The truth about fluids

Todd Fraser on 19-11-2011

I've got a theory that the way we use fluids will completely change in the next 10 years. I can't help but think that we completely over use them. I frequently hear of elderly septic patients being given 7 litres of crystalloid without a response in their blood pressure. I wonder where all that fluid has gone. It now appears fashionable to create methods of calculating whether or not the patient will be "fluid responsive", ie that they will have some improvement in their stroke volume after fluid administration. There are a number of methods out there, some of which may be quite accurate. But here we strike the issue of surrogate endpoints. Sure, we all believe that a better cardiac output is better for the patient, but as Shoemaker found a couple of decades ago, this might not be the case. FACCT found that the drier you can keep an ARDS patient the better they do, and found that contra to accepted wisdom, this did not increase risk of renal dysfunction - in fact, quite the opposite. Perhaps soggy kidneys with increased back pressure don't work so well... A number of studies appear to link overhydration with poor outcomes - a recent review of fluids in pancreatitis stands out. Of course, its hard to separate cause-effect from association. So where will we be in 10 years? Its hard to know, but first step is clearly finding a best approach to improving cardiac output. Eliminating unnecessary administration of fluids would appear to be the first step. As for how much fluid is "too much of a good thing"? Time will tell.

12 Comments

Got something to say?

Log in to your account

Forgotten your password?

Benjamin Moran wrote 12-13-2011 09:13:53 am
It seems that there are increasing limitations on measures of fluid responsiveness with the advances in these measurements. For example, SVV: not validated with spontaneous breathing (even on a ventilator), small Vt (unreliable with 6ml/kg- ARDSNet protocol, which we are all adhering to!), open chest, arrhythmias (AF). Even passive leg raising has its limitations with increased intra-abdominal pressure.

The question then becomes, "What do you do with a mechanically ventilated pt taking the odd spontaneous breath, with lung-protective ventilation (6ml/kg Vt, high PEEP), with intra-abdominal hypertension?" Or even worse-an extubated pt! I resort back to ward-based measures- Trendelenberg or postural drop. I look for a 10-20% increase in MAP (or a 10mmHg drop in postural BP) as a sign of fluid responsiveness.

Benjamin Moran wrote 11-20-2011 12:18:13 pm
I agree Todd. I try to teach the residents and junior registrars that IV fluids should be used like any drug. Know the indications (There seems to be a lot of doctors quoting "insensible losses" as the sole indication for fluid administration, which hurts my head a little). Know the contraindications (apart from CCF/renal failure, there isn't much thought about why you wouldn't give someone too much fluid- i.e. GI surgery). Know the dose (this is the Holy Grail of fluid responsiveness. We now have numerous studies demonstrating dynamic measures (PLR, SVV, PPV) being far superior to static measures (CVP, PAOP). Keeping it simple, passive leg raising or Trendelenberg show the best sensitivity/specificity and has fewer limitations to it's use). Know the complications and toxicology (As already mentioned, too much fluid is not good for you. An elevated CVP and positive fluid balance has an increased risk of mortality in ICU patients. Also, with all the starches, gelatines and albumin solutions, the pharmacodynamics extend beyond the cardiorespiratory system- coagulopathies, acid-base disturbances, anaphylactoid reactions). Then there's the type of fluid used- vehement arguments about which fluid to use still exist despite the SAFE study.

In my mind, the biggest issue is that there is a paucity of evidence to guide our 'dose' of IV fluids. The recent FEAST study demonstrated that surrogate endpoints may not be that useful (external validity issues, I know!). We know th

Jo Butler wrote 11-20-2011 08:39:19 pm
Seems fair enough.

The problem with the Passive Leg Raise (or whatever) is that you still need a well defined, well studied measure of stroke volume to decide if it was effective or not.

And as Todd said, even then you don't know that the fluids are harmful, even if the haemodynamics improve.

Todd Fraser wrote 11-21-2011 03:13:41 pm
I'm curious to know what parameters people are using for assessing fluid responsiveness. I'm used to using stroke volume variation which is available on some pulse-contour analysis setups, and this seems to work okay. Ben referred to some others.

I try to avoid ongoing fluid resuscitation unless there is evidence of end-organ underperfusion or if further major fluid losses are predicted, and there is an indicator that the patient is "fluid responsive".

What do others do?

Jo Butler wrote 11-24-2011 06:11:32 pm
One of my bug-bears is when patients are hammered with fluids post op because of a lowish urine output, despite all other indicators suggesting the patient is euvolaemic. ADH secretion is known to increase dramatically during surgery (50 fold in some cases), so the patient isn't going to make urine no matter what you do.

Urine output is a single indicator - you have to look at the whole picture.

Oliver Arkell wrote 11-24-2011 06:22:51 pm
There was an old study that showed healthy volunteers increased their SV when they had a passive leg raise - being "fluid responsive" is a natural physiological state!

Its probably worthwhile knowing what the pathology is and dealing with that. If the patient is hypotensive due to peripheral vasodilatation, tighten it up!

Alex McKenzie wrote 11-24-2011 06:33:04 pm
I guess the whole point of knowing if they are fluid responsive is that before you give a bolus (presuming you've already decided that it is important to improve stroke volume and therefore cardiac output and therefore oxygen delivery), you know that they are likely to respond. If you can confidently predict they won't, then you wouldn't give it, would you?

Personally, I like stroke volume variation as a marker of fluid responsiveness. I also use a passive leg raise to see if this gets a response before I do, particularly in patients who are at risk of a complication (bad LV etc).

Todd Fraser wrote 11-24-2011 06:45:45 pm
That's a great point Alex - the key is knowing what the risks of fluid administration are (and I suspect this is poorly appreciated, particularly by junior doctors) and making a risk-benefit analysis in your head before you give the fluids.

What methods are you all using for assessment of response to fluids? Echo? Pulse pressure variation? Change in CVP??Other?

Jo Butler wrote 11-24-2011 06:49:39 pm
For patients who are sick enough to require a pulse contour analysis, I use the SVV too.

I would definitely not use change in CVP (and PCWP change is even worse!) - there is almost no correlation between actual CVP or change in CVP, and the responsiveness to fluids.

James O'Connor wrote 11-25-2011 12:05:51 pm
How useful is a collapsing IVC on ultrasound in assessing fluid responsiveness? Patients on ventilators with high PEEPs would probably have dilated IVCs wouldn't they?

Todd Fraser wrote 11-25-2011 03:08:35 pm
Hi James,

I think you're right, there are some problems with this method.

The best results can be obtained when the patient is not spontaneously breathing. Variation of the diameter of the IVC (within 2cm of the right atrium) by >20% correlates well with "preload reserve".

However, I'm not sure the reverse is true - the absence of this variation does not rule out fluid responsiveness. PEEP, pre-existing heart disease, lung pathology, pulmonary embolus and a range of other problems may distend the IVC.

Additionally, raised intra abdominal pressures may also affect measurements.

Vikki Campbell wrote 11-29-2011 05:46:57 pm
I agree we probably have traditionally given too much fluid. I, too, have seen 2 liters given in 500 ml boluses for low UO overnight despite adequate BP.
There is increasing evidence that fliud overload is associated with increased mortality, and worse renal outcomes specifically in AKI...they don't like being waterlogged either.

Also incorporates question of fluid type, and we're probably doing too much Na loading at the same time.

My practice has changed from giving 500-1000 ml fluid boluses until the inotropes are falling, to giving 250-500 ml boluses until the inotropes plateau, and advising residents to avoid the urine chase overnight in the absence of other signs of hypoperfusion.