Personalized medicine of patients with respiratory infections through the measurement of specific blood biomarkers: fact or fiction?

KEYWORDS:

• Biomarker
• procalcitonin
• pneumonia
• respiratory tract infection
• sepsis
• antibiotic guidance

The body’s overwhelming response to a systemic infection in combination with organ failure, a clinical syndrome called sepsis, is responsible for significant morbidity, mortality, and financial burden [1]. Sepsis most frequently originates from infections of the respiratory tract, particularly pneumonia. With this in mind and to prevent such devastating outcomes, physicians may pay particular attention to patients with respiratory infection. This includes the early use of antibiotic treatment and close patient monitoring as an evidence-based guideline recommendation for the patients with pneumonia and sepsis. While such treatment bundles reduce the burden of disease in sepsis, there is a considerable risk for overtreatment. Particularly, due to the lack of reliable microbiological markers (blood cultures and PCR methods still have low sensitivity) and lack of accurate prognostic rules for the individual patient with respiratory infection, patients with mild nonbacterial infections are frequently overtreated with unnecessary use of antibiotics and inpatient care [2]. The consequences include an increase in multiresistant bacteria, high costs for inhospital care, and potential adverse outcomes. Combination of a careful clinical assessment with the measurement of specific blood biomarkers that provide information about the risk for bacterial disease and expected benefit of antibiotic treatment as well as about prognosis and the best site-of-care is a promising avenue to move away from the ‘bundled care approach’ to more personalized treatment decisions in individual patients [3].

1. Biomarkers reflecting bacterial infection and response to treatment

During respiratory infection, pathogens and their antigens stimulate pro- and anti-inflammatory mediators that constitute the host defense to the site of acute infection. These host-response markers are potential biomarkers that provide information about severity, causative organisms of infection, and risk for adverse patient outcomes. Recent studies have provided strong evidence that host-response markers facilitate early recognition of sepsis, enable assessment of its severity, and provide guidance regarding therapeutic decisions in individual patients [4–6]. This may allow for a transition from bundled, nonspecific infection management involving protocols to more individualized management based on the clinical profile of each patient.

Among different candidate markers, observational and interventional clinical trials have provided evidence that the host-response marker procalcitonin (PCT) facilitates improved patient management [7]. There has been a large body of studies evaluating PCT in different types of infections and patient populations (see reference [8] for an overview). As for diagnostic accuracy in sepsis, a meta-analysis from 2013 that included 3244 critically ill patients classified as experiencing sepsis or a systemic inflammatory response syndrome of noninfectious origin pooled the diagnostic power of PCT [9]. Studies between 1996 and 2011 were included and showed a good high discriminatory ability of PCT (area under the curve of 0.85), with pooled sensitivity and specificity of 0.77 and 0.79, respectively [9]. Observational data for pneumonia and general emergency patients with positive blood culture as the gold standard have shown similar results in regard to discrimination [10,11]. More important than observational studies are interventional trials evaluating whether a marker has a positive effect on patient outcomes. For PCT, several trials have been published looking at the effect of this marker on antibiotic consumption and clinical outcomes, particularly for patients with respiratory infections. A 2011 meta-analysis found PCT to be highly effective in reducing antibiotic initiation in low-risk respiratory infections (e.g. bronchitis and chronic obstructive pulmonary disease (COPD) exacerbation) and reducing the duration in pneumonia patients (including sepsis patients in the intensive care unit) with no worse clinical outcomes despite strong reductions in antibiotic use. A recent large trial in the critical care setting also reported a lower mortality (20% vs. 25%) associated with the use of PCT in patients with presumed infection [12]. Importantly, PCT is not a prefect marker for infection, and false positives (e.g. in patients with surgical trauma or cardiac arrest) and false negatives (e.g. in localized or subclinical infections or intracellular pathogens) can occur. Therefore, PCT should only be used in conjunction with a thorough clinical assessment of a patient. More informative than a single level are kinetics of PCT over time. Several studies have found a drop in PCT to indicate good resolution of illness, and de-escalation of antibiotics may be considered. Also, there is need for further research to understand the value of PCT in different types of infection. Although respiratory infections have been in general well studied, most studies looked at community-acquired infection, and it remains somewhat unclear whether hospital-acquired infections have a similar effect on the biomarker.

For an optimal assessment of patients, it would be helpful to combine markers of host response (such as PCT), with improved microbiological markers. Several new assays and technologies that have recently come into clinical use such as gene expression profiling, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, and nucleic acid aptamers among others [2]. A combined assessment looking at both the pathogen and the host response holds great promise for the future for further individualize treatment approaches in respiratory infections. Also, as different markers may provide complimentary infection, a multi-marker profile may be interesting for improved accuracy.

2. Prognostic biomarkers for risk stratification

With the publication of the PORT cohort in 1997 and subsequent guideline recommendations to calculate mortality scores in patients with community-acquired pneumonia, the individual risk assessment was early propagated for respiratory infections [13]. The initial pneumonia severity index as well as the CURB65 score was mainly based on age, clinical parameters on admission (including markers of renal function), and comorbidities. However, in the last 20 years, several new prognostic markers from distinct physiopathological pathways were investigated for their ability to predict mortality and other adverse outcomes in patients with pneumonia and other respiratory infections [14–17]. While most studies focused on short-term (i.e. 30 days) outcomes, some studies also looked at long-term outcomes including mortality, quality of life, among others [18,19]. The search for accurate prognostic markers has been supported by novel approaches including metabolomics. Such an approach is helpful in evaluating a magnitude of novel markers and their degradation products from different pathways in parallel for their prognostic abilities [20]. An improved understanding of the pathophysiology in patients with respiratory infections may also result in novel treatment approaches or a more individualized use of drugs according to a patient’s state of disease [21–23]. For example, while trials found positive effects of corticosteroids on time to recovery in pneumonia patient populations, these effects may or may not be modified by patient’s baseline characteristics (i.e. viral vs. bacterial infection and comorbidities such as diabetes) [24–26]. The association of cortisol level and outcome also shows a time-dependency with high levels predicting adverse outcome at short, but beneficial outcomes at the long run [27].

Importantly, most studies looking at prognostic markers have used observational designs reporting associations of markers and outcomes. More helpful would indeed be interventional studies testing whether the inclusion of prognostic markers improves care of individual patients or lowers cost of care [28]. Without interventional research, we will not understand the true added value of prognostic markers in the care of our patients.

3. Conclusions

There is reason to be optimistic for more personalized patient approaches through the measurement of specific diagnostic and prognostic blood biomarkers being soon available for patients with respiratory infections. With PCT, there is one marker that is already used in common practice in many hospitals as a guide to antibiotic treatment with very positive trial results. Several prognostic markers have been published, but the large clinical trial proofing added value of such an approach is still pending.

Declaration of interest

PS is supported by the Swiss National Science Foundation (SNSF Professorship, PP00P3_150531) and the Forschungsrat of the Kantonsspital Aarau (1410.000.058 and 1410.000.044). PS received research support from BRAHMS/Thermofisher, BioMerieux, Roche, Abbott, Nestle and Novo Nordisk. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Additional information

Funding

This paper was not funded.