Published in: J. Theoret. Biol. vol. 2, pp. 159-164 (1962): 

J. H. Frenster, Department of Growth Physiology, Walter Reed Army Institute of Research, Washington 12, D.C., U.S.A.

(Received 6 November, 1961) 

The tasks of diagnosis include not only the detection of the presence of disease, but also the quantitation of its magnitude in a given subject. Detection of disease is the basis of case-finding; quantitation of disease is an important guide to rational therapeutics and prognostication.

Testing Procedures:

Disordered body processes in many disease states have been detected by the use of increasingly sensitive electronic, radioisotopic, biochemical, immunologic, and cellular techniques. Quantitation of the magnitude of disease present in such processes can be approached by a variety of testing procedures. The testing procedures include tolerance tests, tracer tests, and sensitivity tests. They have as a common feature the estimation of the available capacity of the body process that is tested (Frenster, 1961). Implicit in the use of these tests is the assumption that the magnitude of disease present in a body process is reflected in a proportionate change in the capacity of that process to deal with a normal or an increased load that may be applied to it. (Frenster, 1961).

Tolerance tests are the most useful method for determining the capacity of a body process, since they can measure both the size of the maximal load acted upon by the process, and the rate of its action (Himsworth, 1935). Tracer tests can measure the rates of action of body processes, but since only  minute quantities of a tagged load are administered, they do not yield direct data concerning maximal loads possible (Robertson, 1957). Sensitivity tests can yield only indirect data concerning a given process, since they in fact measure only the responses in other processes which follow upon the action of the tested process (Himsworth, 1935).

The increased loads employed in tolerance tests may not be needed to demonstrate the reduced capacity of  a process if  its capacity is so reduced that failure is evident at normal or physiologic loads to the process. In effect, such physiologic loads are spontaneous tolerance tests, but because such loads are poorly quantitated, they are less useful for quantitating the capacity of the affected process, or for estimating the magnitude of disease.

Tolerance Tests:

Tolerance tests have acquired a widespread variety and application in the early diagnosis of disordered function since their innovation in 1914 (Kleiner & Meltzer, 1915) and their clinical introduction in 1915 (Hopkins, 1915). They have been used to measure the disposition within a sunject of increased loads of such varied substances as glucose (Hopkins, 1915; Hamman & Hirschman, 1917), galactose (Maclagan,, 1940), water (Robinson et al., 1941), salt (Soffer et al,. 1944), bromsulphalein (Bradley et al., 1945), digitalis (Friedman & Bine, 1949), calcium (Howard etal., 1953), ammonium (White et al., 1955), phenylalanine (Hsia et al., 1956), uracil (Carren & Corbo, 1957), fructose (Craig et al., 1958), salicylate (Childs et al., 1959), hemoglobin (Lathem & Worley, 1959), magnesium (Barber et al., 1959), epinephrine and norepinephrine (Cohen et al., 1959), ACTH (Meakin et al., 1959), adrenocorticoids (Peterson, 1960), proconvertin (Hoag et al., 1960), ethanol (Castenfors et al., 1960), insulin (Grodsky & Forsham, 1960), lactic acid (Handler, 1960), triglycerides and chylomicrons (Havel & Gordon, 1960), phosphate (Reynolds et al., 1960), and quinidine (Scherlis et al., 1961). Tolerance tests have assumed decisive importance for case-finding of recessive metabolic lesions in subjects with heterozygous genetic defects (Hsia, 1960; Kirkman, 1960). In addition, their increasing sensitivity permits the earlier detection of many other latent diseases, and therefore the earlier application of preventive measures and of therapy (Backett, 1960).

Tolerance Tests in Diabetes:

Studies of subjects with diabetes mellitus have served as models for studies on the magnitude of disease in other body processes and diseases. Such studies reveal both the difficulties and the usefulness of quantitating the magnitude of diabetes in a given subject. They suggest similar problems and prospects for other diseases and processes.

Diabetes mellitus is characterized by defective glucose utilization (de Bodo et al., 1959), which follows upon a deficiency of endogenous insulin, or an antagonism to its action. The deficiency of insulin may be absolute, or only relative (Haist, 1959) to an increased load of glucose (del Greco & Scapellato, 1953) or to an insulin antagonist. In absolute insulin deficiency, the degree of defective glucose utilization and tolerance is a functional expression of the magnitude of the insulin deficiency. The magnitude of such glucose intolerance can be correlated anatomically with the degree of beta cell degranulation in the islets of Langerhans (Hartroft & Wrenshall, 1955), correlated cheically with the reduced quantity of insulin extractable from the pancreatic tissue (Wrenshall et al., 1954), and correlated immunologically with the reduced quantity of plasma precipitable with heterologous antibodies to insulin (Yalow & Berson, 1960).

The increasing sensitivity of tolerance tests is well illustrated by their development in the study of diabetes mellitus. Historically, diabetes was diagnosed by the simple detection of wasting associated with sweet polyuria and eventual acidosis. Later, fasting hyperglycemia was recognized as an earlier manifestation of the disease. The glucose tolerance test was then found to be abnormal in diabetic subjects who had no significant fasting hyperglycemia (Conn, 1940). Recently, the cortisone-glucose tolerance test has been shown to be capable of detecting or predicting diabetes in subjects with normal glucose tolerance tests (Fajans &Conn, 1959). By the administration both of a glucose load and of cortisone antagonism to the action of insulin, the cortisone-glucose tolerance test is designed to elicit maximal endogenous insulin production and secretion in subjects suspected of early insulin deficience.

In general terms, the cortisone-glucose tolerance test hopes to detect those subjects whose capacity for endogenous insulin production is adequate for normal loads and resistances, but inadequate for increased loads and resistances (Frenster, 1961). During the course of this test, thesubject is temporarily brought to a state of high-output failure of insulin production (Frenster, 1960), manifested functionally as significant hyperglycemia. Since the capacity for insulin synthesis in any subject is ultimately limited by spatial and thermodynamic factors, all subjects would ultimately fail this test if the glucose load and cortisone dosage were high (Frenster, 1961). The load required to produce such failure would be an estimate of the magnitude of the disease in the process of insulin production. It now appears likely, from longitudinal follow-up of tested subjects, that the load of glucose and the dosage of cortisone presently employed in the test are just high enough to detect early or latent disease, but not so high as to include normal subjects (Fajans & Conn, 1959). Further follow-up will yield further data on this point.

Quantitation by Tolerance Tests:

At least four methods have so far been applied to analyze glucose tolerance test with a view to quantitating the magnitude of diabetes present in the tested subject. These methods include:

(1) quantitation of that height to which glucose intolerance proceeds after a standard load of glucose (Himsworth, 1935; Hopkins, 1915' Hamman & Hirschman, 1917);

(2) quantitation of that minimal load of glucose which first evokes the signs of glucose intolerance (Hopkins, 1915; Hamman & Hirschman, 1917);

(3) quantitation of that insulin replacement dosage needed to restore glucose tolerance to normal (Himswoth, 1933; Amatuzio et al., 1954; Marks & Bishop, 1957); and

(4) slope analysis of the blood glucose disapperance curve after a glucose load (Amatuzio e al., 1953; Hlad & Elrick, 1960), with calculation of the disappearance constant (Hlad & Elrick, 1959). Each of these methods has intrinsic limits, and none is completely satisfactory, but each is capable of estimating in a preliminary way the magnitude of diabetes present in the tested subject.

The use of a standard load of glucose rather than a variable load results in a wide continuum of responses in tested subjects, ranging from no significant intolerance in latent diabetes or in normal subjects, to marked or poorly measurable intolerance in diabetic subjects with advanced disease (Hopkins, 1915; Hamman & Hirschman, 1917). It is uncertain as yet whether the magnitude of intolerance recorded should only be the height of the displayed intolerance, or should also include its duration (Hamman & Hirschman, 1917). The administration of large standard loads during tolerance tests of other body processes may be toxic to vulnerable subjects (Kirkman, 1960).

In contrast, the use of a variable load of glucose to determine that minimal load required to first evoke the signs of intolerance involves the use of multiple tests of successively larger loads, which is both inconvenient and perhaps misleading because of the effects of training and adaptation to the loads (Conn, 1940).

Slope analysis of blood disappearance curves optimally requires steady equilibration at a high blood level of the load substance (Hlad & Elrick, 1959), which can perhaps only be obtained by constant perfusion methods of administration. Uncertainty over the proper mathematical analysis of the disappearance has not yet been resolved (Hlad & Elrick, 1959).

The use of increasing doses of insulin replacement therapy to quantitate that minimal dose wgich normalizes glucose tolerance is perhaps the most satisfactory method in studying diabetic subjects, and is the most germane to the therapeutic problem. The results of such titration may be ambiguous in those diabetic subjects who have insulin-binding antibodies in their plasma in significant amounts (Field, 1959; Berson & Yalow, 1959). Furthermore, this method is not available for studying those diseases and body processes for which there exists no current replacement therapy.

The common advantages of these tolerance test methods are that they can be used in serial studies of a subject at various stages of his disease or therapy. They measure directly the disposition of the load substance involved in the diseased process, rather than secondary responses or derived data. In this feature, they do not require the precise knowledge of mechanism nor purity of substance that biochemical or immunolgic methods demand. Finally, their use in vivo makes them superior to methods which require removal of body tissues for in vitro analysis.

Extended Uses for Tolerance Tests:

Tolerance tests are also capable of reflecting the fluctuating magnitude of health or of disease in a given body process. The tolerance of a body process for high loads and resistances is known to be a variable characteristic, determined by such factors as the training or disuse of the process (Frenster, 1961). The changes in available capacity of the process produced by training or disuse are well characterized for such biophysical processes as cardiac and sleletal muscle action (Astrand, 1956), but are less obvious in biochemical processes. That such do occur, however, is well illustrated by the fact that repeated feedings of large loads of glucose to a subject will soon increase the tolerance for glucose, and improve the glucose tolerance test (Conn, 1940), while conversely, deprivation of glucose soon reduces the tolerance for glucose (Peters, 1945).

Such adaptations by the available capacity of a process to the levels of applied load and resistance imply a close responsiveness by the process to its micro-environment (Frenster, 1961). Whether specific biochemical processes in disease states could be rendered less vulnerable by some form of periodic metabolic training has not yet been determined. The genetic balanced polymorphism of the human population (Neel & Schull, 1954) indicates a widespread distribution of undetected heterozygotic individuals with recessive metabolic lesions. Perhaps some of these vulnerable persons could be rendered more resistant to the eventual overt manifestation of their metabolic defects by some form of metabolic training. Tolerance tests would probably be the basis for both the detection of such genetic lesions and for the quantitative estimation of disease in such subjects before and after metabolic training.

References:

Amatuzio DS, Schultz AL, Vanderbilt MJ, Rames ED, & Nesbit S, J. Clin. Invest. 33: 97 (1954).

Amatuzio DS, Stutzman TL, Vanderbilt MJ, & Nesbit S, J. Clin. Invest. 32: 428 (1953).

Astrand PO, Physiol. Rev. 36: 307 (1956).

Backett EM, Lancet 2: 1075 (1960).

Barker ES, Elkinton JR, & Clark JK, J. Clin. Invest. 38: 1935 (1959).

Berson SA, & Yalow RS, J. Clin. Invest. 38: 1996 (1959).

Bradley SE, Ingelfinger FJ, Bradley GP, & Curry JJ, J. Clin. Invest. 24: 890 (1945).

Carren R, & Corbo L, J. Clin. Endocrin. Metab. 17: 1071 (1957).

Castenfors H, Hultman H, & Josephson B, J. Clin. Invest. 39: 776 (1960).

Childs B, Sidbury JB, & Migeon CJ, Pediatrics 23: 903 (1959).

Cohen G, Holland B, Sha J, & Goldenberg M, J. Clin. Invest. 38: 1935 (1959).

Conn JW, Am. J. Med. Sci. 199: 555 (1940).

Craig JW, Miller M, Mackenzie MS, & Woodward H, Jr., J. Clin. Invest. 37: 118 (1958).

de Bodo RC, Altszuler N, Dunn A, Steele R, Armstrong DT, & Bishop JS, Ann. N.Y. Acad. Sci. 82: 431 (1959).

del Greco F, & Scapellato L, Diabetes 2: 457 (1953).

Fajans SS, & Conn JW, Ann. N.Y. Acad. Sci. 82: 208 (1959).

Field JB, J. Clin. Invest. 38: 551 (1959).

Frenster JH, Ann. Int. Med. 53: 647 (1960).

Frenster JH, Perspectives Biol. Med. 4: 152 (1961).

Friedman M, & Bine R, Jr., J. Clin. Invest. 28: 32 (1949).

Grodsky GM, Hultman E, & Josephson B, J. Clin. Invest. 39: 776 (1960).

Haist RE, Ann. N.Y. Acad. Sci. 82: 266 (1959).

Hamman L, & Hirschman II, Arch. Int. Med. 20: 761 (1917).

Handler JS, J. Clin. Invest. 39: 152b (1960).

Hartroft WS, & Srenshall GA, Diabetes 4: 1 (1959).

Havel RJ, & Gordon RS, Jr., J. Clin. Invest. 39: 1777 (1960).

Himsworth HP, Clin. Soc. 1: 1 (1933).

Himsworth HP, Clin. Soc. 2: 67 (1935).

Hlad CJ, Jr., & Elrick H, J. Clin. Endocrin. Metab. 19: 1258 (1959).

Hlad CJ, Jr., & Elrick H, J. Clin. Endocrin. Metab. 20: 1120 (1960).

Hoag MS, Aggeler PM, & Fowell AH, J. Clin. Invest 39: 554 (1960).

Hopkins AH, Am. J. Med. Sci. 149: 254 (1915).

Howard JE, Hopkins TR, & Connor TB, J. Clin. Endocrin. Metab. 13: 1 (1953).

Hsia DY-Y, Metabolism 9: 301 (1960).

Hsia DY-Y, Driscoll KW, Troll W, & Knox WE, Nature 178: 1239 (1956).

Kirkman HN, Metabolism 9: 316 (1960).

Kleiner IS, & Meltzer SJ, Proc. Natl. Acad. Sci. Wash., 1: 338 (1915).

Lathem W & Worley WE, J. Clin. Invest. 38: 1733 (1959).

MacLagan NF, Quart. J. Med. 9: 151 (1940).

Marks PA, & Bishop JS, J. Clin. Invest. 36: 254 (1957).

Meakin JW, Bethune JE, Despointes RH, & Nelson DH, J. Clin. Endocrin. Metab. 19: 1491 (1959).

Neel JV, & Schull WJ, "Human Heredity", Univ. Chicago Press, (1954).

Peters JP, Yale J. Biol. and Med. 17: 705 (1945).

Peterson RE, J. Clin. Invest. 39: 320 (1960).

Reynolds TB, Lenman H, & Tupikova N, J. Clin. Endocrin. Metab. 20: 1136 (1960).

Robertson JS, Physiol. Rev. 37: 133 (1957).

Robinson FJ, Power MH, & Kepler EJ, Proc. Staff Meet. Mayo Clin. 16: 577 (1941).

Scherlis L, Gonzalez LF, & Bessman SP, J. Clin. Invest. 40: 60 (1961).

Soffer LJ, Lesnick G, Sorkin SZ, Sobotka H, & Jacobs M, J. Clin. Invest. 23: 51 (1944).

White LP, Phear EA, Summerskill WHJ, & Sherlock S, J. Clin. Invest. 34: 158 (1955).

Wrenshall GA, Hartroft WS, & Best CH, Diabetes 3: 444 (1954).

Yalow RS, & Berson SA, J. Clin. Invest. 39: 1157 (1960).

Additional References:

Frenster JH, "Load Tolerance as a Quantitative Estimate of Health", Ann. Int. Med. 57: 788 (1962).

Frenster JH, "Analysis of Queueing and Renewal Within Human Systems", Nature 207: 1139 (1965).