenza subtype, and other variables identified with univariate analyses (p<0.25) were entered into logistic regression analysis.

Results

During the study period, 92 (46%) of 201 hospitalized patients who tested positive for influenza were identified as HAI, and the rest (109) were identified as CAI. Univariate analysis showed no differences between survivors and non-survivors in patient characteristics, except non-survivors were more likely to have an ICU admission. The multivariable logistic regression analysis results showed that HAI was associated with 5.6-fold increased odds of mortality (95% confidence interval [CI]=1.6-19.3; p=0.006), after adjustment for age, gender, comorbidity, and ICU admission.

Conclusion

The results of this study revealed that patients who had HAI were more likely to die within 28 days compared to those who had CAI after controlling for key confounders. The high rate of HAI underscores the critical importance of robust infection control measures for hospital-acquired viral infections. Additional research and targeted interventions are necessary to improve the HAI prognosis.

Feasibility and clinical utility of local rapid Nanopore influenza A virus whole genome sequencing for integrated outbreak management, genotypic resistance detection and timely surveillance

Rapid respiratory viral whole genome sequencing (WGS) in a clinical setting can inform real-time outbreak and patient treatment decisions, but the feasibility and clinical utility of influenza A virus (IAV) WGS using Nanopore technology has not been demonstrated. A 24 h turnaround Nanopore IAV WGS protocol was performed on 128 reverse transcriptase PCR IAV-positive nasopharyngeal samples taken over seven weeks of the 2022-2023 winter influenza season, including 25 from patients with nosocomial IAV infections and 102 from patients attending the Emergency Department. WGS results were reviewed collectively alongside clinical details for interpretation and reported to clinical teams. All eight segments of the IAV genome were recovered for 97/128 samples (75.8 %) and the haemagglutinin gene for 117/128 samples (91.4 %). Infection prevention and control identified nosocomial IAV infections in 19 patients across five wards. IAV WGS revealed two separate clusters on one ward and excluded transmission across different wards with contemporaneous outbreaks. IAV WGS also identified neuraminidase inhibitor resistance in a persistently infected patient and excluded avian influenza in a sample taken from an immunosuppressed patient with a history of travel to Singapore which had failed PCR subtyping. Accurate IAV genomes can be generated in 24 h using a Nanopore protocol accessible to any laboratory with SARS-CoV-2 Nanopore sequencing capacity. In addition to replicating reference laboratory surveillance results, IAV WGS can identify antiviral resistance and exclude avian influenza. IAV WGS also informs management of nosocomial outbreaks, though molecular and clinical epidemiology were concordant in this study, limiting the impact on decision-making.

Patient influenza vaccination reduces the risk of hospital-acquired influenza: An incident test negative-case control study in Lyon university hospital, France (2004-2020)

BACKGROUND

Literature is limited on the impact of patient vaccination on the risk of hospital-acquired influenza (HAI). This test negative case-control study nested in a surveillance program aimed at evaluating the effectiveness of influenza vaccination in reducing the risk of HAI in hospitalized patients during 15 influenza seasons (2004-05 to 2019-20).

METHODS

HAI cases were those who developed influenza like illness (ILI) symptoms at least 72 h after hospitalization and had a positive reverse transcriptase-polymerase chain reaction (RT-PCR). Controls were those with ILI symptoms and a negative RT-PCR test. A nasal swab as well as socio-demographic, clinical data and information on influenza vaccination were collected.

RESULTS

Of the 296 patients included, 67 were confirmed HAI cases. Influenza vaccine coverage was significantly higher among controls compared to HAI cases (p = 0.002). The risk of HAI was reduced by almost 60 % in vaccinated patients.

CONCLUSIONS

A better control of HAI can be achieved by vaccinating hospitalized patients.