Response to: Assessment of Fluid responsiveness in the Acute Medical Patient and the Role of Echocardiography
I read the article, ‘Assessment of Fluid responsiveness in the Acute Medical Patient and the Role of Echocardiography’ by Dr Parulekar and Dr Harris with interest. It rightly highlights the challenges posed when assessing for fluid resuscitation in a pressured setting with limited information. This scenario is a routine one for our speciality, which is why it is a concern that the evidence-base outside of intensive care remains limited. Of particular relevance to the Acute Medical specialist, the article acknowledges that performing a focussed-echocardiogram on all acutely-unwell patients is “impractical”. Developing a quick and straightforward approach to fluid resuscitation assessment should be a high research priority for Acute Medicine.
However, there are several statements described in the article which warrant correction, particularly as they reflect misconceptions that are rife in the Intensive Care literature:
- The assumption that fluid responsiveness is equivalent to hypovolaemia:
As the article acknowledges, “No suitably powered RCT has assessed the role of stroke volume guided fluid administration as a resuscitation goal.” Furthermore, fluid responsiveness has been demonstrated in healthy volunteers suggesting it may be a normal physiological condition. Even if there was a unanimous agreement for its use, it is currently a poorly-defined concept with no consensus definition.
Therefore, significant question marks remain about the diagnostic ability of fluid responsiveness and much work is needed before it can be reliably used as a test for hypovolaemia.
2. The contradiction between the stated utility of IVC measurements and that of CVP:
The article states that “IVC [inferior vena cava] size decreases in hypovolaemia” and later notes “a very small collapsing IVC in a shocked patient suggests fluid tolerance”. It goes on to say that “IVC diameter predicts central venous pressure [CVP]”. Then the article contradicts this link by stating that “CVP has little or no role in volume assessment”.
The final statement is based on a systematic review in 2008, as acknowledged in the article. The review was updated in 2013. Both compared the ability of CVP to predict fluid responsiveness, however, they did so based on the assumption that fluid responsiveness predicts hypovolaemia. This is far from proven, as discussed above. Therefore, the utility of CVP remains unclear and should be explored along with IVC measurements as a fluid assessment tool.
3. The meaning of the transient response to a fluid bolus:
Finally, the article suggests that a “fast response device” should be used to measure cardiac response before and after a passive leg raise, because the subsequent “changes in cardiac output may be transient”. This transient change is not unique to a passive leg raise. The haemodynamic improvements that follow an intravenous fluid bolus are similarly temporary. One study found that cardiac output returned to baseline values 90 minutes after a bolus in fluid responsive patients.
Instead of finding a “fast response device” to measure transient haemodynamic improvements, we should be asking if the benefit of fluid resuscitation is also transient, particularly in conditions such as sepsis. We should also question whether rapid fluid boluses cause harm from subsequent oedema, and explore whether this harm persists after the haemodynamic benefit has disappeared. Are we temporarily boosting physiological markers whilst at the bedside, only for the fluid to leak into the interstitium after we have moved on to our next patient?
In conclusion, I applaud the authors for the highlighting the important topic of fluid assessment. Despite nearly two centuries of use, the benefits and harms of intravenous fluid are still poorly understood. This is a vital research topic for our speciality, so I hope they will join me and others in addressing the evidence gaps and research questions that are highlighted by this letter.
BSc (Hons) MBChB MRCP (Acute Medicine)
We share your understanding that fluid responsiveness is NOT equivalent to hypovolaemia, and that it also does not equate to fluid administration being of benefit. We offer our apologies should our article have suggested otherwise. Indeed, this is one of the key reasons why we advocate fluid administration against measurable goals, such as stroke volume, as well as blood pressure and pulse. Research suggests that around 30-70% of patients presenting to the ED or ICU are fluid responders,1-2 i.e. they increase stroke volume by > 10% following a fluid bolus of 250-500 ml delivered in 15-30 minutes3 (most commonly). It is likely that most patients with acute hypovolaemia are fluid responders. Euvolaemic or hypervolaemic patients may or may not be fluid responders. Goal directed therapy was based on the premise that continued fluid administration to hypervolaemic levels, coupled with red cell transfusion and vasopressors, would increases oxygen delivery and that this would improve organ function/reverse organ hypoperfusion. This approach does not improve outcomes in sepsis.4-6 As such we agree that the presence of fluid responsiveness does not imply the requirement for fluids or that fluid administration improves outcomes. Sepsis is also not a hypovolaemic state but accounts for around two thirds of shock seen acutely in the Emergency Departments, and is the patient group from which the bulk of research subjects are recruited. We propose fluid administration against stroke volume as we believe this is the current best approach, in part to at least protect non-responders from excess fluid administration as this is known to be harmful.
In point 2 you quite rightly point out the contradiction between data suggesting central venous pressure (CVP) is not a useful endpoint for fluid resuscitation (although low CVP and low blood pressure do suggest fluid will improve both stroke volume and blood pressure) and some studies suggesting inferior vena cava collapsibility index (IVCCI) may have a role.7 In part this is explained by IVCCI being a dynamic marker whereas most work with CVP was based on it used as a static marker. We suggested that a small and collapsing IVC indicates fluid tolerance – in that fluid will be unlikely to cause harm. We do not suggest the IVC or IVCCI to be used as guides to titrate fluid administration. The literature on IVC and IVCCI is limited and in some cases contradictory. We agree more work is required.
In point 3 you touch on the very important point that if fluid responsiveness is measured directly after fluid administration then measured changes do not allow for subsequent fluid redistribution and excretion. Most studies look at fluid responsiveness immediately post-delivery. A meta-analysis indeed suggests that blood pressure, pulse, stroke volume and cardiac output all decrease post this maximum with time.8 The clinical relevance remains unknown. Recurrent fluid boli will inevitably risk oedema, indeed this is again why we advocate use of ECHO to define the patients where fluid does not improve physiology and may be detrimental.
Fluid administration has been with us for nearly 200 years. Doctors have assumed benefits and fluid responsiveness in patients acutely unwell with sepsis, hypovolemia and organ failure. We are now learning of the risks of fluid administration and overload. We applaud our colleagues in intensive care medicine for their research into fluid therapy and hope to join Dr Seccombe in unravelling some of the issues that apply to patients who are not critically ill but still are managed with fluid administration. We would welcome collaboration and the development of a cross specialty fluid research group.
P. Parulekar & T. Harris