Guest Editorial: What Do the Serum Proteins Tell Us About Our Elderly Patients?

Decision Editor: John E. Morley, MB, BCh ALTHOUGH serum albumin, transferrin, prealbumin, and insulin-like growth factor 1 (IGF-1) are commonly utilized as indicators of nutritional status or disease severity, considerable caution must be exercised when the results of these serum protein measurements are interpreted. With advancing age, it becomes increasingly difficult to differentiate the interrelated effects of natural aging, disease, and nutritive state on the physiologic processes that determine the serum concentrations of these proteins. Consequently, random determinations may have marginal clinical value. However, when used to supplement other data, these measurements often provide valuable clues as to the patient's nutritional status and prognosis and can help in guiding treatment decisions, including the need for nutritional support. The interpretation of these measures is aided by knowledge of what physiologic and pathologic factors influence their serum concentration. Albumin is one of the most abundant and commonly measured of the aforementioned serum proteins. Its concentration in the serum is determined by the balance among production by the liver, distribution within the extracellular fluid compartment, and degradation (1)(2). The primary regulator of albumin production is thought to be serum oncotic pressure. A normal adult on an adequate diet produces approximately 200 mg of albumin per kilogram of body weight per day. The synthetic rate of albumin can exceed 200% of normal during periods of accelerated protein loss (e.g., nephrotic syndrome) and in response to refeeding after brief periods of starvation (2)(3). In health, the total body exchangeable albumin pool is approximately 200 g. Less than 40% of this total is within the intravascular compartment (1)(2). The concentration of albumin is normally much greater within the intravascular space compared with that in the extravascular space (35–50 g/l compared with approximately 10 g/l) (1). However, the intravascular and extravascular pools are in equilibrium, and there is a complete exchange of the intravascular pool every 24 hours. Endothelial cells probably regulate albumin distribution and may also be the site of albumin degradation, which normally occurs at a rate of 3% a day (4). The mechanisms controlling these processes have not been elucidated. When these processes are in equilibrium, albumin has an 18-day half-life. There is little or no decline in the serum albumin concentration with age (5)(6). Outcome studies conducted within communities, hospitals, and long-term care institutions reveal a strong inverse correlation between the serum albumin concentration and the risk of subsequent morbidity and mortality (7)(8)(9). These findings indicate that albumin is an important prognostic indicator that can be utilized in population-based studies for case-mix adjustment. The use of albumin to assess an individual patient's prognosis is more problematic. A random albumin sample, interpreted without regard to clinical context, has low sensitivity and specificity and only limited clinical utility as a prognostic indicator. That is because not all conditions that affect the albumin concentration are associated with adverse consequences. For example, utilizing different laboratory techniques for measuring the albumin concentration can result in a difference of up to 11 g/l in the reported value (10). The actual albumin concentration may drop by up to 5 g/l in a previously ambulatory person after 8 hours of bed rest (11). Fluid shifts may account for this change. During periods of acute physiologic stress (such as major surgery or sepsis), the serum concentration can decline by 15–20 g/l within a few days (12). This effect on the albumin concentration is probably mediated by cytokines (such as tumor necrosis factor and interleukin-6), which are believed to increase vascular permeability to albumin. There is a resultant rapid loss of the normal concentration gradient between the intravascular and extravascular space and an apparent sequestration of albumin in extravascular sites (2)(13)(14). These same cytokines also suppress albumin synthesis and may trigger an increased rate of albumin degradation (2). The prognostic significance of the nadir value and the time it takes the albumin concentration to return to normal following a major stress have not been adequately studied. Prolonged hypoalbuminemia can develop in association with advanced liver disease (cirrhosis), severe congestive heart failure, nephrotic syndrome, and protein-losing enteropathies (2). When these conditions are not present, a persistently low serum albumin probably represents ongoing inflammation and/or nutritional deprivation, and it is associated with a high risk of adverse outcomes. Even though albumin synthesis decreases by 30–50% after only 24–48 hours of protein and energy deprivation, there may be little change in the serum albumin concentration even after a much more prolonged period of fasting in an otherwise healthy individual (15)(16). During periods of inadequate nutrient intake, a decreased rate of albumin degradation and mobilization of albumin from the extravascular space may contribute to the maintenance of a normal serum albumin concentration (4). For these reasons, albumin is not a very sensitive screening test for early stages of nutritional deterioration. Even a low albumin concentration has only moderate specificity as an indicator of protein-energy nutritional status. In addition to nutritional deprivation, chronic inflammation and other disease conditions can contribute to the development and maintenance of a low serum albumin. As it is not currently possible to differentiate the effects of inflammation from those of nutritional deprivation (17), a low albumin rate indicates only that the patient is at risk for being undernourished. Unexplained hypoalbuminemia should trigger a more in-depth evaluation of the patient to identify other evidence of protein-energy undernutrition and to determine whether nutrient intake is adequate. Transferrin is another commonly measured serum protein. It is used most frequently as part of the laboratory evaluation of body iron stores and occasionally as a risk index or nutritional status indicator. In health, transferrin functions primarily to transport iron within the blood, and its production by the liver is closely linked to total body iron stores (18). Besides iron, the serum concentration is also influenced by nutrient intake and various disease states (19). Iron deficiency, acute hepatitis, and estrogen all result in higher serum transferrin levels. Advanced liver disease (cirrhosis), nephrotic syndrome, cancer, chronic infections, and nutritional deprivation often produce a sustained drop in the serum concentration. As is the case with albumin, transferrin is a negative acute-phase reactant, meaning the serum concentration can drop precipitously in response to sepsis, major surgery, and other forms of acute severe physiologic stress (20). The magnitude of any change in transferrin has prognostic significance. Numerous hospital and nursing-home-based studies indicate that transferrin is an important correlate of morbidity and mortality (21)(22). Because its half-life of 8 days is considerably shorter than that of albumin, serum transferrin is considered to be a potentially more sensitive indicator of early changes in nutritional status or illness activity. Offsetting this advantage, the opposing effect of different pathologic conditions on the serum transferrin concentration reduces its sensitivity as a health-risk indicator. Its specificity as a nutritional marker is no greater than that of albumin. Prealbumin (also known as transthyretin) is a transport protein for thyroxine (23). Indirectly, it also plays a role in vitamin A transportation. Prealbumin is a symmetrical tetramer composed of four identical subunits. At physiologic pH, each prealbumin subunit contains one binding site for retinol-binding protein, which is a carrier protein for vitamin A. In health, prealbumin has a half-life of 2 days and a much smaller volume of distribution compared with that of albumin. There is little or no decline in prealbumin as a function of age (24). Like albumin and transferrin, prealbumin and the prealbumin:retinol-binding protein complex are negative acute-phase reactants (25). In response to systemic inflammation, liver production declines and the serum concentration drops rapidly. Low levels are also found in association with end-stage liver disease, iron deficiency, and nutrient deprivation (26)(27). Renal failure (which results in a decreased catabolism of both prealbumin and retinal-binding protein by the kidney) and high-dose steroid therapy are associated with elevated prealbumin concentrations (28). Because of its relative short half-life, prealbumin is more sensitive to changes in nutrient intake and disease activity than are albumin or transferrin. The serum concentration begins to drop after 3–5 days of very low nutrient intake and can decline by as much as 50% subsequent to a major physiologic insult (15). In the latter case, the nadir value is usually reached within 3–5 days and corresponds to the period of maximal negative nitrogen balance (23). With resolution of the inflammatory process, the serum concentration will climb rapidly to the normal range if nutrient intake is adequate. A rising prealbumin correlates with positive nitrogen balance. This fact is used to advantage to assess the adequacy of nutrition support. Failure of the serum concentration to increase by at least 20 mg/l in 1 week is considered an indication of inadequate nutrient intake or ongoing inflammation and should prompt a careful assessment of the patient and the nutrient regimen being used (15). Possibly because it is so sensitive to changes in disease activity and nutrient intake, prealbumin does not correlate as strongly as albumin with mortality. The change in the prealbumin concentration may be more important. Among cancer patients, prealbumin responsiveness to nutrition support has been shown to be a strong predictor of survival (29). Although a relatively nonspecific indicator of health and nutritional status, prealbumin can be used effectively to monitor for change in a patient's clinical status and to detect subclinical nutritional deficits. Insulin-like growth factor 1 (IGF-1) is a low-molecular-weight polypeptide hormone produced and released predominantly by the liver in response to growth hormone stimulation (30). IGF-1 circulates bound to carrier proteins, which play a role in regulating its function and serum concentration (31). The IGF-1 fraction bound to the major carrier protein has a half-life of approximately 15 hours (32). There is little diurnal fluctuation in the plasma level of IGF-1, which is directly related to growth hormone secretion (30). Paralleling the age-related diminution in pulsatile growth hormone release (33), the plasma concentration of IGF-1 declines by 35–60% between the fourth and ninth decades (34). As reported in this issue of this Journal by Dr. Arai and colleagues, centenarians have uniformly low IFG-1 levels compared with those of young adults (35). The finding from this study of a correlation between IGF-1 and cognitive function among centenarians suggests the possibility that the age-associated decline in the growth hormone-IGF-1 axis may have detrimental effects on the central nervous system. Because confounding by illness cannot be excluded, further study of this issue is needed. In frail elderly patients, many factors other than the age-related decline in growth hormone secretion may influence plasma IGF-1 levels. Production of IGF-1 is modulated by other hormones, liver disease, sepsis, severe trauma or illness, and changes in nutritional status (36)(37)(38). In normal volunteers who fasted for 5 days, the plasma IGF-1 concentration dropped by 60–70% and promptly returned to baseline with refeeding (39). Under these circumstances, the fluctuation in IGF-1 correlates with changes in nitrogen (N) balance. During prolonged nutritional deprivation or ongoing inflammation, the plasma IGF-1 concentration remains depressed, correlating with serum albumin, transferrin, and cholesterol; triceps skinfold thickness; and the total lymphocyte count (37)(39). Unless ongoing sepsis or another critical condition persists (36), the plasma concentration of IGF-1 gradually returns to the patient's preillness baseline when adequate nutrient intake is again established (39). The rise in plasma concentration correlates with protein and energy consumption (39), maintenance of positive N balance (39), and improvement in the patient's overall health status. There are currently no age-adjusted norms for IGF-1. For this reason, the change in IGF-1 over time has more significance in elderly patients than the absolute value of any one measurement. Among hospitalized elderly patients, IGF-1 is a strong predictor of serious complications and death (37). Like prealbumin, IGF-1 is a nonspecific but sensitive indicator of changing health status and adequacy of nutrient intake. It can be used as an alternative to prealbumin to monitor the disease activity and nutritional status of patients, particularly those who have renal failure or other conditions that alter the sensitivity of prealbumin (40). Although all of the aforementioned serum proteins have less than optimal sensitivity and specificity as markers of either disease activity or nutritional status, they all can be of value as prognostic indicators and in helping to guide treatment decisions when they are interpreted as part of a patient's overall clinical assessment. Incorporating these parameters into multivariable nutrition assessment or illness severity instruments may provide greater sensitivity and specificity and improve their clinical value. Although much work in this area has already been done, there is clearly a need for more research to determine how serum protein measurements can be utilized most effectively to improve patient care.