Medical hypotheses

Medical Hypotheses
Medical Hypotheses (1994) 42, 183-189 Longman Group Ltd 1994
Mild Adrenocortical Deficiency, Chronic Allergies,
Autoimmune Disorders and the, Chronic Fatigue
Syndrome: A Continuation of the Cortisone Story
W. McK. JEFFERIES
Case-Western Reserve University School of Medicine, Cleveland, Ohio, USA (Correspondence to
WMcKJ, 1208 Bixham Lane, Charlottesville, VA 22901, USA).
Abstract -- The possibility that patients with disorders that improve with administration of large,
pharmacologic dosages of glucocorticoids, such as chronic allergies and autoimmune disorders, might
have mild deficiency of cortisol production or utilization has received little attention. Yet evidence that
patients with rheumatoid arthritis improved with small, physiologic dosages of cortisol or cortisone
acetate was reported over 25 years ago, and that patients with chronic allergic disorders or
unexplained chronic fatigue also improved with administration of such small dosages was reported
over 15 years ago, suggesting that these disorders might be associated with mild adrenocortical
deficiency. The apparent reasons for the failure of these reports to be confirmed or mentioned in
medical textbooks and the facts needed to restore perspective are reviewed, and the need for further
studies of the possible relationship of a mild deficiency of the production or utilization of cortisol and
possibly other normal adrenocortical hormones to the development of these disorders is discussed.
Introduction
pituitary or hypothalamus, might exist has not re- ceived much attention. Also, the possibility that pa- In most endocrine disorders varying degrees of defi- tients with disorders that improve with administration of ciency of hormones are recognized, but deficiency of large, pharmacologic dosages of glucocorticoids, such cortisol is classified only as Addison's disease or as autoimmune disease and chronic allergies, might hypopituitarism, both relatively severe and relatively have mild adrenocortical deficiency apparently has not rare disorders. The possibility that milder degrees of been considered since the introduction of this type of adrenocortical deficiency, either primary in the adre- therapy 40 years ago. At that time methods of studying nals or secondary to inadequate stimulation by the adrenocortical function were limited, and effects that warrant careful study', but still no attempts Kelley and Ely (1) summarized by stating: to extend or confirm these observations appeared. In 1981, with encouragement of two authorities on It appears that neither adrenocortical function nor adrenocortical function, a book was published in which steroid metabolism is completely normal in patients the cortisone story was reviewed and further evidence with connective tissue disease, but the abnormalities was presented of impressive beneficial elfects of small, that exist may be quite subtle and difficult to subreplacement dosages of cortisol or cortisone acetate demonstrate unequivocally. There is as yet no clear in patients with mild adrenal deftciency, either primary demonstration that these abnormalities are of (low adrenal reserve) or secondary to inadequate significance in the pathogenesis of connective tissue stimulation by the pituitary. These beneficial effects disease; however, this remains a distinct possibility that were observed not only in patients with ovarian deserves continuing investigation. dysfunction but also in patients with rheumatoid arthritis, allergic rhinitis, bronchial as- thma, Graves' Since no further evidence for such a possibility was disease, and diabetes mellitus (12). The book was reported, it was concluded that any abnormalities of published by a reputable medical publisher that has adrenocortical function in these disorders were prob- published much of the earlier work on corti- sone, but ably non-specific and unrelated to their etiology. still no subsequent reports of attempts to confirm or Although most clinical experience with glucocorti- extend these studies appeared. The possi- bility that coids, other than in treatment of obvious adrenal defi- some patients with unexplained chronic fatigue (as in ciency in patients with Addison's disease or the chronic fatigue syndrome) might have mild adrenal hypopituitarism, has involved administration of large, deficiency, either primary or secondary, was also pharmacologic dosages with subsequent propensity for suggested. Such a possibility should not be surprising the development of undesirable or even alarming sidg- since fatigue is a characteristic early symptom of effects, beneficial effects of small subreplacement, adrenocortical. insufficiency, but this sug-gestion also safe, physiologic dosages of cortisol were initially apparently received little attention. reported in 1958 in the treatment of women with In the same year, 1981, Poteliakoff(13) reported that ovarian dysfunction and infertility (2). Some had evi- patients with chronic fatigue had lower blood cortisol dence of excess androgen, such as acne or hirsutism, levels than controls matched for age and sex. Still no some had evidence of excess of estrogen, such as attempts to confirm or extend these studies, or even any metropathia hemorrhagica, and some had ovarian dys- function without evidence of either androgen or es- Because one of the most alarming effects of large, trogen, but a majority of women in all groups improved pharmacologic dosages of cortisol or its derivatives is with this therapy (3-9). As experience accumulated, impairment of immunity, and because small, patients with associated allergies or auto-immune physiologic dosages seemed to enhance immunity, a disorders, including allergic rhinitis, bronchial asthma, review of the medical literature on the relationship and rheumatoid arthritis, reported improvement in these between normal adrenocortical function and immunity conditions while taking the small dosages even though was made. This revealed that numerous investigators such small dosages did not produce elevation of blood had reported evidence over the past 40 years that in levels of cortisol above normal. These findings were physiologic amounts cortisol is essential for the presented at the annual meeting of the American development and maintenance of normal immunity in College of Physicians in 1966 and published in 1967 contrast to the well known harmful effects of large, (10), but in subsequent years no reports of attempts by pharmacologic amounts. This review was published in others to confirm or extend them appeared. early 1991 (14), but it also apparently received In 1974, at the invitation of Dr Franz Ingelfinger, a chapter was contributed to Controversies in Internal The report of Demitrack et al later in 1991 (15) of Medicine II entitled, Glucocorticoid Therapy: An evidence of low levels of cortisol in the blood of Overmaligned Reputation With Untapped Potential patients with the chronic fatigue syndrome without any Benefit (11). In this chapter it was reported that, in attempt to treat these patients with a physiologic dosage addition to anti-allergic and anti-rheumatic effects, the of cortisol calls further attention to this unique situation administration of small, safe, physiologic dosages of in which a promising therapeutic approach to relatively cortisol 'appears to have anti-malaise and anti-fatigue common clinical problems apparently is overlooked or Background
dosages was impairment of immunity, causing patients to become more susceptible to infections. This property How could such a situation occur? It has apparently has been used to help patients to tolerate tissue resulted from a unique combination of factors, chief of transplants, and hence has become quite widely known. It has even been suggested that the increased production of cortisol that occurs at the onset of infection may 1. At the time of the discovery of the dramatic clini- serve to limit the immune reaction from overshooting cal effects of cortisol in autoimmune disorders and and hence would be consistent with the anti-immune allergies in 1949 and the early 1950s the normal pro- effects of pharmacologic dosages (16), but a more likely duction rate of cortisol was not known, nor were an explanation of this increased production is that of Ingle optimum dosage, route or schedule of administration. It was found that a dosage of 100-300 mg cortisol intramuscularly, and laler orally, daily was necessary to The increased secretion of adrenal hormones serves produce symptomatic benefit within 24 - 48 h. Later, to meet an increased need during stress and tends to when the normal production rate under unstressed maintain homeostasis rather than to disturb it. The conditions was found to be 18-20 mg daily, it was increased secretion does not cause a state assumed that large, pharmacologic dosages were ofhypercorticism such as develops when the titer of necessary to produce the dramatic therapeutic effects. these hormones is increased artificially in the The possibility that smaller dosages might produce beneficial effects more slowly but more safely was Evidence that cortisol impairs immunity only in large, pharmacologic dosages and that in physiologic amounts 2. Because serious side effects often occurred with this hormone is essential for the development and large dosages of the natural glucocorticoids, deriva- maintenance of normal immunity has been reported by tives of cortisoI or cortisone such as prednisone, pred- investigators over the past 40 years, but largely nisolone, triamcinolone, and dexamethasone were overlooked, as noted above (14). Most physicians are introduced, but, except for less tendency to retain salt still not aware of this important difference between and water, they possessed the potential to produce all of physiologic and pharmacologic effects of cortisol. the other undesirable side-effects of large dosages of the 7.Before patients are given other hormones, tests of the function of the glands that produce those hormones 3. When patents expired on cortisone and cortisol, the are usually performed, but tests of adrenocortical func- more potent derivatives, whose patents persisted, were tion are seldom made on patients before the adminis- promoted more vigorously and the natural hormone tended to be forgotten. Package inserts no longer differentiated between physiologic and pharmacologic dosages of cortisol and it was implied that all of the Present status
grim side-effects might develop at any dosage level. The tendency by some to term all glucocorticoids These factors have led to a unique situation in which a 'cortisone' added to the confusion. normal hormone, one that is essential for life, has developed such a bad reputation that many physicians 4.Most physicians practicing today are therefore under and patients are afraid to use it under any circumstan- the impression that any dosage of cortisol can produce ces. In order to restore perspective, the following facts any of the serious side-effects that occur only with administration of large pharmacologic dosages of this 1.Cortisol (hydrocortisone) is a normal hormone. Cortisone is converted to cortisol after absorption and 5. Reports documenting the safety and effectiveness hence has similar effects, provided there is no inter- of physiologic dosages of cortisol were published in ference with this conversion. The more frequently used reputable medical journals over 25 years ago (2-10), but derivatives of cortisol or cortisone, such as prednisone, computerized reviews of the medical literature, such as prednisolone, triamcinolone, methylprednisolone, and Medline, do not yet cover publications that remote, so dexamethasone, have from 4-30 times the anti- few physicians today are aware of the existence of these inflammatory and glucocorticoid effects of corti- 6.One of the most alarming effects of pharmacologic sol, but, except for less sodium retention than the full replacement dosage, there appears to be no sum- normal hormone, they have equally serious potential mation effect beyond the reaching of an optimum level side-effects, and they are not produced by human or since patients receiving such dosages have not de- veloped hypercortisolism. If they receive a full re- placement dosage for a prolonged period, however, 2.Cortisol is essential for life in humans; its most their adrenals might be suppressed sufficiently to im- obvious effect is to promote gluconeogenesis to pro- pair further their resistance to stress. Patients with vide energy and avoid hypoglycemia in times when inadequate stimulation from the hypothalamus or pi- food intake is limited, but it also helps to protect against tuitary have improved with subreplacement dosage, other stresses including the maintenance of normal but because they do not have proper central control of production of cortisol, optimum treatment for this type 3.In the unstressed state, cortisol is normally pro- duced on a diurnal pattern depending upon the sleep- 8. Because a single intravenous injection of cortisol wake schedule, the highest blood levels occurring after produces metabolic effects that last only about 8 h (20), an optimum schedule for administration of physiologic dosages orally for persistent effects probably should be 4.Cortisol is a very dynamic hormone, with produc- at intervals of 8 h or less. Although medical texts tion and blood levels fluctuating rapidly from minute to recommend two or three times daily schedules of minute, depending upon the degree of stress as well as administration of cortisol to patients with adrenal deft- upon diurnal variation. Because utilization as well as ciency, a three times daily schedule has been found to production varies with stress, blood levels may not maintain afternoon blood levels of cortisol better than a always reflect the degree of stress or rate of produc- two times daily schedule (21), and a four times daily schedule suppresses excessive adrenal androgen ex- cretion better than a two times daily schedule (9). In the 5.To evaluate adrenocortical function, blood ACTH, treatment of adrenal deficiency in our practice, a cortisol and a cosyntropin stimulation test under basal schedule of four times daily before meals and at bed- conditions are helpful. These studies should be time has been found to have several advantages, in- performed on blood specimens drawn preferably in the cluding ease of adherence, less tendency to produce morning before breakfast on patients on a normal sleep- acid indigestion and, in dosages of 5 mg four times wake schedule and not receiving any medication that daily or less, failure to block normal diurnal variation might affect cortisol levels or adrenocortical function. (22). Taking milk or an antacid with the bedtime dose 24h urinary free cortisol also contributes to the helps to avoid acid indigestion in susceptible patients. evaluation of disorders of adrenocortical function For a totally adrenalectomized patient under un- provided collections are made properly. One or more of stressed conditions, a dosage of 10 mg four times daily these tests has been found to be abnormal in at least is adequate, and for lesser degrees of adrenocortical some patients with rheumatoid arthritis, chronic deficiency dosages of 7.5 mg, 5 mg, or 2.5 mg four allergies, or unexplained chronic fatigue. times daily are satisfactory maintenance dosages, de- pending upon the degree of deficiency. Since over 2000 6.Normal ranges for blood cortisol levels and other patient years of experience have been accumulated with tests of adrenal function have been determined on such physiologic dosages (14), therapeutic trials with subjects who did not have obvious adrenocortical ex- such dosages of cortisol would seem to be indicated in cess (Cushing's syndrome), or deficiency (Addison's patients with chronic fatigue syndrome. Because these disease), or panhypopituitarism, or any other apparent studies involve treatment of a hormonal deficiency, the illness, and are rather broad. Hence they may include normal hormone, not any of its derivatives, should be patients with mild deficiency or excess of cortisol. Also it must be remembered that resistance to cortisol may occur because of a defect in receptor function (18,19), 9.As with patients with severe adrenal deficiency, if a so blood cortisol levels in the normal or even patient with mild adrenal deficiency has evidence of an supranormal range do not exclude the possibility of active inflammatory process or infection, a larger symptoms associated with deficiency of cortisol ef- dosage of cortisol, up to 20 mg four times daily, in conjunction with a suitable antibiotic or other type of therapy, is advisable, but when this condition is under 7. If cortisol is administered to patients with mild control, the dosage of cortisol should be tapered to the primary adrenal deficiency in an amount less than a maintenance dosage (between 2.5-7.5 mg four times autoimmune disorders might have mild adrenocortical daily) over 2-4 days. Because cortisol is such a dy- deficiency should not be surprising since cortisol is the namic hormone, with increased production in times of only known substance produced by the body that stress, caution must be exercised not only to provide counteracts the symptoms of both of these disorders. adequate amounts during the stress but also to avoid Furthermore, it is well know that stress, either physical continuing the larger dosage too long, in which case or emotional, often precedes the onset or exacerbation hypercortisolism and hypokalemia with toxic effects of symptoms of allergies, autoimmune disorders, or the may occur. Hypokalemia may also occur if cortisol is chronic fatigue syndrome, and the adrenals are a major administered too rapidly intravenously, so, for parent- component of the body's defense against stress. A eral administration the intramuscular route or a slow deficiency in the adrenals' response to stress might intravenous drip of 100 mg or less in normal saline over therefore in some way contribute to the development 8 h is safer. As with any medication, the admin- and progress of these disorders. Studies are also needed istration of larger dosages must be employed judi- to determine why under increased stress one person might develop chronic allergies, another an autoimmune disorder, and another unexplained chronic fatigue. The 10. During acute exacerebations, patients with aller- familial tendency for the occurrence of these disorders gies or autoimmune disorders need to take larger dos- suggests inherited factors of susceptibility. ages of cortisol (20 mg four times daily is usually The administration of any medication three or four adequate), but as soon as symptoms are under control, times daily might be considered too difficult for patients the dosage should be tapered to a physiologic main- to follow, but this has not been a problem with patients tenance dosage. If this is not possible without a return with mild adrenal deficiency. Their subjective of symptoms, a careful search should be made for improvement has been sufficient to keep most patients obscure infection or other persistent source of stress that taking their medication regularly. Taking cortisol three is preventing the return to a safe maintenance dosage. times daily, or even twice daily, will often produce some improvement, but for optimum benefit the four 11. Patients with mild primary adrenal deficiency times daily schedule has been more helpful and easier to treated with subreplacement dosages of cortisol may follow. Because the bedtime dosage of cortisol tends to have slightly enhanced resistance to additional stress keep the kidneys functioning during the night, some because the supplementary steroid has relieved some of patients prefer to take a smaller dosage at bedtime, e.g. the strain on their impaired adrenals and thereby 2.5 mg, to avoid nocturia. Noticeable improvement provided slightly increased reserve capacity, but they usually occurs within a few hours of the first dose, and seem to tolerate major stresses such as surgery or patients often describe a return of symptoms within a infections better if they are given supplementary cor- few hours of a missed dose. Occasionally improvement tisol just as patients with severe adrenal deficiency. is not noticed until 10-14 days after treatment is begun, Patients with adrenal deficiency secondary to inade- presumably because of a more severe underlying quate stimulation from the pituitary or hypothalamus disorder, so patients should be cautioned regarding this have been treated in a similar fashion with apparently possibility. After a remission occurs, if cortisol is satisfactory results, but further studies of optimum stopped, a return of symptoms may develop after treatment of this type of disorder, as well as factors that varying intervals, sometimes as long as several years. contribute to its development, are advisable. Compliance is helped by giving patients printed instructions, briefly describing the reasons for the 12. Since subreplacement dosages of cortisol do not schedule, what to do if a dose is missed, and, in addition produce an excess of glucocorticoid, they have not to contacting their physician, what to do if they develop tended to promote the development of osteoporosis or a respiratory infection, G-I upset, influenza, or other any other undesirable side-effects that can result from illness (22). Patients have been treated with this administration of large, pharmacologic dosages. If schedule of cortisol or cortisone acetate for as long as patients have a deficiency of estrogen or androgen, 40 years without significant problems. Because some of replacement with physiologic dosages of these normal these patients had ovarian dysfunction and infertility, and because continuation of small, physiologic dosages helped to protect against miscarriages, over 200 babies Discussion
The possibility that patients with chronic allergies or women who continued these physiologic dosages might in turn result in improved adrenocortical func- through their pregnancies and sometimes during their tion, the possibility that the placebo effect might result postpartum (including nursing) periods with no evi- from improved adrenocortical function should be con- dence of harm to mothers or babies (22,23). These findings were published, but the bad reputation The only significant problem that has developed of glucocorticoid therapy and the US Food and Drug related to therapy with physiologic dosages of cortisol Administration (FDA) regulation that requires that arises not from the cortisol itself, but from the filler that when a new use of a medication whose patent has is used in making the commercial tablets. Most, if not expired is found, it should be treated as if it were a new all tablets of cortisol or cortisone acetate contain lactose medication has resulted in these reports receiving little and cornstarch in the filler. In large, pharmacologic attention. Such a regulation is obviously desirable, but dosages, the quantity of cortisol is apparently sufficient without the protection of a patent, pharmaceutical to protect against lactose intolerance or allergy to corn, companies have no incentive to undertake the exten- but small, physiologic dosages are not adequate to sive and expensive studies necessary to determine protect against such sensitivities in some patients, and a whether a medication is effective and safe for its few have developed mild skin rashes or other evidence recommended use for the general public. Without such of a mild allergic reaction to the filler of the tablets. studies, new uses cannot be promoted or advertised by Such patients may need to take a pediatric liquid the manufacturer, and without promotion or preparation of cortisol or capsules prepared with a non-advertisement, a different therapeutic approach, espe- allergenic filler. Because the pediatric preparation is too cially one employing a medication that has achieved sweet for general adult use and specially prepared such a bad reputation as cortisone, is severely handi- capsules are expensive, if these small physiologic capped. It must be remembered, however, that 'The dosages are more widely used, it is hoped that FDA cannot approve or disapprove of how a legally pharmaceutical companies will prepare tablets or marketed drug is used by a physician in his practice. capsules that do not contain lactose, apparently the The agency approves of what the manufacturer may more common offender, or cornstarch, or any other recommend about uses in its labeling (package insert) and advertising' (24). In other words, the physician has There is therefore no reason to fear that physiologic the ultimate responsibility of judging the suitability of a dosages of cortisol will produce any of the harmful medication for his or her patient regardless of whether it side-effects of phannacologic dosages, and, in subjects is patented or whether its use is listed on the label or with mild adrenal deficiency, either primary or secondary, who are not allergic to lactose or cornstarch, Further studies to elucidate these potentially important therapeutic trials with the normal hormone, cortisol, in uses of this normal hormone are obviously advisable, physiologic amounts at proper intervals, should be and it might be questioned whether double-blind, made. Instead of the current custom of prescribing placebo studies should be used for these. Such studies pharmacologic dosages ofglucocorticoids on an were initiated when the use of physiologic dosages of empirical basis, if tests of adrenocortical function are cortisol were found to be helpful in patients with some made prior to initiating glucocorticoid therapy, mild types of ovarian dysfunction (5), but the objective, as degrees of adrenocortical deficiency might be identified well as subjective, benefits were so clear, plus the that might be better treated with safe, physiologic finding that optimum dosage requirements varied within dosages of the normal hormone, cortisol, a treatment the physiologic range from patient to patient and that could be, and possibly should be, continued sometimes from day to day in the same patient, that this indefinitely, rather than being discontinued as soon as a type of study was soon abandoned. Studies of the remission occurs. If this is done, it is possible that effects of other normal hormones have not required, or many, if not all, patients with chronic allergies and even included, double-blind placebo studies, probably at autoimmune disorders will be found to have mild least partly for these reasons. It should also be noted adrenocortical deficiency and hence would benefit from that the cause of the placebo effect has never been persistent administration of safe, physiologic dosages of elucidated, yet double-blind, placebo studies continue to normal adrenocortical hormone instead of being treated be considered essential by many for the evaluation of spasmodically with pharmacologic dosages of synthetic new therapies. Because the placebo effect is apparently related to expected improvement by patients, and It should also be remembered that, in addition to because expected improvement implies a type of cortisol, the human adrenal cortex produce aldosterone, diminished stress, which androgens (chiefly dehydroepiandrosterone [DHEA] and androstenedione) and estrogens (estrone and estradiol). Although effects of aldosterone and estrogens have been studied, the effects of physiologic dysfunction: The endocrine glands other than the dosages of DHEA and androstenedione in human sub- gonads. Clin Obstet Gynec 1965; 8: 73-90. jects are largely unknown. The manner in which these 9. Jefferies WMcK. Glucocorticoids and ovulation. In: adrenal hormones contribute to the welfare of the Greenblatt RB, ed. Ovulation. Philadelphia: individual and possibly to protection against stress therefore also needs further study, especially since a 10. Jefferies WMcK. Low dosage glucocorticoid low or absent excretion of DHEA has been found in therapy. Arch Int Med 1967; 119: 265-278. some patients with rheumatoid arthritis (10). Such 11. Jefferjes WMcK. Glucocorticoid therapy: An studies should be made with physiologic dosages on overmaligned reputation with untapped potential subjects who have a demonstrated deficiency of the benefit. In: Ingelfinger FJ, Ebert RV, Finland M, Relman AS, eds. Controversies in Internal Medicine II Philadelphia: WB Saunders, 1974: 439-445. Conclusions
12. Jefferies WMcK. Safe Uses of Cortisone. With the evidence that at least some patients with 13. Poteliakoff A. Adrenocortical activity and some chronic allergies, autoimmune disorders and unex- clinical findings in acute and chronic fatigue. J plained chronic fatigue, including the 'chronic fatigue syndrome', have mild adrenocortical deficiency, fur- 14. Jefferies WMcK. Cortisol and Immunity. Med ther studies of the above therapeutic approach seem indicated. Such studied will hopefully no longer be 15. Demitrack MA, Dale YK, Strus SEet al. Evidence handicapped by misconceptions that have resulted for impaired activation of the hypothalamic- largely from the unique combination of factors that have pituitary-adrenal axis in patients with chronic fatigue syndrome. J Clin Endocrinol Metab 1991; 75: 1224- References
16. Munck A, Guyre M, Holbrook NJ. Physiological functions of glucocorticoids in stress and their 1. Kelley VC, Ely RS. Production and metabolism of relation to pharmacologic actions. Endocr Rev 1984; adrenocorticosteroids in connective tissue disease. 17. Ingle DJ. Some further studies of the reladonship of 2. Jefferies WMcK, Weir WC, Weir DR, Prouty RL. adrenal cortical hormones in experimentaI diabetes. The use of cortisone and related steroids in irfertility. 18. Vingerhoeds ACM, Thijssen JHH, Schwartz FJ. 3. Jefferies WMcK, Levy RP. Treatment of ovarian Spontaneous hypercortisolism without Cushing's dysfunction with small doses of cortisone or hydro- Syndrome. J Clin Endocrinol Metab 1976; 43:1128- cortiscne. J Clin Endocrinol Metab 1959; 19: 1069- 19. Chrousos GP, Vingerhoeds ACM, Brandon D et al. 4. Jefferies WMcK. Further experience with small doses Primary cortisol resistance in man. J Clin Invest of cortisone and related steroids in infertility associated with ovarian dysfunction. Fertil Steril 20.Jefferies WMcK, Kelly Jr LW, Syndor KL, Levy RP, Cooper GW. Metabolic effects of a single 5. Jeffcries WMcK. Effect of small dosages of cortisone intravenous infusion of hydrocortisone related to upon urinary 17-ketosteroid fractions in patients with plasma levels in a normal versus an adrenally ovarian dysfunction, J Clin Endocrinol Metab 1962; insufficient subject. J Clin Endocrinol Metab 1957; 6. Jefferies WMcK. Effects of low dosage steroid 21. Groves RW, Toms GC, Houghton B J, Monkson JP. therapy on 17-ketosteroid fractions in infertility. Corticosteroid therapy: twice or thrice daily? J R Soc 7. Jefferies WMcK, Michelakis AM. Individual patterns 22. Jeffcries WMcK. Safe Uses of Cortisone. of urinary 17-ketosteroid fractions. Metabolism 1963; 23. Jeffcries WMcK. Safe Uses of Cortisone. 24. Archer JD. The FDA does not approve uses of drugs. (Editorial). JAMA 1984; 252: 1054-1055.

Source: http://www.members.westnet.com.au/pkolb/med_hyp2.pdf