The definition of sepsis has changed several times since 19921516. The SCCM and ESICM revised the definition of sepsis and septic shock in 2016. The new definitions (Sepsis-3) focused on a dysregulated host response to infection and organ dysfunction. Sepsis is defined as infected patients with an increase of ≥2 Sequential Organ Failure Score (SOFA) points17. Septic shock is defined as refractory hypotension requiring vasopressors with concurrent hyperlactemia (>2 mmol/L) despite adequate fluid resuscitation (Figure 1). Severe sepsis was excluded from the guidelines, and quick SOFA (qSOFA), instead of the systemic inflammatory response syndrome (SIRS), was adopted for screening purposes (Figure 2).

The Sepsis-3 definitions were based on a large database and were the first to be tested in derivation and validation datasets. However, the definitions were not endorsed by some organizations and there are several issues associated with them6. First, lactate was not retained in the sepsis definition. Hence, by the Sepsis-3 definitions, patients with an increased lactate level but no hypotension (or compensated septic shock) can be missed. In other words, we may miss patients in the early phase of sepsis. The prevalence of this phenotype (i.e., normotensive patients with hyperlactemia) was 26% in a previous multicenter trial11. In the Sepsis-3 datasets, the prevalence of normotensive hyperlactemia (>4 mmol/L) was 9.9% but their mortality rate was not low (29.9%). Therefore, the validity of the Sepsis-3 definitions is suspect. Second, using the Sepsis-3 definitions, two components (the need for vasopressors and hyperlactemia) are needed concurrently to diagnose septic shock. That is, the lactate level is not a component of the definitions until the patient becomes hypotensive. Also, an infected patient with hypotension might not be considered to be in septic shock unless the lactate level was known. This implies that the utility of the Sepsis-3 definitions is limited in low-resource settings, where lactate levels are not frequently available. Therefore, further prospective studies are needed to demonstrate the validity of the Sepsis-3 definitions. Until then, it seems acceptable to use the pre-existing sepsis definitions.

Sepsis screening is reportedly associated with a decreased mortality rate1819. The surviving sepsis campaign (SSC) guidelines of 2016, as well as those of 2012, emphasize routine screening of potentially infected patients who are likely to be septic to improve the early identification and treatment of sepsis. They recommend that hospitals should have a performance improvement program that involves early recognition and management of sepsis8.

The SIRS criteria have, since 1992, been used to screen and identify sepsis patients20. To diagnose sepsis, at least two of the four SIRS criteria must be met. However, because SIRS can be triggered by a variety of infectious and noninfectious causes, it is insufficiently sensitive, and certainly not specific for sepsis. Hence, some patients who satisfy the SIRS criteria may not have sepsis, and vice versa21. In this context, the Sepsis-3 Task Force discarded the concept of SIRS and introduced, instead, qSOFA for sepsis screening17. The qSOFA is a simplified version of the SOFA score that comprises only three variables, and patients with a qSOFA score of ≥2 should be considered for the possibility of sepsis (Figure 2). The qSOFA is a readily available bedside tool without laboratory tests, and has better performance in non-intensive care unit (ICU) than ICU settings (area under the curve value, 0.81 vs. 0.66)13. The Sepsis-3 Task Force recommended that it be used to identify infected patients outside the ICU who are likely to be septic. However, a recent prospective study showed that a qSOFA score of ≥2 has high specificity (96.1% vs. 61.0% for SIRS ≥2) for organ dysfunctions but its poor sensitivity (29.7% vs. 72.1% for SIRS ≥2) may limit its use as a bedside tool22. The authors insisted that the SIRS criteria can be still useful.

Clinical evidence indicates that patients with acute deterioration or sepsis manifest clinical signs or symptoms several hours before the condition worsens. Early warning scores, such as the Modified Early Warning Score (MEWS), Early Warning Scoring System (EWSS), or National Early Warning Score (NEWS), were developed to screen patients at high risk of deterioration23242526. Although strong evidence based on robust data is lacking, these scores showed a trend toward improved outcomes and, when coupled with an outreach service (i.e., rapid response teams or medical emergency teams), it facilitates timely initiation of the optimal treatments upon recognition of septic patients25. Although respiratory or cardiac problems were the most common trigger for activations of such outreach teams2728, one study showed that sepsis as a cause of activations accounted for 19.9%, and EGDT was undertaken in 22.7%28. Interestingly, Churpek et al.29 compared several early warning scores, including qSOFA, among patients outside the ICU. qSOFA had a higher specificity and lower sensitivity than SIRS, MEWS, and NEWS for predicting in-hospital death or ICU transfer. The SIRS criteria (≥2) were more rapid than qSOFA for identifying patients. Therefore, use of qSOFA may be premature, and the SIRS criteria are a sensitive and useful bedside tool for sepsis screening outside the ICU.

Although sepsis dates back to at least the time of Hippocrates, the term “sepsis” is not well known47. This has led to avoidable mortality and morbidity worldwide. On May 24, 2017, Sir Liam Donaldson, the WHO envoy for patient safety, reported that awareness of sepsis by the public and politicians must be increased during a WHA Side Event on sepsis48. He said further that sepsis is an important issue that has been addressed effectively by clinicians and scientists, but is invisible to the public, political leaders, and leaders of healthcare systems.

Increased awareness, which leads to early presentation to hospital, can decrease the sepsis mortality rate by enabling timely and appropriate treatment. Experts recommend that the term “sepsis” be used frequently by healthcare professionals and patients to increase the level of awareness of the general public49. However, despite the significant impact, public awareness is currently very low; a survey in 2009 reported that 88% of respondents had never heard the term “sepsis”50. Since 2012, the Global Sepsis Alliance (GSA) has organized the annual World Sepsis Day, and many organizations or countries are undertaking sepsis awareness campaigns51. In healthcare facilities, continuous training and education are also needed to increase the level of awareness of sepsis on the part of healthcare providers, who must understand that the condition is a real medical emergency.

Most estimates of sepsis are based on studies in high-income countries, and data on low- and middle-income (or resource-limited) countries are scarce. So, increasing the awareness of sepsis is essential for controlling the sepsis burden in those countries5253. In South Korea, on “World Sepsis Day”, annual symposia and field events have taken place since 201254. However, further promotional or educational activities for the public, as well as support from the political leadership, are required to improve the situation.

Sepsis is frequently handled like a “garbage code” in the Global Burden of Disease Statistics. Most deaths due to sepsis are classified as their underlying infections, rather than sepsis itself48. Therefore, the burden of sepsis is likely to be underestimated. Hence, improving the coding process for sepsis may facilitate estimation of the true burden.

Nationwide (or statewide) registry data collected prospectively can facilitate accurate estimation of the incidence of sepsis, which can be used to improve performance and formulate future policies. A good example is the New York State sepsis registry. Beginning in 2014, all hospitals in New York State adopted sepsis protocols based on Rory's regulations for the early diagnosis and treatment of sepsis548. They are required to report their performance (compliance), as well as clinical information, to the New York State Department of Health548. Surprisingly, after the onset of the initiative, the average rate of protocol compliance has increased progressively from 73.7% in the second quarter of 2014 to 84.7% of adult sepsis patients in the third quarter of 20165. This compliance is in contrast to Asian countries; a large multinational study of Asian ICUs reported rates of compliance with the resuscitation and management bundles of 7.6% and 3.5%, respectively55.

The Core Outcome and Resource Evaluation (CORE) committee is a component of the Australia and New Zealand Intensive Care Society (ANZICS)56. All ICUs in Australia and New Zealand were invited to contribute to the ANZICS CORE registries in 1992. These registries have four registry domains: adult patients, pediatric patients, critical care resources, and central line-associated bloodstream infections. The CORE committee collects comprehensive data on various aspects of ICUs and reports back to the contributing ICUs. They also audit and analyze the performance of ICUs for quality assurance purposes.

Therefore, Asian countries, including South Korea, need to benchmark the successful stories of Western countries. This would eventually lead to the performance improvement of healthcare workers and the reduction of sepsis mortality.

When several individual effective treatments are applied concurrently, we anticipate better outcomes than any of the individual treatments alone; e.g., the ventilator-associated pneumonia or central-line infection prevention bundles5758. In 2004, the SSC group, in partnership with the Institute for Healthcare Improvement, developed the SSC bundle with the goal of reducing the mortality rate by 25%59. In 2012 and 2016, the 3-hour and 6-hour SSC bundles were introduced but, due to the negative results of three RCTs on EGDT, invasive monitoring, such as CVP and ScvO₂, was excluded from the 6-hour bundle in 2016860. More recently, based on the 2016 SSC guidelines, a revised hour-1 bundle (2018 bundle) with five key elements was developed (Table 2)61.

The implementation of sepsis bundles is a cornerstone of sepsis performance improvement programs, which are associated with a significant increase in compliance with the sepsis bundles and a reduction in the mortality rate. Levy et al.62 reported low mortality rates in high-compliance hospitals during a 7.5-year observation. Analysis of the New York State registry also demonstrated that the compliance rate of the 3-hour sepsis bundle was associated with a lower risk-adjusted in-hospital mortality rate5. Among various factors, high-income countries, surgical ICUs, long duration of implementation, and presentation to an Emergency Department were associated with a high rate of compliance5562, and lactate seemed frequently a non-compliant variable6263. However, the rate of compliance should continue to increase during the first 2 years of implementation62, and the mortality rate may decrease even if bundle completion is delayed in sepsis patients6465.

A large Asian study reported a low rate of compliance with SSC bundles. In South Korea, lack of critical care personnel was significantly associated with low compliance rates (e.g., total compliance of 5.6%)66. Thus, sufficient critical care personnel (e.g., intensivists and nurses) is an important factor for improving performance. Further studies should seek to identify methods of improving bundle compliance, as well as ways to overcome other barriers.