Clostridium difficile is a leading cause of nosocomial diarrhoea and one of the most common healthcare-associated infections. A dramatic worldwide increase in the incidence of C. difficile infection has occurred over the past two decades with the emergence of hypervirulent strains. Accurate and timely laboratory diagnosis of C. difficile infection is fundamental to ensure patients receive appropriate treatment and proper counter-infection measures are put in place. However, the availability of many commercial tests with different C. difficile targets contributes to uncertainty and controversy around the optimal diagnostic algorithm.

by Dr Guilherme Grossi Lopes Cançado, Prof. Rodrigo Otávio Silveira Silva and Prof. Eduardo Garcia Vilela

Introduction
Clostridium difficile is a major cause of healthcare-associated diarrhoea, and is linked to significant morbidity, economic burden and even mortality. Disagreement between diagnostic tests is an ongoing barrier to clinical decision-making and epidemiological surveillance. With the increasing number and severity of Clostridium difficile infections (CDI) after the emergence of the epidemic BI/NAP1/027 strain in the 2000s, there has been a renewed interest in optimizing the laboratory diagnosis of this infection.

Testing for C. difficile
Cytotoxin neutralization, toxigenic culture and toxin enzyme immunoassay
The two most accurate methods for CDI diagnosis are cytotoxin neutralization (CTN) and toxigenic culture (TC). CTN has long been used as a reference method for C. difficile detection, although different protocols have been proposed. Basically, this technique consists of inoculating a filtrate of a stool suspension into a cell culture (Vero, Hep2, fibroblasts, CHO or HeLa cells) and observing a cytopathic effect (such as cell rounding) as a consequence of disruption of the cell cytoskeleton 24–48 h later. CDI confirmation is obtained by the addition of a specific antiserum directed against C. difficile or against C. sordellii, neutralizing the toxin effects. Studies performed in the last 10 years, however, demonstrated that the sensitivities of CTN protocols range between 60 and 86%, when compared with toxigenic culture. Although culture was found to be the most sensitive method for detecting C. difficile in feces, it is not very specific owing to the possibility of isolating non-toxigenic strains. In order to overcome this issue, it should always be combined with a toxin detection method, such as enzyme immunoassay (EIA), direct cytopathic effect on cell lines or the identification of toxin-related genes by PCR. This notwithstanding, both techniques are time consuming, laborious and require trained personnel, reasons why they are not frequently used in daily practice [1].

In this context, commercial rapid EIA and DNA-based tests are currently the most widely used tools for CDI diagnosis. EIA for toxin detection is fast, cheap, easy to perform and does not require technical training or special equipment: characteristics which have favored its use in low income countries. However, several studies have reported low detection sensitivities of different commercial kits (approximately 50–70%), making toxin EIAs inadequate as standalone tests [2]. Some authors have suggested the analysis of sequential samples from the same patient, but this practice did not significantly increase the positive predictive value of the test and may even amplify the rate of false positives. Furthermore, toxin tests can be falsely negative, which can be due to previous antibiotic treatment, pre-analytic toxin degradation or sampling error (e.g., ileus, fecal dilution). Consequently, toxin EIA may be sufficient and cost-effective as a screening test in clinical settings where there is a low prevalence of CDI, but not for epidemic scenarios.

Detection of glutamate dehydrogenase
Another method that has been used to diagnose CDI is the detection of glutamate dehydrogenase (GDH), a metabolic enzyme constitutionally produced almost exclusively by C. difficile, which is significantly more sensitive than toxin A and B EIAs. We have recently shown that GDH rapid immunoassay presents 100% sensitivity and negative predictive value, compared with the culture-based method, in accordance with order authors [3]. Cheng et al. have advocated the use of GDH in order to improve the diagnostic capacity and control of potential outbreaks of CDI in developing countries, as this test is five to ten times cheaper than molecular assays [4]. In fact, we have recently reported a significant increase in CDI treatment rates in a university hospital of Brazil (from 53.8% to 100%) simply after replacing the toxin EIA by the rapid GDH immunoassay as a screening test [3]. This finding supports the use of GDH in countries with limited economic resources, as it is simple to perform and, unlike PCR, requires no special facilities or personnel qualifications that might restrict its use. Although sensitive, GDH lacks specificity, because it only indicates the presence of the microorganism, instead of toxin production. In this context, a concern with the overdiagnosis and overtreatment of patients with diarrhoea and C. difficile colonization may arise. A negative test can almost rule out CDI and avoid the need for more expensive toxin testing, but a positive GDH assay should preferably be followed by toxin-based diagnostic methods before one can come to a safe conclusion on infection. Asymptomatic carriage of C. difficile occurs in 5–15% of healthy adults, but may be as high as 90% in newborns and healthy infants, and up to 51% in residents in long-term care facilities [5]. In this way, the incorporation of a two-step strategy, including a sensitive organism-based test for screening, such as GDH, followed by a toxin test for confirmation of clinically significant disease, is a reasonable approach. Combined tests including GDH and toxin detection in one easy-to-use cartridge have been recently developed, but several authors have demonstrated limited sensitivity of the toxin component. In our study, the toxin component presented an even lower sensitivity than conventional toxin EIA (50% vs 58%), compared with toxigenic culture [3]. In this way, GDH+/toxin− samples would still have to be submitted to a third test (in a multistep algorithm) to rule out infection, increasing time to diagnosis and health costs. However, some studies have shown that, even without treatment, patients with toxin-negative stool specimens have shorter diarrhoea duration than those with toxin-positive stool specimens. These findings may suggest a limited need for CDI treatment for GDH-positive patients and toxin-negative stool specimens [6].

Molecular testing and rapid diagnosis
In 2009, facing an epidemic of CDI due to hypervirulent C. difficile strains, the US FDA cleared the first commercial molecular test for rapid CDI diagnosis [7]. Shortly afterwards, hospitals in North America and Europe began switching to DNA-based testing strategies as a method of choice for the diagnosis of C. difficile. Nucleic acid amplification testing (NAAT), including rapid testing PCR and loop-mediated isothermal amplification (LAMP), can detect the tcdA/tcdB genes (regulate toxin A/B production) or the tcdC gene (a negative regulator of toxin A and B production) and identify the presence of toxigenic C. difficile in a single step. It has a higher sensitivity (90–95%) and specificity (95–96%) than toxin EIAs and has a rapid turnaround time, but requires specialized equipment and personnel. It is worth noting that the Gene-Xpert^® system can also simultaneously indicate the presence of the potentially ‘hypervirulent’ ribotype 027 strain, giving important epidemiological information. Nonetheless, although a PCR assay can identify toxin genes, it cannot detect the presence of toxin. Several studies have shown that toxin−/C. difficile+ patients present shorter duration of symptoms and better outcomes than toxin+/C. difficile+ individuals, demonstrating that this subpopulation may have either mild CDI or colonization, and even not need to be treated. In this way, some authors have questioned the use molecular tests as standalone tests because of the high likelihood of overdiagnosis and overtreatment [8]. However, there may be a role for identifying carriers to prevent transmission and this issue should be better addressed in future studies.

C. difficile testing protocol
Using reliable and rapid diagnostic tests, such as NAAT, practitioners could offer appropriate treatment earlier, thereby sparing patients a time-consuming evaluation and unnecessary antibiotic therapy and its complications. Despite worldwide advances in analytical technology, transport systems and computerization, many laboratories in developing countries have difficulties in improving turnaround times and diagnostic capability. The optimal diagnostic algorithm for CDI is yet to be adequately defined and may vary according to the underlying clinical and laboratory circumstances. We believe that the decision about which test(s) to use is determined by a combination of what is practical and feasible in a specific setting. In Brazil, for example, where only toxin EIA was available for CDI diagnosis until 2015, the introduction of the GDH test resulted in a dramatic increase in C. difficile treatment rates [3]. All of the reference methods, CTN, toxigenic culture, or PCR, require advanced infrastructure and expensive testing. Smaller, community-based hospitals, where much of C. difficile testing is currently performed, may not have the financial means to establish these methods nor have the staff to perform the time-consuming, highly complex assays. In this way, using GDH–toxin A/B assays may be an adequate option for diagnostic algorithms in developing countries, whereas molecular techniques, toxigenic culture or CTN may be reserved for discordant samples (Figure 1). Future studies should focus on developing simple diagnostic approaches to accurately distinguish active infection from mere colonization.

References
1. Delmée M. Laboratory diagnosis of Clostridium difficile disease. Clin Microbiol Infect 2001; 7(8): 411–416.
2. Silva RO, Vilela EG, Neves MS, Lobato FC. Evaluation of three enzyme immunoassays and a nucleic acid amplification test for the diagnosis of Clostridium difficile-associated diarrhea at a university hospital in Brazil. Rev Soc Bras Med Trop 2014; 47(4): 447–450.
3. Cançado GGL, Silva ROS, Nader AP, Lobato FCF, Vilela EG. Impact of simultaneous glutamate dehydrogenase (GDH) and toxin A/B rapid immunoassay on Clostridium difficile diagnosis and treatment in hospitalized patients with antibiotic-associated diarrhea in a university hospital of Brazil. J Gastroenterol Hepatol 2017; doi: 10.1111/jgh.13901 [Epub ahead of print].
4. Cheng JW, Xiao M, Kudinha T, Xu ZP, Sun LY, Hou X, Zhang L, Fan X, Kong F, Xu YC. The role of glutamate gehydrogenase (GDH) testing assay in the diagnosis of Clostridium difficile infections: a high sensitive screening test and an essential step in the proposed laboratory diagnosis workflow for developing countries like China. PLoS One 2015; 10(12): e0144604.
5. Furuya-Kanamori L, Marquess J, Yakob L, Riley TV, Paterson DL, Foster NF, Huber CA, Clements AC. Asymptomatic Clostridium difficile colonization: epidemiology and clinical implications. BMC Infect Dis 2015; 15: 516.
6. Yuhashi K, Yagihara Y, Misawa Y, Sato T, Saito R, Okugawa S, Moriya K. Diagnosing Clostridium difficile-associated diarrhea using enzyme immunoassay: the clinical significance of toxin negativity in glutamate dehydrogenase-positive patients. Infect Drug Resist 2016; 9: 93–99.
7. Polage CR, Turkiewicz JV, Cohen SH. The never-ending struggle with laboratory testing for Clostridium difficile infection. J Comp Eff Res 2016; 5(2): 113–116.
8. Polage CR, Gyorke CE, Kennedy MA, Leslie JL, Chin DL, Wang S, Nguyen HH, Huang B, Tang YW, Lee LW, Kim K, Taylor S, Romano PS, Panacek EA, Goodell PB, Solnick JV, Cohen SH. Overdiagnosis of Clostridium difficile infection in the molecular test era. JAMA Intern Med 2015; 175(11): 1792–1801.

The authors
Guilherme Grossi Lopes Cançado*^1,2 MD, Rodrigo Otávio Silveira Silva³ PhD, and Eduardo Garcia Vilela¹ MD, PhD
¹Instituto Alfa de Gastroenterologia,
Federal University of Minas Gerais, Brazil.
²Department of Gastroenterology,
Hospital da Polícia Militar de Minas Gerais, Brazil.
³Veterinary School, Federal University of Minas Gerais, Brazil.

*Corresponding author
E-mail: guilhermegrossi@terra.com.br

Sepsis is a medical emergency that needs rapid identification and treatment to create the best possible outcomes. However, in the early stages it can be very difficult to distinguish sepsis from uncomplicated infection. This article summarizes recent developments in sepsis nomenclature and definitions as well as providing an insight into the role that biomarkers might play in diagnosis and prognosis.

Background
Sepsis is a life-threatening condition associated with high morbidity and mortality, with the risk of death ranging from 30% to 80% depending on the severity of the disease. The World Health Organization estimates that more than 30 million people are affected by sepsis worldwide every year [1], although for reasons discussed by Candel et al., the actual epidemiology of sepsis is difficult to ascertain [2]. In the UK and USA it is thought that sepsis is the cause of around 37 000 and nearly 270 000 deaths per year, respectively [3, 4]. Outcomes of sepsis are better if it is detected and treated early, but despite the large numbers of people affected by it, public awareness of it is still low. In recent years, awareness campaigns have been launched and this year several popular TV and radio programmes in the UK have featured sepsis storylines (Call the Midwife, Coronation Street and The Archers).
Definitions
The difficulties experienced in studying the epidemiology of sepsis are likely to reflect the problems of characterization and diagnosis of the disease, which is in turn a reflection of the complex nature of the condition. Original definitions of sepsis date back to 1991, with the idea that sepsis was caused by systemic inflammatory response syndrome (SIRS) in resulting from infection. In 2001 the definitions were re-examined but left largely unchanged. In 2016, a task force re-evaluated and updated definitions of sepsis and septic shock (Box 1), taking into account improved understanding of the pathobiology of sepsis, which is now recognized to involve early activation of both pro- and anti-inflammatory responses, along with major modifications in non-immunologic pathways such as cardiovascular, neuronal, autonomic, hormonal, bioenergetic, metabolic, and coagulation [5]. A lay definition of sepsis published in 2011 [6] was also accepted by the 2016 task force (Box 1). The definitions created in 1991, 2001 and 2016 have been designated Sepsis-1, Sepsis-2 and Sepsis-3, respectively, to indicate the need for ongoing refinement.

Diagnosis of sepsis
Early diagnosis and treatment of sepsis is associated with improved outcomes, but the difficulty lies in distinguishing sepsis from uncomplicated infection. Identification of patients with sepsis is largely achieved through the use of the Sequential (or Sepsis-Related) Organ Failure Assessment (SOFA) score (Table 1) in the hospital setting or the quick SOFA (qSOFA) score (See Figure 1 “Operationalization of Clinical Criteria Identifying Patients With Sepsis and Septic Shock” in Singer et al. [5]). Commencement of treatment should occur within the first hour of admission and should not be delayed by waiting for results from the lab, as the SOFA score can be applied retrospectively. Management of sepsis also requires (amongst other things) that blood samples are taken before broad spectrum antibiotics are administered and that once the pathogen has been identified antibiotic usage can be refined to aid antimicrobial stewardship (See the Surviving Sepsis Campaign [7] and NICE guidelines [8] for full details of early sepsis management). Sepsis is most commonly caused by bacterial infection, but can also be due to fungal, viral or parasitic infection. However, identification of the pathogen and its antibiotic susceptibility and/or resistance by classic culture techniques is slow and molecular- and proteomic-based approaches, such as matrix-assisted laser desorption/ionization–time of flight (MALDI-TOF) spectroscopy, may improve turnaround times [9].

Biomarkers
The difficulty of distinguishing sepsis from uncomplicated infection has long driven the search for suitable biomarkers to aid sepsis diagnosis. An ideal biomarker would be able to distinguish sepsis from non-infectious causes of critical illness, having a fast and specific increase in sepsis and a rapid decrease after effective therapy. A number of potential biomarkers have been identified, although none are specific enough to be used alone.
Procalcitonin and C-reactive protein
The most-studied biomarkers are procalcitonin and C-reactive protein (CRP). CRP is an acute-phase protein that is secreted from the liver in the response to inflammatory processes and is therefore sensitive but not specific for sepsis. Procalcitonin, again is produced in response to inflammation and infection, and is so far the only biomarker to be used clinically, as it differentiates better than CRP between infectious and non-infectious causes of critical illness. A meta-analysis found that procalcitonin had a mean sensitivity and specificity of around 70% and an area under receiver operator characteristic curve of less than 0.80 [10]. However as levels of procalcitonin are known to be raised after surgery, trauma and viral infection, the Surviving Sepsis Campaign concluded that procalcitonin levels are not adequate to distinguish sepsis from other causes of inflammation [11], although it may be useful for indicating when treatment with antibiotics can end [12].

Interleukin 6 (IL-6)
IL-6 was initially a biomarker of interest for rapid sepsis diagnosis as it has a fast kinetic profile – the concentration increases within 2 hours of onset of sepsis and decreases within 6 hours. However, the results from studies have been mixed, with some suggesting that it was able to discriminate between sepsis and non-infectious illness, whereas others found that procalcitonin was better, hence it has not been added to current guidelines [11].

Promising biomarkers
A number of other biomarkers have been identified that show promise include soluble urokinase-type plasminogen activator receptor, presepsin and proadrenomedullin [2, 13]. Additionally, recently, reduced serum levels of fetuin-A (a major hepatokine) were found to be independently associated with predicting progression to septic shock and higher rates of mortality [14].

Biomarker panels
Even today, no single biomarker has the diagnostic strength to identify patients suffering from sepsis and it is likely that assessing panels of biomarkers will increase the sensitivity and accuracy of diagnosis of sepsis, compared to any individual biomarker (for example, see the study by Kofoed et al. [15]). More recently, the power of mass spectrometry and “-omics studies” is being investigated with some promise, although still suffering from limitations [13].

References
1. Sepsis. World Health Organization 2018; http://www.who.int/news-room/fact-sheets/detail/sepsis.
2. Candel FJ, et al. Current aspects in sepsis approach. Turning things around. Rev Esp Quimioter 2018; 31(4): 298–315.
3. Improving outcomes for patients with sepsis: a cross-system action plan. NHS England 2015; https://www.england.nhs.uk/wp-content/uploads/2015/08/Sepsis-Action-Plan-23.12.15-v1.pdf.
4. Sepsis. Centers for Disease Control and Prevention 2018; https://www.cdc.gov/sepsis/datareports/index.html.
5. Singer M, et al. The Third International Consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016; 315(8): 801–810.
6. Czura CJ. Merinoff symposium 2010: Sepsis – speaking with one voice. Mol Med 2011; 17(1-2): 2–3.
7. Surviving Sepsis Campaign: International guidelines for management of sepsis and septic shock: 2016. Surviving Sepsis Campaign 2016; http://www.survivingsepsis.org/Guidelines/Pages/default.aspx.
8. Sepsis: recognition, diagnosis and early management; NICE guideline [NG51]. National Institutes for Health and Care Excellence 2017; https://www.nice.org.uk/guidance/NG51/chapter/Recommendations#identifying-people-with-suspected-sepsis.
9. Ward KM, Harris R. Sepsis: earlier organism identification using MALDI-TOF. Clin Lab Int 2015; Nov: 14–18.
10. Wacker C, et al. Procalcitonin as a diagnostic marker for sepsis: a systematic review and meta-analysis. Lancet Infect Dis 2013; 13: 426–435.
11. Dellinger RP, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012. Crit Care Med 2013; 41(2): 580–637.
12. Sager R, et al. Procalcitonin-guided diagnosis and antibiotic stewardship revisited. BMC Med 2017; 15: 15.
13. Ludwig KR, Hummon AB. Mass spectrometry for the discovery of biomarkers of sepsis. Mol Biosyst 2017; 13(4): 648–664.
14. Karampela. Karampela I, Kandri E, Antonakos G, Vogiatzakis E, Christodoulatos GS, Nikolaidou A, Dimopoulos G, Armaganidis A, Dalamaga M. Kinetics of circulating fetuin-A may predict mortality independently from adiponectin, high molecular weight adiponectin and prognostic factors in critically ill patients with sepsis: A prospective study. J Crit Care 2017; 41: 78–85.
15. Kofoed K, et al. Use of plasma C-reactive protein, procalcitonin, neutrophils, macrophage migration inhibitory factor, soluble urokinase-type plasminogen activator receptor, and soluble triggering receptor expressed on myeloid cells-1 in combination to diagnose infections: a prospective study. Crit Care 2007; 11(2): R38.