Drug Therapies and the Risk of Malignancy in Crohn's Disease
Drug Therapies and the Risk of Malignancy in Crohn's Disease
Data collected from 6,273 patients followed in the TREAT Registry were available for analysis. Data summarized herein were collected from registry initiation (July 1999) through 23 February 2010. Patient enrollment was complete and closed as of 31 March 2004. The average duration of follow-up was 5.2 years. As of 23 February 2010, 2,710 (43.2%) of enrolled patients were actively participating in the registry and had been followed for a mean of 7.6 years. The remaining 3,563 patients who discontinued from the registry were followed for a mean of 3.4 years. The reasons for discontinuation and distribution of patients by length of follow-up are detailed in Figure 1 and Supplementary Figure S1 online, respectively. Data from both active and discontinued patients are included in this report.
Table 1 summarizes the baseline characteristics of patients. The vast majority (90.8 %) of patients were white; 58.6 % were female, and 41.4 % were male. Th e average age was 43 years.
Slightly more than half of all patients were treated with infliximab during registry participation (n=3,420 or 54.5%), and 3,764 patients received infliximab during or within 1 year before registry participation. Approximately 90% (2,986/3,420) of patients receiving infliximab during the registry had received at least two infusions. An infliximab dose of 5 mg/kg was given in 81.5% of the more than 53,000 infliximab infusions administered.
Infliximab-treated patients were younger (P<0.0001) and also differed from patients receiving other-treatments-only in several other characteristics. Specifically, infliximab-treated patients were more likely to have received treatment with prednisone, immunosuppressants, antibiotics, narcotic analgesics, and antidepressants within the year before registry entry; to have ileal and/or colonic disease; and to have used a biological agent other than infliximab. They also had greater disease severity and were more likely to have required hospital admission for medical or surgical treatment in the year before enrollment (P<0.001 for all comparisons; Table 1 ). The distributions of length of infliximab use and immunosuppressant use are shown in Supplementary Figure S1 online.
The proportions of patients who developed any neoplasm (benign or malignant) were similar between infliximab-treated patients (0.78/100 pt-yrs) and those who received other-treatments-only (0.85/100 pt-yrs), yielding a relative risk (95% CI) of 0.90 (0.69, 1.18; P=0.46) ( Table 2). Similar findings were observed for "any malignancy" and also within malignancy categories. Specifically, lymphoma, NMSC, and solid tumors occurred at rates of 0.05, 0.16, and 0.42, respectively, events/100 pt-yrs in infliximab-treated patients and 0.05, 0.18, and 0.45, respectively, events/100 pt-yrs in patients who received other-treatments-only (P value not significant (NS) for all comparisons). When assessed by immunosuppressant (azathioprine, 6-MP, and/or MTX) use before event onset, no significant difference in these malignancy categories was observed between patients with and without exposure ( Table 2).
Further details of malignancies reported, also based on 100 pt-yrs of follow-up, are provided in Table 3. The infliximab and other-treatments-only cohorts were generally similar with regard to the occurrence of individual types of malignancies. Malignancies that displayed more variation between the treatment cohorts included squamous cell carcinoma (0.05/100 pt-yrs for infliximab vs. 0.08/100 pt-yrs for other-treatments-only), malignancy of the large intestine (0.07 vs. 0.02/100 pt-yrs), lung cancer (0.05 vs. 0.10/100 pt-yrs), and renal cancer (0.01 vs. 0.04/100 pt-yrs; Table 3). For all malignancies shown in Table 3, the incidences were similar between all patients and only those patients who used no biological agent other than infliximab ( Table 3) and between patients who did and did not have immunosuppressant use documented before the event (data not shown).
As discussed above and shown in Table 1, patients in the infliximab-treated cohort had more severe CD and were more commonly treated with several concomitant medications, including corticosteroids and immunosuppressants, at the time of registry enrollment. To adjust for these imbalances and to better assess the relative role of infliximab in malignancies, multivariate analyses were performed using a Cox proportional hazards regression model. Results of these analyses indicated that baseline age (HR=1.59 per 10 years; P<0.001), disease duration (HR=1.64 per 10 years; P=0.012), and smoking (HR=1.38; P=0.045) but neither immunosuppressive therapy alone (HR=1.43; P=0.11), infliximab therapy alone (HR=0.59; P=0.16), nor their combination (HR=1.22, P=0.34), were independently associated with the risk of malignancy ( Table 4).
The occurrence of malignancy did not appear to be increased as the number of infliximab infusions increased (odds ratios (ORs) ranging from 0.41 to 1.06); the only significant difference from the reference was a lower occurrence of malignancy among patients who received 1–4 infliximab infusions (P<0.001). Numerical increases in the occurrence of malignancy were observed for patients who received immunosuppressants during 1–3 (OR=1.82; P=NS), 4–9 (OR=2.83; P=0.016), and >9 (OR=1.63; P=NS) of the 6-month registry periods, as well as for patients with only historical immunosuppressant use (OR=2.3; P=0.049). In this exposure-based analysis, use of immunosuppressants alone (OR=4.19) or in combination with infliximab (OR=3.33) seemed to be associated with a greater risk of malignancy than did treatment with infliximab alone (OR=1.96) relative to treatment with neither, although in no medication group was the numerical difference from the reference statistically significant ( Table 5).
Results of comparisons made between the infliximab and other-treatments-only groups and the SEER database yielded 95% CIs containing 1 for all categories of malignancy assessed, indicating no significant difference, with the exception of breast cancer and prostate cancer ( Table 6). Breast cancer was less common in the TREAT than SEER populations, regardless of treatment group, with a SIR (95% CI) of 0.50 (0.24, 0.92) for infliximab-treated patients and 0.32 (0.12, 0.70) for patients receiving other-treatments-only. Similarly, prostate cancer was less common in the TREAT patients receiving other-treatments-only (0.30 (0.08, 0.76)) than that in the SEER population; this was not the case for infliximab-treated patients ( Table 6).
Comparisons were also made between the TREAT subgroups of patients with and without immunosuppressant use and the SEER database. The results yielded 95% CIs containing 1 for all categories of malignancy assessed, indicating no significant difference, with the exception of breast, bladder, and prostate cancers ( Table 6). Breast cancer was less common in the TREAT than SEER populations, regardless of immunosuppressant use, with SIRs (95% CI) of 0.50 (0.26, 0.87) for patients with immunosuppressant use and 0.28 (0.08, 0.72) for patients with no immunosuppressant use. Similarly, bladder cancer was less common in the TREAT patients with no immunosuppressant use (0.00 (0.00, 0.97)) than that in the SEER population; this was not the case in patients with immunosuppressant use. A similar pattern was observed for prostate cancer ( Table 6). For lymphoma, the 95% CIs surrounding the SIR all contained 1, indicating no significant differences from the general population ( Table 6).
Results
Patient Disposition, Characteristics, and Treatment
Data collected from 6,273 patients followed in the TREAT Registry were available for analysis. Data summarized herein were collected from registry initiation (July 1999) through 23 February 2010. Patient enrollment was complete and closed as of 31 March 2004. The average duration of follow-up was 5.2 years. As of 23 February 2010, 2,710 (43.2%) of enrolled patients were actively participating in the registry and had been followed for a mean of 7.6 years. The remaining 3,563 patients who discontinued from the registry were followed for a mean of 3.4 years. The reasons for discontinuation and distribution of patients by length of follow-up are detailed in Figure 1 and Supplementary Figure S1 online, respectively. Data from both active and discontinued patients are included in this report.
Table 1 summarizes the baseline characteristics of patients. The vast majority (90.8 %) of patients were white; 58.6 % were female, and 41.4 % were male. Th e average age was 43 years.
Slightly more than half of all patients were treated with infliximab during registry participation (n=3,420 or 54.5%), and 3,764 patients received infliximab during or within 1 year before registry participation. Approximately 90% (2,986/3,420) of patients receiving infliximab during the registry had received at least two infusions. An infliximab dose of 5 mg/kg was given in 81.5% of the more than 53,000 infliximab infusions administered.
Infliximab-treated patients were younger (P<0.0001) and also differed from patients receiving other-treatments-only in several other characteristics. Specifically, infliximab-treated patients were more likely to have received treatment with prednisone, immunosuppressants, antibiotics, narcotic analgesics, and antidepressants within the year before registry entry; to have ileal and/or colonic disease; and to have used a biological agent other than infliximab. They also had greater disease severity and were more likely to have required hospital admission for medical or surgical treatment in the year before enrollment (P<0.001 for all comparisons; Table 1 ). The distributions of length of infliximab use and immunosuppressant use are shown in Supplementary Figure S1 online.
Incidence of Cancer
The proportions of patients who developed any neoplasm (benign or malignant) were similar between infliximab-treated patients (0.78/100 pt-yrs) and those who received other-treatments-only (0.85/100 pt-yrs), yielding a relative risk (95% CI) of 0.90 (0.69, 1.18; P=0.46) ( Table 2). Similar findings were observed for "any malignancy" and also within malignancy categories. Specifically, lymphoma, NMSC, and solid tumors occurred at rates of 0.05, 0.16, and 0.42, respectively, events/100 pt-yrs in infliximab-treated patients and 0.05, 0.18, and 0.45, respectively, events/100 pt-yrs in patients who received other-treatments-only (P value not significant (NS) for all comparisons). When assessed by immunosuppressant (azathioprine, 6-MP, and/or MTX) use before event onset, no significant difference in these malignancy categories was observed between patients with and without exposure ( Table 2).
Further details of malignancies reported, also based on 100 pt-yrs of follow-up, are provided in Table 3. The infliximab and other-treatments-only cohorts were generally similar with regard to the occurrence of individual types of malignancies. Malignancies that displayed more variation between the treatment cohorts included squamous cell carcinoma (0.05/100 pt-yrs for infliximab vs. 0.08/100 pt-yrs for other-treatments-only), malignancy of the large intestine (0.07 vs. 0.02/100 pt-yrs), lung cancer (0.05 vs. 0.10/100 pt-yrs), and renal cancer (0.01 vs. 0.04/100 pt-yrs; Table 3). For all malignancies shown in Table 3, the incidences were similar between all patients and only those patients who used no biological agent other than infliximab ( Table 3) and between patients who did and did not have immunosuppressant use documented before the event (data not shown).
As discussed above and shown in Table 1, patients in the infliximab-treated cohort had more severe CD and were more commonly treated with several concomitant medications, including corticosteroids and immunosuppressants, at the time of registry enrollment. To adjust for these imbalances and to better assess the relative role of infliximab in malignancies, multivariate analyses were performed using a Cox proportional hazards regression model. Results of these analyses indicated that baseline age (HR=1.59 per 10 years; P<0.001), disease duration (HR=1.64 per 10 years; P=0.012), and smoking (HR=1.38; P=0.045) but neither immunosuppressive therapy alone (HR=1.43; P=0.11), infliximab therapy alone (HR=0.59; P=0.16), nor their combination (HR=1.22, P=0.34), were independently associated with the risk of malignancy ( Table 4).
The occurrence of malignancy did not appear to be increased as the number of infliximab infusions increased (odds ratios (ORs) ranging from 0.41 to 1.06); the only significant difference from the reference was a lower occurrence of malignancy among patients who received 1–4 infliximab infusions (P<0.001). Numerical increases in the occurrence of malignancy were observed for patients who received immunosuppressants during 1–3 (OR=1.82; P=NS), 4–9 (OR=2.83; P=0.016), and >9 (OR=1.63; P=NS) of the 6-month registry periods, as well as for patients with only historical immunosuppressant use (OR=2.3; P=0.049). In this exposure-based analysis, use of immunosuppressants alone (OR=4.19) or in combination with infliximab (OR=3.33) seemed to be associated with a greater risk of malignancy than did treatment with infliximab alone (OR=1.96) relative to treatment with neither, although in no medication group was the numerical difference from the reference statistically significant ( Table 5).
Results of comparisons made between the infliximab and other-treatments-only groups and the SEER database yielded 95% CIs containing 1 for all categories of malignancy assessed, indicating no significant difference, with the exception of breast cancer and prostate cancer ( Table 6). Breast cancer was less common in the TREAT than SEER populations, regardless of treatment group, with a SIR (95% CI) of 0.50 (0.24, 0.92) for infliximab-treated patients and 0.32 (0.12, 0.70) for patients receiving other-treatments-only. Similarly, prostate cancer was less common in the TREAT patients receiving other-treatments-only (0.30 (0.08, 0.76)) than that in the SEER population; this was not the case for infliximab-treated patients ( Table 6).
Comparisons were also made between the TREAT subgroups of patients with and without immunosuppressant use and the SEER database. The results yielded 95% CIs containing 1 for all categories of malignancy assessed, indicating no significant difference, with the exception of breast, bladder, and prostate cancers ( Table 6). Breast cancer was less common in the TREAT than SEER populations, regardless of immunosuppressant use, with SIRs (95% CI) of 0.50 (0.26, 0.87) for patients with immunosuppressant use and 0.28 (0.08, 0.72) for patients with no immunosuppressant use. Similarly, bladder cancer was less common in the TREAT patients with no immunosuppressant use (0.00 (0.00, 0.97)) than that in the SEER population; this was not the case in patients with immunosuppressant use. A similar pattern was observed for prostate cancer ( Table 6). For lymphoma, the 95% CIs surrounding the SIR all contained 1, indicating no significant differences from the general population ( Table 6).
Source...