Intravenous Hydrocortisone Premedication Reduces Antibodies to Infliximab in Crohn s Disease: A Randomized Controlled Trial

Transcription

1 GASTROENTEROLOGY 2003;124: Intravenous Hydrocortisone Premedication Reduces Antibodies to Infliximab in Crohn s Disease: A Randomized Controlled Trial RICHARD J. FARRELL, MAZEN ALSAHLI, YOON TAE JEEN, KENNETH R. FALCHUK, MARK A. PEPPERCORN, and PIERRE MICHETTI Division of Gastroenterology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts See editorial on page Background & Aims: We assessed the relationshipbetween antibodies to infliximab (ATI) and the loss of response postinfliximab, infusion reactions and, in a randomized trial, investigated whether intravenous hydrocortisone premedication can reduce ATI. Methods: Initially, we prospectively evaluated clinical response, adverse events, and ATI levels in 53 consecutive patients with Crohn s disease who received 199 infliximab (5 mg/kg) infusions. Subsequently, 80 patients with Crohn s disease were randomized to intravenous hydrocortisone 200 mg or placebo immediately before their first and subsequent infliximab infusions. The primary endpoint was reduction in median ATI levels at week 16. Analysis was by intention to treat. Results: Nineteen of our initial 53 patients (36%) developed ATI, including all 7 patients with serious infusion reactions (median ATI level, 19.6 g/ml). Eleven of 15 patients (73%) who lost their initial response were ATI positive compared with none of 21 continuous responders, (8.9 vs. 0.7 g/ml, P < ). Administering a second infusion within 8 weeks of the first (OR, 0.13; 95% CI, ; P ) or concurrent immunosuppressants (OR, 0.19; 95% CI, ; P 0.007) significantly reduced ATI formation. In the placebo-controlled trial, ATI levels were lower at week 16 among hydrocortisone-treated patients (1.6 vs. 3.4 g/ml, P 0.02), and 26% of hydrocortisone-treated patients developed ATI compared with 42% of placebo-treated patients, P Conclusions: Loss of initial response and infusion reactions post-infliximab is strongly related to ATI formation and level. Administering a second infusion within 8 weeks of the first and concurrent immunosuppressant therapy significantly reduce ATI formation. Intravenous hydrocortisone premedication significantly reduces ATI levels but does not eliminate ATI formation or infusion reactions. The emergence of infliximab (Remicade; Centocor, Inc., Malvern, PA), a chimeric (part mouse, part human) monoclonal IgG1 antibody directed against tumor necrosis factor, represents an important advance in the treatment of refractory Crohn s disease. Clinical efficacy of infliximab in patients with refractory luminal Crohn s disease, 1 and fistulous Crohn s disease, 2 was shown in 2 controlled trials, leading to its Food and Drug Administration (FDA) approval in To date, clinical experience with infliximab has been reassuring regarding safety, and clinical response rates have mirrored the efficacy reported in controlled trials. 3 6 Following the ACCENT 1 trial, 7 the FDA approved a 3-dose induction regimen of 5 mg/kg infusions at 0, 2, and 6 weeks, with maintenance dosing every 8 weeks thereafter for patients with refractory luminal Crohn s disease. However, the role of maintenance infliximab therapy is controversial, and, whereas most gastroenterologists currently retreat responsive patients upon clinical relapse, a substantial proportion of patients may lose their initial response on retreatment. Immunogenicity is an emerging issue, which may significantly limit the long-term efficacy of infliximab retreatment through the formation of antibodies to infliximab (ATI). In clinical trials, ATI occur in 13% 18% of patients and are associated with infusion reactions. 8 Although clinical studies suggest that concurrent immunosuppressants can reduce ATI formation, 7,8 there have been no controlled trials assessing the ability of immunosuppressant premedication in preventing ATI. We assessed the relationship between ATI and loss of response post-infliximab therapy and infusion reactions and, in a subsequent randomized trial, investigated whether intravenous hydrocortisone premedication can reduce ATI. Abbreviations used in this paper: 95% CI, 95% confidence interval; ATI, antibodies to infliximab; CDAI, Crohn s disease activity index; ELISA, enzyme linked immunosorbent assay; FDA, Food and Drug Administration; OR, odds ratio by the American Gastroenterological Association /03/$30.00 doi: /gast

2 918 FARRELL ET AL. GASTROENTEROLOGY Vol. 124, No. 4 Patients and Methods Study 1: Response and ATI Formation in Patients Treated With Infliximab Patients. Fifty-three consecutive patients with refractory Crohn s disease referred for infliximab (5 mg/kg) infusions (total infusions, 199; range, 2 13) between November 1998 and October 2000 were included in our initial observational study. Patients with active luminal disease received 1 infusion, and patients with fistulous disease received 3 infusions at weeks 0, 2, and 6. Patients who relapsed were retreated with infliximab on an on-demand basis. Patients with luminal disease who did not respond to their first infusion received a second infusion. All infusions were administered over 2 hours, according to a recommended protocol. 8 Assessments. Crohn s disease activity index (CDAI) was calculated at baseline, and all patients were prospectively evaluated for clinical response, adverse events, and concurrent immunosuppressants. Infusion reactions were defined as any adverse event occurring during infusion or during the 2 hours postinfusion. A serious infusion reaction was defined as any reaction that required infliximab to be discontinued. Concurrent Crohn s disease medications including immunosuppressants were used as clinically indicated. Concurrent immunosuppressive therapy included either prednisone 20 mg/day or azathioprine, 6-mercaptopurine, or methotrexate for at least 3 months at the time of first infusion. Diphenhydramine and acetaminophen premedication were given if patients had prior, serious infusion reactions. Disease activity was determined in all patients 4 weeks after each infusion, using a previously reported 3-point scale: 0 no change or worsening of symptoms, 1 persistence of some clinical symptoms or some fistula drainage or response lasting less than 4 weeks, and 2 absence or near absence of all clinical symptoms without an increase in steroid dose or closure of all fistulas. 4 Patterns of response were stratified as follows: patients who continued to respond to successive infusions (continuous responders), patients who initially responded but subsequently lost their response (past responders), patients who from the outset had only a partial or short-lived response (partial responders), and patients who never responded (nonresponders). Sera were drawn at baseline and approximately 24 weeks (range, 8 52 weeks) following their first infusion to determine ATI and infliximab levels. Study 2: Randomized Trial of Intravenous Hydrocortisone Premedication in Reducing ATI Patients. In a subsequent 16-week, double-blind, placebo-controlled trial performed between October 2000 and December 2001, 80 consecutive patients with refractory Crohn s disease were randomized to 200-mg hydrocortisone IV premedication or placebo immediately prior to receiving their first infliximab 5-mg/kg infusion. The protocol was approved by our institutional review board, and patients gave written consent before enrollment. Randomization was performed by the research pharmacy, and a pharmacist prepared the 200-mg hydrocortisone bolus or an identically appearing placebo. Patients were retreated with infliximab upon clinical relapse on an on-demand schedule, and those who had multiple infusions received either intravenous hydrocortisone or placebo in a double-blind manner before all subsequent infusions as dictated by which group they were randomized to at enrollment. Assessments. All patients were prospectively evaluated at baseline and at weeks 8 and 16 after their first infusion for clinical efficacy, adverse events, concurrent immunosuppressants, and ATI and infliximab concentration levels. Clinical response was assessed by CDAI calculation. All primary and secondary endpoints were predefined, and all analyses were by intention to treat. The primary endpoint was reduction in median ATI concentration levels at week 16 because of less circulating infliximab and consequently fewer indeterminate ATI results at week 16 compared with week 8. Secondary endpoints were reduction in the number of ATI-positive patients at weeks 8 and 16, reduction in median ATI levels at 8 weeks, reduction in baseline CDAI score by 70 points (clinical response) and to 150 points (clinical remission) at weeks 8 and 16, and reduction in adverse events including infusion reactions. ATI and Infliximab Measurement Quantitative ATI and infliximab ELISA immunoassays (Prometheus Laboratories, Inc., San Diego, CA) were performed blindly in duplicate on sera and measured as described previously. 9 Investigators were unaware of the patients ATI status during the study. Because the presence of infliximab interferes with the ATI assay, ATI results were reported as follows: positive if the mean ATI level was 1.69 g/ml and infliximab level was 1.4 g/ml; negative if the mean ATI level was 1.69 g/ml and infliximab level was 1.4 g/ ml, and indeterminate if infliximab levels were 1.4 g/ml. Patients in study 1 whose follow-up sera was indeterminate for ATI had repeat seram drawn at least 8 weeks after their last infusion to clarify their ATI status. Statistical Analysis Data analysis was performed using Statview 4.5 for Macintosh (Abacus Inc. Berkeley, CA). We estimated that 40 patients in each of the hydrocortisone and placebo groups would provide 90% power to detect a true difference of 0.7 population standard deviations in the primary outcome of ATI levels between the 2 groups ( 0.05, 2-sided). Because ATI levels in the various groups were skewed, median levels were calculated, and nonpaired analyses were conducted on the natural log scale using Dunnett s test. We examined associations between other potentially predictable variables and the outcomes using 2 tests or 2-tailed Fisher exact tests for categorical variables and Student t tests for continuous variables. ATI status in study 1 was entered into a multivariable, logistic regression model along with age, sex, concurrent immunosuppressants, fistulous disease, multiple infusions, and

3 April 2003 PREMEDICATION REDUCES ANTIBODIES TO INFLIXIMAB 919 Table 1. Baseline Demographics of Observational Cohort Patients Continuous responders Past responders Partial responders Nonresponders No Age (yr) and sex Mean yr (range) 40 (26 65) 43 (23 62) 46 (23 72) 39 (22 62) Female (%) 13 (61.9) 11 (73.3) 6 (75) 7 (77.7) Location of disease (%) Ileitis 6 (28.6) 4 (26.6) 1 (12.5) 2 (22.2) Ileocolitis 9 (42.9) 7 (46.6) 5 (62.5) 5 (55.5) Colitis 6 (28.6) 4 (26.6) 2 (25) 2 (22.2) Disease type (%) Luminal 9 (42.9) 9 (60) 7 (87.5) 8 (88.8) Fistulous 12 (57.1) 6 (40) 1 (12.5) 1 (11.1) Concurrent therapy at 1st infusion (%) 5-ASA 12 (57) 7 (46.6) 4 (50) 4 (44.4) Prednisone 20 mg/day 2 (9.5) 2 (13.3) 2 (25) 2 (22.2) Prednisone 20 mg/day 7 (33.3) 5 (33.3) 4 (50) 6 (66.6) Azathioprine or 6-MP 8 (38) 3 (20) 2 (25) 3 (33.3) Methotrexate 3 (14.3) 1 (6.6) 1 (12.5) 0 Pred 20 mg, AZA, 6-MP, MTX 16 (76.2) 8 (53.3) 6 (75) 7 (77.8) Smoking history Current 3 (14.3) 4 (26.6) 1 (12.5) 2 (22.2) Former 3 (14.3) 2 (13.3) 2 (25) 3 (33.3) Never 15 (71.4) 9 (60) 5 (62.5) 4 (44.4) Disease activity, mean (range) CDAI 272 ( ) 285 ( ) 305 ( ) 265 ( ) Total number of infusions Number of infusions per patient (range) 4.2 (2 13) 4.4 (2 7) 3.0 (2 9) 2.3 (2 4) NOTE. Study 1. atopy status, because we hypothesized a priori that these might confound ATI formation. An adjustment for multiple comparisons was made for an interim analysis of the data performed when the first 40 patients were enrolled in study Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated. The level was set at All P values were 2-sided. Results Study 1: Response and ATI Formation in Patients Treated With Infliximab The demographics of the 53 patients are shown in Table 1. Forty-five percent of patients received infliximab for fistulous disease, and 36 patients (68%) responded to their initial infusion. At median follow-up of 20 weeks (range, 4 92 weeks), there were 21 continuous responders (40%), 15 past responders (28%), 8 partial responders (15%), and 9 nonresponders (17%). Fortytwo percent of patients who initially responded to infliximab therapy lost their response. Baseline variables, including age, sex, disease location, smoking history, disease type, and concurrent immunosuppressants, were similar among all 4 groups. At follow-up, 19 of 53 patients (36%) developed ATI (median ATI level, 16.1 g/ml; range, g/ml). All 7 patients (4.5%) who developed serious infusion reactions were ATI positive (19.6 g/ml; range, g/ml) with all but 1 of the 7 patients having ATI levels greater than 16.0 g/ml. At follow-up, none of 21 continuous responders were ATI positive compared with 11 of 15 past responders (73%), 2 of 8 partial responders (25%), and 6 of 9 nonresponders (67%) (Figure 1). Continuous responders had lower median ATI levels (0.7 g/ml) than past responders (8.9 g/ml, P ), partial responders (1.0 g/ml, P 0.009), and nonresponders (4.5 g/ml, P ) (Figure 2). Patients on concurrent immunosuppressants at the time of their first infliximab infusion had a significantly lower incidence of ATI formation compared with patients not on immunosuppressants, 24% vs. 63%, respectively, P We found no relationship by univariate analysis between ATI formation and age, sex, current smoking, history of atopy, concurrent aminosalicylates, or prednisone ( 20 mg/day) at the time of first infliximab infusion. In a multivariable, logistic regression model, administration of a second infliximab infusion within 8 weeks of the first (OR, 0.13; 95% CI, ; P ) and concurrent immunosuppressants at the time of first infusion (OR, 0.19; 95% CI, ; P 0.007) were significant independent

4 920 FARRELL ET AL. GASTROENTEROLOGY Vol. 124, No. 4 Figure 1. Relationship between response patterns and ATI formation in initial observational cohort of 53 patients. At follow-up, none of 21 continuous responders (0%) were ATI positive compared with 11 of 15 past responders (73%) (P ), 2 of 8 partial responders (25%) (P 0.02), and 6 of 9 nonresponders (67%), (P ). Study 2: Randomized Trial of Intravenous Hydrocortisone Premedication to Reduce ATI Of 83 patents screened for eligibility, 3 were excluded because corticosteroids were contraindicated (Figure 3). Eighty patients were randomly allocated to receive intravenous hydrocortisone (n 39) and intravenous placebo (n 41). Baseline characteristics between the placebo and hydrocortisone groups were similar, including the number of patients on concurrent immunosuppressants (42% vs. 44%, respectively; P 0.9) and the mean number of infusions per patient (2.1 vs. 1.9, respectively; P 0.8; Table 3). All patients were ATI negative at baseline and had similar ATI levels, (median ATI levels 0.61 vs. 0.64, respectively, g/ml; P 0.4). Follow-up sera at weeks 8 and 16 post-first infusion were available in 34 of 41 placebo-treated patients (83%) and 34 of 39 hydrocortisone-treated patients (87%; P 0.8). Seven patients, none of whom responded to their initial infliximab infusion, and 5 patients from out of state, reflecting our wide tertiary referral base, were lost to follow-up. ATI levels were protective factors against ATI. Further analysis demonstrated that concurrent azathioprine, 6-mercaptopurine, or methotrexate of 3 months duration (OR, 0.16; 95% CI, ; P 0.007) and concurrent prednisone of 20 mg/day (OR, 0.24; 95% CI, ; P 0.01) were significant independent protective factors for developing ATI. To assess the predictive value of an ATI-positive result, we prospectively assessed the clinical outcome in a subgroup of 44 patients who received a total of 68 infliximab infusions after follow-up sera was drawn determining their ATI status (Table 2). Compared with an ATI-negative patient, an ATI-positive patient had a significantly higher nonresponse rate to subsequent infusions (52% vs. 14%, respectively; P ), a significantly shorter median duration of response (28 vs. 61 days, respectively; P 0.007), and a significantly higher incidence of subsequent infusion reactions (40% vs. 4.7%, respectively; P ), including serious infusion reactions (28% vs. 0%, respectively; P ), and ATI levels 8 g/ml prior to a subsequent infusion predicted a significantly higher risk of serious infusion reaction (OR, 4.2; 95% CI, ) and a significant higher risk of nonresponse (OR, 2.0; 95% CI, ). Figure 2. Relationship between response patterns and median ATI concentration levels in initial observational cohort of 53 patients. The boxes indicate the 25th percentiles, median values, and 75th percentiles; the bars indicate the 10th and 90th percentiles; and the circles indicate outliers. Continuous responders had a median ATI level of 0.7 g/ml. By comparison, median ATI levels were significantly higher among past responders (8.9 g/ml, P ), partial responders (1.0 g/ml, P 0.009), and nonresponders (4.5 g/ml, P ).

5 April 2003 PREMEDICATION REDUCES ANTIBODIES TO INFLIXIMAB 921 Table 2. Clinical Outcome and Adverse Events in Subgroup of 44 Patients Who Received 68 Infliximab Infusions After ATI Status Known to be Positive or Negative ATI negative (n 28 patients) ATI status ATI positive (n 16 patients) P value Total number of infusions Clinical outcome Complete response (grade 2) (%) 30 (70) 0 (0) Partial or short response (grade 1) (%) 7 (16.3) 12 (48) No response (grade 0) (%) 6 (13.9) 13 (52) Duration of response, day, (mean SD) Adverse events Early adverse events (%) 18 (41.9) 18 (72) Infusion reactions (%) 2 (4.7) 10 (40) Serious infusion reactions (%) 0 7 (28) Late adverse events (%) 5 (11.6) 4 (16) 0.6 significantly lower in the hydrocortisone group compared with the placebo group at week 8 (2.9 vs g/ml, respectively; P 0.002) and at week 16 (1.6 vs. 3.4 g/ml, respectively; P 0.02) (Figure 4). At weeks 8 and 16 post-first infusion, the hydrocortisone group had fewer ATI-positive patients, 8 of 39 (21%) vs. 12 of 41, respectively, (29%; P 0.14) and 10 of 39 (26%) vs. 17 of 41, respectively (42%; P 0.06; Figure 5). Univariate analysis of the placebo group showed that concurrent immunosuppressants reduced ATI formation at week 8 (OR, 0.22; 95% CI, ; P 0.04) and week 16 (OR, 0.06; 95% CI, ; P ). Although the hydrocortisone group had higher clinical response and remission rates at weeks 8 and 16 compared with the placebo group, and fewer patients who experienced infusion reactions (24% vs. 15%, P 0.06), the differences were not statistically significant (Table 4). There were 2 serious infusion reactions among Figure 3. Trial profile. the placebo group compared with 1 serious infusion reaction in the hydrocortisone group (5% vs. 2.5%, respectively, P 0.6). Discussion The therapeutic benefits of monoclonal antibodies are now firmly established. 11 However, immunogenicity Table 3. Baseline Characteristics of Placebo/Hydrocortisone Trial Patients Characteristic Placebo (n 41) Hydrocortisone (n 39) Age and sex Mean, yr (range) 38 (22 72) 41 (22 69) Female (%) 26 (63) 24 (62) Location of disease Ileitis 9 (22) 10 (26) Ileocolitis 22 (54) 20 (51) Colitis 10 (24) 9 (23) Disease type Luminal 29 (71) 30 (77) Fistulous 12 (29) 9 (23) Concurrent therapy at 1st infusion (%) 5-ASA 20 (49) 18 (46) Prednisone 20 mg/day 8 (20) 9 (23) Prednisone 20 mg/day 6 (15) 8 (21) Azathioprine or 6-MP 13 (32) 13 (33) Methotrexate 1 (2.4) 2 (5) Prednisone 20 mg, AZA, 6-MP, MTX 17 (42) 17 (44) Smoking history (%) Current 6 (15) 7 (18) Former 13 (32) 8 (21) Never 22 (54) 24 (62) Disease activity, mean (range) CDAI 262 ( ) 315 ( ) Total number of infusions at 16 wk Mean number of infusions per patient (range) 2.12 (1 3) 1.85 (1 3) NOTE. Study 2.

6 922 FARRELL ET AL. GASTROENTEROLOGY Vol. 124, No. 4 Figure 4. Median ATI concentration levels at baseline and at 8 and 16 weeks post-first infusion according to treatment (placebo/hydrocortisone) group. The boxes indicate the 25th percentiles, median values, and 75th percentiles; the bars indicate the 10th and 90th percentiles; and the circles indicate outliers. Median ATI levels were significantly lower in the hydrocortisone group compared with the placebo group at 8 weeks, (2.9 vs g/ml, respectively; P 0.002) and at 16 weeks (1.6 vs. 3.4 g/ml, P 0.02). to murine, chimeric, and humanized biologic agents is emerging as an important limitation of retreatment. Although consideration should be given to increasing the infliximab dose from 5 mg/kg to 10 mg/kg in patients who lose their initial response, 7 this strategy remains to be proven effective in clinical practice. Although immunogenicity can be obviated by completely humanized monoclonal antibodies, clinical trials in Crohn s disease with the humanized monoclonal antibody CDP571 suggest that this approach is less efficacious than chimeric antibodies. 12 Immunogenicity may also be obviated by suppression of the immune response, and clinical trial data with infliximab suggest that concurrent immunosuppressant therapy reduces ATI formation and infusion reactions, 7,8 hence the increasing interest in premedication strategies to reduce the immunogenicity of chimeric monoclonal antibodies. 13 Our initial observational study corroborates the established relationship between ATI formation and infusion reactions 7,8 but also suggests that both ATI formation and ATI level play an important role in mediating tachyphylaxis with infliximab retreatment. Forty-two percent of patients lost their initial response to infliximab and almost three fourths of these were ATI positive with a median level of 8.9 g/ml. By comparison, no patients who continued to respond to retreatment developed ATI. ATI levels over 8 g/ml prior to a subsequent Figure 5. ATI formation according to treatment group. Percentage of patients with Crohn s disease who were ATI positive ( ), ATI negative ( ), and ATI indeterminate ( ) at baseline and at 8 and 16 weeks after first infusion according to treatment (placebo/hydrocortisone) group. Patients who were lost to follow-up at 8 and 16 weeks are included in the ATI-indeterminate group. At 8 and 16 weeks after first infusion, the hydrocortisone group had fewer ATI-positive patients: 8 of 39 patients (21%) vs. 12 of 41 patients (29%), P 0.14 and 10 of 39 patients (26%) vs. 17 of 41 patients (42%), P 0.06, respectively. infusion predicted a 4-fold higher risk of serious infusion reaction and 2-fold higher risk of nonresponse. Noman et al. 14 recently demonstrated that ATI levels preinfusion are inversely related to the duration of clinical response and directly related to infusion reactions and reported a Table 4. Clinical Efficacy and Adverse Events in the Placebo/Hydrocortisone Groups Variable Placebo group (n 41) Hydrocortisone group (n 39) P value Clinical efficacy (%) Clinical response at 8 wk 19 (46) 22 (56) 0.5 Clinical response at 16 wk 18 (44) 21 (54) 0.5 Clinical remission at 8 wk 10 (24) 12 (31) 0.6 Clinical remission at 16 wk 10 (24) 11 (28) 0.8 Adverse events (%) No. of adverse events No. of patients with adverse events 19 (46) 16 (41) 0.5 No. of patients with early adverse events 14 (34) 11 (28) 0.5 No. of patients with late adverse events 5 (12) 5 (13) 0.9 No. of patients with infusion reactions 10 (24) 6 (15) 0.3 No. of patients with serious reactions 2 (5) 1 (2.5) 0.6

7 April 2003 PREMEDICATION REDUCES ANTIBODIES TO INFLIXIMAB 923 median preinfusion ATI level of 20.2 g/ml among patients who subsequently experienced an infusion reaction. Our multivariate analysis identified that a second infusion administered within 8 weeks of the first as well as concurrent azathioprine, 6-mercaptopurine, or methotrexate were significant independent protective factors in reducing ATI formation. In the ACCENT 1 trial, patients with Crohn s disease who received a 3-dose induction regimen at 0, 2, and 6 weeks followed by 5 or 10 mg/kg infusions every 8 weeks had significantly higher response and remission rates in addition to a lower incidence of ATI compared with those who received only a single infusion (6% 9% vs. 28%, respectively). 7,15 Furthermore, there was a lower incidence of ATI (and infusion reactions) among ACCENT 1 patients receiving both corticosteroids and immunomodulators (ATI rate 6%) and patients receiving immunomodulators alone (10%) compared with patients receiving neither corticosteroids nor immunomodulators (18%). 7 Concurrent therapy with low-dose methotrexate has also been shown to significantly reduce ATI formation in patients with rheumatoid arthritis receiving repeated infliximab infusions. 9 Pharmacokinetic studies suggest that concurrent methotrexate may improve efficacy of infliximab by delaying its clearance. 9 The true incidence of ATI formation in patients who receive infliximab remains unknown. Although the large ATTRACT trial 16 in rheumatoid arthritis and the AC- CENT 1 trial 7 in Crohn s disease reported an overall incidence of ATI formation in 8% 14% of patients, respectively, Noman et al. 17 reported ATI formation in 68% of a cohort of patients with Crohn s disease who received multiple infliximab infusions. Approximately 33% 35% of patients in our initial cohort study and subsequent trial developed ATI. The large number of patients with luminal disease with single infusions and the longer intervals between infusions as part of the on-demand schedules may account for higher incidence of ATI in both cohort studies. By contrast, all patients in the ATTRACT trial received concurrent methotrexate and were infused every 4 to 8 weeks, whereas most patients in the ACCENT 1 trial received 3 infliximab infusions initially at 0, 2, and 6 weeks and every 8 weeks thereafter. The timing of ATI/infliximab measurement at 8 weeks post-infliximab may also account for the lower ATI rates in the ATTRACT and ACCENT 1 trials, given that the majority of sera when assessed 8 weeks post-infusion are indeterminate for ATI. In our trial, approximately 20% of sera obtained 16 weeks postinfusion were indeterminate compared with over half the sera obtained 8 weeks post-infusion. Consequently, more patients were ATI-positive at 16 weeks compared with 8 weeks postinfusion. The randomized, controlled trial demonstrates that, whereas intravenous hydrocortisone premedication significantly reduces ATI levels, it does not significantly reduce ATI formation or infusion reactions. However, the association observed in study 1 between high ATI levels and both serious infusion reactions and loss of response suggests that intravenous hydrocortisone premedication may still benefit patients by reducing circulating ATI levels. Furthermore, the inability of the premedication schedule to completely abrogate ATI formation may lie in the timing of hydrocortisone administration immediately prior to infusion. Clinical trials have shown that a 2-dose premedication regimen of oral corticosteroids 12 and 2 hours before administration of intravenous contrast can provide significant protection against contrast reactions, whereas administering oral or intravenous corticosteroids 2 hours before contrast administration does not afford the same protection In study 1, we observed that concurrent prednisone of least 20 mg/day was also a significant protective factor in reducing ATI formation. This, combined with the significant reduction in ATI levels afforded by intravenous hydrocortisone premedication, suggests that an alternative strategy of premedicating with oral prednisone in the 24 hours pre-infliximab may be more effective and is worth investigating. However, many patients with Crohn s disease receive infliximab in an attempt to avoid corticosteroid side effects, and, although rare, negative effects of administrating short-term intravenous or oral corticosteroid premedication exist. 21 Currently, the optimum premedication regimen is currently unknown, and further controlled studies are required to evaluate dose and timing variables. The lack of a significant difference in terms of ATI formation, clinical efficacy, and infusion reactions between the treatment groups may also reflect both the relatively short duration of follow-up in the trial, which was primarily designed to evaluate ATI at 16 weeks postinfusion, and the number of patients lost to follow-up in both treatment groups. In conclusion, a substantial proportion of patients lose their initial response to infliximab, and both ATI formation and ATI level are strongly related to loss of initial response and infusion reactions. Our experience demonstrates that close administration of a second infusion close to the first and ensuring that patients are on adequate concurrent immunosuppressants prior to initiating infliximab therapy significantly reduces ATI. Although intravenous hydrocortisone premedication can signifi-

8 924 FARRELL ET AL. GASTROENTEROLOGY Vol. 124, No. 4 cantly reduce ATI levels, it does not eliminate ATI formation or infusion reactions. Whether a 3-dose induction regimen is superior to a 2-dose induction regimen or whether hydrocortisone premedication is superfluous for patients who receive the 3-dose induction regimen and/or are receiving concurrent immunosuppressant therapy is unknown. Although the optimal premedication strategy is currently unknown, and further controlled studies are warranted to evaluate dose, administration, and timing variables, our data suggest that intravenous hydrocortisone premedication should be considered in patients who have only recently commenced immunosuppressant therapy at the time that infliximab therapy is being initiated and in those patients intolerant to all immunosuppressants. References 1. Targan SR, Hanauer SB, van Deventer SJ, Mayer L, Present DH, Braakman T, DeWoody KL, Schaible TF, Rutgeerts PJ. A shortterm study of chimeric monoclonal antibody ca2 to tumor necrosis factor alpha for Crohn s disease. Crohn s Disease ca2 Study Group. N Engl J Med 1997;337: Present DH, Rutgeerts P, Targan S, Hanauer SB, Mayer L, van Hogezand RA, Podolsky DK, Sands BE, Braakman T, DeWoody KL, Schaible TF, van Deventer SJ. Infliximab for the treatment of fistulas in patients with Crohn s disease. N Engl J Med 1999; 340: Farrell RJ, Shah SA, Lodhavia PJ, Alsahli M, Falchuk KR, Michetti P, Peppercorn MA. Clinical experience with infliximab therapy in 100 patients with Crohn s disease. Am J Gastroenterol 2000;95: Cohen RD, Tsang JF, Hanauer SB. Infliximab in Crohn s disease: first anniversary clinical experience. Am J Gastroenterol 2000; 95: Ricart E, Panaccione R, Loftus EV, Tremaine WJ, Sandborn WJ. Infliximab for Crohn s disease in clinical practice at the Mayo Clinic: the first 100 patients. Am J Gastroenterol 2001;96: Sample C, Bailey RJ, Todoruk D, Sadowski D, Gramlich L, Milan M, Cherry R, Ma M, Lalor E, McKaigney J, Sherbaniuk R, Matic K, Switzer C, Fedorak RN. Clinical experience with infliximab for Crohn s disease: the first 100 patients in Edmonton, Alberta. Can J Gastroenterol 2002;16: Hanauer SB, Feagan BR, Lichtenstein GR, Mayer LF, Schreiber S, Colombel JF, Rachmilewitz D, Wolf DC, Olson A, Bao W, Rutgeerts P. Maintenance infliximab for Crohn s disease: the ACCENT 1 randomized trial. Lancet 2002;359: Remicade (infliximab). Centocor Inc., Maini RN, Breedveld FC, Kalden JR, Smolen JS, Davis D, Macfarlane JD, Antoni C, Leeb B, Elliott MJ, Woody JN, Schaible TF, Feldmann M. Therapeutic efficacy of multiple intravenous infusions of anti-tumor necrosis factor a monoclonal antibody combined with low-dose weekly methotrexate in rheumatoid arthritis. Arthritis Rheum 1998;41: Farrell RJ, Alsahli M, Falchuk KR, Peppercorn MA, Michetti P. A randomized, double-blind, placebo-controlled trial of intravenous hydrocortisone in reducing human anti-chimeric antibody following infliximab therapy (abstr). Gastroenterology 2001;120:A Breedveld FC. Therapeutic monoclonal antibodies. Lancet 2000; 355: Sandborn WJ, Feagan BG, Hanauer SB, Present DH, Sutherland LR, Kamm MA, Wolf DC, Baker JP, Hawkey C, Archambault A, Bernstein CN, Novak C, Heath PK, Targan SR. An engineered human antibody to TNF. An engineered human antibody to TNF (CDP571) for active Crohn s disease: a randomized double-blind placebo-controlled trial. Gastroenterology 2001;120: Schreiber S, Campieri M, Colombel JF, van Deventer SJ, Feagan B, Fedorak R, Forbes A, Gassull M, Gendre JP, van Hogezand RA, Lofberg R, Modigliani R, Pallone F, Petritsch W, Prantera C, Rampton D, Seibold F, Vatn M, Zeitz M, Rutgeerts P. Use of anti-tumour necrosis factor agents in inflammatory bowel disease. European guidelines for Int J Colorectal Dis 2001;16: Noman M, Baert F, Vermeire S, Van Assche G, D Haens G, Carbonez A, Rutgeerts P. Post infusion infliximab levels determine duration of response in Crohn s disease and are directly related to infusion reactions. Gastroenterology 2002;(suppl): A Mayer L, Han C, Bala M. Three dose induction regimen of infliximab is superior to a single dose in patients with Crohn s disease. Am J Gastroenterol 2001;96:S Lipsky PE, van der Heijde DM, St. Clair EW, Furst DE, Breedveld FC, Kalden JR, Smolen JS, Weisman M, Emery P, Feldmann M, Harriman GR, Maini RN. Infliximab and Infliximab and methotrexate in the treatment of rheumatoid arthritis. Anti-Tumor Necrosis Factor Trial in Rheumatoid Arthritis with Concomitant Therapy Study Group. N Engl J Med 2000;343: Noman M, Baert F, D Haens G, Van Assche G, Swijsen C, Aerden A, Rutgeerts P. ATI formation after infliximab treatment in Crohn s disease is clearly associated with infusion reactions (abstr). Gastroenterology 2001;120:A Lasser EC, Berry CC, Talner LB, Santini LC, Lang EK, Gerber FH, Stolberg HO. Pretreatment with corticosteroids to alleviate reactions to intravenous contrast material. N Engl J Med 1987;317: Talbot S, Atkins PC, Zweiman B. In vivo effects of corticosteroids on human allergic responses. I. Effects of systemic administrations of steroids. Ann Allergy 1987;58: Ellul-Micallef R, Fenech FF. Intravenous prednisolone in chronic bronchial asthma. Thorax 1975;30: Alexiou C, Kau RJ, Luppa P, Arnold W. Allergic reactions after systemic administration of glucocorticosteroid therapy. Arch Otolaryngol Head Neck Surg 1998;124: Received July 9, Accepted January 9, Address requests for reprints to: Richard J. Farrell, M.D., Department of Gastroenterology, Dana 501, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Harvard Medical School, Boston, Massachusetts rfarrell@caregroup.harvard.edu; fax: (617) Drs. Farrell and Alsahli contributed equally to this work.