Sepsis: recognition and treatment
Key points
The heterogeneous nature of sepsis may result in current definitions of sepsis being too broad for optimal clinical use
Systemic inflammatory response syndrome can be triggered by tissue or cell damage caused by non-microbial insults, such as trauma, necrosis, extremes of temperature, chemicals (including chemotherapeutics) and lack of nutrients/oxygen
Genetic variation in host response to insults (microbial or endogenous) are increasingly recognised as a significant predictor of outcome of sepsis
The combination of physiological early warning systems and biochemical markers serve to aid early recognition of sepsis and therefore improve outcome
Timely delivery of appropriate antibiotics is associated with higher survival rates; it is estimated that every hour of delay equates to a 6% increase in mortality
There is robust evidence for evidence for early goal-directed haemodynamic support
Care bundles are small (3–5) sets of evidenced-based interventions that require healthcare systems to deliver all components consistently, the rationale being that their reliable implementation will yield better results than their inconsistent individual adoption. Recent evaluation of healthcare centres participating in the surviving sepsis campaign shows improvement in performance measures and adjusted odds ratio for mortality
The American College of Chest Physicians and the Society of Critical Care first published definitions for the wide spectrum of presentations described by sepsis syndromes in 1992 (Fig 1).1 The driver for this was to facilitate better clinical trials and consolidate the diverse definitions available. However, because of the heterogeneous nature of sepsis these definitions may be too broad for clinical use. The consensus from the 2001 International Sepsis Definitions Conference suggested that these definitions cause difficulty staging sepsis, prognosticating for host response and can be overly sensitive, especially in the case of systemic inflammatory response syndrome (SIRS).
Systemic inflammatory response syndrome (SIRS): current concepts
Multicellular animals have developed means of surveillance, defence and repair when confronted by microbial challenge.2 The mechanisms behind the activation and co-ordination of this response are complex and not fully elucidated. Exogenous microbial molecules (pathogen-associated molecular patterns (PAMPs)) are recognised by cells of the innate and acquired immune pathway via pattern recognition receptors (PRRs) such as Toll-like receptors (TLRs). These in turn activate cell signalling pathways, triggering a SIRS through secretion of inflammatory mediators and activation of leukocytes (Fig 2).
SIRS can be also triggered by tissue or cell damage caused by non-microbial insults, such as trauma, necrosis, extremes of temperature, chemicals (including chemotherapeutics) and lack of nutrients/oxygen.2 Endogenous molecules (alarmins) are released by tissue damage, triggering a SIRS via PRRs. Alarmins identified to date include uric acid, mitochondrial DNA, advanced glycation end-products and high mobility group box 1(HMGB1). Increased extracellular expression of HMGB1 has been associated with autoimmune conditions such as systemic lupus erythematosus and rheumatoid arthritis. Together, exogenous PAMPs and endogenous alarmins form a larger set of triggering molecules: damage-associated molecular patterns (DAMPs). In addition, genetic variation in host response to insults (microbial or endogenous) is increasingly recognised as a significant predictor of outcome. Each episode of sepsis has elements of systemic inflammation and immune depression throughout its course.3
The severity of SIRS and compensatory anti-inflammatory response syndrome (Fig 3) will vary from individual to individual, depending on their predisposing genetic influences as well as the nature of the pathological insult. For example, genetic polymorphism of TLR 4 predisposes to septic shock in Gram-negative infection.
Extremes of hyper-inflammation or immune depression are associated with poorer outcomes.4 The implications of host response variations cannot be easily quantified, thus complicating design and evaluation of therapeutic sepsis trials. Studies of host genomic response to sepsis (such as the UK-based GAinS study)4 are currently underway but it remains uncertain whether the knowledge gained will translate into effective interventions.
Recognising sepsis
The incidence of sepsis in the US is 3 per 1,000 population per year.5 A large UK clinical database study6 suggests that both the number of critical care admissions with severe sepsis and the total number of deaths from severe sepsis and septic shock are rising. Overall mortality for those admitted with severe sepsis is decreasing through advances in medical care, but it remains unacceptably high (>30%) with approximately 37,000 deaths annually in the UK.7
In a recent international survey investigating the views of physicians regarding sepsis,8 86% of responders felt that the symptoms of sepsis could easily be misattributed to other medical conditions, leading to delayed diagnosis and treatment, while 81% felt that a clear, globally accepted clinical definition would help.
Improving sensitivity of identification
Strategies which improve the sensitivity of identifying patients at risk include complementary systems that capture physiological parameters,9 routine biochemical measurements10 or a combination of both. Scoring systems that allocate weighted points based on the degree of deviation of a patient's physiological parameter (eg heart rate) from the normal range are commonly employed.9
‘Track and trigger systems’ TTS use the following indicators to trigger a review and indicate the level of care and frequency of monitoring:
summation of the points from the scoring systems (often termed Early Warning Score) above a predefined threshold
trends of rising summated points, and/or
degrees of deviation of a single physiological parameter.
Numerous locally derived aggregate weighted TTS are currently in widespread use, with poor evidence for their utility, reliability and validity.11
National standardisation. A national TTS that allows for standardisation of care and homogeneous training of medical personnel is likely to be beneficial.12 The Royal College of Physicians is leading the development of a national early warning scoring system, to be published in 2012.
Biomarkers. The role of biomarkers to prognosticate, stage and diagnose sepsis is a work in progress. Serial serum lactate has prognostic value for mortality13 (40% for lactate >4 mmol/l vs <15% for lactate <2 mmol/l) and predictive value for critical care admission. Similarly, serum procalcitonin14 may be more specific for bacterial-induced severe sepsis and thus aid early diagnosis. Mediators currently being investigated include adrenomedullin, cellular adhesion molecules and interleukin-6.
In the future, the combination of physiological early warning systems and biochemical markers may aid early recognition and treatment.
Treating sepsis
The International Surviving Sepsis Campaign was launched in 2002, with the ambitious aim of reducing sepsis-associated mortality by 25% over five years. The main objectives were to:
increase awareness
define standards of care informed by evidence from clinical trials, and
produce management policy.
These standards were translated into care bundles for implementation in partnership with the Institute for Healthcare Improvement. Care bundles are small (3–5) sets of evidenced-based interventions that require healthcare systems to deliver all the components consistently, the rationale being that failure to implement even one component would yield poor outcomes. Compliance is therefore measured in an ‘all-or-nothing’ manner. A recent study shows improvement in performance measures and adjusted odds ratio for mortality.15
Care bundles
The recommended care bundles were organised in two phases: sepsis resuscitation and sepsis management (Tables 1 and 2), each set of performance measures requiring completion within six and 24 hours, respectively. Broadly speaking, the resuscitation bundle can be divided into two distinct strategies.
Expedient source isolation/control and early administration of empirical broad-spectrum antibiotics
Timely delivery of appropriate antibiotics is associated with higher survival rates. It is estimated that every hour of delay equates to a 6% increase in mortality.16 Source control is vital, but is determined in part by aetiology. For example, a patient with bacterial peritonitis from a perforated diverticular abscess would be likely to require urgent surgical intervention, whilst another with bacterial meningitis would require prompt clinical diagnosis, judicious and expedient antibiotic administration as well as excellent supportive care.
Early goal-directed haemodynamic support with appropriate monitoring
The evidence for early goal-directed therapy stems from a single randomised control trial (RCT) of protocolised haemodynamic support, involving both diagnostic and therapeutic interventions, for patients presenting to an emergency department in the US with severe sepsis or septic shock.17 The 90-day hospital mortality was significantly reduced (46.5–30.5%). The UK-based multicentre PRoMIse RCT is currently underway to investigate further the efficacy of early protocolised management in sepsis, with added examination of cost-effectiveness.
Evolving evidence
The evidence for interventions in the management bundle (Tables 1 and 2) is evolving and some interventions need adjusting. For example, there is robust evidence18 for lung protective strategies in mechanically ventilated patients but, following the multicentre CORTICUS trial, the survival benefit of universal low-dose glucocorticoid therapy in septic shock is uncertain. The recent PROWESS-SHOCK trial failed to show 28-day survival benefit in patients with septic shock, leading to drotrecogin alpha (activated) being withdrawn in October 2011. The NICE-SUGAR trial unexpectedly showed excess mortality for critically ill patients in the tightly controlled arm; earlier guidance has been revised to recommend a higher glucose threshold (<9.9 mmol/l or 180 mg/dl). The current 2008 guidelines are under revision to incorporate recently published evidence. The implementation of sepsis care bundles has shown benefit and they must evolve as new evidence emerges.
The hunt for a magic bullet
The collective term ‘sepsis’ predicates the premise that a common final pathway for systemic inflammation exists which may be targeted and will improve outcome if suppressed.19 Many potential agents have been trialled. Those currently being investigated in sepsis RCTs include:
statins, which seem to have pleiotropic effects extending beyond their favourable action on lipid profiles
talactoferrin alpha, a recombinant form of the protein lactoferrin found in human breast milk, which seems to have immunomodulatory power via action on gut-associated lymphoid tissue.
The central premise of a common final pathway has never been proven; such a pathway may either not exist or be too final a point to be effectively targeted. Sepsis is a broad-based heterogeneous condition consisting of many disease entities, all with individual natural histories and variations in host response. Sepsis syndromes may therefore lack genuine homogeneity and be the wrong point of entry for study – in which case, there may be no magic bullet.
Conclusions
For the moment, and for the foreseeable future, the evidence lies with early recognition, source control, early antibiotic administration and excellent supportive care.
- © 2012 Royal College of Physicians
References
- ↵
- Bone RC,
- Balk RA,
- Cerra FB,
- et al
- ↵
- Bianchi ME
- ↵
- Cavaillon JM,
- Adib-Conquy M,
- Cloëz-Tayarani I,
- Fitting C
- ↵
- ↵
- ↵
- ↵
- Daniels R
- ↵
- ↵
- ↵
- ↵
- ↵
- Royal College of Physicians
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
Article Tools
Citation Manager Formats
Jump to section
Related Articles
- No related articles found.
Cited By...
- Inducible Nitric Oxide Synthase in Circulating Microvesicles: Discovery, Evolution, and Evidence as a Novel Biomarker and the Probable Causative Agent for Sepsis
- Comparison of qSOFA with current emergency department tools for screening of patients with sepsis for critical illness
- Isolated cell behavior drives the evolution of antibiotic resistance
- Youve been bleeped: symptom control in the unwell patient
- A swollen right eye in a child
- Tolerization of Inflammatory Gene Expression