Welcome to the ESPID 2022 Meeting Calendar

Backgrounds:

Infectious diseases play a significant role in the global burden of disease. The gold standard for the diagnosis of bacterial infection, culture of bacteria, can lead to diagnostic delays and unnecessary antibiotic use. The advent of high-throughput microarray and sequencing has led to the discovery of host-based genomic biomarkers, capable of differentiating bacterial from other causes of infection but few have achieved validation for use in a clinical setting.

Methods

A systematic review was performed. PubMed/Ovid Medline, Ovid Embase and Scopus databases were searched for relevant studies from inception up to 21/09/2020 with forward and backward citation searching of key references. Studies which compared the diagnostic performance of host genomic biomarkers of bacterial infection to those with non-bacterial sources of infection were included. Study selection and assessment of quality was conducted by two independent reviewers. Meta-analysis was undertaken for all included genomic signatures using a diagnostic random-effects model. The review was registered with PROSPERO (ID: CRD42021208462).

Results:

Sixty-eight studies which evaluated the performance of 110 biomarkers in 15,299 patients were included. Forty-seven studies examined the performance of biomarkers specific to TB infection and twenty studies were conducted in a paediatric population. The results of pooled sensitivity, specificity, negative and positive likelihood ratio and diagnostic odds ratio of genomic biomarkers of bacterial infection were 0.81 (95% CI 0.78 to 0.83), 0.86 (95% CI 0.84 to 0.88), 0.18 (95% CI 0.15 to 0.21), 5.7 (95% CI 5.0 to 6.5), 31.3 (95% CI 25 to 39), respectively. Significant heterogeneity (I2 77%) was present.

Conclusions/Learning Points:

Host derived genomic biomarkers show significant potential for clinical use as diagnostic tests of bacterial infection however, further validation and attention to test platform is warranted before clinical implementation can be achieved.

Backgrounds:

The ongoing COVID-19 pandemic has led to over five million deaths worldwide highlighting an unprecedented need for rapid diagnostic screening. The gold standard for COVID-19 diagnosis is the collection of a nasopharyngeal swab subsequently processed with an RNA extraction kit requiring electricity and expensive laboratory equipment. Therefore, the diagnosis of COVID-19 in low- and middle-income countries (LMIC) is rarely achievable at a point-of-care (POC) and instead relegated to remote centralized laboratories. To address this need, our team has developed an innovative, rapid and easy to use sample preparation method for RNA extraction allowing for true POC application.

Methods

The SmartLid^TM extraction method utilizes a custom 3D printed magnetic lid, designed to work with standard Eppendorf tubes, to transfer magnetic nanoparticles and attached RNA through three sample preparation steps. This is in contrast to all other manual extraction methods, which require expensive micropipettes, lab training and electrical power. The whole extraction process is performed within five minutes providing pure RNA adequate for downstream applications.

Results:

A total of 410 nasopharyngeal swabs has been tested (including 150 COVID-positive subjects). All clinical isolates were extracted by the SmartLid and the gold standard QIAmp Viral RNA methods and tested by the CDC RT-qPCR assays. The SmartLid method achieved 93.9% sensitivity and 99.5% specificity compared to the QIAmp Viral RNA showing equivalent performance.

Conclusions/Learning Points:

The extraction method presented can compete favourably with conventional laboratory-based extraction techniques which require expensive equipment and electricity, and which incur delays of over 45 minutes. The method has received overwhelmingly positive feedback from collaborators, who have tested it in CAT3 laboratory environments, and from visiting clinicians with experience in LMIC diagnostics. Thus, there is a clear interest for implementation in a domestic clinical/laboratory setting (NHS),and LMIC remote point-of-care setting.

Backgrounds:

Discriminating between viral and bacterial lower respiratory tract infection (LRTI) in children is challenging, leading to an excessive use of antibiotics. Myxovirus resistance protein A (MxA) is a promising biomarker for viral infections. The aim of the study was to assess the difference in blood MxA levels between children with viral and bacterial LRTI and to assess MxA levels in relation to specific respiratory viruses.

Methods

Children with lower respiratory tract infection (LRTI) were enrolled as cases at Sachs’ Children and Youth Hospital, Stockholm, Sweden. Nasopharyngeal aspirates (for respiratory PCR analysis) and blood samples (for analysis of MxA and CRP) were systematically collected from all study subjects in addition to standard laboratory/radiology assessment. Aetiology was defined according to an algorithm based on laboratory and radiological findings. The diagnostic accuracy of MxA was assessed by calculating sensitivity, specificity and area under the curve (AUC) in receiving operator characterstic (ROC) curves.

Results:

Of the 326 cases, 242 had viral aetiology, 11 had mixed viral-bacterial aetiology, 5 had bacterial aetiology, 2 had atypical bacterial aetiology, and 66 cases had undetermined aetiology. MxA levels were higher in children with viral LRTI as compared with bacterial LRTI (p<0.01, AUC 0.92). In the subgroup of children with pneumonia diagnosis, a cut-off of MxA 430µg/l discriminated between viral and bacterial aetiology with 93% sensitivity and 100% specificity (AUC 0.98). The highest MxA levels were seen in cases PCR positive for adenovirus and respiratory syncytial virus (median MxA 1961µg/l and 1226µg/l respectively).

Conclusions/Learning Points:

MxA accurately discriminated between viral and bacterial etiology in children with LRTI, in particular in the group of children with pneumonia diagnosis. Thus, MxA determination might improve rational use of antibiotics in this patient group.

Backgrounds:

Late-onset neonatal sepsis (LONS) is a major complication in preterm neonates. Early recognition, by means of heart rate variability (HRV) monitoring, could help to guide early therapy and thereby improve outcome. The aim of this study was to investigate the association between the implementation of a local HRV-monitoring guideline in a level-IV NICU on mortality, measures of sepsis severity, frequency of sepsis testing and antibiotic usage among very preterm neonates.

Methods

In January 2018 a local guideline was implemented for early detection of LONS using HRV-monitoring combined with determination of inflammatory biomarkers. Data on all patients admitted with a gestational age at birth of <32 weeks were reviewed in the period January 2016-June 2020 (n=1,135; pre-implementation period Jan 2016-Dec 2017 (n=515), and post-implementation period Jan 2018-Jun 2020 (n=620)).

Results:

In the study period, a total of 811 blood cultures in 473 neonates were withdrawn. Of these episodes, 490 (60.4%) were classified as sepsis. In the pre-implementation period, death within 10 days of start of the sepsis episode occurred in 39 (10.3%) episodes and in the post-implementation period it occurred in 34 (7.6%) episodes (P = 0.21). The nSOFA course during a sepsis episode was significantly lower in the post-implementation group (P = 0.01). We observed no statistically significant difference in number of blood cultures drawn and in antibiotic usage between the two periods.

Conclusions/Learning Points:

Implementing HRV-monitoring with determination of inflammatory biomarkers might help identify patients with sepsis sooner, resulting in improved outcome, without an increased use of antibiotics or blood cultures withdrawals.

Backgrounds:

Microscopy is the gold standard for malaria diagnosis but is time intensive and often requires repeat hospital visits. Rapid diagnostic tests (RDT) form the mainstay of diagnosis in many endemic areas. Could an RDT alone rule out imported malaria in children presenting to UK Emergency Departments (ED)?

Methods

UK-based, multi-centre, retrospective, diagnostic accuracy study. Cases: any child <16 years presenting to ED with history of fever and travel to a malaria endemic area, between 01/01/2016 and 31/12/2017. Diagnosis: microscopy for malarial parasites (clinical reference standard) and RDT (index test). UK Health Research Authority approval: 20/HRA/1341.

Results:

51 malaria cases were reported in 1,472 patients documented at 15 sites (prevalence 3.5%), of which 44% female, median age 5 years (interquartile range 2-9 years). There were two deaths but not from malaria. Sensitivity of RDT alone to detect malaria infection was 94.1% (95% CI 83.8%-98.8%), specificity 99.4% (95% CI 98.8-99.7%), positive predictive value (PPV) 84.2% (95% CI 72.1-92.5%) and negative predictive value (NPV) 99.8% (95% CI 99.4-100.0%) (see Figure 1).

Figure 1. Personogram showing expected numbers of RDT results in hypothetical sample of 1000 children (all malaria species).

Forty (78%) cases were due to P falciparum. Sensitivity of RDT alone to detect P falciparum was 100% (95% CI 91.2-100%), specificity 98.8% (95% CI 98.1-99.3%), PPV 70.2% (95% CI 56.6%-81.6%) and NPV 100% (95% CI 99.7-100%).

Conclusions/Learning Points:

Standard RDTs were 100% sensitive in detecting P falciparum malaria in children in this study, with a lower point estimate for all malaria species. RDT alone can likely rule out imported P falciparum malaria in well-appearing children but a prospective study should confirm this, especially with the emergence of pfhrp2/3 gene deletions in the P falciparum parasite.

Backgrounds:

Differentiating bacterial from viral etiologies of pediatric community-acquired pneumonia (CAP) is critical to guiding appropriate therapy. However, current tests are insensitive, invasive, or impractical in children. The objective of this study is to identify candidate metabolomic biomarkers that differentiate bacteria from viral CAP.

Methods

We studied a cohort of children, 3 months-18 years old, with suspected CAP in the emergency department. Patients with chronic medical conditions or who were hospitalized 14 days prior were excluded. Viral and Mycoplasma pneumoniae (Mp) were detected by PCR of nasopharyngeal swabs. Suspected Streptococcus pneumoniae (Sp) was defined as presence of pneumococcal autolysin (lytA) and a procalcitonin of ≥1.5 ng/mL. Urine samples were collected at time of presentation and metabolites were identified and quantified by nuclear magnetic resonance spectroscopy. Metabolomics data were standardized using specific gravity. Demographic and clinical characteristics by patient status ( MP, Sp and viral) were compared using chi-square tests and ANOVA, as applicable. Random forest (RF) was used to determine the most important metabolites and clinical factors to discriminate viral etiology from Mp and Sp.

Results:

Of 160 children, 28 (17.5%) had Mp, 13 (8.1%) had Sp, and 119 (74.4%) had a virus detected by PCR (Table). Most (87%) were between 1-12 years old. The most important variables identified by RF included age, prior history of reactive airways disease, 1-methylnicotinamide, hypoxanthine, tryptophan, quinolinate, valine, trimethylamine-N-oxide, ascorbate, and 4-hydroxybenzoate.

Conclusions/Learning Points:

Metabolites related to inflammatory pathways (e.g., tryptophan, quniolinate) and microbial metabolism (e.g., trimethylamine-N-oxide, 4-hydroxybenzoate) differentiated viral from typical and atypical bacterial CAP in children. Urine metabolomics can identify novel biomarkers to differentiate bacterial from viral CAP in children.

Backgrounds:

Up to 20% of febrile pediatric patients presenting to the emergency department (ED) will not have a source identified by history or physical examination. Most of these patients have self-limiting viral illnesses, but ~10% could be bacterially infected. A host-protein bacterial likelihood score (BV score) based on TRAIL, IP-10 and CRP has demonstrated high performance for differentiating bacterial from viral infections in multiple validation studies. Here we evaluate for the first time BV Score's performance when applied according to instructions for use in routine care of children presenting to the ED with fever without source (FWS).

Methods

A retrospective analysis of patients aged 3 months to 18 years at two medical centers for whom BV score was measured as part of routine care (NCT03075111; 2014–2017). For each patient, the physician documented suspected clinical syndrome at the time of blood draw for BV score. TRAIL and IP-10 were measured, and BV score calculated using ImmunoXpert™. CRP was measured using COBASc501. Reference standard for infection etiology was adjudicated by 3 independent experts based on the patient’s clinical, laboratory and microbiologic data.

Results:

2160 of 3006 patients met the current indication for use for measuring the BV score, of whom 788 were documented by the physician as suspected FWS; 69 patients were adjudicated as bacterial, 518 as viral and 201 as indeterminate. Median age was 2 years (IQR 4.08), 53% were male. The BV score attained sensitivity of 88.1% (95% CI, 77.1%-95.1%), specificity of 93.7% (91.1%-95.8%) and NPV of 98.4% (96.8%-99.2%). The equivocal rate was 13.1%.

Conclusions/Learning Points:

The BV score demonstrated high diagnostic performance when applied according to it indication for use as part of routine care for children presenting to the ED with FWS.

Backgrounds:

The monocyte-to-lymphocyte ratio (MLR) and neutrophil-to-lymphocyte ratio (NLR) are easy to obtain markers from full blood counts. Little is known about the diagnostic accuracy of these ratios in children evaluated for tuberculosis (TB) compared to sick controls.

Methods

Data of two prospective multicentre studies in Switzerland were used: the CITRUS study and the ProPAED study. The CITRUS study included children <18 years with TB exposure (TB-E), TB infection (TB-I) or TB disease (TB-D). The ProPAED study included children 1 month to 18 years of age with fever and lower respiratory tract infection (viral or bacterial) and these were the sick controls (SC).

Results:

A total of 379 children were included in this analysis; 19 with TB-D, 12 with TB-I, 24 TB-E and 324 SC. Median age was 3.08 (IQR [1.37, 6.06]) years and 58% were male. Median NLR was highest in TB-D (2.05 [1.41, 2.64]) and significantly higher compared to children with TB-I (1.08 [0.82, 1.55]), TB-E (0.80 [0.63, 1.33]) and SC (0.31 [0.11, 0.97]) (all p-values < 0.05). Median MLR was similar in TB-D (0.25 [0.18, 0.34]) and SC (0.34 [0.21, 0.58]), but significantly higher in TB-D and TB-I when compared to TB-E (both p-values <0.05). Receiver operating characteristic curves of the ratios were calculated for children with TB-D and SC. NLR and MLR had at cut-off 0.75 and 0.63, an area under the curve of 0.84 and 0.63, sensitivity of 0.94 and 0.94, and specificity of 0.7 and 0.3, respectively. Similar results were obtained after adjustment for age.

Conclusions/Learning Points:

This study shows that NLR and MLR are promising easy-to-obtain diagnostic markers to differentiate children TB-D from non-TB lower respiratory tract infections. These results require confirmation in a larger study sample.