1132 ARTHRITIS & RHEUMATISM Volume 37 Number 8, August 1994, pp 1132-1137 0 1994, American College of Rheumatology

ABNORMAL HYPOTHALAMIC-PITUITARY-ADRENAL AXIS FUNCTION IN RHEUMATOID ARTHRITIS

Effects of Nonsteroidal Antiinflammatory Drugs and Water Immersion

JANE HALL, ERIC F. MORAND, SAM1 MEDBAK, MOLLY ZAMAN, LES PERRY, NICOLAS J. GOULDING, PETER J. MADDISON, and J. PAUL O’HARE

Objective. To investigate the effects of nonsteroi- dal antiinflammatory drug (NSAID) therapy and water immersion on hypothalamic-pituitary-adrenal (HPA) axis function in rheumatoid arthritis (RA).

Methods. Plasma levels of adrenocorticotropic hormone (ACTH) and serum and urine levels of cortisol were compared in untreated RA patients, NSAID- treated RA patients, and healthy control subjects.

Results. ACTH levels were significantly higher in untreated RA patients (mean 2 SEM integrated area 11,377 f 5,246 hours &liter) than in NSAID-treated RA patients (2,285 f 388 hours nghiter) or healthy controls (1,845 f 35.5 hours nghiter) (P < 0.001). Serum and urine cortisol levels were not significantly different between groups. Two-hour head-out water immersion had no effect.

Conclusion. Elevated ACTH levels without hypercortisolemia occur in untreated RA. NSAID ther- apy alters HPA axis response, but immersion has no effect.

Dr. Morand’s work was supported by the National Health and Medical Research Council of Australia. Dr. Goulding’s work was supported by the Arthritis and Rheumatism Council for Re- search, UK.

Jane Hall, M Phil, Grad Dip Phys: Royal National Hospital for Rheumatic Diseases, Bath, UK; Eric F. Morand, MBBS(Hons), FRACP: Monash Medical Centre, Melbourne, Australia; Sami Medbak, MBChB, PhD: St. Bartholomew’s Hospital, London, UK; Molly Zaman, BSc: St. Bartholomew’s Hospital; Les Perry, PhD: St. Bartholomew’s Hospital; Nicolas J. Goulding, PhD: St. Bartho- lomew’s Hospital; Peter J. Maddison, MD, MRCP: Royal National Hospital for Rheumatic Diseases; J. Paul O’Hare, MD, MRCP: Royal United Hospital, Bath, UK.

Address reprint requests to Eric F. Morand, MBBS(Hons), FRACP, Monash Centre for Inflammatory Diseases, Monash Med- ical Centre, Locked Bag No. 29, Clayton, Victoria, 3168, Australia.

Submitted for publication July 6, 1993; accepted in revised form February 17, 1994.

There is considerable evidence to support the hypothesis that the hypothalamic-pituitary-adrenal (HPA) axis contributes to the inhibition of inflamma- tion. For example, adrenalectomy and glucocorticoid antagonists exacerbate inflammation in animal models (1-3). In humans, evidence for the antiinflammatory effects of the HPA axis includes the exacerbation of rheumatoid arthritis (RA) by the administration of metyrapone (4). In addition, HPA axis activity has been shown to be up-regulated in vivo by the proin- flammatory cytokines interleukin-I (IL- 1) and IL-6 (5,6), and the production of these same cytokines is in turn inhibited by glucocorticoids (7,8).

These and other observations have led to a paradigm regarding the role of the HPA axis, which states that the tonic inhibition of inflammation by the HPA axis, and its response to the presence of inflam- mation, act as a negative feedback system (9). It has recently been suggested that dysfunction of the HPA axis may contribute to the development or persistence of inflammatory disease (lo), a hypothesis supported by findings in animal models of RA (3). Few studies of HPA axis function have been performed in humans with RA. Furthermore, there has been considerable variation in the methodology of the published studies, and not surprisingly, the conclusions drawn have been conflicting.

Some studies of HPA axis function in RA have included patients receiving glucocorticoid or nonste- roidal antiinflammatory drug (NSAID) therapy (1 1,12). While it is well known that glucocorticoid therapy may influence HPA axis function, the effects of NSAIDs on HPA axis function are less well described. HPA axis function may also be affected by non-drug interven- tions. Hydrotherapy has been used in the management of musculoskeletal disease for many years, although

HPA AXIS FUNCTION IN RA 1133

experimental evidence for its efficacy remains limited and there is no consensus as to the mechanisms by which it exerts its putative therapeutic effects (13,14). Water immersion per se has been reported to cause a range of physiologic effects, and 2 studies have exam- ined its effects on HPA axis function in healthy sub- jects, again with conflicting results (15,16).

We undertook the present study to investigate HPA axis function in RA, and to examine the effects of NSAID therapy and of immersion. We measured peripheral blood adrenocorticotropic hormone (ACTH) and peripheral blood and urine cortisol levels in pa- tients with untreated RA, NSAID-treated RA patients, and matched healthy control subjects, studied both with and without head-out water immersion.

PATIENTS AND METHODS

Subjects. Age-, sex-, and body mass index-matched patients with RA were recruited from the patient register of the Royal National Hospital for Rheumatic Diseases, Bath, UK. All satisfied the American College of Rheumatology (formerly, the American Rheumatism Association) revised classification criteria for RA (17). Two RA groups were recruited: a group of patients who were receiving daily NSAID therapy and a group of patients who, because of personal preference, were receiving no drug treatment. Disease activity was assessed with the Ritchie articular index (18), and the duration of early morning stiffness was recorded for all patients by a single observer. Plasma vis- cosity and serum C-reactive protein concentration were also recorded. Age, sex, and body mass index-matched healthy control subjects were recruited from among hospital staff. The study protocol was approved by the Bath District Health Authority Research Ethics Committee. Informed consent was obtained from all subjects prior to participation.

Protocol. Urine was collected for the 24 hours imme- diately prior to each subject's first study. All studies began at 9:OO AM. Indwelling venous cannulae were inserted after the application of a topical local anesthetic. Samples of peri- pheral blood were taken at l , 2,3, and 4 hours from baseline. Studies were performed with and without head-out water immersion during hours 2 and 3 of the study, according to the protocol of Epstein (19). During immersion, subjects sat immersed in thermoneutral water (34.5-35°C) to the suprasternal notch. Timed control studies and the first and last hours of the immersion studies were performed with subjects seated adjacent to the immersion tank. To reduce the potential stressor effects of boredom, subjects were invited to watch television.

Study investigations. Serum cortisol was measured by radioimmunoassay (RIA) according to the method of Cun- nah et al (20) and expressed in nmolesfliter. Intra- and interassay coefficients of variation across the working range of the assay (S0-2,OOO nmoles/liter) were <8%. Urinary cortisol was measured by RIA (Coat-A-Count; Diagnostic Products, Los Angeles, CA) in accordance with the manu- facturer's instructions, and expressed in mmoles/24 hours.

Table 1. Characteristics of the 3 groups of subjects*

RA-NS RA-U HC (n = 7) (n = 8) (n = 13)

Age, years 56.9 2 10.3 46.4 f 13.9 53.2 f 12.5 Females/males 611 7/1 11/2 Body mass index 25.0 ? 4.6 22.0 2 2.5 24.5 ? 2.8 Duration of RA, 149.3 f 147.5 107.3 f 122.4 -

months

stiffness, minutes

index

viscosity, mPa

Duration of morning 15.0 f 22.6 31.9 k 39.6 -

Ritchie articular 16.7 f 12.2 10.6 '1 12.9 -

Mean f SEM plasma 1.72 f 0.1 1.69 2 0.1 1.61 0.0

* Except for femaleslmales and plasma viscosity, values are the mean 2 SD. RA-NS = nonsteroidal antiinflammatory drug-treated rheumatoid arthritis patients; RA-U = untreated RA patients; HC = healthy controls.

Plasma ACTH was measured by RIA according to the method of Rees et a1 (21) and expressed in ndliter. The intra- and interassay coefficients of variation across the working range of the assay (10-2,OOO ng/liter) were <lo%.

Statistical analysis. Statistical analyses were per- formed on an Apple Macintosh computer using Statworks and Statview software. Age, sex, and body mass index were compared by Student's t-test. Integrated areas were calcu- lated from the results for the 4 time points studied, as follows (11). The area under the curve for each interval was calcu- lated as the average of the 2 time point results multiplied by the interval (1 hour). These averages were then summed and expressed as hours ng/liter (ACTH) or hours nmole/liter (cortisol). Intergroup variation in plasma ACTH, serum cortisol, and urinary cortisol integrated areas was assessed by Kruskal-Wallis analysis of variance. Values of integrated area obtained during immersion and dry land stages for each study group were compared using the Wilcoxon matched pairs test. The effects of immersion on post-baseline values were assessed using the Friedman test. P values less than 0.05 were considered significant.

RESULTS

Seven RA patients receiving daily oral NSAIDs, 8 patients with untreated RA, and 13 healthy control subjects were studied. Six NSAID-treated RA patients, 4 untreated RA patients, and 12 healthy controls completed both non-immersion and immer- sion studies. The time between immersion and non- immersion studies ranged from 1 week to 36 weeks. Demographic and disease activity data are presented in Table 1. There were no significant differences between groups in age, sex, or body mass index. No significant differences in disease duration or indices of disease activity were found between the NSAID- treated RA and the untreated RA groups. None of the RA patients had received slow-acting antirheumatic

1134

75

50 - loo;

ACTH (no)

25 -

0

HALL ET AL

20000 ACTH integrated area (hr ngn)

15000

t

RA-U ---C RA-NS 1oooo - K

5000 - I lime point 0 subjects

500 - COrtiSOl

: 400 -

300 -

200 -

100

drugs, antidepressants, or glucocorticoids for at least 8 months prior to the study. None of the untreated RA patients or healthy controls were receiving drug ther- apy of any type.

Circulating hormone concentrations were mea- sured at each of 4 time points, and integrated areas calculated. There were no significant changes in plasma ACTH levels during the 4 hours of observation (Figure 1). The mean ? SEM ACTH integrated area in untreated RA patients (1 1,377 * 5,246 hours ngiliter) was approximately fivefold higher than that in either NSAID-treated RA patients (2,285 * 388 hours ng/ liter) or healthy controls (1,845 2 35.5 hours ngiliter) (P < 0.001) (Figure 1). However, serum cortisol concentrations did not reflect the observed differences in plasma ACTH levels. There was no significant

. RA-U - RA-NS

bp-K I I I , time point

difference in cortisol integrated area between the 3 groups (Figure 2). Serum cortisol concentrations tended to be lower in the second half of the experi- mental period, consistent with the diurnal fall in serum cortisol levels, although the changes observed were not statistically significant and were similar in all 3 groups. Similarly, 24-hour urine free cortisol levels were within the normal range in all subjects, and no significant differences between the 3 groups were found (Figure 3).

Immersion had no significant effect on any variable of HPA axis function measured. As was seen in the non-immersion studies, mean ? SEM integrated ACTH was significantly higher in untreated RA pa- tients (12,648 * 3,251 hours ngiliter) than in either NSAID-treated RA patients (2,276 ? 199 hours ng/

Integrated cortlsol

60000 i cortisol integrated area (hr nmolA)

40000

20000

0 subjects HC RA-NS RA-U

Figure 2. Serum cortisol concentrations measured at 4 hourly time points from 9:OO AM, and cortisol integrated areas, in untreated rheumatoid arthritis patients (RA-U), nonsteroidal antiinflammatory drug-treated RA patients (RA-NS), and healthy controls (HC). Serum cortisol concentrations tended to be lower in the second half of the experimental period (P not significant), consistent with the diurnal fall in these levels. There were no significant differences between the 3 groups in either cortisol concentration or cortisol integrated area. Values are the mean ? SEM.

HPA AXIS FUNCTION IN RA 1135

24 hour urine free cortisol

urine free cortisol (micromoles/24 hrs) T vv

40

20

0 subjects HC RA-NS RA-U

Figure 3. Urinary free cortisol excretion measured for 24 hours prior to the other studies, in healthy controls (HC), nonsteroidal antiinflammatory drug-treated rheumatoid arthritis patients (RA- NS), and untreated RA patients (RA-U). All urinary free cortisol excretion values were within the normal range, and there was no significant difference between groups. Values are the mean and SEM.

liter) or healthy controls (1,956 ? 97 hours ng/liter) (P < 0.001). There was no significant difference in integrated cortisol between subject groups, and no significant effect of immersion on individual time point values of ACTH or cortisol was observed.

DISCUSSION

Since the postulation by Munck et a1 that one role of the HPA axis is to limit the response to stress or injury (9), the concept that failure of this system may contribute to the development or persistence of inflammatory disease has gained prominence (10). In RA, evidence for the antiinflammatory role of endog- enous glucocorticoids comes from several observa- tions, including the exacerbation of RA activity by the administration of the 11 P-hydroxylase inhibitor metyrapone (4) and the association between the diur- nal variation of RA activity and the diurnal variation of peripheral cortisol levels (22). The recent demonstra- tion by Sternberg and colleagues of impairment of HPA axis function in arthritis-prone rats (3) has stim- ulated increased interest in the hypothesis that defi- ciencies of HPA axis function may exist in RA.

Abnormalities of serum cortisol diurnal rhythm in patients with active RA have been described (23), but few studies have investigated HPA axis function by comparing RA patients with control subjects. Hill et al reported in 1963 that individuals with RA ex- creted reduced amounts of urinary glucocorticoid me- tabolites in response to ACTH (24). In a more recent

study, Cash et a1 demonstrated increased basal ACTH levels and subtle impairment of the cortisol and ACTH response to exogenously administered ovine corticotropin-releasing factor (CRF) in RA patients (1 1). Interpretation of their results is potentially ham- pered by the fact that all of the RA patients were receiving low-dose glucocorticoids, and many of the observed differences in HPA axis function were most significant during periods of relative glucocorticoid withdrawal. Using insulin-induced hypoglycemic stress induction of HPA axis activity, Struten and Phillips found no difference between cortisol re- sponses in RA patients and controls, but only 4 subjects were included in each group (25). Chikanza et a1 (12) recently reported that RA patients had normal peripheral blood cortisol levels and a normal response to exogenous CRF, but a reduced response to in- creased circulating IL-1 and IL-6 levels induced by surgical stress, compared with healthy subjects and chronic inflammatory disease control groups.

Increased HPA axis activity is a normal re- sponse to the stress of inflammation (9) and may be mediated by central actions of circulating cytokines and other inflammatory mediators (5,6,26,27). Our finding that ACTH levels were greater in untreated RA patients is thus not surprising. The lack of a corre- sponding increase in cortisol levels, however, suggests impairment of the adrenal action of ACTH. Although blood samples were collected only for 4 hours in the morning, similar results were obtained for 24-hour urinary cortisol values, suggesting that the morning values accurately reflect total daily cortisol production.

The significance of the observation of normal cortisol levels in the presence of high ACTH levels is unclear. An impaired response of adrenal cortisol to ACTH may represent a physiologic response to chron- ically elevated ACTH levels. Recent animal studies illustrate that hypothalamic, pituitary, and adrenal responses to acute and chronic stress are highly com- plex (28). Alternatively, there may be a true deficiency of the cortisol response to ACTH. The mechanism of this impairment can only be speculated upon, but could include the presence of circulating inhibitors of ACTH, alterations in adrenal cortex ACTH receptor density, or derangement of a postreceptor event. Nei- ther circulating inhibitors of ACTH nor soluble ACTH-receptor autoantibodies have been described in RA, but their existence would provide a potential explanation of our observations. Further investigation of the mechanisms of the apparent dysregulation of ACTH and cortisol levels in untreated RA patients is clearly warranted.

1136 HALL ET AL

The need to study the effects of NSAID therapy on HPA axis function has recently been emphasized (1 1). In the study by Chikanza et a1 (12), no difference in basal peripheral blood ACTH levels was found between RA patient and healthy control groups, but NSAID therapy was not discontinued in the patients (12). Our finding of a marked difference in ACTH levels between NSAID-treated and untreated RA pa- tients may be explained by the effects of prostaglandin inhibition. Prostaglandins increase the responsiveness of the HPA axis in humans and in experimental models (29,30), while NSAIDs can directly impair the re- sponse of the HPA axis to both prostaglandins (27) and proinflammatory cytokines (3 1). It is unclear whether the neuroendocrine effects of NSAID in humans are primarily mediated centrally (i.e., within the HPA axis) or through the inhibition of circulating pros- taglandins and cytokines. Given the interactions of prostaglandins and proinflammatory cytokines, it is conceivable that the effects of central and peripheral prostaglandins, and hence of NSAIDs, on HPA axis control are altered in the presence of inflammatory disease. The ability of NSAIDs to affect HPA axis function in RA has important implications for the design of future studies in this area, and further specific investigations will be required to determine the level or levels at which NSAIDs act on the HPA axis in humans.

A number of studies have demonstrated that the HPA axis impacts inflammation by mechanisms other than the effects of cortisol. Immuno- and bioactive ACTH and other proopiomelanocortin-derived pep- tides have been shown to be produced by peripheral blood lymphocytes (32,33), and lymphocyte ACTH production is sensitive to stimulation by CRF (34) and bacterial endotoxin (33). The observed elevations in peripheral blood ACTH levels in untreated RA pa- tients could thus conceivably reflect extrapituitary production of this peptide. Some evidence suggests that extrapituitary production of ACTH may be more sensitive to stimulation by eicosanoids than is intra- pituitary production (35). Furthermore, human leuko- cytes bear receptors for ACTH (36), and ACTH has a number of direct immunomodulatory actions in vitro (37). The elevation in ACTH levels we have observed in untreated RA patients may therefore be implicated in the modulation of disease more directly than through the activation of the HPA axis. Furthermore, CRF itself has recently been detected in the synovium of rats with adjuvant- and streptococcal cell wall- induced arthritis (38), and in human rheumatoid synovium (39). Extrahypothalamic CRF production in

RA may thus contribute to dysregulation of HPA axis function by acting on pituitary or extrapituitary sources of ACTH. The difficulty of measuring CRF in peripheral blood samples will continue to make inves- tigation of this question difficult.

Since water immersion has been associated with a range of physiologic effects (40), we hypothe- sized that some of the therapeutic efficacy of hydro- therapy in RA might be attributable to the physiologic effects of immersion per se, mediated through changes in HPA axis function. Coruzzi et a1 reported a de- crease in peripheral ACTH levels in 8 healthy subjects during 2-hour immersion (15), while Epstein et a1 reported no changes in plasma 17-hydroxycorticoster- oid release in healthy subjects during immersion (16). In the current study, immersion was not associated with significant alterations in peripheral ACTH or cortisol concentrations in any group. These findings suggest that it is unlikely that hydrotherapy exerts a therapeutic effect through modulation of HPA axis function.

In conclusion, we have shown that plasma ACTH concentrations are significantly higher in pa- tients with untreated RA compared with healthy con- trols, but that serum and urine cortisol levels fail to reflect this endogenous stimulation. In addition, our finding of significant differences in ACTH concentra- tion in NSAID-treated RA patients suggests that NSAID therapy influences the activity of the HPA axis. This has important implications for the design of future studies of HPA axis function in RA. The appar- ent impairment of the adrenal response to ACTH in RA may represent a mechanism for the putative failure of HPA axis control of inflammation in this disease. Further investigation of this hypothesis, particularly as it relates to early disease, is required.

REFERENCES 1.

2.

3.

4.

5 .

Laue L, Kawai S, Brandon DD, Brightwell D, Barnes K, Knazek RA, Loriaux DL, Chrousos GP: Receptor-mediated effects of glucocorticoids on inflammation: enhancement of the inflammatory response with a glucocorticoid antagonist. J Ste- roid Biochem 29591-598, 1988 Flower RJ, Parente L, Persico P, Salmon JA: A comparison of the acute inflammatory response in adrenalectomised and sham- operated rats. Br J Pharmacol 875742, 1986 Sternberg EM, Hill JM, Chrousos GP, Kamilaris T, Listwak SJ, Gold PW, Wilder RL: Inflammatory mediator-induced hypothalamic-pituitary-adrenal axis activation is defective in streptococcal cell wall arthritis susceptible Lewis rats. Roc Natl Acad Sci U S A 86:2374-2378, 1989 Saldanha C, Tougas G, Grace E: Evidence for anti- inflammatory effect of normal circulating plasma cortisol. Clin Exp Rheumatol4:365-366, 1986 Uehara A, Gottschall PE, Dahl RR, Animura A: Interleukin-1

HPA AXIS FUNCTION IN RA 1137

stimulates ACTH release via an indirect action which requires endogenous corticotropin releasing factor. Endocrinology 12 1 : 1580-1582, 1987

6. Naitoh Y, Fukata J, Tominaga T, Nakai Y, Tamai S, Mori K, Imura H: Interleukin-6 stimulates the secretion of adrenocorti- cotrophic hormone in conscious, freely moving rats. Biochem Biophys Res Commun 155:1459-1463, 1988

7. Kern JA, Lamb RJ, Reed JC, Daniele RP, Nowell PC: Dexa- methasone inhibition of interleukin 1 beta production by human monocytes: posttranscriptional mechanisms. J Clin Invest 81:

8. Nishida T, Nakai S, Kawakami T, Aihara K, Nishino N, Hirai Y: Dexamethasone regulation of the expression of cytokine mRNAs induced by interleukin-1 in the astrocytoma cell line U373MG. FEBS Lett 243:25-29, 1989

9. Munck A, Guyre PM, Holbrook NJ: Physiological functions of glucocorticoids in stress and their relation to pharmacological actions. Endocr Rev 5:25-44, 1984

10. Wilder RL, Sternberg EM: Neuroendocrine hormonal features in rheumatoid arthritis and related conditions. Curr Opin Rheu- matol2:436-440, 1990

1 I . Cash JM, Crofford LJ, Gallucci WT, Sternberg EM, Gold PW, Chrousos GP, Wilder RL: Pituitary-adrenal axis responsiveness to ovine corticotropin releasing hormone in patients with rheu- matoid arthritis treated with low dose prednisolone. J Rheuma-

12. Chikanza IC, Petrou P, Kingsley G, Chrousos GP, Panayi GS: Defective hypothalamic response to immune and inflammatory stimuli in patients with rheumatoid arthritis. Arthritis Rheum

13. Sukenic S, Neumann L, Buskila D, Kleiner-Baumgarten A, Zimlichman S, Horowitz J: Sulphur bath and mud pack treat- ment for rheumatoid arthritis at the Dead Sea. Ann Rheum Dis

14. Danneskiold-Samsde B, Lynberg K, %sum T, Telling M: The effect of water exercise therapy given to patients with rheuma- toid arthritis. Scand J Rehabil Med 19:31-35, 1987

15. Coruzzi P, Ravanetti C, Musiari L, Biggi A, Vescove PP, Novarini A: Circulating opioid peptides during water immersion in normal man. Clin Sci 74:133-136, 1988

16. Epstein M, Fishman LM, Hale BB: Dissociation of aldosterone and 17-OHCS release during water immersion in normal man. Proc SOC Exp Biol Med 138:939-942, 1971

17. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS, Medsger TA Jr, Mitchell DM, Neustadt DH, Pinals RS, Schaller JG, Sharp JT, Wilder RL, Hunder GG: The American Rheuma- tism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 31:315-324, 1988

18. Ritchie DM, Boyle JA, McInnes JM, Jasani MK, Dalakos TG, Grieveson P, Buchanan WW: Clinical studies with an articular index for the assessment of joint tenderness in patients with rheumatoid arthritis. Q J Med 37:393406, 1968

19. Epstein M: Renal effects of head-out water immersion in man: implications for an understanding of volume homeostasis. Phys- iol Rev 58529-581, 1978

20. Cunnah D, Jessop DS, Besser GM, Rees LH: Measurement of circulating corticotroDhin releasing factor in man. J Endocrinol

237-244, 1988

to1 19~1692-1696, 1992

35: 1281-1 288, 1992

49:99-102, 1990

113:123-fi1, 1987 -

21. Rees LH. Cook DM. Kendal JW. Allen CF. Kramer RM. Radcliffe JG, Knight k4: A radioimmunoassay for rat plasma ACTH. Endocrinology 89:254-261, 1971

22. Harkness JAL, Richter MB, Panayi GS, van de Pette K, Unger A, Pownall R, Geddawi M: Circadian variation in disease activity in rheumatoid arthritis. Br Med J 284551-554. 1982

23. Neeck G, Federlin K, Graef V, Rusch D, Schmidt KL: Adrenal

secretion of cortisol in patients with rheumatoid arthritis. J Rheumatol 17:24-29, 1990

24. Hill SR Jr., Ulloa A, Starnes WR, Holley HL: Corticosteroids in rheumatoid arthritis: studies on aldosterone, 17-hydroxycor- ticosteroid, and 17-ketosteroid response to corticotropin in male patients with rheumatoid arthritis. Arch Intern Med 112:603- 611, 1963

25. Struten DHP, Phillips PE: Measurement of the hypothalamic- pituitary-adrenal (HPA) response to hypoglycemic stress in human rheumatoid arthritis (abstract). Arthritis Rheum 34 (suppl9):S156, 1991

26. Dunn AJ, Chuluyan HE: The role of cyclooxygenase and lipoxygenase in the interleukin- I-induced activation of the HPA axis: dependence on the route of injection. Life Sci 51:213-225, 1992

27. Morimoto A, Watanabe T, Morimoto K, Nakamori T, Mu- rakami N: Possible involvement of prostaglandins in psycholog- ical stress-induced responses in rats. J Physiol (Lond) 443:421- 429, 1991

28. Hauger RL, Lorang M, Irwin M, Aguilera G: CRF receptor regulation and sensitization of ACTH responses to acute ether stress during chronic intermitent immobiization stress. Brain Res 532:34-40, 1990

29. Zacharieva S, Borissova AM, Andonova K, Stoeva I, Matoro- zov P: Role of prostaglandin E2 (PGE2) on the coricotropin- releasing hormone (CRH)-induced ACTH release in man. Horm Metab Res 24:336338, 1992

30. Brooks AN, Gibson F: Prostaglandin E2 enhances AVP- stimulated but not CRF-stimulated ACTH secretion from cul- tured fetal sheep pituitary cells. J Endocrinol 132:33-38, 1992

31. Cambronero JC, Rivas FJ, Borelli J , Guaza C: Role of arachi- donic acid metabolism on corticotrophin-releasing factor (CRF)- release induced by interleukin-] from superfused rat hypothal- ami. J Neuroimmunol39:5746, 1992

32. Buzzetti R, McLoughlin L, Lavender PM, Clark AJL, Rees LH: Expression of pro-opiomelanococortin gene and quantifi- cation of adrenocorticotropic hormone-like immunoreactivity in human normal peripheral blood mononucelar cells and lym- phoid and myeloid malignancies. J Clin Invest 83:733-737, 1989

33. Harbour-McMenamin D, Smith EM, Blalock JE: Bacterial lipopolysaccharide induction of leukocyte-derived corticotropin and endorphins. Infect Immun 48:813-817, 1985

34. Smith EM, Momll AC, Meyer WJ 111, Blalock JE: Cortico- tropin releasing factor induction of leukocyte-derived immuno- reactive ACTH and endorphins. Nature 321:881-882, 1986

35. Cudd TA, Wood CE: Prostaglandin E2 releases ovine fetal ACTH from a site not perfused by the carotid vasculature. Am J Physiol 263:R136-R140, 1992

36. Smith EM, Brosnan P, Meyer WJ 111, Blalock JE: An ACTH receptor on human mononuclear leukocytes: relation to adrenal ACTH-receptor activity. N Engl J Med 317:12661269, 1987

37. McGlone JJ, Lumpkin EA, Norman RL: Adrenocorticotropin stimulates natural killer cell activity. Endocrinology 129: 1653- 1658, 1991

38. Crofford LJ, Sano H, Karalis K, Webster EL, Goldmuntz EA, Chrousos GP, Wilder RL: Local secretion of corticotropin- releasing hormone in the joints of Lewis rats with inflammatory arthritis. J Clin Invest 90:2555-2564, 1992

39. Crofford LJ, Sano H, Karalis K, Friedman TC, Epps HR, Remmers E F , Mathern P, Chrousos GP, Wilder RL: Corticotropin-releasing hormone in synovial fluids and tissues of patients wih rheumatoid arthritis. J Immunol 151: 1587-1596, 1993

40. O’Hare JP, Heywood A, Summerhayes C, Lunn G, Evans JM, Walters G, Corral1 RJM, Dieppe PA: Observations on the effects of Bath spa water. Br Med J 291:1747-1751, 1985