<< back to Articles
Low Albumin: Is It Nutritional?
Arthur F. Rosenthal, PhD
(This article appeared in the September 1999 issue of Dialysis & Transplantation.)
Since the classic demonstration by Lowrie1 that low serum albumin levels are closely correlated with mortality in dialysis patients, it has usually been assumed that albumin levels are representative of the nutritional status of the patients.
At first sight, this seems almost self-evident. Albumin is the most abundant plasma protein. Its level is known to reflect protein synthesis in the general population, so that- at an extreme-starvation would be accompanied by very low albumin levels, leading even to sufficient loss of albumin-dependent osmotic function to produce visceral edema.
Like many other assumptions whose origin lies in the direct transfer of conclusions from the general to the dialysis population, this one bears a second look. Consider several other areas of biochemical/physiological uniqueness of dialysis patients:
The extremely high rate of atherosclerotic coronary disease in the face of relatively low total cholesterol levels, and, indeed, the negative correlation of cholesterol with mortality1;
The universal presence of at least one abnormal metabolite in hemodialysis serum2 which interferes with albuminbromcresol purple binding (and which, incidentally, makes this method unreliable for albumin analysis in these patients);
The almost universal presence of hyperhomocysteinemia in dialysis patients, correlated with their very high rate of cardiac events, and the relatively infrequent finding3 of functional folic acid deficiency as a potential cause of elevated homocysteine, very unlike the case in the general population.
It is at least an anecdotal finding, testified to by many dialysis nurse-nutritionists, that at least some patients show persistently low albumin levels despite all types of dietary intervention. Thus, the question arises about what non-nutritional factors might produce a finding of low albumin levels in any given patient.
It has long been known that albumin acts as a negative acute-phase reactant-that is, its blood concentration rapidly decreases under conditions of inflammatory response. It is one of a group of proteins whose levels decrease under such conditions, contrasting with other proteins whose concentration rapidly increases under the same conditions (i.e., positive acute-phase reactants). The interesting thing about these proteins is that they all-except for one- have other clear-cut functions, and the interpretation of their levels is, therefore, ambiguous in the same way as may be true with albumin.
It was reported some time ago that levels of this one unambiguous positive acute-phase marker, C-reactive protein (CRP), are actually more closely correlated with short-term mortality (i.e., ca. a year) in dialysis patients than is albumin.4 Although some question has since been raised about this,5 there is no dispute that CRP is strongly associated with low albumin levels.
The important question of albumin as an inflammatory, rather than a nutritional, marker in many patients should thus have been raised, but seems more often to have been overlooked. CRP, which could be used to discriminate between the two interpretations of low albumin levels, should then have become a common test in dialysis patients, but it has not. Thus, the general idea that albumin is always a nutritional marker seems to have prevailed by default.
In order to develop an appreciation for the concept that albumin could be an inflammatory marker, it is important to understand the variations and characteristics of the other acute-phase marker proteins that fit into this scheme. Without invoking the newer specific complement-cascade analyte levels, it should be appreciated that a number of substances commonly analyzed for other purposes in dialysis patients are found in this category.
Negative Acute-Phase Reactants
Albumin is a relatively slow-reacting negative acute-phase reactant, since its half-life is between 2 and 3 weeks. This means that as either a nutritional or inflammatory marker, albumin would tend to provide only a slow indication of either problem, and that correction of either situation would produce only a slow return to normal levels.
Transferrrn is the main protein that carries iron in the circulation. Its concentration in serum will significantly decrease under both conditions of nutritional deficiency and inflammation. The half-life of transferrin is about 7 days, so that it will serve as a medium-quick marker for either of these conditions.
Prealbumin has the important advantage that it is a very rapid (1-3 day) indicator of onset of either a nutritional or an inflammatory event. Both prealbumin and albumin will remain depressed as long as the triggering pathological event remains uncorrected, but will become normal much more quickly when the problem is corrected.
A very intriguing observation relating to mortality is the strong positive correlation which has been found between initial prealbumin levels and long-term survival in dialysis patients.6 This correlation was found to be much stronger for prealbumin than for initial albumin, PTH, BUN, creatinine, or any other analyte.
Positive Acute-Phase
Reactants
C-reactive proteinis the ideal indicator of inflammatory events, even persistent ones, and it has no interpretation other than such events. In most normal people, it is found at almost undetectable levels, but during inflammation its level will rise hundreds or thousands of times above normal. It is not elevated under conditions of nutritional deficit unless inflammation is also present. As indicated above, CRP is rarely monitored.
Ferritin levels will rise fairly rapidly after the onset of an inflammatory event and stay elevated as long as it persists. It is obvious that the opposite effects of inflammation on transferrin and on ferritin can easily confound the evaluation of iron status, for which these tests are normally performed.
Chronic Inflammatory Response
What is the source of acute-phase events in dialysis patients? In this category, most such events can be considered as inflammatory processes, although inflammation may not be the major, obvious characteristic of some of them. For instance, myocardial infarction, disseminated cancers,
Rheumatoid arthritis, systemic lupus, and surgery have signs and symptoms of their own for which acute-phase markers are generally not used diagnostically. Infection is an important acute-phase event in dialysis patients, but unless it is occult it is also generally diagnosed by other tests.
This seems to leave, as a conceptual residue, some less easily recognized inflammatory process which may show up in elevated CRP levels; low albumin, prealbumin, and transferrin levels; and, most importantly, can put the patient at an increased mortality risk.
The hallmark of this theoretical state is that it is a general, probably chronic, occult inflammatory process unrecognized as a separate condition because its signs (other than CRP) and symptoms blend in with those generally ascribed to the debilitating effects of chronic renal failure and/or other related factors, such as an iron or nutritional deficiency. The possibility must also be seriously considered that it may also be a consequence (primary or contributory) of the dialysis treatment itself.
The specific etiology of this putative condition should next be considered. Some interesting possibilities have been put forward, among which are the following:
It has been suggested that the bioincompatibility of dialysis membranes leads to an activation of the complement cascade, producing a generalized, often subclinical, but chronic inflammatory response. Proponents of this view claim that there is a clear difference in the relative risk of mortality between patients treated with bioincampatibte (essentially, unmodified cellulose) and biocompatible (modified cellulose or totally synthetic) membranes,7,8 but this effect was not seen in other studies?9,10
Although proponents have also claimed very sharp differences in mortality between these membrane types in acute renal failure,11,12 this has also been strongly disputed.13 Obviously, the contribution of membrane incompatibility to chronic inflammation in end-stage renal disease patients awaits further clarifying study before it can be definitively assessed.
It has been proposed that dialysis patients are frequently exposed to very low doses of bacterial endotoxin through their dialysis fluid water, which could produce a chronic subclinical inflammatory response.14,15 This remains highly speculative; given the extreme sensitivity of modern tests for endotoxin in water and dialysis fluid and the rel ative infrequency of finding any detectable endotoxin in these liquids, the factor must be listed as unproven.
A largely neglected possible factor likely to produce occult chronic inflammatory disease in many dialysis patients is chronic acidosis of renal failure.
It is known that CO2 levels in dialysis patients are generally lower than in the normal population. Since CO2 is only a second-order means of estimating blood pH in these patients, a corollary inference is that for many dialysis patients, the dialysis treatments do not provide a complete correction of their metabolic acidosis. A consequence of acidosis is protein catabolism and cellular destruction via activation of lysosomal degradative enzymes. In fact, the original data of Lowrie1 show a strong association of low CO2 levels with mortality risk.
Urumic Acidosis
Although there appears to have been little or no study of these processes in dialysis patients as a mechanism of activation of inflammatory events in response to chronic acidosis, it seems on theoretical grounds to be a distinct possibility as an important contributory factor. In fact, attention to acidosis as a uremic toxin in general has been advocated,16 and it has been proposed as a significant activator of tissue destruction.17
It may be that if this factor is ultimately found to play an important part in producing chronic inflammation- and, thus, is a major mortality risk- the hallmark of adequate dialysis could shift toward the correction of uremic acidosis rather than just the removal of urea or other small molecules.
References:
1. Lowrie EG, Lew NL. Dcath risk in hemodialysis patients: The predictive value of commonly measured variables and an evaluation of death rate differences between facilities. Am J Kidney Dis 1990; 15:458-482.
2. Mabuchi H, Nakasashi H. Underestimation of serum albumin by the bromcresol purple method and a major endogenous ligand in uremia. Clin Chim Acta 1987; 167:89-96.
3. Rosenthal AF, Ginsburg MJ, Crawford JR. Homocysteine and heart disease in dialysis patients. Dial Transpl 1998; 27(10):627-629.
4. Bergstrom J, Heimburger O, Lindholm B, Qurcshi AR. Elevated C-reactive protein as a strong predictor of increased mortality and low serum albumin in hemodialysis (HD) patients. J Am Soc Nephrol 1995; 6:2218/573.
5. Owen WF, Lowrie EG. C-reactive protein as an outcome predictor for maintenance hemodialysis patients. Kidney Int 1998; 54:627-636.
6. Avram MM, Sreedhara R, Mittman N. Long-term survival in end-stage renal disease. Dial Transpl 1998; 27(1):11-22.
7. Bloembergen WE, Hakim RM, Stannard DC, Held PJ, Wolfe RA, Agadoa LY, Port FK. Relationship of dialysis membrane and cause-specific mortality. Am J Kidney Dis 1999; 33:1-10.
8. Hakim RM. Influence of dialysis membrane on outcome of Nationwide patients. Am J Kidney Dis 1998; 32:S71-S75.
9. Bonomini V, Coli L, Scolari MP, Stefoni S. Structure of dialysis membranes and long-term clinical outcome. Am J Nephrol 1995; 15:455-462.
10. Bonornini V, CoIi L, Nanni Costa A. Scolari MP. Long-term comparative evaluation of synthetic and cellulosic membranes in dialysis. Int J Artif Organs 1994; 17:392-398.
11. Schiffl H, Lang SM, Haidcr M. Bioincompatibility of dialyzer membranes may have a negative impact on outcome of acute renal failure, independent of the dose of dialysis delivered: A retrospective multicenter analysis. Am Soc Artif Intern Organs J 1998; 44:M418-M422.
12. Himmelfarb J, Tolkoff Rubin N, Chandran P. Parker RA, Wingard RL, Hakim R. A multicenter comparison of dialysis membranes in the treatment of acute renal failure requiring dialysis. JAm Soc Nephrol 1998; 9:257-266.
13. Kurtal H, von Herrath D, Schaefer K. Is the choice of membrane important for patients with acute renal failure requiring dialysis? Artjf Organs 1995; 19:391-394.
14. Schindler R, Lonnemann G, Schaffer J, Shaldon S, Koch KM, Krautzig S. The effect of ultra-filtered dia]ysate on the cellular content of interleukin-1 receptor in patients on chronic hemodialysis. Nephron 1994; 68:229-233.
15. Weber C, Stummvoll HK, Passon S, Falkenhagen D. Monocyte activation and humoral immune response to endotoxins in patients receiving on-line hemodiafiltration therapy. Int J Artif Organs 1998; 21:335-340.
16. Feriani M, Ronco C, Fabris A, La Greca G. In: Replacement of Rena! Function by Dialysis (4th Ed.), Jacobs C. Kjellstrand CM, Koch KM, Winchester JF (eds.). Boston: Kluwer Acad. Pub., 1996, pp 1014-1015.
17. Mitch WE. Robert H. Herman Memorial Award in Clinical Nutrition Lecture, 1997. Mechanisms causing loss of lean body mass in kidney disease. Ant J Clin Nutr 1998; 67:359-366.
<< back to Articles