2.8.1. Tools for Improved Antimicrobial Stewardship
The most important step in managing a systemic inflammatory response is defining the cause and initiating appropriate treatment; with an infective aetiology, this means administering the correct antibiotic. Current methods have a poor specificity and sensitivity, with a causative agent being identified in 50% of cases of sepsis and they are time-consuming.
New technologies are being developed with the potential to considerably expedite organism identification and antibiotic resistance patterns, providing clinicians with real-time information to guide antibiotic treatment; ensuring immediate appropriate treatment with the promise of reduced antibiotic resistance patterns.
DNA sequencing is a more robust method of bacterial identification [93
] than the more widely used conventional method of microscopy and culture of microbiological samples. The bacterial 16S ribosomal ribonucleic acid (rRNA) gene is a hypervariable DNA sequence of approximately 1500 base pairs that codes for 16S rRNA (a catalytic RNA, which forms part of the 30S ribosomal subunit) and is a molecular marker unique to different bacteria [93
]. It has been applied to the identification of bacteria in clinical samples that are difficult to culture but is rarely available as a point of care test.
Multiplex polymerase chain reaction (PCR) uses this technology and can be used to identify multiple pathogens and their resistance genes, from both positive blood cultures and whole blood samples. The National Institute for Health and Care Excellence (NICE) advocates its use for the diagnosis of Neisseria meningitidis
in under 16-year olds [94
]. However, further evaluation of its clinical performance is required before widespread implementation can be considered [95
]. The technique is currently expensive, but in the long-term this may be offset by a shorter turn-around time for results, focussed treatment and shorter hospital stays. Thus, there is scope for their widespread introduction to routine diagnostics in the future.
Bacterial and fungal organisms can also be identified by identification of their specific protein profile or “fingerprint” using a new technique in mass spectrometry, known as Matrix Assisted Laser Desorption Ionisation Time of Flight Mass Spectrometry or MALDI-TOF MS. Advantages are that it is rapid, accurate and can identify a wide spectrum of organisms from clinical isolates of positive cultures; there is also some evidence to show the accuracy of the technique in identification of organisms directly from biological fluids, e.g., urine despite the considerably lower CFU/mL. As a result, this technology has been adopted by a number of laboratories in developed countries. However, susceptibility testing by identification of resistance phenotypes using MALDI-TOF is not yet clinically applicable and therefore the sole use of MALDI-TOF is not yet feasible [96
]. An additional potential barrier to its widespread use is the expense of the MALDI-TOF instrument [97
Key advantages of results obtained by MALDI-TOF (bacteria and fungi) and PCR based methods are that they are accurate and rapid and theoretically can be applied both to positive blood cultures and whole blood. To et al. demonstrate how MALDI-TOF can be adapted to identify group B streptococcus (GBS) from vaginal swab isolates with a reduced turnaround time and with excellent specificity and sensitivity [98
]. In addition to the clinical relevance of this for women who are pregnant, as referred to previously, significant neonatal morbidity and mortality is associated with this infection. The potential diagnostic applications using other biological fluids is also of particular interest for the obstetric population as the urinary tract is a common source of infection in the pregnant and puerperal patient. Likewise, making a timely diagnosis of chorioamnionitis might be facilitated by analysis of amniotic fluid using proteomic biomarkers reflecting intra-amniotic inflammation [99
]. However, presently there is insufficient data to demonstrate the benefit of using these tools for amniotic fluid. With PCR, the ability to amplify microbial products also means that non-pathogenic organisms will be identified, which may complicate the diagnostic picture.
The ultimate objective is to provide clinicians with the ability to target antibiotic treatment for the pathogen and its sensitivity, to increase efficacy of treatment and to obviate multi-antibiotic resistance.
2.8.2. “Omics Studies”—Biomarkers for the Obstetric Patient
Omics based technologies enable quantitative and qualitative analysis of the molecules involved in biological pathways. Multiple molecular levels including DNA (genome), mRNA (transcriptome), proteins (proteome), and metabolites (metabolome) can be targeted to elicit the pathophysiology of a given disease process. Having identified specific biological molecules using these techniques, their influence in sepsis research can be assessed in the context of their role in the immune and metabolic responses to disease and enable scientists to pinpoint aberrant steps in biological pathways that may contribute to disease.
As sepsis is such a heterogenous condition with multi-systemic impact, it is unlikely that a single biomarker would have significant sensitivity and specificity for the early diagnosis of sepsis; using multiple biomarkers may increase the yield and omic technologies may be a way to achieve this [100
]. The “omics” approach has the potential to help identify “signature” biological markers that can be developed into clinical tools for diagnosis and prognosis. Additionally, the field holds promise in guiding clinicians in the escalation and de-escalation of appropriate therapy (theragnosis) whereby monitoring patient response to treatment allows us to tailor therapy to the individual and enable development of novel therapies [101
]. There is the potential for omics studies to revolutionise sepsis care [103
Transcriptomics, in particular, has been a focus of such studies, analysing the host signalling response in sepsis. The transcriptome of immune cells from patients with sepsis theoretically could provide an early diagnostic tool for sepsis. The set of mRNA markers it represents known as the “transcriptomic signature”, may demonstrate the earliest change triggered by the infective process before translation of the relevant proteins [100
]. Indeed, there is emerging evidence that a transcriptomic signature is a reflection of the host’s sepsis response, and has prognostic value, particularly regarding mortality risk, and this is independent of the causative pathogen or site of infection [104
A recent study identified two distinct transcriptomic profiles of peripheral blood leucocytes amongst a subgroup of patients admitted to the Intensive Care Unit with sepsis secondary to community acquired pneumonia (CAP) [105
]. These profiles correlated with different clinical phenotypes, one of which correlated with markers consistent with an immunosuppressed state (characterised by endotoxin tolerance, T cell exhaustion and downregulation of HLA class II) and was associated with higher patient mortality [105
]. Thus, it can be seen that, with further development, this technology could be used for prognosis of illness severity in patients admitted to ICU with CAP.
Another prospect of transcriptomics is the potential to distinguish bacterial from viral infections and the associated advantage of targeted antibiotic therapy. In a multi-centre prospective observational study of 370 febrile children, a disease risk score (DRS), determined by evaluation of the differential expression of a 2-transcript RNA signature, was able to distinguish bacterial from viral infection with excellent sensitivity and specificity (100% and 96% respectively). There was also good correlation of these findings when the transcript was applied to previously published datasets of RNA expression in adults and children. The transcripts of note, IF144L (interferon-induced protein 44-like) and FAM89A (family with sequence similarity 89, member A), demonstrated increased expression with viral and bacterial infections respectively [106
]. The ability to clinically discriminate widespread inflammation of infectious aetiology from that which is non-infectious is another important tool for stratifying patients for the prudent use of antibiotics and studies have demonstrated that transcriptomics may be a sensitive method to do this [107
]. The “SeptiCyte Lab“ uses a microarray panel of 4 RNA transcripts (CEACAM 4, LAMP1, PLA2G7 and PLAC8) and has demonstrated superior performance (AUROC) distinguishing systemic inflammatory response syndrome (SIRS) from sepsis, when compared to the performance of procalcitonin using whole blood samples, and is one of a number of tests that have been approved by the Food and Drug Administration (FDA) in the United States of America to aid clinical diagnosis [108
]. Along similar lines, an RNA microarray analysis study, with quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) validation, demonstrated that the ratio of gene expression of FAIM3:PLAC8 can be used to determine whether respiratory distress in ITU admissions is secondary to community acquired pneumonia or a non-infectious cause. No prognostic ability was demonstrated [110
Analysis of the genomic expression of peripheral blood white cells in sepsis in pregnancy may demonstrate activation or suppression of specific pathways, which could guide an approach to “precision medicine”. Differential gene expression may be associated with known features of immune tolerance, nitric oxide synthesis, cytokine production or reveal new pathways unique to this population. All of which are potential targets for novel therapies.
In parallel, a growing number of studies, including that of Ferrario et al [101
], demonstrate that plasma metabolite levels reflect the severity of sepsis and septic shock and have been associated with the risk of mortality [112
]. Further, the metabolite signature or metabolome may even prove useful in guiding response to therapy in the acute phase of sepsis [113
]. Using this technique, it may be possible to determine metabolic profiles of plasma and urine in pregnant women with suspected infection that are associated with the development of severe sepsis to further our understanding of the pathophysiology of the syndrome. Urine samples, in particular, also have the potential to provide non-invasive point of care testing to facilitate early diagnosis and treatment [114
A hugely exciting prospective study is the Bioresource in Adult Infectious Diseases (BioAID), a multicentre study in the UK, collecting biological samples from patients who present to the Emergency Department with suspected infection [115
]. Key objectives of the study include analysis of the immunological, proteomic, metabolomic and transcriptomic gene signatures of these patients with the aim of increasing our understanding of the characteristics of host susceptibility to infection and important features of the host-pathogen interaction which may influence the trajectory of disease severity. The resulting data has huge potential to guide biomarker development for the early diagnosis of bacterial infection. Such a study provides a unique opportunity to include pregnant and parturient women presenting to obstetric departments with suspected infection.
Thus, in addition to our current range of assessment tools and investigations, perhaps we can hope to improve our diagnosis and management of sepsis with the addition of immunological point of care tests, echocardiography and omics technologies.