Thanks to everyone who contributed to the discussions about the initial treatment of Mr Watson who had severe sepsis and needed urgent fluid resuscitation. There were some great questions and debates and some really useful guidelines and resources were highlighted by UCLMS students. Millions of people die from sepsis each year worldwide, including 37,000 people a year in the UK, so every doctor in every speciality must keep up to date on latest evidence for best management.
The case and question was set by: Dr Laura-Jane Smith, Medical Registrar and Clinical Teaching Fellow at UCL medical school. This week’s expert was: Prof Mervyn Singer, Professor of Intensive Care Medicine, UCL.
The information provided at the start of the week was:
Mr Philip Watson, a 68 year old man, arrives in hospital via ambulance after the warden at his warden-controlled flat finds him unwell. He has hypertension and diabetes, and performs intermittent self catheterisation for prostatic hypertrophy. He lives alone and is usually independent. He is a retired paper mill foreman, a non-smoker and does not drink alcohol. He has no allergies. His medications include atenolol, perindopril, metformin, gliclazide, simvastation, aspirin and quinine sulphate.
His observations are:
- Temperature: 38.4 C, HR 105bpm, RR 22, GCS 14/15, BP 90/60mmHg, BM 7.1mmol/L, Sats 91% on air.
- A: airway clear
- B: tachypneoic, air entry bilaterally, no crepitations or wheeze
- C: thready pulse, cool peripherally, prolonged capillary refill time of 3 seconds, tachycardic, but heart sounds normal
- D: confused, GCS E3 V4 M6, pupils equal and reactive, no neck stiffness
- E: abdomen soft, no masses, suprapubic tenderness, but no guarding or rebound, bowel sounds present
His VBG shows:
- pH 7.34, pO2 5.4kPa, pCO2 5.6kPa, HCO3 24mEq/L, Lactate 3.7
Other investigations reveal:
- CXR: no focal consolidation
- Urine dip: leucocytes +++ nitrites ++ blood –
- Bloods: Haemoglobin 11.5g/dL, WCC 22, Potassium 4.7mmol/L Sodium 138mmol/L, Urea 14mmol/L, Creatinine 180mmol/L, Platelets 150×10^9, CRP 280
The question prompting discussion on Twitter was: What is the most appropriate fluid regimen for this patient?
Junior doctors are often first to see patients with sepsis which requires rapid identification and urgent treatment. Despite campaigns such as the “surviving sepsis” campaign, reliable delivery of basic urgent treatment such as early fluids and antibiotics is achieved in only 1/8 patients in UK hospitals (as evidenced on audits).
- It is important to give fluids in sepsis to reduce organ dysfunction and multi-organ failure by optimising tissue oxygen delivery and increasing organ perfusion. Fluids increase cardiac output (CO) by increasing venous return to the heart. If you’re a fan of formulae you’ll love this, which is a calculation of oxygen delivery and demonstrates the contributing components: DaO2 = CaO2 x CO, CaO2 = ([Hb] x SaO2 x 1.34) + (PaO2 x 0.225)
- There is no ‘perfect way’ to fluid resuscitate but we advise giving boluses of 250-500ml and, most importantly, reassessing after each one.
- Markers of improvement that you may wish to use include respiratory rate (which will fall), lactate (which will fall), urine output (which will increase >0.5ml/kg/hr), BP (which will rise), and central venous oxygen sats.
- There remains debate over the best type of fluids. Some evidence suggests an advantage for colloids. A recent study suggested advantage for albumin. There is evidence of harm using starches. There remain concerns about using high volumes of saline due to hyperchloraemic acidosis. It is really important to give fluids of some kind and at this point N-saline, Hartmans or albumin would all be justifiable. If what you have to hand is gelofusine this is justifiable too. Remember to reassess frequently.
Of course fluids are not the only treatment for sepsis, but they were the focus of the question this week.
- Surviving Sepsis Guidelines
- SIRS, sepsis and septic shock criteria
- In this paper, of an animal model of sepsis there was no difference to mortality based on type of fluid given
- In this meta-analysis of 17 studies that included patients with sepsis, the use of albumin was associated with a reduction in mortality
- In this paper the harm associated with excess N/saline is highlighted, due to hyperchloraemic acidosis (animal model)
- A Cochrane review from February 2013 found no evidence that colloids reduce the risk of dying compared with crystalloids, and one type of colloid (starches) might increase the risk of death
Prof Mervyn Singer, Professor of Intensive Care Medicine at UCLH says:
Very lively discussion, highlighting the current confusion in the literature!
- Defining sepsis, severe sepsis and septic shock is problematic using current criteria in terms of sensitivity and specificity of both diagnosis and severity of the condition. For example, shock in the true physiological sense is evidence of impaired organ perfusion (e.g. raised lactate + oliguria, etc..) and NOT a low blood pressure. The most important thing is to recognise sepsis and recognise sick patients.
- How much fluid? Answer – enough but not too much. Alas, while there is general agreement on this worthy aim, there is a multitude of different views on specific targets and it is largely an evidence-free zone. Some advocate an arbitrary central venous pressure (CVP) value, while a much more sophisticated physiological approach (i.e. the one I agree with!) is to fluid challenge with 200-250 ml challenges until the top of the Starling curve is reached (n.b. Starling was a UCL man, so must be right!). However, this requires measurement of stroke volume. In the absence of a technique to measure this, a reasonable surrogate is to give a 200-250 ml fluid challenge in 5-10 minutes, wait 5-10 minutes and assess whether the CVP has risen 3 mmHg or more. If yes, no more fluid is needed, if no, can repeat fluid challenges until the rise of 3 mmHg or more is seen. Difficulty arises when no CVP or stroke volume is immediately available, e.g. sick patient arriving in A&E or acutely deteriorating on a general ward. Here, clinical acumen is needed to determine how much fluid is needed, but studies show this is not particularly reliable, even with experienced clinicians. Clues include improvements in BP (going up), heart rate (coming down), urine output (increasing) and biochemical markers such as an improving lactate and base deficit. My message is to start sensible ‘first aid’ therapy with repeated 200-250 ml fluid challenges, and if the patient does not promptly improve then ask for help (usually from ITU). I personally disagree with a 20 ml/kg fluid challenge in adults – this equates to approximately 1500 ml. In some septic patients this may be excessive and it’s easier to give smaller aliquots and repeat as necessary, than to give too much and then worry about how to cure fluid overload!
- Type of fluid? Again, lots of heated debate. Starches are currently ‘out’ because of a reported increased risk of renal failure and, in some studies, mortality. However, I’ve seen an as yet unpublished large multicentre trial from France where the exact opposite was seen with colloids (2/3 of which was starch!). So what is the truth? Albumin is not generally used in the UK for fluid resuscitation because of high costs. Gelatins e.g. gelofusin, geloplasma, are currently viewed as ‘ok’. Crystalloids can be used but approximately 15-20% more needs to be given compared with colloids that hang around longer in the intravascular compartment. 5% glucose is not a resuscitation fluid. N–saline or Hartmann’s can be used – there are pros and cons for each and no head-to-head RCT has yet been performed comparing them. N-saline can cause a hyperchloraemic acidosis if too much is given, whereas it is difficult to add extra potassium to Hartmann’s.
- Surviving sepsis guidelines – these are generic and aimed at populations rather than individual patients. Implementing guidelines arguably leads to more reliable delivery of certain components of care. Again, no specific head-to-head comparison studies exist but temporal cohort studies show improvement over time if some resuscitation is given, compared to doing nothing or doing something belatedly. A previously fit 20 year old will react differently and likely require different goals compared to an 80 year old with co-morbidities. So use guidelines as a guide rather than rules set in stone. If you feel deviation is appropriate and can rationalise good reasons why, don’t be a maverick – call for more senior help and discuss.
A fascinating area of medicine and physiology, and an important condition for junior doctors to recognise and treat promptly. As an aside, I’ve recently been asked to co-chair a European-North American consensus conference to redefine sepsis – watch this space in 2014!
Case 2 will go live on Tuesday 7th May. See you then!