Skip to main content

The Japanese Clinical Practice Guidelines for Management of Sepsis and Septic Shock 2020 (J-SSCG 2020)

Abstract

The Japanese Clinical Practice Guidelines for Management of Sepsis and Septic Shock 2020 (J-SSCG 2020), a Japanese-specific set of clinical practice guidelines for sepsis and septic shock created as revised from J-SSCG 2016 jointly by the Japanese Society of Intensive Care Medicine and the Japanese Association for Acute Medicine, was first released in September 2020 and published in February 2021. An English-language version of these guidelines was created based on the contents of the original Japanese-language version. The purpose of this guideline is to assist medical staff in making appropriate decisions to improve the prognosis of patients undergoing treatment for sepsis and septic shock. We aimed to provide high-quality guidelines that are easy to use and understand for specialists, general clinicians, and multidisciplinary medical professionals. J-SSCG 2016 took up new subjects that were not present in SSCG 2016 (e.g., ICU-acquired weakness [ICU-AW], post-intensive care syndrome [PICS], and body temperature management). The J-SSCG 2020 covered a total of 22 areas with four additional new areas (patient- and family-centered care, sepsis treatment system, neuro-intensive treatment, and stress ulcers). A total of 118 important clinical issues (clinical questions, CQs) were extracted regardless of the presence or absence of evidence. These CQs also include those that have been given particular focus within Japan. This is a large-scale guideline covering multiple fields; thus, in addition to the 25 committee members, we had the participation and support of a total of 226 members who are professionals (physicians, nurses, physiotherapists, clinical engineers, and pharmacists) and medical workers with a history of sepsis or critical illness. The GRADE method was adopted for making recommendations, and the modified Delphi method was used to determine recommendations by voting from all committee members.

As a result, 79 GRADE-based recommendations, 5 Good Practice Statements (GPS), 18 expert consensuses, 27 answers to background questions (BQs), and summaries of definitions and diagnosis of sepsis were created as responses to 118 CQs. We also incorporated visual information for each CQ according to the time course of treatment, and we will also distribute this as an app. The J-SSCG 2020 is expected to be widely used as a useful bedside guideline in the field of sepsis treatment both in Japan and overseas involving multiple disciplines.

Introduction

Approximately 50 million people worldwide die from sepsis each year. Sepsis is a serious illness that affects all age groups, and the social significance of the creation of a high-quality guideline with the objective of providing medical support for this illness is high. The Surviving Sepsis Campaign Guideline (SSCG) [1, 2] has been revised as an international sepsis clinical practice guideline every 4 years since 2004. In 2012, the Japanese version of the Surviving Sepsis Campaign Guideline (J-SSCG), which considered the actual circumstances of Japanese clinical settings, was first published by the Japanese Society of Intensive Care Medicine (JSICM) [3, 4]. At the time of the 2016 revision (J-SSCG 2016), JSICM and the Japanese Association for Acute Medicine (JAAM) worked together to create a high-quality guideline that is easy to understand even for general clinicians, aiming for widespread dissemination. J-SSCG 2016 actively took up new domains not covered in SSCG 2016, such as imaging diagnosis, body temperature regulation, ICU-acquired weakness (ICU-AW), and post-intensive care syndrome (PICS), providing medical guidelines.

In this current revision (J-SSCG 2020), the two societies have once again cooperated with one another with the aim of providing support not only to specialists and general clinicians but also multidisciplinary medical professionals to make appropriate decisions to improve the prognosis of patients with sepsis. In addition to the 26 committee members and directors in charge selected from both societies, we received the participation and support of a total of 226 individuals, comprising 85 working group members that included multiple professions (nine nurses, four physiotherapists, two clinical engineers, and two pharmacists) and those with a history of sepsis or critical illness (two, one of which was a nurse) and 115 systematic review members. The participation of multiple professions and experienced patients as working group members in particular expanded the perspective of our work and enabled a more flexible evaluation, which was a great step forward from the J-SSCG 2016. Furthermore, systematic reviews were conducted by the working group members and systematic review members, and there was a certain degree of independence from the committee members who formulated the recommendations.

Four new topics were incorporated in the J-SSCG 2020 in addition to the domains in the previously mentioned J-SSCG 2016: neuro-intensive care, patient- and family-centered care, sepsis treatment system, and stress ulcers. The J-SSCG 2020 also included a section on children after considering the fact that there are few pediatric intensive care units in Japan, and the situation is such that medical professionals who primarily treat adult sepsis patients must treat pediatric sepsis patients. With these additions, this guideline comprised a total of 22 topics and 118 CQs. The GRADE system was incorporated to prepare the recommendations, and the modified Delphi method was used to decide recommendations by voting from all committee members. Responses to the CQs were as follows: 79 GRADE-based recommendations, 5 Good Practice Statements (GPS), 18 expert consensuses, 27 answers to background questions (BQs), and definition and diagnosis of sepsis. We will also incorporate visual information for each CQ according to time axes such as medical care flow charts as a new attempt. Each CQ will be clinically positioned, and we will also distribute this as an app.

The J-SSCG 2020 original Japanese version was first released in the official society websites of the JSICM and JAAM in September 2020, followed by the publication in their official journals the Journal of JSICM [2021; Volume 28 (Supplement)] https://doi.org/10.3918/jsicm.27S0001 and Journal of Japanese Association for Acute Medicine [2021; Volume 32, S1] https://doi.org/10.1002/jja2.S0024 in February 2021. It was then translated into English and released on the societies’ websites in April, in advance of the simultaneous publication in their English-language official journals Journal of Intensive Care and Acute Medicine and Surgery.

Overview and basic principles of these guidelines

Name

The English name of this guideline is the Japanese Clinical Practice Guidelines for Management of Sepsis and Septic Shock 2020, and the abbreviation used was J-SSCG 2020 in consideration of the comparison made with the international version (SSCG).

Overall objective of this guideline

The objective of this guideline is to provide support for medical professionals to make appropriate decisions in order to improve the prognosis of patients in the clinical treatment of sepsis and septic shock.

Target patient populations

This guideline targets patients with or who are suspected of sepsis or septic shock, ranging from children to adults. This includes patients who receive diagnoses and treatment not only in the intensive care unit but also in the general ward and emergency outpatient departments. However, sepsis patients require advanced systemic management, so we emphasize that it is desirable for those with or who are strongly suspected of sepsis to be promptly transferred to intensive care units as circumstances allow and undergo management there.

Target users (users of this guideline)

All medical professionals such as specialists, general clinicians, nurses, pharmacists, physiotherapists, clinical engineers, and registered dietitians who are engaged in or involved in sepsis treatment.

Participation of representatives of associated expert groups and support for guideline creation experts

In addition to the 26 committee members and directors in charge selected from the Japanese Society of Intensive Care Medicine and the Japanese Association for Acute Medicine, the J-SSCG 2020 received the participation and support of a total of 226 individuals, comprising 85 working group members that included multiple professionals (nine nurses, four physiotherapists, two clinical engineers, and two pharmacists) and those who had an experience of sepsis or critical illness (two; one of which was a nurse) and 115 systematic review members.

As guideline creation experts, these individuals reviewed and confirmed the work process at each stage of the guideline creation process under the guidance of the EBM Medical Information Department of the Japan Council for Quality Health Care and in accordance with the principles of the GRADE system. Specialists from the EBM Medical Information Department participated in committee meetings and responded to questions from the guideline creation managers in order to directly solve problems.

Methods to reflect the values of the target populations (e.g., patients, general public)

Two medical professionals and researchers who had sepsis were added as committee members or working group members in order to reflect the values and hopes of patients and patient families. This point was considered useful in reflecting values and hopes from the position of patients and families after understanding the complexity, severity, and pathology of sepsis, which requires wide-ranging and advanced medical knowledge.

Peer review and public comments

Transparency during the creation of the J-SSCG 2020 was considered to be crucial. Official mailing lists (ML) were created for discussions among members of each team. Core members joined the MLs established by each team as read-only members. Through these measures, we aimed to increase the transparency of team discussions, and by implementing the appropriate interventions, we were able to coordinate the directions taken by each team and achieve consistency throughout the entirety of the guidelines. Mutual peer review was conducted for various work processes by external team members across the region. Work products from each group were repeatedly edited and revised, and each revised draft was discussed by the Guideline Creation Committee.

The initial draft of the CQs received public comments over the Internet. Answer for each CQ also had public comments. Public commenters were requested to disclose any conflicts of interest.

Disclosure of conflicts of interest (COIs) and members’ roles

Financial and academic COIs as well as the role(s) of each committee member are disclosed in the Additional file 1 (https://www.jsicm.org/pdf/guidelineEN/Additionalfile1.pdf). Financial COIs were disclosed in accordance with the standards used by the Japanese Association of Medical Sciences from 2017 through 2019.

Funding

These guidelines were prepared with financial support from the Japan Society of Intensive Care Medicine and the Japanese Association for Acute Medicine. No member of the Guideline Creation Committee received any form of financial compensation during the preparation of these guidelines. The views and interests of these societies were not reflected in the preparation of the guidelines’ recommendations.

Guideline dissemination strategy

The Japanese version of these guidelines is open access. To promote ease of use, the digest version of the guidelines booklet is available. In addition, the app version of the guideline is available for use to support the clinical setting. We will strive to make these guidelines available at various academic meetings and seminars and also monitor activities related to sepsis practice as well as the spread of these guidelines throughout the target medical community.

Planned revisions

These guidelines are scheduled to undergo revision every 4 years. The next revision will occur in 2024. Should important new information warranting revision be obtained beforehand, partial revision will be considered.

Methods used for creating this guideline

The J-SSCG 2020 was created through the three following processes: 1) planning a clinical question (CQ); 2) searching, collecting, and integrating evidence through a systematic review and evaluating its certainty; and 3) formulating a recommendation. Relevant information for a recommendation based on GRADE and expert consensus were available at https://www.jsicm.org/pdf/J-SSCG2020_supplementary_appendix01.pdf.

Planning a CQ

Clinical practice guidelines should cover the basic knowledge of clinical practice and contribute to the construction of a standard clinical practice system. For this reason, important CQs were extracted from each domain regardless of presence or absence of evidences, and important CQs taken up in previous guidelines were adopted in this guideline. Based on the rules of planning a CQ, committee members and working group members collaborated to create a draft CQ in their area of responsibility, an opinion extracted from mutual peer review by committee members was reflected, and a CQ list was created by the Guideline Creation Committee. Public comments were solicited online for these CQs. The CQs were then revised using these public comments received, and a total of 118 CQs were ultimately decided by the committee.

CQ classifications

CQs include background questions (BQs) and foreground questions. BQs indicate CQs that inquire about what is well known as general knowledge, such as diseases, diagnoses, and treatment. Meanwhile, foreground questions are CQs that inquire about information specialized to various situations in clinical settings and can influence decision-making in clinical practice (Table 1).

Table 1 CQ classifications

Formulating answers to BQs

BQs aim to present information that summarizes general knowledge such as illnesses, diagnoses, and treatment. Each area group prepared draft recommendations for the CQs, which were amended and revised repeatedly until the approval rate in the committee exceeded 95% for consensus.

Formulating answers to foreground questions

Foreground questions include (1) GPS, which are CQs that are extremely common and of which all medical personnel should be aware, and (2) CQs that are subject to systematic review and for which recommendations are formulated. The latter CQ was given a recommendation based on GRADE or on expert consensus depending on whether target articles were present or absent, respectively.

Formulating GPS

GPS was displayed for CQs, which handled themes that were extremely common and for which randomized controlled trials were theoretically impossible. These were amended and revised repeatedly until the approval rate in the committee exceeded 95% for consensus.

Searching, collecting, and integrating evidence through systematic review

A comprehensive literature review was conducted for each CQ in the foreground questions except for GPS, from which randomized controlled trials (RCTs) were extracted. As a general rule, the methodology was based on the Grading of Recommendations Assessment, Development and Evaluation (GRADE).

  • Step 1: Literature review

Literature reviews were conducted using the search engines of CENTRAL, PubMed, and Ichushi-Web.

The search equations were created by two or more independent reviewers using Medical Subject Headings (MeSH) terms and free search terms. Searches on PubMed used the sensitive-maximizing version of search strategies created by Cochrane as a general ruler for research design filters that specified RCTs. The publication date of the subject articles was not restricted. The languages of the manuscript were limited to Japanese and English. After confirming that the key RCTs specified in advance were included, the literature review equations underwent a final decision, and the literature review date and number of articles found in each search engine were recorded.

  • Step 2: Primary screening

All the titles and abstracts specified in Step 1 were downloaded. The automatic duplicate deletion function of the literature management software EndNote (Clarivate Analytics, USA) or Mendeley (Mendeley Ltd., UK) were used to remove duplicates, with duplicate articles further deleted manually. Article screening was conducted online using Rayyan (https://rayyan.qcri.org/welcome). Two independent reviewers reviewed the titles and abstracts of the literature and excluded research methods and PICO criteria, which were clearly not within the target. If there was any possibility that it was a target article, it was not excluded.

  • Step 3: Secondary screening

The full text of the remaining articles from Step 2 were ordered, and two reviewers selected articles whose research design and PICO criteria conformed to the CQ, and they confirmed them as target articles. Articles for which the opinions of the two reviewers did not match were sent to a third reviewer and discussed among the three reviewers. Articles excluded at this stage were provided a reason for exclusion. The process from literature review to target article selection is summarized in the PRISMA flow diagram.

  • Step 4: Evaluation of the certainty of evidence for CQs where evidence existed

Risk evaluations were conducted for the certainty of evidence (A-D) of the CQ undergoing systematic review for which each group was responsible. The definitions for the certainty of evidence as set by the GRADE system adopted in this guideline are as follows.

Definition of the certainty of evidence

  • High: Highly confident in the estimated value of effects

  • Medium: Moderate confidence in the estimated value of effects

  • Low: Limited confidence in the estimated value of effects

  • Very low: Almost no confidence in the estimated value of effects

  • Step 5: Data extraction, bias risk evaluation

Data extraction was performed by two independent reviewers, and a standardized data extraction form was used. In cases where insufficient information was recorded in the reference, this was stated as such, and the authors were not contacted.

  • Step 6: Meta-analysis and evaluation of the certainty of evidence

Qualitative and quantitative evaluations of the references to be adopted were performed. The qualitative evaluations used RevMan 5 whenever possible to conduct meta-analyses. This was summarized so that each area group could create evaluations of the certainty of evidence.

Handling of CQs with network meta-analysis

Indirect and network estimate values were calculated using a frequency-based analysis method for CQs with network meta-analyses (Confidence in Network Meta-Analysis [CINeMA] from R package netmeta used). The surface under the cumulative ranking curve (SUCRA) was used for rankings (calculated as Stata mvmeta command). The quality of evidence was evaluated based on the GRADE working group methods (ref). Network meta-analyses were conducted on CQ9–2 and CQ9–6 of this guideline.

Handling of CQs with qualitative research as evidence

The GRADE-Confidence in the Evidence from Reviews of Qualitative research (CERQual) approach was adopted as an evidence extraction method for CQs, where qualitative research was thought to be an appropriate research method. This was used in CQ20–3, “Should physical binding (restraints) be avoid during intensive care?”, in this guideline.

Formulation of proposed recommendations

The committee members and working group collaborated to create an evidence to decision (EtD) table in advance of deciding the recommendations. They then considered four factors (certainty of evidence, balance of effects, values, and cost/resource utilization) and formulated recommendations in consultation with the committee. The strengths of the recommendations shown in the GRADE system are classified as recommended, suggested, not suggested, and not recommended.

=Description methods for the strength of recommendations=

Strength of recommendation “1”: recommended.

Strength of recommendation “2”: suggested.

Committee members and the working group collaborated to create an EtD table for foreground question type CQs, for which insufficient evidence was obtained through comprehensive literature reviews conforming to the PICO criteria and formed an expert consensus based on this EtD. Recommendations in this EtD took into consideration the expert-proposed factors of the balance between the desired and undesired effects of each intervention, values, and costs/resource utilization, conducted in consultation with the committee. Recommendations with these expert consensuses were “suggestions”, and “(expert consensus: insufficient evidence)” was added at the end of the text so that this could be distinguished from the above-mentioned recommendations based on GRADE.

Consensus building in CQs in accordance with GRADE and CQs showing expert consensus

The modified Delphi method was used for consensus building among committee members.

  • Step 1: Voting

Each committee member anonymously voted online in an independent manner using a point system ranging from 1 to 9 (1: disagree, 9: agree). The median, interpercentile range (IPR), interpercentile range adjusted for symmetry (IPRAS), and disagreement index (DI) of the obtained scores were calculated.

  • Step 2: Panel meeting

Panel meetings were conducted based on the aggregated results as shown below to reach a consensus.

  1. 1.

    When median < 7.5 and DI ≥0.2

Discussions were held within the committee, after which amendments were made to the EtD and recommended text, and a second vote was held.

  1. 2.

    When median ≥ 7.5 or DI < 0.2

    1. A

      When a serious opinion was present during voting for a comment or recommendation presented by committee member

      Discussions were held within the committee, and a consensus was reached. CQs for which a consensus was not reached within the committee resulted in amendments to the EtD and recommended text, after which a second vote was held.

    2. B

      When no serious opinions were present during voting for a comment or recommendation presented by a committee member.

The voting results were confirmed among the committee members, and a consensus was reached.

Quick reference list of CQ&As

CQ1: Definition and diagnosis of sepsis

CQ1-1: Definition of sepsis

Summary: According to the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3), sepsis is defined as “life-threatening organ dysfunction caused by a dysregulated host response to infection.” Septic shock is defined as a subset of sepsis in which the underlying circulatory and cellular/metabolic abnormalities profoundly increase the risk of mortality.

CQ1-2: Diagnosis of sepsis and septic shock

Summary: A diagnosis of sepsis is confirmed when the Sequential Organ Failure Assessment (SOFA) score of 2 points or more acutely increase in the presence of a clear infection or suspected infection. Patients with septic shock can be identified with a clinical construct of sepsis with persisting hypotension requiring vasopressors to maintain mBP ≥ 65 mmHg and having a serum lactate level > 2 mmol/L (18 mg/dL) despite adequate volume resuscitation. In out-of-hospital, emergency department, or general hospital ward settings, adult patients with suspected infection can be rapidly identified as more likely to have poor outcomes typical of sepsis if they have at least two of the following clinical criteria that together constitute the quick SOFA (qSOFA) score: a respiratory rate of 22 breaths/min or higher, altered consciousness, and a systolic blood pressure of ≤100 mmHg. The qSOFA criteria can be used to prompt clinicians to further investigate organ dysfunction, initiate or escalate therapy as appropriate, and to consider referral for critical care. Ultimately, an acutely increased SOFA score of 2 or more points confirms the diagnosis of sepsis. Daily routine screening for sepsis is recommended to support the early diagnosis and treatment of sepsis.

CQ2: Diagnosis of infection

CQ2-1: When should a blood culture be taken?

Answer: Take two or more sets before administering the antibacterial drug (Good Practice Statement).

CQ2-2: When should culture specimens other than blood be collected?

Answer: Each cultured specimen other than blood should be collected as needed prior to the administration of antibacterial drugs (Good Practice Statement).

CQ2-3: Is Gram staining useful in the selection of antimicrobial agents before obtaining culture results?

Answer: We suggest referencing Gram staining findings of the culture specimen when selecting an antibacterial drug to use for empirical treatment (expert consensus: insufficient evidence).

CQ2–4-1: What are the positions of C-reactive protein (CRP), procalcitonin (PCT), presepsin (P-SEP), and interleukin 6 (IL-6) as biomarker tests for sepsis diagnosis in general ward and emergency rooms (ER)?

Answer: Sensitivity and specificity in biomarker tests when sepsis was suspected in general ward and ER visits were as follows: CRP, 59, 79%; PCT, 74, 81%; P-SEP, 75, 74%; IL-6, 78, 78%. As such, sepsis diagnosis with biomarkers alone is generally thought to be difficult, and its use should be seen as supplemental to any observations of general conditions (Provision of information for background question).

CQ2–4-2: What are the positions of C-reactive protein (CRP), procalcitonin (PCT), presepsin (P-SEP), and interleukin-6 (IL-6) as biomarker tests for sepsis diagnosis in the intensive care unit?

Answer: Sensitivity and specificity in biomarker tests when sepsis was suspected in the intensive care unit were as follows: CRP, 74, 70%; P-SEP, 82, 73%; IL-6, 72, 76%. As such, sepsis diagnosis with biomarkers alone is generally thought to be difficult, and its use should be supplemental to any observations of general conditions (Provision of information for background question).

CQ3: Source control

CQ3-1: Should imaging tests be conducted in patients suspected of sepsis in order to search for the source of infection?

Answer: Imaging tests should be conducted when the source of infection is unclear in order to search for the source of infection (Good Practice Statement).

CQ3-2: Should whole-body contrast-enhanced CT tests be conducted at an early stage for sepsis patients with unknown source of infection?

Answer: We suggest conducting whole-body contrast-enhanced CT tests as soon as possible for sepsis patients with unknown source of infection (expert consensus: insufficient evidence).

CQ3-3: Should the source of infection be controlled by surgery/invasive drainage in patients with sepsis due to intraperitoneal infection?

Answer: We suggest controlling the source of infection as soon as possible with surgery/invasive drainage (including abscess drainage, biliary tract/gallbladder drainage) for patients with sepsis due to intraperitoneal infection (expert consensus: insufficient evidence).

CQ3-4-1: Should the source of infection be controlled with invasive interventional therapy during the early period of infectious pancreatic necrosis?

Answer: We suggest against controlling the source of infection with invasive interventional therapy during the early period of infectious pancreatic necrosis (GRADE 2C: certainty of evidence = “low”).

CQ3-4-2: Should the source of infection be controlled with low-invasive interventional therapy for infectious pancreatic necrosis?

Answer: We recommend controlling the source of infection with less invasive interventional therapy for patients with sepsis caused by infectious pancreatic necrosis (GRADE 2B: certainty of evidence = “moderate”).

CQ3-5: Should the source of infection be controlled with invasive drainage for patients with sepsis due to acute pyelonephritis caused by ureteral obstruction?

Answer: We suggest controlling the source of infection as soon as possible with transurethral ureteral stent implantation or percutaneous nephrostomy in patients with sepsis due to acute pyelonephritis caused by ureteral obstruction (expert consensus: insufficient evidence).

CQ3-6: Should source control be achieved by means of surgical debridement for sepsis patients due to necrotic soft tissue infection?

Answer: We suggest controlling the source of infection as soon as possible by means of surgical debridement for sepsis patients due to necrotic soft tissue infection (expert consensus: insufficient evidence).

CQ3-7: Should the source of infection be controlled with catheter removal in patients with sepsis where catheter-related bloodstream infections are suspected?

Answer: We suggest controlling the source of infection as soon as possible with catheter removal in patients with sepsis where catheter-related bloodstream infections are suspected (expert consensus: insufficient evidence).

CQ3-8: Should the source of infection be controlled through invasive drainage in patients with sepsis due to empyema?

Answer: We suggest controlling the source of infection as soon as possible with percutaneous thoracic drainage or surgical intervention in patients with sepsis due to empyema (expert consensus: insufficient evidence).

CQ4: Antimicrobial therapy

CQ4-1: How should empirical antimicrobial therapy be selected?

Answer: Antimicrobials can be selected by estimating the causative microorganism based on suspected infectious foci, patient background, epidemiology and rapid microbial diagnostic tests, and by considering the tissue penetration properties of drugs and the probabilities of resistant bacteria (see Table 2 for reference). (Provision of information for background question).

Table 2 Empiric therapeutic agents for each infectious disease

CQ4-2: Under what circumstances should carbapenems be used in empirical antimicrobial therapy?

Answer: Carbapenems can be included in the empirical antimicrobial regimen when the use of carbapenem is considered to be particularly effective; ESBL-producing Enterobacteriaceae or Pseudomonas aeruginosa or Acinetobacter species with limited susceptibility for carbapenems (Provision of information for background question).

CQ4–3: Under what circumstances should empirical antimicrobial therapy be selected for MRSA and non-bacterial pathogens (e.g., Candida, Viruses, Legionella, Rickettsia, or Clostridioides difficile )?

Answer: Each microorganism can be covered by empirical antimicrobial regimen if highly suspected by suspected infectious foci, patient background and culture results (Provision of information for background question).

CQ4-4: Should empirical antimicrobial therapy be suspended if culture results were negative?

Answer: We suggest stopping any empiric antimicrobials where sepsis is excluded by negative culture results and after careful consideration of clinical progress (expert consensus: insufficient evidence).

CQ4-5: Under what circumstances should an infectious disease specialist or antimicrobial stewardship team be consulted?

Answer: An infectious disease specialist and/or antimicrobial stewardship team can be consulted when 1) the cause of sepsis is unknown, 2) involvement of extensively drug-resistant bacteria is suspected, 3) emerging, re-emerging, or imported infectious diseases are suspected, or 4) in cases of Staphylococcus aureus bacteremia or candidemia (Provision of information for background question).

CQ4-6: Should empirical antibacterial drugs for sepsis begin within 1 h upon identification of sepsis?

Answer: We suggest that antibacterial drugs be administered as soon as possible upon identification of sepsis or septic shock, but we suggest against using the target time of less than 1 h (GRADE 2C: certainty of evidence = “low”).

CQ4-7: Should continuous or extended infusion of β-lactam antibiotics be used for sepsis?

Answer: We suggest using continuous or extended infusion of β-lactam antimicrobials (GRADE 2B: certainty of evidence = “moderate”).

CQ4-8: Should de-escalation antimicrobial therapy be used for sepsis?

Answer: We suggest applying de-escalation antimicrobial therapy for sepsis (GRADE 2D, certainty of evidence = “very low”).

CQ4-9: Should procalcitonin be used as an indicator for stopping antimicrobial therapy for sepsis?

Answer: We suggest using procalcitonin as an indicator for stopping antimicrobial therapy for sepsis (GRADE 2B, certainty of evidence = “moderate”).

CQ4-10: Should relatively short-term (i.e. within 7 days) antimicrobial therapy be applied for sepsis?

Answer: We suggest applying relatively short-term (i.e. within 7 days) antimicrobial therapy for sepsis (GRADE 2D: certainty of evidence = “very low”).

CQ4-11: What should be used as a reference for adjusting the dose for renal-excretion antimicrobial drugs?

Answer: Changes in bodily fluid volume and the presence of renal replacement therapy and other extracorporeal circulation therapies in addition to renal function test values (e.g., serum Cr level, eGFR level) measured at multiple time points are informative (Provision of information for background question).

CQ5: Intravenous immunoglobulin therapy

CQ5-1: Should intravenous immunoglobulin (IVIG) be administered to adult patients with sepsis?

Answer: We suggest against administering IVIG to patients with sepsis (GRADE 2B: certainty of evidence = “moderate”).

CQ5-2-1: Should IVIG be administered to patients with streptococcal toxic shock syndrome (STSS)?

Answer: We suggest administering IVIG to patients with STSS (GRADE 2D: certainty of evidence = “very low”).

CQ5-2-2: Should IVIG be administered to patients with staphylococcal toxic shock syndrome (staphylococcal TSS)?

Answer: We suggest against administering IVG to patients with staphylococcal TSS (expert consensus: insufficient evidence).

CQ6: Initial resuscitation/inotropes

CQ6-1: Should echocardiography be conducted in patients with sepsis?

Answer: We suggest, following initial fluid resuscitation, conducting cardiac function and hemodynamics assessments with echocardiography in patients with sepsis/septic shock (GRADE 2D: certainty of evidence = “very low”).

CQ6-2: Is EGDT recommended for initial resuscitation in patients with sepsis?

Answer: We suggest against conducting EGDT as initial resuscitation in patients with sepsis/septic shock (GRADE 2C: certainty of evidence = “low”).

CQ6-3: Should vasopressors be used simultaneously or in the early stage (within 3 h) of initial fluid resuscitation in adult patients with sepsis?

Answer: We suggest administering vasopressors simultaneously or in the early stages (within 3 h) of initial fluid resuscitation in patients with sepsis/septic shock who have difficult maintaining hemodynamics (GRADE 2C: certainty of evidence = “low”).

CQ6-4: Should lactate levels be used as an indicator for initial resuscitation in adult patients with sepsis?

Answer: We suggest using lactate levels as an indicator of tissue hypoperfusion during initial resuscitation in patients with sepsis/septic shock (GRADE 2C: certainty of evidence = “low”).

CQ6-5: What is the initial fluid infusion rate and volume in adult patients with sepsis?

Answer: There is an opinion that the initial fluid resuscitation in patients with reduced intravascular volume due to sepsis should be administered over 30 mL/kg of crystalloid solution within 3 h, aiming to optimize the circulating blood volume. It is important during initial fluid resuscitation to carefully observe vital signs and to avoid excessive fluid loads by using lactate clearance and echocardiography while conducting tissue oxygen metabolism and hemodynamics assessments (Provision of information for background question).

CQ6-6: How should fluid responsiveness be assessed in adult patients with sepsis?

Answer: Fluid responsiveness is significant increase in stroke volume (SV) after fluid infusion, and multiple parameters, including static and dynamic parameters, should be used to predict fluid responsiveness. Static parameters, including central venous pressure (CVP) and pulmonary capillary wedge pressure (PCWP), are measured at a point. Dynamic parameters include changes in cardiac output by passive leg raising (PLR) and fluid challenge, pulse pressure variation (PPV) and stroke volume variation (SVV) during mechanical ventilation (Provision of information for background question).

CQ6-7: Should albumin solution be used for initial resuscitation in adult patients with sepsis?

Answer: We suggest against administering albumin solution as a standard treatment at the beginning of initial fluid resuscitation in patients with sepsis (GRADE 2C: certainty of evidence = “low”). Albumin solution can be used in patients with sepsis when patients do not respond to standard treatment and require substantial amounts of crystalloids (expert consensus: insufficient evidence).

CQ6-8: Should artificial colloids be used for initial resuscitation in adult patients with sepsis?

Answer: We suggest against administering artificial colloids in patients with sepsis/septic shock (GRADE 2D: certainty of evidence = “very low”).

CQ6-9-1: Should noradrenaline, dopamine, or phenylephrine be used as a first-line vasopressor in adult patients with sepsis? noradrenaline vs. dopamine

Answer: Between noradrenaline and dopamine, we suggest administering noradrenaline as a first-line vasopressor in adult patients with sepsis (GRADE 2D: certainty of evidence = “very low”).

CQ6-9-2: Should noradrenaline, dopamine, or phenylephrine be used as a first-line vasopressor in adult patients with sepsis? noradrenaline vs. phenylephrine

Answer: Between noradrenaline and phenylephrine, we suggest administering noradrenaline as a first-line vasopressor in adult patients with sepsis (GRADE 2D: certainty of evidence = “very low”).

CQ6-10-1: Should adrenaline be used as a second-line vasopressor in adult patients with sepsis?

Answer: We suggest against using adrenaline as a second-line vasopressor in patients with sepsis/septic shock (GRADE 2D: certainty of evidence = “very low”).

CQ6-10-2: Should vasopressin be used as a second-line vasopressor in adult patients with sepsis?

Answer: We suggest using vasopressin as a second-line vasopressor in patients with sepsis/septic shock (GRADE 2D: certainty of evidence = “very low”).

CQ6-11: Should inotropes be used in adult patients with sepsis accompanied by cardiogenic shock?

Answer: We suggest administering inotropes (adrenaline, dobutamine) in adult patients with septic shock accompanied by cardiac dysfunction (expert consensus: insufficient evidence).

CQ6-12: Should β-blockers be used in adult patients with sepsis?

Answer: We suggest administering short-acting β1-adrenoceptor antagonists in patients with sepsis/septic shock while being monitored with the objectives of managing tachycardia which cannot be controlled with standard therapy like initial fluid resuscitation (GRADE 2D: certainty of evidence = “very low”). Administering short-acting β1-adrenoceptor antagonists can induce hemodynamic fluctuations, so they should be administered under the supervision of a physician with expertise in cardiovascular management in the intensive care unit (expert consensus: insufficient evidence).

CQ6-13: What are the indications of assisted circulation in adult patients with septic shock?

Answer: There is insufficient evidence for the effects of assisted circulation such as veno-arterial extracorporeal membrane oxygenation (V-A ECMO) and intra-aortic balloon pump (IABP) for cardiac dysfunction in septic shock, and its applications are still under investigation (Provision of information for background question).

CQ7: Corticosteroid therapy

CQ7-1: Should low-dose corticosteroids (hydrocortisone) be administered to adult patients with septic shock who do not respond to initial fluid resuscitation and vasopressors?

Answer: We suggest administering low-dose corticosteroids (hydrocortisone) to adult patients with septic shock who do not respond to initial fluid resuscitation and vasopressors for the purpose of withdrawing from shock (GRADE 2D: certainty of evidence = “very low”).

CQ7-2: Should hydrocortisone and fludrocortisone be administered to patients with septic shock who do not respond to initial fluid resuscitation and vasopressors?

Answer: We suggest concomitant administration of hydrocortisone and fludrocortisone to adult patients with septic shock who do not respond to initial fluid resuscitation and vasopressors (GRADE 2C: certainty of evidence = “low”).

CQ7-3: Should corticosteroids (hydrocortisone) be administered to patients with sepsis without shock?

Answer: We suggest against administering hydrocortisone to patients with sepsis without shock (GRADE 2D: certainty of evidence = “very low”).

CQ8: Blood transfusion therapy

CQ8-1: How should blood transfusion be conducted during the initial resuscitation of septic shock?

Answer: We suggest starting blood transfusion at a hemoglobin level of less than 7 g/dL during initial resuscitation for patients with septic shock (GRADE 2C: certainty of evidence = “low”).

CQ8-2: How should blood transfusion be conducted during hemodynamically stable sepsis?

Answer: We suggest starting blood transfusion at a hemoglobin level of less than 7 g/dL in patients with hemodynamically stable sepsis (expert consensus: insufficient evidence).

CQ8-3: How should fresh frozen plasma be administered in patients with sepsis?

Answer: We suggest administering fresh frozen plasma in patients with sepsis when hemorrhaging tendencies are observed. If surgical/invasive interventions are required, we suggest administering when PT/APTT is extended (PT is over INR 2.0 or activity level of less than 30%; APTT is over two times the upper limit of standards at each medical institution or activity level less than 25%) or when fibrinogen levels are less than 150 mg/dL (expert consensus: insufficient evidence).

CQ8-4: How should platelet transfusion be conducted for patients with sepsis?

Answer: We suggest conducting platelet transfusion in patients with sepsis and platelet counts of less than 10,000/μL, or less than 50,000/μL when accompanied by hemorrhaging symptoms (expert consensus: insufficient evidence). We suggest conducting platelet transfusion so as to maintain a platelet count of over 50,000/μL when active hemorrhaging is observed or when surgical/invasive procedures are needed (expert consensus: insufficient evidence).

CQ9: Respiratory management

CQ9-1: What is the S P O 2 range for respiratory management in adult patients with sepsis?

Answer: We suggest against setting a high target SPO2 (98–100%) during respiratory management in adult patients with sepsis (GRADE 2B: certainty of evidence = “moderate”).

Remarks: This does not apply in cases where there is the possibility of a disruption in the oxygen supply/demand balance due to severe anemia or increased metabolism due to infection in cases where hemodynamics are unstable.

CQ9-2: Should non-invasive ventilation (NIV) or nasal high-flow therapy (NHFT) be conducted for early respiratory failure in adult patients with sepsis?

Answer: We suggest conducting non-invasive ventilation (NIV) or nasal high-flow therapy (NHFT) for early respiratory failure in adult patients with sepsis (GRADE 2A: certainty of evidence = “high”).

CQ9-3: Should protective ventilation strategies be implemented for ventilation management in adult patients with sepsis?

Answer: We suggest implementing protective ventilation strategies for ventilation management in adult patients with sepsis (GRADE 2B: certainty of evidence = “moderate”).

CQ9-4: Should high PEEP settings be utilized for ventilation management in adult patients with sepsis?

Answer: We suggest against utilizing high PEEP settings (PEEP over 12 cm H2O) for the initial stage of ventilation management in adult patients with sepsis (GRADE 2B: certainty of evidence = “very low”).

CQ9-5: Should spontaneous breathing trials (SBT) be conducted prior to extubation in adult patients with sepsis placed under ventilation management?

Answer: We suggest utilizing weaning protocols from ventilators, including spontaneous breathing trials (SBTs) prior to extubation in adult patients with sepsis placed under ventilation management (GRADE 2D: certainty of evidence = “very low”).

CQ9–6: Should preventative non-invasive ventilation (NIV) or nasal high-flow therapy (NHFT) be conducted after extubation for adult patients with sepsis placed under ventilation management?

Answer: We suggest conducting preventative non-invasive ventilation (NIV) or nasal high-flow therapy (NHFT) over standard oxygen therapy following extubation for adult patients with sepsis placed under ventilation management (GRADE 2B: certainty of evidence = “moderate”).

CQ10: Management of pain, agitation, and delirium

CQ10-1: Should management based on analgesia-first sedation protocol be used for adult patients with sepsis on mechanical ventilation?

Answer: We suggest using management based on analgesia-first sedation protocol in adult patients with sepsis on mechanical ventilation (GRADE 2C: certainty of evidence = “low”).

CQ10-2: Should propofol or dexmedetomidine be prioritized over benzodiazepines as sedatives for adult patients with sepsis on mechanical ventilation?

Answer: We suggest using propofol or dexmedetomidine over benzodiazepines as sedatives for patients with sepsis on mechanical ventilation (GRADE 2D: certainty of evidence = “very low”).

CQ10–3: Should light sedation through the interruption of sedatives once a day or sedative adjustments based on protocol be used for adult patients with sepsis on mechanical ventilation?

Answer: We suggest using light sedation through the interruption of sedatives once a day or sedative adjustments based on protocol for patients with sepsis on mechanical ventilation (GRADE 2C: certainty of evidence = “low”).

CQ10-4: Should drug therapy be used to prevent delirium in adult patients with sepsis?

Answer: We suggest administering dexmedetomidine for delirium prevention in adult patients with sepsis (GRADE 2C: certainty of evidence = “low”). We suggest against the administration of haloperidol (GRADE 2B: certainty of evidence = “moderate”). We suggest against the administration of atypical antipsychotics (GRADE 2C: certainty of evidence = “low”). We suggest against the administration of statins (GRADE 2D: certainty of evidence = “very low”).

Remarks: We recommend against the routine administration of dexmedetomidine to patients who do not require sedation. Furthermore, dexmedetomidine administration can cause hemodynamic fluctuations, so this should ideally be administered under the supervision of a physician who is experienced with systematic management in an intensive care unit (expert consensus).

CQ10-5: Should drug therapy be used to treat delirium in adult patients with sepsis?

Answer: We suggest against administering dexmedetomidine for delirium treatment in adult patients with sepsis (GRADE 2D: certainty of evidence = “very low”). We suggest against administering haloperidol (GRADE 2C: certainty of evidence = “low”). We suggest against administering atypical antipsychotics (GRADE 2B: certainty of evidence = “moderate”).

Remarks: The use of dexmedetomidine, haloperidol, or atypical antipsychotics should not be prevented when the patient’s life or body is at risk due to hyperactive delirium.

CQ10-6: Should non-drug therapy be used to prevent delirium in adult patients with sepsis?

Answer: We suggest using non-drug therapy to prevent delirium in adult patients with sepsis (GRADE 2C: certainty of evidence = “low”).

CQ11: Acute kidney injury/blood purification

CQ11-1: Should furosemide be used to prevent or treat septic AKI?

Answer: We suggest against using furosemide for preventing or treating septic AKI (GRADE 2C, certainty of evidence = “low”).

CQ11-2: Should atrial natriuretic peptide (ANP) be used to prevent or treat septic AKI?

Answer: We suggest against using ANP for preventing or treating septic AKI (GRADE 2D, certainty of evidence = “very low”).

CQ11-3: Should dopamine be used to prevent or treat septic AKI?

Answer: We suggest against using dopamine for preventing or treating septic AKI (GRADE 2C, certainty of evidence = “low”).

CQ11-4: Should continuous renal replacement therapy (RRT) rather than intermittent RRT be used for the management of septic AKI?

Answer: Either continuous or intermittent RRT can be selected for septic AKI (GRADE 2C, certainty of evidence = “low”). Continuous RRT should be used for hemodynamically unstable patients (Good Practice Statement).

CQ11-5-1: Should RRT be initiated early for septic AKI (Stage 2 vs. Stage 3 or absolute indications)?

Answer: We make no recommendation on whether RRT should be initiated early at Stage 2 for patients with septic AKI.

CQ11-5-2: Should RRT be initiated early for septic AKI (Stage 3 vs. absolute indications)?

Answer: We suggest against initiating RRT at Stage 3 for patients with septic AKI rather than absolute indication (GRADE 2D, certainty of evidence = “very low”).

CQ11-6: Should a large RRT dose be delivered for septic AKI?

Answer: We suggest against increasing a RRT dose beyond the standard dose for patients with septic AKI (GRADE 2C, certainty of evidence = “low”).

CQ11-7: Should PMX-DHP be used for patients with septic shock?

Answer: We suggest against using PMX-DHP for patients with septic shock (GRADE 2B, certainty of evidence = “moderate”).

CQ12: Nutrition support therapy

CQ12-1: Should either enteral nutrition or parenteral nutrition be given for nutrition administration in septic patients?

Answer: We suggest enteral nutrition be administered for septic patients. (GRADE 2D: certainty of evidence = “very low”).

CQ12-2: Should hemodynamically unstable septic shock patients receive enteral nutrition?

Answer: We suggest against administering enteral nutrition in hemodynamically unstable septic shock patients (GRADE 2D: certainty of evidence = “very low”).

CQ12-3: When should enteral nutrition be initiated in septic patients?

Answer: We suggest initiating enteral nutrition at an early period of acute phase (within 24–48 h following the start of treatment to critical illness) for septic patients (GRADE 2D: the certainty of evidence = “very low”).

CQ12-4: Should the septic patients receive enteral nutrition less than their energy expenditure in the acute phase?

Answer: We suggest the septic patients receive enteral nutrition less than their energy expenditure in the acute phase. (GRADE 2B: certainty of evidence = “moderate”).

CQ12-5: Should parenteral nutrition be combined with enteral nutrition in septic patients?

Answer: We suggest supplemental parenteral nutrition be combined in septic patients receiving insufficient amount of enteral nutrition (GRADE 2D: certainty of evidence = “very low”).

CQ12-6: What is the optimal protein dose in the acute phase for septic patients?

Answer: We suggest providing less than 1 g/kg/day of protein (peptides, amino acids) to septic patients in the acute phase (GRADE 2D: certainty of evidence = “very low”).

CQ12-7-1: Should vitamin C be actively provided to septic patients in the acute phase?

Answer: We suggest providing vitamin C to septic patients (GRADE 2D: certainty of evidence = “very low”).

CQ12-7-2: Should vitamin D be actively provided to septic patients in the acute phase?

Answer: We suggest against providing vitamin D in septic patients (GRADE 2D: certainty of evidence = “very low”).

CQ12-8: What are the methods for determining enteral nutrition initiation and monitoring intolerance in septic patients?

Answer: Findings such as bowel sounds, which indicate contractility of the gastrointestinal tract, at the start of enteral nutrition should not be required. Meanwhile, various findings show intolerance following the initiation of enteral nutrition, including the lack of intestinal sounds, abnormal intestinal sounds, vomiting, intestinal dilation, diarrhea, gastrointestinal bleeding, and excessive gastric residue. Excessive gastric residue suggests intolerance, but the gastric residue volume criteria for determining the presence of intolerance are unknown (Provision of information for background question).

CQ12-9: What nutrition support therapy should be provided to septic patients after the acute phase?

Answer: Provision of energy that meets the goals (around 25–30 kcal/kg/day, including protein) is thought to be needed when the patients overcome the clinical conditions of acute phase, or where about 1 week has passed following the onset of critical illness. Some experts are of the opinion that protein dose of over 1 g/kg/day is ideal in this phase. However, there are other expert opinions that the energy dose should be increased at an earlier phase for patients with malnutrition prior to exacerbation of the disease (Provision of information for background question).

CQ13: Blood glucose management

CQ13-1: Should blood glucose be measured using a glucometer with capillary blood in septic patients?

Answer: We suggest against the use of a glucometer with capillary blood in patients with sepsis (GRADE 2A: certainty of evidence = “high”).

CQ13-2: What is the optimal blood glucose target level in septic patients?

Answer: We suggest an optimal target blood glucose range of 144–180 mg/dL in septic patients (GRADE 2D: certainty of evidence = “very low”).

CQ14: Body temperature control

CQ14-1: Should antipyretic therapy be applied to sepsis patients presenting with fever?

Answer: We suggest against conducting antipyretic therapy to sepsis patients presenting with fever (GRADE 2A: certainty of evidence = “high”).

CQ14-2: Should rewarming therapy be applied to hypothermic sepsis patients?

Answer: We suggest attempting to correct the body temperature of hypothermic (core body temperature < 35 °C) sepsis patients while considering hemodynamic stabilization when hemodynamic disorders and coagulation abnormalities related to hypothermia are observed (expert consensus: insufficient evidence).

CQ15: Diagnosis and treatment of disseminated intravascular coagulation in patients with sepsis

CQ15-1: What is the diagnosis method for septic disseminated intravascular coagulation (DIC)?

Answer: There are multiple diagnostic criteria for conducting DIC diagnosis. The acute DIC diagnostic criteria are widely used in Japan, while the ISTH overt-DIC is used as the international standard. It is difficult to determine the superiority between diagnostic criteria, and these should be used according to the purpose (Provision of information for background question).

CQ15-2: What are differential diseases for patients where septic DIC is suspected?

Answer: Thrombotic thrombocytopenic purpura (TTP), hemolytic uremic syndrome (HUS) and heparin-induced thrombocytopenia (HIT) are common DIC-like pathological conditions. These types of diseases require managements that are different from that of DIC (Provision of information for background question).

CQ15-3: Should antithrombin replacement therapy be administered in sepsis-associated DIC?

Answer: We suggest antithrombin replacement therapy for patients with sepsis-associated DIC (GRADE 2C, certainty of evidence = “low”).

CQ15-4: Should heparin or heparin analogs be administered in sepsis-associated DIC?

Answer: We suggest against administering heparin or heparin analogs as a standard treatment for patients with sepsis-associated DIC (GRADE 2D, certainty of evidence = “very low”).

CQ15-5: Should recombinant thrombomodulin be administered to patients with sepsis-associated DIC?

Answer: We suggest administering recombinant thrombomodulin for patients with sepsis-associated DIC (GRADE 2C, certainty of evidence = “low”).

CQ15-6: Should protease inhibitors be administered to patients with sepsis-associated DIC?

Answer: We suggest against administering protease inhibitors as standard treatment for patients with sepsis-associated DIC (GRADE 2D, certainty of evidence = “very low”).

CQ16: Venous thromboembolism countermeasures

CQ16-1: Should mechanical prophylaxis (elastic stockings, intermittent pneumatic compression) be used to prevent deep vein thrombosis during sepsis?

Answer: We suggest using mechanical prophylaxis (elastic stockings, intermittent pneumatic compression) to prevent deep vein thrombosis in patients with sepsis (expert consensus: insufficient evidence).

CQ16-2: Should anticoagulation therapy (unfractionated heparin, low-molecular-weight heparin, warfarin, NOAC/DOAC) be conducted to prevent deep vein thrombosis during sepsis?

Answer: We suggest conducting anticoagulation therapy to prevent deep vein thrombosis in patients with sepsis (expert consensus: insufficient evidence).

CQ16-3: For how long should VTE prophylaxis be conducted in patients with sepsis?

Answer: We suggest conducting venous thromboembolism (VTE) prophylaxis in patients with sepsis until they are able to walk or discharged from the hospital (expert consensus: insufficient evidence).

CQ17: ICU-acquired weakness and early rehabilitation

CQ17-1: Should early rehabilitation be implemented to prevent PICS?

Answer: We suggest conducting early rehabilitation to prevent PICS in patients with sepsis (GRADE 2D, certainty of evidence = “very low”).

CQ17-2: Should passive joint exercise therapy be conducted to prevent ICU-AW in patients with sepsis?

Answer: We suggest conducting passive mobilization as standard treatment for patients with sepsis (GRADE 2D: certainty of evidence = “very low”).

CQ17-3: Should neuromuscular electrical stimulation be used to prevent ICU-AW?

Answer: We suggest against using neuromuscular electrical stimulation as a standard treatment to prevent ICU-AW in patients with sepsis (GRADE 2D: certainty of evidence = “very low”).

CQ18: Pediatric considerations

CQ18-1: Should the initial resuscitation algorithm be used for pediatric sepsis?

Answer: We suggest using the initial resuscitation algorithm for pediatric sepsis (GRADE 2D: certainty of evidence = “very low”).

CQ18-2: How should empirical antibacterial drugs be selected for pediatric sepsis where the source of infection is difficult to estimate?

Answer: Antibacterial drugs which cover the possible microorganisms should be selected with consideration of the site of occurrence (e.g., community, hospital, intensive care unit) and patient background (e.g., immune status, treatment history) (see Table 3 for reference) (Provision of information for background question).

Table 3 Thresholds and limits of dynamic indicators

CQ18-3: Under what scenarios should anti-herpetic agents be included in empirical treatment for pediatric sepsis?

Answer: There are cases where a central nervous system infection is suspected or a bacterial source of infection cannot be specified in neonates, because the prevalence of the herpes simplex virus is higher and they can easily become severe once infected (Provision of information for background question).

CQ18-4: What is the optimal blood pressure for hemodynamic management in pediatric sepsis?

Answer: Suitable values for the optimal blood pressure are unknown, and this should be set with consideration to age and organ perfusion. The median value for the mean blood pressure “55 + age x 1.5 mmHg” and the 5th percentile value “40 + age x 1.5 mmHg” in healthy children are used as a reference (Provision of information for background question).

CQ18-5: What is the method for assessing fluid responsiveness during the management of pediatric sepsis?

Answer: Assessments for fluid responsiveness include clinical findings (changes in pulse rate, blood pressure, temperature difference between peripheral and central skins, strength of pulsation, and capillary refill time (CRT)) and test values (e.g., lactate clearance, echocardiography findings) (Provision of information for background question).

CQ18-6: What is the initial fluid infusion rate and volume for pediatric sepsis?

Answer: In children with sepsis not complicated by heart failure, there is a method for repeating a bolus administration 10–20 mL/kg at a time while assessing response to an initial fluid resuscitation. Meanwhile, the occurrence of clinical findings which suggest fluid overload or a blunted fluid response should serve as a reference for suspending fluid resuscitation. There is no high-quality evidence regarding the upper limits of fluid infusion rate or volume (Provision of information for background question).

CQ18-7: Should dopamine be used as a first-line vasoactive agent in children with septic shock?

Answer: We suggest against using dopamine ad a first-line vasoactive agent in children with septic shock, and instead suggest selecting either adrenaline or noradrenaline according to hemodynamics (for adrenaline - GRADE 2D: certainty of evidence = “very low”; for noradrenaline - expert consensus: insufficient evidence).

CQ18-8: Should vasopressin be used as a vasoactive agent in children with septic shock?

Answer: We suggest against using vasopressin as a vasoactive agent in children with septic shock (GRADE 2D: certainty of evidence = “very low”).

CQ18-9: Should corticosteroids be administered to children with septic shock when they do not respond to initial fluid resuscitation and inotropic agents?

Answer: We suggest against the routine administration of corticosteroids in children with septic shock when they do not respond to initial fluid resuscitation and inotropic agents (GRADE 2D: certainty of evidence = “very low”).

CQ18-10: When should blood infusions be started in hemodynamically stable children with sepsis?

Answer: We suggest starting blood transfusions with a hemoglobin level of 7.0 g/dL as a threshold for critical, hemodynamically stable children with sepsis (GRADE 2C: certainty of evidence = “low”).

CQ18-11: Should blood purification therapy (including plasma exchange) be used to treat children with sepsis without acute kidney injury?

Answer: We suggest against using blood purification therapy to treat children with sepsis without acute kidney injury (GRADE 2D: certainty of evidence = “very low”).

CQ18-12: Should intravenous immunoglobulin (IVIG) therapy be administered in children with sepsis?

Answer: We suggest against administering IVIG for children with sepsis (expert consensus: insufficient evidence).

CQ18-13: Should blood glucose level be controlled tightly in children with sepsis?

Answer: We suggest against controlling blood glucose level tightly in children with sepsis (GRADE 2C: certainty of evidence = “low”).

CQ19: Neuro intensive care

CQ19–1: What are the differential diseases and its testing methods in sepsis patients where brain damage is suspected due to symptoms such as disturbances in consciousness, convulsions, and paralysis?

Answer: Intracranial lesions (e.g., stroke) and potential causes (e.g., metabolic disorders) are first differentiated with the assumption that there may be compound causes for brain damage. Tests include neuroimaging, continuous electroencephalography (EEG) monitoring, biochemical tests, confirmation of the causative agent, and cerebrospinal fluid examination if necessary. Neuroimaging are performed urgently if focal neurologic signs were observed (Provision of information for background question).

CQ20: Patient- and Family-Centered Care

CQ20-1: What are methods for providing information regarding PICS and PICS-F to patients and their families?

Answer: Providing accurate yet continuous information regarding PICS and PICS-F to patients and their families is thought to be important. There are increasing tendencies among medical staff working with the patient to provide handouts at the time of ICU admission/discharge and providing appropriate information. There are initiatives which continuously provide information, such as rounds after discharge from the ICU and the establishment of follow-up outpatients (Provision of information for background question).

CQ20-2: Should ICU diaries be kept by patients with sepsis or those undergoing intensive care?

Answer: We suggest keeping an ICU diary for adult patients with sepsis or those undergoing intensive care (GRADE 2D: certainty of evidence = “very low”).

CQ20-3: Should physical restraints be avoided during intensive care?

Answer: We suggest avoiding physical restraints during intensive care for adult patients with sepsis or those undergoing intensive care (GRADE 2C: certainty of evidence = “low”).

CQ20-4-1: Should ventilation support be provided for sleep care?

Answer: We suggest adding ventilation support as part of sleep care for adult patients with sepsis or those undergoing intensive care (GRADE 2D: certainty of evidence = “very low”).

CQ20-4-2: Should non-pharmacological sleep management (earplugs, eye-masks, music therapy) be used for sleep care?

Answer: We suggest non-pharmacological sleep management for adult patients with sepsis or those undergoing intensive care (GRADE 2D: certainty of evidence = “very low”).

CQ20-5: Should family visiting restrictions be relaxed for the ICU?

Answer: We suggest relaxing family visiting restrictions for adult patients with sepsis or those undergoing intensive care (GRADE 2D: certainty of evidence = “very low”).

CQ20-6: What are methods for supporting decision-making which respects the value systems and ways of thinking in the patient?

Answer: There are methods which support decision making which respects the value systems and ways of thinking of the patient through repeated multi-disciplinary conferences including patients and their families. Methods which carefully identify surrogate intention-estimating individuals (e.g., families) who estimate the intentions of the patient themselves have been proposed when the intentions of the patient are unclear. It is important to respect the intentions of the patients as well as to provide medically accurate information to patients and their families (Provision of information for background question).

CQ21: Sepsis Treatment System

CQ21-1: What methods are there for detecting sepsis at an early stage in the general ward and ER?

Answer: Screening tools such as qSOFA and the early warning score are available as methods which can detect sepsis at an early stage in general wards and in the ER (Provision of information for background question).

CQ21-2: What is the role of a rapid response system (RRS) which acts against changes in the condition of patients in the general ward where sepsis is suspected?

Answer: The rapid response system (RRS) is a system which detects and responds to changes in the condition of patients in the hospital, and there is an opinion where its introduction is expected to improve prognosis of patients even for sepsis (Provision of information for background question).

CQ21-3: Where should sepsis which does not respond to initial fluid resuscitation be managed?

Answer: Sepsis which does not respond to initial fluid resuscitation should be managed in a facility where intensive care can be conducted (Good Practice Statement).

CQ21-4: What quality indicators are there for initial treatment of sepsis?

Answer: Quality indicators for initial treatment of sepsis include implementation rates for each indicator, such as blood culture collection, lactate level measurement, early administration of antimicrobial drug, initial fluid resuscitation, and repeated intravascular volume/cardiac function assessment (Provision of information for background question).

CQ21-5: What kinds of activities raise awareness for sepsis?

Answer: There have been events like “World Sepsis Day” for the general public and seminars for healthcare professionals held, taking the lead by the Global Sepsis Alliance and World Health Organization (WHO) (Provision of information for background question).

CQ22: Stress Ulcer Prophylaxis

CQ22-1: Should antiulcer drugs be administered to septic patients to prevent gastrointestinal bleeding?

Answer: We suggest administering antiulcer drugs to septic patients to prevent gastrointestinal bleeding (GRADE 2B: certainty of evidence = “moderate”).

CQ22-2: How should the suspension of antiulcer drugs be determined for septic patients?

Answer: The specific decision criteria for suspending antiulcer drugs are unclear. Clinical decision criteria include when bleeding risk factors have decreased, side effects such as pancytopenia or liver dysfunction have occurred, and when sufficient enteral nutrition was able to be administered (Provision of information for background question).

CQ1: Definition and diagnosis of sepsis

CQ1-1: Definition of sepsis

Summary: According to the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3), sepsis is defined as “life-threatening organ dysfunction caused by a dysregulated host response to infection.” Septic shock is defined as a subset of sepsis in which the underlying circulatory and cellular/metabolic abnormalities profoundly increase the risk of mortality.

Commentary: Sepsis is defined according to Sepsis-3 [19] in the J-SSCG 2020, similar to the J-SSCG-2016 [3, 4].

In 1992, the definition of sepsis (Sepsis-1) with the concept of systemic inflammatory response syndrome (SIRS) [20] was provided by the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference. The SIRS criteria is widely accepted worldwide, including Japan. According to Sepsis-1, sepsis is defined as SIRS due to infection. However, the Sepsis-1 definition had a low ability to predict the progression of organ damage and low diagnostic specificity for prognosis. Thus, the Sepsis-3 [19] definition adopted in the J-SSCG 2020 guideline focuses on the progression of organ injury in infectious diseases.

In the J-SSCG 2020, sepsis is defined as a condition in which organ dysfunction newly develops after infection. Septic shock is defined as a condition in which sepsis is accompanied by cardiovascular dysfunction, cellular damage, and severe metabolic abnormality. The definition focuses on organ dysfunction associated with infection and assesses the progression of organ dysfunction in infectious diseases that do not meet the criteria for SIRS [20].

CQ1-2: Diagnosis of sepsis and septic shock

Summary: A diagnosis of sepsis is confirmed when the Sequential Organ Failure Assessment (SOFA) score of 2 points or more acutely increase in the presence of a clear infection or suspected infection. Patients with septic shock can be identified with a clinical construct of sepsis with persisting hypotension requiring vasopressors to maintain mBP ≥ 65 mmHg and having a serum lactate level > 2 mmol/L (18 mg/dL) despite adequate volume resuscitation. In out-of-hospital, emergency department, or general hospital ward settings, adult patients with suspected infection can be rapidly identified as more likely to have poor outcomes typical of sepsis if they have at least two of the following clinical criteria that together constitute the quick SOFA (qSOFA) score: a respiratory rate of 22 breaths/min or higher, altered consciousness, and a systolic blood pressure of ≤100 mmHg. The qSOFA criteria can be used to prompt clinicians to further investigate organ dysfunction, initiate or escalate therapy as appropriate, and to consider referral for critical care. Ultimately, an acutely increased SOFA score of 2 or more points confirms the diagnosis of sepsis. Daily routine screening for sepsis is recommended to support the early diagnosis and treatment of sepsis.

Commentary: In the Japanese clinical practice guidelines for the J-SSCG 2020, the severity of sepsis is classified into two categories: sepsis and septic shock according to the Sepsis-3 definition [19]. The diagnosis and treatment of sepsis involves the progression of organ dysfunction in cases of suspected infection. The diagnosis of sepsis is based on agreement with various guidelines, such as the Sepsis-3 definition [19], the J-SSCG 2016 [3, 4], and the SSCG2016 [1, 2]. The qSOFA tool is advantageous as it enables the early evaluation of sepsis. The SOFA score [21] is used for the final diagnosis of sepsis, similar to the J-SSCG 2016 [3, 4]. On the other hand, the low sensitivity of the qSOFA tool for the diagnosis of sepsis and mortality outcome, and evaluation of its utility as an early alert system for sepsis are issues to be resolved in the future [22, 23]. Updates of the SOFA score remain an important issue considering current practices in the treatment of sepsis [24, 25].

CQ2: Diagnosis of infection

Introduction

It is important to diagnose the cause of infection in the treatment of sepsis/septic shock. Identifying pathogenic microorganisms by collecting samples is of utmost importance when diagnosing infections, and this also leads to appropriate treatment. The source of infection should be narrowed down as soon as possible using information from the medical history, physical examination findings, the results of imaging tests, etc., and culture samples should be collected appropriately along with blood cultures from the estimated infection site. Blood culture is the most important test among cultures. Many reports have described the importance of blood culture, which has a high clinical significance in identifying pathogenic microorganisms that cause bacteremia, regardless of the presence of good evidence. However, the method and timing of blood sample collection are not yet well known; thus, we decided to cover this topic in the present guideline [3, 4].

The positivity rate of blood culture tests among patients with septic shock is reported to be 69%. However, there are limits to blood cultures since the positivity rate did not increase despite performing blood culture tests for fever. There is no evidence that an improved prognosis resulted from collecting samples from sites where the possible source of infection could not be ruled out on the basis of clinical images prior to the initiation of antibacterial drugs; however, this is recommended by expert consensus in many guidelines. Describing various culture tests other than blood culture was extremely important in the present guideline as well.

Antibacterial drugs are selected without waiting for blood culture results in clinical practice; however, the practice of referring to Gram stain findings when selecting antibacterial drugs is widespread, and is valid to some extent from the perspective of pathophysiology [3, 4]. Describing the benefit of Gram staining was extremely important in the present guideline as well.

Furthermore, it is important to confirm the effectiveness of these biomarkers for the diagnosis of infection. Four biomarkers (C-reactive protein, procalcitonin, presepsin, and interleukin 6) are currently used to assist in the diagnosis of sepsis. The evaluation of non-severely ill patients, such as emergency outpatients and those in general wards, differs from that of severely ill patients, such as those admitted to the ICU. Thus, these have been discussed separately. Clinical flow of these CQs is shown in Fig. 1.

Fig. 1
figure1

CQ2: Diagnosis of infection (clinical flow)

CQ2-1: When should a blood culture be taken?

Answer: Take two or more sets before administering the antibacterial drug (Good Practice Statement).

Rationale

Bacteremia is generally caused by infections such as endocarditis, central venous catheter infection, pneumonia, abscesses, osteomyelitis, intraperitoneal infection, and urinary tract infections, resulting in a high mortality rate [26]. Various rapid diagnostic methods have been developed [27]; however, at present, blood cultures are the standard test method in the diagnosis of bacteremia. There is no high-quality evidence regarding the timing of blood culture collection, and we have not made a clear recommendation in this CQ.

It has been recommended that sepsis should be suspected in the presence of symptoms indicative of bacteremia (e.g., fever, shivering, hypotension, and tachypnea), hypothermia with an unknown cause, hypotension, altered state of consciousness, increased/decreased white blood cell count, and metabolic acidosis, as well as respiratory failure, acute kidney injury (AKI), and acute liver dysfunction in immunodeficient patients. In these cases, it is recommended that two or more sets of blood cultures be collected as rapidly as possible when the patient has a temperature greater than 38.5 °C or is shivering [28]. However, some reports have indicated that blood cultures do not need to be obtained exclusively for the reasons of fever or an increased white blood cell count, which indicate a low possibility of sepsis [29].

As a general rule, it is important to collect sets before administering antibacterial drugs, while keeping in mind not to delay the initiation of antibacterial drug treatment. This is because the sensitivity of detection often decreases after drug administration, and the bacteria may not be detected [30]. During antibacterial therapy, samples should be collected near the trough of the antibacterial drug concentration, or in other words, immediately before the administration of the next round of antibacterial drugs. Furthermore, samples should be collected again when the patient responds poorly to treatment, and the anti-bacterial drug is changed.

With regard to the amount of sample to collect, it is known that larger collection volumes increase the likelihood of bacterial identification [31]. However, increasing the collection volume can increase the risk of iatrogenic anemia; thus, it is generally recommended that a collection volume of 20–30 mL be used per set. In Japan, the commonly used blood culture bottle often has a capacity of 10 mL, so 20 mL is typical for a single set. Cheruvanky et al. reported that from a clinical economy perspective, 20 mL was better than 30 mL [32].

Reports regarding the number of sets to collect indicated that just one set was characterized by negative results due to a lower sensitivity and an inability to exclude contamination, indicating that two sets (three if possible) were ideal [31, 33]. In reality, it has been said that the blood culture positivity rate is only 5–13%, and that 20–56% of samples are contaminated [34]. A report has indicated that increasing the number of sets would increase the sensitivity (approximately 80, 89, and 98% for one, two, and three sets, respectively) [30]. No increases in sensitivity was seen when four or more sets were collected, and this should be avoided, as it increases the burden on the patient.

Appropriate skin disinfection and the collection of multiple sets are necessary to reduce the likelihood of contamination. It is unclear which among 1% chlorhexidine gluconate, povidone iodine, and 70% alcohol is the optimal antiseptic suitable for skin disinfection; however, there is no doubt regarding the importance of using these agents to ensure an accurate aseptic procedure [35].

CQ2-2: When should culture specimens other than blood be collected?

Answer: Each cultured specimen other than blood should be collected as needed prior to the administration of antibacterial drugs (Good Practice Statement).

Rationale

Blood cultures are a standard diagnostic tool for diagnosing bloodstream infections and bacteremia. Patients with septic shock have been reported to have a blood culture positivity rate of 69%; however, there are limits to blood cultures since the presence of a fever alone does not result in a high positivity rate even with blood culture tests [29]. Identifying infected organs and causative microorganisms is extremely difficult, particularly in cases of sepsis caused by urinary tract infections, pneumonia, and meningitis. Despite showing no evidence of improved prognosis, many guidelines recommend that specimens be collected from areas where the source of infection cannot be ruled out based on clinical findings prior to the administration of anti-bacterial drugs as much as possible [36,37,38].

The diagnosis and treatment of pneumonia can vary depending on the underlying pathology, although diagnoses via sputum culture can be useful. However, as sputum samples have an increased risk of contamination in evaluating the upper respiratory tract, care should be taken in interpreting its test results when they are inconsistent with those of pleural effusion and blood culture. Critically ill patients who have undergone tracheal intubation for mechanical ventilation should have their endotracheal sputum collected and quantitatively cultured; if the bacterial count is found to be over 10 [4] CFU/mL (sputum prior to antibacterial drug administration, sensitivity of 90%, specificity of 77%), then a high possibility of infection with causative bacteria is suspected [39]. Furthermore, a report on the diagnosis of ventilator-associated pneumonia indicated that the probability of non-isolation of causative bacteria was 94% when bacteria were not isolated from endotracheal sputum [40]. Furthermore, searching for microorganisms in bronchoalveolar lavage fluid is also important for deciding the treatment policy for acute respiratory distress syndrome (ARDS) with pneumonia as either a cause or complication, and this is effective for excluding pneumocystis pneumonia or pulmonary mycosis when the immune system of the patient is weakened [41].

Most urinary tract infections are of the ascending type, caused by indigenous bacteria in the colon, and a urine culture test should be performed prior to administering antibacterial drugs in order to isolate the causative bacteria and investigate drug sensitivity. Antibacterial drugs should be administered in recurrent or refractory diseases, and urine culture tests should be performed between drug withdrawals lasting 2–3 days [37, 42].

No RCTs have confirmed the efficacy of blood/cerebrospinal fluid cultures for the diagnosis of bacterial meningitis. However, it is ideal to collect cerebrospinal fluid in all patients with suspected meningitis due to the presence of headaches and altered consciousness so long as cerebral hernias are not suspected based on cranial computed tomography (CT) scans or clinical findings, and lumbar punctures are not contraindicated [38]. However, antibacterial drug administration should be prioritized in cases where cerebrospinal fluid collection takes time. The cerebrospinal fluid culture positivity rate is 70–80% without treatment and less than 50% following antimicrobial therapy [43]. Regarding the cerebrospinal fluid positivity rate for bacterial meningitis, an increased collection volume and centrifugation speed (1500–2500×g, 15 min) increases the detection rate [44].

CQ2-3: Is Gram staining useful in the selection of antimicrobial agents before obtaining culture results?

Answer: We suggest referencing Gram staining findings of the culture specimen when selecting an antibacterial drug to use for empirical treatment (expert consensus: insufficient evidence).

Rationale

The desired effect of Gram staining may be helpful in selecting antibacterial drugs for use in empiric therapy. The 2019 Infectious Diseases Society of America (IDSA) guidelines for community-acquired pneumonia [45] stated that pre-treatment sputum Gram staining and culture should be performed. This should be done when there is severe pneumonia, empiric therapy was commenced for methicillin-resistant Staphylococcus aureus or Pseudomonas aeruginosa, or when oral antibacterial drugs were administered during hospitalization or within 90 days.

The 2015 Japanese Association for Infectious Disease/Japanese Society of Chemotherapy infection treatment guideline [37] for urinary tract infections and male genital infections have shown that urinary Gram staining was deemed useful in estimating the causative organism in cases of catheter-related urinary tract infections. The selection of antibacterial drugs based on Gram stain findings leads to suitable empiric therapy and often leads to definitive therapy. Furthermore, Gram staining has been reported to evaluate bacterial meningitis in that the results can be obtained in a simple and prompt manner, with a sensitivity of 50–90%, specificity of 60–90%, and minimum detection sensitivity of 105 cfu/mL [12].

Selecting antibacterial drugs based only on the results of this test alone has an inherent risk of selecting inappropriate narrow-range antimicrobial drugs regardless of the severity of the patient’s condition. Sensitivity and specificity are also influenced by the tester, and there is a risk of selecting inappropriate antibacterial drugs. The balance between benefits and harms are thought to vary according to the patient’s condition. Gram staining can be performed in a simple yet prompt manner and is also inexpensive; thus, it is thought that the benefits of performing it while understanding its utility and limits outweigh its harms.

Meanwhile, its undesirable effects are as follows. Selecting the antibacterial drug based solely on these test results has the risk of selecting inappropriate narrow-range antimicrobial drugs regardless of the severity of the patient’s condition. Sensitivity and specificity are also influenced by the tester, and there is a risk of selecting inappropriate antibacterial drugs (there is the possibility of the tester using inappropriate testing methods, or the possibility of arriving at false positive/false negative results due to insufficient testing experience). The 2019 IDSA guidelines for community-acquired pneumonia [45] also recommended against Gram staining for sputum obtained after treatment due to the fact that the bacterial strain results could change due to the administration of antibacterial drugs.

Based on the above, it is thought that the balance between benefits and harms vary according to the patient’s condition. However, Gram staining can be performed in a simple and prompt manner and is also inexpensive; thus, it is thought that the benefits of performing Gram staining while understanding its utility and limits outweigh its harms.

CQ2–4-1: What are the positions of C-reactive protein (CRP), procalcitonin (PCT), presepsin (P-SEP), and interleukin 6 (IL-6) as biomarker tests for sepsis diagnosis in general ward and emergency rooms (ER)?

Answer: Sensitivity and specificity in biomarker tests when sepsis was suspected in general ward and ER visits were as follows: CRP, 59, 79%; PCT, 74, 81%; P-SEP, 75, 74%; IL-6, 78, 78%. As such, sepsis diagnosis with biomarkers alone is generally thought to be difficult, and its use should be seen as supplemental to any observations of general conditions (Provision of information for background question).

Rationale

How CQ2–4-1 and CQ2–4-2 became BQs

CQ2–4-1 and CQ2–4-2 were initially grade-based CQs, as follows: “Which among C-reactive protein (CRP), procalcitonin (PCT), presepsin (P-SEP), and interleukin 6 (IL-6) should be used as a biomarker for infectious disease diagnosis?” However, the target infectious diseases varied extremely; thus, in light of the characteristics of this guideline, we focused on sepsis, which is a critical condition that negatively affects general physiological conditions. A comprehensive literature search was conducted as part of a systematic review, with a focus on the diagnostic accuracy of dividing the extracted articles into “general ward or emergency rooms (ERs)” (CQ2–4-1) or “ICUs” (CQ2–4-2). A total of 11 articles were included in the category “general ward or ER”, and the number of papers assessed via a meta-analysis on each biomarker were as follows: CRP, eight articles [46,47,48,49,50,51,52,53]; PCT, 11 articles [1,2,3,4, 19,20,21,22,23,24,25]; P-SEP, four articles [51, 52, 54, 55]; IL-6, four articles [46, 48, 49, 56]. Furthermore, a total of nine articles were included in the category “ICUs”, and the number of papers assessed via a meta-analysis on each biomarker were as follows: CRP, seven articles [57,58,59,60,61,62,63]; PCT, nine articles [57,58,59,60,61,62,63,64,65]; P-SEP, four articles [57, 61, 62, 64]; and IL-6, six articles [58,59,60,61, 63, 65].

An evidence profile and EtD were summarized based on these results, and the following responses were presented: “The diagnostic accuracies of PCT, P-SEP, and IL-6 are thought to be relatively high; however, we do not recommend the use of each biomarker, including CRP, in the diagnosis of sepsis, because this antagonizes the balance of effects against important outcomes among patients and their families” for “general wards and ER” (CQ2–4-1), and “We suggest that the levels of CRP, PCT, and P-SEP be measured as biomarkers for the diagnosis of sepsis in the ICU. We do not recommend the measurement of IL-6 levels” for “ICUs” (CQ2–4-2). A committee vote was then held.

The results of two rounds of voting did not yield any consensus for either CQ, and for CQ2–4-1, committee members indicated that “the role of biomarkers alone is ultimately supplemental for the diagnosis of sepsis but not infectious diseases”, and “this may be interpreted as indicating that the levels of CRP, PCT, and P-SEP, which have until now been widely measured on a regular basis, are no longer necessary, with a concern that biomarker measurements may no longer be conducted”. Furthermore, for CQ2–4-2, there were opinions that “suggesting the usefulness of CRP at the same level as PCT and P-SEP, and suggesting against only IL-6, were inappropriate”. The results of repeated discussions within the committee ultimately resulted in CQ2–4-1 and CQ2–4-2 being handled as BQs.

Explanation: The following explanation was provided using the EP (Tables 4, 5, 6 and 7) created as a result of systematic review and the grade recommendation process.

Table 4 Evidence profile (CRP in general wards or the ER)
Table 5 Evidence profile (PCT in general wards or the ER)
Table 6 Evidence profile (P-SEP in general wards or the ER)
Table 7 Evidence profile (IL-6 in general wards or the ER)

The results of the systematic review for this CQ in terms of the respective sensitivities and specificities of biomarker tests when sepsis was suspected in the general ward or ER were as follows: CRP, 59, 79%; PCT, 74, 81%; P-SEP, 75, 74%; and IL-6, 78, 78%. In actual clinical settings, there are facilities that can only measure CRP levels as well as other facilities that can measure multiple biomarkers. For these reasons, it is worth noting that CRP has an inferior sensitivity to those of PCT, P-SEP, and IL-6 when used as a supplement for the suspicion of sepsis among patients. Based on the above results of systematic review, in facilities in which the levels of the biomarkers PCT, P-SEP, and IL-6 can be measured in addition to CRP, they can be used as a reference to aid the suspicion of sepsis. In these ways, these biomarkers have the potential to bring about significant results in some patients; however, care must be taken as the interpretation of these measurements differ under various conditions depending on patients’ conditions, time of blood sample collection, and location. For these reasons, we decided to specifically display the sensitivities and specificities obtained in the meta-analysis and to leave this to the discretion of the readers in their various respective circumstances.

CQ2–4-2: What are the positions of C-reactive protein (CRP), procalcitonin (PCT), presepsin (P-SEP), and interleukin-6 (IL-6) as biomarker tests for sepsis diagnosis in the intensive care unit?

Answer: Sensitivity and specificity in biomarker tests when sepsis was suspected in the ICU were as follows: CRP, 74, 70%; P-SEP, 82, 73%; IL-6, 72, 76%. As such, sepsis diagnosis with biomarkers alone is generally thought to be difficult, and its use should be supplemental to any observations of general conditions (Provision of information for background question).

Rationale

The background and recommendation making process was described in the rationale for CQ2–4-1. The following rationale was created in reference to the evidence profile (Tables 8, 9, 10 and 11) created as a result of an systematic review and the grade recommendation process.

Table 8 Evidence profile (CRP in the ICU)
Table 9 Evidence profile (PCT in the ICU)
Table 10 Evidence profile (P-SEP in the ICU)
Table 11 Evidence profile (IL-6 in the ICU)

The results of the systematic review for this CQ showed that the respective sensitivities and specificities of the biomarker tests when sepsis was suspected in the ICU were as follows: CRP, 71, 61%; PCT, 74, 70%; P-SEP, 82, 73%; and IL-6, 72, 76%. Based on these results, it cannot be determined whether the sensitivities and specificities were sufficiently high or low.

The biomarker tests suggested significant results for the diagnosis of sepsis in individual articles assessed in the systematic review [57,58,59,60,61,62,63,64,65]. Meanwhile, care must be taken because the results of biomarker tests can change or can be influenced by the bacterial type or location of the infection depending on various factors such as patient status or time of blood sample collection. For these reasons, we have specifically displayed the sensitivities and specificities obtained in the meta-analyses and have left this to the discretion of the individual readers in their respective circumstances.

CQ3: Source control

Introduction

The importance of initiating treatment for sepsis at an early stage is widely accepted. Among early treatment modalities, controlling the source of infection is one that exhibits its effectiveness by cutting off and “controlling” the “infection source” that is at the root of sepsis, and forms the basis of initial treatment. Diagnostic imaging is essential to promptly control the source of infection. Therefore, two CQs on diagnostic imaging were first incorporated, after which seven CQs on controlling the source of infection were incorporated.

The first CQ on diagnostic imaging that was incorporated was “CQ3-1: Should imaging tests be performed in patients with suspected sepsis to identify the source of infection?” Diagnostic imaging modalities for identifying the source of infection include simple radiography, ultrasonography, CT scans, and magnetic resonance imaging (MRI) scans, and highly useful test methods vary by site. The explanations in this CQ include a table on diagnostic imaging methods thought to be specific for each organ/illness in order to be of use in actual clinical practice.

The second CQ on diagnostic imaging is that regarding full-body contrast CT scans: “CQ3-2: Should full-body contrast-enhanced CT tests be performed at an early stage in patients with sepsis and an unknown source of infection?” Identifying the source of infection early when the source is unknown is essential for formulating a treatment policy. Performing CT scans, which are diagnostic imaging modalities that have seen widespread use in Japan, are important for local diagnosis as well as for determining the severity of the source of infection. Thus, this was taken up as a CQ.

Subsequent discussions on the selection of CQs regarding the control of the source of infection resulted in the following six sources of infection that were thought to be of particular importance and set as CQs: 1) intraperitoneal infection, 2) infectious pancreatic necrosis, 3) acute pyelonephritis secondary to ureteral obstruction, 4) necrotic soft tissue infection, 5) catheter-related bloodstream infections, and 6) empyema.

It is universal knowledge that the basic concept underlying the control of the source of infection is to do so “promptly” and “appropriately.” The best methods are those that are minimally invasive, have a low incidence of complications, and have sufficient expected effects. Furthermore, the source of infection should generally be controlled promptly; however, we also suggest that elective interventions may be considered for patients with infectious pancreatic necrosis. Clinical flow of these CQs is shown in Fig. 2.

Fig. 2
figure2

CQ3: Source control (clinical flow)

CQ3-1: Should imaging tests be conducted in patients suspected of sepsis in order to search for the source of infection?

Answer: Imaging tests should be conducted when the source of infection is unclear in order to search for the source of infection (Good Practice Statement).

Rationale

Controlling the source of infection at an early stage is an important treatment strategy that is linked to an improved outcome among patients with sepsis. For this reason, it is important to assess early whether there is a source of infection that needs to be controlled among patients with suspected sepsis, and imaging tests need to be considered for this procedure. Imaging tests useful for identifying the source of infection include plain radiography, ultrasonography, CT scans, and MRI scans. The most effective testing modality varies with the site of suspected infection. Diagnostic imaging modalities considered characteristic of each organ/disease are shown in Table 12.

Table 12 Diseases that require control of the source of infection and imaging tests
  1. (1)

    Head and neck

Cerebral abscess: CT scans are easier to conduct in an emergency relative to MRI scans; thus, the former is often prioritized in its implementation. Contrast-enhanced MRI scans are the most recommended imaging modality because of their ability to detect the spread of inflammation to the capsule or tissue surrounding the abscess [66].

Cervical abscess (descending mediastinitis): Cervical abscesses near the surface of the body can be detected via ultrasonography; however, there are limits to the detection of deep cervical abscesses, and CT scans are considered effective. Contrast-enhanced CT scans are recommended because they can clearly differentiate between fluid retention due to infection and structures such as blood vessels [67].

  1. (2)

    Chest

Empyema: Plain X-ray imaging and ultrasonography are first-line evaluation modalities. Contrast-enhanced CT scans are effective for controlling the source of infection or as an indicator for assessing the course of treatment when an empyema is suspected.

Infectious endocarditis: One of the two major categories in the diagnostic criteria for infectious endocarditis (the Duke diagnostic criteria) [68] is based on the findings of echocardiography, and transthoracic echocardiography should be implemented as a first-line evaluation modality for all patients when infectious endocarditis is suspected. The accuracy of transesophageal echocardiography for the diagnosis of infectious endocarditis is superior relative to the transthoracic variation; therefore, we recommend that additional transesophageal echocardiography should be performed when necessary [69].

  1. (3)

    Abdomen

Intestinal perforation/peritonitis: Plain X-ray imaging and ultrasonography should be performed first. CT scans should be subsequently performed when further assessments are needed. We recommend that contrast-enhanced CT scans be performed when detailed assessments of phenomena such as the presence of ischemia in organs or the intestinal tract needs to be determined [70].

Cholecystitis/cholangitis: Ultrasonography and CT scans are the most recommended evaluation modalities. Contrast-enhanced CT scans can be used to identify important findings. We also recommend MRI/magnetic resonance cholangiopancreatography as alternative imaging modalities [71].

Obstructive urinary tract infection: Ultrasonography should be performed as a first-line assessment modality. We recommend that CT scans should be performed to carefully evaluate the causes of obstruction if the clinical findings are suggestive of obstructive urinary tract infection [72].

  1. (4)

    Others

Necrotizing soft tissue infection: A contrast-enhanced CT scan should be performed because of its ability to detect the swelling and fluid retention in soft tissue. However, a definitive diagnosis of necrotizing fasciitis cannot be made with a contrast-enhanced CT scan alone; such a diagnosis requires surgical examination of the subcutaneous tissue/fascia and direct observation of the fascia/muscle [73].

Imaging modalities are beneficial for the selection of the optimal treatment method. Meanwhile, the risk of exposure to X-rays or utilization of contrast agents, particularly the risk of sudden changes while transferring critically ill patients to the examination room, must be recognized.

CQ3-2: Should whole-body contrast-enhanced CT tests be conducted at an early stage for sepsis patients with unknown source of infection?

Answer: We suggest conducting whole-body contrast-enhanced CT tests as soon as possible for sepsis patients with unknown source of infection (expert consensus: insufficient evidence).

Rationale

Appropriate therapeutic interventions at an early stage against the source of infection are recommended for sepsis [74]. Searching for the source of infection at an early stage when it is unknown is also essential to formulating a treatment plan. The use of CT scans, which are widespread diagnostic imaging modalities in Japan, is essential for local diagnosis and determining the severity of the source of infection.

The results of a systematic review showed that there were no RCTs conforming to the PICO criteria that were conducted on patients who satisfied the sepsis diagnostic criteria or who were undergoing intensive care.

It is possible that improvements in general conditions are not achieved even with standard therapy in cases of sepsis in which the sources of infection are unclear. Therefore, efforts must be made to perform whole-body contrast-enhanced CT scans at an early stage and clarify the source of infection to improve vital prognosis, and it is thought that a desirable therapeutic intervention for the patient could be possible. It is feared that patients with complications of shock will have experience destabilization of hemodynamics accompanied by moving them. Furthermore, it is feared that the use of contrast agents will result in the onset of allergies to iodine or contrast agent-induced nephropathy.

At the very least, it is possible that the source of infection could be clarified by performing whole-body contrast-enhanced CT scans. It is thought that the benefits outweigh the harms, such as destabilized hemodynamics accompanied by moving, contrast agent-induced nephropathy, and allergies to iodine.

Japan has the highest number of CT scanning devices per capita worldwide, and there are many facilities in which sepsis can be treated and this is thought to be possible.

Contrast-enhanced CT scans are not necessarily useful for all organs when searching for the source of infection. In some cases, specific inspection methods should be prioritized for each organ, and further investigations are necessary on the usefulness of contrast-enhanced CT scans according to organs involved in sepsis with an unknown source of infection.

CQ3-3: Should the source of infection be controlled by surgery/invasive drainage in patients with sepsis due to intraperitoneal infection?

Answer: We suggest controlling the source of infection as soon as possible with surgery/invasive drainage (including abscess drainage, biliary tract/gallbladder drainage) for patients with sepsis due to intraperitoneal infection (expert consensus: insufficient evidence).

Rationale

The potential benefits of rapidly controlling the source of infection among patients is considered large in cases of sepsis due to intraperitoneal infection such as generalized peritonitis due to the perforation of the lower gastrointestinal tract, where the possibility of improvements with only typical antibacterial drug treatment without controlling the source of infection is extremely low. Possible harms that can occur in actual clinical practice include bleeding, organ damage, deteriorating general conditions due to biological invasion, and infection. There were no RCTs conforming to the PICO criteria, and the balance of effects is unclear. It is thought that the benefits outweigh the harms, even when comparing the advantages obtained via surgical intervention by way of drainage (including abscess and biliary drainage) for sepsis due to intraperitoneal infection, and the harms of bleeding, organ damage, deteriorating general conditions due to biological invasion, and infection due to surgery or drainage.

CQ3-4-1: Should the source of infection be controlled with invasive interventional therapy during the early period of infectious pancreatic necrosis?

Answer: We suggest against controlling the source of infection with invasive interventional therapy during the early period of infectious pancreatic necrosis (GRADE 2C: certainty of evidence = “low”).

Answer: We suggest against controlling the source of infection with invasive interventional therapy during the early period of infectious pancreatic necrosis (GRADE 2C: certainty of evidence = “low”).

Rationale

Necrotic tissue is a cause of infection, and early intervention is a general principle underlying treatment. However, pancreatic necrosis does not fall under this general principle of early intervention. Furthermore, RCTs that incorporated minimally invasive and effective methods to control the sources of infection have been conducted; thus, the timing of intervention for this disease is an important CQ.

The results of a systematic review confirmed a single RCT conforming to the PICO criteria with a small sample size (early intervention, 25 patients; late intervention, 11 patients). The mortality rates were 56 and 27% for the early and late intervention groups, respectively. The estimated value of effects yielded a risk difference (RD) of 286 more per 1000 (95% confidence interval [CI]: 71 fewer to 1000 more), and no desired effects related to vital outcomes were observed in the early intervention group compared to the late intervention group [75]. No investigations have been conducted on adverse effects or medical costs, and the desired effects in the early intervention group are unknown. The mortality rate of the late intervention group was lower than that of the early intervention group; thus, it is likely that the benefits of late intervention outweigh its harms.

CQ3-4-2: Should the source of infection be controlled with low-invasive interventional therapy for infectious pancreatic necrosis?

Answer: We recommend controlling the source of infection with less invasive interventional therapy for patients with sepsis caused by infectious pancreatic necrosis (GRADE 2B: certainty of evidence = “moderate”).

Rationale

Infectious pancreatic necrosis is a condition that requires the removal of the source of infection with some types of interventional treatment. A number of treatment strategies have been reported in recent years, such as (1) surgical drainage, (2) endoscopic drainage, (3) percutaneous drainage (mainly via the retroperitoneal route), and (4) the step-up approach, which becomes incrementally more invasive according to the treatment effect. The relationship between treatment invasiveness and treatment effect is therefore an important CQ.

The results of systematic reviews confirmed the existence of two RCTs (less invasive methods, 94 patients; invasive methods, 92 patients) [76, 77]. The data used in these two RCTs showed that the onset of complications when the source of infection was controlled with less invasive methods (drainage methods) was lower than that when invasive methods were used RD of 187 fewer per 1000 (95%CI: 305 fewer to 55 more). Based on the above results, the desired effects of less invasive interventional treatment are considered moderate.

In terms of mortality outcomes, researchers investigated the three timings of short-term (6 months), medium-term (3 years), and long-term (10 years) outcomes. It was possible to pool data from the 2 RCTs (less invasive methods, 94 patients; invasive methods, 92 patients) using only mortality within six months and the number of effects of mortality outcomes yielded a RD of 40 more per 1000 (95%CI: 48 fewer to 211 more). Furthermore, the number of effects for the length of stay in the ICU and in-hospital stay each yielded a mean difference (MD) of 19.74 days longer (95%CI: 20.84 shorter to 60.31 longer) and 7.76 days shorter (95%CI: 27.86 shorter to 12.34 longer), respectively. The number of effects varied widely and the undesired effects of controlling the source of infection with less invasive interventional methods when compared to invasive interventional methods were unclear.

The invasiveness of procedures for controlling the source of infection, their timing, the range over which debridement is to be performed, and the necessity of repeated debridement needs to be investigated alongside the general conditions of patients, and this is not recommended for standard treatment among all cases.

CQ3-5: Should the source of infection be controlled with invasive drainage for patients with sepsis due to acute pyelonephritis caused by ureteral obstruction?

Answer: We suggest controlling the source of infection as soon as possible with transurethral ureteral stent implantation or percutaneous nephrostomy in patients with sepsis due to acute pyelonephritis caused by ureteral obstruction (expert consensus: insufficient evidence).

Rationale

The results of a systematic review showed that there were no RCTs that conformed to the PICO criteria. Patients with acute pyelonephritis secondary to ureteral obstruction are unlikely to recover from sepsis unless transurethral stenting or percutaneous nephrostomy is performed to eliminate the cause. Therefore, it is thought that the potential benefits of rapidly controlling the source of infection are high among these patients. There was no significant difference between patients who underwent percutaneous renal fistula construction and transurethral ureteral stenting, which are methods of providing emergency relief for ureteral obstruction. Complications associated with invasive procedures include bleeding, organ damage, and the spread of infection to the retroperitoneal space (cavity). However, it is thought that the benefits outweigh the harms, even when considering complications or the burden of transferring a patient to a facility in which rapid specialized treatment modalities (transurethral ureteral stenting or percutaneous renal fistula) can be performed when such treatments cannot be offered.

CQ3-6: Should source control be achieved by means of surgical debridement for sepsis patients due to necrotic soft tissue infection?

Answer: We suggest controlling the source of infection as soon as possible by means of surgical debridement for sepsis patients due to necrotic soft tissue infection (expert consensus: insufficient evidence).

Rationale

Necrotic soft tissue infection is a condition that requires early surgical control of the source of infection, and the need for debridement is difficult to determine with imaging tests. Performing surgical debridement of the necrotic tissue (soft tissue) that causes sepsis can reliably control the source of infection, and desirable effects such as an increased survival rate and a shortened therapeutic duration can be obtained. Meanwhile, most patients require surgery under general anesthesia, and there is concern about further anesthesia-induced destabilization due to unstable hemodynamics, as well as effects on hemodynamics due to hemorrhaging, and in some patients, multiple sessions of surgical debridement are necessary. There were no RCTs that conformed to the PICO criteria, and the balance of effects was unclear. The benefits of surgically removing the source of infection are thought to outweigh the harms, even when the harm caused by surgical treatment is compared.

CQ3-7: Should the source of infection be controlled with catheter removal in patients with sepsis where catheter-related bloodstream infections are suspected?

Answer: We suggest controlling the source of infection as soon as possible with catheter removal in patients with sepsis where catheter-related bloodstream infections are suspected (expert consensus: insufficient evidence).

Rationale

Vascular catheter infections may not be improved with normal antibacterial drug treatment alone without controlling the source of infection. There have been cases in which the prognosis or mortality rate worsened if the cause was not resolved; therefore, it is thought that promptly controlling the source of infection has a high potential of yielding benefits among patients. This desirable effect is influenced by the accuracy of diagnosis of catheter infections. Patients who require vascular catheters do not only require the removal of the vascular catheter but also its re-insertion when controlling the source of infection. This may yield complications associated with vascular catheter insertion and affect the risks associated with re-insertion. Furthermore, frequent route exchanges increase costs and work burden. There are no RCTs that conform to the PICO criteria, and the balance of effects is unknown. It is thought in the case of vascular catheter infection that the benefits obtained by controlling the source of infection (catheter removal) outweigh the harms of complications relating to vascular catheter removal.

CQ3-8: Should the source of infection be controlled through invasive drainage in patients with sepsis due to empyema?

Answer: We suggest controlling the source of infection as soon as possible with percutaneous thoracic drainage or surgical intervention in patients with sepsis due to empyema (expert consensus: insufficient evidence).

Rationale

The results of a systematic review showed that there were no RCTs that conform to the PICO criteria. Encapsulated empyema cannot be improved with conventional antibacterial drug treatment; thus, the possibility of recovery from sepsis is low without resolving the source. Therefore, the potential benefits of promptly controlling the source of infection are thought to be high for patients. It is thought that patients could be rapidly transferred to facilities capable of performing open chest drainage when parenchymal organs are present in the drainage route due to tissue adhesion and when percutaneous drainage is difficult. Possible harms associated with invasive damage include bleeding, lung injury, and pain in the wound or around the drain. Open chest drainage is highly invasive compared to percutaneous drainage and likely has a greater degree of undesired effects. However, the benefits of open chest and subcutaneous drainage are thought to outweigh its harms in cases of sepsis due to empyema, even when considering complications such as hemorrhage and lung injury or the rapid transfer to a facility capable of performing drainage procedures.

CQ4: Antimicrobial therapy

Introduction

Antimicrobial therapy for underlying infectious diseases is an essential aspect of sepsis treatment. The importance of antimicrobial therapy is that not only it is directly associated with an outcome, but it is also related to the global concern regarding antimicrobial resistance and the associated risk of reducing effective therapeutic options in the future. The judicious use of antimicrobial agents that fully incorporates the concepts of antimicrobial stewardship [78] is required.

This guideline targets the treatment of sepsis and will not delve into the details of drug selection. The basic principles underlying drug selection for patients with sepsis are the same as those for general infection treatment. In other words, antimicrobial agents to be administered are selected by assuming specific microorganisms and drug resistance as much as possible based on patients’ backgrounds, suspected infectious foci, epidemiological information on the region or facility, and recent antimicrobial use. However, it is important to promptly administer effective antimicrobials against causative pathogens in septic patients compared to non-critically ill patients.

With regard to antimicrobial therapy for patients with sepsis, empiric antimicrobials should initially be selected after assuming the underlying microorganism, which should then be optimized to targeted antimicrobial agent(s) after the causative pathogens and their susceptibility patterns have been determined.

The appropriateness of empiric antimicrobials is associated with mortality outcomes [79]. The underlying microorganism should be determined for each suspected source of infection based on patients’ background, epidemiology, and rapid diagnostic tests, and the drug should be selected in consideration of the properties of drug distribution/tissue penetration and antimicrobial resistance. Indications for carbapenems and pathogens that require antimicrobial drugs other than β-lactams have been described. The timing of initiation of empiric antimicrobial drug administration has also been described.

With regard to interventions after culture results are obtained, 1) the possibility of termination when culture results are negative, 2) the significance of de-escalation to target antimicrobial agents with narrower spectrum, 3) procalcitonin guidance as a reference for the discontinuation of antimicrobial drugs, and 4) the possibility of setting up a relatively short duration (within 7 days) of antimicrobial therapy are provided. These reflect fundamental concepts of antimicrobial stewardship.

For the selection and administration of drugs, 1) when to consult the antimicrobial stewardship team, 2) continuous or prolonged infusion of β-lactams based on the pharmacokinetic-pharmacodynamic theory, and 3) dose adjustment of renally excreted antimicrobials are discussed.

Clinical flow of these CQs is shown in Fig. 3.

Fig. 3
figure3

CQ4: Antimicrobial therapy (clinical flow)

CQ4-1: How should empirical antimicrobial therapy be selected?

Answer: Antimicrobials can be selected by estimating the causative microorganism based on suspected infectious foci, patient background, epidemiology and rapid microbial diagnostic tests, and by considering the tissue penetration properties of drugs and the probabilities of resistant bacteria (see Table 2 for reference). (Provision of information for background question).

Rationale

The selection of empiric antimicrobial therapy should include the determination of the causative microorganisms for each suspected source of infection based on the patient’s background and the epidemiology of the disease. This should be done according to the tissue penetration properties of drugs, antibacterial spectrum (including the possibility of resistant bacteria), clinical evidence, and the results of rapid diagnostic testing if available.

Table 2 (Empiric therapeutic agents for each infectious disease) shows a list of empiric antimicrobial therapy selections for each combination of common sources of infection and patient background based on expert opinions. It is assumed that this table will serve as a reference for decision-making by adding information such as an individual patient’s circumstances and the local/regional epidemiological factors and using them alongside antimicrobial therapy guidelines in each region or medical facility. Furthermore, antimicrobial therapy guidelines for each region or facility can be created using this table as a foundation if such guidelines do not exist.

The causative microorganisms can be determined based on the epidemiology of each source of infection. As such, the identification of the source of infection is important not only for surgical drainage, but also for specimen collection to select appropriate antimicrobial therapy. Two epidemiological studies conducted in Japan (2010–2011: 15 facilities; 2016–2017: 59 facilities) indicated that common sources of sepsis were respiratory infections, intra-abdominal infections, urinary tract infections, and soft tissue infections, all of which accounted for approximately 90% of cases [80, 81] Similar trends were observed in multiple international studies [82,83,84,85,86]. Meanwhile, reports have also indicated that a source of infection was not identified in approximately 1/6th of patients with sepsis [82,83,84,85,86]. Infectious diseases that should be considered when a specific source of infection could not be identified included diseases in which specific findings are difficult to determine (e.g., infectious endocarditis, catheter-related bloodstream infections) and systemic infections in which a source of infection did not form or was unclear (e.g., fulminant infection following splenectomy, purpura fulminans, rickettsial infection, febrile neutropenia with unknown source, etc.). Caution should be taken in evaluating implantable device-related infections (e.g., catheter-related bloodstream infections, prosthetic valve endocarditis, cerebrospinal fluid shunt-related meningitis/ventriculitis, and prosthetic joint infection) since specific findings are difficult to determine [87,88,89,90].

The causative microorganisms can also be determined based on patient background. There are two factors: 1) external factors such as history of exposure (including healthcare exposure or travel history), and 2) internal factors – the patient’s own conditions (including age, sex, and underlying diseases). The classification of patient background factors for selecting antimicrobial therapy varies depending on the source of infection. Community-acquired infections generally have causative microorganisms that differ from those of healthcare-associated infections, and Pseudomonas aeruginosa does not need to be routinely covered as a community-acquired pathogen. Exposures, which serve as risk factors for healthcare-associated infections, include invasive procedures or devices (surgery, transplantation, intravascular catheters, urinary catheters, endotracheal tubes, enteral feeding tubes, etc.) and antimicrobial therapy history. For patients with sepsis with a travel history, there is a need to consider systemic infections such as malaria, meningococcal infections, viral hemorrhagic fever, rickettsial diseases, and infections due to drug-resistant bacteria [91, 92]. Sepsis due to rickettsial infection (Japanese spotted fever and scrub typhus) or severe fever with thrombocytopenia syndrome (SFTS) should be included in the differential diagnosis if the patient has a history of travel to endemic areas of tick-borne infectious diseases in Japan [93]. Furthermore, age is an important patient factor because the causative bacteria in meningitis differ depending on whether the patient is older than 50 years [94]; more than 90% of cases of Legionnaires’ disease leading to pneumonia occur in patients older than 50 years [95]. Urinary tract infections and soft tissue infections are common among diabetic patients [96]. Pseudomonas aeruginosa and/or methicillin-resistant Staphylococcus aureus (MRSA) should be considered in neutropenic sepsis [97]. Pneumocystis pneumonia should be included in the differential diagnosis of pneumonia in patients with cellular immunodeficiency, such as human immunodeficiency virus infection [98].

Rapid diagnostic testing should be implemented if possible after the causative microorganisms have been determined from epidemiological information relating to the source of infection and the patient’s background. Gram staining can aid in the identification of significant microorganisms by determining whether local inflammation is present through the presence of leukocytes in the collected specimen. It is important to examine whether the coverage of empiric antimicrobial therapy is sufficient while considering the quality of the specimen when performing Gram staining [73].

Antimicrobial therapy covering inferred or confirmed microorganisms should be selected with due consideration to tissue penetration properties of drugs, antimicrobial resistance, and clinical evidence. Caution with regard to the tissue penetration and in situ activity of the antimicrobial therapy are shown as follows: ceftriaxone, cefepime, and meropenem can be used as β-lactams in the treatment of meningitis; however, cefazolin should be avoided due to inappropriate cerebrospinal fluid penetration. Daptomycin should be avoided in the treatment of pneumonia because it is deactivated by alveolar surfactants [99].

Drug resistance is an increasingly widespread problem globally and constitutes a threat to the treatment of sepsis [100,101,102,103,104]. The susceptibility rates of antimicrobial therapy vary according to time and place (country, region, facility, and hospital ward), and it is important to determine the local data by region or facility via methods such as antibiograms [105]. As antibiograms are the collected data of specimens that were submitted for various objectives, caution should be taken in their usage as reported resistance rates may be higher than the actual rates of limited specimens prior to antimicrobial therapy [106]. The previous culture testing results of an individual patient are also important. Previously colonizing or infecting bacteria do not always become the causative microorganisms of sepsis; however, the detection of resistant bacteria is a risk factor, and their coverage should be considered.

Empiric antimicrobial therapy should be selected to minimize the lapse in coverage of the inferred causative microorganisms and to anticipate a later transition to targeted or definitive antimicrobial agents. Changes in drug therapy need to be implemented rapidly if coverage is deemed insufficient. Targeted antimicrobial therapy should be selected to maximize the treatment effect and minimize adverse effects and collateral damage (i.e., negative influences on indigenous microbiota) [107]. It is beneficial to consider the targeted antimicrobial agents that can be used later when selecting empiric antimicrobial agents. For example, in Japan, cefazolin is the first-line treatment for methicillin-sensitive Staphylococcus aureus (MSSA) bacteremia with no intracranial dissemination, and its treatment performance against MSSA bacteremia was superior to that of vancomycin, which is frequently used when the presence of methicillin resistance is not known [108, 109]. With this in mind, the concomitant use of cefazolin should be considered when using vancomycin as an empiric antimicrobial agent with the objective of covering MRSA if the possibility of MSSA is deemed high. In this way, Table 13 (Target therapeutic agents by causative microorganism) shows a list of targeted antimicrobial therapies likely to be encountered in the treatment of sepsis by susceptibility result patterns. When focusing on spectrums in shifting from empirical to targeted antimicrobial therapy, changing from wide-spectrum to narrow-spectrum antimicrobial therapy is referred to as de-escalation, while the opposite is referred to as escalation [1, 3, 110] Refer to CQ4–8 for information on whether de-escalation is an effective strategy against sepsis.

Table 13 Target therapeutic agents by causative microorganism

Various investigations have been conducted to improve prognosis by optimizing the selection of antimicrobial therapy in sepsis. The J-SSCG 2016 recommended against routine empiric combination therapy due to the lack of evidence of improved prognosis and treatment harm, including renal injury(1B) [3]. Carbapenem is often selected for the treatment of sepsis; however, some initiatives have strategically implemented carbapenem-sparing regimens that avoid excessive carbapenem use as the threat of drug resistance has become a global issue. The use of appropriate antimicrobial therapy in supporting additional payments were established in 2018 in Japan, and the establishment of consultation systems of infectious disease specialists and support systems for proper antimicrobial therapy (i.e., antimicrobial stewardship teams) has been promoted. Refer to CQ4–2 (Under what circumstances should carbapenems be used in empiric antimicrobial therapy?), CQ4–3 (Under what circumstances should empiric antimicrobial therapy be selected for MRSA and non-bacterial pathogens (e.g., Candida, viruses, Legionella, Rickettsia, and Clostridioides difficile)), and CQ4–5 (Under what circumstances should an infectious disease specialist or antimicrobial stewardship team be consulted?).

Finally, the complexity of the body of specialized knowledge of infectious diseases continues to grow, including the diversification of the causes of sepsis, the global threat of antimicrobial resistance, the decline in antimicrobial agent development, the constraints of drug supply, and multiple revisions to the criteria of susceptibility tests. There are also problems in clinical practice that hinder the optimization of antimicrobial therapy, such as “culture-negative sepsis” (CQ4–4) which refers to the fact that even if the proper test is performed in cases of sepsis, 30–50% of culture tests yield negative results. It is important to faithfully practice the basic principles underlying infectious disease management –“estimating the causative microorganisms based on the source of infection, patient background, epidemiology, and rapid diagnostic testing results, as well as considering the tissue penetration properties of drugs, antimicrobial resistance, and clinical evidence”, in order to effectively use limited antimicrobial therapy resources.

CQ4-2: Under what circumstances should carbapenems be used in empirical antimicrobial therapy?

Answer: Carbapenems can be included in the empirical antimicrobial regimen when the use of carbapenem is considered to be particularly effective; ESBL-producing Enterobacteriaceae or Pseudomonas aeruginosa or Acinetobacter species with limited susceptibility for carbapenems (Provision of information for background question).

Rationale

Carbapenems that are currently available in Japan for intravenous injection include meropenem, doripenem, imipenem/cilastatin, panipenem/betamipron, and biapenem. The antibacterial spectrum of all of these drugs is virtually the same and wide-ranging, from Gram-positive to Gram-negative bacteria. However, methicillin-resistant Staphylococcus and Enterococcus species, Stenotrophomonas maltophilia, and fungi are not sensitive to these agents.

Several RCTs have compared the effects of carbapenems and other wide-spectrum β-lactams, but were not designed to distinguish between empiric and target therapy for sepsis. The treatment effects of carbapenems were identical to those of β-lactams alone or concomitant use of β-lactams, aminoglycosides, or metronidazole [142]. RCTs of patients with severe infections showed that carbapenems had an efficacy comparable to that of tazobactam/piperacillin in the treatment of pneumonia [143,144,145]. Carbapenems also had an efficacy comparable to that of tazobactam/piperacillin [146] or quinolones [147] for intraperitoneal infection, and to third-generation cephalosporins [148, 149] for meningitis. Taken together, the routine use of carbapenems in patients with sepsis has not yet been determined to be superior.

There is an opinion that carbapenems should only be used selectively if a specific microorganism is suspected as a causative pathogen. Currently, the increase in the number of strains that produce extended-spectrum β-lactamase (ESBL) among the Enterobacteriaceae is a concern [150]. Apart from carbapenems, other treatment options for ESBL-producing strains include broad-spectrum penicillin with β-lactamase inhibitors combinations, cephamycin, and aminoglycosides. Observational studies have shown that these agents are not inferior to carbapenems [151], while some RCTs have shown that carbapenems are superior [152, 153]. As empiric therapy, carbapenems are likely the first-line treatment option, particularly in critically ill patients, such as those with sepsis/septic shock. Furthermore, the number of resistant Pseudomonas aeruginosa and Acinetobacter species with sensitivity only to carbapenems has been increasing. It is logical to select carbapenems when these microorganisms are suspected. However, these types of resistant strains are rarely encountered in Japanese clinical settings.

Meanwhile, the issue of carbapenem-resistant Gram-negative bacilli is becoming a global problem. The resistance rate of Pseudomonas aeruginosa due to antimicrobial exposure is the highest, particularly to carbapenems [154, 155]. The use of carbapenems has been found to be the commonest risk factor for multidrug-resistant Pseudomonas aeruginosa or Acinetobacter species [156]. A meta-analysis showed that the odds ratio (OR) of the occurrence of carbapenem-resistant Pseudomonas aeruginosa due to the use of carbapenems was 7.09 (95%CI 5.43 to 9.25) [157]. The proportions of carbapenem-resistant Pseudomonas aeruginosa detected in the Japanese Nosocomial Infection Surveillance (JANIS) study were high, at 11 and 17% for meropenem and imipenem, respectively [158]. Furthermore, carbapenem use is a risk factor for the identification of carbapenem-resistant Enterobacteriaceae, including carbapenemase-producing strains [159]. The increase in the number of resistant bacteria worldwide, particularly in developing countries, has become a concern. The proportion of Gram-negative bacilli in the group of carbapenem-resistant Enterobacteriaceae is still low (at less than 0.5% according to the JANIS data) [158]; however, this is expected to increase in the future with globalization. The presence of resistant bacteria increase the inappropriateness of empiric antimicrobial therapy and are associated with poor outcomes [160,161,162]. Taken together, carbapenems should be used when appropriate, being aware of the risk of development of drug-resistant strains.

The emphasis on the appropriate use of carbapenem leads to the use of carbapenem in limited cases in which the causative bacteria are one of the aforementioned microorganisms with carbapenem-limited sensitivity. This is the most conservative option when using carbapenems. This guideline supports this conservative option from the viewpoint of prioritizing antimicrobial stewardship and the current situation of frequent use of carbapenems in Japan. Specifically, carbapenems can be selected when ESBL-producing Gram-negative bacilli, multidrug-resistant Pseudomonas aeruginosa, or Acinetobacter species with limited sensitivity to carbapenem are suspected.

Multiple studies including systematic reviews have reported on the risk factors for infections caused by ESBL-producing strains [150, 163, 164], third-generation cephalosporin-resistant Enterobacteriaceae [165], and multidrug-resistant Pseudomonas aeruginosa [159]. Although there are differences according to the microorganism, the primary risk factors shared by many studies were a history of administration of antimicrobial agents and colonization by any resistant pathogen.

Rottier et al. [165] assessed the risk factors for infection due to the third-generation cephalosporin-resistant Enterobacteriaceae among Gram-negative bacilli bacteremia. The authors showed that using carbapenems selectively in cases of colonization with multidrug-resistant bacteria could avoid the excessive use of carbapenems or aminoglycosides without reducing their appropriateness.

Lambregts et al. [166] extracted the risk factors for the presence of second-generation cephalosporins and aminoglycoside-resistant bacteria as colonization and history of using those antimicrobials in Enterobacteriaceae bacteremia. The authors showed that carbapenem use could be decreased and the appropriateness of empiric therapy increased when carbapenem was administered selectively for cases with risk factors.

Based on these results, (i) colonization or a history of infection with a resistant pathogen, or (ii) a history of administration of antimicrobial agents can be listed as risk factors for infections with ESBL-producing bacteria, multidrug-resistant Pseudomonas aeruginosa or Acinetobacter species that have sensitivity only to carbapenems, against which carbapenems can be considered as empiric therapy.

However, antimicrobial agents that can be used as alternatives to carbapenems and their resistance patterns can vary according to the country, region, facility, and department. Thus, consideration should be given to each clinical setting.

CQ4–3: Under what circumstances should empirical antimicrobial therapy be selected for MRSA and non-bacterial pathogens (e.g., Candida, Viruses, Legionella, Rickettsia, or Clostridioides difficile)?

Answer: Each microorganism can be covered by empirical antimicrobial regimen if highly suspected by suspected infectious foci, patient background and culture results (Provision of information for background question).

Rationale

The onset of infectious diseases and the risks of exacerbation should be considered when selecting empiric antimicrobials for the treatment of infection with MRSA and other specific bacteria, as described here.

  • MRSA

Reports have indicated that 30 and 50% of adults are temporary or permanent carriers, respectively, of Staphylococcus aureus [167, 168]. Bacterial loads in permanent carriers and the risk of S. aureus-based infection were particularly high [169]. The risks of carriage among patients with diabetes, hemodialysis, peritoneal dialysis, atopic dermatitis, medical exposure, recurrent S. aureus skin infections, human immunodeficiency virus (HIV) infection, and drug addiction were found to be high [