Safety of Infliximab in Inflammatory Bowel Disease Needs to Be Debated

With great interest we read the article by Caspersen et al1 in which they investigated the long-term safety of infliximab in patients with inflammatory bowel disease (IBD) in a national Danish population-based cohort. They scored adverse events in up to 47% of all patients and concluded that infliximab was generally well tolerated. Their conclusion based on these numbers is debatable, and is also based on a retrospective study, with subsequent nonstandardized underreporting of adverse events as compared with prospective clinical trials. Moreover, the data of the Danish IBD cohort was gathered over a relatively short follow-up period.

In Table 1, we compare their study with 3 other recently published retrospective cohort studies about the long-term safety of infliximab in patients with IBD from Belgium, Scotland, and The Netherlands.2, 3, 4 The Denmark IBD population treated with infliximab developed fewer malignancies (4; 0.6%) compared with the cohorts from Belgium (21; 2.8%), Scotland (6; 3.0%), and The Netherlands (9; 6.1%). Most likely this could be explained by a relatively short follow-up period (median follow-up period, 29.1 mo) compared with the cohorts from Belgium and The Netherlands (median follow-up periods, 58 and 59 mo, respectively), and therefore underestimation of the true rate of malignancies. The raised concern of developing malignancies in patients treated with anti–tumor necrosis factor, in combination with the fact that carcinogenesis is a process requiring many years of exposure to disease or treatment, makes it necessary to have a long-term follow-up period to detect malignancies. This is emphasized further with the significant association with duration of disease and the development of malignancies in the studies from Belgium and The Netherlands. This indicates that thorough follow-up evaluation of patients with IBD for the development of malignancies is warranted.

Table 1. Characteristics of the Four Recently Reported Studies on the Long-Term Safety of Infliximab

		Caspersen et al1	Fidder et al2	Lees et al3	de Vries et al4
	Number of treated patients	651	734	202	147
	Median age at start of therapy, y	31.6(range,9.7–92.4)	35(IQR,25–35)	30.8(IQR,22.0–42.9)	38(SD,12)
	Median disease duration before first infusion, y	5(range,0–38)	7(IQR,2–15)	4.9(IQR,1.2–11.0)	16(range,0–41)
	Number of infusions	3351	7276	718	1941
	Median number of infusions per patient	3	6	3 for CD, 1 for UC	10
	Follow-up period (patient-years)	—	3775	620	674
	Median follow-up period, mo	29.1	58	28.8	59
	Number of infections, %a	69(63patients)	59(48patients)	60(42patients)	57(36patients)
	Annual infection rate, %	—	1.3	7	5
	Number of malignancies, %	4(0.6)	21(2.8)	6(3.0)	9b(6.1)
	Mortality, %	13(2.0)	12(1.6)	7(3.5)	8(5.4)
	Annual mortality rate, %	—	0.3	1.1	1.2

IQR, interquartile range; CD, Crohn's disease; UC, ulcerative colitis.

a In some of the studies (de Vries et al), only those patients who underwent hospitalization for their infections were included. In others, hospitalization was not required in case of an infection. Therefore, the given numbers are not completely comparable since inclusion criteria are different between these studies. b There were 9 malignancies in 8 patients.

The question remains whether the high rate of malignancies is related to disease, treatment, or a synergistic effect. The Belgian cohort did include a control group of IBD patients treated in the same center and period. They compared IBD patients treated with infliximab with patients not receiving any biological therapy and found no differences in terms of mortality, infection rates, and malignancies.2 These findings need to be reproduced in other centers, but give us some first impressions regarding causality with IBD.