Division of Endocrinology Bnai Zion Medical Center Haifa, 31048 Israel
In the study by Mayenknecht et al. (1), the low (1 µg) and high (250 µg) ACTH tests were compared in 35 subjects without pituitary or adrenal disease, and in 40 patients with pituitary disease, in most of whom the insulin tolerance test (ITT) was used to define normalcy or abnormality of the pituitary-adrenal axis. Their data show that, in normals, the low dose was as specific as the high dose in elucidating an adrenal response, with no cases who did not respond to the low dose test. In patients with pituitary disease, they show that in their group of patients, both the low dose and the high dose tests showed a sensitivity of 100% in picking up 14 patients with clearly subnormal reference test. However, they define a group of 9 patients with "slightly subnormal insulin test," in whom peak cortisol responses to ITT were between 1820 µg/100 mL (500550 nmol/L).
One should be very careful in defining such a response to ITT as pathologic. Four papers were published in which ITT were done in normal volunteers. Nelson and Tindall (2) showed that 5 out of 24 (21%) normal volunteers (see Fig. 1) and 13 out of 44 (29.5%) control subjects (see Fig. 2) had peak cortisol levels of 20 or less µg/100 mL. Evans et al. (3) found a mean peak cortisol level of somewhat less than 500 nmol/L (see Fig. 1). Avgerinos et al. (4) defined the lower limit of normal response at about 460 nmol/L (see Fig. 1). In the most impressive study (5), ITT was done twice in 16 normal volunteers. Ten of the 32 tests (31%) had peaks of less than 550 nmol/L, despite nadir blood glucose of less than 2.0 mmol/L. More interesting is the fact that, in 4 out of the 16 volunteers tested twice, the peak cortisol in one of the tests was less than 550 nmol, while in the other test it was higher than 600, despite good CV(%) of 812. Using Mayenknechts criteria, these 4 normal volunteers would quite easily pass the test one day and flunk it a few days later, or vice versa. It is therefore clear that cortisol responses to ITT between 1820 µg/100 mL (500550 nmol/L) cannot be considered even slightly abnormal. Actually, if we add groups 1 (normal ITT responses, 21 patients) and 3 (peak cortisols of 1820 µg/dL, 9 patients) in the study, we can see that group 3 constitutes 30% of normal responders, which is a similar percentage to the studies quoted above.
All the investigators show is that in their group of patients the low dose ACTH test (LDT) was not much more sensitive than the high dose test (HDT). The reason is probably that in this group of patients the sensitivity of both tests was high. This sensitivity of the HDT here may be explained by a long duration of pituitary disease. Indeed, out of the 14 patients defined as having clearly pathological reference tests, 5 actually had no reference tests, as the morning cortisol level was suppressed and should not have been included at all. Others (number not defined) were on steroid treatment for months to years (otherwise not specified), as proof of long-standing adrenal suppression by the disease. In such patients the superiority of the low dose versus the high dose ACTH test also can not be evaluated. The longer the disease, the higher the adrenal suppression and the sensitivity of even the high dose ACTH test. Even with these limitations, we find 2 patients in the "clearly pathological" group who showed false negative normal response to the HDT, but correct abnormal response to the LDT. The authors admit to one of these cases in the results section, while the other case had a 60 min cortisol of about 26 µg/dL, which can by no means be called an abnormally low response. Both these patients had cortisol responses of less than 18 µg/dL to the LDT. If we look at the data of the 9 patients who really needed a stimulation test, the sensitivity of the LDT was 100%, while the sensitivity of the HDT was 78%.
The authors also measured ACTH levels in the blood after iv injection of 1 µg ACTH and found a very high level of 1900 pg/mL. They claim that this proves that the low dose ACTH test actually uses a pharmacological dose, which compromises its physiologic value. However, their ACTH levels were much higher than those found by Graybeal and Fang (6) in response to 0.05 µg/kg (about 3.0 µg) ACTH, which were lower than 800 pg/mL, and similar to ACTH levels achieved after insulin hypoglycemia. A recent study (7) found ACTH levels 2 min after the low dose ACTH test to be 546 ± 70 pg/mL. We have lately shown that ACTH doses of less than 1.0 µg fail to stimulate the adrenal maximally (8). All this suggests that the 1-µg ACTH dose is physiologic rather than pharmacological. If the ACTH level reached in the 1-µg test is indeed as high as the authors claim, how can the lowest normal cortisol response be defined as lower than the one needed in the ITT, where ACTH peak levels are much lower408 pg/mL (6)?
The low dose ACTH test was designed to induce an ACTH test that will reduce false negative responses without creating false positive ones (9). To the papers suggesting that this is the case, another one was added lately (10). In the present series, the 0% false positive resultsboth in normals and in pituitary patientswas proven again. Lowering false negative results was moderate (22%). Unlike other studies, most patients with compromised pituitary-adrenal axis in the present series showed an abnormal response even to the high dose test. However, sensitivity to the low dose test was 100%.
Footnotes
Address correspondence to: Gabriel Dickstein, Division of Endocrinology, Bnai Zion Medical Center, 47 Golomb Street, P.O. Box 4940, Haifa, 31048 Israel.
Received July 9, 1998.
References