|
|
| |||||||||||||||||||||||||||||||
Background Long-term corticosteroid therapy for Crohn's disease is associated with important types of morbidity, such as osteoporosis. Safe and effective alternative treatments are required. Although a short-term benefit of cyclosporine in active Crohn's disease has been suggested, the long-term safety and efficacy of this treatment have not been established.
Methods We conducted a randomized, double-blind, placebo-controlled evaluation of the effect of 18 months of low-dose cyclosporine treatment on the course of Crohn's disease. Adult patients whose disease had been active within the previous two years were randomly assigned to receive cyclosporine (151 patients) or placebo (154 patients) in addition to their usual therapy. Randomization was stratified according to center and score on the Crohn's Disease Activity Index (193 patients had scores of 150 or less, and 112 had scores greater than 150). The primary outcome measure was clinically important worsening of Crohn's disease, defined as a 100-point increase in the Crohn's Disease Activity Index from the patient's base-line value. Secondary outcomes were the use of prednisone and 5-aminosalicylates, mean score on the Crohn's Disease Activity Index and mean quality-of-life score, and the need for surgery.
Results The condition of more patients worsened with cyclosporine than with placebo (91 of 151, or 60.3 percent, vs. 80 of 154, or 51.9 percent; P = 0.10). The median time to worsening of disease in patients receiving cyclosporine was 338 days, as compared with 492 days in patients receiving placebo (P = 0.25; relative risk, 1.22; 95 percent confidence interval, 0.86 to 1.72). Analyses of the mean Crohn's Disease Activity Index and quality-of-life scores and of the use of prednisone and 5-aminosalicylates also failed to demonstrate benefit.
Conclusions In our patient population, the addition of low-dose cyclosporine to conventional treatment for Crohn's disease did not improve symptoms or reduce requirements for other forms of therapy.
Cyclosporine, a cyclic oligopeptide that modulates T-cell function, is effective in some steroid-responsive inflammatory diseases, such as rheumatoid arthritis,12 psoriasis,13 and Behcet's disease14. In 1989 Brynskov et al. reported a beneficial effect in Crohn's disease of short-term treatment with cyclosporine at relatively high doses (median dose, 7.6 mg per kilogram of body weight daily)15. However, the efficacy and potential for adverse effects16,17,18 such as nephrotoxicity and gastrointestinal toxicity with long-term cyclosporine treatment in a larger, more heterogeneous patient population have not been established.
In this study, we assessed the effect of 18 months of low-dose cyclosporine as compared with placebo in patients whose Crohn's disease was active or in remission, but who were at substantial risk for worsening.
Methods
The study was conducted at seven Canadian centers from July 1988 through March 1992 with the approval of the investigational review board at each center.
Study Population
To be eligible for the study, patients had to be from 18 to 65 years of age, with active Crohn's disease within the preceding two years, as defined by the occurrence of symptoms requiring treatment with prednisone or a 5-aminosalicylate. Each patient's records, radiographs, and pathological reports were reviewed, and patients with a confirmed diagnosis2 of Crohn's disease were eligible for the study, whether their disease was in remission or active. Patients with previous neoplasia or other serious medical problems were not eligible for the study. People with any of the following risk factors for cyclosporine nephrotoxicity were also excluded16: estimated creatinine clearance less than 78 µmol per minute; continuous use of nonsteroidal antiinflammatory drugs; hypertension (blood pressure greater than 165/90 mm Hg or the use of antihypertensive drugs); or diabetes mellitus. All patients gave written informed consent in accordance with the provisions of the Second Helsinki Declaration.
The Crohn's Disease Activity Index19,20 incorporates eight items: the number of liquid or very soft stools, abdominal pain, general well-being, extraintestinal manifestations of Crohn's disease, use of certain opiates to treat diarrhea, abdominal mass, hematocrit, and body weight. The index provides a composite score ranging from 0 to approximately 600. Higher scores indicate more disease activity. Patients with scores below 150 are considered to have inactive disease. Patients with scores above 450 are considered to be seriously ill.
Scores on the Crohn's Disease Activity Index were determined for potentially eligible patients. This determination was repeated six weeks later, and patients whose scores had increased by 100 points or more (indicating rapidly deteriorating disease) were excluded from the study.
Eligible patients were randomly assigned to receive cyclosporine or placebo. The randomization was stratified according to center and disease activity (as determined by a score of 
150 or >150 on the Crohn's Disease Activity Index).
Intervention
Cyclosporine (Sandimmune, Sandoz) was given orally at an initial daily dose of 2.5 mg per kilogram in two divided doses with meals. Placebo was given in capsules identical in appearance to the active drug. After one week, the dose was increased to 5 mg per kilogram. This stepped dose escalation was chosen to minimize minor toxic effects, such as gastrointestinal upset. Thereafter, the dose was increased as needed to a maximal daily dose of 15 mg per kilogram in order to attain a whole-blood trough concentration of 200 ng per milliliter. If a patient could not take medications orally or required surgery, the study drug was resumed when the patient was again able to take oral medication. Intravenous cyclosporine was not used.
Other Treatments for Crohn's Disease
The treatment of Crohn's disease was standardized. Patients received prednisone or a 5-aminosalicylate (sulfasalazine [Salazopyrin], mesalamine [Asacol or Salofalk], or olsalazine [Dipentum]) to control symptoms. The maximal daily doses were as follows: prednisone, 60 mg; sulfasalazine, 8 g; and other 5-aminosalicylates, 4 g. The prednisone dose was reduced with improvement in the patient's condition and increased with worsening of disease. Prednisone tapering after disease flares followed an eight-week schedule. The use of hydrocortisone enemas was permitted for Crohn's disease distal to the splenic flexure of the colon.
Treatment with metronidazole, azathioprine, mercaptopurine, or methotrexate was not permitted. Up to three two-week courses of nonsteroidal antiinflammatory drugs were allowed yearly. Drugs with pharmacokinetic or toxic interactions with cyclosporine, such as erythromycin, were not permitted, and patients who required such medications were withdrawn from the study. Surgery for complications of Crohn's disease or refractory disease was performed as required.
Follow-up
Patients were seen every two months or as necessary. The physicians treating them were not provided with the patients' scores on the Crohn's Disease Activity Index. At each visit a score on this index, the quality-of-life score,18,21 and the global ratings of disease activity by the physician and the patient were obtained, and any use of other medications was recorded. The cyclosporine doses were adjusted by clinicians who were aware of the group assignments but who had no contact with the patients. Cyclosporine and creatinine levels were monitored every two weeks to minimize nephrotoxicity and maintain a trough blood cyclosporine level of 200 ng per milliliter. The dose of the study drug was reduced by 50 percent, regardless of the blood level, if the patient's serum creatinine concentration increased by more than 30 percent from base line or if the blood pressure exceeded 160/90 mm Hg. Patients were withdrawn from the study if two consecutive dose reductions did not resolve the abnormality. Matching dose adjustments were made in the placebo group to preserve blinding. Cyclosporine was measured by a radioimmunoassay of whole blood (Incstar)22 that was standardized by a central laboratory.
Outcome Measures
The primary outcome measure was a 100-point increase in the Crohn's Disease Activity Index19,20. A blinded adjudication committee assessed the clinical validity of all potential primary outcomes, using the patient's records and the global ratings by the patient and the physician (on seven-point Likert scales). The records of patients who underwent surgery or who were withdrawn from the study because of possible adverse drug reactions were also reviewed. The secondary outcome measures were the mean scores on the Crohn's Disease Activity Index, the mean quality-of-life scores, and the mean doses of prednisone and 5-aminosalicylates.
Statistical Analysis
All the statistical comparisons were performed according to the intention-to-treat principle. Base-line characteristics measured on a nominal or an ordinal scale were compared by chi-square procedures; continuous measures were compared by t-tests. Important base-line variables were entered as covariates.
Kaplan-Meier curves for the time to withdrawal from the study, worsening of disease, and a requirement for surgery were compared by the likelihood-ratio test,23 with adjustment for base-line covariates by the Cox proportional-hazards model24. A repeated-measures analysis25 was applied to the Crohn's Disease Activity Index and the quality-of-life measure. Total amounts of prednisone and 5-aminosalicylates used were compared by linear models26. The incidence of adverse events was compared by Fisher's exact test. Statistical tests were two-sided, with an alpha error of 0.05. The analyses were performed with SAS27 and BMDP28 software.
An analysis of efficacy was performed in which patients were excluded if a primary outcome event occurred within the first 28 days or if patients in the cyclosporine group had blood cyclosporine levels below 150 ng per milliliter more than 50 percent of the time. The randomization of 300 patients gave the study 90 percent power to detect an absolute risk reduction of 20 percent for the primary outcome, assuming that the rate of worsening in the placebo group was 50 percent over the study period29.
Results
A total of 1448 patients were screened, and 376 potentially eligible patients were identified. The most common reasons for ineligibility were refusal to participate by the patient or the patient's physician (38 percent), inactivity of disease for more than two years (26 percent), use of contraindicated medication (15 percent), and renal insufficiency (9 percent). Seventy-one patients who were initially included in the study were not randomized: 35 were found to be ineligible on further review, 24 declined randomization, and 12 had a worsening of their disease in the period before randomization. The age, sex, duration of disease, and site of disease of these 71 patients did not differ significantly from those of the patients who were randomized.
The base-line characteristics of the cyclosporine and placebo groups were similar (Table 1). There was a significant difference between groups in the percentage of patients who had undergone previous surgery (cyclosporine, 42 percent; placebo, 60 percent; P = 0.03). This difference was taken into account in the multivariate model.
|
No patients were lost to follow-up. More patients in the cyclosporine group than in the placebo group withdrew from the study (58 of 151 receiving cyclosporine, or 38.4 percent, as compared with 41 of 154 receiving placebo, or 26.6 percent; P = 0.03). More patients receiving cyclosporine had worsening of disease (91 of 151 receiving cyclosporine, or 60.3 percent, as compared with 80 of 154 receiving placebo, or 51.9 percent; P = 0.10). The median time to the worsening of disease was 338 days in the cyclosporine group, as compared with 492 days in the placebo group (relative risk, 1.22; 95 percent confidence interval, 0.86 to 1.72; P = 0.25) (Figure 1). Separate life-table analyses of the patients with a high or low level of activity of disease also did not show a benefit of cyclosporine treatment (Figure 1). More patients had worsening of their condition with cyclosporine than with placebo in the low-activity stratum (64 of 99 receiving cyclosporine, or 64.6 percent, as compared with 47 of 94 receiving placebo, or 50.0 percent; P = 0.11). The proportion of patients with worsening disease in the high-activity stratum was similar for both treatments (27 of 52 receiving cyclosporine, or 51.9 percent, as compared with 33 of 60 receiving placebo, or 55.0 percent; P = 0.73). A total of 39 patients were withdrawn from the study because of worsening of disease (21 of 151 receiving cyclosporine, or 13.9 percent, as compared with 18 of 154 receiving placebo, or 11.7 percent; P = 0.56).
|
Effect of Cyclosporine Treatment on Secondary Outcomes
No differences were detected in the overall comparisons of mean scores on the Crohn's Disease Activity Index (P = 0.31) and mean quality-of-life scores (P = 0.49). These results are presented according to level of disease activity in Figure 2. In the low-activity stratum, the patients in the placebo group had a significantly lower mean score on the Crohn's Disease Activity Index (P = 0.02).
|
During the study 49 patients required surgery, most commonly for refractory symptoms (18 patients) or bowel obstruction (12 patients). Similar numbers of patients in the cyclosporine and placebo groups underwent a single operation (25 of 151 receiving cyclosporine, or 16.6 percent, as compared with 24 of 154 receiving placebo, or 15.6 percent; P = 0.39). Three patients (two receiving cyclosporine and one receiving placebo) underwent a second operation. The indications for surgery were similar in the two groups.
Cyclosporine Dose and Blood Levels, and Renal Function
Table 2 shows the daily dose and blood levels of cyclosporine and the serum creatinine concentration. The mean dose of cyclosporine (4.8 ±1.6 mg per kilogram) was associated with an overall blood cyclosporine level of 182 ±61 ng per milliliter and a maximal increase from base line in the creatinine level of 16.9 percent. The mean target blood cyclosporine level of 200 ng per milliliter was not achieved because dose reductions were required to minimize nephrotoxicity. Fifteen patients (9.9 percent) had malabsorption of cyclosporine, defined as a blood cyclosporine level below 75 ng per milliliter on more than 50 percent of determinations, despite dose increases and confirmation of compliance.
|
Adverse Events and Cyclosporine Toxicity
Cyclosporine was well tolerated. Four malignant neoplasms developed: one lung cancer and one Dukes' B colon cancer in the cyclosporine group, and one metastatic adenocarcinoma and one testicular carcinoma in the placebo group. The patient with colon carcinoma was randomized in error and already had a villous adenoma, a lesion with known malignant potential. Three patients died, one in the cyclosporine group (from lung cancer) and two in the placebo group (one by suicide and one from metastatic adenocarcinoma).
Withdrawal due to adverse events was more frequent in the cyclosporine group (22 of 151, or 14.6 percent, as compared with 5 of 154 receiving placebo, or 3.2 percent; P = 0.003). Seven of these withdrawals were required by the protocol because of increases in the serum creatinine level or the development of hypertension. Although paresthesias, headaches, and hypertrichosis occurred more commonly with cyclosporine, these problems caused the withdrawal of only eight patients. Two cyclosporine-treated patients had uncomplicated pregnancies.
Discussion
We found that the addition of low-dose cyclosporine to conventional treatment for Crohn's disease did not improve symptoms or reduce requirements for other forms of therapy. Given the size of this study and the comprehensive nature of the outcome measures, it is unlikely that a clinically important benefit of cyclosporine went undetected. Although the small increase in disease activity observed in the low-activity stratum in the cyclosporine group as compared with the placebo group may have been a chance occurrence, it is also plausible that cyclosporine increased patients' symptoms through drug-related side effects or by direct action on the disease process.
The primary outcome was a 100-point increase in the Crohn's Disease Activity Index, a well-validated measure. One potential problem with this index is that it includes the hematocrit among the items evaluated, with a lower hematocrit indicating higher disease activity. Cyclosporine may produce a mild decrease in the hematocrit. Brynskov et al.15. used a modified grading score that did not include the hematocrit7. It is likely, however, that our negative result was due to a lack of drug efficacy. The absolute difference in the mean hematocrit was small (39.7 percent in the cyclosporine group vs. 38.1 percent in the placebo group, P = 0.001). Moreover, an additional analysis performed with the values for the hematocrit deleted from the index did not alter the result.
Our findings might be explained by the relatively low dose of cyclosporine used. The active-treatment group had a mean blood cyclosporine level of 182 ±61 ng per milliliter, slightly below the target level of 200 ng per milliliter. No beneficial effect of cyclosporine was demonstrated, however, in a subgroup of 83 patients who had a mean cyclosporine level (218 ±45 ng per milliliter) higher than the target level. Brynskov et al.15. showed a small benefit of cyclosporine in Crohn's disease with a higher dose (median dose, 7.8 mg per kilogram) that produced higher blood cyclosporine levels. The median blood cyclosporine level of the patients in that study who had improvement was substantially higher than that of the patients who did not respond (425 vs. 212 ng per milliliter). Therefore, the lack of effect in patients with low blood levels is consistent in that study and ours.
Cyclosporine is effective in other diseases at doses and blood levels similar to those we selected12,13,14. Chronic cyclosporine nephrotoxicity is dose-dependent, with the risk increasing as the daily dose exceeds 5 mg per kilogram30. We cannot exclude the possibility that higher doses of cyclosporine or intravenous administration of the medication may be beneficial, as has been suggested in an uncontrolled trial31. Nevertheless, we believe that a higher dose is unlikely to be useful for the long-term management of Crohn's disease because of the risk of nephrotoxicity.
Low-dose cyclosporine was well tolerated. Serious nephrotoxicity was avoided by the use of the drug in a low-risk population and by careful monitoring16. The tumors that occurred in the cyclosporine group, including fatal lung cancer in a smoker, were not those generally attributable to immunosuppression.
Further research is needed on the role of immunosuppression in Crohn's disease. The most effective drugs -- prednisone and, to a lesser extent, purine analogues -- have broad-based immunosuppressive activity32. This may be a prerequisite for the control of Crohn's disease, in which a wide array of immune and inflammatory pathways are activated. In established disease, nonspecific immune mechanisms may be critical and resistant to the action of cyclosporine, a drug that modulates T-cell function.
Supported under a University Industry Award Grant (1875) from Sandoz Pharma, Dorval, Que., Canada, and by the Medical Research Council of Canada.
We are indebted to the patients who participated in the study and to the following companies for supplying study medications: Sandoz Pharma (cyclosporine, placebo), Incstar Corporation (cyclosporine radioimmunoassay kits), Janssen Pharmaceutica (loperamide hydrochloride), Norwich Eaton Pharmaceuticals (mesalamine), Pharmacia (Canada) (sulfasalazine), and the Upjohn Company of Canada (prednisone).
Source Information
From the Department of Medicine, Division of Gastroenterology, University of Alberta, Edmonton (R.N.F.); the Department of Medicine, Division of Gastroenterology, McMaster University, Hamilton, Ont. (E.J.I.); the Department of Medicine, Division of Gastroenterology, Queen's University, Kingston, Ont. (A.G.); the Departments of Medicine (B.G.F., J.W.D.M., B.V.) and Epidemiology and Biostatistics (B.G.F., J.R.), University of Western Ontario, London; the Department of Medicine, Division of Gastroenterology, McGill University, Montreal (D.K.); the Faculty of Medicine, University of Montreal, Montreal (A.A.); the Department of Medicine, Division of Gastroenterology, University of Toronto, Toronto (F.S.); and the Departments of Medicine (A.L., R.G.) and Epidemiology and Community Medicine (A.L.), University of Ottawa, Ottawa, Ont. -- all in Canada. The Canadian Crohn's Relapse Prevention Trial Investigators are listed in the Appendix.
Address reprint requests to Dr. Feagan at 6 OF 11 University Hospital, 339 Windermere Rd., London, ON N6A 5A5, Canada.
References
The Canadian Crohn's Relapse Prevention Trial Investigators were as follows.
Steering Committee: A. Archambault, B. Feagan, R. Fedorak, R. Groll, E.J. Irvine, D. Kinnear, A. Laupacis, J.W.D. McDonald (chair), J. Rochon, F. Saibil, and B. Valberg; Adjudication Committee: B. Feagan, R. Gillies, and J.W.D. McDonald (chair); Inclusion Criteria Review Panel: E.J. Irvine (chair), R. Riddell, T. Seaton, and S. Somers; Operations Committee: B. Feagan (chair), J.W.D. McDonald, J. Rochon, and B. Valberg; External Advisory Committee: J. Dirks, G. Feutren, K. Jeejeebhoy, and D. Sackett (chair); Unblinded clinicians: R. Dandavino, C.N. Ghent (chair), J.R. Grynoch, A.M. Holbrook, B.A. Kiberd, N. Kneteman, M. Levine, M. Manuel, N.N. Muirhead, C.S. Saiphoo, and P.J. Somerville; Coordinating center: L. Cameron, T. Lockwood, E. Seglenieks, K. Taylor-Dolmer, and B. Valberg.
The following centers and investigators participated in the study: University of Alberta, Edmonton (68 patients): R. Cherry, R. Fedorak, D. Fisher, P. Kirdeikis, V. Mahachai, T. Sedens, R. Sherbaniuk, A. Thompson, and R. Wensel; McMaster University, Hamilton, Ont. (65 patients): M. Castelli, S. Collins, K. Croitoru, S. Crowe, M. Donnelly, R. Goodacre, R. Hunt, E.J. Irvine, B. Lumb, R. Rossman, B. Salena, T. Seaton, and B. Stoskopf; Queen's University, Kingston, Ont. (16 patients): I. Beck, L. Dacosta, D. D'Amore, W. Depew, A. Groll, J. Hudacin, W. Paterson, and J. Simon; University of Western Ontario, London (92 patients): P. Adams, B. Anderson, W. Barnett, M. Belsheim, D. Bondy, W. Ebers, B. Feagan, P. Gilmore, M. Hopkins, J. Howard, A. Laupacis, D. Lloyd, J.W.D. McDonald, T. Ponich, I. Prokopiw, R. Reynolds, C. Stiller, L. Valberg, and W. Watson; McGill University, Montreal (33 patients): A. Aronoff, S. Blum, D.P. Cleland, D. Daly, H. Daoud, W.D. Dauphinee, C. Goresky, M. Jabbari, S. Katz, J. Kessler, D. Kinnear, M. Lichter, D. Mills, S. Mishkin, P. Mlynaryk, M. Poleski, P. Szego, A. Szilagyi, and G. Wild; University of Montreal, Montreal (13 patients): A. Archambault, G. Aumais, S. Cote, M.-F. Courteau, S. Dube, F. Heyen, G. Jobin, J. Marchand, R. Mousseau, and F. Thibert; University of Toronto, Toronto (18 patients): J. Andrews, L. Cohen, S. Pearen, and F. Saibil; Sandoz Pharma: B. von Graffenried, G. Feutren, and L. Beauregard-Zollinger.
| |||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
