Sepsis is due to infection, it is a life-threatening condition with a high mortality even when treated quickly. Patients continue to die from sepsis. So after 15 years of the Surviving Sepsis campaign are we diagnosing and treating sepsis better? I believe the Surviving Sepsis campaign has done a brilliant job in improving the recognition of sepsis and the subsequent management of these patients. However it has also led to a dramatic over diagnosis of sepsis in patients who are not really septic. Locally, sepsis has become the most commonly documented indication for antibiotics! This is ahead of UTI, community acquired pneumonia and cellulitis. Yet sepsis is not nearly as common as these infections, so are we actually getting better at diagnosing sepsis? It appears not, and if we are not, is the answer to redefine what sepsis is?!? It appears so!!! There seems to have been a flurry of papers in 2016 related to sepsis and how to define it; but does this actually add to the management of patients and are these actually even new definitions of sepsis!? The papers have been written by “The Third International Consensus Definitions for Sepsis and Septic Shock (sepsis 3)” and are published in the Journal of the American Medical Association (JAMA)…but these are not free to access for all…mutter, mutter, mutter.
So how do you reduce the over diagnosis of sepsis whilst not missing the truly septic patient? We do need better clinical definitions of sepsis and I was looking forward to these papers…however I was left disappointed. These new papers, in my opinion, cause further confusion and do not solve the current problem in sepsis management; distinguishing the septic patient from all the others in a busy emergency department.
Currently sepsis is diagnosed by answering 3 questions:
1. Does the patient have a potential source of infection? e.g. pneumonia, empyema, UTI, acute abdomen, meningitis, infective endocarditis, skin/soft tissue infection, bone or joint infection, wound infection, CVC infection etc
2. Does the patient have new signs or symptoms of infection? TWO or more of the following:
• Hyperthermia >38.3°C
• Tachycardia >90bpm
• Leucopaenia <4x109/L
• Altered mental state
• Hypothermia <36°C
• Tachypnoea >20 bpm
• Leucocytosis >12x109/L
• Hyperglycaemia >6.5mmol/L
3. Does the patient have evidence of organ dysfunction remote to the site of infection?
• Systolic blood pressure <90mmHg OR Mean arterial pressure <65mmHg
• Systolic blood pressure >40mmHg below baseline
• Bilateral pulmonary infiltrates PLUS O2 required to keep O2 saturations >90%
• Bilateral pulmonary infiltrates PLUS PaO2/FiO2 ratio <300*
• Bilirubin >34 mmol/L
• Creatinine >175mmol/L OR urine output <0.5ml/kg/hour for more than 2 hours
• Coagulopathy INR >1.5 OR APTT >60 seconds
• Platelet count <100x109/L
• Lactate >2mmol/L
*PaO2 measured in mmHg (1kPa = 7.5mmHg), FiO2 as % converted into a decimal e.g. 32% = 0.32
If YES to questions 1, 2 and 3 then the patient meets the criteria for SEVERE SEPSIS and requires urgent antibiotics and supportive treatment.
The new definition presented by the consensus paper states that sepsis should now be defined as “life-threatening organ dysfunction caused by a dysregulated host response to infection”. However they also recognise “that no current clinical measures reflect the concept of a dysregulated host response”. This is not very helpful in a clinical setting. I also question why the host’s response to the infection has to be dysregulated (abnormal or impaired) in order for the patient to be septic. The patient’s response may be unable to compensate adequately for the damage being caused by the infection or damage may be being caused as a consequence of keeping the patient alive, whether this response is working or not.
In a clinical setting, the consensus paper states that sepsis can be diagnosed using either:
Currently sepsis is diagnosed by answering 3 questions:
1. Does the patient have a potential source of infection? e.g. pneumonia, empyema, UTI, acute abdomen, meningitis, infective endocarditis, skin/soft tissue infection, bone or joint infection, wound infection, CVC infection etc
2. Does the patient have new signs or symptoms of infection? TWO or more of the following:
• Hyperthermia >38.3°C
• Tachycardia >90bpm
• Leucopaenia <4x109/L
• Altered mental state
• Hypothermia <36°C
• Tachypnoea >20 bpm
• Leucocytosis >12x109/L
• Hyperglycaemia >6.5mmol/L
3. Does the patient have evidence of organ dysfunction remote to the site of infection?
• Systolic blood pressure <90mmHg OR Mean arterial pressure <65mmHg
• Systolic blood pressure >40mmHg below baseline
• Bilateral pulmonary infiltrates PLUS O2 required to keep O2 saturations >90%
• Bilateral pulmonary infiltrates PLUS PaO2/FiO2 ratio <300*
• Bilirubin >34 mmol/L
• Creatinine >175mmol/L OR urine output <0.5ml/kg/hour for more than 2 hours
• Coagulopathy INR >1.5 OR APTT >60 seconds
• Platelet count <100x109/L
• Lactate >2mmol/L
*PaO2 measured in mmHg (1kPa = 7.5mmHg), FiO2 as % converted into a decimal e.g. 32% = 0.32
If YES to questions 1, 2 and 3 then the patient meets the criteria for SEVERE SEPSIS and requires urgent antibiotics and supportive treatment.
The new definition presented by the consensus paper states that sepsis should now be defined as “life-threatening organ dysfunction caused by a dysregulated host response to infection”. However they also recognise “that no current clinical measures reflect the concept of a dysregulated host response”. This is not very helpful in a clinical setting. I also question why the host’s response to the infection has to be dysregulated (abnormal or impaired) in order for the patient to be septic. The patient’s response may be unable to compensate adequately for the damage being caused by the infection or damage may be being caused as a consequence of keeping the patient alive, whether this response is working or not.
In a clinical setting, the consensus paper states that sepsis can be diagnosed using either:
- Sequential [Sepsis-related] Organ Failure Assessment (SOFA)
- quickSOFA (qSOFA)
Click for larger image
The SOFA score is complicated and based upon the measurement of factors relating to respiratory, coagulation, cardiovascular, liver, CNS and renal parameters. A score of ≥ 2 points is considered to be sepsis. However all these measurements take time, something a doctor treating a septic patient does not have. Therefore, I believe, SOFA is not useful in diagnosing sepsis; in fact the consensus paper states it is measured retrospectively, so hardly useful in a clinical setting.
The qSOFA is however much more practical and is based on whether the patient has ≥ 2 of the following:
So my disappointment is that they haven’t given us a useable tool which can be applied in a clinical setting within the “Golden Hour”; the hour in which sepsis should be diagnosed and treatment started.
Neither of these “definitions” of sepsis (SOFA or qSOFA) actually defines sepsis at all; they are criteria that predict mortality from infection. A SOFA score of ≥ 2 is associated with a 10% mortality; the prediction of mortality for the qSOFA score is not given but the consensus paper does admit it “is less robust” than the SOFA score. The lack of practical application of the new definitions and criteria for predicting increased mortality suggest that the “new definitions” are really only of use for ensuring better comparison across future studies into sepsis and I believe this is supported by the following statements that appear at the end of the consensus paper:
“Neither qSOFA nor SOFA is intended to be a stand-alone definition of sepsis. It is crucial, however, that failure to meet 2 or more qSOFA or SOFA criteria should not lead to a deferral of investigation or treatment of infection or to a delay in any other aspect of care deemed necessary by the practitioners… The [consensus group] wishes to stress that SIRS criteria may still remain useful for the identification of infection… The proposed criteria should aid diagnostic categorization once initial assessment and immediate management are completed. qSOFA or SOFA may at some point be used as entry criteria for clinical trials”.
One of the main reasons for wanting a quick diagnosis of sepsis is to identify those who may have septic shock, as it has an associated mortality of 40%. The consensus paper defines septic shock “as a subset of sepsis in which underlying circulatory, cellular, and metabolic abnormalities are associated with a greater risk of mortality than sepsis alone”. The consensus paper states that septic shock is hypotension requiring vasopressor therapy to maintain mean blood pressure ≥ 65mmHg AND serum lactate > 2mmol/L after fluid resuscitation. However, I don’t think their new definition of septic shock is practical, as septic shock can only be diagnosed AFTER a patient has been fluid resuscitated and started on vasopressor therapy. This does not help doctors diagnose septic shock in a clinical setting; it takes too long and requires specialist intervention. The current method of 3 questions at least allows a diagnosis of sepsis and septic shock (severe sepsis) to be made at the bedside quickly by junior staff.
In reality the diagnosis of sepsis boils down to experience, a gut feeling almost. A diabetic patient with a UTI and ketoacidosis, a bronchiectatic patient with exacerbation of COPD and chronic renal failure or a myelodysplastic patient with cellulitis may all technically fulfil the criteria for severe sepsis without actually being septic. Many senior doctors cannot accurately define what sepsis is or who is in septic shock, they just know! It is not a fixed parameter but a collection of factors, no one has yet defined sepsis more accurately than the lay version…“a life-threatening condition that arises when the body’s response to an infection injures its own tissues and organs” and no current diagnostic tool works efficiently enough not to over diagnose sepsis.
Currently we tend to over diagnose because over treating is deemed less harmful than missing a septic patient. However Microbiologists are acutely aware of the consequences to more patients being over diagnosed with sepsis:
How have the new definitions for sepsis and septic shock been created?
These new definitions are based upon a systematic review of 44 studies of the outcome of the management of patients with sepsis, a total of 166,470 patients! A large study should produce robust results, however if the original studies used a definition of sepsis which the consensus paper now deems inadequate further analysis of these studies is subject to the same flaws in the definition.
It is unusual for a systematic review to analyse the validity of the selection criteria for the studies. Most studies compare treatments for a specific clinical condition; they don’t look at the criteria to define the clinical condition. This consensus looks at the selection criteria for the condition, states that it is wrong but then still uses the selected patients to look for the new parameters to define the condition. The study should go back to the whole population to define the parameters not just those filtered or chosen within the set parameters within the old studies. This is a bit of a tricky concept so let’s consider an analogy. To define handsomeness old studies considered men with beards. Some believe new definitions of handsomeness are now required. However if only the old studies are used and non-bearded men are not represented, a biased conclusion might be drawn - short beards make a man most handsome. This ignores the fact that non-bearded men were not represented in the analysis and could in fact be even more attractive and handsome than short bearded men. Okay, this is an odd analogy, but hopefully it makes the point clearer.
Do the new definitions help in the diagnosis of sepsis and septic shock?
Well… I have to say I’m a little underwhelmed by the new definitions. I don’t really think that they will help us improve the current diagnosis and management of patients. I think their value lies in setting specific measurable criteria which can be used to compare research into sepsis which may in the future lead to better diagnosis and management of patients.
A potential danger in applying the new definitions of sepsis based on SOFA is that more patients with infection will be under diagnosed and under-treated and therefore not start appropriate therapy until they fulfil the sepsis criteria, by which stage they will have a mortality of 10-40%.
Alternatively if qSOFA is used we will see even more patients being diagnosed with sepsis and the current over diagnosis problems will just get worse.
There is clearly still a lot more work to be done. The problem lies in the distinction between defining what sepsis is and providing an effective tool for diagnosing sepsis quickly in patients. I like the lay definition that the consensus paper puts forwards for sepsis as “a life-threatening condition that arises when the body’s response to an infection injures its own tissues and organs”. It’s simple and it’s understandable. What we need now is an accurate diagnostic tool.
The references
The qSOFA is however much more practical and is based on whether the patient has ≥ 2 of the following:
- Respiratory rate ≥ 22/min
- Altered mentation (i.e. altered Glasgow Coma Score)
- Systolic blood pressure ≤ 100mmHg
So my disappointment is that they haven’t given us a useable tool which can be applied in a clinical setting within the “Golden Hour”; the hour in which sepsis should be diagnosed and treatment started.
Neither of these “definitions” of sepsis (SOFA or qSOFA) actually defines sepsis at all; they are criteria that predict mortality from infection. A SOFA score of ≥ 2 is associated with a 10% mortality; the prediction of mortality for the qSOFA score is not given but the consensus paper does admit it “is less robust” than the SOFA score. The lack of practical application of the new definitions and criteria for predicting increased mortality suggest that the “new definitions” are really only of use for ensuring better comparison across future studies into sepsis and I believe this is supported by the following statements that appear at the end of the consensus paper:
“Neither qSOFA nor SOFA is intended to be a stand-alone definition of sepsis. It is crucial, however, that failure to meet 2 or more qSOFA or SOFA criteria should not lead to a deferral of investigation or treatment of infection or to a delay in any other aspect of care deemed necessary by the practitioners… The [consensus group] wishes to stress that SIRS criteria may still remain useful for the identification of infection… The proposed criteria should aid diagnostic categorization once initial assessment and immediate management are completed. qSOFA or SOFA may at some point be used as entry criteria for clinical trials”.
One of the main reasons for wanting a quick diagnosis of sepsis is to identify those who may have septic shock, as it has an associated mortality of 40%. The consensus paper defines septic shock “as a subset of sepsis in which underlying circulatory, cellular, and metabolic abnormalities are associated with a greater risk of mortality than sepsis alone”. The consensus paper states that septic shock is hypotension requiring vasopressor therapy to maintain mean blood pressure ≥ 65mmHg AND serum lactate > 2mmol/L after fluid resuscitation. However, I don’t think their new definition of septic shock is practical, as septic shock can only be diagnosed AFTER a patient has been fluid resuscitated and started on vasopressor therapy. This does not help doctors diagnose septic shock in a clinical setting; it takes too long and requires specialist intervention. The current method of 3 questions at least allows a diagnosis of sepsis and septic shock (severe sepsis) to be made at the bedside quickly by junior staff.
In reality the diagnosis of sepsis boils down to experience, a gut feeling almost. A diabetic patient with a UTI and ketoacidosis, a bronchiectatic patient with exacerbation of COPD and chronic renal failure or a myelodysplastic patient with cellulitis may all technically fulfil the criteria for severe sepsis without actually being septic. Many senior doctors cannot accurately define what sepsis is or who is in septic shock, they just know! It is not a fixed parameter but a collection of factors, no one has yet defined sepsis more accurately than the lay version…“a life-threatening condition that arises when the body’s response to an infection injures its own tissues and organs” and no current diagnostic tool works efficiently enough not to over diagnose sepsis.
Currently we tend to over diagnose because over treating is deemed less harmful than missing a septic patient. However Microbiologists are acutely aware of the consequences to more patients being over diagnosed with sepsis:
- The use of unnecessary antibiotics and other drugs can cause damage to patients resulting in harm e.g. Clostridium difficile associated disease, renal failure secondary to nephrotoxic drugs or bone marrow suppression
- Antibiotics for sepsis are primarily targeted against Gram-negative bacteria and therefore are not appropriate for all types of infections e.g. community acquired pneumonia, cellulitis, meningitis, etc and inadvertently fail to treat the infection the patient actually has
- Increased use of empirical broad-spectrum antibiotics drives antibiotic resistance e.g. extended-spectrum beta-lactamases, carbapenemases, glycopeptide-resistant Enterococcus
- Over use of laboratory or radiological investigations as part of sepsis pathways leads to the risk of false positive results. Reacting to the result because it’s a septic patient rather than dismissing a result which would normally be regarded as a contaminant may adversely affect the patients care
How have the new definitions for sepsis and septic shock been created?
These new definitions are based upon a systematic review of 44 studies of the outcome of the management of patients with sepsis, a total of 166,470 patients! A large study should produce robust results, however if the original studies used a definition of sepsis which the consensus paper now deems inadequate further analysis of these studies is subject to the same flaws in the definition.
It is unusual for a systematic review to analyse the validity of the selection criteria for the studies. Most studies compare treatments for a specific clinical condition; they don’t look at the criteria to define the clinical condition. This consensus looks at the selection criteria for the condition, states that it is wrong but then still uses the selected patients to look for the new parameters to define the condition. The study should go back to the whole population to define the parameters not just those filtered or chosen within the set parameters within the old studies. This is a bit of a tricky concept so let’s consider an analogy. To define handsomeness old studies considered men with beards. Some believe new definitions of handsomeness are now required. However if only the old studies are used and non-bearded men are not represented, a biased conclusion might be drawn - short beards make a man most handsome. This ignores the fact that non-bearded men were not represented in the analysis and could in fact be even more attractive and handsome than short bearded men. Okay, this is an odd analogy, but hopefully it makes the point clearer.
Do the new definitions help in the diagnosis of sepsis and septic shock?
Well… I have to say I’m a little underwhelmed by the new definitions. I don’t really think that they will help us improve the current diagnosis and management of patients. I think their value lies in setting specific measurable criteria which can be used to compare research into sepsis which may in the future lead to better diagnosis and management of patients.
A potential danger in applying the new definitions of sepsis based on SOFA is that more patients with infection will be under diagnosed and under-treated and therefore not start appropriate therapy until they fulfil the sepsis criteria, by which stage they will have a mortality of 10-40%.
Alternatively if qSOFA is used we will see even more patients being diagnosed with sepsis and the current over diagnosis problems will just get worse.
There is clearly still a lot more work to be done. The problem lies in the distinction between defining what sepsis is and providing an effective tool for diagnosing sepsis quickly in patients. I like the lay definition that the consensus paper puts forwards for sepsis as “a life-threatening condition that arises when the body’s response to an infection injures its own tissues and organs”. It’s simple and it’s understandable. What we need now is an accurate diagnostic tool.
The references
- The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). Singer M, Deutschman C, Seymour C, et al. JAMA 2016; 315(8): 801-810
- Developing a New Definition and Assessing New Clinical Criteria for Septic Shock for the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). Shankar-Hari M, Phillips G, Levy M, et al. JAMA 2016; 315(8): 775-787
- New Definitions for Sepsis and Septic Shock, Continuing Evolution but with much still to be done. Abraham E. JAMA 2016; 315(8): 757-759
RSS Feed
