Why measure serum albumin levels?

The death rate among patients with end-stage renal disease (ESRD) requiring maintenance hemodialysis is significantly greater in the United States than in other industrialized countries,1 even when factored for comorbid conditions such as age, cardiovascular disease, and diabetes. Hypoalbuminemia is the most powerful predictor of death in these ESRD patients.1 Serum albumin concentration is strongly influenced by inflammation, which suppresses albumin synthesis,2 increases its fractional catabolic rate,3 and can alter partition of albumin between the vascular and extravascular spaces.4, 5 It is also strongly dependent on nutrition, especially the adequacy of dietary protein intake.6, 7, 8, 9 Loss of albumin into the urine10, 11, 12 or into either peritoneal dialysate13, 14 or hemodialysate15 is statistically related to the serum albumin concentration and directly decreases serum albumin concentration despite an increase in albumin synthesis that occurs in response.9, 11

Some risk factors associated with hypoalbuminemia, such as smoking and diabetes, are outside of provider control, although with respect to smoking the provider should offer advice. Other factors, such as type of vascular access choice and dialyzer membrane choice,16 are clearly directly controllable. Water quality also has been found to be a source of inflammation, most likely as a result of leak of dialysate across the dialyzer as well as interaction of mononuclear cells with lipopolysaccharide across the membrane. These nonnutritional factors can contribute to inflammation, and by inference to cardiovascular risk, can be addressed by changing clinical practice.

Ayus and Sheikh-Hamad17 have identified clotted asymptomatic polytetrafluoroethylene (PTFE) grafts as a source of cryptic infection. Identifying this source may require radionuclide scans; however, removal of these clotted accesses has revealed infection, and after removal serum albumin concentration increased. Clearly observing that serum albumin concentration is low in the presence of a clotted PTFE graft should at least raise the suspicion of cryptic infection and trigger a clinical response. Beyond these points, nutritional intervention clearly has a role in managing dialysis patients.

Although dietary protein can vary considerably in healthy individuals, the presence of infection or inflammation significantly alters the relationship between dietary protein and serum albumin concentration (Fig 1).

Although women with anorexia nervosa may have normal serum albumin concentration despite profound reduction in lean body mass,18 a decrease in albumin concentration in these patients is a very poor prognostic sign.19 Once infection or other trauma occurs, the need for amino acid supplementation increases.20 Nitrogen balance can be protected, or at least loss can be minimized by increasing protein intake to greater than 1.0 g/kg. Within the dialysis patient population, cross-sectional studies show a relative discontinuity in the relationship between normalized protein catabolic rate (nPCR), assuming that this represents protein intake, and serum albumin concentration (and creatinine concentration as well). That break point is at approximately 1.0 g/kg/d and also is associated with serum albumin concentration longitudinally. This would predict that augmenting dietary protein would increase albumin concentration. One problem has been compliance with regimens. Supplements are expensive, and patients may not remain on them even if they are provided. However, some patients do, and the reasons responsible for failed compliance have only recently been subjected to scientific rigor.21

Measurement of albumin as a surrogate marker for inflammation does have limitations. The effect of inflammation (C-reactive protein [CRP] concentration) is nearly undetectable when CRP values are less than 13 mg/dL. Thus, significant inflammation can occur without triggering an effect on albumin concentration. Although the insensitivity of albumin to low but clinically significant levels of CRP could be used as an argument to measure CRP concentration frequently (currently not reimbursed by Medicare), the opposite argument, that is not to measure albumin concentration because it fails to identify some of the patients at risk of inflammation, makes no sense whatever.

Clearly identifying a source of inflammation, be it an otherwise-unrecognized graft infection, a decubitus ulcer, or a dialysis-related process, and treating it effectively is reason enough to measure albumin levels routinely. What can be said about the effects of nutritional supplementation?

Clinical intervention by a trained renal dietitian has been shown in a small prospective randomized trial to improve albumin concentration in dialysis patients.22 Leon et al22 identified specific barriers to adequate nutrition in hypoalbuminemic patients, including inadequate knowledge regarding which foods were high in protein, and needs for access to adequate nutrition, such as shopping or help with cooking. Correction of these social barriers and others resulted in a significant improvement in albumin concentration. Although these may be outside the scope of direct care by a physician, they clearly affect patients and are the responsibility of a dialysis unit and care organization.

Double-blind prospective studies of the effect of nutritional supplementation in dialysis patients are lacking. Nevertheless, well-analyzed retrospective studies of the effects of intradialytic parenteral nutrition (IDPN) suggest that those patients with the lowest albumin concentrations are those on whom IDPN will have the greatest effect not only on subsequent albumin concentration, but also on survival.23 These retrospective observations are entirely consistent now with longitudinal prospective studies of the relationship between nPCR and subsequent albumin levels.24 Thus increasing nPCR, presumably either by increasing dietary intake or by IDPN, results in increased albumin levels. Whether this will ultimately lead to an increase in survival will require further study, but it would be surprising if it did not, and the onus falls on those who would take a nihilistic approach either to nutritional intervention or to diligently uncovering a cause of inflammation and removing it if possible.

Although serum albumin concentration decreases with age and is lower in women, the primary processes that cause large deviations in albumin concentration are nutrition, inflammation, and external loss, and these processes can and should be controlled. Not identifying problems by failing to look for them will not prolong survival of dialysis patients.