Professor Emeritus of Medicine SUNY Health Science Center Syracuse, New York 13210
Address correspondence and requests for reprints to: David H. P. Streeten, Dept. of Medicine, SUNY Health Science Center, 750 E. Adams St., Room CWB-322, Syracuse, New York 13210.
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() |
---|
However, it has been recognized for several years that, in some patients who had adrenal insufficiency secondary to hypothalamic or pituitary failure, the normal result of the short ACTH test with 250 µg did not necessarily indicate normal function of the entire hypothalamic-pituitary-adrenal (HPA) system. This evidence and the fact that plasma ACTH levels rose to extremely high levels after 250 mg of ACTH124 led to attempts to determine whether the elevation of plasma ACTH to more physiologically increased concentrations by the injection of lower doses of ACTH124, might provide a more reliable means of diagnosing moderate or mild forms of corticotropin deficiency. The fact that 1 µg and 250 µg doses of ACTH124 resulted in very similar plasma levels of glucocorticoids at 30 min (3) was substantiated in several subsequent measurements, particularly those of Dickstein et al. (4). These clinical investigators found, in six patients who had been on long-term steroid therapy, that serum cortisol concentration rose to 649 (SD 91) nmol/L, similar to the responses of normal subjects (to 707, SD 55, P > 0.3) after the 250 µg dose of ACTH124, but rose subnormally after 1 µg of ACTH124 in five of the six steroid-treated patients, to 458 (SD 86) compared with 704 (SD 72) nmol/L in normal subjects (P < 0.01). In contrast to these abnormal responses to the 1 µg dose of ACTH124, only one of the six steroid-treated patients responded subnormally to the 250 µg dose. These impressive data stimulated several more recent publications by Dicksteins and other groups (5, 6, 7, 8), in which it was reported that some patients with suspected or established ACTH deficiency responded subnormally to the 1 µg dose of ACTH124 (in some instances confirming the conclusions from an abnormal insulin tolerance test) but normally to the 250 µg dose. Such results might reasonably have been expected now that it has been shown that, unlike the usual up-regulation of end-organ receptors in response to reduced concentrations of the appropriate agonist, the ACTH receptors on human adrenocortical cellsat least during culture in vitroare up-regulated when exposed to increasing ACTH concentrations (10) and would be expected therefore, to be down-regulated by the subnormal plasma concentrations of endogenous ACTH in patients with ACTH deficiency.
The article in this issue of JCE&M by Abdu et al. (see page 0000) (9) reports another comparison of cortisol responses to the 1 and 250 µg ACTH124 tests and to the insulin tolerance test in a series of 64 patients with suspected or proven pituitary diseases of various types. Using 600 nmol/L (22 µg/dL) as the lower limit of the normal serum cortisol response to 1 µg ACTH124, the authors found a sensitivity of 100% compared with responses to the insulin tolerance test. Eleven percent of their patients were abnormal by the low dose test without other clinical or biochemical evidence of cortisol deficiency. The low dose test was found to be superior to the 250 µg ACTH test, and the authors concluded that, as a screening procedure, the low dose test was an attractive alternative to the insulin tolerance test that could be reserved for patients in whom careful clinical evaluation failed to dispel reasonable doubts of possible HPA axis impairment. Used in this way, as a screening test for HPA insufficiency and with the recognition that the insulin tolerance test would be required in patients with borderline responses, the low dose test might eventually prove to be a useful diagnostic procedure.
Regrettably, the apparently encouraging evidence supporting the usefulness of the low dose ACTH test has not eliminated all uncertainties. At least two problems indicate that the 1-µg ACTH test cannot provide a conclusive means of excluding the presence of mild or incomplete deficiencies in the functional capacity of the HPA axis. The first problem to be addressed is the fact that the lower limit of the normal cortisol response to 1 µg of ACTH124 has been variously considered to be 600 nmol/L (22 µg/dL) (9), 560 nmol/L (20 µg/dL) (4), 535 nmol/L (19 µg/dL) (11), 500 nmol/L (18 µg/dL) (6), and 480 nmol/L (17 µg/dL) (5). Obviously, the interpretation of normality and of the specificity of the low dose test will be very dependent on the determination of the lower limit of the normal range. Moreover, this determination is going to vary widely from one laboratory to another and from determinations performed with one RIA kit compared with another (12). Unless and until a generally acceptable lower limit of the normal range can be defined, the low dose procedure is likely to prove more confusing than diagnostically helpful to clinicians. And body size will probably influence the results of serum cortisol responses to such low doses of ACTH, as some investigators have appreciated (11, 12, 13). Therefore, if limits of the normal range of the 30-min plasma cortisol concentration can be agreed upon, it would be preferable to have these expressed in terms of serum cortisol responses to a dose of ACTH124 expressed in µg/m2 or µg/1.73 m2.
There is another difficult problem that will confront physicians who allow their diagnostic decisions to depend entirely on the low dose ACTH or even the insulin tolerance test. This results from the evidence, already clearly available, that mild deficiencies of ACTH secretion do not always prevent a normal rise in serum cortisol concentration stimulated by injected ACTH, thereby failing to indicate the potential danger of adrenocortical insufficiency during exposure to severe stress. This has been strongly suggested by the findings of a normal response to the 250 µg dose of ACTH124 in individuals with severely abnormal responses to the insulin tolerance test (14, 15) or the metyrapone test (16). The potentially life-threatening consequences of relying on a "normal" response to 250 µg of ACTH124 given i.v. have been dramatically illustrated in four patients (17). The partial impairment in their ACTH response to severe stress was disguised by the presence of a plasma cortisol concentration 30 min after 250-µg injection of ACTH124, normal for a mildly stressed individual but inadequate for the severe stresses to which these patients had been exposed. Their hypotension (exclusively orthostatic in two and present in recumbency in the other two patients) and other evidence of adrenal insufficiency were associated with a brisk response in plasma cortisol to i.v. ACTH124 (250 µg) and were rapidly responsive to intravenous steroid therapy. These patients were later shown to have subnormal responses to the overnight metyrapone test. How often this clinical picture occurs in our intensive care units and other hospital beds cannot be reliably estimated. Even at autopsy in such patients it is very difficult to establish whether the presence of immunoreactive ACTH in the pituitary and its release in response to stress were adequate to prevent death from secondary adrenocortical insufficiency.
The evidence that mild deficiency of ACTH release is a real entity and, like mild deficiency of thyroxine and several other hormones, may be as common as severe deficiency, has been supported by the meticulous studies of Mayenknecht et al. (11). These investigators found, in patients with clear evidence of moderate or severe ACTH deficiency (>10% reduction below lower limits of normal responses to the insulin tolerance or the metyrapone test), that the serum cortisol responses to ACTH124 were subnormal in all 14 patients when low dose ACTH124 was used (0.5 µg/m2), and in 13 of the 14 patients when the 250 µg dose was used. However, among 9 additional patients whose insulin tolerance or metyrapone tests were only mildly impaired (<10% below the lower normal limits of these tests), responses to ACTH124 were normal at 30 min in 7 of the 9 patients during the 250 µg dose and in 8 of the 9 patients during the 1 µg dose of ACTH124. The authors concluded that there was no significant difference in the responses to the high and low dose tests. Even more important, perhaps, is their demonstration that 8 of their 23 patients (35%) with ACTH deficiencies, all with evidence of mildly subnormal responses to the insulin tolerance or metyrapone test would have been passed as having a functionally normal HPA axis if either the high- or low-dose ACTH test had been relied upon. Additional evidence that the low dose test is no more successful than the 250 µg test in identifying patients with functional impairment of the HPA axis was reported by Weintrob et al. (18).
![]() |
What should we do in the face of these diagnostic predicaments? |
---|
![]() ![]() ![]() ![]() |
---|
In patients seen when under stress or who are candidates for stress (e.g. surgery), the short ACTH test, using either 250 or 1 µg doses, continues to be a reliable indicator of the need for steroid therapy if the serum cortisol response is subnormal; but unfortunately, it cannot be relied upon if the response is normal. It is known that the normal increase in adrenocortical secretion during stress is roughly proportional to the severity of the stress, as Chernow et al. (19) have shown. These authors reported that, in response to such mild surgical stresses as laparoscopy and repair of an inguinal hernia, serum cortisol rose 1 h after the incision to only slightly above the level reached normally after a short 250-µg ACTH test: 500630 nmol/L (or 1823 µg/dL). More stressful procedures, such as cholecystectomy, appendectomy, and vaginal hysterectomy, raised serum cortisol to mean values above 900 nmol/L (32 µg/dL), and the most stressful forms of surgery evaluated (including colectomy, gastrectomy, and aorto-femoral bypass) increased serum cortisol to approximately 1430 nmol/L (or 52 µg/dL), 1 h after the onset of surgery.
When measured hourly (in other studies) for longer periods of time after the initial incision, plasma cortisol concentrations have increased steadily from mean values of 750950 nmol/L (2734 µg/dL) at 30 min, to peak concentrations of 12701360 nmol/L (4649 µg/dL) at or after the conclusion of uncomplicated cholecystectomy (17, 20). There is little doubt that far higher cortisol concentrations are reached after even more severe and long-lasting stress and are probably maintained for as long as the stress continues at levels considerably in excess of 500630 nmol/L (1823 µg/dL), which is required for a normal response to the short ACTH test. The presence of normal responses to either the short ACTH test or to the insulin tolerance test give us no assurance that a patients serum cortisol concentration will reach and be maintained at the far higher concentrations normally seen after severe, continuing stresses such as multiple traumatic injuries, infections, and burns. The only test that will indicate the capacity of the adrenals to raise serum cortisol concentrations to between 800 and 2200 nmol/L (2980 µg/dL) is measurement of serum cortisol every 2 h during an i.v. infusion of ACTH124, 250 µg over 8 h (21). However, even a normal response of this type does not ensure adequacy of the patients hypothalamic and pituitary capacity to stimulate an increase in serum cortisol concentrations to these levels during severe stress. Until a test becomes available to evaluate the ability of the entire HPA axis to raise serum cortisol far above 630 nmol/L (23 µg/dL), we have no recourse other than to rely on clinical assessment of the severity of stress and an "educated" guess of the adequacy of the measured serum cortisol concentration. When the adequacy of this response is in doubt, our only potentially life-saving response is to prescribe a therapeutic trial of i.v. hydrocortisone, 200 µg/day, by continuous i.v. infusion, or by six-hourly i.v. injections, or of dexamethasone 4 mg six-hourly injections i.v., for at least 34 days, with reduction in the dosage if and when clinical improvement or abatement of the stress occurs.
Experience over the past 45 yr has shown that, when given such steroid therapy, most patients with Addisons disease or those who have undergone bilateral adrenalectomy will survive severe surgical, traumatic, or infective stresses about as well as patients with intact HPA axes, and without recognizable, serious side-effects of the steroids administered. And there is evidence that this clinical, albeit "unscientific" approach, has saved lives in individuals whose partial HPA insufficiency could not be immediately substantiated with laboratory procedures but was later confirmed with a metyrapone test (17) without convincing evidence of any harmful consequences to date. On the other hand, the administration of far larger, perhaps inappropriately excessive doses of glucocorticoids (methylprednisolone 30 mg/Kg every 6 h for 1 day) have not improved the outcome in patients with severe stress and shock in the absence of known HPA failure (22).
Received December 14, 1998.
Accepted December 18, 1998.
![]() |
References |
---|
![]() ![]() ![]() ![]() |
---|