Hypoalbuminemia Associated with Diffuse Hypergammaglobulinemia in Chronic Diseases: Lack of Diagnostic Specificity

Transcription

1 Hypoalbuminemia Associated with Diffuse Hypergammaglobulinemia in Chronic Diseases: Lack of Diagnostic Specificity ALBERT A. KESHGEGIAN, M.D., PH.D. Although elevated gamma globulin is known to produce hypoalbuminemia both experimentally and in disease, a low albumin concentration in chronic liver disease often is assumed to reflect impaired liver synthetic function. Albumin and gamma globulin measurements in a series of 200 patients with a variety of chronic diseases (including cirrhosis, connective tissue disease, chronic inflammation, and malignancy) associated with diffuse hypergammaglobulinemia were combined with similar measurements from a previous study (Am J Med 1959; 29: ). The mean serum albumin concentration correlated inversely with mean gamma globulin, irrespective of disease category. Double reciprocal plot analysis showed that the relationship fits a rectangular hyperbola (r = , P < 0.001), with the mean albumin concentration approaching 2.31 g/dl at infinite gamma globulin. This suggests that serum albumin decreases to a similar extent in various chronic diseases and that hypoalbuminemia has no diagnostic implications, except to the extent that it reflects the severity of hypergammaglobulinemia. (Key words: Hypoalbuminemia; Diffuse hypergammaglobulinemia; Liver disease; Cirrhosis; Oncotic pressure; Plasma viscosity) Am J Clin Pathol 1984;81: THE CONCENTRATION of albumin often is decreased in conditions associated with diffuse hypergammaglobulinemia. This phenomenon has been shown both experimentally in animals 4 ' 4 and in human patients with a variety of diseases, including chronic liver disease.' 3 ' 8,9 ' 2 ' 8 The explanation usually advanced has been that albumin decreases to offset the increased contribution to oncotic pressure by elevated gamma globulin However, since the liver is the site of albumin synthesis, a prominent decrease in serum albumin often is considered to reflect liver disease, in the absence of malnutrition or protein-losing states. Similarly, in known chronic liver disease the albumin concentration is often used as a measure of synthetic capability and thus severity of disease. Two current textbooks, 7 " for example, state that a low serum albumin concentration is an indicator of impaired synthetic function in the absence of other causes of hypoalbuminemia; in neither is hypergammaglobulinemia included as a cause of hypoalbuminemia. Received June 3, 1983; received revised manuscript and accepted for publication October 21, Address reprint requests to Dr. Keshgegian: Department of Pathology, Lankenau Hospital, Lancaster Avenue West of City Line, Philadelphia, Pennsylvania Department of Pathology, Lankenau Hospital, Philadelphia, Pennsylvania This study correlates the serum albumin and gamma globulin concentrations in patients with a variety of chronic diseases, using data from an early report of serum protein changes in various diseases 9 and a current group of patients. In these patients, the serum albumin concentration inversely follows the gamma globulin concentration, irrespective of the disease category, in a relationship fitting a rectangular hyperbola. This shows a quantitative relationship between albumin and gamma globulin that does not seem to be affected substantially by disease category. It reinforces the concept that a depressed serum albumin concentration is not specific for liver disease and that decreased albumin in liver disease does not necessarily imply impairment of synthetic capacity. Patient Selection Materials and Methods Serum protein electrophoresis patterns were reviewed and the medical diagnoses of all patients having greater than 2 g/dl gamma globulin were obtained. The diagnoses were grouped into categories of disease known to be associated with diffuse hypergammaglobulinemia: chronic infection, connective tissue (collagen-vascular) disease, malignancy, micronodular cirrhosis/alcoholic liver disease, macronodular cirrhosis/chronic hepatitis, and other various chronic liver diseases. An overall miscellaneous category also was included. The mean albumin and gamma globulin concentrations were calculated for each category. These data were correlated with similar data extracted from the work of Ogryzlo and associates. 9 Those authors included all patients in various categories of disease. For the present study, data were extracted from those disease categories in which the mean gamma globulin concentration was increased. 477

2 478 KESHGEGIAN A.J.C.P. April 1984 Table 1. Disease Categories and Mean Albumin and Gamma Globulin Concentrations in Patients with Greater than 2 g/dl Gamma Globulin (Present Series) Mean Concentration (g/dl) Disease Normal mean Chronic inflammation Malignancy Connective tissue disease Chronic liver disease: Alcoholic/micronodular cirrhosis Chronic active hepatitis/macronodular cirrhosis Not specified Miscellaneous disease or multiple diagnoses Protein Measurements Total protein was measured in serum by an automated biuret technic. Cellulose acetate protein electrophoresis was performed on Sepraphore HI membranes (Gelman) in a Beckman Microzone system, according to standard methods. The membrane was stained with Ponceau S dye, and the protein fractions were quantitated by scanning on a Gelman ACD-18 densitometer. Ogryzlo and associates 9 measured total protein by the biuret technic and performed electrophoresis on paper, with amido-schwarz staining, and quantitation with an electronic scanning device using reflected light. Results Table 1 shows the mean albumin and gamma globulin concentrations in the current series of patients in whom Table 2. Mean Albumin and Gamma Globulin Concentrations in Disease Categories with Elevated Mean Gamma Globulin (Data from Ogryzlo and associates 9 ) Disease Normal Rheumatoid arthritis Systemic lupus erythematosus Discoid lupus erythematosus Scleroderma Dermatomyositis Polyarteritis "Collagen disease" Rheumatic fever Spondylitis Reiter's syndrome Portal cirrhosis Infectious hepatitis Sarcoidosis Lupoid hepatitis Patients Mean Concentration (g/dl) Albumin Gamma Globulin Patients _ Albumin Gamma Globulin gamma globulin clearly was increased (greater than 2 g/ dl). The patients are grouped into disease categories, and the mean of the established normal range is included for comparison. The highest mean gamma globulin levels are seen in patients with chronic liver disease and connective tissue (collagen-vascular) disease. Albumin levels are all depressed in comparison with the normal mean. Table 2 shows similar data, from patients with comparable disease categories, extracted from the work of Ogryzlo and associates. 9 Since these authors included all patients within disease categories, the mean gamma globulin concentrations are lower. However, the results are qualitatively similar, in that the mean gamma globulin concentration is elevated compared with the normal mean. Likewise, albumin levels are decreased to varying degrees. The data in Tables 1 and 2 are presented graphically in Figure 1. Interestingly, even though the two studies were performed using different methods, the normal values and trend of albumin values in disease states appear essentially superimposable. The striking feature is the correlation between the mean albumin and gamma globulin concentrations, irrespective of disease category. Minor differences do exist, some of which are statistically significant. The most extreme example is the mean albumin concentration (3.00 g/dl) in patients from the current series with connective tissue disease, compared with the mean albumin concentration (2.57 g/dl) in patients with alcoholic/micronodular cirrhosis (/> < 0.01 by the Wilcoxon rank sum test for unpaired samples), even though the gamma globulin concentration is relatively similar in the two groups (2.93 and 3.09 g/dl, respectively). However, these differences are small compared with the extent of deviation of the albumin concentration in all disease categories from the normal mean. The main factor influencing the albumin concentration appears to be the gamma globulin concentration, irrespective of disease category. To investigate the quantitative relationship between mean albumin and gamma globulin levels further, a rect-

3 Vol. 81 No. 4 HYPOALBUMINEMIA IN CHRONIC DISEASE 479 angular hyperbolic relationship was assumed to exist between albumin and gamma globulin levels, similar to that of classical enzyme kinetics, 19 except that in this case the A7 relationship is an inverse one. The formula alb = K + 7 can be applied, where alb = albumin, 7 = gamma globulin, and A and K are constants, analogous to the formula Vmax [S] 10 v = T, T^i of enzyme kinetics with a double recip- Km 4- [S] rocal form of 1/alb = 1/A + K/A (I/7). In double reciprocal plot analysis of 1/alb versus 1/7 (Fig. 2), the data points from Tables 1 and 2 fit a straight line relationship with a regression coefficient (slope ± standard error) of ± (P < 0.001, 95% confidence limits to ) and an intercept at the 1/alb axis (± standard error) of ± The correlation coefficient (r) of the line is (P < 0.001). This relationship generates the linear equation 1/alb = (1 /y), from which can be derived the constant A = 2.31 ± 0.07 (reciprocal of 0.433; 95% confidence limits ), representing the concentration of albumin at infinite gamma globulin concentration. The constant K. = can be derived from the slope (K/A), generating the equation alb =. This signifies that an inverse rectangular hyperbolic relationship between albumin and gamma globulin does exist with reasonable certainty, with albumin decreasing to a level of 2.31 g/dl at infinite gamma globulin. A rectangular hyperbola generated from the equation obtained from the double reciprocal analysis has been superimposed on the data in Figure 1 and approximates the data well. Discussion The serum albumin concentration depends on the rate of albumin synthesis, the distribution between intravascular and extravascular spaces and loss or catabolism. 15 In the absence of diseases producing altered distribution or increased loss, such as edema, nephrotic syndrome, burns, or enteropathy, the rate of albumin synthesis is the major determinant of serum albumin concentration. Many factors can influence albumin synthesis, including nutrition, hormones, environment, oncotic pressure, and the presence of disease. 15 In chronic liver disease, in particular, serum albumin levels are decreased. A common assumption, reflected in current textbooks, 7 " is that serum albumin levels are a guide to synthetic function in liver disease (and thus severity of disease) in the absence of albumin loss or redistribution. However, it has been shown clearly that albumin synthesis may be low, normal, or even high in cirrhosis and that the synthetic rate does not correlate with the serum albumin concentration CO GAMMA GLOBULIN (g/dl) FIG. 1. Mean serum albumin and gamma globulin concentrations in patients with chronic diseases. The mean concentrations in different disease categories from Tables 1 and 2 are re-expressed in graphic form. = data of Orgryzlo and associates'; = data from current series. The line represents a rectangular hyperbola generated from the equation 2.3I7 alb as described in the text. D CO CD Previous studies have documented correlations between hypoalbuminemia and diffuse hypergammaglobulinemia both in clinical disease ' 218 and in experimental stud < I /GAMMA GLOBULIN FIG. 2. Double reciprocal analysis of mean serum albumin and gamma globulin concentrations. The data points from Figure 1 are re-expressed in double-reciprocal form. The straight line shown was generated from linear regression analysis, giving the equation 1/alb = (\/y). Further details are described in the text.

4 480 KESHGEGIAN A.J.C.P. April 1984 ies The present study confirms these previous reports. Furthermore, this study compares the actual albumin level in different diseases and demonstrates that the albumin level correlates inversely with the degree of hypergammaglobulinemia irrespective of the disease process. The relationship can be expressed as a rectangular hyperbola with albumin asymptotically decreasing to a level of 2.31 g/dl at infinite gamma globulin. Thus, mean albumin concentrations are decreased essentially equally according to the gamma globulin concentration, no matter what the disease. Although some differences do exist, such as between the mean albumin levels in connective tissue disease and alcoholic/micronodular cirrhosis, these differences are minor compared with the decrease in all disease states from the normal albumin level. The differences among disease states are probably not clinically significant and could not be used to distinguish among disease entities. The current results reinforce the concept that the decreased albumin levels in chronic disease are neither restricted to liver disease nor substantially more prominent in liver disease. The results are consistent with the idea that in most cases of liver disease synthetic capacity is not a major factor in determining serum albumin levels. 17 Certainly if liver disease were far advanced to the point of liver failure, albumin synthesis would be impaired. Concomitantly, though, the synthesis of other proteins also would be impaired and the serum levels of other proteins, such as alpha globulins, also would be decreased. Although such patients do exist, they appear to be uncommon. Ascites also may have an independent effect on the serum albumin level. 1 The few patients in the current series who had ascites had serum albumin levels generally above the mean. What is the mechanism of hypoalbuminemia? While the demonstration of a correlation between albumin and gamma globulin levels does not necessarily imply the presence of a direct mechanism, a number of clinical studies have suggested that the albumin concentration decreases in order to maintain constant oncotic pressure by offsetting the contribution to oncotic pressure of elevated gamma globulin These clinicalfindingshave been supported by experimental studies in animals. In rabbits made hypergammaglobulinemic with repeated injections of pneumococcal antigens, the serum albumin level 2 and albumin synthetic rate 14 are decreased. Similarly, infusion of dextran into rabbits results in decreased albumin synthesis. 16 Intravenous administration of albumin itself does not change significantly the rate of albumin synthesis. 13 When rat livers are directly perfused, increasing oncotic pressure in the perfusate by the addition of gamma globulin causes albumin synthesis to decrease. 4 Taken together, these experiments suggest that oncotic pressure regulates albumin synthesis by the liver and that hypergammaglobulinemia, whether passive or active, can contribute to this effect. The present findings of an inverse hyperbolic relationship between albumin and gamma globulin suggest the presence of a mechanism that operates effectively at slightly elevated concentrations of gamma globulin to depress the albumin concentration, but that becomes less and less efficient at higher gamma globulin levels and approaches a lower limit of albumin concentration (2.31 g/dl). Although the slope of a rectangular hyperbola varies continuously, estimates of the slope in the region where gamma globulin is only slightly elevated may give insights into the operative mechanism. Using linear regression analysis on the data below arbitrary cutoff levels of gamma globulin ranging from 1.5 to 2.0 g/dl gives slopes ranging from to The contribution of gamma globulin (measured as g/dl) to oncotic pressure would be expected to be less than that of albumin 6 (since there are more molecules of albumin per gram due to its lower molecular weight); hence the slope of the relationship should be less than one. Thus, the observed slopes greater than one are inconsistent with an oncotic pressure hypothesis. Another physical property of proteins, to which gamma globulin would contribute more than albumin on a weight basis, is viscosity. Plasma viscosity is relatively constant in health and is characteristically slightly increased in disease when gamma globulin and/or fibrinogen are increased. 5 The ratio of the intrinsic viscosity of gamma globulin/albumin is , and the ratio of viscosity at concentrations observed in vivo is also in the same range. 6 This is consistent with the observed slope values and suggests that viscosity may be the regulatory factor for albumin concentration in the presence of diffuse hypergammaglobulinemia. Such a hypothesis also could explain the results of previous clinical observations and experiments involving both gamma globulin and dextran and is amenable to testing by a variety of methods. References 1. Bjorneboe M: Studies on the serum proteins in hepatitis. I. The relation between serum albumin and serum globulin. Acta Med Scand 1946; 123: Bjorneboe M, Schwartz M: Investigations concerning the changes in serum proteins during immunization. The cause of hypoalbuminemia with high gamma globulin values. J Exp Med 1959; 110: Brackenridge CI, Csillag ER: A quantitative electrophoretic survey of serum protein fractions in health and disease. Acta Med Scand 1962; 172(Suppl):383: Dich J, Hansen SE, Thieden HID: Effect of albumin concentration and colloid osmotic pressure on albumin synthesis in the perfused rat liver. Acta Physiol Scand 1973; 89: Harkness J: The viscosity of human blood plasma; its measurement in health and disease. Biorheology 1971; 8:

5 Vol. 81 No. 4 HYPOALBUMINEMIA IN CHRONIC DISEASE Hughes WL: Interstitial proteins: The proteins of blood plasma and lymph, The proteins, Vol. IIB. Edited by H Neurath, K Bailey. New York, Academic Press, 1954, pp 668, Jeffries GH: Diseases of the liver, Cecil textbook of medicine. Edited by PB Beeson, W McDermott, JB Wyngaarden. Philadelphia, WB Saunders, 1979, p Jencks WP, Smith ERB, Durrum EL: The clinical significance of the analysis of serum protein distribution by filter paper electrophoresis. Am J Med 1956; 21: Ogryzlo MA, Maclachlan M, Dauphinee JA, Fletcher AA: The serum proteins in health and disease. Filter paper electrophoresis. Am J Med 1959; 27: Oncley JL, Scatchard G, Brown A: Physical-chemical characteristics of certain of the proteins of normal human plasma. J Phys Chem 1947; 51: Price CP, Alberti KGMM: Biochemical assessment of liver function, Liver and biliary diseases. Edited by R Wright, KGMM Alberti, S Karran, GH Millward-Sadler. London, WB Saunders, 1979, p Post J, Patek AJ Jr: Serum proteins in cirrhosis of the liver. I. Relation to prognosis and to formation of ascites. Arch Intern Med 1942; 69: Rothschild MA, Oratz M, Evans C, Schreiber SS: Alterations in albumin metabolism after serum and albumin infusions. J Clin Invest 1964;43: Rothschild MA, Oratz M, Franklin EC, Schreiber SS: The effect of hypergammaglobulinemia on albumin metabolism in hyperimmunized rabbits studied with albumin-i 131. J Clin Invest 1962; 41: Rothschild MA, Oratz M, Schreiber SS: Albumin synthesis. N Engl J Med 1972; 286: , Rothschild MA, Orate M, Wimer E, Schreiber SS, Quesenberry R, Faison D: Studies on albumin synthesis: the effects of dextran and cortisone on albumin metabolism in rabbits studied with albumin-i 131. J Clin Invest 1961; 40: Rothschild MA, Oratz M, Zimmon D, Schreiber SS, Weiner I, VanCaneghem A: Albumin synthesis in cirrhotic subjects with ascites studied with carbonate- 14 C. J Clin Invest 1969; 48: Sunderman FW Jr, Sunderman FW: Clinical applications of the fractionation of serum proteins by paper electrophoresis. Am J Clin Pathol 1957;27: Tietz NW: Fundamentals of clinical chemistry. Philadelphia, WB Saunders, 1976, pp