Tom Gallacher:
I am interested in peoples opinions on the concept of keeping patients in danger of developing acute lung injury "dry" i.e fluid restriction. Since the primary problem is capillary leak and this is due to an alteration in the filtration co-efficient and not to any change in capillary pressure the knock on effects of relative hypovolaemia on other organ systems I believe is detrimental. However, we have all encountered colleagues with patients in ICU who look at the fluid balance chart and note a large positive balance and the signs of pulmonary oedema on the chest x-ray and blame the fluid.
If we do not have a large positive balance and still get pulmonary oedema with a normal wedge pressure how do they explain that? Simple - "it's ARDS". My own analogy is that of two buckets containing identical fluids with identical size holes. The pressure above the fluid in both buckets is kept constant (i.e. we will not tolerate a mean arterial pressure below say 50mmHg) and neither bucket can be allowed to empty (since gross hypovolaemia is fatal) then the fluid leak from both buckets must be identical. This is the model in it's simplest form and I am interested to see what peoples response is as the analogy of the buckets can be further developed.
Rolando Berger:
T.Gallacher's theoretical points about fluid-balance and permeability pulmonary edema (ALI/ARDS) are well taken. However, animal and clinical data based on lung-water measurements do show that total lung water content increases with increased hydrostatic pressure... which is presumably related to intravascular volume if the LV properties and function are preserved. The assumption then becomes that more water (i.e. edema) is bad for you. No prospective randomized study has ever been done to absolutely prove this point... nor is it likely to ever be done: Can you imagine proposing to intentionally fluid overload a patient with ALI/ARDS?
However, the next best thing is to correlate outcome with fluid balance in patients with comparable degrees of ALI/ARDS. Prospectively, although no patient would be intentionally overloaded, some could be "dried" quite vigorously to see if "drier is better", even within usually accepted ranges of PWP/CVP. Alternatively, ARDS/ALI series could be analyzed to see if fluid balance/PWP values correlated with outcome: length of mechanical ventilation, total ICU stay, and even mortality. These data, of course, would not be conclusive and definitive but can give a reasonable idea of the potential validity of the hypothesis that "drier is better". This is a summary of some of this type of data:
None of these studies offers perfectly convincing evidence that should put this issue to rest.... but, in my view, the weight of the available data is enough to suggest that, within reason, trying to run a "dry ship" in ARDS is worth the effort. In my view, the burden of proof has now shifted to those who think it is not. Of course, I would not go to absurd extremes of massive diuresis or ultrafiltration, but trying to maintain a reasonable I:O balance, and a ":lowish" hydrostatic capillary pressure in the lungs, seems easy enough to do in most ARDS cases.
Malvolm Fisher:
Rolando Berger outlines the compelling arguements for keeping the lungs dry which is my personal bias. On the other hand Loren Nelson has data on using SuperPeep and deluging the body with crystalloid. I would add the brilliant old paper by Worthley and Fisher in the brilliant old and disappeared disease of fat embolim syndrome gave a convincing case for forgetting fat and embolism and treating pulmonary oedema with rapid resolution and avoiding ventilation at all. It is as Dastardly Dick suggests inextricably linked to the crystalloid v colloid arguement.
One of the repeatable observations we have made in patients with acute severe membrane pulmonary oedema(i.e. envenomation,anaphylaxis,aspiration) is that there is a stage when capilliaries leak crystaloid and not colloid.. PEEP stops the frothing and improves A-a gradient which implies the leak can be controlled. In very rare circumstances such as the rare severe unexplained pulmonary oedema following bypass we have seen the phenomenon described in the literature where Plasma leaks through the lung surface and requires reopening to prevent tamponade. The fluid shifts are complex but ARDS is pulmonary oedema and that can be reduced by lowering filling pressures at least in the early stages. My bias is to keep dry (and never need superpeep) based on the studies Rolando cites.
Dick Burrows:
I think that this is an argument somewhat akin to crystalloid/colloid.
I also use the leaky bucket analogy but I also consider the bucket to be highly plastic and active in the sense that at that level (the microvasculature) there are forces acting in a manner which we seldom consider - the bucket is actually always leaky but the leakiness is controlled by the filtration and reflection coefficients (which are not constants) refs - Simonescue & Simonescue, Hills BA, Michel CC, Cotran & Renke etc etc Presumably too nitric oxide and calcitonin gene related peptide (and lots of others) affect membrane function at that level and will effect changes in membrane permeability.
In any event we have a simple protocol which is to bring the CVP to approx 10-15 cmsH2O and hold it there. Then look at the amount of fluid you have to give in order to maintain a constant CVP (which is, I admit, a crude index of blood volume and not meant to be an absolute) A normal fluid requirement (post op for example) would be in the region of 3mls/kg/hr (using crystalloid). There is trouble (leakiness) when it is in the region of 8-12mls/kg/hr (previous deficits accounted for) and there is deep deep trouble at 15mls/kg/hr and above. This gives an idea of the leakiness.
I also take an idea of membrane function from the urine output (among others) which often settles on 1-2mls/kg/hr (no lasix) Creatinine & electrolytes seem to remain normal with this unless the kidneys have really taken a hiding. Oftentimes, reversal of the situation is heralded by a brisk diuresis. Also I'm presupposing that there is no pre-existing organ disease which will complicate the picture further.
We force the fluids believing that not to force them allows the development of hypovolaemia which perpetuates a vicious cycle which eventually will be lethal. Incidentally I've seen cases of frank pulmonary oedema settle by forcing fluids - esp near drowning. I've also had it put to me that this improvement is due to treatment of a concommitant cardiac failure which is only detectable by echo - fair enough.
Trouble is that If the leakiness is persistent, perhaps because the underlying problem has not been identified and treated in the first place or because the condition of leakiness has itself rapidly become self-sustaining and pathological in it's own right and thus unbreakable by any treatment.
It is my feeling that this is the sort of patient who will not survive whatever you do - dry/wet crystalloid/colloid, this/that. The situation is compounded by the likelihood of a re-perfusion injury which will also cause havoc with capillary permeability. But then you run that risk anyway - whether you use crystalloid or colloid or drive 'em with inotropes.
So you pays your money and takes your pick. Running them wet can kill or cure them as can running them dry. I don't know when the trial is going to be designed to sort this one out - especially when we do not really understand the physiology let alone the pathophysiology.
It's not really nonsense - we just behave nonsensically.
Tom Gallacher:
I do not disagree with what you say but the other interesting arguement is that if we could maintain an oncotic pressure difference between plasma and interstitial fluid we could remove some of the leaked fluid/prevent fluid from leaking. The crystalloid/colloid controversy was looked at using albumin, or Haemacell or Gelofusine. It is well known that albumin can pass freely into the lung and take with it 18 mls of water/g albumin in a chelated complex. I am not sure if the synthetic colloids leak or not but I would be surprised if they did not. However, in our arguments for colloids we concentrate on size which determines the length of time they spend in the intravascular compartment, however if you try to measure the oncotic pressure of synthetic colloids you will find that it is unmeasurable - i.e. it has no oncotic pressure.
Surprisingly, 4.5% human albumin solution has an oncotic pressure of 13 which is roughly half that of plasma, and so, in giving this we should not be surprised that water/sodium leaks into the lung down it's concentration gradient.In addition the albumin is small enough to pass through the pores of injured endothelium. However, the starches on the other hand are too big to pass through the endothelial pores even in "capillary leak syndrome" and may partially occlude these pores thus preventing albumin from passing through. Their oncotic pressure is about 60 and thus they have the effect of pulling water out of the lung. This to me seems like a more rational approach to the problem of dealing with rather than treating "capillary leak" since no-one despite 20 years of searching has found a convincing way of reversing the leak though there have been many false gods!
The problem which remains is that the manufacturers of the starches will not produce starches of a significantly narrow molecular weight spectrum, which is about 200 - 300 kDa. This is the optimum size to balance uptake by reticulo-endothelial cells which may be detrimental and maximising oncotic effect. Anyone interested in persuading the manufacturers to make the recipe and participating in a multi-centre trial?
Dick Burrows:
To me this is the nub of the argument. Leakiness to colloids (or anything else) is a function of the membrane rather than the particular colloidal substance. Pappenheimer in 1954 showed that the size of an albumin molecule (not weight) is half that of a so called pore. Therefore it should be free to equilibrate, in equal concentrations, across the membrane. that there is a difference is a function of membrane rather than albumin per se. How therefore do we stop the leakiness - I got no idea. I do think that a sort of shotgun therapy of using large molecules seems to be somewhat self defeating as it is vital for cellular function to allow selective movement of substances both ways across the membrane and stopping the leaks in this way sounds like the Berlin wall to me - stops everything but does not address the cause.
I doubt the size of the 'holes' matters so much. I would put my money more on the metabolic functions of the membrne being able to maintain gradient forces of a quantum nature at that level but I don't think anybody has any real idea how this metabolic function works let alone if it exists but it makes more sence to me that at that level quantum forces are coming into play.
if ARDS is a sole expression of a multiple system disease which affects only the lungs then OK to aim for the dry solution (scuse the pun). But all too often it seems to be part of a more extensive disease process involving many organs where not only is there increased permeability but also decreased function alongside the permeability change.. The fact that we are using analogies from different situations seems to underscore our ignorance about capillary permeability. How often have you read journals where the author talks about x or y causing an increase in capillary permeability with no reference whatsoever to the nature or mechanism of the increased permeability.
Graham Baldock:
As a big user of starch I'd be very interested.
Another point which sometimes gets overlooked is the amount of sodium that patients get whether you use crystalloid or colloid. It is easy to forget that all colloids (with the exception of 20% albumin) contain the equivalent of normal saline, ~150mmol/l of sodium (Haemaccel slightly less). The positive sodium balances achieved by some patients run into hundreds (occasionally thousands) of mmol so it is hardly surprising that they become oedematous. I have been asked by a representative of one of the starch producers whether a salt free solution might be useful. On the face of it it sounds a promising idea but the idea has not been tested to my knowledge. Any opinions, thoughts etc.?
Ian Tan: I think the issue has not been resolved. For those who titrate fluids to hemodynamics, the sicker the patient, the more fluid the patient needs. The correlation between positive fluid balance & poor outcome is obvious. None of the retrospective data can be believed. Underresuscitation can lead to further multiple organ damage and has also been found to worsen the ARDS/pulmonary edema. (Herndon DN, et al. Surgery 1986;100:248-251) (that was a sheep though).