An Empirical Regression Equation Relating Total Serum Calcium to Serum Albumin and Globulins

Abstract

Conclusions

The above considerations suggest that the total serum calcium is composed of at least 4 fractions: 1. Calcium bound to and varying with albumin; 2. calcium bound to a globulin fraction, relatively constant in amount irrespective of the total globulin level; 3. calcium bound to another globulin fraction, varying with that globauin fraction, and though small, becoming siginificant at high total globulin levels; 4. calcium not bound to protein, relatively constant because cases with gross variations are excluded by definition. These fractions are represented in the following general regression equaiton:

IV. Total Ca = m ₁. albumin + m ₂. “globulin II” + m ₂. “globulin I” + b.

Where b = 5.8 ± 0.2 mg. per 100 gm. serum H₂O, analysi of our data suggests that m ₁ is of the order 0.7-0.9 mg. Ca per gram albumin; the product m ₂. “globulin II” is a constant of the order 1.0-1.5 mg. Ca per 100 gm. serum H₂O; m ₃ is of the order 0.1-0.2 mg. Ca per gram “globulin I” where “globulin I” is defined arbitrarily as all globulin in excess of 3.0 gm. total globulin.

While the method does not warrant further precision of constants, it is of interest that a number of formulae based on equation IV may be devised which tend to obviate the systematic divergencies of equations I, II and III. Table I and Fig 1 illustrate the results where:†

V. Total Ca = .80 albumin + 7.0 + 0.2. (total globulin—3.0).

Calcium is expressed in mg., albumin and globulin in gm. per 100 gm. serum H₂O. The constant 7.0 is the sum of b and m ₂. “globulin II”. The last term is used only if the total globulin content exceeds 3.0 gm.

The fact that equation IV applies more generally to our data is regarded as supoort for the inferences drawn concerning calcium bound to the several protein fractions binding calcium and the approximate amounts so bound. It should be emphasized that no special signigicance is attached to the particular values used by us in equation V except as they are indicative of approximate orders of magnitude. We do not believe that a discrepancy between observed calcium values and those calculated by this or other equations necessarily implies the presence of a primary disturbance in calcium metabolism. As a prediction formula, equation V is of academic interest only, since total serum calcium may be determined directly with precision.