Albumin and oedema
IV Albumin in Critically ill Patients David Brewster 11.09.98
Albumin and oedema André Briend 23.12.98
1999 Steve Collins  
Re: Albumin and oedema Kerin Gould 05.01.99
IV Albumin and protein intake Michael Golden 06.01.99
RE: Albumin and oedema Steve Collins 15.01.99
high protein and loss of appetite References Michael Golden 15.01.99
RE: Albumin and oedema Buford Nichols 16.01.99
RE: Albumin and oedema Steve Collins 18.01.99

From: David Brewster at health_rdh

Date: 11/9/98 3:15PM

Subject: IV Albumin in Critically ill Patients



I sent these comments out when the network was down, so I do not

believe they were received. I still wonder whether they have some

relevance to the debate over high protein diets in severely ill

patients (adults and children with kwashiorkor).

I would like to bring to the attention of the Network the important

Cochrane systematic review on human albumin in critically ill

patients. This review included 24 randomised controlled trials of

intravenous human albumin in critical ill patients with hypovolaemia,

burns or hypoalbuminaemia in which deaths occurred. Overall, they

found that the relative risk of death increased 1.7 fold with the use

of IV albumin solutions. I have summarised the findings of the

systematic review below.

The authors of this review suggest that the pathophysiological

mechanisms underlying the increased risk of death with IV albumin

infusion are related to circulatory overload, leakage of albumin into

extravascular spaces from increased capillary permeability, impairment

of sodium and water excretion which worsened renal failure and impaired

anticoagulant activity with increased blood loss. I would like to ask

Michael and Barbara Golden, as well as others on the Network, whether

they feel that this finding could be related to the well-known adverse

effects of high protein intake in severe malnutrition, particularly

kwashiorkor. Is it possible that intravenous albumin in critically ill

patients results in toxic metabolites as a result of metabolic blocks

in critically ill patients (and severely malnourished patients)? Is

there a parallel with high protein diets in these patients?








* Serum albumin concentration is inversely related to the risk of

death (24-56% increase in risk per 2.5g/L decrement in serum albumin)

* Low serum albumin may be due to leakage, increased metabolism or

insufficient synthesis

* Human albumin is used for the emergency treatment of shock, burns

and hypoproteinaemia

* Human albumin costs 30 times more than crystalloid solutions (e.g.

Ringer's lactate)



* Systematic review: 30 RCTs

* Comparison:

albumin to crystalloid solution or no administration

* Subjects:

1419 critically ill patients with hypovolaemia, burns or


* Outcome: mortality


Statistical Methods:

* Mantel-Haenszel method for relative risks

* intention to treat

* regression approach to assessing funnel plot asymmetry (publication

or selection bias)

* identified 32 RCTs by computerised and hand searches, contacting

authors and companies

* only 24 trials had deaths

* only 13 trials had adequate group allocation (by pharmacy or




Condition n Relative Risk of Death

Hypovolaemia 13 1.46 (0.97-2.22)

Burns 3 2.40 (1.11-5.19)

Hypoalbuminaemia 8 1.69 (1.07-2.67)

Pooled (total) 24 1.68 (1.26-2.23)

Good studies 13 1.61 (1.09-2.38)


Pooled increased risk of death with albumin = 6% (3-9%),

Number Needed to Treat = 17

Date: Wed, 23 Dec 1998 11:07:03 +0100 (CET)

From: (Andre' BRIEND)

Subject: Albumin and oedema


Dear NOG'nuts,

Sorry for giving a break in the first flurry of new year E mails.

Thanks David for flagging this interesting review on IV albumin and survival. David, can you kindly send us the reference of this meta analysis ? To me, the quesiton of safe protein intake during oedematous malnutrition is not so clear. The 'proof' that very low protein intake are beneficial at the beginning of tt of PEM would be a trial showing any favourable income on a group of patients on a low protein diet compared to another group receiving a similar diet, differing only by a higher protein content. These data are lacking. Most studies quoted to say that very low protein intakes are beneficial usually are not really experimental and compare groups of children receiving diets also differing in Na content or osmolality among other things (see the AJCN editorial commenting Steve's paper). Mike's Lancet paper (Lancet 1982; i: 1261-5) on protein intake and odema shows absolutely no correlation between protein intake and rate of oedema disappearance.

This issue may look a little bit academic, since the best way to lower Na and osmolarity in milk based diet... is to lower their protein content anyway. Still, if Na is an issue (which I think is not ruled out) we should be careful too not to over use ORS (whether WHO or resomal).

David's approach to look around in other types of conditiions is worth pursuing. Does any one know of malnutrition (or oedema) associated conditions where very low protein diets (0.7 g/kg/day) make a difference compared to 1.5 to 2 g/kg/day ? The series of paper on IV albumin is interesting, but has no straightforward relevance to protein intake PER OS, which is quite different.

Recently, I came across a paper on nephrotic syndrome oedema which just discussed the issue of protein intake :

Palmer BF. Nephrotic Edema - pathogenesis and treatment. Am J Med Sc 1993; 306: 53-67.

Please find below a scan of the most relevant part of this paper; Intgerestingly, it claims that low proteins diets help because of their inhibiting effect on the renin angiotensin system, a mechanism quite different from the usually supposed toxic action of amino acid metabolims (I anticipate - and look forward to - a 12 page minimum litterature review on this topic from Mike when he is back). This renin angiotensin issue draws back to the Na. Anyway, other examples of beneficial low protein diets most welcome +++.

Merry Christmas and Happy new year to all.




Palmer BF. Nephrotic Edema - pathogenesis and treatment. Am J

Med Sc 1993; 306: 53-67.


Dietary Protein.


In the past, patients with the nephrotic syndrome were routinely prescribed high-protein diets in an attempt to prevent protein malnutrition resulting from urinary-protein losses. This strategy has not been effective in increasing serum protein levels, primarily because the resultant enhancement in proteinuria negates the beneficial effect of increased albumin synthesis.

In contrast, low-protein diets have been shown in experimental and clinical studies to reduce the magnitude of proteinuria in nephrotic states (88,89).

The institution of a low-protein diet can lead to favorable effects on the size-selective properties of the glomerular basement membrane, as reflected by a decrease in the fractional clearance of large-sized dextrans.

The beneficial effect of a low-protein diet is not related to changes in the glomerular filtration rate and the filtered load of protein. Rather, the clinical effect is similar to that observed with angiotensin-converting enzyme inhibitors, suggesting that these two maneuvers may share a common mechanism. Low-protein diets are associated with decreases in pluma-renin activity and in expression of the renin gene (90,91). Therefore, a decrease in angiotensin II levels may underlie the antiproteinuric effects of both low-protein diets arid converting-enzyme inhibitors. In support of this possibility, it has been shown that in rats with subtotal nephrectomy, acute blockade of the renin-angiotensin system reduces proteinuria only in animals ingesting a high-protein diet (92,93). Similarly, in rats with immunecomplex nephropathy, the administration of an angiotensin-converting enzyme inhibitor reduces proteinuria only in the setting of high-protein intake and has no effect in animals fed a low-protein diet (94). Finally, preliminary clinical findings in patients with a transplant suffering from chronic rejection have shown a greater antiproteinuric effect with angiotensin-converting enzyme inhibitors in the setting of high-protein intake (95). Based on these findings,-a reasonable dietary-protein intake for patients with the nephrotic syndrome is 0.8 g/kg/day. The protein should be of high biologic value. The adequacy of nutrition should be carefully monitored by following parameters such as serum albumin, transferrin, and skin-fold thickness.

Dr. Andre' Briend , Groupe Nutrition Santé, CNAM/INSERM

5 rue du Vertbois, 75003 Paris, France

tel 33-1-53 01 80 36

Date: Tue, 05 Jan 1999 18:06:36 +0000

From: kerin gould <>

Subject: Re: Albumin and oedema


Dear NGO'nuts:

I have a question in this discussion of high and low protein diets: has anyone compared the difference between animal and vegetable sources of protein of the same amounts? Is it possible that some of the benefits or disadvantages come partly from what goes with the protein - for example, fats and carbohydrates, different amino acids and enzymes, toxins, etc?


Date: Wed, 06 Jan 1999 15:36:05 +0000

From: Michael Golden <>

Subject: IV Albumin and protein intake


The issue of the amount of dietary protein and of intravenous treatment are

separate: both are important in the management of oedematous malnutrition.


In terms of IV albumin (and blood transfusion, and salt sensitivity) the

timing of the treatment, with respect to the cardiovascular system, seems

to be critical. Brian Wharton in his classical paper on heart failure

[Wharton BA, Howells GR, McCance RA. 1967 Cardiac failure in kwashiorkor.

Lancet 2:384-387. Wharton BA, Balmer SE, Somers K, Templeton AC. 1969 The

myocardium in kwashiorkor. Quarterly Journal of Medicine 38:107-116] makes

the point that it is OK to have a high sodium intake during the first 2/3

days after admission but thereafter it will induce heart failure. This is

my experience also - but it also applies to blood transfusion and albumin

administration! For example, when I went to Haiti to advise on treatment

there were two children whom the local staff had kept alive without showing

any progress - I advised judicious blood transfusion (10ml/kg slowly), they

were both dead in 12 hours. I sat with them during this time and they died

from heart failure despite giving the usual treatment. Subsequently, back

in Jamaica, I did immediate post-mortem transthoracic cardiac puncture to

see how much blood I could retrieve from the hearts of 12 children - the

volumes recovered were over 20% of calculated total blood volume! (60 to

120 ml in children 4-7kg body weight) - the conclusion was that many of the

children we were treating were dying of undiagnosed heart failure, at least

as a terminal event. Then in several children who were in extremis, I

tried venesection of about 20ml of blood acutely; this led to rapid

clinical improvement (I made no measurements at all of cardiovascular

function so it is entirely clinical observation). When I wrote the WHO

manual I initially advocated exchange transfusion only and put in a section

about venesection if the child was in extremis - this advice was taken out

on the grounds that it would be dangerous in less experienced hands and

could well lead to staff or parental accusation of doing harm. But the

problem remains of how to deal with these individuals. With modern imaging

methods and sophisticated cardiovascular investigation this whole subject

should be reopened for serious investigation in centres where such

procedures are possible.

The question of timing is important. When the patient initially presents

there is often haemoconcentration and relative hypovolaemia. Then with

treatment the defects in the cell membranes are corrected and Na is

exported from the cells and K imported. There is also correction of the

defects of the intercellular ground-substance so that fluid shifts from the

interstitial to the intravascular compartment - a fall in haematocrit at

this time is almost universal. In a substantial proportion of children the

renal defect is not repaired at the same rate as the general somatic

defects so that the fluid in the vascular compartment cannot be excreted.

Such a differential rate of recovery of the organs leads to a disequlibrium

syndrome resulting in volume overload heart failure. The heart failure is

usually misdiagnosed as respiratory tract infection and treated with

antibiotics! The falling haematocrit leads to a diagnosis of severe

anaemia and the patient may be transfused because of this! At this

critical stage anything which will increase the circulatory overload will

lead to increased mortality, be it albumin transfusion, blood transfusion,

a high salt diet or even force feeding the standard diet. The same

phenomena occur in intensive care patients. The patient who is initially

anaemic will be at a much greater risk of heart failure, particularly when

the metabolic rate rises and tissue oxygen demand increases dramatically in

early recovery.


The nature of the renal abnormality does not seem at all clear.

The table shows the Fractional sodium excretion (Klahr S & Alleyne GAO,

Kidney International 1973;3:129-141)

  malnourished recovered
control 0.50 1.23
volume expansion 0.82 11.07

It is quite clear that the kidney of the malnourished child cannot respond

to volume expansion with a naturesis.


The renin levels are very high as are the aldosterone levels - but a word

of caution in interpretation - during the naturesis that occurs later

during treatment the levels of aldosterone rise even further before

returning to normal levels later in recovery. I do not understand this at all.

[Lurie AO, Jackson WPU. 1962 Aldosteronuria and the edema of kwashiorkor.

American Journal of Clinical Nutrition 11 edn. :115-126. Leonard PJ,

MacWilliam KM. 1965 The binding of aldosterone in the serum in

kwashiorkor. American Journal of Clinical Nutrition 16 edn. :360-362.

Godard CM, Munoz M, Sanchez MA, Vallotton MB, Riondel A. 1986 A study of

the renin-angiotensin-aldosterone system in severe infantile malnutrition.

International Journal of Pediatric Nephrology 7 edn. :39-44. Van Der

Westhuysen JM, Kanengoni E, Jones JJ, Van Niekerk CH. 1975 Plasma renin

activity in oedematous and marasmic children with protein energy

malnutrition. South African Medical Journal 49 edn. :1729-1731].


To my knowlege the Atrial Natretic Peptides have not been measured, neither

have any of the hormonal receptors or ADH. And follow up renal

physiological measurements have not been made since the 1970s. We need

this information in order to design the correct diets and timing of their

introduction for these patients. Several of the receptors use sulphydryl

binding [s-s bonding] so that the level of SH group oxidation that we have

shown to be affected elsewhere, may be important. The other finding we

have made is that nitrate excretion (measure of NO production) is very high

(hence the confusion of toxic-shock with dehydration0 BUT it does not seem

to diminish during the first phase of treatment. There are some exciting

results with the use of intravenous reductants (methylene blue) in animal

models - I think the data are becoming sufficiently convincing for there to

be a clinical trial.


There is a particular problem in the patient who is anaemic. Anaemia is

associated with a much higher mortality and a higher prevalence of heart

failure. In my opinion, if such a patient is to be transfused then this

should be done in the first day or so after admission, before the

volume/electrolyte disequlibrium occur - if possible by exchange

transfusion, exactly as one would do for a neonate, and slowly. Later if

it is necessary to give blood then this should be by exchange transfusion,

all dietary intake should be stopped for the duration of, and for 12-24

hours after the transfusion, and other measures to treat heart-failure

should be given prophylactically.


I do not see any reason to give an albumin infusion, and I believe it to be

dangerous - the Chochrane review strengthens this view.


Another corollary of the dynamics of change during early recovery is that

ORS or ReSoMal may be reasonable when the child first presents, but use of

these fluids from days 3-14 may be much more dangerous. Such a situation

can arise if there is some osmotic diarrhoea due to sudden increases in

dietary intake above 100 kcal/kg/d which is then treated with "rehydration

solutions" and continuing high dietary intakes - instead of by reducing the

amount of the diet that is given (putting the child back to phase one



Also very rapid rates of recovery (metabolic change), such as we are

achieving with modern diet therapy, may make patients more vulnerable to

errors of judgement or management decisions with regard to treatment of

anaemia or mild osmotic diarrhoea.


Prof. Michael H.N.Golden

Dept of Medicine and Therapeutics, Univ of Aberdeen, Foresterhill

AB9 2ZD. Scotland, (UK)

INTERNET, Tel +44 (1224) 681 818 ext 52793/53014, Tel(direct) +44 (1224) 663 123 527 93, Fax +44 (1224) 699 884

Fri, 15 Jan 1999 12:43:30 GMT

From: "steve collins" <>

Subject: RE: Albumin and oedema



I don’t think that there is necessarily a contradiction between Mikes findings: that energy, not protein, content was related to resolution of oedema in children, and ours: that the protein content of diets in adults was related to resolution of oedema (and mortality). I think that differences in appetite of adults on our Higher Protein and Lower Protein diets readily explain this apparent anomaly. Although our study didn’t collect any data on actual intakes, my qualitative impressions was that on the higher protein diets appetites were very poor in most oedematous patients who therefore ate little. As I remember it the reduction in intake covered all elements of the diet. Within a day or so after the introduction of the LP diet almost all the oedematous patients reported an increase in appetite and the development of diarrhoea. The development of diarrhoea after the switch to the lower protein diet, in patients who had been in the centre receiving the HP diet for a month or so, supports the impression that their food intake had increased. This diarrhoea appeared in almost all oedematous patients who switched to the LP diet, it was watery and resolved within a few days. I reckoned that it was a refeeding diarrhoea most probably a result of poor colonic mucosal function. This indicated a failure of regeneration over the past month with the HP diet. As I understand it this regeneration occurs when the colon is exposed to food and a failure would indicate that such exposure hadn’t taken place i.e they weren’t eating.

The adults that we were treating would not tolerate NG tubes and thus if they were anorexic they ate little or nothing. This is different to Mikes children, where, I presume, NG tubes were used in all anorexic cases.

So maybe the the suppression of appetite by the HP diet rather than (or together with) any direct effect of protein should be investigated? I have seen an abstract of a review (haven’t got the paper) that increasing plasma levels of certain amino acids such as Histidine will suppress appetite.

Maybe other NGOnuts know of other evidence of protein / amino-acids suppressing appetite? Interestingly this might to an extent tie appetite suppressing effect of higher protein diets with the angio-tensin - renin hypothesis for the causation of oedema, as both are concerned with histamine, (the synthesis of which is rate limited by brain [histidine]).



Date: Fri, 15 Jan 1999 14:53:46 +0000

From: Michael Golden <>

Subject: high protein and loss of appetite References:


Steve Collin's observations fit in very will with the observations of experimental animals with zinc deficiency. These animals develop a cyclical eating pattern with two days or so of normal intake followed by two days of food aversion. This eating behavior is abolished if the dietary protein level is reduced. If such an animal is force-fed protein it dies. Much the same type of pattern is seen when diets with very unbalanced amino-acid patterns are given. (Alf Harper did most of the amino acid work - Mark Hegsted did a lovely experiment on monkeys)

I interpret these data to suggest that IF the protein synthetic machinery is not functioning properly (lacking an essential component for making a protein, such as an essential amino-acid or zinc) then the amino acids absorbed from digested protein cannot be disposed of correctly - that is by being incorporated into the body protein or metabolised. They will then be toxic, just as in a person with an inborn error of amino acid metabolism, and cause loss of appetite. The metabolic machinery has to be repaired before a diet "high" in anything is given - and an unbalanced diet should never be given.

I later expanded and generalised these thoughts to develop the type I and II classification of nutrients. So the concept is that if any of the type II nutrients is missing or grossly unbalanced then the others cannot be correctly used and will accumulate in the body if they cannot be disposed of rapidly enough. If any of these are toxic then the response will be a poor appetite, and great danger if the patient is force fed the diet. What I think happened with Steve's patients is that when the protein was reduced the levels came into line with the other components of the diet, such as zinc, phosphorus, magnesium etc and so the diet could be used. Imagine if you were building a house instead of a tissue, and the delivery firm kept sending truck loads of bricks but no cement - you would soon not be able to work because of the mountains of bricks burying you whilst you could not even put one layer of them where they are meant to be!

Poor appetite in most situations is due to one of the following: 1) Type II nutrient imbalance, 2) inflammatory response mediators (infection, endotoxicosis etc) or 3) liver disease (cf infectious hepatitis). I think that these are probably not separate underlying "causes" but aetiologies for a similar underlying pathogenesis at the level of disordered intermediary hepatic metabolism - which fits in with other anoretic states such as giving anti-cancer drugs, acid-base disturbance etc. Many of the "weight-loss" diets prescribed for obesity rely on being unbalanced in type II nutrients to suppress appetite, many are high protein and several of these have been associated with mortality!

Nevertheless, the malnourished individual usually has elements of each of the three common proximate causes.

In my Lancet paper comparing protein and energy intakes with the rate of loss of oedema - do not interpret association between energy intake and rate of loss of oedema as being a causal relationship (this part of the discussion was removed from the submitted paper by the editors at the time). It is really only a proxy for the AMOUNT of the diets that the children took - so that some other, unmeasured, component could have accounted for the result - it should probably be interpreted to show that those with the best appetites lost oedema the fasted!

I would like to hear Vernon Young's comments upon unbalanced amino-acid diets (and other type II nutrients) and appetite.

Best wishes,



Prof. Michael H.N.Golden

Date: Sat, 16 Jan 1999 15:16:46 +0000

From: "Buford Nichols, M.D." <>

Subject: RE: Albumin and oedema



A low protein diet in your circumstances is probably a high carbohydrate

diet. I would like to know which carbohydrates were in the diet of your

adult subjects who developed diarrhea when they were switched to the LP

diets. I assume that this was not a lactose containing diet but composed

of sucrose or starch energy sources. Can you fill me in on these details.

We have a submitted manuscript which proves by molecular biological methods

that villus atrophy causes a 60% loss of small intestinal sucrase and

maltase activities in malnourished Brazilian children. The sucrase and

glucose transported messages were increased in individual enterocytes. This

is a sufficient reduction of activity to cause proximal CHO malabsorption

but the diarrheal response varied with the individual subject's ability to

salvage this energy via colonic mechanisms.

Studies by C. Lifschitz and F. Carrazza, from the same population which

developed diarrhea after feeding CHO, had a defect in the ability to absorb

acetate from the colon. These subjects with diarrhea also failed to produce

hydrogen in response to a CHO load.

>From this I argue that two problems have occurred in the subject with

diarrhea after a carbohydrate load: 1. Small intestinal sucrose, starch

(and in some cases lactose) malabsorption. 2. Colonic salvage failure, and

in your adults adaptation and recovery of salvage. Breath H2 tests would

be informative in the adult subjects because colonic function probably

recovered long before small intestinal mucosal digestion.


Buford Nichols

From: "steve collins" <>

Subject: RE: Albumin and oedema

Date: Mon, 18 Jan 1999 12:24:35 -0000




Thanks for your comments. The LPP diet's enery was from lactose, sucrose and fat (I enclose the diets spread sheets as a seperate indivual mail so as not to block e-mails). I agree that there was also a failure of absorption in the SI caused by villi atrophy there to cause proximal malabsorption. I was interested in your reference to the studies of C. Lifschitz and F.

Carrazza; does this mean that in malnourished there is both a change in the flora reducing the capacity to break down CHO and decrease in the the colons ability to absorb the short chain fatty acids produced?