Predictive Medicine XIII. Tolerance Testing

ABSTRACT: The prime purpose of tolerance testing is to elicit an exaggerated response to a particular stimulus, as when glucose loading is used to determine the glucose tolerance pattern. Examples are presented to show the relationship of age to alveolar bone loss in terms of glucose tolerance tests (glucose stress), and of age to blood glucose concentration (no glucose stress). The amplitudes of fluctuation in clinical and biochemical signs are compared in healthy versus sick subjects. In a predictive medicine system, the aim is to ascertain a predisposition to, or early evidence of pathosis. Therefore, data obtained under ordinary circumstances (no stress) serve adequately if it is recognized that physiologic values and normal values are not synonymous and that subtle changes may be predictive.

 

“Testing under load or stress conditions may be likened to getting a young doctor drunk the morning he takes his state board examinations. The results may be revealing. However, the purpose of the state board is to establish how he will fare under the usual practice conditions, and it is hoped that he will usually not be intoxicated.” – Cheraskin

 

There is increasing interest in the diagnostic utility of testing under load or stress conditions. This is emphasized by the various exercise procedures prior to electrocardiography and by the stressing techniques (preparatory diet, glucose load, glucose and steroid load) employed for assessment of glucose tolerance. There is no question but that these pre-test conditions have yielded much valuable diagnostic information.

However, in a predictive medicine program designed to anticipate rather than identify disease, tolerance testing is less important. First, tolerance test circumstances are abnormal, whereas an objective of a predictive medicine system is to learn how the subject is coping with his everyday problems. Second, tolerance testing simply exaggerates the picture and makes the diagnosis easier or more convincing. Third, the identical information derived from tolerance testing can be obtained without loading if one is willing to accept the fact that physiologic and normal values are not the same and that small changes in a particular determinant are significant. These concepts have been discussed in earlier reports.^1-3

An attempt is made in this report to show the relative utility of tolerance versus “non-tolerance” (no load) tests for predictive purposes, employing blood glucose as the experimental model.

Tolerance Tests

The most popular load test in conventional medicine today is the glucose tolerance test. Figure 1 is an analysis of this procedure in 120 presumably healthy subjects.⁴ Fasting blood glucose groups were differentiated by increments of 10 mg per 100 ml. One subject in the category of 50-59 mg per 100 ml had a fasting blood glucose level of 58 mg per 100 ml, and 18 subjects in the category of 60-69 mg per 100 ml had a mean fasting blood glucose level of 65 mg per 100 ml. Figure 1 shows that the subsequent blood glucose levels at half an hour, one, two and three hours after glucose loading were directly related to the fasting concentration. The higher the fasting score, the higher the values following loading with glucose. Hence, the concentration of fasting blood glucose is predictive of the glucose tolerance pattern. Other reports have confirmed these findings with the classical glucose tolerance test5 and the cortisone glucose tolerance test.⁴

Fig. 1. Results of the classic glucose tolerance test (blood glucose, mg/100 ml) in 120 presumably healthy subjects. The pre-load (fasting) scores are predictive of the post-load values at half, one, two and three hours.

 

The additional problem was to determine what portion of the glucose tolerance pattern can serve best for predictive purposes. In order to establish the answer, one must make several assumptions. First, that both hyper- and hypoglycemia are pathologic, and that normoglycemia is physiologic.² Second, that there is a relationship between blood glucose concentration and oral pathosis, a subject previously reviewed.⁶

Figures 2-6 portray the relationship between age and alveolar bone loss in terms of the five phases of the classic glucose tolerance test.⁷ The data in Figure 2 indicate: 1) the highest and most significant correlation (r = +0.501, P < 0.01) between age and alveolar bone loss was in the hypoglycemic group; 2) the next highest correlation (r =+0.429, P < 0.01) was in the hyperglycemic group; and 3) there was even a significant correlation (r = +0.359, P < 0.05) in the normoglycemic group. Thus, fasting blood glucose values can serve in a predictive system.

Fig. 2. The relationship of age to mean alveolar bone loss in terms of fasting blood glucose concentration. There were statistically significant correlations in all three groups (hyper-, normo-, and hypoglycemic). The highest correlation was in the hypoglycemic, and the lowest in the normoglycemic group.

 

In Figure 3, age and alveolar bone loss are related to blood glucose concentration at thirty minutes after the glucose load. Three points warrant special: attention: 1) the pattern was virtually that observed for fasting blood glucose (Fig. 2); 2) the correlations with hyper- and hypoglycemia were higher than those observed in the fasting state; and 3) the relationship of normoglycemia to age and alveolar bone loss was not statistically significant (r = +0.243, P> 0.05). Hence, it would appear that the thirty-minute blood glucose value is more predictive than the fasting value.

Fig. 3. The relationship of age to mean alveolar bone loss in terms of the half-hour (after loading) blood glucose concentration. Statistically significant correlations were noted only in the hypo- and hyperglycemic groups. The highest correlation was in the group with a relatively low blood glucose level.

 

Figure 4 shows the relationship of age and alveolar bone loss to the blood glucose level one hour after glucose loading. The highest and most significant correlation was found in the normoglycemic group (r = +0.483, P < 0.01). Obviously the one-hour blood glucose value has no prognostic significance.

Fig. 4. The relationship of age to mean alveolar bone loss in terms of the one-hour (after loading) blood glucose concentration. There were statistically significant correlations in all three groups. The highest correlation was in the normoglycemic, and the lowest in the hyperglycemic subjects.

 

In Figure 5, age and alveolar bone loss are related to the blood glucose level. There were significant correlations with both hyper and hypoglycemia, by none with normoglycemia.

Fig. 5. The relationship of age to mean alveolar bone loss in terms of the two-hour (after loading) blood glucose concentration. Statistically significant correlations were found only in the hyper- and hypoglycemic subjects, with the highest correlation in the low blood glucose group.

 

The three-hour observations showed the highest correlation with low blood glucose levels, and the lowest correlation with high blood glucose levels (Fig. 6).

 

Fig. 6. The relationship of age to mean alveolar bone loss in terms of the three-hour (after loading) blood glucose concentration. There were statistically significant correlations in all three groups, with the highest correlation in the hypoglycemic, and the lowest in the hyperglycemic subjects.

 

On the basis of the data in Figures 2-6, it would appear that the most predictive temporal point is two hours after glucose loading (Fig. 5). The reasons are: 1) the rate of alveolar bone loss with age was not statistically significant for the normoglycemic group (r = +0.245, P> 0.05); 2) the hypoglycemic correlation (r = +0.531, P < 0.01) was exceeded only by the correlations at thirty minutes and three hours; 3) the hyperglycemic correlation (r = +0.500, P < 0.01) was higher than at any other temporal point. It is noteworthy that the findings for the two-hour load period were similar to the findings for the two-hour no-load postprandial period recognized by many authorities as the best point at which to measure blood glucose.

“Non-Tolerance” (No Load) Tests

Many years ago, Claude Bernard pointed out that life and death are functions of homeostasis, the steady state.² Since then, he and others have shown that this state is indeed steady. Health is characterized by only small variations, but disease is characterized by wide fluctuations in clinical and biochemical signs.

A graph of the concept of homeostasis is shown in Figure 7. For example, in a healthy subject the temperature varies minimally during the day, but in a sick person it may fluctuate widely. In the healthy person, the psychic state varies only slightly (from minimal elation to marginal depression); in the sick, it may vary in abrupt manic-depressive cycles. The patterns for blood pressure, peristalsis, and other clinical and biochemical signs vary in similar fashion.

Fig. 7. In health, factors such as temperature, blood pressure, psychic state, peristalsis, or blood sugar concentration can be represented by small daily fluctuations. In disease, the amplitudes of the fluctuations are much greater.

 

Figure 8 represents an extension of Figure 7 over a period of years. In the very early stages of chronic disease, the amplitudes increase. With the passage of time, the large undulations become smaller and slowly rise toward a plateau, e.g., the course of hyperglycemia in diabetes mellitus, and of hypertension in heart disease.

Fig: 8. This is an extension of Figure 7 over a period of years. In the early stages of chronic disease, the amplitudes increase. With additional time, the broad undulations become narrower and the baseline slowly rises towards a plateau.

 

On this basis, it follows that relatively young persons usually show relatively small variations in clinical and biochemical signs. With advancing age, the fluctuations increase and the basic levels slowly rise, as indicated in Figure 9.

Fig. 9. This is a supplement of Figure 8, showing that, with aging, the amplitudes of fluctuation become larger and the mean scores slowly rise.

 

To obtain data on this point, 272 presumably healthy persons were studied in terms of blood glucose concentration at 8:30 a.m., 10:30 a.m., 12:30 p.m., 2:30 p.m., and 4:30 p.m., without glucose loading.⁷ Figure 10 shows the means and standard deviations for the diurnal blood glucose patterns in the relatively young (less than 40 years old), in those of intermediate age (40-49 years), and in the elderly (50+ years). With advancing age, the means scores slowly rose, and there was an increase in variance. This is a quantitative demonstration of the patterns delineated in Figures 8 and 9.

Fig. 10. Means and standard deviations (mg/100 ml) for diurnal blood glucose (without loading) in different age groups. The mean values slowly rose and the variances (amplitudes) slowly increased.

 

Figure 11 shows the difference in diurnal blood glucose levels between elderly subjects who were relatively well (0-5 clinical signs and symptoms) and those who were relatively sick (40 or more clinical signs and symptoms). The mean values varied more and the fluctuations were much greater in the sick subjects. This supports the concepts depicted in Figure 8 and 9.

Fig. 11. Means and standard deviations (mg/100 ml) for diurnal blood glucose (without loading) in subjects aged 50 or older. The comparison is between subjects with only 0.5 general symptoms and signs (Cornell Medical Index, CMI) and those with 40 or more symptoms and signs. In the relatively sick group, the mean values were more variable and the variances (amplitudes) were distinctly greater.

 

References Cited:

1. Cheraskin, E. and Ringsdorf, W. M., Jr.: “Predictive medicine. IV. The gradation concept,” J. Am. Geriatrics Soc. 19: 511 (June) 1971.
2. Cheraskin, E., and Ringsdorf, W. M., Jr.: “Predictive medicine. V. Linear versus curvilinear functions,” J. Am. Geriatrics Soc. 19: 721 (Aug.) 1971.
3. Cheraskin, E., and Ringsdorf, W. M., Jr.: “Predictive medicine. VI. Physiologic versus normal values,” J. Am. Geriatrics Soc. 19: 729 (Aug.) 1971.
4. Cheraskin, E.: Ringsdorf, W. M., Jr.; Setyaadmadja, A. T. S. H., and Barrett, R. A.: “Clinical chemistry and predictive medicine,” J. M. A. Alabama 36: 1337 (May) 1967.
5. Frethem, A. A.: “Clinics on endocrine and metabolic diseases. 10. Relation of fasting blood glucose level to oral glucose tolerance curve,” Proc. Staff Meet. Mayo Clin. 38: 110 (March 13) 1963.
6. Cheraskin, E., and Ringsdorf, W. M., Jr.: “Predictive medicine. XII. The oral cavity,” J. Am. Geriatrics Soc. (in press).
7. Unpublished data, Department of Oral Medicine, University of Alabama Medical Center, Birmingham, Alabama.