Assessment of hypothalamic-pituitary-adrenal axis function after corticosteroid therapy in inflammatory bowel disease

THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 2002 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.

Vol. 97, No. 7, 2002 ISSN 0002-9270/02/$22.00 PII S0002-9270(02)04142-4

Assessment of Hypothalamic-Pituitary-Adrenal Axis Function After Corticosteroid Therapy in Inflammatory Bowel Disease J. Desrame´, M.D., J. M. Sabate´, M.D., Ph.D., R. Agher, Ph.D., C. Bremont, M.D., M. Gaudric, M.D., D. Couturier, M.D., Ph.D., and S. Chaussade, M.D., Ph.D. Services d’He´pato-gastroente´rologie and d’Endocrinologie, Hoˆpital Cochin; and De´partement de Recherche Clinique, Groupe Hospitalier Pitie´-Salpe´trieˆre, Paris, France

OBJECTIVES: In patients with inflammatory bowel disease (IBD), little is known about the effect of long term corticosteroid therapy (CT) on the hypothalamic-pituitary-adrenal (HPA) axis function. Our aim was to assess HPA axis function in IBD, before the end of CT, during the tapering phase. METHODS: HPA axis function was assessed with cortisol (ng/ml) measurement before (T0) (normal ⬎ 100) and 60 min (normal ⬎ 210) after 0.25 mg tetracosactide (Synacthen imme´diat) injection (T60) in 55 consecutive cases of IBD attacks. Abnormal response was defined as a T60 ⬍210. The attacks were separated into two groups according to the result of the Synacthen test (ST). RESULTS: In all, 36 of 55 ST were abnormal. The time for recovery normal HPA axis function was 7.2 ⫾ 1.3 months. Duration of disease since onset, past history of surgical or immunosuppressive treatment, severity and extension of the attack, need for surgical or immunosuppressive treatment, total cumulative and mean daily corticosteroid dose, total duration of CT, and steroid dose at the time of ST were not significantly different in the two groups. In multivariate analysis a past history of CT was predictive of abnormal ST (OR ⫽ 8.4, 95% CI ⫽ 2.2–31.5, p ⫽ 0.0009). Among patients with a past history of CT, the time (months) elapsed between the last course of CT was significantly longer in those with normal ST than in those with abnormal ST (45.5 ⫾ 13.5 vs 15.4 ⫾ 6.0; p ⫽ 0.02), and in multivariate analysis a duration free of CT ⬍15 months was predictive of abnormal ST (OR ⫽ 15.00, CI ⫽ 1.23–183.00, p ⫽ 0.03). CONCLUSIONS: In all, 65% of the ST were abnormal. These results suggest that ST should be performed before corticosteroid withdrawal, especially in patients with recent past history of CT. (Am J Gastroenterol 2002;97:1785–1791. © 2002 by Am. Coll. of Gastroenterology)

J.D. and J.M.S. contributed equally to this article.

INTRODUCTION Corticosteroid therapy (CT) remains the mainstay of acute treatment for moderate-to-severe inflammatory bowel disease (IBD) attacks (1, 2). One limitation of long term steroid therapy is the frequency of different side effects. Glucocorticoids have a suppressive action on the hypothalamic-pituitary-adrenal (HPA) axis by the mean of negative feedback. This suppression of corticotropin (ACTH) and adrenal steroid secretion can produce adrenal insufficiency (3–9). The recovery of HPA axis function is usually observed during the tapering phase but sometimes persists at the end of treatment (10, 11). Some studies tried to evaluate the frequency of this complication in patients receiving long term steroid therapy, the factors that could explain it, and the time necessary to recover normal HPA axis function; these studies showed contradictory results (12–16). In IBD patients, little is known about the effect of long term glucocorticoid therapy on HPA axis function. Our aim was to assess adrenal function of these patients without clinical adrenal insufficiency, before the end of steroid therapy, during the tapering phase, in a retrospective study and to look for risk factors of this side effect.

MATERIALS AND METHODS Selection of Patients All patients with IBD who were treated with CT and followed by one physician of the hepato-gastroenterology unit of Cochin Hospital between January, 1997, and January, 1999, were studied retrospectively. For this study, only one attack per patient was evaluated to avoid statistical bias. If, during the period of the study, a patient had more than one attack treated by CT, only the most recent attack was evaluated and the others were excluded. All the patients had a Synacthen test (see later here) during the tapering phase of the treatment, before the withdrawal of CT, in the absence of clinical signs of adrenal insufficiency. The diagnosis of IBD was done on a clinical, endoscopic, and histological basis. Crohn’s disease (CD) activity was measured with the Crohn’s Disease Activity Index (CDAI) (17) and Truelove

1786

Desrame´ et al.

criteria were used for ulcerative colitis (UC) (18). The standard dose of glucocorticoid was the one developed by the Groupe d’Etude The´ rapeutique des Affections Inflammatoires Digestives (1). Patients were started on p.o. CT (prednisolone or prednisone) at 1 mg/kg body weight in single daily dose. In patients with clinical remission (CDAI ⱕ150), after 3 wk of treatment at the initial dosage, a tapering schedule was started as follows: decrease in daily dose of 10 mg every 10 days until half dose was reached; then decrease of 5 mg every 10 days to a daily dose of 15 mg; then a 2.5-mg decrease every 10 days to a daily dose of 10 mg; then a 0.5-mg decrease every 7 days to a daily dose of 7.5 mg. Below these dosages, either the tapering schedule was carried on with a 0.5 mg decrease every 7 days until the Synacthen test (ST) was performed or prednisolone or prednisone was replaced by hydrocortisone at a dose of 20 or 30 mg daily and the Synacthen test was performed 7 days after. If a clinical relapse was observed during the tapering schedule, the dose of glucocorticoid was increased and, after a new remission, the tapering schedule was started again. Synacthen Test The ST was performed either when a daily dose ⱕ7.5 mg of prednisolone or prednisone was reached, or when the patient was treated by a hydrocortisone dose of 20 or 30 mg daily. If a surgical procedure was performed during the course of disease, prednisolone or prednisone was replaced by hydrocortisone; the ST was then performed when the daily dose was ⬍30 mg of hydrocortisone. If immunosuppressive treatment was given during the attack of IBD, the ST was performed under the same conditions. For all patients, the ST and plasma cortisol dosages were performed in the laboratory of endocrinology department of Cochin hospital as follows: the test was performed after an overnight fast and after the usual morning dose of exogenous steroid therapy was withheld. An i.v. catheter was inserted, after which a baseline cortisol level was obtained (T0). Then 0.25 mg of B1–24 corticotrophin (Te´ tracosactide ⫽ Synacthe`ne imme´ diat, Novartis Pharma, Rueil-Malmaison, France) was injected. Another sample for cortisol was collected 60 min after the injection. The cortisol levels were determined by radioimmunocompetition on transcortine using the Murphy technique (19). This technique has low cross-reactivity with other glucocorticoids, with half lives of hydrocortisone, prednisone, and prednisolone 90, 60, and 200 min, respectively, 24 h after the last dose of exogenous steroid. The cortisol level determined by the test is only the expression of endogenous secretion (20). The ST results were considered to be pathological if the T60 cortisol level was ⬍210 ng/ml, whatever the T0 cortisol level, normal values of which were between 100 and 200 ng/ml (to convert plasma cortisol values concentrations from ng/ml to nmol/L, multiply by 2.759). The methodology of our ST has been validated previously in the endocrinology unit of our hospital (21). In patients with pathological test results, a control was performed every 2 or 3 months until recovery.

AJG – Vol. 97, No. 7, 2002

Attack Assessment For each attack, the following data were collected: sex, age, body weight, height, type of IBD (CD or UC), duration of disease since onset, number of previous attacks, past history of surgical treatment, corticosteroid or immunosuppressive therapy, CDAI or Truelove score before treatment and CDAI score performed at the time of the test, extension of disease assessed by endoscopy including segmental data (ileum, ileum and colon, colon), and need for surgical procedure or immunosuppressive treatment during the attack. Concerning the corticosteroid treatment, the total cumulative dose (mg), the mean cumulative dose (in milligrams/ day), the dose at the time of the ST (mg) and the total duration of CT (days) has been calculated. The drugs used were prednisone, prednisolone, methylprednisolone, and hydrocortisone. The doses are given as prednisone equivalent doses (1.0 mg methylprednisolone, 1.0 mg prednisolone, and 1.0 mg hydrocortisone equivalent to 1.2 mg, 1.0 mg, and 0.25 mg prednisone, respectively) (20). The time (in months) elapsed since the last CT and, for patients with pathological test results, the time for recovery of normal values was also noted. Patients were separated in two groups according to the ST results: those with normal ST and those with pathological ST. Finally, the same analysis has been performed after stratification according the kind of corticosteroid treatment at the time of the ST. Statistical Analysis We used SAS software, version 6.11 (SAS Institute, Cary, NC) for statistical analysis. Results were expressed as mean ⫾ SEM. The two groups of patients were compared with each other using Pearson’s ␹2, Pearson ␹2 with Yates correction, and Mann-Whitney test as appropriate. Pearson’s correlation coefficient (r) was calculated with linear regression analysis. A p value ⬍0.05 was considered to be statistically significant. A power analysis was performed for the comparison of factors related to CT (total dose, duration, mean cumulative dose). Risk factors associated with pathological ST results were determined by multivariate logistic regression analysis for all patients (n ⫽ 55) and for patients with a past history of CT (n ⫽ 37). Initially, all parameters were analyzed using univariate logistic regression. In a second step, variables that reached the 20% significance level in univariate analysis were considered for multivariate logistic regression analysis (enter and remove limits 0.2 and 0.05, respectively). It was decided to include the variable “mean cumulative corticosteroid dose per day” into multivariate analysis irrespective of the result of univariate analysis for this variable because of data in the literature (9, 11, 13).

RESULTS Between December, 1996, and December, 1998, for the 55 consecutive patients with IBD who were treated by CT and followed by one physician of the hepato-gastroenterology

AJG – July, 2002

HPA Axis and Steroid Therapy in IBD

1787

Table 1. Descriptive Data About 55 IBD Attacks: Epidemiology and Past History

Variable

Normal Synacthen Test

Pathological Synacthen Test

p

Sex (male/female) Age, yr (mean ⫾ SEM) Weight, kg (mean ⫾ SEM) Height, cm (mean ⫾ SEM) Crohn’s disease (N) Ulcerative colitis (N) Duration of the disease since onset, mo (mean ⫾ SEM) Number of previous attacks (mean ⫾ SEM) Past history of surgical treatment (N) Past history of immunosuppressive therapy (N) Past history of corticosteroid therapy (N) Time elapsed since the last corticosteroid therapy, mo (mean ⫾ SEM)

(N ⫽ 19) 7/12 31.5 ⫾ 2.4 60.0 ⫾ 3.0 168.5 ⫾ 1.9 16 3 62.8 ⫾ 19.5 1.5 ⫾ 0.4 4 2 7 45.5 ⫾ 13.5

(N ⫽ 36) 12/24 34.3 ⫾ 2.5 61.0 ⫾ 2.2 167.0 ⫾ 1.6 29 7 82.6 ⫾ 15.4 2.3 ⫾ 0.3 7 8 30 15.4 ⫾ 6.0

NS NS NS NS NS NS NS NS NS NS 0.0004 0.02

unit of Cochin Hospital, 62 attacks treated with CT have been registered in the data base. A total of 55 attacks for 55 patients are included and seven attacks excluded. The main characteristics of the 55 attacks according to the result of ST are displayed and compared in Tables 1 and 2. There were no statistically significant differences between the two groups of patients in term of sex, age, body weight, height, type of IBD (CD or UC), duration of disease since onset, number of previous attacks, past history of surgical treatment or immunosuppressive treatment therapy, CDAI or Truelove score before treatment and CDAI performed at the time of the ST, extent of disease as assessed by endoscopy, and need for surgical procedure or immunosuppressive treatment during the attack. The past history of corticosteroid treatment was more frequent in the group with patho-

logical ST results: 83% (30 of 36) versus 40% (seven of 19), p ⫽ 0.0004. The total dose, mean cumulative daily dose, dose at the time of the ST, and duration of CT were not statistically significantly different between the two groups and were not correlated with T0 and T60 cortisol levels. In contrast, there was a strong correlation between T60 and T0 cortisol level (r ⫽ 0.8, p ⬍ 10⫺5). The power (1 ⫺ ␤) for the comparison of the total dose, mean cumulative dose, and duration of CT between the two groups was 73%, 87%, and 99%, respectively. In the subgroup previously treated with corticosteroids, the mean time (months) elapsed since the withdrawal of the last corticosteroid treatment was significantly lower for patients with pathological ST results than for those with normal ST results (15.4 ⫾ 6.0 vs 45.5 ⫾ 13.4, p ⫽ 0.02). The mean duration for a normal ST recovery

Table 2. Descriptive Data About 55 IBD Attacks: Data Concerning the Last Attack

Variable Crohn’s disease: CDAI at begining of attack (mean ⫾ SEM) Crohn’s disease: CDAI at time of Synacthen test (mean ⫾ SEM) Ulcerative colitis: Truelove score (minor/moderate/severe) Extent of disease Ileum (N) Ileum and colon (N) Colon (N) Duration of corticosteroid treatment (mean in days ⫾ SEM) Total cumulative dose of corticosteroid (mean in mg ⫾ SEM)* Mean cumulative dose of corticosteroid per day (mg/j ⫾ SEM)* Corticosteroid treatment at the time of Synacthen test Hydrocortisone (N) Prednisolone (N) Prednisone (N) Dose of corticosteroid at the time of Synacthen test Mean (mg/j ⫾ SEM)*/95% CI/extremes Number of surgical procedure during the attack (N) Number of immunosuppressive treatment during the attack (N) Time for recovery normal Synacthen test (mean in mo ⫾ SEM)† * Data available for 25 of 37 patients with pathological Synacthen test. † Dose in prednisone equivalent.

Normal Synacthen Test (N ⫽ 19)

Pathological Synacthen Test (N ⫽ 36)

p

314.3 ⫾ 31.4 114.1 ⫾ 18.3 0/3/0

288.8 ⫾ 13.8 107.5 ⫾ 12.1 0/4/3

NS NS NS

2 10 7 217.6 ⫾ 26.1 6293.1 ⫾ 1005.8 28.6 ⫾ 1.93

5 18 14 257.2 ⫾ 38.5 6956.5 ⫾ 791.6 30.2 ⫾ 1.7

NS NS NS NS NS NS

8 8 3

15 20 1

5.5 ⫾ 0.2/5.1–6.0/4–7.5 4 6

5.8 ⫾ 0.1/5.3–6.0/4–7.5 14 12 7.2 ⫾ 1.3

NS NS NS NS NS NS

1788

Desrame´ et al.

AJG – Vol. 97, No. 7, 2002

Table 3. Univariate and Multivariate Logistic Regression Analysis for 55 Patients and for 37 Patients With Past History of Corticosteroid Therapy Variable

Univariate p

Age Sex Weight Height Type of IBD Duration of disease since onset Past history of surgical treatment Past history of immunosuppressive therapy Past history of corticosteroid therapy Ileum Ileum and colon Colon Corticosteroid treatment (hydrocortisone, prednisolone, prednisone) at time of ST Total cumulative dose of corticosteroid Mean cumulative dose of corticosteroid per day Mean daily dose at time of Synacthen test Mean duration of corticosteroid treatment Surgical procedure during attack Imunosuppressive treatment during the attack Duration free of corticosteroid therapy ⱕ 15 mo†

0.470 0.790 0.780 0.560 0.750 0.430 0.880 0.540 0.0009 0.700 0.990 0.790 0.46 0.760 0.320 0.440 0.930 0.180 0.840 0.01

Multivariate* p

Odds Ratio (95% CI)

0.0009

8.40 (2.20–31.50)

0.54 0.21

2.59 (0.57–11.75)

0.03‡

15.00 (1.23–183.00)

* n ⫽ 55, R ⫽ 0.20; p ⫽ 0.003. n ⫽ 37, R ⫽ 0.25; p ⫽ 0.02. † For the subgroup of patients who had a past history of corticosteroid therapy (n ⫽ 37). ‡ Multivariate logistic regression analysis considering past history of surgical treatment (univariate p ⫽ 0.09) corticosteroid treatment at the time of the ST (univariate p ⫽ 0.18), mean cumulative dose of corticosteroid per day (univariate p ⫽ 0.45) and duration free of corticosteroid therapy lower or equal to 15 mo. 2

2

(available for 25 of 36 patients) was 7.2 ⫾ 1.3 months. In univariate logistic regression analysis, two parameters reached the 20% level of statistical significance (Table 3). These were past history of corticosteroid therapy (p ⫽ 0.0009) and need for surgical procedure during the attack (p ⫽ 0.18). These parameters, as well as the mean cumulative daily corticosteroid dose, although this variable did not reach statistical significance level (p ⫽ 0.32), were considered for multivariate analysis. Their corresponding OR and statistical significance are displayed in Table 3. It seemed that past history of corticosteroid treatment was the only risk factor associated with pathological ST. Moreover, in the subgroup of patients who had a past history of CT (n ⫽ 37), in univariate and multivariate logistic regression analysis, it seemed that a duration free of CT of ⱕ15 months was the only risk factor associated with pathological ST results (OR ⫽ 15.00, CI 1.23–183.00; p ⫽ 0.03) (Table 3). After stratification according to the kind of corticosteroid treatment (hydrocortisone, prednisolone) at the time of the ST, the results were unchanged. In univariate analysis, the kind of corticosteroid treatment (hydrocortisone, prednisolone, prednisone) was not predictive of the result of the ST.

DISCUSSION Suppression of the HPA axis is a well known complication of long term CT. However, little is known about its frequency and about the factors that can produce such side effect in IBD patients. In our retrospective study of about 55

consecutive attacks of IBD, 36 of 55 ST were pathological in long term corticosteroid–treated patients whose corticosteroid treatment had been tapered to physiological doses. In those patients, a past history of corticosteroid treatment was significantly more frequent. Long term CT can cause inhibition of the HPA axis and adrenal atrophy secondary to prolonged suppression of synthesis and secretion of corticotrophin-releasing hormone and ACTH. This adrenal atrophy produces a diminution of the adrenal ability to produce endogenous steroids (3). This state of latent adrenal insufficiency is compensated by low doses of glucocorticoid given during the tapering phase. However, clinical adrenal insufficiency can be related to stress, which requires an increase in cortisol requirements (3–9). Several studies have shown in long-term corticosteroids-treated patients that daily doses equivalent to ⬎10 mg of prednisone consistently caused some degree of HPA axis suppression. In contrast, daily doses ⬍10 mg equivalent prednisone produced HPA axis suppression only in 34 –70% of patients (11–16). In IBD, only a few studies looked for the effect of long term corticosteroid therapy on the HPA axis. In the National Cooperative Crohn’s Disease Study it was not assessed (22). In patients treated for 2– 8 wk by corticosteroid enema with mucosal absorption (hydrocortisone, betamethasone phosphate, or prednisolone), HPA axis suppression was observed in 30 –100% (23–27). Only the budesonide enema did not produce HPA axis impairment (24, 26). Four studies with p.o. CT showed that prednisolone and, to a lesser degree, budesonide produced an inhibition of the HPA axis (28 –31).

AJG – July, 2002

Campieri et al. showed that in a group of patients treated with prednisolone, after 2 wk with 40 mg/day followed by 6 wk with 30 mg/day, 84% of patients had pathological ST results, whereas in the budesonide group at a dose of 9 mg/day, the proportion was 42% with a single daily dose and 50% with 4.5 mg b.i.d. (31). In our study, 65% of ST were pathological. This frequency may be increased because in 10 attacks the dose at the time of the test was equivalent to ⬎6 mg of prednisone. Indeed, the dose of prednisone equivalent to physiological secretion is around 5 mg/day, and the impairment of the HPA axis begins with doses between 7.5 and 10 mg/day (20). However, these 10 attacks were equally distributed in the groups of patients with normal and pathological ST (four vs six, p ⬎ 0.05). Moreover, the mean dose at the time of the test was not significantly different in the two groups (5.5 ⫾ 0.2 vs 5.8 ⫾ 0.1, p ⬎ 0.05) and was not correlated with the T60 cortisol value. In our series, the frequency of HPA axis impairment was similar to that observed in the literature for patients treated with corticosteroids long term versus any disease other than IBD, i.e., 34 –70% (11–16). Comparison between different studies remains difficult because the therapeutic schedule and, especially, the test used to explore adrenal function are very varied. In most studies, adrenal function was evaluated only by basal cortisol levels (23, 24, 26, 29, 30). Serum cortisol is not an accurate predictor of the integrity of the HPA axis because it does not provide any information about the ability of the HPA axis to respond to stress (12, 32). Usually, during stress, there is a prolonged increase in plasma cortisol levels (7, 33). Several studies have shown that in patients treated with corticosteroids or in those with a past history of steroid therapy, stress response can be blunted with a normal basal cortisol level (11, 12, 32). In our study, there was a good correlation between the T60 and the T0 cortisol levels (r ⫽ 0.8, p ⬍ 10 – 6). However, in 13 cases the stimulation test was pathological, with a plasma cortisol level within the normal range (33). In France, the most widely used test for assessment of HPA axis and its ability to respond to acute stimulation is the Synacthen test. Jasani et al. showed good correlation between cortisol plasmatic levels measured with the ST and those during a surgical procedure (33). In this test, plasma cortisol measurements are performed at 8 AM and both 30 and 60 min after an injection of 0.250 mg of Synacthen. In our study, only basal and 60-min values of plasma cortisol were measured. However, a strong correlation was evident between the insulin hypoglycemia test and the standard ACTH stimulation test at both 30 and 60 min (r ⫽ 0.85 and 0.89, respectively; p ⬍ 0.0001) (34). In another study, the sensitivity and specificity of ST were 95% and 100%, respectively, using the Metopyrone test as the gold standard (21). Thus, the ST is considered as a reliable, safe, and easily performed initial assessment of the HPA axis in patients receiving corticosteroids long term; the insulin hy-

HPA Axis and Steroid Therapy in IBD

1789

poglycemia test remains a valuable test for those who fail the ST, but may be hazardous (35). However, the clinical benefit of ST for the physician in charge of IBD patients is questionable. During the tapering phase, this test can show latent adrenal insufficiency still compensated by low doses of exogenous steroids, which can be revealed during acute stress after corticosteroid withdrawal (3–9). The real frequency of such complication has never been assessed. No clinical adrenal insufficiency was observed after corticosteroid withdrawal in trials in which corticosteroids were stopped after immunosuppressive drugs were introduced, but any search for this side effect was not mentioned in the protocol for these trials (36, 37). Deaths secondary to adrenal failure during a surgical procedure in patients treated long term with corticosteroids have been described by Fraser et al. (4), Salassa et al. (5), and Hayes (6). In a retrospective study of 53 operated patients, Hayes found 15 cases of cardiovascular collapses in the group of 28 patients previously treated with steroids, and none in the 25 patients with no previous steroid therapy. Among these 28 patients, 13 were treated by corticosteroids for IBD (6). It is interesting to note that the clinical presentation of adrenal insufficiency could be misleading with fatigue, weakness, fever, diarrhea, and abdominal pain (3, 38). In IBD patients, these symptoms can be misinterpreted as a relapse of the disease. An increase in steroid therapy for these patients treats the adrenal insufficiency and leads to misdiagnosed cases. One patient in our series with pathological test results secondarily developed adrenal failure with hypoglycemic attacks after sudden withdrawal of hydrocortisone. Another case of adrenal insufficiency in an IBD patient has been recently described after sudden withdrawal of steroid enemas (39). Some studies have suggested a relationship between the duration or the total dose of corticosteroid and the HPA impairment (9, 11, 13, 40). These studies all indicated trend but presented no data reaching statistical significance. In our study, as in the others with statistical analysis (12, 14, 15), neither the duration, the total dose, nor the mean cumulative daily dose of corticosteroid predicted the final result of ST. We cannot eliminate the possibility of a type 2 error because we included only 55 patients. Nevertheless, the power in our study is higher than 80% except in the case of the factor “total cumulative dose of CT.” Moreover, in a study in which 279 patients have been included, the conclusions of the authors were similar (12). Thus, our results reinforce the need for stimulation tests to assess HPA axis function. An individual susceptibility to steroid therapy has also been advocated in HPA axis impairment (41). In IBD, variations in steroid absorption related to mucosal alterations (42) as well as interaction between cytokines and the HPA axis (43– 45) could also play a role. In rheumatoid arthritis, tumor necrosis–␣, interleukin-1, and interleukin-6 stimulate the secretion of corticotrophin at the hypothalamic level, whereas at peripheral level they inhibit endogenous steroid synthesis (43, 45).

1790

Desrame´ et al.

Consequences of repeated courses of steroid treatment have never been studied in corticosteroid-treated patients. We found that pathological ST results were more frequent in patients previously treated with corticosteroids. Past history of CT was the only risk factor for pathological ST results in multivariate analysis. It is difficult to interpret these results in the absence of ST after the end of previous steroid treatment. Indeed, after steroid withdrawal, recovery of normal HPA axis function is progressive and can take several months or years (10). In our series, when these data were available for 25 patients, the mean time for normal test recovery was 7.2 ⫾ 1.3 months (range 1–18 months). Moreover, for the patients with a past history of CT, the time elapsed between the last course of steroid treatment was significantly greater in patients with normal ST than in those with pathological ST (45.5 ⫾ 13.5 vs 15.4 ⫾ 6.0 months, p ⫽ 0.02) and a duration free of CT of ⱕ15 months was the only risk factor associated with pathological ST (OR ⫽ 15.00, CI ⫽ 1.23–183.00; p ⫽ 0.03). This suggests that patients with pathological ST may not have sufficient time to recover normal HPA function after the last course of corticosteroids. Interestingly, when the results of ST were reviewed according to past history of CT, although total duration of CT was not a predictive factor of the result of ST in logistic regression, we found no pathological ST results in the 10 patients with previous CT of ⬍5 months. In conclusion, long term CT in IBD patients was associated with HPA axis impairment in 65% of our patients. The only risk factor for HPA axis impairment was a past history of steroid therapy. These results need to be confirmed in a prospective study and suggests performing stimulation tests during the tapering phase before steroid withdrawal, especially in patients with past history of CT. Patients with HPA axis impairment must be carefully followed and treated with hydrocortisone until HPA axis recovery (46). Reprint requests and correspondence: J. Desrame´ , M.D., 15 rue du Dr Roux 75015, Paris, France. Received June 12, 2001; accepted Dec. 4, 2001.

REFERENCES 1. Modigliani R, Mary JY, Simon JF, et al. Clinical, biological and endoscopic picture of attacks of Crohn’s disease: Evolution on prednisolone. Gastroenterology 1990;98:811– 8. 2. Summers RW, Switz DM, Sessions JT Jr, et al. National Cooperative Crohn’s Disease Study: Results of drug treatment. Gastroenterology 1979;77:847– 69. 3. Oelkers W. Adrenal insufficiency. N Engl J Med 1996;335: 1206 –12. 4. Fraser CG, Preuss FS, Bigford WD. Adrenal atrophy and irreversible shock associated with cortisone therapy. JAMA 1952;149:1542–3. 5. Salassa RM, Bennett WA, Keating FR Jr, et al. Postoperative adrenal cortical insufficiency. Occurrence in patient previously treated with cortisone. JAMA 1953;152:1509 –15. 6. Hayes MA. Surgical treatment as complicated by prior adrenocortical steroid therapy. Surgery 1956;40:945–50.

AJG – Vol. 97, No. 7, 2002

7. Lamberts SWJ, Bruining HA, De Jong FH. Corticosteroid therapy in severe illness. N Engl J Med 1997;337:1285–92. 8. Sampson PA, Brooke BN, Winstone NE. Biochemical confirmation of collapse due to adrenal failure. Lancet 1961;1:1377. 9. Sampson PA, Winstone NE, Brooke BN. Adrenal function in surgical patients after steroid therapy. Lancet 1962;1:322–5. 10. Graber AL, Ney RL, Nicholoson WE, et al. Natural history of pituitary-adrenal recovery following long-term suppression with corticosteroids. J Clin Endocrinol Metab 1965;25:11– 6. 11. Livanou T, Ferriman D, James VHT. Recovery of hypothalamic-pituitary-adrenocortical function after corticosteroid therapy. Lancet 1967;2:856 –9. 12. Schlaghecke R, Kornely E, Santen RT, et al. The effect of long-term glucocorticoid therapy on pituitary-adrenal responses to exogenous corticotropin-releasing hormone. N Engl J Med 1992;326:226 –30. 13. Wood JB, Frankland AW, James VHT, et al. A rapid test of adrenocortical function. Lancet 1965;1:243–5. 14. La Rochelle GE, La Rochelle AG, Ratner RE, et al. Recovery of the hypothalamic-pituitary-adrenal (HPA) axis in patients with rheumatic diseases receiving low-dose prednisone. Am J Med 1993;95:258 – 64. 15. Daly JR, Myles AB, Bacon PA, et al. Pituitary adrenal function during corticosteroid withdrawal in rheumatoid arthritis. Ann Rheum Dis 1967;26:18 –25. 16. Myles AB, Bacon PA, Daly JR. Single daily dose corticosteroid treatment. Effect on adrenal function and therapeutic efficacy in various diseases. Ann Rheum Dis 1971;30:149 –53. 17. Best WR, Becktel JM, Singleton JW, et al. Development of a Crohn’s Disease Activity Index: National Crohn’s Disease Study. Gastroenterology 1976;70:439 – 44. 18. Truelove SC, Witts LJ. Cortisone in ulcerative colitis: Final report on a therapeutic trial. Br Med J 1955;2:1041– 8. 19. Murphy BE. Studies of the protein binding of steroids and their applications to the routine of micro and ultra-micro measurements of various steroids in body fluid by competitive protein radioassay. J Clin Endocrinol 1967;27:973–90. 20. Swartz SL, Dluhy RG. Corticosteroids: Clinical pharmacology and therapeutic use. Drugs 1978;16:238 –55. 21. Chaieb L, Vidal-Trecan G, Laudat MH et al. Le test rapide a` la ␤ 1–24 corticotrophine dans l’exploration de l’axe corticotrope. Soixante et onze observations. Presse Med 1984;13: 2183–7. 22. Singleton JW, Law DH, Kelley ML, et al. National Cooperative Crohn’s Disease Study: Adverse reactions to study drugs. Gastroenterology 1979;77:870 – 82. 23. Danielsson A, Hellers G, Lyrenas E, et al. A controlled randomized trial of budenoside versus prednisolone retention enemas in active distal ulcerative colitis. Scand J Gastroenterol 1987;22:987–92. 24. Bianchi Porro G, Prantera C, Campieri M, et al. Comparative trial of methylprednisolone and budesonide enemas in active distal ulcerative colitis. Eur J Gastoenterol Hepatol 1994;6: 125–30. 25. Luman W, Gray RS, Pendek R, et al. Prednisolone metasulphobenzoate foam retention enemas suppress the hypothalamo-pituitary-adrenal axis. Aliment Pharmacol Ther 1994; 8:255– 8. 26. Lo¨ fberg R, Thomsen OO, Langholz E, et al. Budesonide versus prednisolone enemas in active distal ulcerative colitis. Aliment Pharmacol Ther 1994;8:623–9. 27. Lee DA, Taylor GM, James VH, et al. Plasma prednisolone levels and adrenocortical responsiveness after administration of prednisolone-21-phosphate as a retention enema. Gut 1979; 20:349 –55. 28. Greenberg GR, Feagan BG, Martin F, et al. Oral budesonide for active Crohn’s disease. N Engl J Med 1994;331:836 – 41.

AJG – July, 2002

29. Rutgeerts P, Lo¨ fberg R, Malchow H, et al. A comparison of budesonide with prednisolone for active Crohn’s disease N Engl J Med 1994;331:842–5. 30. Lo¨ fberg R, Danielsson A, Suhr O, et al. Oral budesonide versus prednisolone in patients with active extensive and leftsided ulcerative colitis. Gastroenterology 1996;110:1713– 8. 31. Campieri M, Ferguson A, Doe W, et al. Oral budesonide is as effective as oral prednisolone in active Crohn’s disease. Gut 1997;41:209 –14. 32. Blevins LJ. Serum cortisol is not an accurate predictor of the integrity of the hypothalamic-pituitary-adrenocortical axis. Clin Endocrinol 1995;42:101–2. 33. Jasani MK, Freeman PA, Boyle JA, et al. Studies of the rise in plasma 11-hydroxycorticosteroids in corticosteroid-treated patients with rheumatoid arthritis during surgery: Correlations with the functional integrity of the hypothalamo-pituitaryadrenal axis. Q J Med 1968;37:407–21. 34. Rasmuson S, Olsson T, Ha¨ gg E. A low dose ACTH test to assess the function of the hypothalamo-pituitary-adrenal axis. Clin Endocrinol 1996;44:151– 6. 35. Kane KF, Emery P, Sheppard MC, et al. Assessing the hypothalamo-pituitary-adrenal axis in patients on long term glucocorticoid therapy: The short synacthen versus the insulin tolerance test. Q J Med 1995;88:263–7. 36. Feagan BG, Rochon J, Fedorak RN, et al. Methotrexate for the treatment of Crohn’s disease. N Eng J Med 1995;332:292–7. 37. Present DH, Korelitz BI, Wisch N, et al. Treatment of Crohn’s disease with 6-mercaptopurine. N Engl J Med 1980;302: 981–7.

HPA Axis and Steroid Therapy in IBD

1791

38. Tobin MV, Albridge SA, Morris AI, et al. Gastrointestinal manifestations of Addison’s disease. Am J Gastroenterol 1989;84:1902–5. 39. Bili H, Algayres JP, Revel F, et al. Insuffisance surre´ nale secondaire a` un traitement par lavements de corticoı¨des. Gastroenterol Clin Biol 1998;22:1113– 4. 40. Jasani MK, Boyle JA, Greig WR, et al. Corticosteroid-induced suppression of the hypothalamo-pituitary-adrenal axis: Observations on patients given oral corticosteroids for rheumatoid arthritis. Q J Med 1967;143:261–76. 41. Sarna S, Hoppu K, Neuvonen PJ, et al. Methylprednisolone exposure, rather than dose, predicts adrenal suppression and growth inhibition in children with liver and renal transplants. J Clin Endocrinol Metab 1997;82:75–7. 42. Shaffer JA, Williams SE, Turnberg LA, et al. Absorption of prednisolone in patients with Crohn’s disease. Gut 1983;24: 182– 6. 43. Masi AT, Chrousos GP. Hypothalamic-pituitary-adrenal-glucocorticoid axis function in rheumatoid arthritis. J Rheumatol 1996;23:596 – 602. 44. Balfour Sartor R. Cytokines in intestinal inflammation: Pathophysiological and clinical considerations. Gastroenterology 1994;106:533–9. ¨ berg K, et al. Intact adreno45. Gudbjo¨ rnsson B, Skogseid B, O corticotropic hormone secretion but impaired cortisol response in patients with active rheumatoid arthritis. Effect of glucocorticoids. J Rheumatol 1996;23:577– 81. 46. Christy NP. Pituitary-adrenal function during corticosteroid therapy. N Engl J Med 1992;326:266 –7.