As noted in a recent post (Assessing and discussing risk of lymphoma in IBD), diagrams can be useful to convey the absolute risks for IBD medicines; this risk is often poorly understood by patients and their families. Two articles add additional insight that may help with counseling. (Thanks to Ben Gold for forwarding these articles.)
- Am J Gastroenterol 2012; 107: 964-70.
- Am J Gastroenterol 2012; 107: 1051-63.
The first article reviews the risk of lymphoma with regard to inflammatory bowel disease treatment decisions. Important points in this article include the following:
1. Relative risks may appear large while absolute risk may be quite low. The estimate for absolute risk for lymphoma:
- For azathioprine or 6-mercaptopurine: 1 additional case for every 4357 persons treated
- For anti-TNF (infliximab, adalimumab, certolizumab): 1 additional case for every 2380 persons treated between ages 20-29.
2. Framing the discussion with regards to risk of continued active disease, continued exposure to steroids, and potential need for surgery can be helpful:
- “It would take only three patients discontinuing their azathioprine to cause one additional relapse of IBD per year”
- Seven patients stopping anti-TNF therapy would result in an additional hospitalization each year and 14 patients stopping anti-TNF therapy would result in an additional abdominal surgery each year
- While the ‘relative risks of lymphoma associated with these medications may sound inappropriately large…A more appropriate comparison is the absolute risk of lymphoma versus the absolute risk of active/untreated disease or corticosteroid therapy.’ An example: if 2000 patients with IBD took one of these medications (eg. anti-TNF agent), probably one patient will develop a lymphoma; however, if none of those patients took an anti-TNF agent, it would result in more than 100 hospitalizations and nearly 60 surgeries.
3. “Patients can also be very sensitive to numerators and pay relatively insufficient attention to denominators.” This has been called “anchoring and adjustment” or “base-rate neglect.” This issue has to do with “numeracy;” this is defined as the basic math skill needed for health-related activities. To help families, consider the following:
- avoid labels such as “very low” when describing risk
- use absolute risk data
- use similar denominators when comparing risks
- use visual aids
The second study cited is a pooled analysis of infections, malignancy, and mortality risks associated with infliximab and immunomodulator treatment in adult IBD patients. This study collected safety data from 10 previous trials. Five of these trials were randomized, controlled studies. Table 3 and Table 4 detail extensive safety data for immunomodulators (eg. azathioprine) and for infliximab respectively.
With regard to immunomodulators, the combined studies enrolled 947 on immunomodulators in comparison to 1170 without immunomodulators. With ulcerative colitis (UC) patients but not with Crohn’s disease (CD), immunomodulator use increased the risk of infections (120/100 patient years versus 92.5 among placebo-treated patients). CD patients, but not UC patients, had an increased risk of malignancy with immunomodulator use (1.84/100 patient years compared with 0/100 patient years in the control group). With the exception of the SONIC study, use of immunomodulator was not randomly assigned. So, some of the increased risk in these patients could be due to having more severe IBD rather than due to the medication.
With regard to infliximab, and in contradiction to the previous article’s estimated risks, infliximab treatment did not appear to affect the incidence of infection, mortality, or malignancy. This study and several others have not demonstrated an increased risk of malignancy with infliximab. This could be that even with this pooled data it is difficult to detect a rare adverse outcome. More prospective studies and more long-term followup will be needed to truly determine the risk.
While it is known that these agents may increase the risk of some infections, the limited increase in infections which was detected only with immunomodulator use in UC is likely due to a lowered risk of infection when active inflammatory bowel disease is controlled. A much bigger risk factor for infection is the use of corticosteroids. When effective IBD medications are administered, this helps control inflammation and allows tapering of corticosteroids.