Anion Gap and Immunoglobulin Concentration

Anion Gap and Immunoglobulin Concentration ALBERT A. KESHGEGIAN, M.D., PH.D. Keshgegian, Albert A.: Anion gap and immunoglobulin concentration. Am J Clin Pathol 74: 22-26,190. The serum anion gap is often decreased in patients who have multiple myeloma or monoclonal gammopathy, presumably owing to the unmeasured contribution of cations by the paraprotein. In 3 patients who had moderate-to-severe diffuse (polyclonal) elevations in immunoglobulin concentration (>3 g/dl), the mean serum anion gap was significantly lower than in patients who had normal concentrations of immunoglobulin. In the subgroup with immunoglobulin levels greater than 4 g/dl, the mean anion gap was.7 meq/l, compared with 11.9 for the control group (). Known causes of decreased anion gap, such as hyperkalemia or hypercalcemia, were absent. The albumin concentration had a minimal effect on the anion gap. The mean anion gap was independent of the patient's diagnosis and the relative contribution of each immunoglobulin class to the total immunoglobulin. Thus, patients who have diffusely elevated immunoglobulin concentrations, as well as those who have monoclonal elevations, have a significantly lower mean anion gap. This low anion gap should be considered in the evaluation of the acid-base status of these patients. (Key words: Anion gap; Immunoglobulin; Polyclonal gammopathy.) THE ANION GAP is defined as the difference between the serum concentration of sodium ions and the sum of chloride and bicarbonate ions. It represents the net effect of unmeasured cations and anions in the serum. The diagnostic utility of an increased anion gap in the evaluation of acid-base disorders is well known. 310 The significance of a decreased anion gap has also recently become apparent. Causes of a decreased gap include hypoalbuminemia, hypercalcemia, hyperkalemia, hypermagnesemia, hyperviscosity, and laboratory error. An interesting association is that between monoclonal immunoglobulin gammopathies and a decreased anion gap. Murray and associates 9 reported decreased anion gaps to occur in patients who had multiple myeloma, and attributed the decrease to the contribution of unmeasured cations by the myeloma protein. The decrease has been reported to be more prominent in those patients having IgG myeloma than in those having IgA myeloma, presumably because IgG molecules tend to be more cationic than IgA molecules. 2 A decreased anion gap has also been described to occur in association with asymptomatic monoclonal gam- Received October 2, 1979; accepted for publication November 13, 1979. Address reprint requests to Dr. Keshgegian: Department of Pathology, Hospital of the University of Pennsylvania, 3400 Spruce Street, Philadelphia, Pennsylvania 19104. William Pepper Laboratory, Hospital of the University of Pennsylvania and Division of Laboratory Medicine, Department of Pathology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania mopathy, 11 with no difference between the effects of IgG and IgA monoclonal proteins. In contrast to these studies dealing with monoclonal gammopathy, several authors have failed to find a correlation between anion gap and polyclonal gammopathy, or diffuse hypergammaglobulinemia.' M - In a preliminary study 7 1 presented evidence that moderateto-severe diffuse hypergammaglobulinemia is associated with a decrease in the anion gap, apparently independent of the albumin concentration. This report examines in more detail the relationship between the concentration of individual immunoglobulin classes and the anion gap. Methods The serum protein electrophoresis studies of all patients at the Hospital of the University of Pennsylvania performed over a two-year period were reviewed, and all patterns showing diffusely increased gamma globulin with concentrations of 3 g/dl or greater were selected. A number of patterns showing mild-tomoderate diffuse hypergammaglobulinemia, hypogammaglobulinemia, and normal gamma globulin concentrations were also randomly selected for review. In most cases, multiple determinations of electrolytes had been performed, including the specimen used for protein electrophoresis. The following conditions excluded a case from further review: abnormal potassium or calcium concentrations, blood urea nitrogen concentration greater than 50 mg/dl (normal, -1), or hyperviscosity. Serum electrolytes were determined on a Technicon SMA 6/60. Total protein was measured by the biuret method on a Centrifichem analyzer. The albumin concentration was determined by densitometric scanning of the protein electrophoretic pattern on cellulose acetate (Beckman Microzone System) stained with ponceau S dye. IgG, IgA, and IgM concentrations were measured by radial immunodiffusion* on the same specimen that was used for protein electrophoresis. * Meloy Laboratories Inc., Springfield, Virginia. 0002-9173/0/0900/022 $00.75 American Society of Clinical Pathologists 22

Vol. 74 No. 3 ANION GAP AND IMMUNOGLOBULIN CONCENTRATION 23 Results The relationship between the anion gap and the total serum immunoglobulin concentration is shown in Figure 1. At low, normal, or slightly elevated immunoglobulin concentrations, the ranges of the anion gap appear similar and independent of immunoglobulin concentration. However, at concentrations of immunoglobulin above 3 g/dl, the anion gap is decreased. The decrease is most prominent in sera containing 5 g/dl or more of immunoglobulin. These data are summarized in Table 1, in which the sera are divided into seven groups on the basis of immunoglobulin concentration, and the mean anion gap is given for each group. The group with 1.00 to 1.99 g/dl of immunoglobulin is used as a reference group representing normal immunoglobulin levels. The mean anion gap for this group (11.9 meq/1) agrees well with normal values previously reported 310 and with the reference group in the preliminary study. 7 In the groups with low immunoglobulin concentration (0-0.99 g/dl) and mildly elevated immunoglobulin concentration (2.00-2.49 and 2.50-2.99 g/dl) the mean anion gaps are slightly higher and lower, respectively, than the mean in the reference group, but the differences are not Table 1. Relation between Mean Anion Gap and Total Immunoglobulin Concentration Immunoglobulin (g/dl) 0-0.99 1.00-1.99 2.00-2.49 2.50-2.99 3.00-3.49 3.50-3.99 4.00+ Number of Patients 23 47 26 35 17 2 3 Mean Anion Gap (meq/1 ± SEM) 12.2 ± 0.5 11.9 ± 0.3 10.9 ± 0.5 11.3 ± 0.4 10.4 ± 0.3 9.4 ± 0.5.7 ± 0.4 Significance* P > 0.05 P = 0.05 P > 0.05 P < 0.01 * y-value was determined by comparison with the group having 1.00-1.99 g/dl immunoglobulin, using Student's /-test. Table 2. Relation of Anion Gap and Protein Parameters in Patients Having Immunoglobulin Levels Greater Than 2 g/dl Protein Total immunoglobulin IgG IgA IgM Total protein Albumin Alb/(alb + Ig)t r* -0.39-0.36-0.19 0.001-0.13 0.24 0.20 Significance 0.02 < P < 0.05 0.9 <P < 1.0 0.1 < P < 0.2 P = 0.01 0.01 < P < 0.02 * The value ofr (correlation coefficient) was determined using standard linear-regression analysis. t Represents the ratio of the concentration of albumin and the sums of the concentrations of albumin and immunoglobulin. 20 S 16 _E nion Gap «< x. Vt» (T 4 1 1 ' 1 > 1. 1, I 2 3 4 5 6 7 Immunoglobulin (g/dl) FIG. 1. Relation between anion gap and immunoglobulin concentration. Values were determined as described in Methods. statistically significant. However, the mean anion gaps of the groups with 3 g/dl or more of immunoglobulin are significantly lower than the mean gap of the reference group, and there is an inverse correlation between the concentration of total immunoglobulin and the mean anion gap (e.g., the mean anion gap in the group with more than 4 g/dl of immunoglobulin is lower than the mean gap in the group with 3-3.49 g/dl, P < 0.002 by Student's /-test). To investigate this relationship further, an inverse linear relationship was assumed to exist between total immunoglobulin concentration and anion gap in all patients having immunoglobulin concentrations greater than the reference group (2 g/dl or more). Table 2 shows that the correlation coefficient for such an association (-0.39) is highly significant. Most of this correlation is contributed by IgG (-0.36), but there is also a correlation with IgA concentration (-0.19). There is almost no correlation between the anion gap and IgM concentration. The anion gap has been reported to be dependent upon albumin and total protein concentration. 3 It is possible that the apparent correlation of immunoglobulin concentration and anion gap is not direct, but rather reflects a decrease in the albumin concentration or an increase in total protein concentration. Table 2 also shows the correlation coefficients for the relation of anion gap and several protein parameters: albumin, total protein, and the ratio albumin/(albumin + immunoglobulin), which reflects the relative contribution of albumin to total protein compared with immunoglobulin. There is no significant correlation of the anion gap with total protein concentration. The correlations of the anion gap with albumin and the ratio albumin/ (albumin + immunoglobulin) are statistically signifi-

24 KESHGEGIAN A.J.C.P. September 190 Table 3. Relation of Anion Gap and Protein Parameters in Patients Having Immunoglobulin Levels Greater Than 4 g/dl Protein IgG/(IgA + IgM)t IgG/IgA IgG - (IgA + IgM) Total protein Albumin Alb/Igt Alb/(alb + Ig) r* -0.06 0.12-0.16-0.10 0.11 0.17 0.20 Significance 0.7 <P < 0. 0.4 <P < 0.5 0.3 < P < 0.4 0.5 <P <0.6 0.5 <P <0.6 0.3 < P < 0.4 0.2 < P < 0.3 * The value of r (correlation coefficient) was determined using standard linear-regression analysis. t Represents the ratio of the concentration of IgG and the sum of the concentrations of IgA and IgM. t Represents the ratio of the concentration of albumin and the total concentration of immunoglobulin. Represents the ratio of the concentration of albumin and the sum of the concentrations of albumin and immunoglobulin. cant, but weaker than the correlation between anion gap and immunoglobulin concentration. These results suggest that IgA, as well as IgG, contributes to the decrease in anion gap, and that the decreased gap at high immunoglobulin concentrations is not due solely to decreased albumin. However, the group of patients having immunoglobulin levels greater than 2 g/dl is quite heterogeneous; accordingly, the subset of patients having immunoglobulin levels greater than 4 g/dl was studied separately. Within the group of patients having immunoglobulin levels greater than 4 g/dl, there is variation in the value of the anion gap (Fig. 1). If this variation were due to differences in the relative concentrations of the immunoglobulin classes or differences in the concentrations of albumin or total protein, then there would be a correlation between the anion gap and immunoglobulin ratios or nonimmunoglobulin protein concentrations. Table 3 shows several relationships among immunoglobulin classes and the degree of correlation of these parameters with the anion gap. None of them are significant. Similarly, within this subset of patients having very high immunoglobulin concentrations, the correlations between albumin, albumin/(albumin + immunoglobulin), albumin/immunoglobulin, or total protein and the anion gap are similar to or smaller than those in the group having immunoglobulin levels greater than 2.0 g/dl. None of the correlations are statistically significant in the subset of patients having immunoglobulin levels of 4 g/dl or more. To investigate further the relationship between the anion gap and high immunoglobulin concentrations, the medical charts of all patients in those subgroups that had significantly low mean anion gaps (more than 3 g/dl of immunoglobulin) were reviewed, and the diagnosis was recorded. For patients with more than one diagnosis, that condition known to cause diffuse hypergammaglobulinemia was recorded (e.g., alcoholic liver disease, rather than ischemic heart disease). The diagnoses were divided into five categories: alcoholic liver disease/micronodular cirrhosis, chronic active hepatitis/macronodular cirrhosis, collagen-vascular disease, chronic infection or inflammation, and malignancy. In five patients, two of the above categories of disease were present (footnote to Table 4). In addition, eight patients had liver diseases that were not included in the categories listed above. For four pateints there was no diagnosis or the chart was unobtainable. In each of the five diagnostic groups the mean total immunoglobulin concentration and mean anion gap were determined, as were the ratios of IgG/IgA and IgG/IgM. The highest mean total immunoglobulin concentrations were found in macronodular cirrhosis (or chronic active hepatitis) and collagen-vascular diseases. These values were higher than in the group with malignancy (P < 0.05 by the Wilcoxon sum of ranks test). The macronodular cirrhosis group also had a higher mean immunoglobulin concentration than the group with infection (P = 0.05). The IgG/IgA ratio in the alcoholic liver disease group is significantly lower than the ratio in the other four categories (P < 0.05 for malignancy, P < 0.01 for other groups). This reflects the known high IgA levels occurring in alcoholic cirrhosis. 6 The IgG/IgM ratio in the alcoholic liver disease group is also lower than the ratio in the collagen-vascular disease and infection groups (P < 0.05). With regard to the mean anion gaps in each group, the only significant difference is between the groups with macronodular cirrhosis and infection (P < 0.05 by both the Wilcoxon sum of ranks test and Student's Mest). This correlation is not strong, and is probably explained by the higher mean total immunoglobulin concentration in the group with macronodular cirrhosis. These results show again that the relative contribution of immunoglobulin class does not affect the mean anion gap. They also show that the diagnostic category of the patient does not affect the gap many types of disease known to produce elevations in immunoglobulin concentration act similarly in lowering the mean anion gap. Discussion Although the contribution of proteins to the ionic milieu of serum has been recognized for many years, 13 the effects of globulins have only recently been emphasized. Murray and colleagues 9 reported that increased levels of monoclonal immunoglobulin occur-

Vol. 74 No. 3 ANION GAP AND IMMUNOGLOBULIN CONCENTRATION Table 4. Diagnoses of Patients in Groups with Low Mean Anion Gap 25 Number Total Immunoglobulin (g/dl) IgG/IgA IgG/IgM Anion Gap (meq/1) Alcoholic liver disease or micronodular cirrhosis Chronic active hepatitis and/or macronodular cirrhosist Collagen-vascular disease Chronic infection or inflammation Malignancy Miscellaneousli Liver disease, other Two diagnosestt No diagnosis or chart not available 15 19 9 7 5 4 4.03 ± 0.5* 5.22 ± 1.7 4.55 ± 0.95 3.97 ± 0.72 3.73 ± 0.40 4.0 ± 0.95 3.4 ± 0.7 4.29 ± 0.4 3.6 ± 1.2 11.7 ± 10.2 15.4 ± 11.5 7.1 ± 6.1 7.5 ± 4.3 5.6 ± 3.6 5. ± 3.0 11.3 ±.2 9. ± 5.2 15. ± 14.7* 1.2 ± 11.3 1.1 ± 12.0 17.6 ± 9.9 2.2 ± 26.2** 14.1 ± 7.7 24.2 ± 16.6 9.2 ± 2.7.0 ± 1.7 9.5 ± 2. 9. ± 1. 9.2 ± 2.6 9.3 ± 1.4 9.4 ± 2. 9.9 ± 1.1 * All numerical values represent mean ± SD. t Includes two patients who also had hepatocellular carcinoma. t Excludes one patient having a ratio of 219. Systemic lupus erythematosus, four; rheumatoid arthritis, polymyositis, scleroderma and mixed connective tissue disease, one each. ring in multiple myeloma are associated with a decrease in the anion gap, and that the decrease is proportional to the concentration of the monoclonal protein. They hypothesized that the decrease in the gap was due to the net cationic charge of monoclonal immunoglobulin at physiological ph, resulting in unmeasured cations that reduce the apparent anion gap. Consistent with that hypothesis, they showed that several monoclonal proteins had isoelectric points above physiologic ph, and that sera containing monoclonal proteins bound more anion in vitro than normal serum. These observations were extended by DeTroyer and co-workers, 2 who found that the gap was decreased in IgG, but not IgA, multiple myeloma. They hypothesized that IgG is cationic, whereas IgA is slightly anionic at physiologic ph, thereby accounting for the difference in anion gap between the two groups of myeloma patients. Determinations of the isoelectric points of several IgG and IgA monoclonal proteins were consistent with this hypothesis. Schnur and associates 11 showed that asymptomatic monoclonal gammopathy is also associated with a decreased anion gap. They failed, however, to find a correlation with either the class of immunoglobulin or the concentration of monoclonal protein. Another approach to the study of the effect of immunoglobulins on the anion gap is to examine the effect of polyclonal gammopathy. Several studies have failed to find a correlation between polyclonal gammopathy and anion gap, 34 and there has been speculation that monoclonal proteins may have some special effect on the anion gap that polyclonal immunoglobulins lack. 3 However, as early as 1936, Gutman and colleagues 5 reported a decrease in the "acid-base equivalence" of blood in disease associated with diffuse hyperglobulinemia, as well as multiple myeloma. H Sarcoidosis, two; sickle cell anemia, organic brain syndrome, aortic aneurysm, stroke, idiopathic pulmonary fibrosis, one each. ** Excludes one patient having a ratio of 161. tt Chronic infection and neoplasia, three; chronic infection and alcoholic liver disease, two. In a preliminary study, 7 I determined that diffuse polyclonal hypergammaglobulinemia is indeed associated with a decrease in the anion gap. The inability of other studies to demonstrate such an increase may be due to quantitative, rather than qualitative, differences; the gamma globulin concentration, which was not stated in those studies, may not have been high enough to affect the anion gap. This study confirms and extends the preliminary findings. Moderate-to-severe increases in immunoglobulin concentration are associated with a significantly low anion gap. The association becomes statistically significant at immunoglobulin levels of 3 g/dl and is dependent on the concentration of immunoglobulin. At levels of 4 g/dl or higher, the mean anion gap is.7 meq/1, which compares favorably with the mean gap of 9.2 meq/1 that Murray and colleagues 9 found for patients who had multiple myeloma. High levels of immunoglobulin are often associated with high concentrations of total protein and low concentrations of albumin. The association of anion gap and immunoglobulin level, however, is independent of total protein. There may be a small association between albumin concentration and anion gap in patients who have high immunoglobulin levels, but this correlation is weaker than the correlation between the immunoglobulin concentration itself and the anion gap. With regard to the components of total immunoglobulin concentrations, the correlation of anion gap is strongest with the IgG concentration (Table 2). There is also a significant correlation between IgA arid anion gap. Whether this is direct, or whether it reflects the inclusion in this study of a high proportion of patients who had liver disease, in whom IgA elevations accompany IgG elevations, cannot be determined. At high concentrations of total immunoglobulin, the

26 KESHGEGIAN A.J.C.P. September 190 decrease in anion gap is independent of the immunoglobulin classes elevated (Table 3). Since IgG was elevated in all cases, no conclusion can be drawn about the anion gap in cases of high IgA or IgM with normal IgG. However, the results do show that the relative contributions of IgA and IgM to the total immunoglobulin concentration in the presence of elevated IgG have no effect on the anion gap. Thus, the effect of high immunoglobulin concentration on the anion gap is not affected by the immunoglobulin classes involved. These results are consistent with the findings of Schnur and co-workers 11 for IgG and IgA monoclonal gammopathy, and differ from the conclusions of DeTroyer and associates 2 for multiple myeloma. Since the latter authors did not state the concentration of IgA monoclonal proteins studied, the lack of an effect of IgA monoclonal proteins may be related tq a lower concentration of these proteins compared with IgG monoclonal proteins. The mechanisms producing a decrease in anion gap in patients who have high immunoglobulin concentrations (monoclonal or polyclonal) remain to be completely elucidated. The effect of serum proteins on the anion gap depends upon the total concentration of protein, the composition of total protein, and theph. The usual net contribution of total serum protein to the ion pool is as unmeasured anions. Thus, in hyperproteinemia or hypoproteinemia the contribution of anions would increase or decrease, respectively, causing a respective increase or decrease in the anion gap.,,3,s A change in ph would render the proteins more or less anionic, and thus would alter the anion contribution of serum protein. 1, These mechanisms do not appear to be pertinent within the context of this study, since there were presumably noph changes and since the gap was not significantly correlated with the total protein concentration (Tables 2 and 3). Variation in the composition of protein in the presence of constant ph and total protein concentration can also affect the net charge contribution of protein. Van Slyke and colleagues 13 showed that globulin binds much less base (i.e., is less anionic) than albumin. Van Leeuween 12 also determined that at ph 7.4 albumin has a large "net cation equivalency" (i.e., is very anionic), whereas gamma globulin has zero to slightly negative "net cation equivalency" (i.e., gamma globulin itself is slightly cationic). These findings are in accord with more modern determinations of isoelectric points. The fact that IgA as well as IgG can affect the anion gap suggests that the immunoglobulin need not be.cationic, as previous studies 2,9 have suggested. Instead, a larger contribution of only slightly charged protein, either anionic or cationic, to the total protein can decrease the net negative charge or total protein, and thereby decrease the anion gap. The clinical significances of low anion gap in the presence of a high immunoglobulin concentration, whether monoclonal or polyclonal, are similar. The presence of a low gap may be the initial indication of increased immunoglobulin. A normal anion gap, on the other hand, does not exclude the presence of high immunoglobulin levels. Finally, in a patient known to have a high concentration of immunoglobulin, an apparently normal anion gap may mask the presence of high anion gap metabolic acidosis. References 1. Adrogue HJ, Brensilver J, Madias NE: Changes in the plasma anion gap during chronic metabolic acid-base disturbances. Am J Physiol 235:F291-F297, 197 2. DeTroyer A, Stolarczyk A, DeBeyl DZ, et al: Value of aniongap determination in multiple myeloma. N Engl J Med 296: 5-60, 1977 3. Emmett M, Narins RG: Clinical use of the anion gap. Medicine 56:3-54, 1977 4. Frohlich J, Adam W, Golbey M, et al: Decreased anion gap associated with monoclonal and pseudomonoclonal gammopathy. CMA Journal 114:231-232, 1976 5. Gutman AB, Gutman EB, Jillson R, et al: Acid-base equivalence of the blood in diseases associated with hyperglobulinemia; with special reference to lymphogranuloma inguinale and multiple myeloma. J Clin Invest 15:475-44, 1936 6. Hobbs JR: Immunoglobulins in clinical chemistry. Adv Clin Chem 14:219-317, 1971 7. Keshgegian AA: Decreased anion gap in diffuse polyclonal hypergammaglobulinemia. N Engl J Med 299:99-100, 197. Madias NE, Ayus JC, Adrogue JH: Increased anion gap in metabolic alkalosis. The role of plasma-protein equivalency. N Engl J Med 300:1421-1423, 1979 9. Murray T, Long W, Narins RG: Multiple myeloma and the anion gap. N Engl J Med 292:574-575, 1975 10. Oh MS, Carroll HJ: The anion gap. N Engl J Med 297:14-17, 1977 11. Schnur MJ, Appel GB, Karp G, et al: The anion gap in asymptomatic plasma cell dyscrasias. Ann Intern Med 6:304-305, 1977 12. Van Leeuwen AM: Net cation equivalency (base binding power) of the plasma proteins. Acta Med Scand [Suppl 422] 176:3-212, 1964 13. Van Slyke DD, Hastings AB, Hiller A, et al: Studies of gas and electrolyte equilibria in blood, XIV. The amounts of alkali bound by serum albumin and globulin. J Biol Chem 79:769-70, 192