TREATMENT OF CROHN'S DISEASE WITH AZATHIOPRINE: A CONTROLLED EVALUATION

Transcription

1 GASTROENTEROLOGY 66: , 1974 Copyright 1974 by The Williams & Wilkins Co. Vol. 66, No.5 Printed in U.S.A. TREATMENT OF CROHN'S DISEASE WITH AZATHIOPRINE: A CONTROLLED EVALUATION MADELINE KLEIN, M,D., HENRY J. BINDER, M.D., MALCOLM MITCHELL, M.D,, ROBERT AARONSON, M,D., AND HOWARD SPIRO, M.D, The Department of Internal Medicine and Pharmacology, Yale University School of Medicine, New Haven, Connecticut To determine whether azathioprine is an effective drug in the treatment of Crohn's disease a double-blind study was performed, Twenty-seven patients with involvement of either small or large intestine or both were randomized to either a treatment group (azathioprine 3 mg per kg of body weight) or a placebo group for 4 months. During the subsequent 4-month period, the treatment schedule of the two groups was reversed. Although some patients improved dramatically on azathioprine, others improved during the placebo period. This study does not provide evidence to support the effectiveness of azathioprine in the treatment of Crohn's disease. Treatment of Crohn's disease of the small and large intestine is often unsatisfactory, a variety of medications having been utilized empirically with variable degrees of success. Suppression of symptoms and spontaneous remission, rather than cure of the disease, is the general pattern. Since the cause of Crohn's disease is unknown, it is not surprising that an immunological etiology has been proposed with only modest experimental support, and recently immunosuppressive therapy has been reported effective in Crohn's1-1o colitis as well as in ulcerative colitis The initial reports were impressive but anecdotal since they were not supported by appropriate control studies. Of all the Received August 2, Accepted November 20, Address requests for reprints to: Dr. Henry J. Binder, 333 Cedar Street, Yale University School of Medicine, New Haven, Connecticut 06510, Supported in part by a grant from the John A. Hartford Foundation, Inc. The patients were studied on the General Clinical Research Center supported by a grant from the United States Public Health Service, Division of Research Resources (RR-125). Azathioprine (Imuran) and placebo tablets were kindly supplied by Burroughs Wellcome Company, Research Triangle, North Carolina. 916 studies thus fa.r published on the use of azathioprine in the treatment of Crohn's disease, only two have been controlled,7. 8 and azathioprine was found effective in only one. Crohn's disease is a disorder characterized by spontaneous remissions in which fistulae occasionally heal spontaneously so that randomized studies are mandatory in the evaluation of any new therapeutic regimen in this disease. We have, therefore, initiated such a randomized double blind cross-over study of the effect of azathioprine, an immunosuppressive agent, in patients with Crohn's disease. We studied the clinical effectiveness of azathioprine as compared with a placebo in patients with extensive Crohn's disease, and also the effectiveness of azathioprine in suppressing humoral and cellular immunity in these patients. Patients and Methods For this study we chose patients with Cr<lhn's disease: (1) unresponsive to other forms of medical management; (2) with intestinal complications that would require extensive extirpative surgery; or (3) unable to tolerate steroids because of the severity of their side effects. Therefore, subjects in this study represent a group of patients seriously ill with Crohn's

2 May 1974 AZATHIOPRINE AND CROHN'S DISEASE 917 disease, some for as long as 32 years and some for as short a period as 18 months. Twenty-seven patients with active disease entered the study. Information concerning the primary clinical features, extent of previous surgery, and location of present disease for all patients is summarized in table 1. Ten patients had fistulae (enterocutaneous, enterovesicular, or perirectal) that failed to heal with conventional medical management. Two patients had had complications secondary to steroid therapy: osteoporosis or cataract formation. The other patients had active disease with weight loss, diarrhea, and abdominal pain refractory to other forms of therapy. Thirteen patients had had previous resection of either small or large bowel, and 1 patient had had a bypass procedure. Only 3 of the patients had normal small bowel by X-ray at the time of entering the study, and only 8 had a normal colon by barium enema. Two patients had had total colectomies and therefore were considered to have only small intestinal disease. Before starting the study, all patients understood the nature of this investigation. The patients were admitted to the General Clinical Research Center for sigmoidoscopy with biopsy, X-ray studies ofthe entire alimentary tract, and an evaluation which included protein electrophoresis, immunoelectrophoresis, liver function studies, and absorptive studies (d-xylose, carotene, and Schilling test with intrinsic factor). Subjects were randomized to a treatment or a control (placebo) group. The treatment group received. azathioprine 3 mg per kg of body weight orally each day, while the control subjects took a similar amount of dummy tablets (placebo). Patients were seen weekly for 2 weeks, then biweekly, and finally monthly. At each visit, a complete blood count was done; liver function was tested monthly and blood was drawn weekly for 6 weeks for immunological studies. Symptoms were recorded, specific note being taken of well being, appetite, weight change, number and character of stools, rectal bleeding, abdominal pain, joint or eye symptoms, and the state of perirectal or fistulous disease. Changes in medication were also noted. The initial and final prednisone dose was recorded for each patient. The initial dose represented the average dose received during the week before and the week after initiation of the azathioprine. The final dose represents the average of the dose of prednisone at 3 and 4 months. In addition, the patients were seen by a private physician at whatever interval he thought appropriate, and other medications were changed by that physician depending on his evaluation of the course of disease, attempts specifically being made to decrease steroids as rapidly as consistent with well being. Four patients were on a fixed dose of salicylazosulfapyridine throughout the entire study. No patient was started on either salicylazosulfapyridine or antibiotics during the study. The results of the blood counts were controlled by one of the investigators (M. M.) who did not evaluate these patients clinically. This same physician altered the azathioprine dose whenever indicated. Studies of immunoreactivity included an attempt to sensitize with 2, 4-dinitrochlorobenzene (DNCB) and Candida and Trichophyton antigens. Patients also received subcutaneous injections of Escherichia coli Vi antigens (70!J.g) and tetanus toxoid (7.5 flocculation units) at the time of initiation of drug therapy, and serial samples of blood were collected at weekly intervals to test the primary and secondary humoral immunoresponses, respectively. At the end of a 4-month period, repeat X-rays and blood studies were carried out. Patients were then "crossed over" to the other drug regardless of how they were functioning. Four months later, all studies again were repeated and the study was considered complete at 8 months. Determination of the sequence of drug administration was not made until 27 patients had completed the study. Results Ofthe 27 patients who entered the study, 1 was excluded from the protocol because of medication error. Thirteen patients (group A) received azathioprine during the first 4 months of the study (period 1) and 13 patients (group B) received placebo during period 1. At the given dose at 3 mg per kg, there was no change in the white blood count necessitating an adjustment in drug dosage. A comparison between patients entered into groups A and B (table 2) shows no significant difference between the two groups. Groups A and B were evaluated during period 1 (table 3). Of the 13 patients in group A, 2 developed apparently allergic reactions to azathioprine with chills, fever, and acute arthritis 5 days and 14 to 21 days after starting the drug. Both developed recurrent symptoms with reinstitution of azathioprine. Therefore, they could not

3 - - -,-- TABLE 1. Summary of clinical information a Patient Age Sex Duration of Main clinical feature Previous surgery Site of bowel disease' symp toms.vr 1 27 M 5 Ankylosing spondylitis Total colectomy TI severe osteoporosis 2 31 M 8 Diarrhea, fistula Ileotransverse colos- TC + pre anastomotic tomy ileum 3 23 F 9 Anorexia, weight loss, None Entire colon with rectal diarrhea, abdominal ulcers pain, arthralgias 4 30 M 6 Rectal fistulae Ileosigmoidostomy Preanastomotic ileum; rectal ulcers 5 26 M 4 Abdominal pain, diarrhea None DC, TC, TI 6 22 F 10 Rectal bleeding Left hemicolectomy DC, cecum 7 57 F 25 Perirectal fistulae and Ileotransverse colos- TI, entire colon abscesses, diarrhea tomy 8 23 M 5 Anorexia, abdominal Ileal resection Upper jejunum, TI pain, fever 9 22 F 4 Anorexia, diarrhea, ab- None Entire colon, TI dominal pain, perirectal fistula, low back pain M 2 Diarrhea, abdominal pain None TI, TC, DC M 17 Anorexia, abdominal pain Multiple small bowel Jejunum resections M 2 Enterocutaneous fistula Ileal resection Preanastomotic ileum F 5 Perirectal fistulae and None Total colitis, TI, ileoabscesses cecal fistul ae M 5 Anorexia, weight loss, Ileotransverse colos- Entire colon fever, diarrhea, fistala tomy F 10 Diarrhea, multiple peri- Ileotransverse colos- Rectum, DC, TC, + TI rectal fistulae, and to my abscesses F 8 Weight loss. arthralgias None Entire colon, TI, rectal ulcers M 4 Weight loss, diarrhea None TI F 3 Pneumaturia, diarrhea None Mid-small bowel, DC, rectovesical fistula M 1.5 Anorexia, weight loss. None Multiple areas in small abdominal pain, low bowel sigmoid back pain M 9 Anorexia. abdominal Ileotransverse colos- TI pain, diarrhea to my M 3 Abdominal pain None Multiple areas of stricture in jejunum M 4 Anorexia, abdominal Ileal resection Jejunum pain, low back pain M 12 Diarrhea, arthralgia Right hemicolectomy Entire colon. preanastomotic ileum M 32 Diarrhea, steroid related Iieoproctostomy Distal small bowel cataracts F 14 Diarrhea, abdominal pain Ileotransverse colos- TI tomy M 11 Perirectal fistulae and Multiple small bowel TI abscesses resections F 5 Anorexia, abdominal pain Right hemicolectomy Preanastomotic ileum a Group A includes patients 1 to 13; group B patients 14 to 26, and patient 27 was excluded due to medication error. o TI, terminal ileum; DC, descending colon; TC, transverse colon; AC, ascending colon. 918

4 May 1974 AZATHIOPRINE AND CROHN'S DISEASE 919 continue in the study and were excluded from additional consideration. None of the patients had recognizable changes in absorptive or radiological studies. Six patients in group A felt that they had improved on azathioprine, while 5 patients noted no subjective improvement. Objective evidence of improvement consisted of weight gain of 24, 12, and 8 lb in 3 patients, decrease in diarrhea (with or without improvement in abdominal pain) in 3 of 6 patients, cessation of rectal bleeding in 1 patient, and improvement of arthritis in 2 of 4 patients. One patient had complete healing of an enterocutaneous fistula. Perirectal fistulae improved but did not completely heal in 3 patients. A large ischiorectal abscess in 1 of these patients decreased in size but did not heal. One patient showed no improvement of severe perirectal disease. A total of 9 patients received prednisone during period 1. The initial average dose of 13.3 mg per day decreased to the final average dose of 6.3 mg per day. Two patients were able to discontinue use of prednisone completely and 1 required its addition. Of 13 patients in group B, 2 had to leave the study after 1 and 2 months because of surgical emergencies: in 1 patient a free perforation ofthe cecum, and, in the other, the development of a rectovesical fistula TABLE 2. Comparison of patients receiving azathioprine (group A) and placebo (group B) during initial 4-month (period 1) Number of patients.. Mean age (yr) Sex (M :F) Duration of illness (yr). Extent of disease Small intestine only Large intestine only... Small and large intestine. Symptoms Fistulae Perirectal Enterocutaneous Enterovesicular... Abdominal pain and/or diarrhea. Arthritis.' Rectal bleeding Group Group A B :5 9: TABLE 3. Summary of objective and subjective parameters in patients receiving azathioprine (group A) or placebo (group B) during period 1 Group A Group B Completed period 1 11/13a. b 11/13< Subjective assessment Improvement. 6 6 No change. 5 5 Objective assessment Change in X-ray. 0/ 11 0/11 Change in absorption.. 0/11 0/11 Change in weight d Gain 3/11 3/11 Loss 0/11 1/11 Change in joint symptoms Better. 2/4 1/3 No Change. 1/4 1/3 Worse. 1/4 1/3 Change in pain/diarrhea Improvement.. 3/7 3/11 Cessation of rectal bleeding 1/1 Change in fistulae Healed. 1/5 1/5 Improved.. 3/5 1/5 Unchanged 1/5 3/5 Average prednisone dose Initial 13.3 ± ± 5.9 Final ± ± 3.0 a Demoninator refers to total patient population at risk. Two patients developed allergic reaction to drug and were dropped from study. < Two patients developed surgical emergencies. d Greater than 5 lb. with severe urinary tract symptoms. They were, therefore, considered treatment failures. No patient had changes in absorptive or radiological studies. Six patients felt better, and 5 noted no symptomatic change. Weight gain occurred in 3 patients; 1 of 3 patients noted improvement of arthritis; 3 of 9 patients noted decreased diarrhea and abdominal pain. A perirectal fistula healed completely in 1 patient and 1 patient noted decrease in fistula drainage; 3 other patients noted no change in their fistulae. Eight patients required prednisone during period 1. The initial average dose was 19.8 mg per day; the final average dose was 7.8 mg per day. Two patients were.

5 920 KLEIN ETAL. Vol. 66, No.5 able to discontinue the prednisone, but 1 required its addition to the regimen. A comparison between periods 1 and 2, after crossover, for both groups was also made. Of 11 patients in group A who completed period 1 (azathioprine period) all went on to complete period 2 (placebo period). Minimal radiological changes were noted in 3 of these patients during period 2: 2 patients improved and 1 worsened. No changes in absorption were noted. Six patients felt better during period 1, 3 felt better during period 2, and 2 noted no change. Objective changes during period 2 consisted of weight gain in 2 patients and weight loss in 1; increase in pain and diarrhea in 4 patients, decrease in only 1; decrease in fistula size in 3 of 4 patients and increase in arthritic symptoms in 3 of 4. The initial average dose of prednisone was 5.2 mg per day; the final average dose was 5.0 mg per day. Six patients required steroids during period 2; 2 patients discontinued the use of prednisone and 1 required reinstitution of steroid therapy. Eleven of 13 patients in group B, who received placebo first, completed period 1. One patient required a resection because of repeated bouts of small bowel obstruction at the end of period 1. Of the 10 remaining patients who entered period 2 (azathioprine period for group B), 9 completed the study and 1 required surgery 2 weeks after crossover because of small bowel obstruction. No changes in absorptive study were noted. In 1 patient, radiological improvement was noted during period 2. Six patients noted subjective improvement in period 1, however, none felt improved during period 2. Objective changes during period 2 consisted of increased pain and diarrhea in 2 patients and increased arthritis symptoms in 2 patients with improvement in none. There was no improvement in perirectal disease. The initial average dose of prednisone was 4.4 mg per day; the final average dose was 1.7 mg per day. Three of 5 patients on prednisone were able to discontinue its use. Immunological Tests Humoral immunity. Twenty-seven patients were tested for their ability to respond to an injection of Vi antigen (primary response) and tetanus toxoid (secondary response) during the first period of therapy. Fourteen patients received azathioprine and 13 received placebo. Of the patients receiving the active drug, only 2 responded normally to both antigens. One of these was a patient whose drug was discontinued early because of an adverse reaction, 2 others responded only to tetanus, and another to Vi alone. The remaining 8 patients failed to respond to either antigen. Interpretation of whether azathioprine was truly immunosuppressive in these patients is clouded by the finding that, of 13 patients receiving placebo, 7 failed to respond to both antigens, 4 manifested only a secondary response to tetanus, and another patient responded only to Vi, giving an equivocal tetanus response. Only 1 patient of the 13 gave a completely normal response to both antigens. There was no consistent relationship between immunosuppression and administration of prednisone in the patients on placebo. Delayed hypersensitivity (cell-mediated immunity). Thirteen patients were tested for their ability to be sensitized to DNCB-7 receiving azathioprine and 6 receiving placebo. Four of the 7 patients taking azathioprine could be sensitized even though the induction and testing of skin reactivity occurred during therapy in 3 of them. The 4th patient was the one whose drug was discontinued prematurelywhose challenge dose of DNCB was given in the absence of azathioprine. Five of the 6 patients receiving the placebo were sensitized to DNCB. Patients with regional enteritis, regardless of whether they were receiving "immunosuppressive" therapy, failed to respond to antigens leading to humoral immunity in most instances. They were able to respond to a skin-sensitizing agent even in the presence of the drug. Discussion Bean 11 first reported the use of an immunosuppressive agent in the treatment of inflammatory bowel disease in The initial reports reported success in the treat-

6 May 1974 AZATHIOPRINE AND CROHN'S DISEASE 921 ment of ulcerative colitis when nitrogen mustard, 6-mercaptopurine, or azathioprine were used. Although multiple reports of the use of azathioprine and other suppressive agents have been reported in ulcerative colitis to date,11-15 there is no objective evidence of its efficacy. Furthermore, it must be remembered that in diseases of a chronic nature, the natural course of the disease is associated with frequent exacerbations and remissions often requiring the use of careful controlled studies to ascertain the effectiveness of a new therapeutic regimen. A recent preliminary communication indicated that azathioprine was not effective in preventing exacerbation in a group of glucocorticosteroid-treated patients with ulcerative colitis. 16 In 1969, Brooke et al. 3 reported that azathioprine induced the closure of enterocutaneous fistulae in 6 patients with Crohn's disease. This dramatic success prompted the use of azathioprine in additional patients. These subsequent reports have, on the whole, been encouraging. Since we initiated this study, two control studies have been reported: one demonstrated the failure of azathioprine to induce remission in a group of patients with exacerbatio.lls of Crohn's disease,7 and the other study demonstrated its effectiveness in preventing exacerbations in patients presently in remission. 8 Our results permit several tentative conclusions. Comparison of group A and group B patients during the initial 4 months of study showed no significant difference between the treatment and the control group, except that 4 patients who were receiving placebo first required operation during or immediately after the placebo period. Improvement during treatment with azathioprine certainly occurred, and the requirement of prednisone was generally reduced, but improvement was not uniformly seen. That some patients improved while receiving placebo testifies to the importance of a control study. Moreover, when we tried to find common characteristics in 6 patients who improved most dramatically on azathioprine, we were unable to find any set of characteristics that would allow us to predict, in advance, the subgroup who might be anticipated to do well on azathioprine. Although in a chronic disease, evaluation of a second period may be unduly influenced by the events of the initial period, some additional information can be obtained by a comparison of the crossover groups. Group A patients received azathioprine and then placebo, while group B patients received placebo first and then azathioprine. It appears that while some patients in group A relapsed when placed on placebo, other patients in group B who improved on placebo also relapsed when crossed over to azathioprine. Immunological investigations revealed that patients with regional ileitis were poorly responsive to antigens that led to antibody synthesis. They were deficient in their ability to manifest a primary or a secondary (anamnestic) immune response, regardless of whether they were receiving azathioprine. This characteristic was clearly seen even in patients who were able to be sensitized to DNCB. The latter preservation of cell-mediated immunity was found in both treated and untreated patients. It is unlikely that azathioprine had any immunosuppressive effects upon these patients. Usually it is cell-mediated immunity that is most sensitive to immunosuppressive agents,17 although exceptions have been noted Our results with DNCB are, for inexplicable reasons, at odds with earlier reports that patients with ileitis were less able to be sensitized to this chemical than normal subjects. 21 However, other studies have reported intact delayed hypersensitivity in Crohn's disease. 22 The immunosuppressive evaluation carried out on these patients is of interest since azathioprine has anti-inflammatory properties as well as being an immunosuppressant. Although some patients showed suppression of their delayed and immediate hypersensitivity systems before drug treatment, no patients treated with azathioprine demonstrated immunosuppressive when appropriately challenged. Several conclusions appear warranted from this study. Over-all, azathioprine does not appear to provide significant im-

7 922 KLEIN ETAL. Vol. 66,No.5 provement in the treatment of all patients with diffuse Crohn's disease during a 4- month period. It should be noted, however, that the only patients in this study who required operation during the study were 4 patients who were not receiving azathioprine. That some patients respond dramatically to azathioprine during the administration of the drug and during additional periods of up to 8 months afterwards (M. Klein, H. J. Binder, and H. M. Spiro, unpublished observations) suggests that some "subset" needs to be detected, but at present, we can find no common denominator that will predict which patients may respond to azathioprine. We conclude, therefore, that while azathioprine may be an effective agent in the treatment of some patients with Crohn's disease, it is most likely effective in a small group of patients of uncertain characterization at this time. It seems unlikely that improvement is secondary to immunosuppression. If this group represents no more than even 30 to 50% of all patients with Crohn's disease, it will probably be impossible to demonstrate the effectiveness of azathioprine by studying all patients with the disease in a double-blind fashion. Indeed, if the slight improvement observed during period 1 in the azathioprine-treated patients continued, we estimate that 75 patients would be required in each group in order to establish statistically that azathioprine is effective in the treatment of Crohn's disease. Therefore, an attempt to identify an appropriate subgroup will first be required followed by a control study of this subgroup. Those patients that do respond to azathioprine may represent a single nosological entity, especially since, at present, Crohn's disease probably represents several etiological conditions. REFERENCES 1. Patterson JF, Norton RA, Schwartz RS: Azathioprine treatment of ulcerative colitis, granulomatous colitis and regional enteritis. Am J Dig Dis 16: , Brooke BN, Javett SL, Davison OW: Further experience with azathioprine for Crohn's disease. Lancet 2: , Brooke BN, Hoffmann DC, Swarbreck ET: Azathioprine for Crohn's disease. Lancet 2: , Drucker WR, Jeejeebhoy KN: Azathioprine: an adjunct to surgical therapy of granulomatous enteritis. Ann Surg 172: , Brown CH, Achkar E: Azathioprine therapy for inflammatory bowel disease. Am J Gastroenterol 54: , Jones FA, Brown P, Lennard-Jones JE: Azathioprine for Crohn's disease. Lancet 2:795, Rhodes J, Bainton D, Beck P, et al: Controlled trial of azathioprine in Crohn's disease. Lancet 2: , Willoughby JMT, Beckett J, Kumar PJ, et al: Controlled trial of azathioprine in Crohn's disease. Lancet 2: , Jones RA: Immunosuppressive therapy in Crohn's disease. Proc Roy Soc Med 64: , Korelitz BI, Wisch N: Long term therapy of ulcerative colitis with 6-mercaptopurine: a personal series. Am J Dig Dis 17: , Bean RHD: Treatment of ulcerative colitis with antimetabolites. Br Med J 1: , Bowen GE, Irons GV, Rhodes JB, et al: Early experiences with azathioprine in ulcerative colitis. JAMA 195: , MacKay IR, Wall AJ, Goldstein G: Response to azathioprine in ulcerative colitis. Am J Dig Dis 11: , Theodor E, Gilon E, Waks U: Treatment of ulcerative colitis with azathioprine. Br Med J 4: , Jones RA: Immunosuppressive therapy in ulcerative colitis. Proc Roy Soc Med 62: , Rosenberg JL, Wall AJ, Settles RH, et al: A controlled trial of azathioprine in the treatment of chronic ulcerative colitis (abstr). Gastroenterology 64:793, Makinodan T, Santos GW, Quinn RP: Immunosuppressive drugs. Pharmacol Rev 22: , Mitchell MS, Wade ME, DeConti RC, et al: Immunosuppressive effects of cytosine arabinoside and methotrexate in man. Ann Intern Med 70: , Friedman RM, Buckler CE, Baron S: The effect of amino methyl pteroylglutamic acid on the development of skin hypersensitivity and antibody formation in guinea pigs. J Exp Med 114: , Friedman RM, Buckler CE, Baron S: Effect of methotrexate on the development of skin hypersensitivity and on antibody formation in guinea pigs (abstr). Fed Proc 20:258, Jones JV, Housley J, Ashurst PM: Development of delayed hypersensitivity to dinitrochlorobenzene in patients with Crohn's disease. Gut 10:52-56, Binder HJ, Spiro HM, Thayer WR Jr: Delayed hypersensitivity in regional enteritis and ulcerative colitis. Am J Dig Dis 11: , 1966