Management of Sepsis

The following is a scenario based learning discussion on the management of sepsis.
This article was submitted to ACAP by:

Dr Paul Robertson, MRCP MD
Specialty Doctor in Microbiology, Monklands District General Hospital, Monkscourt Avenue, Airdrie, ML6 0JS
A call to assess a febrile in-patient is a common scenario for junior medical staff or advanced nurse practitioners (ANP). Nursing and medical tradition would see this as an opportunity to administer an anti-pyretic and obtain blood cultures. Sadly these are often the only interventions that occur, with insufficient thought being given to identifying and treating the cause of the fever. I’ll argue that a call to assess a febrile inpatient represents an opportunity for early identification and treatment of sepsis, and that a shift in hospital culture – particularly out of hours – is needed to improve sepsis management. I hope to illustrate this point using a case presentation based on a mixture of real life scenarios and reference to a significant national audit of sepsis management – the Scottish Trauma Audit Group (STAG) audit into sepsis management in Scotlandi. Although this audit focused on the first 48hours of hospital admission, many of its learning points can be extrapolated to hospital- acquired sepsis.
Jen is an FY1 in surgery covering several surgical wards on a Saturday night shift as part of a hospital at night team. She receives a call from a general surgical ward to tell her that Mr. MacLeod, a 67 year old man who is five days post op following a colectomy has a temperature of 38.5°C. The staff nurse has given him some paracetamol, but has asked for blood cultures to be taken. Jen adds this to her list of jobs and gets round to seeing Mr. MacLeod half an hour later. The patient doesn’t look too bad so Jen takes a set of blood cultures and moves on to the rest of her job list.
Can you effectively triage your patient?
Any phone call about a febrile patient cannot be properly triaged without a full set of observations and some brief clinical detail. Patients with severe sepsis frequently have little in the way of symptoms and do not attract attention in the same way as other clinical emergencies such as myocardial infarction or GI bleeding. Patients commonly look “too well” to have severe sepsis, perhaps because early vasodilatation gives them a flushed appearance. Therefore, the severity of the patient’s illness may be under-appreciated by the referring staff member. The initial phone call provides an opportunity to more accurately establish the nature of the problem and, therefore, effectively prioritise the call.
In addition to Mr. MacLeod’s fever, he had a heart rate of 120 bpm, respiratory rate of 24/min and a blood pressure of 85/50. He was lucid and in no pain.
How sick is your patient?
Here, the patient fits at least three of the sepsis criteria, can you name them?  He has at least one feature consistent with severe sepsis, can you name it? These sepsis criteria are by no means universally known by medical and nursing staff, or incorporated into clinical practiceii. Had this information be obtained over the phone Jen could have realized that her patient potentially had severe sepsis and prioritised this as a clinical emergency. Although there are many causes of fever in hospital in-patients, infection is by far the most common. This infection is often in the context of sepsis, a pathophysiological state carrying a high mortality rate. Therefore, assessment of the febrile patient should be targeted towards identifying and effectively treating sepsis. This includes thorough clinical assessment and may mean obtaining blood tests looking for evidence of new organ dysfunction (such as FBC, Co-ag, U+E, LFT, lactate, ABG): as well as requesting tests to identify a source of infection (such as CXR, blood culture, urine culture, sputum culture, and culture of wound or line sites). A severity assessment – looking specifically for features of severe sepsis – should be performed in all patients having blood cultures performed (though severity assessment should not be limited to the taking of blood cultures).
Jen goes to the ward where she recognizes that Mr. MacLeod has severe sepsis. Clinical examination reveals no unexpected findings. His surgical scar appears to be healing nicely. He has a central line placed in the right femoral vein. The site does not appear inflamed. A CXR possibly show increased opacity at the right lung base. A catheter specimen of urine has one plus of protein and blood on dipstick. Jen sends blood and urine for culture. She makes a diagnosis of possible chest sepsis and prescribes co-amoxiclav (1.2g tid IV) and clarithromycin (500mg bd po). She alerts her senior colleague and moves on to her next call.
Have you started supportive therapy?
Correcting abnormal physiology is at least as important as giving antibiotics to patients with sepsis. Jen should have given this patient high flow oxygen, an intravenous fluid bolus (such as 20ml/kg of suitable crystalloid) and observed the response. The STAG audit demonstrated that failure to give appropriate fluid to hypotensive patients is common. Persisting hypotension or elevated lactate following adequate fluid challenge is diagnostic of septic shock and requires a prompt decision by senior medical staff to outline an appropriate ceiling of therapy. Assuming no contraindications to aggressive treatment, further fluid resuscitation – often guided by CVP measurement – and inotropic support would be indicated. This is best carried out in a CCU/HDU setting.
Are you sure where the source of sepsis is?
In this scenario, Jen decided that the patient had pneumonia. However, there’s no convincing evidence pinpointing infection at any one site. Minor CXR abnormalities are common in post-operative patients and need not represent pneumonia. Similarly, blood and protein in a catheter sample of urine cannot be taken as evidence of infection. The absence of erythema at the central line site does not rule out the possibility of line sepsis. Local audit data in Monklands suggest that the both the presence of sepsis AND the source of hospital acquired sepsis at the time of blood culture is less likely to be correctly identified than for community-acquired bacteraemia. Overlooking the possibility of vascular access device infection or intra-abdominal infection in favour of tentative diagnoses of respiratory or urinary tract infection is the commonest error. Misdiagnosis of the first assessing doctor is not unusualiii.
Getting the anatomical source wrong is important for two main reasons. First, necessary investigations and interventions may be delayed. Second, giving an effective antibiotic regime may be delayed. In the case above, co-amoxiclav and clarithromycin, while good empirical treatment for a community acquired pneumonia, is not adequate for a patient with hospital-acquired sepsis following recent abdominal surgery with a central line in place. Infection due to Methicillin-resistant Staphylococcus aureus (MRSA) or gram negative organisms resistant to co-amoxiclav in this situation is a real concern.
In reality bacteraemic infection can occur with no obvious source of infection, even after thorough clinical assessment. This is due to the fact that sepsis is an evolving process, but also that definitive diagnostic tests (such as CT scanning) may not be immediately available. In the management of many septic patients, therefore, there can a period of diagnostic uncertainty until culture results or definitive investigations to identify sources are undertaken. Where the source of sepsis is not apparent after careful initial assessment, staff should feel comfortable with the diagnosis of “sepsis, source undetermined”.
How are you going to diagnose the infection?
Establishing the microbiological cause of infection cannot normally be done without sampling body fluid or tissue. Prior antibiotic administration reduces the likelihood of culturing organisms, so samples – particularly blood cultures – are best taken before giving antibiotics. For patients with severe sepsis, two pairs of peripheral blood cultures are recommended to optimise the yield, as well as cultures from the ports of vascular-access devices. In patients where endocarditis is suspected, three sets of blood cultures from different sites should be obtained. Molecular diagnostic tests (antigen and PCR-based tests) are not culture-dependant and are therefore less affected by
antibiotic administration. Serological tests are not affected by antibiotic therapy, but typically require paired samples a week apart to give any useful information and so usually only give a diagnosis retrospectively.
Diagnostic Testing:
  • 2 sets of peripheral cultures 1 set of blood cultures from each central line lumen
  • 1 catheter specimen of urine for culture Chest X Ray
  • If Mr MacLeod’s wound or central line site had appeared erythematous or purulent, swabs (or better pus in a universal container) for culture would have been appropriate.
Thoughtful sampling is important in patient management for several reasons. First, it can establish a diagnosis. Second, a positive culture can allow targeted antibiotic treatment, reducing the unnecessary use of costly broad-spectrum antibiotics that select for Clostridium difficile or MRSA. Third, the organism isolated can provide a diagnostic clue as to the source of infection. Fourth, knowing the epidemiology and resistance patterns of the common pathogens in a hospital informs its antibiotic policy. Lastly, clusters of unusual pathogens (such as Legionella pneumophila) or unusually resistant organisms (such as carbapenem resistant Klebsiella sp.) may indicate an outbreak requiring infection control or public health intervention.
How quickly have the antibiotics been given?
The Surviving Sepsis Campaigniv advocates that intravenous antibiotics be given to patients with severe sepsis or septic shock within one hour of recognition. This is based predominantly on a single paper examining the relationship between mortality and time to antibiotics after the onset of hypotension in sepsisv. In reality it can be difficult to deliver antibiotics this quickly even when severe sepsis is recognised. ANPs are well placed to have a pivotal role in HAN teams for ensuring the prompt delivery of antibiotic therapy.
There is very little evidence to guide an ideal time within which to give antibiotics to patients who have non-severe sepsis. Pragmatically a 4 hours cut off might be a reasonable aim. It is likely that the longer the time to effectively treat sepsis, the longer the hospital stay would be anticipated to be. In a climate of increasing efficiency in delivering health care, prompt treatment of non-severe sepsis might be warranted in economic grounds, even if mortality benefits are uncertain.
Is the initial antibiotic regime effective?
Even when prompt antibiotics are given, it is not uncommon for them to have no in vitro activity against the isolated pathogen (around 25% of bacteraemias are treated with ineffective antibiotics in Monklands). Hospital-acquired sepsis is more frequently associated with organisms resistant to antibiotics commonly used in the setting of community-acquired infection, such as MRSA, extended spectrum β-lactamase (ESBL) producing gram negative bacilli, and the various Candida species. The combination of co-amoxiclav and clarithromycin would fail to cover any of these. The ideal empirical regime should take into account the severity of illness, the source of infection, and a judgement into the likeliness of resistant organisms (based on prior history, duration of stay and prior antibiotic exposure). Precise regimes will vary between hospitals, but one suitable regime might be intravenous vancomycin, gentamicin and metronidazole. This would provide active agents against most gram positive organisms including MRSA, most gram negative organisms, and anaerobes. Gaps in its cover would include vancomycin-resistant enterococci, many ESBL-producing bacteria and Candida species. Hospital antibiotic policies are carefully designed to factor in this complexity. Departure from the hospital antibiotic policy is more likely to lead to an ineffective antibiotic being given. Where there is doubt, or in the presence of significant allergies, seeking the advice from an infection specialist may be needed to ensure optimal first-line antibiotic treatment.
Is the team working together?
Managing a patient with severe sepsis cannot be done successfully without good teamwork between different healthcare disciplines and hospital departments. ANPs represent a vital component of that team – both in providing a thorough initial assessment, beginning prompt physiological support, obtaining appropriate specimens, and ensuring prompt delivery of antibiotics. Given their permanent presence in a HAN team, they are ideally placed to support and educate nursing and junior medical staff on sepsis management, helping to lead the cultural change required to improve sepsis management.
Sepsis is often under-recognised. When recognised, sepsis management rarely occurs with the attention to detail and urgency that its associated morbidity and mortality should mandate. Both sepsis recognition and management are poorer in hospital-acquired sepsis. Staff education and systematic improvements in enabling better sepsis recognition and care are needed. In the current financial climate, any systemic changes will likely need to be cost neutral. One such solution might to be use the blood culture as the route in to a brief, formalised sepsis assessment (or bundle). The aim should be to move Hospital-at-night culture from “I have taken blood cultures and given paracetamol for fever” to “I have identified this febrile patient as having severe sepsis, have taken appropriate diagnostic tests, escalated clinical care, and have promptly given antibiotics likely to be effective.” ANP have the training and clinical role to help ensure that this change occurs.
i STAG Audit Group (2010) Sepsis Management in Scotland accessed at
ii Ziglam, HM, et al. (2006) Knowledge about sepsis among training-grade doctors Journal of Antimicrobial Chemotherapy 57: 963-5
iii Bhandari, S (2009) A single-centre audit of junior doctors’ diagnostic activity in medical admissions Journal of the Royal College of Physicians 39: 307-12
iv Dellinger, RP, et al. (2008) Surviving Sepsis Campaign: International guidelines for management of severe sepsis and septic shock. Critical Care Medicine 36: 296-327
v Kumar, A, et al. (2006) Duration of hypotension prior to initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Critical Care Medicine 34:1589-96

5 thoughts on “Management of Sepsis

Leave a Reply